Progress in Brain Research Volume 163 The Dentate Gyrus: A Comprehensive Guide to Structure, Function, and Clinical Implications

PROGRESS IN BRAIN RESEARCH VOLUME 163 THE DENTATE GYRUS: A COMPREHENSIVE GUIDE TO STRUCTURE, FUNCTION, AND CLINICAL IMPLICATIONS

Other volumes in PROGRESS IN BRAIN RESEARCH Volume 127: Neural Transplantation II. Novel Cell Therapies for CNS Disorders, by S.B. Dunnett and A. Bjo¨rklund (Eds.) – 2000, ISBN 0-444-50109-6. Volume 128: Neural Plasticity and Regeneration, by F.J. Seil (Ed.) – 2000, ISBN 0-444-50209-2. Volume 129: Nervous System Plasticity and Chronic Pain, by J. Sandku¨hler, B. Bromm and G.F. Gebhart (Eds.) – 2000, ISBN 0-44450509-1. Volume 130: Advances in Neural Population Coding, by M.A.L. Nicolelis (Ed.) – 2001, ISBN 0-444-50110-X. Volume 131: Concepts and Challenges in Retinal Biology, by H. Kolb, H. Ripps and S. Wu (Eds.) – 2001, ISBN 0-444-50677-2. Volume 132: Glial Cell Function, by B. Castellano Lo´pez and M. Nieto-Sampedro (Eds.) – 2001, ISBN 0-444-50508-3. Volume 133: The Maternal Brain. Neurobiological and Neuroendocrine Adaptation and Disorders in Pregnancy and Post Partum, by J.A. Russell, A.J. Douglas, R.J. Windle and C.D. Ingram (Eds.) – 2001, ISBN 0-444-50548-2. Volume 134: Vision: From Neurons to Cognition, by C. Casanova and M. Ptito (Eds.) – 2001, ISBN 0-444-50586-5. Volume 135: Do Seizures Damage the Brain, by A. Pitka¨nen and T. Sutula (Eds.) – 2002, ISBN 0-444-50814-7. Volume 136: Changing Views of Cajal’s Neuron, by E.C. Azmitia, J. DeFelipe, E.G. Jones, P. Rakic and C.E. Ribak (Eds.) – 2002, ISBN 0-444-50815-5. Volume 137: Spinal Cord Trauma: Regeneration, Neural Repair and Functional Recovery, by L. McKerracher, G. Doucet and S. Rossignol (Eds.) – 2002, ISBN 0-444-50817-1. Volume 138: Plasticity in the Adult Brain: From Genes to Neurotherapy, by M.A. Hofman, G.J. Boer, A.J.G.D. Holtmaat, E.J.W. Van Someren, J. Verhaagen and D.F. Swaab (Eds.) – 2002, ISBN 0-444-50981-X. Volume 139: Vasopressin and Oxytocin: From Genes to Clinical Applications, by D. Poulain, S. Oliet and D. Theodosis (Eds.) – 2002, ISBN 0-444-50982-8. Volume 140: The Brain’s Eye, by J. Hyo¨na¨, D.P. Munoz, W. Heide and R. Radach (Eds.) – 2002, ISBN 0-444-51097-4. Volume 141: Gonadotropin-Releasing Hormone: Molecules and Receptors, by I.S. Parhar (Ed.) – 2002, ISBN 0-444-50979-8. Volume 142: Neural Control of Space Coding, and Action Production, by C. Prablanc, D. Pe´lisson and Y. Rossetti (Eds.) – 2003, ISBN 0-444-509771. Volume 143: Brain Mechanisms for the Integration of Posture and Movement, by S. Mori, D.G. Stuart and M. Wiesendanger (Eds.) – 2004, ISBN 0-444-513892. Volume 144: The Roots of Visual Awareness, by C.A. Heywood, A.D. Milner and C. Blakemore (Eds.) – 2004, ISBN 0-444-50978-X. Volume 145: Acetylcholine in the Cerebral Cortex, by L. Descarries, K. Krnjevic´ and M. Steriade (Eds.) – 2004, ISBN 0-444-51125-3. Volume 146: NGF and Related Molecules in Health and Disease, by L. Aloe and L. Calza` (Eds.) – 2004, ISBN 0-444-51472-4. Volume 147: Development, Dynamics and Pathology of Neuronal Networks: From Molecules to Functional Circuits, by J. Van Pelt, M. Kamermans, C.N. Levelt, A. Van Ooyen, G.J.A. Ramakers and P.R. Roelfsema (Eds.) – 2005, ISBN 0-444-51663-8. Volume 148: Creating Coordination in the Cerebellum, by C.I. De Zeeuw and F. Cicirata (Eds.) – 2005, ISBN 0-444-51754-5. Volume 149: Cortical Function: A View from the Thalamus, by V.A. Casagrande, R.W. Guillery and S.M. Sherman (Eds.) – 2005, ISBN 0-444-51679-4. Volume 150: The Boundaries of Consciousness: Neurobiology and Neuropathology, by Steven Laureys (Ed.) – 2005, ISBN 0-444-51851-7. Volume 151: Neuroanatomy of the Oculomotor System, by J.A. Bu¨ttner-Ennever (Ed.) – 2006, ISBN 0-444-51696-4. Volume 152: Autonomic Dysfunction after Spinal Cord Injury, by L.C. Weaver and C. Polosa (Eds.) – 2006, ISBN 0-444-51925-4. Volume 153: Hypothalamic Integration of Energy Metabolism, by A. Kalsbeek, E. Fliers, M.A. Hofman, D.F. Swaab, E.J.W. Van Someren and R. M. Buijs (Eds.) – 2006, ISBN 978-0-444-52261-0. Volume 154: Visual Perception, Part 1, Fundamentals of Vision: Low and Mid-Level Processes in Perception, by S. Martinez-Conde, S.L. Macknik, L.M. Martinez, J.M. Alonso and P.U. Tse (Eds.) – 2006, ISBN 978-0-444-52966-4. Volume 155: Visual Perception, Part 2, Fundamentals of Awareness, Multi-Sensory Integration and High-Order Perception, by S. Martinez-Conde, S.L. Macknik, L.M. Martinez, J.M. Alonso and P.U. Tse (Eds.) – 2006, ISBN 978-0-444-51927-6. Volume 156: Understanding Emotions, by S. Anders, G. Ende, M. Junghofer, J. Kissler and D. Wildgruber (Eds.) – 2006, ISBN 978-0444-52182-8. Volume 157: Reprogramming of the Brain, by A.R. Møller (Ed.) – 2006, ISBN 978-0-444-51602-2. Volume 158: Functional Genomics and Proteomics in the Clinical Neurosciences, by S.E. Hemby and S. Bahn (Eds.) – 2006, ISBN 9780-444-51853-8. Volume 159: Event-Related Dynamics of Brain Oscillations, by C. Neuper and W. Klimesch (Eds.) – 2006, ISBN 978-0-444-52183-5. Volume 160: GABA and the Basal Ganglia: From Molecules to Systems, by J.M. Tepper, E.D. Abercrombie and J.P. Bolam (Eds.) – 2007, ISBN 978-0-444-52184-2. Volume 161: Neurotrauma: New Insights into Pathology and Treatment, by J.T. Weber and A.I.R. Maas (Eds.) – 2007, ISBN 978-0444-53017-2. Volume 162: Neurobiology of Hyperthermia, by H.S. Sharma (Ed.) – 2007, ISBN 978-0-444-519269.

PROGRESS IN BRAIN RESEARCH

VOLUME 163

THE DENTATE GYRUS: A COMPREHENSIVE GUIDE TO STRUCTURE, FUNCTION, AND CLINICAL IMPLICATIONS EDITED BY HELEN E. SCHARFMAN Associate Professor of Clinical Pharmacology (in Neurology) Columbia University, College of Physicians and Surgeons and Center for Neural Recovery and Rehabilitation Research, Helen Hayes Hospital, New York State Department of Health, West Haverstraw, NY 10993-1195, USA

AMSTERDAM – BOSTON – HEIDELBERG – LONDON – NEW YORK – OXFORD PARIS – SAN DIEGO – SAN FRANCISCO – SINGAPORE – SYDNEY – TOKYO

Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands Linacre House, Jordan Hill, Oxford OX2 8DP, UK

First edition 2007 Copyright r 2007 Elsevier B.V. All rights reserved No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; e-mail: [email protected]. Alternatively you can submit your request online by visiting the Elsevier web site at http://www.elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

ISBN: 978-0-444-53015-8 (this volume) ISSN: 0079-6123 (Series)

For information on all Elsevier publications visit our website at books.elsevier.com

Printed and bound in The Netherlands 07 08 09 10 11

10 9 8 7 6 5 4 3 2 1

List of contributors D.G. Amaral, Department of Psychiatry and Behavioral Sciences, The M.I.N.D. Institute and the California National Primate Research Center, UC Davis, 2825 50th Street, Sacramento, CA 95817, USA L. Acsa´dy, Institute of Experimental Medicine, Hungarian Academy of Sciences, PO Box 67, 1450 Budapest, Hungary C.A. Barnes, Arizona Research Laboratories Division of Neural Systems, Memory & Aging, University of Arizona, Tucson, AZ and Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, AZ and Departments of Psychology and Neurology, University of Arizona, Tucson, AZ, USA I. Bechmann, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany D.K. Binder, Department of Neurological Surgery, University of California, Irvine, CA 92868-3298, USA M. Blaabjerg, Anatomy and Neurobiology, Institute of Medical Biology, University of Southern Denmark, Winslowparken 21, DK-5000 Odense C, Denmark C.R. Bramham, Department of Biomedicine and Mental Health Research Center, University of Bergen, Jonas Lies vei 91, N-5009 Bergen, Norway G. Buzsa´ki, Center for Molecular and Behavioral Neuroscience, Rutgers, The State University of New Jersey, 197 University Avenue, University Heights, Newark, NJ 07102, USA G.C. Carlson, The Children’s Hospital of Philadelphia, Abramson Pediatric Research Center, Room 410D, 3516 Civic Center Boulevard, Philadelphia, PA 19104-4318, USA C. Chavkin, Department of Pharmacology, University of Washington, Seattle, WA 98195, USA M.K. Chawla, Arizona Research Laboratories Division of Neural Systems, Memory & Aging, University of Arizona, Tucson, AZ and Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, AZ, USA B.J. Claiborne, Department of Biology, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249, USA W.F. Colmers, Department of Pharmacology, University of Alberta, 9-36 Medical Sciences Building, Edmonton, AL T6G 2H7, Canada D.A. Coulter, The Children’s Hospital of Philadelphia, Abramson Pediatric Research Center, Room 410D, 3516 Civic Center Boulevard, Philadelphia, PA 19104-4318, USA D. Del Turco, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany T. Deller, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany B.E. Derrick, Department of Biology, The Cajal Neuroscience Research Institute, The University of Texas at San Antonio, 6900 N. Loop 1604 West, San Antonio, TX 78249-0662, USA L. DeToledo-Morrell, Department of Neurological Sciences, Rush University Medical Center, Chicago, IL 60612, USA C.T. Drake, Division of Neurobiology, Department of Neurology and Neuroscience, Weill-Cornell Medical College, 411 East 69th Street, New York, NY 10021, USA

v

vi

M.R. Drew, Department of Neuroscience and Psychiatry, Division of Integrative Neuroscience, Columbia University, New York, NY 10032, USA F.E. Dudek, Department of Physiology, University of Utah, Salt Lake City, UT, USA E. Fo¨rster, Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104 Freiburg, Germany C.J. Frazier, Department of Pharmacodynamics and Department of Neuroscience, University of Florida, College of Medicine, JHMHC 100487, 1600 S.W. Arder Road, Gainesville, FL 32610, USA M. Frotscher, Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104 Freiburg, Germany R. Gutie´rrez, Department of Physiology, Biophysics and Neurosciences, Center for Research and Advanced Studies, Mexico City Apartado Postal 14-740, Mexico D.F. 07000, Mexico T. Hajszan, Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University School of Medicine, 333 Cedar Street, FMB 313, New Haven, CT 06520, USA T. Hamilton, Department of Pharmacology, University of Alberta, 9-36 Medical Sciences Building, Edmonton, AL T6G 2H7, Canada C.W. Harley, Department of Psychology, Memorial University of Newfoundland, St. John’s, NL A1B 3X9, Canada R. Hen, Departments of Neuroscience, Psychiatry, and Pharmocology, Division of Integrative Neuroscience, Columbia University, New York, NY 10032, USA D.A. Henze, Neuroscience Drug Discovery Merck Research Laboratories, 770 Sunney town Pike, P.O.B. 4 WP 26A-2000, West Point, PA 19486, USA C.R. Houser, Department of Neurobiology, David Geffen School of Medicine at UCLA, 73-235 CHS, 10833 Le-Conte Avenue, Los Angeles, CA 90095, USA D. Hsu, Department of Neurology, University of Wisconsin, 600 Highland Avenue, H6/526, Madison, WI 53792, USA D.B. Jaffe, Department of Biology, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249, USA M. Joe¨ls, Swammerdam Institute of Life Sciences, Center for NeuroScience (SILS-CNS), University of Amsterdam, Kruislaan 320, 1098 SM Amsterdam, The Netherlands S. Ka´li, Institute of Experimental Medicine, Hungarian Academy of Sciences, PO Box 67, 1450 Budapest, Hungary R.P. Kesner, Department of Psychology, University of Utah, 380 S. 1530 E., Room 502, Salt Lake City, UT 84121, USA P. Lavenex, Department of Medicine, Unit of Physiology, University of Fribourg, 1700 Fribourg, Switzerland C. Leranth, Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University, School of Medicine, 333 Cedar Street, FMB 312, New Haven, CT 06520, USA G. Li, Department of Neurology, Programs in Neuroscience, Developmental Biology, University of California, San Francisco, 1550 4th Street, Rock Hall Room 448C, San Francisco, CA 94158, USA J.E. Lisman, Department of Biology and Volen Center for Complex Systems, Brandeis University, Waltham, MA 02454, USA T.A. Milner, Department of Neurology and Neuroscience, Division of Neurobiology, Weill-Cornell Medical College, 411 East 69th Street, New York, NY 10021, USA R.J. Morgan, Department of Anatomy and Neurobiology, 193 Irvine Hall, University of California, Irvine, CA 92697, USA J.J. O’Connor, UCD School of Biomolecular and Biomedical Science, Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland

vii

T.G. Ohm, Institute of Integrative Neuroanatomy, Department of Clinical Cell and Neurobiology, Charite´ CCM, 10098 Berlin, Germany J.M. Parent, Department of Neurology, University of Michigan Medical Center, 109 Zina Pitcher Place, 5021 BSRB, Ann Arbor, MI 48109-2200, USA P.R. Patrylo, Department of Physiology, Southern Illinois University School of Medicine, Carbondale, IL 62901, USA M. Pickering, UCD School of Biomolecular and Biomedical Science, Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland S.J. Pleasure, Department of Neurology, Programs in Neuroscience, Developmental Biology, University of California, San Francisco, 1550 4th Street, Rock Hall Room 448C, San Francisco, CA 94158, USA B. Po¨schel, Department of Cell Biology and Anatomy, New York Medical College, Valhalla, NY 10595, USA O. Rahimi, Department of Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, TX 78229, USA A. Rappert, Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany C.E. Ribak, Department of Anatomy and Neurobiology, University of California at Irvine, Irvine, CA 92697-1275, USA A Sahay, Departments of Neuroscience and Psychiatry, Division of Integrative Neuroscience, Columbia University, New York, NY 10032, USA V. Santhakumar, Department of Neurology, David Geffen School of Medicine, University of California, Los Angeles, CA, USA H.E. Scharfman, Departments of Pharmacology and Neurology, Columbia University, College of Physicians and Surgeons, New York, NY, USA and Center for Neural Recovery and Rehabilitation Research, Helen Hayes Hospital, New York State Department of Health, Rte 9W, West Haverstraw, NY 10993-1195, USA Present address: Nathan Kline Institute for Psychiatric Research, 140 Old Orangeburg Rd., Bldg. 35, Orangeburg, NY 10962, USA L. Seress, Central Electron Microscopic Laboratory, Faculty of Medicine, University of Pe´cs, Szigeti str. 12, 7624 Pe´cs, Hungary L.A. Shapiro, Department of Anatomy and Neurobiology, University of California at Irvine, Irvine, CA 92697-1275, USA I. Soltesz, Department of Anatomy and Neurobiology, 193 Irvine Hall, University of California, Irvine, CA 92697, USA G. Sperk, Department of Pharmacology, Medical University Innsbruck, Peter-Mayr-Str. 1a, 6020 Innsbruck, Austria P.K. Stanton, Department of Cell Biology and Anatomy, New York Medical College, Valhalla, NY 10595, USA T.R. Stoub, Department of Neurological Sciences, Rush University Medical Center, Chicago, IL 60612, USA T.P. Sutula, Department of Neurology H6/570 CSC, University of Wisconsin, 600 Highland Avenue, Madison, WI 53792, USA M.K. Tallent, Department of Pharmacology and Physiology, Drexel University College of Medicine, 245 N. 15 Street, Philadelphia, PA 19102, USA C. Wang, Department of Neurological Sciences, Rush University Medical Center, Chicago, IL 60612, USA A. Williamson, Department of Neurosurgery, Yale University School of Medicine, New Haven, CT 06518, USA

viii

M.P. Witter, Department of Anatomy and Neurosciences, VUMC, MF-G102C, P.O.Box 7057, 1007 MB Amsterdam, The Netherlands S. Zhao, Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104 Freiburg, Germany J. Zimmer, Anatomy and Neurobiology, Institute of Medical Biology, University of Southern Denmark, Winslowparken 21, DK-5000 Odense C, Denmark

Preface

Part I. Why is a guide to the dentate gyrus necessary? As more knowledge is acquired about the nervous system, one sometimes is lost in a sea of literature. The current generation surveys the literature primarily from PubMed, which provide a vast resource for data, but often there is little context for it, and even less of a guide to seminal but pre-1990 papers. Presumably this context, and a historical overview, can be gained by courses, lectures, and from reviews, but this is often difficult. One example of this problem is our understanding of the dentate gyrus. This part of the brain has always been a topic of intense interest, and pioneering studies from the distant and recent past attest to it. But where in a textbook or review can one find a comprehensive overview some basic information about the dentate gyrus? The primary goal of the book is to address this problem by bringing together in an organized manner a series of chapters that address, in an approachable format, the fundamental concepts about dentate gyrus structure, function, and their implications. Some have more detail than others, some provide distinct perspectives on a similar topic, but to some extent this is welcome because each reader may desire different degrees of detail, and some may want to compare viewpoints of different experts in the field. Often a region of the brain is best understood by first outlining its components, so the first section of the book deals with the fundamental structure of the dentate gyrus. In the first chapter, David Amaral, Helen Scharfman, and Pierre Lavenex describe the neuronal organization and intrinsic circuitry. La´szlo´ Seress discusses this further by comparing the dentate gyrus of different species, emphasizing those that are primarily used in research (mouse, rat, monkey). Menno Witter outlines the topography of arguably the most important afferent system, the input from entorhinal cortex (the perforant path). Csaba Leranth and Tibor Hajszan discuss other extrinsic afferents, such as those from septum, mammillary bodies, and other areas that are poorly understood compared to the perforant path, particularly from the physiological perspective. Morten Blaabjerg and Jens Zimmer discuss in detail one of the most impressive and complex aspects of the dentate gyrus, the mossy fibers axons of the dentate gyrus granule cells. The remarkable structural complexity of the mossy fiber system, including the unique specializations that characterize their terminal arbors, is like no other in the hippocampus. Following this overview is a complementary discussion of the expression, plasticity, and physiology of the mossy fiber system, by David Jaffe and Rafael Gutierrez. This section then ends with two overviews about the development of the dentate gyrus, the first by Michael Frotscher, Shanting Zhao, and Eckart Fo¨rster, and the second by Guangnan Li and Sam Pleasure. The emphasis of both is on the unique set of signals that orchestrate the development of the normal laminar organization and cytoarchitecture of the adult dentate gyrus. The second section of the book addresses the major neuronal cell types in the dentate gyrus, mostly in the rodent. Charles Ribak and Lee Shapiro begin this section with an overview of the structural characteristics of the neuronal cell types, emphasizing their unique ultrastructural ‘‘signatures.’’ Omid Rahimi and Brenda Claiborne discuss the granule cell from a developmental and morphological perspective. Anne Williamson and Peter Patrylo review characteristics of granule cells from an electrophysiological viewpoint, both in rodent and in human tissue studies. The non-granule cells are then addressed in two chapters that cover the two primary subtypes of non-granule cells: the hilar mossy cells and the GABAergic interneurons. Darrell

ix

x

Henze and Gyuri Buzsa´ki provide an overview about the mossy cells. Carolyn Houser discusses the GABAergic neurons, which are heterogeneous in anatomy, expression, and functional organization. Section three attempts to cover the numerous transmitters and neuromodulators that influence the dentate gyrus. Glutamatergic inputs are discussed in other parts of the book (the perforant path in Section 1, extrinsic inputs in Section 1, the CA3 input in Section 5), so this leaves the first chapter to consider GABA (Doug Coulter and Greg Carlson), followed by a series of reviews about non-classical neurotransmitters. Carrie Drake, Charlie Chavkin, and Teri Milner review the expression and organization of the opioid system, a complex topic because of the numerous types of opioid peptides, receptors, and actions. Melanie Tallent addresses the neuropeptide somatostatin, which is normally co-localized with + GABA in a subset of interneurons. Gunther Sperk, Trevor Hamilton, and William Colmers review neuropeptide Y and its receptors, functions, and plasticity. Carolyn Harley provides an overview of norepinephrine, a neuromodulator that has been known to have a robust influence in the dentate gyrus for decades. Jason Frazier addresses a topic that has been appreciated relatively recently, endocannabinoids in the dentate gyrus. Mark Pickering and John O’Connor discuss the pro-inflammatory cytokines, neuromodulators that have not been widely acknowledged until recent years. Marian Joe¨ls reviews the glucocorticoids, and Devin Binder discusses the neurotrophins, a family of ‘‘growth factors’’ that are expressed in high concentration in the dentate gyrus, and modulate function in a profound manner. The reproductive steroids, estrogen, progesterone, and androgen, are addressed by Tibor Hajszan, Teri Milner, and Csaba Leranth. One of the more remarkable characteristics of the dentate gyrus is its plasticity, and this is the focus of the 4th part of the volume. Brian Derrick begins with a perspective that ties together studies of in vitro and in vivo synaptic plasticity, circuitry, and the behaving animal. Clive Bramham reviews long-term potentiation (LTP) and its candidate mechanisms. Beatrice Po¨schel and Patric Stanton address long-term depression (LTD). Other examples of plasticity follow, such as lesion-induced plasticity. One of the most widely studied examples involves the response to a lesion of the entorhinal cortex, and this large literature is reviewed by Thomas Deller, Domenico Del Turco, Angelika Rappert, and Ingo Bechmann. They also provide a modern context by updating the field with what is currently understood from studies of mice. On a separate subject, Jack Parent provides a review of one of the most exciting developments in the history of dentate gyrus research, and could not illustrate plasticity any better: that neurogenesis occurs in the dentate gyrus throughout the lifespan of mammals. Finally, another remarkable type of plasticity, is discussed by Tom Sutula and Ed Dudek: mossy fiber sprouting. This reorganization of the mossy fiber pathway has captured the attention of many, because it involves dramatic anatomical and physiological changes, and can occur in response to many types of insults or injury. In the 5th part of the volume, network considerations are addressed. This is a rich and diverse area of research that is reflected by many perspectives, each suggesting distinct roles of the dentate gyrus as a structure. Ray Kesner begins, followed by La´szlo´ Acsa´dy and Szabolcs Ka´li. David Hsu discusses the concept that the dentate gyrus is a ‘‘gate’’ or ‘‘filter.’’ John Lisman turns to the question of the perforant path input, and specifically how the medial vs. lateral perforant path provide distinct roles. Helen Scharfman discusses the evidence that CA3 is a major input to the dentate gyrus, by virtue of axon collaterals that innervate diverse dentate neurons. This section ends with a comprehensive model of the dentate gyrus network, presented by Robert Morgan, Vijy Santhakumar, and Ivan Soltesz. In the last section of the book, an effort is made to review some of the literature that addresses how abnormalities within the dentate gyrus contribute to disease, or aging. Monica Chawla and Carol Barnes discuss how granule cell function changes with age, drawing upon the wealth of information from electrophysiology, imaging, and other approaches. Peter Patrylo and Anne Williamson also provide a review of this large field, but with a different perspective. The potential role of dentate gyrus neurogenesis in depression is reviewed by Amar Sahay, Michael Drew, and Rene Hen, whose lab was one of the first to provide evidence that deficits in neurogenesis in the dentate gyrus might be a reason for depression. Thomas

xi

Ohm discusses the diverse anatomical changes in the dentate gyrus that occur in Alzheimer’s disease. Leyla DeToledo-Morrell, Travis Stoub, and Changsheng Wang review their studies of Alzheimer’s disease, providing a persuasive argument that a major contributing factor is pathology that develops in the entorhinal area. Finally, Ed Dudek and Tom Sutula review a large and active area of research: epileptogenesis in the dentate gyrus. Although a book such as this can never cover all the topics and all information, and certainly cannot do justice to the multitude of contributors both from the past and present, its purpose is primarily as a guide, and readers will be led to references that can address topics that are not covered. Particularly for those who are not well versed in ‘‘hippocampology,’’ it is hoped that this volume can help – and in doing so, provide new impetus to resolve the many outstanding questions that remain about the dentate gyrus. Helen E. Scharfman Columbia University and Helen Hayes Hospital Part II. A guide to terminology and orientation The dentate gyrus is a complex, three-dimensionally curved structure that is coupled to another similarly curved structure, the hippocampus. As such, it is often difficult to understand its orientation in sections taken from any one plane (e.g., coronal, horizontal, etc.). Moreover, some of the terms used to describe the components of the dentate gyrus are confusing, and not completely defined in any textbook, to our knowledge. Therefore, here we provide a general guide to the three-dimensional organization of the dentate gyrus, and to terminology. The dentate gyrus in rodents and in primates is a composed of a compact layer containing densely packed granule cells. Their dendrites mainly lie on one side of the cell layer, and their axons on the other side. The dendritic layer, called the molecular layer, is bounded by the hippocampal fissure and the granule cell layer. The layer containing granule cell axons is a polymorphic zone, called the hilus. It includes two main classes of non-granule cells, GABAergic interneurons, and glutamatergic mossy cells (Fig. 1A). The borders of the hilus in the rat are shown in Fig. 1. In primates, the border between the hilus and CA3c is difficult to define, because the hilus is relatively small, and the large mossy cells of the hilus are close to the modified pyramidal cells of CA3. The tri-laminar dentate gyrus is folded such that in a cross-section it looks like a V- or C-shaped structure, having two ‘‘blades’’ and an area where these two blades meet (Fig. 1B). Moreover, it also curves as a whole, along a dorsal-ventral and medio-lateral axis (Figs. 1C, D). In the rat, the dorsal dentate gyrus is medial and anterior, and as it extends posteriorly, it moves to a more ventral and lateral position, eventually curving anteriorly again in its most ventrolateral position (Fig. 1D). Because the result is a structure that is curved in three-dimensions, it is suggested here that the best way to discuss the ‘‘blades’’ of the dentate gyrus cell layer is not the way that is most common. As shown in Table 1, the portion of the dentate gyrus that is closest to CA1 is often called the ‘‘upper’’ or ‘‘dorsal’’ blade. In actuality, this same part of the granule cell layer becomes ventral and inferior as the dentate gyrus curves from dorsal to ventral along the longitudinal axis. Therefore, it is more consistent to use the term ‘‘enclosed’’ or ‘‘hidden’’ for this portion of the cell layer. In other words, it is always enclosed in the CA1–CA3 pyramidal cell regions of the hippocampal formation. Similarly, the opposite part of the granule cell layer, typically referred to as a ‘‘lower’’ or ‘‘ventral’’ blade, always lies ‘‘exposed’’ or ‘‘open’’ because it is not embedded in the pyramidal cell layers of the hippocampal formation. Therefore, the terminology that applies to all parts of the dentate gyrus is ‘‘exposed’’ vs. ‘‘enclosed.’’ In the rodent, the curve of the tri-laminar dentate gyrus blades is slightly different in dorsal and ventral levels. This is reflected in the fact that sections cut in the horizontal plane through the middle-third of the dentate gyrus, show a cell layer that is folded into a ‘‘C’’ shape, yet coronal sections in the dorsal area show a more ‘‘V’’ like structure (Fig. 1B). At the point that is most ventral and anterior (also referred to as

xii

Fig. 1. The three-dimensional organization of the dentate gyrus. (A) The fundamental organization of the neurons in the dentate gyrus is shown, with the main categories of neurons (granule cells and non-granule cells, the latter including GABAergic interneurons and glutamatergic mossy cells). (B) The two-dimensional perspective of the dentate gyrus changes with septotemporal position and plane of section. In the dorsal, coronal plane, the layers are folded similar to a ‘‘V’’, with the apex or crest as the location of the point of the V. In the middle of the hippocampus cut in horizontal section, the granule cell layer is curved. At extreme septal locations, dorsal sections demonstrate an arc to the ‘‘V’’; at extreme temporal sites, the exposed blade is relatively small. The hilus of the rodent is illustrated schematically for two orientations. (C) Coronal sections from rostral to caudal levels illustrate the changes in the blades and illustrate the lack of consistency of terms such as ‘‘dorsal blade’’ across the entire length of the hippocampus. (D) Comparative views of the rodent and primate hippocampus illustrate differences in relative locations schematically.

temporal pole), the dentate appears as if almost no exposed blade is present. The most dorsal and anterior region (also referred to as septal pole) is distinct again, because the curve is flattened and the crest twisted upward, toward the dorsal aspect of the brain. Some of these changes impact the way the layers are discriminated, particularly the separation between the hilus and CA3. In the dorsal dentate gyrus, the CA3c

xiii Table 1 Formal term

Synonyms

A. Common nomenclature Stratum moleculare Stratum granulosum Polymorphic layer (zone) Subgranular layer (zone) Granule cell GABAergic neuron Glutamatergic hilar neuron Granule cell axons Enclosed blade Exposed blade

Molecular layer Granule cell layer Hilus, hilar region, CA4, zone 4 of Amaral (1978) 50–100 mm of the hilus, next to the granule cell layer Granular cell Interneuron Mossy cell Mossy fibers Upper, dorsal, inner, hidden, superior blade Lower, ventral, outer, open, inferior blade

Rodents

Primates

B. Species-specific nomenclature Dorsal, septal Ventral, temporal Rostral, anterior, dorsal pole

Dorsal, posterior Ventral, temporal Rostral, anterior, temporal pole

pyramidal cell layer extends far into the dentate gyrus, almost reaching the point where the blades meet (the apex or crest). At temporal levels, the CA3c pyramidal cell layer curves toward the enclosed blade. These issues stress the importance of clear descriptive criteria to differentiate between subfields and layers based on Nissl-stained sections, in conjunction with other staining methods, as needed. Another confusing set of terms describes the parts of the longitudinal axis of the dentate gyrus (and the entire hippocampal formation, for that matter). In rodents, this axis is most commonly referred to as the septal-to-temporal axis or dorsal-to-ventral axis (Fig. 1). In primates, the longitudinal axis is generally referred to as the posterior-to-anterior axis, where posterior is comparable to dorsal in rodents. The confusion lies in the fact that the dentate is positioned differently in the primate, with the more anterior tip located in the ventral part of the brain (Fig. 1). Helen Scharfman Columbia University and Helen Hayes Hospital Menno Witter Vrije Universiteit, Amsterdam and Norwegian University of Science and Technology, Trondheim

This page intentionally left blank

Contents

List of Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

Section I. Functional Organization and Development 1.

2.

3.

4.

5.

6.

7.

8.

The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies) D.G. Amaral, H.E. Scharfman and P. Lavenex (Davis, CA and West Haverstraw, NY, USA and Fribourg, Switzerland) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

Comparative anatomy of the hippocampal dentate gyrus in adult and developing rodents, non-human primates and humans L. Seress (Pe´cs, Hungary) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

23

The perforant path: projections from the entorhinal cortex to the dentate gyrus M.P. Witter (Trondheim, Norway and Amsterdam, The Netherlands). . . . . . . . . . .

43

Extrinsic afferent systems to the dentate gyrus C. Leranth and T. Hajszan (New Haven, CT, USA and Szeged, Hungary) . . . . . . .

63

The dentate mossy fibers: structural organization, development and plasticity M. Blaabjerg and J. Zimmer (Odense, Denmark) . . . . . . . . . . . . . . . . . . . . . . . . .

85

Mossy fiber synaptic transmission: communication from the dentate gyrus to area CA3 D.B. Jaffe and R. Gutie´rrez (San Antonio, TX, USA and Mexico City, Mexico) . . .

109

Development of cell and fiber layers in the dentate gyrus M. Frotscher, S. Zhao and E. Fo¨rster (Freiburg, Germany) . . . . . . . . . . . . . . . . . .

133

Genetic regulation of dentate gyrus morphogenesis G. Li and S.J. Pleasure (San Francisco, CA, USA) . . . . . . . . . . . . . . . . . . . . . . . .

143

Section II. Cellular Analyses 9.

Ultrastructure and synaptic connectivity of cell types in the adult rat dentate gyrus C.E. Ribak and L.A. Shapiro (Irvine, CA, USA) . . . . . . . . . . . . . . . . . . . . . . . . .

xv

155

xvi

10.

11.

12.

13.

Morphological development and maturation of granule neuron dendrites in the rat dentate gyrus O. Rahimi and B.J. Claiborne (San Antonio, TX, USA) . . . . . . . . . . . . . . . . . . .

167

Physiological studies of human dentate granule cells A. Williamson and P.R. Patrylo (New Haven, CT and Carbondale, IL, USA) . . . .

183

Hilar mossy cells: functional identification and activity in vivo D.A. Henze and G. Buzsa´ki (West Point, PA and Newark, NJ, USA). . . . . . . . . .

199

Interneurons of the dentate gyrus: an overview of cell types, terminal fields and neurochemical identity C.R. Houser (Los Angeles, CA, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

Section III. Neurotransmitters and Neuromodulators 14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Functional regulation of the dentate gyrus by GABA-mediated inhibition D.A. Coulter and G.C. Carlson (Philadelphia, PA, USA) . . . . . . . . . . . . . . . . . . .

235

Opioid systems in the dentate gyrus C.T. Drake, C. Chavkin and T.A. Milner (New York, NY and Seattle, WA, USA)

245

Somatostatin in the dentate gyrus M.K. Tallent (Philadelphia, PA, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

265

Neuropeptide Y in the dentate gyrus G. Sperk, T. Hamilton and W.F. Colmers (Innsbruck, Austria and Edmonton, AB, Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

285

Norepinephrine and the dentate gyrus C.W. Harley (St. John’s, NL, Canada) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

299

Endocannabinoids in the dentate gyrus C.J. Frazier (Gainesville, FL, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

319

Pro-inflammatory cytokines and their effects in the dentate gyrus M. Pickering and J.J. O’Connor (Dublin, Ireland) . . . . . . . . . . . . . . . . . . . . . . . .

339

Role of corticosteroid hormones in the dentate gyrus M. Joe¨ls (Amsterdam, The Netherlands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

355

Neurotrophins in the dentate gyrus D.K. Binder (Irvine, CA, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

371

Sex steroids and the dentate gyrus T. Hajszan, T.A. Milner and C. Leranth (New Haven, CT and New York, NY, USA and Szeged, Hungary) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399

xvii

Section IV. Plasticity 24.

Plastic processes in the dentate gyrus: a computational perspective B.E. Derrick (San Antonio, TX, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

417

25.

Control of synaptic consolidation in the dentate gyrus: mechanisms, functions, and therapeutic implications C.R. Bramham (Bergen, Norway) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453

26.

Comparison of cellular mechanisms of long-term depression of synaptic strength at perforant path–granule cell and Schaffer collateral–CA1 synapses B. Po¨schel and P.K. Stanton (Valhalla, NY, USA) . . . . . . . . . . . . . . . . . . . . . . .

473

Structural reorganization of the dentate gyrus following entorhinal denervation: species differences between rat and mouse T. Deller, D. Del Turco, A. Rappert and I. Bechmann (Frankfurt/Main, Germany). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

501

Adult neurogenesis in the intact and epileptic dentate gyrus J.M. Parent (Ann Arbor, MI, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

529

Unmasking recurrent excitation generated by mossy fiber sprouting in the epileptic dentate gyrus: an emergent property of a complex system T.P. Sutula and F.E. Dudek (Madison, WI and Salt Lake City, UT, USA) . . . . . .

541

27.

28.

29.

Section V. The Dentate Gyrus Network 30.

31.

32.

A behavioral analysis of dentate gyrus function R.P. Kesner (Salt Lake City, UT, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

567

Matching computational arguments with morphological and physiological properties to understand dentate functions L. Acsa´dy and S. Ka´li (Budapest, Hungary) . . . . . . . . . . . . . . . . . . . . . . . . . . . .

577

The dentate gyrus as a filter or gate: a look back and a look ahead D. Hsu (Madison, WI, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

601

33.

Role of the dual entorhinal inputs to hippocampus: a hypothesis based on cue/action (non-self/ self) couplets J. Lisman (Waltham, MA, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615

34.

The CA3 ‘‘backprojection’’ to the dentate gyrus H.E. Scharfman (West Haverstraw, New York, NY, USA) . . . . . . . . . . . . . . . . .

627

Modeling the dentate gyrus R.J. Morgan, V. Santhakumar and I. Soltesz (Irvine and Los Angeles, CA, USA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

639

35.

xviii

Section VI. The Dentate Gyrus in Aging and Disease 36.

Hippocampal granule cells in normal aging: insights from electrophysiological and functional imaging experiments M.K. Chawla and C.A. Barnes (Tucson, AZ, USA). . . . . . . . . . . . . . . . . . . . . . .

661

The effects of aging on dentate circuitry and function P.R. Patrylo and A. Williamson (Carbondale, IL and New Haven, CT, USA) . . . .

679

Dentate gyrus neurogenesis and depression A. Sahay, M.R. Drew and R. Hen (New York, NY, USA) . . . . . . . . . . . . . . . . .

697

The dentate gyrus in Alzheimer’s disease T.G. Ohm (Berlin, Germany) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

723

Hippocampal atrophy and disconnection in incipient and mild Alzheimer’s disease L. deToledo-Morrell, T.R. Stoub and C. Wang (Chicago, IL, USA) . . . . . . . . . . .

741

Epileptogenesis in the dentate gyrus: a critical perspective F.E. Dudek and T.P. Sutula (Salt Lake City, UT and Madison, WI, USA) . . . . . .

755

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

775

37.

38.

39.

40.

41.

See Color Plate Section at the end of this book

SECTION I

Functional Organization And Development

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 1

The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies) David G. Amaral1,, Helen E. Scharfman2 and Pierre Lavenex3 1

Department of Psychiatry and Behavioral Sciences, The M.I.N.D. Institute and the California National Primate Research Center, UC Davis, Davis, CA, USA 2 Departments of Pharmacology and Neurology, Columbia University, New York, NY 10032 and the Center for Neural Recovery and Rehabilitation Research, Helen Hayes Hospital, New York State Department of Health, Rte 9W, West Haverstraw, NY 10993-1195, USA 3 Department of Medicine, Unit of Physiology, University of Fribourg, 1700 Fribourg, Switzerland

Abstract: The dentate gyrus is a simple cortical region that is an integral portion of the larger functional brain system called the hippocampal formation. In this review, the fundamental neuroanatomical organization of the dentate gyrus is described, including principal cell types and their connectivity, and a summary of the major extrinsic inputs of the dentate gyrus is provided. Together, this information provides essential information that can serve as an introduction to the dentate gyrus — a ‘‘dentate gyrus for dummies.’’ Keywords: neuroanatomy; circuits; connections; cell types One of the unique features of the hippocampal formation is that many of its connections are unidirectional (Fig. 1). Thus, the entorhinal cortex provides the major input to the dentate gyrus via fibers called the perforant path (see Chapter 3). However, the dentate gyrus does not return a projection to the entorhinal cortex. Since the entorhinal cortex is the source of much of the cortical sensory information that the hippocampal formation uses to carry out its functions, and since the dentate gyrus is the major termination of projections from the entorhinal cortex, it is reasonable to consider the dentate gyrus as the first step in the processing of information processing that ultimately leads to the production of episodic memories. Moreover, the unique neuroanatomy of the dentate gyrus predicts that it carries out a specific information-processing task with the information

Introduction The dentate gyrus is a simple cortical region that is an integral portion of the larger functional brain system called the hippocampal formation (Fig. 1) (Amaral and Lavenex, 2007). The regular organization of its principal cell layers coupled with the highly ordered laminar distribution of many of its inputs has encouraged its use as a model system for many facets of modern neurobiology. In this chapter, we present the fundamental principles of neuroanatomical organization of the dentate gyrus, its principal cell types and their connectivity, and a summary of the major extrinsic inputs of the dentate gyrus. Corresponding author. Tel.: +1 916 703 0225; Fax: +1 916 703 0287; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63001-5

3

4

5

that it receives from the entorhinal cortex and ultimately conveys to the CA3 field of the hippocampus. In this chapter, we will first give a general overview of the neuroanatomical organization of the dentate gyrus. We will then put the spotlight on three of its most important neurons: the dentate granule cell, the dentate pyramidal basket cell and the mossy cell. We will summarize the major structural features of these neurons as well as the connections they make within and beyond the dentate gyrus. We will then very briefly discuss some of the other neurons located in the dentate gyrus. Finally, we briefly summarize the origin and termination of the major extrinsic inputs to the dentate gyrus. Many of these topics will be discussed in greater detail in other chapters of this book. But, this chapter provides a synoptic view in order to serve as an introduction to the dentate gyrus — this is dentate gyrus for dummies.

Overall organization The dentate gyrus has three layers (Figs. 1–2). There is a relatively cell-free layer called the molecular layer which, in the rat, is approximately 250 mm thick and is occupied by, among other things, the dendrites of the dentate granule cells. The other major occupants of the molecular layer are the fibers of the perforant path that originate in the entorhinal cortex. There are also a small

number of interneurons that reside in the molecular layer and fibers from a variety of extrinsic inputs terminate there. The principal cell layer, the granule cell layer, is made up largely of densely packed granule cells. The thickness of the granule cell layer ranges from 4 to 8 neurons or
60 mm. While the granule cell layer is mainly made up of granule cells, there are some other neurons that are located at the boundary of the granule and polymorphic layers. The cell body of the dentate pyramidal basket cell, for example, is often located just within the granule cell layer at its border with the polymorphic layer. The granule cell layer encloses a cellular region, the polymorphic cell layer, which constitutes the third layer of the dentate gyrus (Fig. 1). A number of cell types are located in the polymorphic layer but the most prominent is the mossy cell that we will describe below. The dentate gyrus along with many other fields of the hippocampal formation create a bananashaped structure that, in the rat, runs from the septal nuclei rostrally to the temporal cortex caudally. Thus, the long axis is typically called the septotemporal axis. The axis at right angles to the septotemporal axis is typically called the transverse axis. The dentate gyrus has a relatively similar structure at all septotemporal levels of the hippocampal formation and is not generally divided into subregions. However, it does tend to have more of a ‘‘V’’ shape septally and more of a ‘‘U’’ shape temporally (Fig. 3). In discussing features of the dentate gyrus, it is often useful to refer to a

Fig. 1. The rat hippocampal formation. (A) Nissl-stained horizontal section through the hippocampal formation of the rat. The major fields are indicated. Projections (1) originate from layer II of the entorhinal cortex (EC) and terminate in the molecular layer of the dentate gyrus (DG) and in the stratum lacunosum-moleculare of the CA3 field of the hippocampus. An additional component of the perforant path originates in layer III and terminates in the CA1 field of the hippocampus and the subiculum. Granule cells of the DG give rise to the mossy fibers (2) that terminate both within the polymorphic layer of the DG and within stratum lucidum of the CA3 field of the hippocampus. The CA3 field, in turn, gives rise to the Schaffer collaterals (3) that innervate the CA1 field of the hippocampus. Pyramidal cells in CA1 project to the subiculum and to the deep layers of the EC. The subiculum also gives rise to projections to the deep layers of the EC. (B) Line drawing to illustrate the major regional and laminar organization of the DG. The DG is divided into a molecular layer (ml) a granule cell layer (gcl) and a polymorphic layer (pl). The molecular layer is divided into three sublayers based on the laminar organization of inputs. The hippocampus is divided into CA3, CA2 and CA1 subfields. Within CA3, a number of layers are defined. The main cell layer is the pyramidal cell layer (pcl). Deep to the pyramidal cell layer is the stratum oriens (so); deep to this is the white matter of the alveus (al). Superficial to the pyramidal cell layer is stratum lucidum (sl), stratum radiatum (sr) and stratum lacunosum-moleculare (sl-m). Fields CA2 and CA1 have the same layers as CA3 except for stratum lucidum. The remainder of the hippocampal formation is made up of the subiculum (Sub), presubiculum (Pre), parasubiculum (Para) and EC. Layers of these latter structures are indicated with roman numerals. Additional abbreviations: ab, angular bundle; fi, fimbria; hf, hippocampal fissure. (C) Schematic illustration of DG and hippocampus to illustrate position of suprapyramidal blade, infrapyramidal blade and crest of the DG. This model is used in subsequent illustrations to demonstrate the major cell types and connections of the DG. (See Color Plate 1.1 in color plate section.)

6

Fig. 2. Comparative Nissl and Timm’s staining of the rat and monkey dentate gyrus. Approximately the same portions of the rat (A & B) and monkey (C & D) hippocampal formation are illustrated. The magnification of the paired images are different. The increased complexity of the hilar region in the monkey compared to the rat is obvious in these photomicrographs. In the monkey, the CA3 field inserts more deeply into the dentate gyrus and accounts for a much greater area of the hilar region. While the darkly stained polymorphic layer is pronounced in the rat, it is much thinner in the monkey brain. The condensed layer of mossy fibers is also much more pronounced in the rat compared to the monkey. Also note that the strict lamination of the molecular layer into thirds in the rat is not as sharp as in the monkey. This corresponds to a more gradient-like termination of the medial and lateral perforant paths in the monkey compared to the rat. Calibration, 250 mm (A, B); 350 mm (C, D).

particular transverse portion of the ‘‘V’’- or ‘‘U’’shaped structure. The portion of the granule cell layer that is located between the CA3 field and the CA1 field (separated by the hippocampal fissure) is called the suprapyramidal (above CA3) blade and the portion opposite to this, the infrapyramidal (below CA3) blade. The region bridging the two blades (at the apex of the ‘‘V’’ or ‘‘U’’) is called the crest (Fig. 3).

Comparative neuroanatomy of the dentate gyrus The comparative neuroanatomy of the dentate gyrus is covered in detail in Chapter 2. However, we would like to raise a few issues for which there remains substantial confusion in the literature. First, the basic trilaminar structure of the dentate gyrus is common across all species studied. And,

relative to other hippocampal structures such as the CA1 field of the hippocampus or the entorhinal cortex, there has not been substantial phylogenetic modification. For example, in the rat, there are approximately one million granule cells in the dentate gyrus. There are approximately ten times more granule cells in the monkey than in the rat, a ratio that parallels the overall volume differences. But there are only 15 times more dentate granule cells in humans when compared to rats, while the volume of the dentate gyrus plus the hippocampus is
100 times larger in humans than in rats (reviewed in more detail in Amaral and Lavenex, 2007). Despite the similarities, there are also some obvious differences. In the rat, the CA3 field of the hippocampus inserts into the dentate gyrus and abuts the cells of the polymorphic layer. By using stains such as the Timm’s stain that identifies

7

Fig. 3. Horizontal sections through the rat hippocampal formation. This figure illustrates a more dorsally situated (A) and a more ventrally situated (B) horizontal section through the rat hippocampal formation. The approximate level of the section is illustrated on a 3D reconstruction of a magnetic resonance image series of the rat brain. Subtle differences in the cytoarchitectonic organization are seen throughout the hippocampal formation. The dentate gyrus, takes on a more ‘‘V’’ shape dorsally and a more ‘‘U’’ shape ventrally. Calibration bar ¼ 250 mm. (See Color Plate 1.3 in color plate section.)

depositions of metals such as zinc (Fig. 2), the heavily labeled polymorphic layer can be easily differentiated from the CA3 field. This differentiation is not so clear in the nonhuman primate or in the human dentate gyrus. This is because the CA3 field is much more prominent within the confines of the dentate gyrus and the polymorphic layer has thinned to a very narrow subgranular region (Fig. 2). In the early nomenclature espoused by Lorente de No´, he used a term ‘‘CA4’’ that was vaguely defined but appears to have been intended for the part of CA3 that inserts within the dentate gyrus. However, sometimes Lorente de No´ appears to have used the CA4 term for the polymorphic layer. We have recommended that this term be abandoned and that even the part of CA3 that enters the limbs of the dentate gyrus be called CA3. There are a variety of other differences in the rat, monkey and human dentate gyrus. The granule cells, which we will describe in much more detail shortly, only have apical dendrites in the rat. But in the monkey and human, many granule cells also have basal dendrites (Seress and Mrzljak,

1987). Similarly, the mossy cells of the rat have very rare dendrites that enter the molecular layer. Yet, in the macaque monkey, many of the mossy cells give rise to numerous dendrites that enter the molecular layer (Buckmaster and Amaral, 2001). The implication of this is that rat mossy cells get little if any perforant path input whereas monkey mossy cells could get a quite substantial perforant path input. Unfortunately, many of the connections of the monkey and human dentate gyrus have yet to be investigated. One example of a projection that has been studied in both species is the commissural connection. While cells of the polymorphic layer (mainly the mossy cells) give rise to a very robust commissural projection to the molecular layer of the contralateral dentate gyrus in the rat, this projection does not exist in the nonhuman primate brain or in the human brain. As connections of the primate and human dentate gyrus are better studied, it is undoubtedly the case that there will be other fundamental differences that will affect not only normal function but also pathological processes. This raises the caveat that while the rodent

8

dentate gyrus is an extremely valuable tool for evaluating function and potential mechanisms of pathology, caution must be exercised when these results are extrapolated to the human dentate gyrus. For the remainder of this chapter, we will focus on neuroanatomical findings from the rat brain.

Major cell types of the dentate gyrus — and their connections The dentate granule cell The principal cell type of the dentate gyrus is the granule cell (Figs. 4 and 5). The dentate granule cell has an elliptical cell body with a width of approximately 10 mm and a height of 18 mm (Claiborne et al., 1990). The granule cell bodies are tightly packed together and, in most cases, there is no glial sheath interposed between cells. The granule cell has a characteristic cone-shaped tree of

spiny apical dendrites. The branches extend throughout the molecular layer and the distal tips of the dendritic tree end just at the hippocampal fissure or at the ventricular surface. The total length of the dendritic trees of granule cells located in the suprapyramidal blade are, on average, larger than those of cells in the infrapyramidal blade (3500 mm vs. 2800 mm, respectively). Dendrites of cells in the suprapyramidal blade have 1.6 spines/mm, whereas dendrites in the infrapyramidal blade have 1.3 spines/mm (Desmond and Levy, 1985). Thus, an estimate for the number of spines on the average suprapyramidal granule cell would be around 5600 and for an infrapyramidal cell 3640. Since virtually all of the excitatory inputs to granule cells are on these dendritic spines, these numbers indicate the approximate number of excitatory synapses that dentate granule cells receive from all sources. As noted earlier, the total number of granule cells in one dentate gyrus of the rat is
1.2  106 (West et al., 1991; Rapp and Gallagher, 1996). Although neurogenesis in

Fig. 4. The dentate granule cell. The characteristic features of the dentate granule cell are illustrated, including its axonal arbor. A collateral plexus gives rise to numerous (
200) typical synapses on cells located within the polymorphic layer. Most of these synapses are onto the dendrites of inhibitory interneurons. Some of the large mossy fiber expansions are also distributed in the polymorphic layer. Many of these terminate on the proximal dendrites of mossy cells. The mossy fiber axons ultimately enter the CA3 field where they travel through the full transverse extent of the field. On their course, they terminate with mossy fiber expansions on a small number (15–20) of CA3 pyramidal cells. Additional abbreviations: gc, granule cell; pc, pyramidal cell. (See Color Plate 1.4 in color plate section.)

9

Fig. 5. The dentate granule cell. A photomicrograph (A) and line drawing (B) of a prototypical granule cell that was filled with Lucifer yellow in a hippocampal slice. The dendrites arise primarily from the apical surface of the cell body, and the axon emerges from the basal surface. The spiny dendrites extend into the molecular layer until the hippocampal fissure, and the axon collateralizes profusely within the polymorphic layer. Calibration bar ¼ 25 mm (A), 20 mm (B).

the dentate gyrus persists into adulthood, and appears to be under environmental control, modern stereological studies have shown that the total number of granule cells does not vary in adult animals (Rapp and Gallagher, 1996). This implies that there is a steady state turnover of granule cells rather than a continuous accretion. The packing number of granule cells and their ratio to CA3 pyramidal cells varies along the septotemporal axis (Gaarskjaer, 1978b); the packing density is higher septally than temporally. Since the packing density of CA3 pyramidal cells follows an inverse gradient, the net result is that at septal levels of the hippocampal formation, the ratio of granule cells to CA3 pyramidal cells is something on the order of 12:1, whereas at the temporal pole the ratio drops to 2:3. Since the CA3 pyramidal cells are the major recipients of granule cell innervation, and the number of mossy fiber synapses is roughly the same along the septotemporal axis, contact probability is much lower septally than temporally.

The mossy fibers — projections to the polymorphic layer The granule cells give rise to distinctive unmyelinated axons which Ramo´n y Cajal called mossy

fibers. The mossy fibers have unusually large boutons that form en passant synapses with the mossy cells of the polymorphic layer and with the CA3 pyramidal cells of the hippocampus. Less appreciated is the fact that the mossy fiber axons give rise to a distinctive set of collaterals that heavily innervate cells within the polymorphic layer of the dentate gyrus. Each principal mossy fiber (which is on the order of 0.2–0.5 mm in diameter) gives rise to about seven thinner collaterals within the polymorphic layer before entering the CA3 field of the hippocampus. As much as 2300 mm of collateral axonal plexus is generated by a single mossy fiber in the polymorphic layer (Claiborne et al., 1986). Within the polymorphic layer, the mossy fiber collaterals branch extensively and the daughter branches bear two types of synaptic varicosities. Numerous small (approximately 2 mm) spherical synaptic varicosities are distributed unevenly along these collaterals. There are
160–200 of these varicosities distributed throughout the axonal collateral plexus of a single granule cell, and these form contacts on dendrites located in the polymorphic layer. At the end of many of the collateral branches there are also larger (3–5 mm diameter), irregularly shaped varicosities that resemble, although are smaller than, the mossy fiber boutons found in CA3. The mossy fiber terminals in the polymorphic layer

10

establish contacts with the proximal dendrites of the mossy cells (Ribak et al., 1985), the basal dendrites of the pyramidal basket cells as well as with other, unidentified cells. Acsady et al. (1998) reported that the vast majority of mossy fiber collaterals in the polymorphic cell layer terminate on GABAergic interneurons. Since there are 160–200 such varicosities (compared with
20 of the larger thorny excrescences), mossy fiber axons synapse on a larger number of interneurons than mossy cells or CA3 pyramidal cells. Mossy fiber collaterals occasionally enter the granule cell layer, but they rarely enter the molecular layer under normal conditions. The collaterals that enter the granule cell layer appear to terminate preferentially on the apical dendritic shafts of pyramidal basket cells. The lack of mossy fiber innervation of the molecular layer changes dramatically in pathological conditions and sprouting of mossy fibers is one of the major hallmarks of temporal lobe epilepsy (see Chapter 29).

The mossy fibers — projections to CA3 The rat dentate gyrus does not project to any brain region other than the CA3 field of the hippocampus (but see Chapter by Zimmer for exceptions in other species). The mossy fibers terminate in a relatively narrow zone mainly located just above the CA3 pyramidal cell layer (Blackstad et al., 1970; Swanson et al., 1978; Gaarskjaer, 1978a; Claiborne et al., 1986). In the proximal portion of CA3, mossy fibers are also located below and within the pyramidal cell layer. The layer of mossy fiber termination located just above the pyramidal cell layer is called stratum lucidum. There is no indication that dentate neurons other than the granule cells project to CA3. In particular, cells in the polymorphic layer do not project to the hippocampus, at least in the rodent. The dentate projection to CA3 stops near the border of CA3 with CA2, and the lack of granule cell input is one of the main features that distinguishes CA3 from CA2 pyramidal cells. As far as can be determined, all dentate granule cells appear to project to CA3 and the axon trajectory is partially correlated with the position of

the parent cell body. In the proximal portion of CA3 (closer to the dentate gyrus), mossy fibers are distributed below, within and above the pyramidal cell layer. The fibers located below the layer, i.e., those that are in the area occupied primarily by basal dendrites, are generally called the infrapyramidal bundle (Fig. 2). The fibers located within the pyramidal cell layer are called the intrapyramidal bundle and those located above the pyramidal cell layer (in the area occupied mainly by proximal apical dendrites) the suprapyramidal bundle. The suprapyramidal bundle occupies the stratum lucidum. At mid and distal portions of CA3, the intra- and infrapyramidal bundles are largely eliminated and those fibers that were in these regions cross the pyramidal cell layer and join the other mossy fibers within stratum lucidum. Granule cells at all transverse positions within the granule cell layer generate mossy fibers that extend for the full transverse extent of CA3 (Gaarskjaer, 1981). Cells located in the infrapyramidal blade of the granule cell layer have axons that tend to enter CA3 in the infrapyramidal bundle, but ultimately cross the pyramidal cell layer to enter the deep portion of stratum lucidum. The axons of granule cells located in the crest of the dentate gyrus tend to enter CA3 in the intrapyramidal bundle and also ultimately ascend into stratum lucidum. Cells located in the suprapyramidal blade of the dentate gyrus give rise to axons that enter CA3 in the stratum lucidum and continue within the most superficial portion of stratum lucidum (Claiborne et al., 1986). Early Golgi anatomists indicated that the mossy fiber axons were mainly oriented perpendicular to the long axis of the hippocampus. Blackstad and colleagues confirmed the largely transverse trajectory of the mossy fibers using degeneration track tracing methods. Mossy fibers originating from any particular septotemporal level travel through the full transverse extent of CA3 in a very narrow septotemporal zone of
1 mm, which has been described as a lamella. There is one peculiarity of the mossy fiber projection, however, that has eluded explanation. Lorente de No´ (1934) and McLardy (McLardy and Kilmer, 1970) indicated that the mossy fibers

11

actually do change their course and take a longitudinal direction, but not until they reach the distal portion of CA3. The extent of this distal longitudinal projection was further clarified by Swanson and colleagues, using the autoradiographic method of neural tract tracing (Swanson et al., 1978). They showed that granule cells located at septal levels give rise to mossy fibers that travel throughout most of the transverse extent of CA3 at the same septotemporal level. But just at the CA3/CA2 border, they abruptly change course and travel toward the temporal pole for nearly 2 mm. The extent of this longitudinal component, however, appears to depend on the septotemporal location of the cells of origin. Granule cells in the mid to temporal portions of the dentate gyrus have mossy fibers that exhibit only a slight temporal inclination at their distal extremity. And mossy fibers that originate at the extreme temporal pole of the dentate gyrus barely extend to the CA3/CA2 border and have little or no longitudinal component. On the face of it, this indicates that some CA3 pyramidal cells that are located very close to the border with CA2 may be contacted by mossy fiber axons from granule cells spread out over a much broader septotemporal extent of the dentate gyrus. They might, therefore, form a special class of integrator CA3 cells. An elegant neuroanatomical verification of this has come from the work of Acsady et al. (1998) who stained dentate granule cells by using in vivo intracellular techniques. Not only do mossy fibers travel temporally for 2–3 mm, but they also demonstrate typical mossy fiber expansions along this portion of their trajectory. There is substantial evidence indicating that the granule cells use glutamate as their primary transmitter, and the asymmetric contacts made between the mossy fiber expansions and the thorny excrescences would tend to confirm this notion. The mossy fibers are, nonetheless, also immunoreactive for several other neuroactive substances. At least some of the mossy fibers demonstrate immunoreactivity for the opioid peptide dynorphin and they are also immunoreactive for GABA (Walker et al., 2002). The chemical neuroanatomy of the dentate granule cell will be dealt with in more detail in Chapter 6.

The dentate pyramidal basket cell As with most other brain regions, some neurons of the dentate gyrus, such as the granule cells, are excitatory whereas others are inhibitory. The most intensively studied type of inhibitory interneuron in the dentate gyrus is the pyramidal basket cell (Figs. 6 and 7; Ribak et al., 1978; Ribak and Seress, 1983). These cells are generally located along the interface between the granule cell layer and the polymorphic layer. They have pyramidalshaped cell bodies that are substantially larger (25–35 mm in diameter) than the granule cells (10–18 mm). Ramo´n y Cajal first described the pyramidal basket cells as having a single, principal aspiny apical dendrite directed into the molecular layer that divides into several aspiny branches, and several basal dendrites that divide and extend into the polymorphic cell layer. The number of basket cells is not constant throughout either the transverse or septotemporal extents of the dentate gyrus (Seress and Pokorny, 1981). At septal levels, the ratio of basket cells to granule cells is 1:100 in the suprapyramidal blade and 1:180 in the infrapyramidal blade. At temporal levels, the number is 1:150 for the suprapyramidal blade and 1:300 for the infrapyramidal blade. The basket portion of the name refers to the fact that the axon of these cells forms pericellular plexuses, like the covering of a chianti bottle, that form encompassing synapses with the cell bodies of granule cells. The dentate pyramidal basket cell along with other basket cells located just below the granule cell layer contribute to a very dense terminal plexus that is confined to the granule cell layer (Struble et al., 1978; Sik et al., 1997). The terminals in this basket plexus are GABAergic and form symmetric, inhibitory contacts, located primarily on the cell bodies and proximal dendritic shafts of the apical dendrites of the granule cells. Analysis of Golgi-stained axonal plexuses from single basket cells in the rat indicates that they extend for distances greater than 900 mm in the transverse axis and
1.5 mm in the septotemporal axis. This widely distributed axonal plexus would allow a single basket cell to influence as many as 10,000 (1%) of the granule cells.

12

Fig. 6. The pyramidal basket cell. The cell body of the pyramidal basket cell is located at the interface between the granule cell layer and the polymorphic cell layer. The axon (arrow) emerges from the apical dendrite. Collaterals of this axon form a curtain of terminals that synapse with the granule cell bodies. Additional abbreviations: pbc, pyramidal basket cell. (See Color Plate 1.6 in color plate section.)

Fig. 7. The pyramidal basket cell. A prototypical pyramidal basket cell is shown after intracellular injection of Neurobiotin. The montage that was created after visualization of the cell (A) and line drawing (B) illustrate the characteristics of this cell type. An arrow points to the axon. Calibration bar ¼ 25 mm (A), 40 mm (B).

13

Fig. 8. The mossy cell. A line drawing of a mossy cell (mc) in the polymorphic layer. The axon (arrow) develops a plexus within the polymorphic layer, and also an ipsilateral projection to the inner molecular layer, known as the associational pathway. The main axon also projects contralaterally to the inner molecular layer, forming the commissural pathway. The ipsilateral projection increases in density with distance from the cell body of origin. (See Color Plate 1.8 in color plate section.)

Fig. 9. The mossy cell. A montage of several focal planes (A), and line drawing (B) of a classic mossy cell, illustrating the characteristic thorny excrescences and dendritic tree of this cell type. Thorny excrescences are present proximal to the soma. The dendrites of the cell extend throughout nearly the entire polymorphic region, but few enter either the granule cell or molecular layers. Calibration bar ¼ 25 mm (A), 50 mm (B).

The mossy cell The polymorphic layer harbors a variety of neuronal cell types. The most common, and certainly the most impressive, is the mossy cell (Figs. 8 and 9) This cell type is probably what Ramo´n y Cajal referred to as the ‘‘stellate or triangular’’ cells

located in his ‘‘subzone of fusiform cells’’, and is undoubtedly what Lorente de No´ referred to as ‘‘modified pyramids’’. The mossy cell received its current name from Amaral (1978) who studied this and other neurons of the polymorphic layer using the Golgi method. The cell bodies of the mossy cells are large (25–35 mm) and are often triangular

14

or multipolar in shape. Three or more thick dendrites originate from the cell body and extend for long distances within the polymorphic layer. Each principal dendrite bifurcates once or twice and generally gives rise to a few side branches. While most of the dendritic branches remain within the polymorphic layer, an occasional dendrite pierces the granule cell layer and enters the molecular layer. The mossy cell dendrites virtually never enter the adjacent CA3 field in the rat. The most distinctive feature of the mossy cell is that all of its proximal dendrites are covered by very large and complex spines, the so-called thorny excrescences (Fig. 9). These are the distinctive sites of termination of the mossy fiber axons. While thorny excrescences are also observed on the proximal dendrites of CA3 pyramidal cells, they are never as dense or as complex as the ones on the mossy cells. The distal dendrites of the mossy cell have typical pedunculate spines that appear to be less dense than those on the distal dendrites of the hippocampal pyramidal cells.

Mossy cell projections The inner third of the molecular layer (Fig. 2) receives a projection that originates exclusively from neurons in the polymorphic layer (Laurberg and Sorensen, 1981; Frotscher et al., 1991; Buckmaster et al., 1992, 1996). Since, in the rat, this projection originates both in the ipsilateral and contralateral sides of the dentate gyrus, it has been called the associational/commissural projection. As noted above, the commissural portion of this projection does not exist in the primate brain. There is substantial evidence from neural track tracing studies that this projection arises almost exclusively from cells in the polymorphic layer and mainly from the mossy cells. The fact that the mossy cells are immunoreactive for glutamate (Soriano and Frotscher, 1994) adds credence to the notion that the associational/commissural projection is excitatory. A direct electrophysiological demonstration of the excitatory nature of the mossy cell synapse on the granule cell has been provided by Scharfman (1994). Scharfman (1995) has also demonstrated that mossy cells innervate both granule

cells and GABAergic interneurons. It should also be noted for the sake of completeness that a small component of the commissural connection in the rat does appear to arise from GABAergic neurons (Ribak et al., 1986). There are a number of interesting features about the ‘‘feedback’’ projection from the mossy cells to the granule cells. First, the projection from mossy cells located at any particular level of the dentate gyrus is distributed widely along the longitudinal axis, both septally and temporally from the point of origin. Axons from any particular septotemporal point in the dentate gyrus may innervate as much as 75% of the long axis of the dentate gyrus (Amaral and Witter, 1989). Second, the projection to the molecular layer at the septotemporal level of origin is very weak, but gets increasingly stronger at levels that are progressively more distant from the cells of origin. Remembering that mossy cells are the recipients of massive innervation from the granule cells at their same level (via the mossy fiber collaterals into the polymorphic layer), it would appear that the mossy cells pass on the collective output of granule cells from one septotemporal level to granule cells located at distant levels of the dentate gyrus.

Other neurons of the dentate gyrus There has been an explosion in the number of interneurons identified in the rat hippocampal formation. A detailed overview of the characteristics of the various hippocampal interneurons has been published by Freund and Buzsaki (1996) and is also covered in Chapter 13 of this book. Many of the cell types can be distinguished on the basis of the distribution of their axonal plexus. Some have axons that terminate on cell bodies, whereas others have axons that terminate exclusively on the initial segments of other axons. Interneurons have also been distinguished on the basis of their inputs. Some are preferentially innervated, for example, by the serotonergic fibers originating from the raphe nuclei. Interneurons can also be differentiated from principal cells on the basis of their electrophysiological characteristics. At least some interneurons have high rates of spontaneous

15

activity and fire in relation to the theta rhythm. For this reason, interneurons are often called theta cells (Sik et al., 1997). Within the same subgranular region occupied by the cell bodies and dendrites of the pyramidal basket cells are several other cell types with distinctly different somal shapes, as well as different dendritic and axonal configurations. Some of these cells are multipolar with several aspiny dendrites entering the molecular and polymorphic layers, while others tend to be more fusiform-shaped with a similar dendritic distribution. As Ribak and colleagues have pointed out, many of these cells share fine structural characteristics such as infolded nuclei, extensive perikaryal cytoplasm with large Nissl bodies and intranuclear rods. Moreover, it appears that all of these cells give rise to axons that contribute to the basket plexus in the granule cell layer. Most of these neurons are immunoreactive for GABA, form symmetrical synaptic contacts with the cell bodies, proximal dendrites and occasionally with axon initial segments of granule cells, and therefore function as inhibitory interneurons. These cells are not neurochemically homogeneous, however, since subsets appear to colocalize distinct categories of other neuroactive substances (Ribak, 1992).

the location of the cell body and to the region where the axon is distributed. Unfortunately, this terminology has not been embraced and more widely applied to neurons of the hippocampal formation. Frotscher and colleagues have described a second type of neuron in the molecular layer that resembles the chandelier or axo-axonic cell originally found in the neocortex (Soriano and Frotscher, 1989). These cells are generally located immediately adjacent or even within the superficial portion of the granule cell layer. The axoaxonic cell is named for the fact that its axon descends from the molecular layer into the granule cell layer, collateralizes profusely and then terminates, with symmetric synaptic contacts, exclusively on the axon initial segments of granule cells. Thus, their shape resembles that of a chandelier. Each axo-axonic cell may innervate the axon initial segments of as many as 1000 granule cells. Since these cells are immunoreactive for markers of GABAergic neurons and make symmetrical synapses, it is likely that they provide an additional means of inhibitory control of granule cell output.

Neurons of the polymorphic cell layer Neurons of the molecular layer The molecular layer is occupied primarily by dendrites of the granule cells, pyramidal basket cells and polymorphic layer cells, as well as axons and terminal axonal arbors from the entorhinal cortex and other sources. At least two neuron types are also present in the molecular layer. The first is located deep in the molecular layer, has a multipolar or triangular cell body and gives rise to an axon that produces a substantial terminal plexus largely limited to the outer two thirds of the molecular layer. This neuron has aspiny dendrites that remain mainly within the molecular layer and has been called the MOPP cell (molecular layer perforant path-associated cell). This terminology was proposed by Han et al. (1993) to bring some order to naming interneurons in the hippocampal formation. The lettering system refers to

Besides the mossy cell, there are a number of fusiform cells in the polymorphic layer. The main difference between the fusiform cell types is whether they have spines or not and the characteristic shapes and sizes of the spines. One type, the long-spined multipolar cell first described by Amaral (1978) has recently been called the HIPP cell (hilar perforant path-associated cell) (Fig. 10; Han et al., 1993). The conspicuous feature of this cell is the distribution of copious, long and often branched spines over its cell body and dendrites. Intracellular staining techniques demonstrate that these cells have axons that ascend into the outer two-thirds of the molecular layer (i.e., the perforant path zone) and terminate with symmetrical and presumably inhibitory synapses on the dendrites of granule cells. An amazing feature of these neurons is that their axonal plexus can extend for as much as 3.5 mm along the septotemporal axis of

16

Fig. 10. The long-spined cell. A montage (A) and line drawing (B) of a long-spined cell in the polymorphic layer. The extremely long spines that characterize this cell type are marked by arrowheads, and can be proximal as well as distal to the cell body, although in this example they were primarily located along distal dendrites. The axon of this cell collateralized in the molecular layer. Some of the longspined cells correspond to the GABAergic interneurons that colocalize somatostatin, so-called HIPP cells. Calibration bar ¼ 25 mm (A), 50 mm (B).

the dentate gyrus (the entire length of the dentate gyrus in the rat is only
10 mm) and may generate as many as 100,000 synaptic terminals. Since inhibitory interneurons typically have aspiny dendrites and relatively local axonal plexuses, this long spined multipolar/HIPP cell is a very atypical interneuron! At least some of these HIPP cells appear to correspond to the somatostatin/GABA cells that give rise to the somatostatin innervation of the outer portion of the molecular layer. Antibodies directed against the peptide somatostatin have revealed that neurons such as the HIPP cells scattered throughout the polymorphic layer are immunoreactive for this peptide, and account for approximately 16% of the GABAergic cells in the dentate gyrus (Morrison et al., 1982;

Bakst et al., 1986; Freund and Buzsaki, 1996; Sik et al., 1997; Boyett and Buckmaster, 2001). The somatostatin-positive cells all colocalize with GABA, and are the source of the somatostatin immunoreactive fibers and terminals in the outer two-thirds of the molecular layer. This system of fibers, which forms contacts on the distal dendrites of the granule cells, provides a third means for inhibitory control of granule cell activity, in addition to the GABAergic basket cell plexus and axo-axonic terminals of the chandelier cells. Since electron microscopic studies have demonstrated that the somatostatin cells are contacted by mossy fiber terminals, the projection to the outer molecular layer thus constitutes a local feedback inhibitory circuit.

17

Fig. 11. Neurons of the polymorphic region. The summary figure reprinted from Amaral (1978) illustrates the diversity of cells types within the dentate gyrus and proximal CA3 region of the rat. Based on Golgi staining and cameral lucida drawings of individual neurons from multiple preparations. See original paper for description of cell types.

Interestingly, unlike the mossy cell associational projection that terminates more heavily at distant levels of the dentate gyrus, the GABA/somatostatin projection terminates most heavily around the level of the cells of origin, and termination rapidly decreases within approximately 1.5 mm septally and temporally to the cells of origin. Thus, the mossy cell projection and the somatostatin/GABA cell projection have terminal fields that are spatially complementary in both radial and septotemporal axes. The distribution suggests that the two cell types mediate distal excitation and local inhibition, respectively. There are also multipolar or triangular cells in the polymorphic layer with thin, aspiny dendrites that extend both within the hilus and within the molecular layer. The axons of these HICAP cells (hilar commissural-associational pathway related cells) extend through the granule cell layer and branch profusely in the inner third of the molecular layer. There is a variety of other neuron types

in the polymorphic layer of the dentate gyrus whose axonal plexus have not yet been well described. A summary of the neurons in the polymorphic region, to some extent still incompletely understood, is shown in Fig. 11. Extrinsic afferents Afferent projections Although this topic is covered in more details in other chapters (see Chapters 3 and 4), we will provide a brief general synopsis of the afferent projections of the dentate gyrus. Entorhinal cortex projection to the dentate gyrus The dentate gyrus receives its major input from the entorhinal cortex, via the so-called perforant pathway (Ramo´n y Cajal, 1893). The projection to the

18

dentate gyrus arises mainly from cells located in layer II of the entorhinal cortex (Fig. 1), although a minor component of the projection also comes from layers V and VI (Steward and Scoville, 1976; Deller et al., 1996). The entorhinal terminals are strictly confined to the outer two-thirds of the molecular layer where they form asymmetric synapses that account for nearly 85% of the total axospinous terminations (Nafstad, 1967; HjorthSimonsen and Jeune, 1972). These contacts occur primarily on the dendritic spines of granule cells, although a small number of perforant path fibers also form asymmetric synapses on the shafts of GABA-positive interneurons. The perforant pathway can be divided into two parts based on the region of origin, pattern of termination and appearance in histochemical and immunohistochemical preparations. In the rat, the two divisions have been called the lateral and medial perforant paths because they originate from the lateral and medial entorhinal areas, respectively. Interestingly, while the cells of origin for these projections look very similar and the light microscopic appearance of their projections is indistinguishable, they do demonstrate substantial histochemical and chemical neuroanatomical differences which remain largely unexplored. Perforant path fibers originating in the lateral entorhinal area terminate in the most superficial third of the molecular layer, whereas the fibers originating from the medial entorhinal area terminate in the middle third of the molecular layer (see Chapter 3 for more detail on the perforant path projection). These terminal zones are readily distinguished by the classical Timm’s stain method for visualization of heavy metals which demonstrates dense staining in the outer third of the molecular layer, a near absence of staining in the middle third and a dark staining in the inner third that is associated with the commissural/associational connection (Fig. 2). Projections from both areas of the entorhinal cortex innervate the entire transverse extent of the molecular layer. The thin axon branches (0.1 mm) in the molecular layer of the dentate gyrus show periodic varicosities with a thickness of 0.5–1.0 mm. Most of entorhinal cortex layer II spiny stellate cells project up to 2 mm in the septotemporal direction forming a sheet-like

axon arbor in the molecular layer (Tamamaki and Nojyo, 1993). While it is often assumed that the perforant path fibers from the entorhinal cortex are the only hippocampal input reaching the dentate gyrus, it is now clear that at least minor projections also arise in the presubiculum and parasubiculum (Kohler, 1985). These fibers enter the molecular layer of the dentate gyrus and ramify in a zone that is interspersed between the lateral and medial perforant path projections. The presubicular axons tend to be thicker than those from the entorhinal cortex and give rise to collaterals that take a radial course in the molecular layer. Virtually nothing is currently known about which cells these fibers innervate or what type of transmitter they use. Since the presubiculum receives the only direct input from the anterior thalamic nucleus, these fibers provide a potential link by which thalamic information could reach the dentate gyrus.

Basal forebrain inputs: projections from the septal nuclei The dentate gyrus receives relatively few inputs from subcortical structures. The most robust is the projection from the septal nuclei (Mosko et al., 1973; Swanson, 1978; Amaral and Kurz, 1985). The septal projection arises from cells of the medial septal nucleus and the nucleus of the diagonal band of Broca. Septal fibers heavily innervate cells of the polymorphic layer, particularly in a narrow region just subjacent to the granule cell layer. Septal fibers are lightly distributed throughout the molecular layer. A major portion of the fibers of the septal projection to the dentate gyrus are cholinergic. Thirty to fifty percent of the cells in the medial septal nucleus and 50–75% of the cells in the nucleus of the diagonal band that project to the hippocampal formation are cholinergic. Many of the other septal cells that project to the dentate gyrus are GABAergic. The most interesting facet of this heterogeneous septal projection is that the cholinergic and GABAergic components target different cell types. Fibers of the septal GABAergic projections terminate preferentially on other GABAergic

19

nonpyramidal cells, such as the basket pyramidal cells of the dentate gyrus, and they form symmetrical, presumably inhibitory contacts. The heaviest GABAergic septal termination is on interneurons located in the polymorphic layer. The cholinergic septal projection to the dentate gyrus, in contrast, terminates mainly on granule cells, making asymmetric, presumably excitatory contacts on dendritic spines, chiefly in the inner third of the molecular layer. Only 5–10% of the septal cholinergic synapses are on dentate interneurons. The large mossy cells are innervated by cholinergic fibers (Lu¨bke et al., 1997).

Supramammillary and other hypothalamic inputs The major hypothalamic projection to the dentate gyrus arises from a population of large cells in the supramammillary area that caps and partially surrounds the medial mammillary nuclei ( Wyss et al., 1979; Dent et al., 1983; Vertes, 1993; Magloczky et al., 1994). The supramammillary projection mainly terminates in a narrow zone located just superficial to the granule cell layer and lightly in the polymorphic layer or the remaining portion of the molecular layer. The vast majority of the supramammillary fibers terminate on the proximal dendrites of granule cells. This projection is excitatory and is likely using glutamate as a primary neurotransmitter (Kiss et al., 2000). Most, but not all, of the glutamatergic supramammillary neurons that project to the dentate gyrus also colocalize calretinin; some of these cells also colocalize substance P (Borhegyi and Leranth, 1997).

Brainstem inputs The dentate gyrus receives a particularly prominent noradrenergic input from the nucleus locus coeruleus (Pickel et al., 1974; Swanson and Hartman, 1975; Loughlin et al., 1986). The noradrenergic fibers terminate mainly in the polymorphic layer of the dentate gyrus and extend into the stratum lucidum of CA3, as if preferentially terminating in the zones occupied by mossy fibers.

The dentate gyrus receives a minor and diffusely distributed dopaminergic projection that arises mainly from cells located in the ventral tegmental area. The dopaminergic fibers terminate mainly in the polymorphic layer. The serotonergic projection that originates from median and dorsal divisions of the raphe nuclei also terminates most heavily in the polymorphic layer in an immediately subgranular portion of the layer (Conrad et al., 1974; Moore and Halaris, 1975; Ko¨hler and Steinbusch, 1982; Vertes et al., 1999). A number of GABAergic interneurons appear to be preferentially innervated by the serotonergic fibers. The targets are often the pyramidal basket cells. Fusiform neurons in the region, particularly those that are stained for the calcium binding protein calbindin, are also very heavily innervated. As with the cholinergic projection from the septum, many of the cells in the raphe nuclei that project to the hippocampal formation appear to be nonserotonergic, but their transmitter is not known.

Conclusions In the remainder of this book, the reader will receive far greater detail on many of the neuroanatomical features of the dentate gyrus that were only briefly touched on in this chapter. Every effort will also be made to correlate the neuroanatomy of the dentate gyrus with both its electrophysiological and functional attributes. We will close with a short reflection on the position of the dentate gyrus both within the hippocampal formation and within the brain at large. It has been fashionable at times to highlight the simple neuroanatomy of the dentate gyrus and to claim that it provides a heuristic for studying and understanding the much more complicated neocortex. This is probably a misguided strategy. There are numerous features of the dentate gyrus that make it absolutely unique from a neuroanatomical point of view and thus presumably from a functional point of view as well. The largely unidirectional nature of its inputs and outputs is one distinctive feature. The distinctive mossy fibers that give rise to massive synaptic endings that have

20

as many as 40 active sites onto postsynaptic neurons is another. Our view is that the distinctive neuroanatomy of the dentate gyrus has been sculpted by evolution to play a precise and unique role in the hippocampal information processing that ultimately leads to the production of declarative memories. Whatever computations the dentate gyrus executes will undoubtedly be different from, and will work in concert with, the computations carried out by field CA3 and other portions of the hippocampal formation. The challenge for the future will be to understand enough about the input/output patterns of the dentate gyrus to identify the specific contribution it makes to memory formation. There are other unique features of the dentate gyrus. Why, for example, is there the conspicuous production during adult life of new granule cells and how do these support the unique function of the dentate gyrus? Does the formation of these new neurons have significance for the broader questions of stem cell recovery of function or does it reflect again the unique attributes of one particular brain region? Despite the ‘‘simplicity’’ of the dentate gyrus, it will undoubtedly remain the topic of extensive research for decades to come. And, the chapters that constitute the remainder of this book will provide the reader both with a valuable summary of the state of dentate gyrus science and guidance for research into the future.

References Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Amaral, D.G. and Kurz, J. (1985) An analysis of the origins of the cholinergic and noncholinergic septal projections to the hippocampal formation of the rat. J. Comp. Neurol., 240: 37–59. Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182: 851–914. Amaral, D.G. and Lavenex, P. (2007) Hippocampal neuroanatomy. In: Andersen P., Morris R., Amaral D., Bliss T. and O’Keefe J. (Eds.), The Hippocampus Book. Oxford University Press, New York, p. 872.

Amaral, D.G. and Witter, M.P. (1989) The three dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31: 571–591. Bakst, I., Avendan˜o, C., Morrison, J.H. and Amaral, D.G. (1986) An experimental analysis of the origins of the somatostatin immunoreactive fibers in the dentate gyrus of the rat. J. Neurosci., 6: 1452–1462. Blackstad, T.W., Brink, K., Hem, J. and Jeune, B. (1970) Distribution of hippocampal mossy fibers in the rat. An experimental study with silver impregnation methods. J. Comp. Neurol., 138: 433–450. Borhegyi, Z. and Leranth, C. (1997) Distinct substance P- and calretinin-containing projections from the supramammillary area to the hippocampus in rats — a species difference between rats and monkeys. Exp. Brain Res., 115: 369–374. Boyett, J.M. and Buckmaster, P.S. (2001) Somatostatinimmunoreactive interneurons contribute to lateral inhibitory circuits in the dentate gyrus of control and epileptic rats. Hippocampus, 11: 418–422. Buckmaster, P.S. and Amaral, D.G. (2001) Intracellular recording and labeling of mossy cells and proximal CA3 pyramidal cells in macaque monkeys. J. Comp. Neurol., 430: 264–281. Buckmaster, P.S., Strowbridge, B.W., Kunkel, D.D., Schmiege, D.L. and Schwartzkroin, P.A. (1992) Mossy cell axonal projections to the dentate gyrus molecular layer in the rat hippocampal slice. Hippocampus, 2: 349–362. Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwartzkroin, P.A. (1996) Axon arbors and synaptic connections of hippocampal mossy cells in the rat in vivo. J. Comp. Neurol., 366: 271–292. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus. J. Comp. Neurol., 246: 435–458. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1990) A quantitative three-dimensional analysis of granule cell dendrites in the rat dentate gyrus. J. Comp. Neurol., 302: 206–219. Conrad, L.C.A., Leonard, C.M. and Pfaff, D.W. (1974) Connections of the median and dorsal raphe nuclei in the rat: an autoradiographic and degeneration study. J. Comp. Neurol., 156: 179–206. Deller, T., Martinez, A., Nitsch, R. and Frotscher, M. (1996) A novel entorhinal projection to the rat dentate gyrus — direct innervation of proximal dendrites and cell bodies of granule cells and gabaergic neurons. J. Neurosci., 16: 3322–3333. Dent, J.A., Galvin, N.J., Stanfield, B.B. and Cowan, W.M. (1983) The mode of termination of the hypothalamic projection to the dentate gyrus: An EM autoradiographic study. Brain Res., 258: 1–10. Desmond, N.L. and Levy, W.B. (1985) Granule cell dendritic spine density in the rat hippocampus varies with spine shape and location. Neurosci. Lett., 54: 219–224. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus [Review]. Hippocampus, 6: 347–470. Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E. (1991) The mossy cells of the fascia dentata: a comparative

21 study of their fine structure and synaptic connections in rodents and primates. J. Comp. Neurol., 312: 145–163. Gaarskjaer, F.B. (1978a) Organization of the mossy fiber system of the rat studied in extended hippocampi. II. Experimental analysis of fiber distribution with silver impregnation methods. J. Comp. Neurol., 178: 73–88. Gaarskjaer, F.B. (1978b) Organization of the mossy fiber system of the rat studied in extended hippocampi. I. Terminal area related to number of granule and pyramidal cells. J. Comp. Neurol., 178: 49–72. Gaarskjaer, F.B. (1981) The hippocampal mossy fiber system of the rat studied with retrograde tracing techniques. Correlation between topographic organization and neurogenetic gradients. J. Comp. Neurol., 203: 717–735. Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 395–410. Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. J. Comp. Neurol., 144: 215–232. Kiss, J., Csaki, A., Bokor, H., Shanabrough, M. and Leranth, C. (2000) The supramammillo-hippocampal and supramammillo-septal glutamatergic/aspartatergic projections in the rat: a combined. Neuroscience, 97: 657–669. Kohler, C. (1985) Intrinsic projections of the retrohippocampal region in the rat brain. I. The subicular complex. J. Comp. Neurol., 236: 504–522. Ko¨hler, C. and Steinbusch, H. (1982) Identification of serotonin and non-serotonin-containing neurons of the mid-brain raphe projecting to the entorhinal area and the hippocampal formation. A combined immunohistochemical and fluorescent retrograde tracing study in the rat brain. Neuroscience, 7: 951–975. Laurberg, S. and Sorensen, K.E. (1981) Associational and commissural collaterals of neurons in the hippocampal formation (hilus fasciae dentate and subfield CA3). Brain Res., 212: 287–300. Lorente, de No´ R. (1934) Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system. J. Psychol. Neurol., 46: 113–177. Loughlin, S.E., Foote, S.L. and Bloom, F.E. (1986) Efferent projections of nucleus locus coeruleus: topographic organization of cells of origin demonstrated by three-dimensional reconstruction. Neuroscience, 18: 291–306. Lu¨bke, J., Deller, T. and Frotscher, M. (1997) Septal innervation of mossy cells in the hilus of the rat dentate gyrus — an anterograde tracing and intracellular labeling study. Exp. Brain Res., 114: 423–432. Magloczky, Z., Acsady, L. and Freund, T.F. (1994) Principal cells are the postsynaptic targets of supramammillary afferents in the hippocampus of the rat. Hippocampus, 4: 322–334. McLardy, T. and Kilmer, W.L. (1970) Hippocampal circuitry. Am. Psychol., 25: 563–566.

Moore, R.Y. and Halaris, A.E. (1975) Hippocampal innervation by serotonin neurons of the midbrain raphe in the rat. J. Comp. Neurol., 164: 171–184. Morrison, J.H., Benoit, R., Magistretti, P.J., Ling, N. and Bloom, F.E. (1982) Immunohistochemical distribution of prosomatostatin-related peptides in hippocampus. Neurosci. Lett., 34: 137–142. Mosko, S., Lynch, G. and Cotman, C.W. (1973) The distribution of septal projections to the hippocampus of the rat. J. Comp. Neurol., 152: 163–174. Nafstad, P.H.J. (1967) An electron microscope study on the termination of the perforant path fibres in the hippocampus and the fascia dentata. Zeitsch. Zellforsch. Mikrosk. Anat., 76: 532–542. Pickel, V.M., Segal, M. and Bloom, F.E. (1974) A radioautographic study of the efferent pathways of the nucleus locus coeruleus. J. Comp. Neurol., 155: 15–42. Ramo´n y Cajal, S. (1893) Estructura del asta de Ammon y fascia dentata. Ann. Soc. Esp. Hist. Nat., 22. Rapp, P.R. and Gallagher, M. (1996) Preserved neuron number in the hippocampus of aged rats with spatial learning deficits. Proc. Natl. Acad. Sci. U.S.A., 93: 9926–9930. Ribak, C.E. (1992) Local circuitry of GABAergic basket cells in the dentate gyrus. Epilepsy Res. Suppl., 7: 29–47. Ribak, C.E. and Seress, L. (1983) Five types of basket cell in the hippocampal dentate gyrus: a combined Golgi and electron microscopic study. J. Neurocytol., 12: 577–597. Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The development, ultrastructure and synaptic connections of the mossy cells of the dentate gyrus. J. Neurocytol., 14: 835–857. Ribak, C.E., Seress, L., Peterson, G.M., Seroogy, K.B., Fallon, J.H. and Schmued, L.C. (1986) A GABAergic inhibitory component within the hippocampal commissural pathway. J. Neurosci., 6: 3492–3498. Ribak, C.E., Vaughn, J.E. and Saito, K. (1978) Immunocytochemical localization of glutamic acid decarboxylase in neuronal somata following colchicine inhibition of axonal transport. Brain Res., 140: 315–332. Scharfman, H.E. (1994) Evidence from simultaneous intracellular recordings in rat hippocampal slices that area CA3 pyramidal cells innervate dentate hilar mossy cells. J. Neurophysiol., 72: 2167–2180. Scharfman, H.E. (1995) Electrophysiological evidence that dentate hilar mossy cells are excitatory and innervate both granule cells and interneurons. J. Neurophysiol., 74: 179–194. Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule cells are normal features of the fetal and adult dentate gyrus of both monkey and human hippocampal formations. Brain Res., 405: 169–174. Seress, L. and Pokorny, J. (1981) Structure of the granular layer of the rat dentate gyrus. A light microscopic and Golgi study. J. Anat., 133: 181–195. Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in the hippocampal dentate gyrus — an in vivo intracellular study. Eur. J. Neurosci., 9: 573–588.

22 Soriano, E. and Frotscher, M. (1989) A GABAergic axo-axonic cell in the fascia dentata controls the main excitatory hippocampal pathway. Brain Res., 503: 170–174. Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat fascia dentata are glutamate-immunoreactive. Hippocampus, 4: 65–69. Steward, O. and Scoville, S.A. (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol., 169: 347–370. Struble, R.G., Desmond, N.L. and Levy, W.B. (1978) Anatomical evidence for interlamellar inhibition in the fascia dentata. Brain Res., 152: 580–585. Swanson, L.W. (1978) The anatomical organization of septohippocampal projections. In: Gray A. (Ed.), Functions of the Septo-Hippocampal System. Elsevier North Holland, Amsterdam, pp. 25–48. Swanson, L.W. and Hartman, B.K. (1975) The central adrenergic system. An immunofluorescence study of the location of cell bodies and their efferent connections in the rat utilizing dopamine-beta-hydroxylase as a marker. J. Comp. Neurol., 163: 467–506. Swanson, L.W., Wyss, J.M. and Cowan, W.M. (1978) An autoradiographic study of the organization of intrahippocampal

association pathways in the rat. J. Comp. Neurol., 181: 681–716. Tamamaki, N. and Nojyo, Y. (1993) Projection of the entorhinal layer-II neurons in the rat as revealed by intracellular pressure-injection of neurobiotin. Hippocampus, 3: 471–480. Vertes, R.P. (1993) PHA-L analysis of projections from the supramammillary nucleus in the rat. J. Comp. Neurol., 326: 595–622. Vertes, R.P., Fortin, W.J. and Crane, A.M. (1999) Projections of the median raphe nucleus in the rat [Review]. J. Comp. Neurol., 407: 555–582. Walker, M.C., Ruiz, A. and Kullmann, D.M. (2002) Do mossy fibers release GABA? Epilepsia, 43: 196–202. West, M.J., Slomianka, L. and Gundersen, H.J.G. (1991) Unbiased stereological estimation of the total number of neurons in the subdivisions of the rat hippocampus using the optical fractionator. Anat. Rec., 231: 482–497. Wyss, J.M., Swanson, L.W. and Cowan, W.M. (1979) Evidence for an input to the molecular layer and the stratum granulosum of the dentate gyrus from the supramammillary region of the hypothalamus. Anat. Embryol., 156: 165–176.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 2

Comparative anatomy of the hippocampal dentate gyrus in adult and developing rodents, non-human primates and humans La´szlo´ Seress Central Electron Microscopic Laboratory, Faculty of Medicine, University of Pe´cs, Szigeti str. 12, 7624 Pe´cs, Hungary

Abstract: There has been substantial progress in our understanding of the hippocampus in the past 70 years. During this time, it has become clear that the hippocampus is not an olfactory-related structure alone, but plays critical roles in other functions that do not necessarily depend on olfaction, such as learning and memory. In addition, it has become clear how important the hippocampus is to a wide variety of neurological disorders and psychiatric illness. Animal models have provided a great resource in such studies, but a frequent question is whether the data from laboratory animals is relevant to man. Keywords: hippocampus; rat; monkey; postnatal development example, the memory-related function of the hippocampal formation is preserved from rodents to primates (including humans), supporting the principle that rodent models are relevant to humans. However, in some cases, care should be taken when selecting the proper animal model. The archicortical hippocampal formation is one of the phylogenetically oldest cortical areas. The anatomical equivalent of the hippocampus and the dentate gyrus were first identified in reptiles (Lopez-Garcia and Martinez-Guijjaro, 1988), but the characteristic morphological features of dentate gyrus, appear only in mammals. The dentate gyrus has a number of distinguishing features. First, it is a peculiar part of the hippocampal formation; it caps the end of the pyramidal cell layer of Ammon’s horn, and the granule cells form a unidirectional pathway that innervates the CA3 portion of Ammon’s horn. More important from a functional perspective, all of the cortical afferents that innervate the dentate gyrus also project to the

In this chapter, a comparison is made between the most common laboratory animals, rodents, with non-human primates (monkey), and the human hippocampus. The focus is hippocampal anatomy and development. It appears that there are many differences between rodents, non-human primates, and humans; phylogenesis has its own turns and twists either in development, in neuroanatomy, or both. However, the hippocampal formation has standard features also, and therefore in some ways the data from animal models is more readily related to the human. It is essential to understand the ways the hippocampus is similar across mammals, and the ways it differs, in order to best interpret the results from studies of animal models. In addition to similarities in neuroanatomy and development, functions of the hippocampus appear to be comparable across mammals. For Corresponding author. Tel.: +36-72-536-060; Fax: +36-72-536000/1510; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63002-7

23

24

pyramidal cells of the hippocampus (CA1–3 areas). In addition, the number of principal neurons (granule cells) of the dentate gyrus is larger than the total number of pyramidal cells of Ammon’s horn (Seress, 1988). Therefore, from the structure and connections of the dentate gyrus, the following functional conclusion can be assumed: it modifies the effect of all cortical input to the hippocampus proper (CA1–3 areas). However, the function of the dentate gyrus is more complex, since lesions of the dentate gyrus result in a partial or total functional lesion of the hippocampus. It should be noted that there is a vast reserve of granule cells in the dentate gyrus, because approximately half of the number of granule cells must be lost before functional effects can be detected (Czurko´ et al., 1997; Cze´h et al., 1998, 2001). However, this is not unique in the brain, since it is known that clinical manifestations of Parkinson’s disease occurs when nigrostriatal dopamine loss reaches approximately 50% (Bernheimer et al., 1973; Fearnley and Lees, 1991; Brooks, 1998). Considering the growing complexity of the brain and the increasingly complex function of the hippocampal formation through evolution, one may raise the question what kind of changes can be observed in morphology and in afferent and efferent connections of the hippocampus between rodents and primates? Interestingly enough the hippocampus proper (CA1–3 areas) does not show much difference between experimental animals, like mouse, rat, and the different monkey strains. The neurons including principal pyramidal cells and interneurons are larger, their total number increases and instead of the almost one-cell layer thick pyramidal cell layer found in rodents, pyramidal cells form an increasingly thicker multilayered structure in non-human primates and in humans. In contrast, the dentate gyrus shows many differences across species even in macroscopic evaluation: general cytoarchitecture, cell numbers, anatomical features of its principal cells and local circuit neurons, neurochemical content of neurons, afferent and efferent connections, and in development. This chapter will discuss these morphological differences between the rodent and primate dentate gyrus, which are likely to play an

important role in different levels of functional complexity between rodents and primates.

Granule cells Granule cells form the majority of principal cells of the dentate gyrus, composing 90% of the total number of neurons, which is similar for all mammalian species. The number of local circuit neurons is exceptionally low in the dentate gyrus, being less than 10% (Cze´h et al., 2005). The number of hilar neurons is approximately 1/20th of the number of granule cells. In mouse and rats, the total number of granule cells varies between 600,000 and 1 million (Boss et al., 1985; Seress, 1988), whereas in non-human primates the number is approximately 5 millions (Seress, 1988). In humans, depending on the method used to count cells, the numbers vary between 9 and 18 millions (Seress, 1988; West and Gundersen, 1990). Therefore, it is obvious that number of granule cells increases with the growing brain size, although this increase is not proportional. The weight of the rat brain is 2 g and the average brain weight of the human is around 1400 g, which is a 700-fold increase, whereas the granule cell number increases by at most 20-fold. This approximately corresponds to the weight increase, because the wet weight of an isolated hippocampus of the rat is approximately 0.1 g, whereas the weight of a human hippocampus is approximately 2 g. Granule cells of the rodent dentate gyrus exhibit relatively uniform morphological features. They have a small cell body (on average, 10 mm in diameter), a characteristic unipolar dendritic tree, and significant variations in dendritic length (Claiborne et al., 1990). The total dendritic length of granule cells found in the suprapyramidal blade is significantly larger than those of the infrapyramidal blade (Amaral et al., 1990). Somata of granule cells form a compact layer, the granule cell layer of the dentate gyrus (Fig. 1A). Their apical dendrites extend into a molecular layer of even width along the septo-temporal extent of the hippocampus. Basal dendrites were not found on mature granule cells of the rodent dentate gyrus, except in young rats of 5–10 days of age (Seress and Pokorny, 1981;

25

Fig. 1. Photomicrographs of Timm-stained sections of the dentate gyrus (A and B) and Ammon’s horn (C) in adult rat. Timm-stained fibers are evenly distributed throughout the hilus (A), whereas they form a bundle in stratum lucidum of the CA3 area (A and C). Timm-stained puncta outline dendrites inside the granule cell layer (g) in B. Timm-stained mossy fibers terminate at the border of CA3/ CA2 areas (arrow), but a few Timm-labeled terminals (curved open arrows) are in the pyramidal layer (p) of CA2 area. There are no Timm-labeled terminals in the CA1 area. Double arrows mark the border between the CA2 and CA1 areas (C). Scale bar ¼ 100 mm for A, 20 mm for B and C. (See Color Plate 2.1 in color plate section.)

26

Spigelman et al., 1998), meaning that basal dendrites of granule cells in rodents can be considered as transitional developmental features. In contrast, in primates the granule cells and the width of granule cell- and molecular layers are quite variable (Fig. 2A). In both the monkey and the human dentate gyrus, the granule cell layer

exhibits several deep convolutions, accompanied with significant variations in the thickness of the molecular layer (Fig. 2A). As a consequence, there is an obvious difference in the length of apical dendrites among granule cells, depending on their location in the convoluted granule cell layer (for more details, see Seress and Frotscher, 1990;

Fig. 2. Photomicrographs of Timm-stained sections of the dentate gyrus (A and B) and Ammon’s horn (C and D) in adult rhesus monkey. The hilus is outlined by the Timm-stained fibers (A). Timm-stained puncta outline dendrites (arrows) in the granule cell layer (g), and a few puncta (small arrows) are always found in the molecular layer (B). Timm-stained fibers occupy not only stratum lucidum (lm) but also the entire width of the pyramidal layer (p) of the CA3 area (C). In the pyramidal layer (p) and in stratum lucidum (lm) Timm-stained puncta clearly outline dendrites (C). In stratum oriens (o) some puncta form clusters (arrows) (D). Scale bar ¼ 200 mm for A, 20 mm for B, 100 mm for C, 20 mm for D. (See Color Plate 2.2 in color plate section.)

27

Seress, 1992). The width of the granule cell layer also varies, and is formed by 4–20 rows of granule cells. The width of the molecular layer can vary from 150 to 1000 mm (Fig. 2A). An additional difference is that a subpopulation of granule cells displays thicker dendrites and higher spine density

than the others, even if the dendritic length is similar (Seress and Frotscher, 1990). In addition to characteristics of rat granule cells that are similar to those in monkeys and in humans (Fig. 3A and D), there are differences, and one is that a large subpopulation of granule

Fig. 3. Photomicrographs of Golgi-impregnated granule cells without basal dendrites (A and D) and with basal dendrites (B, C and D) in the dentate gyrus of adult rhesus monkey. Some basal dendrites (arrows) branch close to the cell body (B), and others extend into the hilus (arrows on C and white arrows on D). Morphology of dendritic tree of conventional granule cells differs depending on the location of the soma within the granule cell layer. Those cells at the hilar border (h) have one main dendrite and narrower dendritic tree (open curved arrow on D), whereas those at the border with the molecular layer have multiple dendrites that arise from the soma (open arrowhead on D). Basal dendrites are similarly covered by spines as the apical dendrites (C). Axons usually originate from the basal pole of the soma (A and D). Two Golgi-impregnated pyramidal cells in the pyramidal layer (p) of the CA3 area display thorny excrescences (arrows) on both the apical and basal dendrites (E). Scale bar ¼ 20 mm. (See Color Plate 2.3 in color plate section.)

28

cells in primates display basal dendrites (Fig. 3B and C). Basal dendrites enter the hilus, whereas recurrent basal dendrites originate at the basal pole of the granule cells and turn back toward the molecular layer by crossing the granule cell layer (Seress and Mrzljak, 1987; Seress and Frotscher, 1990). The number of granule cells with basal dendrites is higher in the human than in the monkey dentate gyrus. Golgi preparations of granule cells have shown that the morphological features of the basal dendrites, such as dendritic branching and spine density (Fig. 3B and C) are similar to those of apical dendrites (Seress and Mrzljak, 1987; Seress and Frotscher, 1990). The frequency of granule cells with basal dendrites is difficult to estimate, because Golgi-impregnation only reveals a randomly-selected group of neurons. Granule cells are well defined by the calcium binding protein, calbindin, in both the rodent and primate hippocampus. Usually, calbindin immunohistochemistry reveals larger numbers of granule cells with basal dendrites in the human than in the monkey dentate gyrus. However, distribution of granule cells with basal dendrites is not uniform throughout the granule cell layer; cells with basal dendrites appear to accumulate in certain parts of the dentate gyrus, such as the crests of the granule cell layer (Fig. 3D). Basal dendrites are involved in the excitatory circuitry that is formed among granule cells in non-epileptic primates (Seress and Frotscher, 1990). However, it was described that such cells are found in higher numbers in the dentate gyrus of the epileptic patients (Scheibel and Scheibel, 1973). Therefore, similar to rodent models of epilepsy, where basal dendrites appear and form synapses with the axons of granule cells, the large number of basal dendrites in humans may play a role in epileptogenesis (Spigelman et al., 1998; Ribak et al., 2000). Regarding the ultrastructure of granule cells, there are a few minor differences between the rodents and primates. Somata of primate granule cells display somatic spines and lower number of axosomatic synapses, both in non-human primates and in humans (Seress and Ribak, 1992; Seress et al., 2004b).

Mossy cells Mossy cells populate the hilus of the dentate gyrus exclusively. Mossy cells were first described by Amaral (1978) in rodents, by Frotscher et al. (1991) in non-human primates and by Seress and Mrzljak (1992) in humans, and are large, densely spiny neurons with characteristic large complex spines on their somata and main dendrites, ‘‘baptized’’ by Amaral (1978) as thorny excrescences. The major excitatory input of mossy cells is from granule cells, and in turn, their major target neuronal group is also the granule cells (Ribak et al., 1985; Amaral and Witter, 1989). However, it has been shown that mossy cells are likely to also innervate local circuit neurons (Sloviter et al., 2006), a topic that has previously been emphasized for its potential importance by Sloviter (1991). In terms of the quantity of axonal connections, however, it appears that granule cells and mossy cells form a mainly excitatory circuitry inside the dentate gyrus. This recurrent excitatory circuit is not, however, to the same cells: granule cells are not innervated by a mossy cell from which they receive afferents (Amaral and Witter, 1989). Morphology of mossy cells differs between rodents and primates in two respects: (i) Thorny excrescences of mossy cells are much larger in monkeys than in rodents, and even larger in humans than in monkeys (Fig. 4A and B) (Frotscher et al., 1991; Seress and Mrzljak, 1992). (ii) Dendrites of mossy cells respect the borders of the hilus, and only a few penetrate the granule cell layer, in rodents as revealed with Golgi studies (Ribak et al., 1985). However, with more modern intracellular labeling techniques it was demonstrated that a proportion of mossy cells have dendrites penetrating the molecular layer (Scharfman et al., 2001). In monkeys and humans, dendrites of most mossy cells penetrate the granule cell layer (Fig. 4A), and reach the upper third of the molecular layer (Frotscher et al., 1991; Seress and Mrzljak, 1992). The dendritic segment inside the granule cells layer is aspiny, whereas the dendritic portion inside the molecular layer is similarly covered with spines as the proximal dendritic segments in the hilus (Fig. 4C and D). The difference in dendritic arbor suggests that mossy cells

29

Fig. 4. Photomicrographs of Golgi-impregnated mossy cells in the hilus of the dentate gyrus of adult rhesus monkey (A) and humans (B). Thorny excrescences are considerably larger on the human mossy cells than in rats and monkey. In primates, dendrites of many mossy cells penetrate the molecular layer (m) (A). The dendrite (arrows) is covered by spines in the molecular layer (m), but the segment inside the granule cell layer (g) is practically spine free (C). The proximal dendrites and soma (arrow) of the hilar mossy cell is covered by thorny excrescences (D). Scale bar ¼ 100 mm for A, 50 mm for B, 25 mm for C and D. (See Color Plate 2.4 in color plate section.)

of the primate dentate gyrus receive afferents from the perforant path, while such connections of mossy cells are rare or absent in rodents. It is interesting to note that spine density and morphology of individual spines do not differ across these species, both for the dendritic segments in the hilar region and in the molecular layer, although afferents that form synapses with those spines are different. One possible reason is that the size and morphology of presynaptic boutons determine the shape and size of the

postsynaptic spines, independent of the possible neurochemical content and the postsynaptic receptor structure. It is likely that the diameter of the dendrites of mossy cells and not only the large mossy terminals influence spine generation, because large mossy fiber terminals in primates connect only thorny excrescences of large diameter dendrites, whereas giant mossy fiber boutons innervate the shaft of the distal dendrites forming exclusively conventional spines (Frotscher et al., 1991).

30

Pyramidal cells of Ammon’s horn show few of the differences between rodents and primates that have been described above for mossy cells. This point supports the argument that mossy cells are not simply modified pyramidal cells, although they do have similarities to the CA3 pyramidal cells. The discussion of Frotscher et al. (1991) strengthen this notion: (i) unlike CA3 pyramidal cells, mossy cells have no septal efferent connections or Schaffer collaterals; (ii) CA3 pyramidal cells receive their dendritic afferents in a laminated fashion, whereas mossy cells not; (iii) mossy cells receive input from granule cells to all parts of their dendrites, whereas CA3 pyramidal cells only are innervated by mossy fibers within stratum lucidum in rats; in primates, there is an additional input to the proximal part of their basal dendrites, but never on the distal dendrites; (iv) electrophysiological properties of CA3 pyramidal cells differ from those of mossy cells (Scharfman and Schwartzkroin, 1988; Scharfman, 1993). In addition, a fifth argument can be added regarding the neurochemical differences between mossy cells and CA3 pyramidal cells in both primates and rodents. Mossy cells in the human and monkey dentate gyrus contain cocaine- and amphetamine-regulated transcript peptide (Figs. 5 and 6), whereas CA3 pyramidal neurons do not (Seress et al., 2004a; A´braha´m et al., 2005). In the rat, mossy cells are immunoreactive for calcitonin gene-related peptide but not the CA3 pyramidal cells (Freund et al., 1997). Similarly, the pyramidal cells of the CA1–3 areas of Ammon’s horn express neurogranin, but the mossy cells of the hilus are neurogranin negative (Singec et al., 2004).

Local circuit neurons Morphology of local circuit neurons of the dentate gyrus appears to be stable through the mammalian evolution. Similar subpopulations of local circuit neurons, namely, the axosomatic, axoaxonic and axodendritic cells, are present in primates as in rodents (Freund and Buzsa´ki, 1996). Their axonal branching fields are probably similar between

primates and the rat, as described elsewhere (Halasy and Somogyi, 1993). Only one difference in the local circuit neurons of rat and primate has been described so far: axodendritic, calretinin-positive interneurons that were found in the rat dentate gyrus (Gulya´s et al., 1996) have not been described either in monkeys or in humans. Instead, the non-human primate dentate gyrus contains calretinin-positive neurons which innervate dendrites of granule cells and local circuit neurons, as well as the somata of granule cells (Seress et al., 1993). Similarly, calretinin-positive neurons in the human hippocampus innervate somata of principal cells, not only the dendrites (Urba´n et al., 2002; Maglo´czky and Freund, 2005).

Chemical neuroanatomy Relatively few differences have been described for the neurochemical content of neurons in the rat and primate dentate gyrus. Granule cells contain calcium binding protein calbindin in both the rodent and monkey as well as in the human dentate gyrus (Freund and Buzsa´ki, 1996). However, in the rat, a subpopulation of granule cells and their axons reveal CART peptide immunoreactivity, whereas granule cells do not contain CART in primates (Seress et al., 2004a; A´braha´m et al., 2005). Perisomatic basket and axoaxonic cells contain parvalbumin or cholecystokinin (CCK), whereas most axodendritic cells express calretinin (Freund and Buzsa´ki, 1996). In the hilus of the rat dentate gyrus, large number of GABAergic axodendritic neurons express somatostatin (Katona et al., 1999). Similar cells in the monkey dentate gyrus express substance-P in addition to somatostatin (Seress and Leranth, 1996). Among the afferents of the granule cells, the parent neurons of the supramammillary terminals slightly differ in their neuropeptide content (Borhegyi and Leranth, 1997). In different rodent species, mossy cells differ in their neurochemical content, because in mouse and hamster, they express calretinin, but the same cell type in rat does not (Blasco-Ibanez and Freund, 1997; Murakawa and Kosaka, 1999). Mossy cells in the monkey dentate gyrus also express

31

Fig. 5. CART-immunoreactive mossy cells and their axons in the dentate gyrus of adult rhesus monkey. Somata of mossy cells are located in the hilus (h), whereas their axons form the associational pathway and project mainly to inner third of the molecular layer (m) of the dentate gyrus (A). Somata of mossy cells at higher magnification (B). CART-positive puncta inside the granule cell layer (g) shows the ascending axons from the hilus to the molecular layer (B). Large magnification of CART-positive mossy cell displays thorny excrescences (arrowheads) in the monkey dentate gyrus (C). Scale bar ¼ 250 mm for A, 50 mm for B, 10 mm for C. (See Color Plate 2.5 in color plate section.)

32

Fig. 6. High magnification photomicrographs of a CART-immunostained mossy cell displaying large and complex thorny excrescences in the hilus of the human dentate gyrus (A and B). Scale bar ¼ 10 mm for A and B. (See Color Plate 2.6 in color plate section.)

calretinin, at the uncal subdivision of the dentate gyrus (Sloviter et al., 1996). In contrast, mossy cells in the human dentate gyrus do not contain calretinin. It is very difficult to assume whether the differential expression of calretinin has any functional importance, but it is important to note, that neurochemical content of a neuronal type is not identical through evolution.

Afferent and efferent connections: connections between the dentate gyrus and Ammon’s horn The major connection between the two main parts of the hippocampal formation is formed by the axons of granule cells, the mossy fiber pathway. In fact, this is the only axonal pathway that provides direct granule cell input to pyramidal cells of

33

Ammon’s horn. Axon collaterals of CA3 pyramidal cells project into the hilus of the dentate gyrus, but those axons innervate hilar neurons and not granule cells (Ishizuka et al., 1990; Scharfman, 1994). This ‘‘back-projection’’ is relatively sparse compared to the mossy fiber pathway. Terminals of mossy fibers contain zinc, and it allows visualization of the mossy fiber terminals using the Timm’s stain or its modification (Timm, 1958; Danscher, 1981; Seress and Gallyas, 2000). Mossy fibers of granule cells innervate hilar neurons, and neurons of the CA3 area of Ammon’s horn (both pyramidal and non-pyramidal cells). Termination of the mossy fibers occurs at the border between the CA3 and CA2 (Fig. 1C). However, even though CA2 is not innervated by large mossy fiber terminals, Timm-labeled fibers forming small terminals can be observed in the CA2 area (Fig. 1C). There is a major difference in the mode of termination of mossy fibers on the pyramidal cells of CA3 area in rodents and primates. In rodents, mossy fibers occupy only stratum lucidum of the CA3 area (Fig. 1C), although small intra- and infrapyramidal bundles occur within the CA3c area, i.e., the portion of CA3c that is located between the blades of granule cells layers (Fig. 1A). In contrast, the mossy fiber bundle in primates spans the entire width of the pyramidal cell layer, stratum lucidum, and also the proximal dendritic zone of the basal dendrites of CA3 pyramidal cells (Fig. 2C and D). Therefore, both apical and basal dendrites of pyramidal neurons are postsynaptic targets of the large mossy fiber terminals in the monkey and human hippocampus (Fig. 2C). In accordance with the distribution of mossy fibers, the thorny excrescences of pyramidal cells are present only on apical dendrites of the rodent CA3 pyramidal neurons, whereas thorny excrescences are common on both the apical and basal dendrites of the primate pyramidal cells (Fig. 3E). In fact, in well-stained Timm preparations, the Timm-stained puncta, representing zinc-containing large mossy fiber terminals, outline dendrites in strata lucidum and pyramidale (Fig. 2C) and form large groups of terminals in stratum oriens (Fig. 2D) of the primate hippocampus.

In the dentate gyrus, a few mossy fibers cross the granule cell layer (Figs. 1B and 2B), following dendrites of local circuit neurons in adult rodents and primates (Seress, 1992; Blasco-Ibanez et al., 2000; Seress et al., 2001a, b). However, they do not innervate the apical dendrites and somata of granule cells, because Timm-labeled axon terminals did not establish synapses either with somata or apical dendrites of granule cells in control rats (Seress et al., 2001a). In the hilus, mossy fibers form synapses with spines of the basal dendrites of granule cells of non-human primates (Frotscher et al., 1991). In contrast to a larger postsynaptic target zone of excitatory associational connections of granule cells, the commissural connection between the two sides of the dentate gyrus is reduced in primates. In rodents, the hilar mossy cells have strong commissural connections and terminate directly on granule cells and local circuit neurons (Seress and Ribak, 1984; Ribak et al., 1985; Frotscher et al., 1991). In contrast, the hilar commissural connection is almost completely absent in primates (Amaral et al., 1984). Mossy cells form extensive local axonal branches in the hilus in both rodents (Swanson et al., 1981) and primates (Frotscher et al., 1991; Seress and Ribak, 1995b), but commissural connections are minimal in monkeys (Amaral et al., 1984). Electrophysiological evidence indicates the lack of a direct commissural connection between the two sides of dentate gyrus in humans (Wilson et al., 1991). Although hypothetical, this anatomical difference may represent the increasing strength of hippocampal associational connections in primates, and may be a sign of lateralization, similar to cortical lateralization, and could be relevant to memory-related lateralization in primates. Non-hippocampal afferent connections to the dentate gyrus (entorhinal, septal, supramammillary) develop perinatally or early postnatally in rodents (Supe´r and Soriano, 1994; Ceranic et al., 2000), whereas the equivalent afferent connections are established in the first half of gestation in primates. In humans, septal connections are probably the first afferent to the hippocampus, arriving around the 15th gestational week (Kostovic et al., 1989), whereas entorhinal and

34

supramammillary connections are established by midgestation (Hevner and Kinney, 1996; Berger et al., 2001).

Comparative developmental features of the rodent and primate dentate gyrus Neuronal development is a complex process consisting of cell proliferation, differentiation, migration and positioning, dendritic maturation, and synapse formation. As it is described in the following paragraphs, there are species differences even at the first step of development of the dentate gyrus, i.e., in the generation of neurons. It should be emphasized that species differences in development are not necessarily analogous to evolutionary differences. In addition, the developmental differences between rodents and primates are not completely explained merely by the development of a larger structure in primates; there is more structural complexity as well. In rodents, granule cells of the dentate gyrus started to be formed only 2–3 days before birth, on the 17th embryonic day. As a consequence, hilar cells are formed prenatally, whereas 85% of the dentate granule cells are formed postnatally (Angevine, 1975; Bayer, 1980). Importantly, generation of granule cells continues at a low level into adulthood. It is interesting to note, however, that adult granule cell formation does not follow the sequence observed during the embryonic and early postnatal period, since new cells are generated with no obvious pattern in both blades of the dentate gyrus (Kempermann et al., 1997a, b). In contrast, neurogenesis in the non-human primate dentate gyrus takes place relatively early in prenatal development. In Rhesus monkeys, the first neurons appear almost simultaneously in the different subregions of the hippocampal formation, from the entorhinal cortex to the dentate gyrus, between embryonic days 36 and 38 (Rakic and Nowakowski, 1981). Except for the dentate gyrus, cell formation ceases during the first half of pregnancy, between embryonic days 62 and 65. Granule cell formation lasts until the end of the first postnatal month, with only 15% of the granule cells being formed postnatally (Rakic and

Nowakowski, 1981). Thus, granule cell formation is a prenatal event in the non-human primates, with its peak at the middle and last thirds of embryonic development. Proliferation of granule cells continues in the dentate gyrus of adult monkeys, although to a very limited extent (Gould et al., 1999; Kornack and Rakic, 1999). In humans, the entorhinal cortex, subiculum, and Ammon’s horn, are discernable at the 10th gestational week, whereas the dentate granule cell layer appears at the 11.5th gestational week (Humphrey, 1967). In the early descriptions of the hippocampal fissure, it was mentioned that hippocampal formation and the future dentate gyrus form a straight line. The dentate gyrus begins at the peak and starts to bend away from the ventricular wall at the 12th gestational week (Hines, 1922; Humphrey, 1967). With the aid of the mitotic marker Ki-67 (MIB-1), it has been shown that both germinative matrices, the ventricular zone and hilus, contain a large number of proliferating cells as early as the 14th gestational week. Until around midgestation, groups of dispersed MIB-1 positive cells occur close to the ventricular wall; however, the major stream of the proliferating cells locates alongside the CA3 pyramidal cells. Later, the major site of cell proliferation is shifted to the hilar region, and proliferating cells disappear from the future stratum oriens of the CA3 area, and the ventricular layer becomes thin and contains few dividing cells. In newborns, the ventricular germinal layer contains 5–6 cell layers of loosely packed cells, and rarely displays MIB-1 immunoreactivity. Thus, in the second trimester of pregnancy, the hilus of the dentate gyrus has high rate of cell proliferation; however, the number of dividing cells in the hilus decreases rapidly after the 24th gestational week, and less than 0.1% MIB-1 positive cells can be observed during the first six postnatal months (Seress, 2001). Only a proportion of the labeled cells might be neuronal precursors, whereas the others are neuroglial and endothelial cells (Seress et al., 2001; A´braha´m et al., 2004). In conclusion, granule cell formation in the human dentate gyrus is a very early phenomenon starting around the 11th gestational week, and it sharply decreases in the second half of the embryonic

35

development between the 24th and 28th gestational weeks. It has been demonstrated that granule cell proliferation is long-lasting in the dentate gyrus of the human hippocampus (Eriksson et al., 1998). However, in surgically removed adult hippocampi from individuals with temporal lobe epilepsy, we failed to find more than a few dividing cells. The seizures did not appear to be the explanation because the same observations were made in specimens resected because of a benign tumor. We only occasionally found MIB-1-labeled granule cells in the dentate gyrus of children who were older than 1 year, suggesting that a few neurons may preserve their capability for proliferation (Seress et al., 2001b). In rodents, virtually all cortical inhibitory GABAergic neurons, including interneurons of hippocampus and dentate gyrus, are generated in the ganglionic eminence and migrate to their destination via tangential migration and ventricleguided migration (Anderson et al., 1997; Nadarajah and Parnavelas, 2002). However, in humans, only 35% of the neocortical interneurons have been shown to originate from the ganglionic eminence, and 65% of GABAergic cells are formed in the neocortical ventricular and subventricular zones and migration is guided by radial glial (Letinic et al., 2002). Although it is not proven, the possibility cannot be excluded that inhibitory neurons of the dentate gyrus partly originate from the hippocampal ventricular zone and not from the ganglionic eminence (Letinic et al., 2002). Another developmental (anatomical) difference between rodents and primates is the transitional appearance of the dorsal hippocampus during ontogeny in primates, which morphologically corresponds to the rodent dorsal hippocampus (Stephan, 1975; A´braha´m et al., 2004). This transitional dorsal hippocampus includes the main parts, e.g., Ammon’s horn and the dentate gyrus. It gradually disappears in the second trimester of the human pregnancy, and its remnants become the indusium griseum. Granule cells, hilar mossy cells, and CA3 pyramidal cells are in an advanced stage of development at birth in monkeys (Seress and Ribak, 1995a, b). In monkeys, groups of small cells with thin cytoplasm and dark cell nucleus persist in the

subgranular zone of the dentate gyrus throughout the first postnatal year (Eckenhoff and Rakic, 1988). Not all of the undifferentiated, immaturelooking cells express glial fibrillary acidic protein (GFAP) in the 6-month- and 1.5-year-old monkeys, suggesting that these immature cells may be neurons, and may differentiate into granule cells later in development (Eckenhoff and Rakic, 1988). In contrast to monkeys, a large proportion of human principal cells are in an early stage of dendritic and spine development at birth (Purpura, 1975; Seress, 1992; Seress and Mrzljak, 1992). In newborn children, fully matured granule cells display varicose, stubby, short, and spineless dendrites that terminate in growth cones (Seress, 1992). Therefore, the diversity in maturation of granule cell development reported by Purpura (1975) in the 33-week-old fetus still persists at birth. A few immature-looking granule cells are still present in the 18-month-old child, but they are rare, again suggesting that the continued proliferation of granule cells in the human after birth is not striking (Seress, 1992). In contrast to monkeys (but similar to rodents) the formation of complex spines of hilar mossy cells from human occurs exclusively postnatally; no mossy cells with thorny excrescences have been observed at birth (Seress and Mrzljak, 1992). The first small, thorn-like excrescences appeared on human mossy cells by the third postnatal month. At the age of seven months, thorny excrescences were frequent on mossy cells, but their number and size continued to increase up to the 3rd year (Seress and Mrzljak, 1992). The thorny excrescences of the adult human mossy cells are much larger than the thorny excrescences of mossy cells in either rodents or monkeys (Frotscher et al., 1991). Interneurons in rodents, as well as in primates, are formed early, parallel with the proliferation of pyramidal cells of Ammon’s horn. In the rodent dentate gyrus, the development of GABAergic neurons occurs mainly postnatally. GABAergic or glutamic acid decarboxylase (GAD)-positive somata appear around birth, whereas the first identifiable basket cells, and their perisomatic synapses, appear during the first postnatal week (Seress and Ribak, 1988;

36

Fig. 7. Photomicrographs of Golgi-impregnated basket cells in the granule cell layer (g) of the dentate gyrus in newborn (A and B) and 1month-old (C) rhesus monkey. Soma of a fusiform type basket cell located in the granule cell layer with an axon (arrow) originating from the main apical dendrite (A). High magnification photomicrograph of the basket cell shows the point of origin of the axon (arrows), as well as the well-developed axonal branches among the granule cells (B). A multipolar basket cell at the hilar border (h) has dendrites that extend into the molecular layer, and an axon (arrow) with many branches that ramify within the granule cell layer. Scale bar ¼ 20 mm for A, B, C. (See Color Plate 2.7 in color plate section.)

37

Rosenberg et al., 1989; Seress et al., 1989). With the Golgi method, well-developed basket cells could only be demonstrated in 16-day-old rats, whereas with intracellular-labeling, local

circuit neurons with an extensive axonal arbor could be shown between the 7th and 9th days (Seress and Ribak, 1990; Seay-Lowe and Claiborne, 1992).

Fig. 8. Photomicrographs of parvalbumin-immunostained basket cells and their axons in the human dentate gyrus: (A) 1 month-old; (B) 8 years old; (C) 10 year old. In the dentate gyrus of a 1 month-old child, there are no parvalbumin-stained somata, but a few axonal branches were visible (A). At older ages, axons arborize among the granule cells (B) in a manner characteristic of adults (C). Scale bar ¼ 20 mm for A, B, C. (See Color Plate 2.8 in color plate section.)

38

In contrast to rodents, basket cells are well developed at birth in monkeys, as observed in Golgiimpregnated preparations (Fig. 7A–C). These local circuit neurons display mature-like apical and basal dendrites and extensive axon braches (Fig. 7B). Similarly, the neurochemical marker, parvalbumin (expressed by interneurons providing axosomatic or axoaxonic inhibition) appears very early in embryonic development. This suggests that functional circuits between local circuit and projecting neurons are formed in non-human primates in the first half of gestation (Berger and Alvarez, 1996; Berger et al., 1999). In humans, parvalbumin expression starts late in development, in contrast to the calretinin- and calbindin-containing axodendritic populations of interneurons. In the late fetal period and even at birth, parvalbumin-immunoreactive cells are not yet detectable in the human hippocampal formation. In 1-month-old infants, parvalbuminimmunoreactive cells still display poorly developed dendritic branches, but a few parvalbumin-positive axons and a few axon-terminal-like boutons have developed (Fig. 8A). In the next few months, both dendritic and axonal arborization expands. In a 2-year-old child, the morphology of the dendrites and axonal branching, as well as the terminal boutons of the parvalbumin-immunoreactive cells, are less developed than in 8–10-year-old children (Fig. 8B and C). In our experience, an 8-year-old child demonstrated cells with an adult-like morphology of parvalbumin expression. Although the limited nature of human material does not allow us to define the development of parvalbuminimmunoreactivity precisely, the data suggest that maturation of these cells is completed between the 2nd and 8th year, probably around school age similarly as the maturation of hilar mossy cells, that reach an adult-like appearance between the 3rd and 5th years of age (Seress and Mrzljak, 1992; Seress, 2001). The delayed expression of parvalbumin in the perisomatic inhibitory cells coincides with the prolonged maturation of the principal cells, i.e., the granule and mossy cells of the human dentate gyrus (Seress, 1992; Seress and Mrzljak, 1992). Similarly, the mossy fiber pathway forms during a prolonged postnatal period, although all cell types

that are involved in the connections are generated in the first half of pregnancy. There is no explanation at the present time what causes this delayed development of mossy fibers. However, this delay, and the long postnatal development of both principal and inhibitory cells, may explain the long-lasting cognitive development of children.

Abbreviations CA1–3 g h lm m o p

subregions of Ammon’s horn granule cell layer of the dentate gyrus hilus of the dentate gyrus stratum lucidum of the CA3 area of Ammon’s horn molecular layer of the dentate gyrus stratum oriens of Ammon’s horn pyramidal layer of Ammon’s horn

Acknowledgments The author wishes to thank Dr. Hajnalka A´braha´m for critically reading and commenting the manuscript. The author acknowledges Mrs. Emese Papp for her excellent technical assistance in the histological preparations of the tissue. This work was supported by the Hungarian National Science Fund (OTKA) with grant no. T047109.

References A´braha´m, H., Perez-Garcia, C.G. and Meyer, G. (2004) p73 and Reelin in Cajal-Retzius cells of the developing human hippocampal formation. Cerebr. Cortex, 14: 484–495. A´braha´m, H., Cze´h, B., Fuchs, E. and Seress, L. (2005) Mossy cells and different subpopulations of pyramidal neurons are immunoreactive for cocaine- and amphetamine-regulated transcript peptide in the hippocampal formation of nonhuman primates and tree shrew (Tupaia Belangeri). Neuroscience, 136: 231–240. Amaral, D.G. (1978) A Golgy study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182: 851–914.

39 Amaral, D.G., Insausti, R. and Cowan, M. (1984) The commissural connections of the monkey hippocampal formation. J. Comp. Neurol., 224: 307–336. Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons, numbers and the hippocampal network. In: Storm-Mathisen J., Zimmer J. and Ottersen O.P. (Eds.), Understanding the Brain Through the Hippocampus: Progress Brain Research, Vol. 83. Elsevier, Amsterdam, New York, Oxford, pp. 1–11. Amaral, D.G. and Witter, M. (1989) The three dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31: 571–591. Anderson, S.A., Eisenstat, D.D., Shi, L. and Rubenstein, J.L. (1997) Interneuron migration from basal forebrain to neocortex: dependence on Dlx genes. Science, 278: 474–476. Angevine, J.B. (1975) Development of the hippocampal formation. In: Isaacson R.L. and Pribram K.H. (Eds.), The Hippocampus. Plenum Press, New York, pp. 61–90. Bayer, S.A. (1980) The development of the hippocampal region in the rat. I. Neurogenesis examined with H3-thymidine autoradiography. J. Comp. Neurol., 190: 87–114. Berger, B. and Alvarez, C. (1996) Neurochemical development of the hippocampal region in the fetal rhesus monkey. III: calbindin-D28k, calretinin ad parvalbumin with special mention of Cajal-Retzius cells ad the retrosplenial cortex. J. Comp. Neurol., 366: 474–499. Berger, B., DeGrissac, N. and Alvarez, C. (1999) Precoccious development of parvalbumin-like immunoreactive interneurons in the hippocampal formation ad entorhinal cortex of the fetal Cynamogus monkey. J. Comp. Neurol., 403: 309–331. Berger, B., Esclapez, M., Alvarez, C., Meyer, G. and Catala, M. (2001) Human and monkey fetal brain development of the supramammillary-hippocampal projections: a system involved in the regulation of theta activity. J. Comp. Neurol., 429: 515–529. Bernheimer, H., Birkmayer, W., Hornykiewicz, O., Jellinger, K. and Seitelberger, F. (1973) Brain dopamine and the syndromes of Parkinson and Huntington. Clinical, morphological and neurochemical correlations. J. Neurol. Sci., 20: 415–455. Blasco-Ibanez, J.M. and Freund, T.F. (1997) Distribution, ultrastructure, and connectivity of calretinin-immunoreactive mossy cells of the mouse dentate gyrus. Hippocampus, 7: 307–320. Blasco-Ibanez, J.M., Martinez-Guijjaro, F.J. and Freund, T.F. (2000) Recurrent mossy fibers preferentially innervate parvalbumin-immunoreactive interneurons in the granule cell layer of the rat dentate gyrus. Neuroreport, 11: 3219–3225. Borhegyi, Z. and Leranth, C. (1997) Distinct substance-P and calretinin-containing projections from the supramammillary area to the hippocampus in rats: a species difference between rats ad monkeys. Exp. Brain Res., 115: 369–374. Boss, B., Peterson, G.M. and Cowan, W.M. (1985) On the number of neurons in the dentate gyrus of the rat. Brain Res., 338: 144–150. Brooks, D.J. (1998) The early diagnosis of Parkinson’s disease. Ann. Neurol., 44(Suppl. 1): 10–18.

Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1990) Quantitative, three-dimensional analysis of granule cell dendrites in the rat dentate gyrus. J. Comp. Neurol., 302: 206–219. Ceranic, K., Zhao, S. and Frotscher, M. (2000) development of the entorhino-hippocampal projection: Guidence by CajalRetzius cell axons. In: Scharfman, H., Witter, M.P. and Schwartz, R. (Eds.), The parahippocampal region. Annals of the New York Academy of Sciences, Vol. 911, pp. 43–54. Cze´h, B., Hajnal, A. and Seress, L. (2005) NADPH-diaphorase positive neurons of the rat hippocampal formation: Regional distribution, total number and colocalization with calcium binding proteins. Prague Med. Rep., 106/3: 261–274. Cze´h, B., Seress, L., Nadel, L. and Bures, J. (1998) Lateralized fascia dentata lesion and blockade of one hippocampus: Effect on spatial memory in rats. Hippocampus, 8: 647–650. Cze´h, B., Stuchlik, A., Wesierska, M., Cimadevilla, J.M., Pokorny, J., Seress, L. and Bures, J. (2001) Effect of dentate gyrus lesion on allothetic and idiothetic navigation in rats. Neurobiol. Learn. Mem., 74: 190–213. Czurko´, A., Cze´h, B., Seress, L., Nadel, L. and Bures, J. (1997) Severe spatial navigation deficit in the Morris water maze after single high dose of neonatal X-ray irradiation in the rat. Proc. Natl. Acad. Sci. U.S.A., 94: 2766–2771. Danscher, G. (1981) Histochemical demonstration of heavy metals. Histochemistry, 71: 1–6. Eckenhoff, M.F. and Rakic, P. (1988) Nature and fate of proliferative cells in the hippocampal dentate gyrus during life span of the rhesus monkey. J. Neurosci., 8: 2729–2747. Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M., Nordborg, C., Peterson, D.E. and Gage, F.H. (1998) Neurogenesis in the adult human hippocampus. Nat. Med., 4: 1313–1317. Fearnley, J.M. and Lees, A.J. (1991) Aging and Parkinson’s disease: Substantia nigra regional selectivity. Brain, 114: 2283–2301. Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Bubl, E. (1991) The mossy cells of the fascia dentata: A comperative study of their five structure and synaptic connections in rodents and primates. J. Comp. Neurol., 312: 145–163. Freund, T.F. and Buzsa´ki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Freund, T.F., Ha´jos, N., Acsa´dy, L., Go¨rcs, T.J. and Katona, I. (1997) Mossy cells of the rat dentate gyrus are immunoreactive for calcitonin-gene-related peptide (CGRP). Eur. J. Neurosci., 9: 1815–1830. Gould, E., Reeves, A.J., Fallah, M., Tapanat, P., Gross, C.G. and Fuchs, E. (1999) Hippocampal neurogenesis in adult Old World primates. Proc. Natl. Acad. Sci. U.S.A., 96: 5263–5267. Gulya´s, A.I., Ha´jos, N. and Freund, T.F. (1996) Interneurons containing calretinin are specialized to control other interneurons in the rat hippocampus. J. Neurosci., 16: 3397–3411. Halasy, K. and Somogyi, P. (1993) Subdivision in the multiple GABAergic innervation of granule cells in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 411–429.

40 Hevner, R.F. and Kinney, H.C. (1996) Reciprocal entorhinalhippocampal connections established by human fetal midgestation. J. Comp. Neurol., 372: 384–394. Hines, H. (1922) Studies in the growth and differentiation of the telencephalon in man. The fissura hippocampi. J. Comp. Neurol., 34: 73–170. Humphrey, T. (1967) The development of the human hippocampal fissure. J. Anat., 101: 655–676. Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization of intrahippocampal projections originating from the CA3 pyramidal cells in the rat. J. Comp. Neurol., 295: 580–623. Katona, I., Acsa´dy, L. and Freund, T.F. (1999) Postsynaptic targets of somatostatin-immunoreactive interneurons in the rat hippocampus. Neuroscience, 88: 37–55. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1977a) More hippocampal neurons in adult mice living in an enriched environment. Nature, 386: 493–495. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1977b) Genetic influence on neurogenesis in the dentate gyrus of adult mice. Proc. Natl. Acad. Sci. U.S.A., 94: 10409–10414. Kornack, D.R. and Rakic, P. (1999) Continuation of neurogenesis in the hippocampus of the adult macaque monkey. Proc. Natl. Acad. Sci. U.S.A., 96: 5768–5773. Kostovic, I., Seress, L., Mrzljak, L. and Judas, M. (1989) Early onset of synapse formation in the human hippocampus: a correlation with Nissl-Golgi architectonics in 15- and 16.5week-old fetuses. Neuroscience, 30: 105–116. Letinic, K., Zoncu, R. and Rakic, P. (2002) Origin of GABAergic neurons in the human neocortex. Nature, 417: 645–649. Lopez-Garcia, C. and Martinez-Guijjaro, F.J. (1988) Neurons in the medial cortex give rise to Timm-positive boutons in the cerebral cortex of lizards. Brain Res., 463: 205–217. Maglo´czky, Z. and Freund, T.F. (2005) Impaired ad repaired inhibitory circuits in the epileptic human hippocampus. Trends Neurosci., 28: 334–340. Murakawa, R. and Kosaka, T. (1999) Diversity of the calretinin immunoreactivity in the dentate gyrus of gerbils, hamsters, guinea pigs, and laboratory shrews. J. Comp. Neurol., 411: 413–430. Nadarajah, B. and Parnavelas, J.G. (2002) Modes of neuronal migration in the developing cerebral cortex. Nat. Rev. Neurosci., 3: 423–432. Purpura, D.S. (1975) Normal and aberrant neuronal development in the cerebral cortex of human fetus and young infant. In: Brazier M. (Ed.), Brain Mechanisms in Mental Retardation. UCLA Forum in Medical Sciences. Academic Press, New York, pp. 141–169. Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The development, ultrastructure and synaptic connections of the mossy cells of the dentate gyrus. J. Neurocytol., 14: 835–857. Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and Nadler, J.V. (2000) Status epilepticus-induced hilar basal dendrites on rodent granule cells contribute to recurrent excitatory circuitry. J. Comp. Neurol., 428: 240–253. Rakic, P. and Nowakowski, R.S. (1981) The time of origin of neurons in the hippocampal region of the rhesus monkey. J. Comp. Neurol., 196: 99–128.

Rosenberg, F., Robain, O., Jardin, L. and Ben-Ari, Y. (1989) Distribution of GABAergic neurons in late fetal ad early postnatal rat hippocampus. Brain Res. Dev. Brain Res., 50: 177–187. Scharfman, H. (1993) Spiny neurons of area CA3/c in rat hippocampal slices have similar electrophysiological characteristics ad synaptic responses despite morphological variation. Hippocampus, 3: 9–28. Scharfman, H. (1994) Evidence from simultaneous intracellular recordings in rat hippocampal slices that area CA3 pyramidal cells innervate dentate hilar mossy cells. J. Neurophysiol., 72: 2167–2180. Scharfman, H. and Schwartzkroin, P.A. (1988) Electrophysiology of morphologically identified mossy cells of the dentate hilus recorded in guinea pig hippocampal slices. J. Neurosci., 8: 3412–3421. Scharfman, H., Smith, K.L., Goodman, J.H. and Sollas, A.L. (2001) Survival of dentate hilar mossy cells after piulocarpininduced seizures ad their synchronized burst discharges with area CA3 pyramidal cells. Neuroscience, 104: 741–759. Scheibel, M.E. and Scheibel, A.B. (1973) Hippocampal pathology in temporal lobe epilepsy. A Golgy-survey. In: Brazier M.A. (Ed.), Epilepsy. Its phenomena in man, UCLA Forum in medical Sciences, Vol. 17. Academic Press, New York, pp. 311–337. Seay-Lowe, S.L. and Claiborne, B.J. (1992) Morphology of intracellularly labeled interneurons in the dentate gyrus of the immature rat. J. Comp. Neurol., 324: 23–36. Seress, L. (1988) Interspecies comparison of the hippocampal formation shows increased emphasis on the regio superior in the Ammon’s horn of the human brain. J. Hinforschung (J. Dev. Brain Res.), 29: 335–340. Seress, L. (1992) Morphological variability and developmental aspects of primate granule cells: Differences between the rodent and primate gentate gyrus. In: Ribak C.E., Gall C.M. and Mody I. (Eds.), The Dentate Gyrus and Its Role in Seizures. Epilepsy Research, Supplement 7. Elsevier Science Publishers, Amsterdam, pp. 3–28. Seress, L. (2001) Morphological changes in the human hippocampal formation from midgestation to early childhood. In: Nelson C.A. and Luciana M. (Eds.), Handbook of Developmental Cognitive Neuroscience. MIT Press, Cambridge, MA, pp. 45–58. Seress, L., A´braha´m, H., Do´czi, T., La´za´r, G. and Kozicz, T. (2004a) Cocaine- and amphetamine-regulated transcript peptide (CART) is a selective marker of rat granule cells and of human mossy cells in the hippocampal dentate gyrus. Neuroscience, 125: 13–24. Seress, L., A´braha´m, H., Do´czi, T., Pokorny, J., Klemm, J. and Bakay, R.A. (2004b) Axomatic synapses on granule cells are preserved in human non-infiltrating tumour or lesion-related and mesial temporal sclerotic epilepsy, but markedly reduced in tumour-infiltrated dentate gyrus with or without epilepsy. Prague Med. Rep., 105/4: 357–368. Seress, L., A´braha´m, H., Paleszter, M. and Gallyas, F. (2001a) Granule cells are the main source of excitatory input to a subpopulation of GABAergic hippocampal neurons as

41 reveiled by electron microscopic double staining for zinc histochemistry ad parvalbumin immunocytochemistry. Exp. Brain Res., 136: 456–462. Seress, L., A´braha´m, H., Torno´czky, T. and Kosztola´nyi, Gy. (2001b) Cell formation in the human hippocampal formation from midgestation to the late postnatal period. Neuroscience, 105/4: 831–843. Seress, L. and Frotscher, M. (1990) Morphological variability is a characteristic feature of granule cells in the primate fascia dentata: a combined Golgi/electron microscope study. J. Comp. Neurol., 293: 253–267. Seress, L., Frotscher, M. and Ribak, C.E. (1989) Local circuit neurons in both the dentate gyrus and Ammon’s horn establish synaptic connections with principal neurons in five day old rats: a morphological basis for inhibition in early development. Exp. Brain Res., 78: 1–9. Seress, L. and Gallyas, F. (2000) The use of a sodium tungstate developer markedly improves the electron microscopic localization of zinc by Timm method. J. Neurosci. Method, 100: 33–39. Seress, L. and Leranth, C. (1996) Distribution of substance-P immunoreactive neurons and fibers in the monkey hippocampal formation. Neuroscience, 71: 633–650. Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule cells are normal features of the fetal and adult dentate gyrus of both monkey and human hippocampal formations. Brain Res., 405: 169–174. Seress, L. and Mrzljak, L. (1992) Postnatal development of mossy cells in the human dentate gyrus: a light microscopic Golgi study. Hippocampus, 2: 127–143. Seress, L., Nitsch, R. and Leranth, C. (1993) Calretinin immunoreactivity in the monkey hippocampal formation. I. Light and electron microscopic characteristics and colocalization with other calcium-binding proteins. Neuroscience, 55: 775–796. Seress, L. and Pokorny, J. (1981) Structure of the granular layer of the rat dentate gyrus. A light microscopic and Golgi study. J. Anat., 133/2: 181–195. Seress, L. and Ribak, C.E. (1984) Direct commissural connections to the basket cells of the hippocampal dentate gyrus: anatomical evidence for feed-forward inhibition. J. Neurocytol., 13: 215–225. Seress, L. and Ribak, C.E. (1988) The development of GABAergic neurons in the rat hippocampal formation. An immunocytochemical study. Dev. Brain Res., 44: 21–26. Seress, L. and Ribak, C.E. (1990) Postnatal development of the light and electron microscopic features of basket cells in the hippocampal dentate gyrus of the rat. Anat. Embryol., 181: 547–566. Seress, L. and Ribak, C.E. (1992) Ultrastuctural features of primate granule cell bodies show important differences from those of rats: axosomatic synapses, somatic spines and infolded nuclei. Brain Res., 569: 353–357. Seress, L. and Ribak, C.E. (1995a) Postnatal development and synaptic connections of hilar mossy cells in the hippocampal

dentate gyrus of rhesus monkeys. J. Comp. Neurol., 355: 93–110. Seress, L. and Ribak, C.E. (1995b) The postnatal development of CA3 pyramidal neurons and their afferents in the Ammon’s horn of rhesus monkeys. Hippocampus, 5: 217–231. Singec, I., Knoth, R., Ditter, M., Volk, B. and Frotscher, M. (2004) Neurogranin is expressed by principal cells but not interneurons in the rodent ad monkey neocortex ad hippocampus. J. Comp. Neurol., 479: 30–42. Sloviter, R.S. (1991) Permanently altered hippocampal structure, excitability ad inhibition after experimental status epilepticus in the rat: the dormant basket cell hypothesis and its possible relevance to temporal lobe epilepsy. Hippocampus, 1: 41–66. Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E., Goodman, J.H., Sollas, A.L. and Martin, D.L. (1996) Basal expression and induction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal dentate gyrus. J. Comp. Neurol., 373: 593–618. Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M. (2006) Kainic acid-induced recurrent mossy fiber innervation of dentate gyrus inhibitory interneurons: possible anatomical substrate of granule cell hyper-inhibition in chronically epileptic rats. J. Comp. Neurol., 494: 944–960. Spigelman, I., Yan, X.X., Obenaus, A., Lee, E.Y., Wasterlain, C.G. and Ribak, C.E. (1998) Dentate granule cells form novel basal dendrites in a rat model of temporal lobe epilepsy. Neuroscience, 86: 109–120. Stephan, H. (1975) Allocortex. In: Bargman W. (Ed.), Handbuch der mikroskopischen Anatomie des Menchen. Nervensystem, Vol. IV/9. Springer-Verlag, Berlin. Supe´r, H. and Soriano, E. (1994) The organization of the embryonic ad early postnatal murine hippocampus. II. Development of entorhinal, commissural ad septal connections studied with the lipophilic tracer DiI. J. Comp. Neurol., 344: 101–120. Swanson, L.W., Sawchenko, P.E. and Cowan, W.M. (1981) Evidence for collateral projections by neurons in Ammon’s horn, the dentate gyrus, ad subiculum — a multiple retrograde labeling study in the rat. J. Neurosci., 1: 548–559. Timm, F. (1958) Zur histochemie der schwermetalle, das sulfidsilber-verfahre. Deutsch Z. Gerichtl. Med., 46: 706–711. Urba´n, Z., Maglo´czky, Z. and Freund, T.F. (2002) Calretinincontaining interneurons innervate both principal cells ad interneurons in the CA1 region of the human hippocampus. Acta Biol. Hung., 53: 205–220. West, M.J. and Gundersen, H.J.G. (1990) Unbiased stereological estimation of the number of neurons in the human hippocampus. J. Comp. Neurol., 296: 1–22. Wilson, C.L., Isoka, M., Babb, T.L., Crandall, P.H., Levesque, M.F. and Engel Jr., J. (1991) Functional connections in the human temporal lobe. II. Evidence for a loss of functional linkage between contralateral limbic structures. Exp. Brain Res., 85: 174–187.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 3

The perforant path: projections from the entorhinal cortex to the dentate gyrus Menno P. Witter1,2, 1

Institute for Clinical and Experimental Neurosciences, Department of Anatomy & Neurosciences, VU University Medical Center, MF-G102C, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands 2 Centre for the Biology of Memory, Norwegian University of Science and Technology, Trondheim, Norway

Abstract: This paper provides a comprehensive description of the organization of projections from the entorhinal cortex to the dentate gyrus, which together with projections to other subfields of the hippocampal formation form the so-called perforant pathway. To this end, data that are primarily from anatomical studies in the rat will be summarized, complimented with comparative data from other species. The analysis of the organization of any of the connections of the hippocampus, including that of the entorhinal cortex to the dentate gyrus, is severely hampered because of the complex three-dimensional shape of the hippocampus. In particular in rodents, but to a lesser extent also in primates, all traditional planes of sectioning will result in sections that at some point or another do not cut through the hippocampus at an angle that is perpendicular to its long axis. To amend this, we will describe own unpublished tracing data obtained in the rat with the use of the so-called extended preparation. A number of issues will be addressed. First, data will be summarized which will clarify the laminar origin of the perforant pathway within the entorhinal cortex. Second, we will discuss whether or not a radial organization, along the proximo-distal dendritic axis of granule cells, characterizes the entorhinal-dentate projection. Third, we will discuss whether this projection is governed by any transverse organization, and fourth, we will focus on the organization along the longitudinal axis. Finally, the synaptic organization and the contralateral entorhinal-dentate projection will be described briefly. Taken together, the available data suggest that the projection from the entorhinal cortex to the dentate gyrus is a fairly well conserved connection, present in all species studied, exhibiting a grossly similar organization. Keywords: hippocampus; anatomy; parahippocampal region; topographical organization; subiculum dentate gyrus with its major cortical input. Second, the perforant pathway is among the pathways in the brain that have been most actively investigated, beginning with the earliest pioneers in neuroscience who described its remarkable structure and organization. These studies were followed by others, which clarified the central importance of the perforant pathway to hippocampal function and plasticity, studies which continue to this day.

Introduction A description of the projection from the entorhinal cortex to the dentate gyrus is an indispensable part of any work on the dentate gyrus for at least two reasons. First, this pathway, generally referred to as the perforant pathway, provides the Corresponding author. Tel.:+31 20 4448048; Fax:+31 20 4448054; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63003-9

43

44

The elaborate Golgi studies of Ramo´n y Cajal (1911) and Lorente de No´ (1933), first demonstrated that the entorhinal cortex is the origin of an immensely strong projection to the dentate gyrus. These observations were subsequently corroborated and extended in a seemingly continuous stream of tracing studies using axonal degeneration techniques (Blackstad, 1958; Hjorth-Simonsen, 1972; HjorthSimonsen and Jeune, 1972), followed by transport of radioactively labeled amino acids and horseradish peroxidase (Van Hoesen and Pandya, 1975; Steward, 1976; Wyss, 1981; Witter and Groenewegen, 1984; Witter, 1989; Witter et al., 1989b) and finally the more recently introduced sensitive tracing with lectines and dextran-amines (Ko¨hler, 1985; Witter, 1989; Tamamaki and Nojyo, 1993; Deller et al., 1996; Deller, 1998). In the same period, a number of retrograde tracing studies, again using a variety of different tracers, have further contributed to our current understanding of the entorhinaldentate projection (Ruth et al., 1982, 1988; Witter et al., 1989b; Dolorfo and Amaral, 1998). Although the dentate gyrus is the ‘‘traditional’’ target of the entorhinal-hippocampal fibers, there is ample evidence that the entorhinal cortex also projects to the hippocampal fields CA1–CA3, and to the subiculum (Steward, 1976; Steward and Scoville, 1976; Witter et al., 1989a; Desmond et al., 1994; Naber et al., 2001; Baks-Te-Bulte et al., 2005). In all species, the hippocampus and entorhinal cortex show a complex three-dimensional orientation and relationship. In rats in particular, this has lead to a general tendency to restrict hippocampal and entorhinal studies to those parts that are most easily accessible, such as dorsal hippocampus and central parts of entorhinal cortex. Initially this resulted in rather coarse descriptions of the overall functionality and connectivity of entorhinal-dentate relationships, and it was only after analyses started to include the full extent of both entorhinal cortex and dentate gyrus that we became aware of the complex topographical organization of this connection. It is now more widely appreciated that the hippocampus and entorhinal cortex are not as homogeneous in their organization as initially conceived, (cf. Witter and Groenewegen, 1990). A major reason for this new perspective was the

use of tools to circumvent a major problem with hippocampal studies in rodents, the fact that the hippocampal formation and the entorhinal cortex are curved structures. Cutting through such curved structures in any plane will result in sections that at some point or another do not cut through the hippocampus and entorhinal cortex at an angle that is perpendicular to the respective long axes. Initially described by Gaarskjaer (1978), and advocated by Ishizuka (2001), and likewise by Amaral and Witter (1989), this problem can be amended by using the so-called extended preparation, and in a number of studies this approach has resulted in improved understanding of the connectional organization of the system (Amaral and Witter, 1989; Ishizuka et al., 1990; Ishizuka, 2001). Many of the disputes about functions of the entorhinal-dentate projection, and entorhinal-hippocampal connectivity more generally, are likely to be related to this spatial problem, at least in part. A proper understanding of the complex three-dimensional organization of the system is essential to evaluate data from functional studies which suggest heterogeneity, as has been suggested for the dorsal versus the ventral hippocampus, for example (Witter et al., 1989a; Moser et al., 1993). Another, more recent example deals with the rather confusing data addressing the functional relevance of the entorhinal cortex. Here, again, studies have benefited significantly from taking into account the complex three-dimensional organization of cortico-entorhinal-hippocampal connections (Fyhn et al., 2004; Hafting et al., 2005; Hargreaves et al., 2005; Steffenach et al., 2005; Sargolini et al., 2006). Over the years, one of our aims has been to further our understanding of the system through systematic descriptions of its complex anatomical organization. For the present paper, the focus will be on the entorhinal-dentate projection; in a recently published parallel paper, the focus was on entorhinal projections to the subiculum (Witter, 2006). The data support a currently prevailing view that there are two differentially organized components within the entorhinaldentate projection, originating from the lateral and medial entorhinal cortex respectively.

45

Nomenclature Dentate gyrus In all species, the dentate gyrus represents part of the cortex situated closest to the free rim of the original cortical anlage, but the final position and overall structural organization of the dentate gyrus and of the hippocampus as a whole may be quite different among species (Stephan, 1975; Voogd et al., 1998). To give a striking example, one just has to compare the position of the hippocampus in the rat with that in primates. In the rat the structure extends from a dorsomedial position, in close proximity to the most caudodorsal part of the septal complex to a ventrolateral position in close proximity to the most caudomedial parts of the amygdaloid complex. In contrast, in primates, this same structure starts from a dorsocaudal position, close to the splenium of the corpus callosum, extending ventrorostrally all the way again to apposition in close proximity to the amygdaloid complex. Moreover, in the rat, the dorsal portion of the dentate gyrus, plus parts of the hippocampus proper, seem to fold backwards, continuing into a structure generally referred to as the gyrus fasciolaris (fasciolaris cinereum), whereas in primates such a folding seems to be the main characteristic of the most ventral and anterior part, referred to as genu and uncal portions, spatially associated with the amygdaloid complex. A second example refers to the overall connotation that the principal cell layer of dentate gyrus contains granule cells, i.e. cells without basal dendrites but with an extensive apical tuft. In contrast, in the human and monkey dentate, a sizeable proportion, in humans up to 30%, of the granule cells do have basal dendrites (Seress and Mrzljak, 1987; Lim et al., 1997). Irrespective of such differences, it is useful to develop a coherent and generally applicable nomenclature, for example to be able to distinguish one portion of the granule cell layer from another. For this paper, I will call the portion of the granule cell layer that is adjacent to CA1, the enclosed blade (which is synonymous with suprapyramidal, dorsal, or inner blade/limb), and the opposite

portion of the granule cell layer will be called the exposed blade (which is synonymous with infrapyramidal, ventral, outer, or free blade/limb). The region of the ‘‘V’’ or ‘‘U’’ that unites the two blades will be referred to as the crest, also called apex or vertex. Aside from the longitudinal or dorsoventral axis, a description of the organization of the dentate gyrus is strongly facilitated by defining two additional axes, a transverse and a radial axis. The transverse axis is oriented perpendicular to the long axis, similar to the term ‘lamellar axis’ because a section through the transverse axis would reveal the classic lamellar arrangement of the hippocampus. The radial axis refers to an axis that is essentially perpendicular to the transverse axis. It can be defined as a perpendicular to the granule cell layer, extending from molecular layer through the granule cell layer to the hilus. It therefore represents the orientation of the granule cell dendrites from their most distal apical extent to their origin at the granule cell soma. For the present description we will refer to the three layers of the dentate gyrus as molecular layer, granule layer, and the hilus of the dentate gyrus as the third, inner, or polymorph layer. The molecular layer will be further subdivided into inner, middle, and outer one-thirds.

Entorhinal cortex In view of the classical notion that the entorhinal cortex is the source of the projection to the dentate gyrus (Ramo´n y Cajal, 1911), we and others have used this to delineate the entorhinal cortex of the rat and the monkey (Insausti et al., 1997). The entorhinal cortex comprises six layers, including the two cell-sparse layers I and IV, respectively called the molecular layer and lamina dissecans. There are several reasons to use this nomenclature for entorhinal cortical layers instead of the initially proposed terms defined by Lorente de No´ (1933), who considered the lamina dissecans as the deep, sparsely populated part of layer III, and used the designation IV for the layer of large pyramidal cells which is layer Va in the currently employed

46

terminology (Witter et al., 1989a; Witter and Amaral, 2004). In general, the entorhinal cortex can be subdivided into two components generally referred to as lateral and medial entorhinal cortex or Brodman’s areas 28a and 28b respectively (for a more detailed description and comparison of different nomenclatures used in the rat and in different species the reader is referred to a number of reviews (Witter et al., 1989a). In the rat, and likewise in the mouse, the entorhinal cortex has been further subdivided into dorsolateral (DLE), dorsal-intermediate (DIE), ventral-intermediate (VIE), caudal (CE), and medial (ME) subdivisions (Insausti et al., 1997; van Groen et al., 2003). Initially, we included a sixth domain called amygdalo-entorhinal cortex (AE) as part of the entorhinal cortex because our own tracing data indicated that this region also contributed to the projections to the dentate gyrus (Insausti et al., 1997). In a recent and more detailed analysis of the efferent connectivity of this region, it was shown that it is quite unlikely that this so-called area AE indeed contributes projections to the dentate gyrus (Kemppainen et al., 2002; Majak and Pitkanen, 2003). Therefore, we exclude this region in the present description. In monkeys, humans, and in other species in which the entorhinal cortex was described, such as cat, dog, and guinea pig (Krettek and Price, 1977; Amaral et al., 1987; Witter et al., 1989a; Insausti et al., 1995; Uva et al., 2004; Woznicka et al., 2006) comparable partitioning schemes into multiple subdivisions of the entorhinal cortex have been proposed. However, in those species there is also a tendency to consider the entorhinal cortex as composed of two primary components, most likely reflecting functional differences. For the present description we will therefore maintain a subdivision of the entorhinal cortex into two main subdivisions, the lateral (LEC) and medial (MEC) entorhinal cortices respectively. Studies of the perforant pathway projection to the dentate gyrus are most numerous and most detailed in the rat. Therefore, the descriptions will be initially provided for material gathered from experiments in rats. When relevant, data from other species will be added in a comparative fashion.

Fiber pathways After leaving the entorhinal cortex, perforant path fibers enter the underlying white matter and the angular bundle. They then traverse the pyramidal cell layer of the subiculum and cross the hippocampal fissure to enter the dentate gyrus, or distribute to the molecular layer of the subiculum and the hippocampus. The entorhinal cortex fibers also take alternative routes such as projecting through the alveus before entering the hippocampus, or by traversing the molecular layers of the entorhinal cortex, pre and parasubiculum. The general orientation of entorhinal fibers within hippocampal subfields is parallel to the principal cell layers, and it has been described that entorhinal fibers in the molecular layer of the CAfields continue around the tip of the enclosed blade of the dentate gyrus to enter the dentate molecular layer (Ramo´n y Cajal, 1911; Witter, 1989).

Layers of origin in the entorhinal cortex In rat, mouse, or monkey, the ipsilateral projection to the dentate gyrus appears to arise mainly, if not exclusively from layer II of the entorhinal cortex (Steward and Scoville, 1976; Schwartz and Coleman, 1981; Ruth et al., 1982, 1988; Witter et al., 1989b; van Groen et al., 2003; Chrobak and Amaral, 2006). In humans, fetal material indicates a similar origin (Hevner and Kinney, 1996). It is of interest to note that in the brain of Alzheimer patients, the entorhinal cortex layer II neurons are among the ones preferentially implicated in the disease, such that up to 50% of those neurons apparently disappear. This rather selective loss of layer II neurons has been associated with changes in the outer part of the molecular layer, such as a decrease in free glutamate and a decrease in markers associated with the perforant path, suggesting that in humans also the origin, and obviously termination, of the entorhinal-dentate projection is similar to that reported in other species (Hyman et al., 1986, 1987, 1988; Morys et al., 1994). There is evidence that a minor component of the projection to the dentate gyrus also comes from the deep layers (IV–VI) of the entorhinal cortex (Ko¨hler, 1985; Witter and Amaral, 1991; van Groen et al., 2003). In rat and

47

monkey, neurons in layer II not only project to the dentate gyrus but also to CA3 (Steward and Scoville, 1976; Witter and Amaral, 1991). Intracellular tracing in the rat revealed that a single layer II cell may innervate not only the dentate gyrus and CA3 but also the subiculum (Tamamaki and Nojyo, 1993), but whether this holds true for the monkey as well is currently unknown. In contrast, in the mouse, at least in one of the strains (C57BL/6J), layer II cells appear to project only to the dentate gyrus and does not extend collaterals to other hippocampal subfields (van Groen et al., 2003). Although this appears a striking species difference, emphasized by several authors, it is currently not known whether this is a general phenomenon in all mouse strains or typical for this one strain. Moreover, our own material has indicated that in the rat layer III cells contribute to the projection to CA3. As illustrated in Fig. 1A, a ventrally positioned injection in layers II and III of CE (case 88246) results in strong labeling of the dentate gyrus, all CA fields and the subiculum. In contrast, a similarly positioned injection confined to layer III (case 89339) does not result in any labeling in the dentate gyrus, as expected on the basis of the selective origin of the dentate component of the perforant path in layer II. However, in this case we did find some labeling in CA3, as well as strong labeling in CA1 and the subiculum (Fig. 1B, D). Likewise, this can be observed in case of an injection in layer III of VIE (case 88178; Fig. 1C, E). On the basis of these data, it thus seems safe to conclude that in all species, layer II is the predominant if not exclusive source of the entorhinal projections to the outer two-thirds of the molecular layer of the ipsilateral dentate gyrus (cf. Kunzle, 2002). In addition, these data indicate that in the rat, the projection to CA3 originates predominantly in layer II with a minor component arising from layer III. The layer III projection to area CA3 is apparently more prominent in mice.

Radial or layered terminal organization in the dentate gyrus One of the more convincing arguments to differentiate between the lateral and medial subdivisions

of the entorhinal cortex, aside from overall cytoarchitectonic differences, was based on observations that fibers originating in the LEC terminate in the outer one-third of the molecular layer and fibers from the MEC terminate in the middle onethird of the molecular layer (Hjorth-Simonsen and Jeune, 1972; Hjorth-Simonsen, 1972). This initial idea was strengthened further by the striking laminar staining pattern that was revealed by Timm stain for heavy metals, (Haug, 1976; Stanfield and Cowan, 1979) and by immunocytochemistry using antibodies against enkephalin and cholecystokinin (Fredens et al., 1984). Although most consider a non-overlapping distribution in the dentate of LEC and MEC fibers, it is important to note that there may be a continuum in these fiber systems. This was originally suggested on the basis of experiments with anterograde transport of tritiated amino acids by Steward (1976), who suggested that the perforant path should actually be subdivided into lateral, intermediate, and medial components. The intermediate pathway was thought to originate in the intermediate entorhinal area, an area most likely including VIE and ME, and to terminate in the molecular layer between the projections from LEC and MEC. In an elegant series of anterograde tracing experiments using small injections of tritiated amino acids in rats, Wyss (1981) reported that there is a striking radial gradient in the terminal distribution of entorhinal fibers depending on the position of the source in the entorhinal cortex. Lateral portions of LEC project close to the pial surface and successively more medial portions project closer to the granule cells. However, the data from this study also supported an intermediate perforant pathway, distributing to the border region between the typical lateral and medial perforant path terminal zones in the outer and middle one-thirds of the molecular layer, respectively. A comparable conclusion was reached on the basis of a study using very small injections of the anterogradely transported tracer DiI in portions of the entorhinal cortex close to the rhinal fissure, i.e. projecting to the dorsal part of the dentate gyrus (Tamamaki, 1997). When looking at a summary figure taken from this particular study (Fig. 2), two conclusions are apparent. First, the radial

48

Fig. 1. Layer II and to a lesser extent layer III contribute to the projections to CA3. Sections taken from unfolded preparations (Gaarskjaer, 1978), stained for the presence of BDA and counterstained for Nissl-substances with cresyl violet. (A) Section from a case with an injection centered in layers II and III of MEC (Fig. 1, case 89246). Strong labeling is present in the dentate gyrus, as well as in CA3 (and CA1 and subiculum). (B) Section from a case with an injection of which the location corresponds to that illustrated in A, but no tracer has been taken up by layer II neurons (case 88339). Note the absence of labeling in the dentate gyrus, strong labeling in CA1 and subiculum but also weak labeling in CA3 (boxed area shown in D). (C) Section from a case with an injection in layer III of LEC (Fig. 2, case 88178). Similar to the case shown in B, labeling is present in CA1 and subiculum, absent in dentate gyrus, but weak labeling can be seen in CA3 (boxed area shown in E). (D) Higher magnification of boxed area marked in B, showing the light terminal plexus in CA3. (E) Higher magnification of boxed area marked in C, showing the light terminal plexus in CA3. Scale bar in A equals 1 mm; also holds for B and C. Scale bars in D and E equal 100 mm.

extent of the labeled terminal field in the dentate gyrus almost never covers one-third of the molecular layer: one-fifth to one-sixth would be a more appropriate description. Second, the terminal fields show a gradual transition from a very distal to a more proximal position along the apical dendrites of dentate granule cells. These observations led to the proposition that the entorhinaldentate projection is a continuum rather than subdivided into two or three discrete components. Interestingly, analysis of intracellularly labeled

cells in layer II of the medial entorhinal cortex which projected to the dorsal part of the dentate gyrus revealed yet another pattern, where the axon is confined to the middle one-third in the enclosed blade but extends into the outer one-third of the exposed blade, i.e. the terminal zone becomes thicker in the exposed blade (Tamamaki and Nojyo, 1993). This finding was supported by the results of anterograde tracing using DiI. Injection of DiI into lateral entorhinal cortex led to a thick band of terminal labeling in the enclosed blade but

49

Fig. 2. Confocal microphotographs showing anterograde labeling in the enclosed blade of the dentate gyrus following a series of anterograde tracer injections (DiI) along the rostro-caudal extent of the rhinal fissure. A is taken from the most anterior injection, which is in LEC, F is taken from the most posterior position, which is in MEC. Cases D and E are most likely along the border between LEC and MEC, although this has not been assessed in the original paper. Arrows demarcate laminar boundaries. Scale bar equals 100 mm. Adapted with permission from Tamamaki (1997).

comparatively thin labeling in the exposed blade. Following DiI injections in the medial entorhinal cortex, the reverse pattern was observed. Our own data, based on visualizing the projections with PHA-L and BDA (Witter, 1989, 1990), to a large extent corroborates the observations reported by Tamamaki and Wyss (Wyss, 1981; Tamamaki, 1997). Whereas large injections generally label the full width of the outer or middle one-third of the dentate molecular layer, small injections never result in consistent labeling throughout the entire outer or middle one-third. In these cases, the labeled domain is closer to one-fifth or one-sixth of the entire width of the molecular layer, similar to the observations of Tamamaki (1997). There is a weak topographical arrangement emerging from these experiments. Injections in the lateral part of LEA or dorsolateral part of MEA more densely target the more distal portions of the dendrite in their respective terminal zone, whereas more ventrally and medially positioned injections tend to distribute fibers preferentially to more proximal portions

of their corresponding one-third of the molecular layer. What remains to be discussed is whether or not the perforant path is organized as a continuum, as proposed by Tamamaki (1997), whether it comprises two non-overlapping systems — lateral and medial, or whether there are three components — lateral, intermediate and medial. To address this point, we compared the radial distribution of anterogradely labeled fibers following two closely spaced injections, centered around the border between VIE and ME in the rat (Figs. 3 and 4). The injection in ME (case 88246), which is part of MEC, clearly produces a narrow terminal field in a narrow distal part of the middle one-third of the molecular layer (Fig. 3A). In contrast, a spatially close injection, but in VIE (case 88334), which is part of LEC, labels a narrow terminal field in a proximal part of the outer one-third of the molecular layer (Fig. 3B). One can easily imagine that a slightly larger injection around the border region between VIE and ME would result in a terminal

50

Fig. 3. Radial distribution of the entorhinal-dentate projections. (A) High power photomicrograph of a small portion of the enclosed blade of the dentate gyrus, taken from a case with an injection in MEC. Indicated is the border between the molecular layer and lacunosum-moleculare of CA1 (white line). Arrows indicate the measures of the width of the molecular layer (a), expressed as 100% (see also the Y-axis in D), the position of the outer border or the terminal field (b) and that of the inner border (c), as measured from the outer border of the molecular layer (see also C). (B) High power photomicrograph of a small portion of the enclosed blade of the dentate gyrus, taken from a case with an injection in LEC. The picture is size-matched and merged with A to illustrate the striking difference in position of the two terminal fields (see also D). (C) Section with terminal labeling in the dentate gyrus, illustrating the measuring protocol applied in this study. The center of the line has been derived, and equally distanced lines, perpendicular to this center-axis, have been generated to measure the position of the terminal field as indicated in A. This procedure is carried out in all sections with a labeled terminal field. (D) Graphic representation of four representative cases with terminal fields confined to either the outer 35% (cases 88183R and 88334; see also Fig. 2), or the middle portion between 35 and 66.5%. Exposed and enclosed blades have been analyzed separately in view of the overall fluctuations in density and position of the terminal fields between the two blades. Note that two closely positioned injections on either side of the border between LEC and MEC (cases 88334 and 89246 respectively) show completely non-overlapping, though adjacent, terminal fields. Scale bar in C, D equals 100 mm.

51

Fig. 4. Longitudinal and transverse distribution of the entorhinal-dentate projections. (A) Unfolded representations of the dorsoventral extent of the dentate gyrus and terminal labeling resulting from injections in LEC (upper row) and MEC (lower row). (B) Unfolded representation of the entorhinal cortex with injections in lateral and medial subdivisions. Case 88178 (light grey) is an injection in layer III, illustrated in Fig. 1C. Abbreviations: AE, amygdalo-entorhinal area; LEC, lateral entorhinal cortex; MEC, medial entorhinal cortex; PaS, parasubiculum.

52

pattern that is best described as an intermediate pathway (see for example Fig. 2D, E). A problem with all such types of analyses is that the data are taken from a single level of a highly divergent projection. Note that most studies indicated differences between exposed and enclosed blades (see above) or between different longitudinal levels. Therefore, we aimed to assess our terminal distributions on the basis of all available data collected throughout the entire terminal domain. As illustrated schematically in Fig. 3C, we calculated the centerline of the terminal field for multiple positions and subsequently measured the radial thickness of the molecular layer, taken perpendicular to this centerline (measure a). We also measured the outer (measure b) and inner border (measure c) of the terminal field at all these radial points throughout all sections of the extended dentate gyrus that showed anterogradely labeled perforant path fibers (Figs. 4 and 5). For both illustrated experiments, as well for all other experiments illustrated throughout this paper, we collected such a population of outer and inner border measures. We subsequently established that the outer border measures for the medial perforant pathway where significantly different from the inner border measures for the lateral perforant pathway (Fig. 3D; Po0.05). At this point, one may wonder what we know about this issue in other species. In the mouse, no detailed studies with respect to the radial differences in the distribution of the entorhinal projections to the dentate are available. The only detailed account in mice seems to support a separation between fibers originating in LEC and MEC (van Groen et al., 2003). Irrespective of their origin in either DLE, DIE or VIE, LEC fibers terminate throughout the extent of the outer onethird of the molecular layer. Fibers arising from MEC preferentially terminate throughout the middle one-third, although in this pathway, thinner terminal zones have been noted, similar to what has been observed in the rat. In contrast, the situation in the macaque monkey is in support of a more continuous organization of the entorhinaldentate projection (Witter et al., 1989b; Witter and Amaral, 1991). Irrespective of the origin in EC, at all levels of the dentate gyrus innervated, labeling

was present throughout the extent of the outer two-thirds of the molecular layer. Subtle indications for differences between lateral and medial perforant pathways were observed however. Rostral parts of EC, most likely homologous to LEC of the rat, project most densely to the outer onethird, whereas projections originating in the most caudal parts of the monkey EC, most likely an area comparable to MEC of the rat, showed a preference for the middle one-third. Interestingly, also the typical differentiation between the terminal zones of the two perforant path components, as seen with the Timm stain, is largely absent in the monkey dentate, with the exception of the anterior genu and uncal portion (Witter and Amaral, 1991). Taken together, the data do not allow for a conclusive statement whether or not the radial organization of the entorhinal-dentate projections supports the differentiation into two discrete lateral and medial components. One question is whether such distinctions influence function. We do know that lateral and medial entorhinal cortex most likely convey different types of information to the dentate molecular layer, and the most striking difference seem to relate to spatial modulation, with the cells in the medial entorhinal cortex being more strongly spatially modulated relative to those in the lateral entorhinal cortex (Hafting et al., 2005; Hargreaves et al., 2005). We also know that these pathways are differently modulated and that the overall postsynaptic effects are different (see the section on synaptic organization below). More important, however, is the conclusion that in all species, information from functionally different entorhinal domains converges onto a single population of dentate granule cells (see section on synaptic organization below).

Transverse organization in the dentate gyrus There are conflicting papers on the transverse distribution of the perforant path projection. In earlier studies no differences were reported (HjorthSimonsen, 1971, 1972; Hjorth-Simonsen and Jeune, 1972; Steward, 1976). Wyss (1981) reported that the lateral perforant pathway preferentially

53

Fig. 5. Representative examples of longitudinal, transverse, and radial terminal distributions of the entorhinal-dentate projections. (A–C) Injections in LEC. (D–F) Injections in MEC. For details see text.

54

projects to the enclosed blade of the dentate gyrus, whereas the medial component either does not show a preference or predominantly targets the exposed blade. It should be stressed, however, that dimensions of injection sites may lead to different results, similar to what was described with respect to the radial organization. The findings of Wyss were supported by the observation that the terminal zone of LEC fibers is wider in the enclosed blade, covering almost the entire outer one-third of the enclosed blade, whereas the terminal field in the exposed blade is much thinner (Tamamaki, 1997). For the MEC, the reverse appears to be the case. Importantly, individual neurons show extensive variation in their target projection, demonstrated by the axon arbors of layer II cells that were intracellularly labeled in vivo. An individual layer II cell may send axon collaterals along the entire transverse extent of the dentate gyrus, or project preferentially to either one of the two blades. These data are important because they suggest that the EC projections as a whole may provide homogeneous innervation of the dentate, but individual components could have specific and selective functional influence. Indeed, using the extended preparations and subsequent unfolding, we noticed no systematic organization in a total of 21 experiments. In most cases, when in the core of the projection, labeling was present in both blades, although it was unequal in density in some instances (Figs. 4 and 5). In case of injections in LEC, we generally observed, in the dorsal and ventral domains of the terminal field, a clear preference for the enclosed blade, whereas even in the core, labeling was denser in the enclosed blade (Fig. 5A–C). In particular, more laterally placed injections in LEC (cases 90170, 881813R, 89156) showed extensive labeling in the dorsal part of the enclosed blade, exclusively (Fig. 4, top row), in line with the observations published by Wyss (1981). Regarding the projections originating from MEC (Fig. 4, lower row; Fig. 5D–F) no striking preference for either blade was apparent although differences in density do occur between experiments (compare Fig. 5D, E and F) as well as within experiments (Fig. 5E). In the mouse and the monkey, no transverse organization has been described. Irrespective of

the origin of the projection, when labeling is present in the dentate molecular layer, it appears to be distributed equally along both blades (Witter et al., 1989b; van Groen et al., 2003). In conclusion, the entorhinal-dentate projection appears to distribute along the transverse axis of the dentate gyrus in a homogeneous fashion. Although a subtle topographical arrangement cannot be excluded, and individual cells may differ in their particular target projection, there is no obvious pattern. It is worth to emphasize though that convergent evidence points to functional differences between the enclosed and exposed blades of the dentate gyrus. For example in guinea pigs, cells in the enclosed blade as well as the associated hilar mossy cells show morphological sex differences whereas such differences have not been observed in the exposed blade and associated hilar mossy cells (Bartesaghi et al., 2003; Guidi et al., 2006). Differences between the two blades have further been reported with respect to sensitivity for hypoxia (Hara et al., 1990), neurogenesis (Choi et al., 2003), efficacy in activating hippocampal circuits (Scharfman et al., 2002), and loss of granule cells following adrenalectomy (Jaarsma et al., 1992); (see also Chawla et al., this volume).

Longitudinal organization of entorhinal-dentate projections Entorhinal projections show a striking organization along the longitudinal axis of the dentate gyrus. Originally described in the cat (Witter and Groenewegen, 1984), it was later discovered to be a general governing principle in all species studied (Ruth et al., 1982, 1988; Witter et al., 1989b; Dolorfo and Amaral, 1998; van Groen et al., 2003). In the rat, cells located laterally in the entorhinal cortex project to dorsal parts of the dentate gyrus while cells located progressively more medially project to more ventral levels of the dentate gyrus (Fig. 4). This organization leads to a pattern of connections such that the dorsal part of the dentate gyrus receive inputs from lateral parts of the LEC and lateral and caudal parts of MEC, whereas the ventral portions of the dentate gyrus receive input from more medial portions of both

55

LEC and MEC (Ruth et al., 1982, 1988; Witter et al., 1989a, 2000). This topographical organization has been convincingly demonstrated in a series of retrograde tracing experiments, in which discrete injections in dorsal, mid-dorsoventral, and ventral levels of the dentate gyrus resulted in labeled populations of entorhinal neurons, in both LEC and MEC, with a different lateral to medial position (Dolorfo and Amaral, 1998). Although the domains of the entorhinal cells projecting to these different longitudinal levels do not show much overlap, one should take into account that a single entorhinal layer II neuron, as shown with intracellular filling, may distribute an axon along as much as 2 mm (20–25%) of the dorsoventral extent (Tamamaki and Nojyo, 1993) and that restricted injections in EC in almost all cases result in labeling extending over at least one-half to up to two-thirds of the long axis (Witter et al., 1989a; Tamamaki, 1997), such that overlapping terminal zones in the dentate gyrus of these populations of layer II neurons are likely to exist. This widespread distribution along the long axis demonstrated by anterograde tracing seems at odds with the rather restrictive labeling in the entorhinal cortex resulting from retrograde tracing methods. One may therefore pose the question whether this longitudinal topographical organization actually holds true (Tamamaki, 1997). However, it is essential to take into account the effect of injection size, which quite often is not easily determined in anatomical tracing experiments. Most data support the idea that small foci of entorhinal layer II cells adhere to the overall 25% longitudinal axonal distribution, as described on the basis of single cell axonal distributions. The unfolded approach unmasks another feature of this projection that has not been appreciated before. Irrespective of the size of the injection, the position of the origin in the entorhinal cortex, and thus the overall terminal distribution along the longitudinal axis, appears critical. In case of an injection in the entorhinal cortex, which strongly labels the extreme dorsal or ventral tips of the dentate gyrus (Fig. 4, cases 90170, 88334, 89361, and 88400), the terminal fields do not extend over more than about 20–25% of the long axis. In contrast, following injections that produce strong

labeling in the intermediate levels of the hippocampus, the longitudinal extent is generally larger (Fig. 4, cases 88183R, 89156, 89247, and 89246). The longitudinal organization of the entorhinaldentate projection in the mouse, cat, and monkey is strikingly comparable. In these species again, lateral and caudal parts of the entorhinal cortex, encompassing a band that parallels the lateral border between entorhinal and perirhinal/parahippocampal cortex and thus including parts of both the lateral and medial entorhinal cortex preferentially project to the dorsal (posterior) dentate gyrus. Progressively more medial portions of EC, again encompassing portions of both LEC and MEC project to increasingly more ventral (anterior) portions of the dentate gyrus (Witter et al., 1989a, b; van Groen et al., 2003). We may thus conclude that of all the principles governing the organization of the perforant pathway projection to the dentate gyrus, the longitudinal one is most reliable throughout the animal kingdom. Whether this also holds true for the human is currently unknown but it seems plausible given the number of studies supporting functional differences along the long axis that would be likely to reflect this longitudinal topography. It is clear, however, that further research is needed to fully understand and assess the potential functional implications of these findings. Complicating factors are the strong longitudinal associational pathways present within several of the hippocampal subfields (Witter and Amaral, 2004). Since we do not yet appreciate the functional relevance of those strong longitudinal associational connections, functional differences along the long axis may be hard to assess.

Contralateral projections In the rat, the entorhinal cortex projects to the ipsilateral dentate gyrus, and also gives rise to a crossed projection to the contralateral dentate gyrus, as well as the contralateral CA3 and CA1 subfields, and the subiculum. The crossed entorhinal projection is most prominent to the more dorsal portions of the hippocampal subfields and rapidly diminishes in density at more temporal

56

levels (Goldowitz et al., 1975; Steward, 1976). With respect to the laminar origin of the crossed projections, Steward and Scoville (Steward and Scoville, 1976) reported that it matches the ipsilateral origin. The crossed dentate projection, that thus originates from layer II cells, mainly takes the more common perforant path trajectory, i.e. crossing the midline through the ventral hippocampal commissure. In the mouse, almost no crossed projection to the dentate gyrus have been observed, in contrast to the rat (van Groen et al., 2003). In rabbit and cat, a distinct projection to the contralateral DG, mainly limited to the dorsal part, has been described, similar to that in the rat (Goldowitz et al., 1975; Hjorth-Simonsen and Zimmer, 1975; Steward and Scoville, 1976; Wyss, 1981). In the monkey, this projection is very modest, limited to the uncal/genu portion only (Witter and Amaral, 1991).

Synaptic organization of the entorhinal-dentate projection In the molecular layer of the dentate gyrus in the rat, the terminals of the perforant path fibers in the outer two-thirds of the molecular layer make up at least 85% of the total synaptic population (Nafstad, 1967; Matthews et al., 1976). Entorhinal fibers form mainly, if not exclusively, asymmetric synapses (Nafstad, 1967; Matthews et al., 1976; Deller and Leranth, 1990; Leranth et al., 1990). These occur most frequently on the dendritic spines of dentate granule cells, although a small proportion of perforant path fibers terminate on non-spiny dendrites of presumed interneurons. These include parvalbumin/ GABA-immunoreactive (Zipp et al., 1989), somatostatin-positive, and NPY-positive neurons with cell bodies located in the hilus and apical dendrites which extend into the outer portions of the molecular layer (Scharfman, 1991; Soriano and Frotscher, 1993). Since at least half of the NPY-positive neurons also stain for somatostatin (Swanson and Ko¨hler, 1986) and some co-localization of NPY and parvalbumin has been reported (Deller and Leranth, 1990), it cannot be excluded that what may seem as three different population of interneuronal targets, may in fact be the population of interneurons

known to distribute local axonal plexi to the perforant path terminal zone (Bakst et al., 1985, 1986; Halasy and Somogyi, 1993; Boyett and Buckmaster, 2001). It is most likely that the so-called molecular layer perforant path-associated cells (MOPP) as well as the hilar commissural-associational pathway related cells (HICAP)(Han et al., 1993) are also among the postsynaptic targets of entorhinal axons, but this remains to be established. Note that in monkeys, NPY-positive cells in the hilus do not extend dendrites into the molecular layer (Nitsch and Leranth, 1991). Whether this implies a species difference in connectivity or in the expression of certain proteins is not clear. In addition to the main innervation arising from layer II cells in the entorhinal cortex, a projection originating from deep layers has been described above (Ko¨hler, 1985). This projection preferentially distributes to the inner portion of the molecular layer, the granule cell layer, as well as the subgranular zone, where it establishes asymmetrical synapses onto granule cell dendrites as well as on their somata and onto spine-free dendrites in the subgranular zone. The latter most likely represent dendrites of interneurons (Deller et al., 1996). The observation that most of the entorhinal synapses in the dentate gyrus are asymmetric strongly suggests that the pathway is largely excitatory, most likely using glutamate as primary transmitter (White et al., 1977). Although not studied in detail, prominent differences in these respects between the lateral and medial perforant pathway are unlikely (Nafstad, 1967; Matthews et al., 1976), similar to observations recently reported for the lateral and medial entorhinal projections to the subiculum (Baks-Te-Bulte et al., 2005). However, regarding the chemical characteristics, differences between the two pathways have been described. Terminals of the lateral perforant pathway are also enkephalin immunoreactive, whereas those of the medial pathway are immunoreactive for CCK and dynorphin (Fredens et al., 1984; van Abeelen, 1989). Medial perforant path fibers are immunoreactive for the metabotropic glutamate receptor, mGLUR 2/3, whereas the lateral fibers are not. There is also convincing evidence that neurons in LEC and MEC that project to the

57

dentate gyrus are markedly different with respect to their electrophysiological properties (van Der Linden and Lopes da Silva, 1998; Wang and Lambert, 2003). Surprisingly, there is, as yet, no distinctive marker at the level of the entorhinal cortex for cells that give rise to the lateral and medial perforant path projections. In view of the extensive focus on whether or not the two pathways predominantly target different segments of granule cell dendrites (see section on radial organization above), there has been surprisingly less interest in the question whether or not it is a general feature for lateral and medial perforant path fibers to converge onto a single granule cell. Although this seems likely in view of the overall ‘‘en passant’’ type of termination of both pathways and the fact that both indeed distribute along the entire transverse extend of the dentate gyrus, thus increasing the likelihood that each dentate granule cells receives convergent input from multiple entorhinal layer II cells, there is no strong anatomical data to support the idea of convergence. In view of the surprising observation of an apparent lack of convergence in similarly organized pathways in the subiculum (Cappaert et al., 2005), this issue of convergence at the single cell level remains to be resolved.

Summary and functional comments The main origin of the projection from the entorhinal cortex to the dentate gyrus is throughout layer II cells in EC. Entorhinal axons of layer II cells have their terminals predominantly, if not exclusively, in the outer two-thirds of the molecular layer of the dentate gyrus. An additional projection arises from cells in deeper layers of the entorhinal cortex, mainly deep V and VI. In the rat, and likely in the mouse as well, these deeply located cells appear to be the source of most if not all of the perforant path fibers that distribute outside the ‘‘traditional’’ terminal zone in the outer molecular layer, showing a more dispersed terminal distribution in the hilar region, inner molecular and granular layers of the DG. However, these minor projections have received little emphasis, and remain largely unexplored, and most would

suggest that they play a minor role in the entorhinal input to the dentate gyrus compared to the primary projection from layer II neurons. Although there is converging evidence that the entorhinal-dentate projection can be subdivided into two, functionally different systems, the precise spatial relationship between those two systems and the radial terminal distribution in the molecular layer of the dentate gyrus is not entirely clear at the present time. Although generally accepted in rodents, and quite often used as a model system to understand mechanisms for lamina-specific axon outgrowth (see chapter by M. Frotscher in this volume), this radial differentiation is less apparent in the monkey. Taken together with the other species differences in the two pathways discussed above, the question can be raised whether these species differences in the topographic organization of the perforant path impart functional differences to distinct species. Physiological differences would be expected if the organization of the projections is not the same, given the evidence that the lateral and medial pathways differentially affect dentate activity upon stimulation. Moreover, the effects of either pathway on dentate activity can be manipulated with a number of pharmacological tools that appear selective for only one or the other (see chapter by Bramham in this volume). A critical question that seems yet to be addressed is whether or not the position of a single entorhinal synapse along the proximo-distal extent of the apical dendrite of a granule cells is a necessary element to explain the function of that input. Furthermore, it is not clear whether fibers from LEC and MEC must target individual granule cells in specific convergent patterns in order to transfer information from entorhinal cortex normally. In view of the primate organization, this seems an unlikely proposition. Surprisingly, there appears no solid anatomical evidence for the generally accepted view that all granule cells receive convergent inputs from both pathways. No clear conclusions can be derived with respect to a transverse organization of this projection. From this overview, the generally accepted view that there is no transverse organization appears acceptable. In contrast, the entorhinal-dentate system shows a striking organization along the longitudinal axis of

58

the dentate gyrus such that lateral and caudal portions of the entorhinal cortex project preferentially to the dorsal (posterior in primates) part of the dentate gyrus, whereas more medial and anterior portions of the entorhinal cortex projects preferentially to more ventral (anterior in primates) portions of the dentate gyrus. It is of interest that the projections to the dorsal (posterior) part generally show a much more dispersed terminal distribution, up to almost 60% of the entire length, whereas those to the ventral (anterior) dentate appear more restricted to maximally 30% of the long axis. This appears to be in correspondence to the differential longitudinal spread of the intrinsic CA3 associative system. This clear cut organization along the long axis has triggered numerous experiments trying to relate differences in inputs to the different lateral-to-medial bands in entorhinal cortex to a functional differentiation along the long axis of the hippocampus, as initially proposed (Witter et al., 1989a). It has been established in rats that dorsal hippocampal lesions selectively interfere with certain forms of spatial learning and memory (Moser and Moser, 1998). Likewise, lesions of the related entorhinal input zone produces spatial deficits (Steffenach et al., 2005). In contrast, ventral hippocampal lesions do not result in clear spatial deficits but lead to more motivational deficits (Kjelstrup et al., 2002) and again comparable behavioral effects have been reported to result from lesions of the corresponding entorhinal input zone (Steffenach et al., 2005). In non-human primates this issue has not been addressed in any detail, but human functional imaging experiments indeed did report functional differences along the longitudinal axis. However, these experimental data mainly support differences in encoding and retrieval processes and as such are not easily comparable to the rodent data. In order to reconcile this current apparent mismatch between animal and human data, we most likely have to take the strong intrinsic wiring of both the hippocampus and the entorhinal cortex into account (Witter and Amaral, 2004; Witter and Moser, 2006). In particular the intrinsic interactions of the hippocampus, both along the transverse and the long axis may complicate our understanding of this issue. In order to be able to value the relevance of the organizational and

topographical issues described here, we have to understand the unique, and most likely complementary, contributions of each subfield of the hippocampus and the entorhinal cortex, as well as the roles of the longitudinal associational connections.

Acknowledgments This paper is based on large quantities of work carried out by many colleagues over the years. I am greatly indebted to them and their well-prepared publications on this subject. Moreover, for the experimental data described, I owe my former technician, Mrs. B. Jorritsma-Byham, who performed most of the analyses, assisted by a sizable group of under-graduate students. Dr. G. Doctor developed the statistical routines and performed the measurements on the radial position of terminal fields. Finally, I want to thank Helen Scharfman for her valuable discussions and contributions to the manuscript. This work has been supported in part by grants from HFSPO and the Dutch Organization for Scientific Research (NWO).

References van Abeelen, J.H. (1989) Genetic control of hippocampal cholinergic and dynorphinergic mechanisms regulating novelty-induced exploratory behavior in house mice. Experientia, 45(9): 839–845. Amaral, D.G., Insausti, R. and Cowan, W.M. (1987) The entorhinal cortex of the monkey. I. Cytoarchitectonic organization. J. Comp. Neurol., 264(3): 326–355. Amaral, D.G. and Witter, M.P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31(3): 571–591. Bakst, I., Avendano, C., Morrison, J.H. and Amaral, D.G. (1986) An experimental analysis of the origins of somatostatin-like immunoreactivity in the dentate gyrus of the rat. J. Neurosci., 6(5): 1452–1462. Bakst, I., Morrison, J.H. and Amaral, D.G. (1985) The distribution of somatostatin-like immunoreactivity in the monkey hippocampal formation. J. Comp. Neurol., 236(4): 423–442. Baks-Te-Bulte, L., Wouterlood, F.G., Vinkenoog, M. and Witter, M.P. (2005) Entorhinal projections terminate onto principal neurons and interneurons in the subiculum: a quantitative electron microscopical analysis in the rat. Neuroscience, 136(3): 729–739.

59 Bartesaghi, R., Guidi, S., Severi, S., Contestabile, A. and Ciani, E. (2003) Sex differences in the hippocampal dentate gyrus of the guinea-pig before puberty. Neuroscience, 121(2): 327–339. Blackstad, T.W. (1958) On the termination of some afferents to the hippocampus and fascia dentata. An experimental study in the rat. Acta Anat., 35: 202–214. Boyett, J.M. and Buckmaster, P.S. (2001) Somatostatinimmunoreactive interneurons contribute to lateral inhibitory circuits in the dentate gyrus of control and epileptic rats. Hippocampus, 11(4): 418–422. Cappaert, N.L.M., Wadman, W.J. and Witter, M.P. (2005) Spatiotemporal analyses of interactions between entorhinal and CA1 projections to the subiculum of the rat. Soc. Neurosci. Abstr., 73.13. Choi, Y.S., Lee, M.Y., Sung, K.W., Jeong, S.W., Choi, J.S., Park, H.J., Kim, O.N., Lee, S.B. and Kim, S.Y. (2003) Regional differences in enhanced neurogenesis in the dentate gyrus of adult rats after transient forebrain ischemia. Mol. Cells, 16(2): 232–238. Chrobak, J.J. and Amaral, D.G. (2006) The entorhinal cortex of the monkey. VII. Intrinsic connections. J. Comp. Neurol., 500(4): 612–633. Deller, T. (1998) The anatomical organization of the rat fascia dentate — new aspects of laminar organization as revealed by anterograde tracing with Phaseolus vulgaris-leucoagglutinin (Pha-L). Anat. Embryol., 197(2): 89–103. Deller, T. and Leranth, C. (1990) Synaptic connections of neuropeptide Y (NPY) immunoreactive neurons in the hilar area of the rat hippocampus. J. Comp. Neurol., 300(3): 433–447. Deller, T., Martinez, A., Nitsch, R. and Frotscher, M. (1996) A novel entorhinal projection to the rat dentate gyrus: direct innervation of proximal dendrites and cell bodies of granule cells and GABAergic neurons. J. Neurosci., 16(10): 3322–3333. van Der Linden, S. and Lopes da Silva, F.H. (1998) Comparison of the electrophysiology and morphology of layers III and II neurons of the rat medial entorhinal cortex in vitro. Eur. J. Neurosci., 10(4): 1479–1489. Desmond, N.L., Scott, C.A., Jane, J.A. and Levy, W.B. (1994) Ultrastructural identification of entorhinal cortical synapses in CA1 stratum lacunosum-moleculare of the rat. Hippocampus, 4(5): 594–600. Dolorfo, C.L. and Amaral, D.G. (1998) Entorhinal cortex of the rat: topographic organization of the cells of origin of the perforant path projection to the dentate gyrus. J. Comp. Neurol., 398(1): 25–48. Fredens, K., Stengaard-Pedersen, K. and Larsson, L.I. (1984) Localization of enkephalin and cholecystokinin immunoreactivities in the perforant path terminal fields of the rat hippocampal formation. Brain Res., 304(2): 255–263. Fyhn, M., Molden, S., Witter, M.P., Moser, E.I. and Moser, M.B. (2004) Spatial representation in the entorhinal cortex. Science, 305(5688): 1258–1264. Gaarskjaer, F.B. (1978) Organization of the mossy fiber system of the rat studied in extended hippocampi. I. Terminal area

related to number of granule and pyramidal cells. J. Comp. Neurol., 178(1): 49–72. Goldowitz, D., White, W.F., Steward, O., Lynch, G. and Cotman, C. (1975) Anatomical evidence for a projection from the entorhinal cortex to the contralateral dentate gyrus of the rat. Exp. Neurol., 47(3): 433–441. van Groen, T., Miettinen, P. and Kadish, I. (2003) The entorhinal cortex of the mouse: organization of the projection to the hippocampal formation. Hippocampus, 13(1): 133–149. Guidi, S., Severi, S., Ciani, E. and Bartesaghi, R. (2006) Sex differences in the hilar mossy cells of the guinea-pig before puberty. Neuroscience, 139(2): 565–576. Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser, E.I. (2005) Microstructure of a spatial map in the entorhinal cortex. Nature, 436(7052): 801–806. Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple GABAergic innervation of granule cells in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5(5): 411–429. Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5(5): 395–410. Hara, H., Onodera, H., Kogure, K. and Akaike, N. (1990) The regional difference of neuronal susceptibility in the dentate gyrus to hypoxia. Neurosci. Lett., 115(2–3): 189–194. Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J. (2005) Major dissociation between medial and lateral entorhinal input to dorsal hippocampus. Science, 308(5729): 1792–1794. Haug, F.M. (1976) Sulphide silver pattern and cytoarchitectonics of parahippocampal areas in the rat. Special reference to the subdivision of area entorhinalis (area 28) and its demarcation from the pyriform cortex. Adv. Anat. Embryol. Cell Biol., 52(4): 3–73. Hevner, R.F. and Kinney, H.C. (1996) Reciprocal entorhinalhippocampal connections established by human fetal midgestation. J. Comp. Neurol., 372(3): 384–394. Hjorth-Simonsen, A. (1971) Hippocampal efferents to the ipsilateral entorhinal area: an experimental study in the rat. J. Comp. Neurol., 142(4): 417–437. Hjorth-Simonsen, A. (1972) Projection of the lateral part of the entorhinal area to the hippocampus and fascia dentata. J. Comp. Neurol., 146(2): 219–232. Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. J. Comp. Neurol., 144(2): 215–232. Hjorth-Simonsen, A. and Zimmer, J. (1975) Crossed pathways from the entorhinal area to the fascia dentata. I. Normal in rabbits. J. Comp. Neurol., 161(1): 57–70. Hyman, B.T., Kromer, L.J. and Van Hoesen, G.W. (1988) A direct demonstration of the perforant pathway terminal zone in Alzheimer’s disease using the monoclonal antibody Alz-50. Brain Res., 450(1–2): 392–397. Hyman, B.T., Van Hoesen, G.W. and Damasio, A.R. (1987) Alzheimer’s disease: glutamate depletion in the hippocampal perforant pathway zone. Ann. Neurol., 22(1): 37–40.

60 Hyman, B.T., Van Hoesen, G.W., Kromer, L.J. and Damasio, A.R. (1986) Perforant pathway changes and the memory impairment of Alzheimer’s disease. Ann. Neurol., 20(4): 472–481. Insausti, R., Herrero, M.T. and Witter, M.P. (1997) Entorhinal cortex of the rat: cytoarchitectonic subdivisions and the origin and distribution of cortical efferents. Hippocampus, 7(2): 146–183. Insausti, R., Tunon, T., Sobreviela, T., Insausti, A.M. and Gonzalo, L.M. (1995) The human entorhinal cortex: a cytoarchitectonic analysis. J. Comp. Neurol., 355(2): 171–198. Ishizuka, N. (2001) Laminar organization of the pyramidal cell layer of the subiculum in the rat. J. Comp. Neurol., 435(1): 89–110. Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization of intrahippocampal projections originating from CA3 pyramidal cells in the rat. J. Comp. Neurol., 295(4): 580–623. Jaarsma, D., Postema, F. and Korf, J. (1992) Time course and distribution of neuronal degeneration in the dentate gyrus of rat after adrenalectomy: A silver impregnation study. Hippocampus, 2(2): 143–150. Kemppainen, S., Jolkkonen, E. and Pitkanen, A. (2002) Projections from the posterior cortical nucleus of the amygdala to the hippocampal formation and parahippocampal region in rat. Hippocampus, 12(6): 735–755. Kjelstrup, K.G., Tuvnes, F.A., Steffenach, H.A., Murison, R., Moser, E.I. and Moser, M.B. (2002) Reduced fear expression after lesions of the ventral hippocampus. Proc. Natl. Acad. Sci. U.S.A., 99(16): 10825–10830. Ko¨hler, C. (1985) A projection from the deep layers of the entorhinal area to the hippocampal formation in the rat brain. Neurosci. Lett., 56(1): 13–19. Krettek, J.E. and Price, J.L. (1977) Projections from the amygdaloid complex and adjacent olfactory structures to the entorhinal cortex and to the subiculum in the rat and cat. J. Comp. Neurol., 172(4): 723–752. Kunzle, H. (2002) Distribution of perihippocampo-hippocampal projection neurons in the lesser hedgehog tenrec. Neurosci. Res., 44(4): 405–419. Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Afferent and efferent synaptic connections of somatostatin-immunoreactive neurons in the rat fascia dentata. J. Comp. Neurol., 295(1): 111–122. Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997) Connections of the hippocampal formation in humans. I. The mossy fiber pathway. J. Comp. Neurol., 385(3): 325–351. Lorente de No´, R. (1933) Studies on the structure of the cerebral cortex. J. Psychol. Neurol., 45(6): 26–438. Majak, K. and Pitkanen, A. (2003) Projections from the periamygdaloid cortex to the amygdaloid complex, the hippocampal formation, and the parahippocampal region: a PHAL study in the rat. Hippocampus, 13(8): 922–942. Matthews, D.A., Cotman, C. and Lynch, G. (1976) An electron microscopic study of lesion-induced synaptogenesis in the dentate gyrus of the adult rat. I. Magnitude and time course of degeneration. Brain Res., 115(1): 1–21.

Morys, J., Sadowski, M., Barcikowska, M., Maciejewska, B. and Narkiewicz, O. (1994) The second layer neurones of the entorhinal cortex and the perforant path in physiological ageing and Alzheimer’s disease. Acta Neurobiol. Exp. (Wars.), 54(1): 47–53. Moser, E., Moser, M.B. and Andersen, P. (1993) Spatial learning impairment parallels the magnitude of dorsal hippocampal lesions, but is hardly present following ventral lesions. J. Neurosci., 13(9): 3916–3925. Moser, M.B. and Moser, E.I. (1998) Functional differentiation in the hippocampus. Hippocampus, 8(6): 608–619. Naber, P.A., Lopes da Silva, F.H. and Witter, M.P. (2001) Reciprocal connections between the entorhinal cortex and hippocampal fields CA1 and the subiculum are in register with the projections from CA1 to the subiculum. Hippocampus, 11(2): 99–104. Nafstad, P.H. (1967) An electron microscope study on the termination of the perforant path fibres in the hippocampus and the fascia dentata. Z. Zellforsch. Mikrosk. Anat., 76(4): 532–542. Nitsch, R. and Leranth, C. (1991) Neuropeptide Y (NPY)immunoreactive neurons in the primate fascia dentata; occasional coexistence with calcium-binding proteins: a light and electron microscopic study. J. Comp. Neurol., 309(4): 430–444. Ramo´n y Cajal, S. (1911) Histologie du SystemeNerveux de l’Homme et des Vertebres. Maloine, Paris. Ruth, R.E., Collier, T.J. and Routtenberg, A. (1982) Topography between the entorhinal cortex and the dentate septotemporal axis in rats. I. Medial and intermediate entorhinal projecting cells. J. Comp. Neurol., 209(1): 69–78. Ruth, R.E., Collier, T.J. and Routtenberg, A. (1988) Topographical relationship between the entorhinal cortex and the septotemporal axis of the dentate gyrus in rats. II. Cells projecting from lateral entorhinal subdivisions. J. Comp. Neurol., 270(4): 506–516. Sargolini, F., Fyhn, M., Hafting, T., McNaughton, B.L., Witter, M.P., Moser, M.B. and Moser, E.I. (2006) Conjunctive representation of position, direction, and velocity in entorhinal cortex. Science, 312(5774): 758–762. Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the molecular layer have lower thresholds for synaptic activation by perforant path than granule cells. J. Neurosci., 11(6): 1660–1673. Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and Goodman, J.H. (2002) Structural and functional asymmetry in the normal and epileptic rat dentate gyrus. J. Comp. Neurol., 454(4): 424–439. Schwartz, S.P. and Coleman, P.D. (1981) Neurons of origin of the perforant path. Exp. Neurol., 74(1): 305–312. Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule cells are normal features of the fetal and adult dentate gyrus of both monkey and human hippocampal formations. Brain Res., 405(1): 169–174. Soriano, E. and Frotscher, M. (1993) GABAergic innervation of the rat fascia dentata: a novel type of interneuron in the

61 granule cell layer with extensive axonal arborization in the molecular layer. J. Comp. Neurol., 334(3): 385–396. Stanfield, B.B. and Cowan, W.M. (1979) The morphology of the hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol., 185(3): 393–422. Steffenach, H.A., Witter, M., Moser, M.B. and Moser, E.I. (2005) Spatial memory in the rat requires the dorsolateral band of the entorhinal cortex. Neuron, 45(2): 301–313. Stephan, H. (1975) Allocortex. In: Borgman W. (Ed.), Handbuch der Mikroskopischen Anatomie des Menschen. SpringerVerlag, Berlin. Steward, O. (1976) Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat. J. Comp. Neurol., 167(3): 285–314. Steward, O. and Scoville, S.A. (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol., 169(3): 347–370. Swanson, L.W. and Ko¨hler, C. (1986) Anatomical evidence for direct projections from the entorhinal area to the entire cortical mantle in the rat. J. Neurosci., 6(10): 3010–3023. Tamamaki, N. (1997) Organization of the entorhinal projection to the rat dentate gyrus revealed by Dil anterograde labeling. Exp. Brain Res., 116(2): 250–258. Tamamaki, N. and Nojyo, Y. (1993) Projection of the entorhinal layer II neurons in the rat as revealed by intracellular pressureinjection of neurobiotin. Hippocampus, 3(4): 471–480. Uva, L., Gruschke, S., Biella, G., de Curtis, M. and Witter, M.P. (2004) Cytoarchitectonic characterization of the parahippocampal region of the guinea pig. J. Comp. Neurol., 474(2): 289–303. Van Hoesen, G.W. and Pandya, D.N. (1975) Some connections of the entorhinal (area 28) and perirhinal (area 35) cortices of the rhesus monkey. III. Efferent connections. Brain Res., 95(1): 39–59. Voogd, J., Nieuwenhuis, R., van Dongen, P.A.M. and Ten Donkelaar, H.J. (1998) Mammals. In: Dubbeldam J.L., van Dongen P.A.M. and Voogd J. (Eds.), The Central Nervous System of Vertebrates, Vol. 3. Springer-Verlag, Berlin, pp. 1637–2098. Wang, X. and Lambert, N.A. (2003) Membrane properties of identified lateral and medial perforant pathway projection neurons. Neuroscience, 117(2): 485–492. White, W.F., Nadler, J.V., Hamberger, A., Cotman, C.W. and Cummins, J.T. (1977) Glutamate as transmitter of hippocampal perforant path. Nature, 270(5635): 356–357.

Witter, M.P. (1989) Connectivity of the rat hippocampus. In: Chan-Palay V. and Ko¨hler C. (Eds.), The Hippocampus — New Vistas. Allen R. Liss, New York, pp. 53–69. Witter, M.P. (1990) Organization of the entorhinal projections to the hippocampal formation of the rat. Soc. Neurosci. Abstr., 16: 124. Witter, M.P. (2006) Connections of the subiculum of the rat: topography in relation to columnar and laminar organization. Behav. Brain Res., 174(2): 251–264. Witter, M.P. and Amaral, D.G. (1991) Entorhinal cortex of the monkey. V. Projections to the dentate gyrus, hippocampus, and subicular complex. J. Comp. Neurol., 307(3): 437–459. Witter, M.P. and Amaral, D.G. (2004) Hippocampal formation. In: Paxinos G. (Ed.), The Rat Nervous System (3rd Ed.). Elsevier Academic Press, San Diego, CA, pp. 635–704. Witter, M.P. and Groenewegen, H.J. (1984) Laminar origin and septotemporal distribution of entorhinal and perirhinal projections to the hippocampus in the cat. J. Comp. Neurol., 224(3): 371–385. Witter, M.P. and Groenewegen, H.J. (1990) The subiculum: cytoarchitectonically a simple structure, but hodologically complex. Prog. Brain Res., 83: 47–58. Witter, M.P., Groenewegen, H.J., Lopes da Silva, F.H. and Lohman, A.H. (1989a) Functional organization of the extrinsic and intrinsic circuitry of the parahippocampal region. Prog. Neurobiol., 33(3): 161–253. Witter, M.P. and Moser, E.I. (2006) Spatial representation and the architecture of the entorhinal cortex. Trends Neurosci., 29(12): 671–678. Witter, M.P., Van Hoesen, G.W. and Amaral, D.G. (1989b) Topographical organization of the entorhinal projection to the dentate gyrus of the monkey. J. Neurosci., 9(1): 216–228. Witter, M.P., Wouterlood, F.G., Naber, P.A. and van Haeften, T. (2000) Anatomical organization of the parahippocampalhippocampal network. Ann. N.Y. Acad. Sci., 911: 1–24. Woznicka, A., Malinowska, M. and Kosmal, A. (2006) Cytoarchitectonic organization of the entorhinal cortex of the canine brain. Brain Res. Rev., 52(2): 346–367. Wyss, J.M. (1981) An autoradiographic study of the efferent connections of the entorhinal cortex in the rat. J. Comp. Neurol., 199(4): 495–512. Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989) Entorhinal fibers form synaptic contacts on parvalbuminimmunoreactive neurons in the rat fascia dentata. Brain Res., 495(1): 161–166.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 4

Extrinsic afferent systems to the dentate gyrus Csaba Leranth1,2, and Tibor Hajszan1,3 1

Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University School of Medicine, 333 Cedar Street, FMB 312, New Haven, CT 06520, USA 2 Department of Neurobiology, Yale University School of Medicine, New Haven, CT 06510, USA 3 Department of Biophysics, Biological Research Center, Hungarian Academy of Sciences, H-6726 Szeged, Hungary

Abstract: The dentate gyrus is the first stage of the intrahippocampal, excitatory, trisynaptic loop, and a primary target of the majority of entorhinal afferents that terminate in a laminar fashion on granule cell dendrites and carry sensory information of multiple modalities about the external world. The electric activity of the trisynaptic pathway is controlled mainly by different types of local, GABAergic interneurons, and subcortical and commissural afferents. In this chapter we will outline the origin and postsynaptic targets in the dentate gyrus of chemically identified subcortical inputs. These systems are afferents originating from the medial septum/diagonal band of Broca GABAergic and cholinergic neurons, neurochemically distinct types of neurons located in the supramammillary area, serotonergic fibers from the median raphe, noradrenergic afferents from the pontine nucleus, locus ceruleus, dopamine axons originating in the ventral tegmental area, and the commissural projection system. Because of the physiological implications, these afferents are discussed in the context of the glutamatergic innervation of the dentate gyrus. One common feature of the extrinsic dentate afferent systems is that they originate from a relatively small number of neurons. However, the majority of these afferents are able to exert a powerful control over the electrical activity of the hippocampus. This strong influence is due to the fact that the majority of the extrinsic afferents terminate on a relatively small, but specific, populations of neurons that are able to control large areas of the hippocampal formation. Keywords: septum; supramammillary area; median raphe; commissural connection; acetylcholine; dopamine; noradrenaline; serotonin

Even if only a fraction of these connections are active at the same time, a network with virtually endless numbers of possible active patterns emerge (Freund and Buzsaki, 1996). Thus, even if we will be able to understand the complete molecular and biophysical properties of a single hippocampal neuron, in order to understand the function of the hippocampal formation, including the dentate gyrus, as a whole, it is indispensable to elucidate both the intrinsic and extrinsic connections of these cells.

Introduction and anatomical overview It has been known for more than 100 years that neurons present in any cortical area are far from being uniform regarding their morphology and intrinsic and extrinsic connections (Ramon y Cajal, 1893, 1911). This suggests that they are able to interact with each other in a complex fashion. Corresponding author. Tel.: +1 203 785 4748; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63004-0

63

64

The major cell types and the majority of intrinsic and extrinsic connections of the hippocampal formation, which comprises the cornu ammonis, divided into three subfields, CA1–CA3 (Lorente de No, 1934) and the dentate gyrus (fascia dentata and hilus), have been well known since the studies of Ramon y Cajal (1893) and Lorente de No (1934) and have been reviewed extensively (e.g., Amaral and Witter, 1989, 1995; Lopes de Silva et al., 1990). It should be noted, however, that the term hippocampal formation sometimes is used to include the subicular complex and entorhinal cortex (Amaral and Witter, 1995). The three dimensional position of the hippocampal formation in the brain is rather complex. In the majority of species, the hippocampal formation is an elongated structure with its long axis extending in a C-shaped fashion from the septal nuclei of the basal forebrain rostrodorsally, over and behind the diencephalon, to the incipient temporal lobe caudoventrally (see more details in Amaral and Witter, 1995). Although there is still a debate about the classification of hippocampal neurons, the simple terms of principal and non-principal neurons are accepted (Freund and Buzsaki, 1996). The hippocampal principal neurons are the pyramidal and granule cells located in the CA1–CA3 subfields of the ammon’s horn and dentate gyrus, respectively. In addition, the mossy cells of the dentate gyrus can be referred to as principal cells, as discussed further below. The granule cells and pyramidal neurons, with their intrahippocampal connections, are the major components of the so-called trisynaptic circuit of the hippocampus, which is unidirectional and glutamatergic. Layer 2 pyramidal cells of the entorhinal cortex project to granule cells (and to some extent, to CA3 pyramidal cells) via the perforant pathway. Granule cells, via their mossy fiber projection, organized in a laminar fashion, terminate on CA3 pyramidal neurons, which send their Schaffer collaterals to the CA1 pyramidal cells. CA1 pyramidal neurons, in turn, project back to the entorhinal cortex, via the subiculum. As mentioned above, the dentate gyrus is the first stage of the intrahippocampal, excitatory, trisynaptic loop, being the primary target of the majority of entorhinal afferents that terminate in a

laminar fashion on granule cell dendrites (see Chapter 1) and carry sensory information of multiple modalities about the external world. The principal cells of the dentate gyrus are the granule cells,
1 million in rats and 5 million in nonhuman primates (Claiborne et al., 1986; Seress, 1988), and the mossy cells (Amaral, 1978). The small cell bodies of the granule cells (8–12 mm) form the granule cell layer. Granule cells have two main, radially oriented, spiny dendrites emitting several fine branches, which reach the hippocampal fissure. The axons of the granule cells, the mossy fiber axons, originate from the opposite pole of the soma relative to the dendrites, and enter the dentate hilus, where they give rise to several collaterals (Claiborne et al., 1986). Recurrent collaterals periodically enter the granule cell layer, climb along the cell bodies and dendrites of presumed basket cells, and form synapses (Ribak and Peterson, 1991). The main axonal projection of the granule cells leaves the hilar region and courses through the stratum lucidum of the CA3 subfield, where it forms the giant mossy terminals synapsing on the proximal dendrites of pyramidal cells. The dentate hilus is located subjacent to the granule cell layer and extends to the border of the dendritic layer of CA3 that is interposed between the upper (suprapyramidal) and lower (infrapyramidal) blades of the dentate gyrus. The principal and most numerous cell type in the hilus is the mossy cell. These neurons are characterized by their densely spiny dendrites and several thorny excrescences on both the cell body and proximal dendritic shafts and their dendrites are mostly confined to the hilus (Amaral, 1978). However, single mossy dendrites were observed penetrating the stratum moleculare (Scharfman, 1991; Scharfman, 1995; Soltesz and Mody, 1994). Axons of mossy cells innervate the inner third of the dentate molecular layer of both the ipsi- and contralateral dentate gyrus and also have collaterals in the hilus (Amaral, 1978; Laurberg and Sorensen, 1981; Ribak et al., 1985; Buckmaster et al., 1996). These intrahilar collaterals of mossy cells terminate on unidentified dendrites in the hilus and on dendrites of hilar interneurons (Frotscher and Zimmer, 1983a, b; Frotscher et al., 1984; Ribak et al., 1985;

65

Buckmaster et al., 1996). Because both the CA3 pyramidal cells and mossy axons form asymmetric, glutamatergic, excitatory synapses, many authors consider the mossy cells as modified CA3 pyramidal neurons (Soriano and Frotscher, 1994; Scharfman, 1995, 1999). Therefore, mossy cells are considered as excitatory principal cells that provide long-range ipsilateral and commissural projections into the dentate gyrus (Amaral, 1978). The electric activity of the aforementioned excitatory signal loop is controlled mainly by different types of local, GABAergic interneurons (for review, see Freund and Buzsaki, 1996) and subcortical and commissural afferents. In this chapter we will outline our recent knowledge regarding the origin and postsynaptic targets in the dentate gyrus of chemically identified subcortical inputs, including afferents originating from the medial septum/diagonal band of Broca (MSDB) GABAergic and cholinergic neurons, the dentate projection of neurochemically different types of neurons located in the supramammillary area (SUM), serotonergic fibers from the median raphe (MR), noradrenergic afferents from the pontine nucleus, locus ceruleus, dopamine axons originating in the ventral tegmental area, and the commissural system.

Septo-hippocampal connections In the 1970s, it had been shown that anterogradely-transported radiolabeled amino acids, injected into the MSDB, could be detected in hippocampal neurons (Rose and Schubert, 1977). Studies in the seventies and eighties, using the combination of retrograde tracer technique and immunostaining for choline acetyltransferase (ChAT) and glutamate decarboxylase (GAD) showed that there were two major populations of MSDB neurons projecting to the hippocampus (for review, see Leranth and Frotscher, 1989). Later, experiments using anterograde labeling with the lectin PHAL revealed two types of septohippocampal fibers. One has large boutons that occur in clusters, whereas the other has small boutons and arborizes more diffusely (Nyakas et al., 1987). Type 1 axons are immunoreactive for GABA and are processes

of GABAergic, parvalbumin (PA)-containing (Freund, 1989) neurons located in the midline of the MSDB (Kiss et al., 1990). The type 2 axons are GABA-negative (Freund and Antal, 1988), but are immunoreactive for ChAT, the synthesizing enzyme of acetylcholine (Frotscher and Leranth, 1985, 1986; Leranth and Frotscher, 1987). The parent neurons of these axons are also in the MSDB and are positioned in a way so that they surround the septo-hippocampal GABAergic cells (Kiss et al., 1990).

MSDB cholinergic innervation of the dentate gyrus Basic and clinical studies have long recognized the importance of cholinergic mechanisms in cognitive function (Givens and Sarter, 1997), and drugs which increase synaptic acetylcholine levels are currently the most common for the treatment of cognitive deficits associated with disorders such as Alzheimer’s disease, albeit, with limited effectiveness (Benzi and Moretti, 1998). The septo-hippocampal pathway (SH), which originates in the MSDB and shows progressive degeneration in Alzheimer’s disease (Whitehouse et al., 1982), has specifically been implicated in cognitive function. Lesions of the fimbria-fornix, which carry SH cholinergic (Lewis and Shute, 1967) fibers to the hippocampal formation, including the dentate gyrus, interfere both with learning and memory tasks and with generation of the theta rhythm in rats (Brito and Brito, 1990). These deficits can be attenuated by grafting acetylcholine-producing cells to the hippocampus (Dunnett et al., 1982; Dickinson-Anson et al., 1998). Electrophysiological experiments have demonstrated that excitatory, inhibitory, and disinhibitory responses can be recorded as a result of the iontophoretic application of acetylcholine in the dentate gyrus (Wheal and Miller, 1980; Fricke and Prince, 1984). It is not clear, however, from these experiments whether acetylcholine exerts these effects at the level of the primary dendrites, directly onto the perikaryon of principal cells or indirectly, via interneurons (Hounsgaard, 1978). Histochemical staining procedures to visualize putative cholinergic neurons and fibers applying

66

acetylcholinesterase (ACHE) reaction were first used in the dentate gyrus in the early 1960’s (Storm-Mathisen and Blackstad, 1964). These authors described a laminar pattern of ACHE fiber staining in the dentate gyrus; the densest band occurring at the interface between the granule cell layer and the molecular layer. Shute and Lewis (1966) modified this ACHE histochemical technique to permit examination of the dentate gyrus, stained for ACHE, using the electron microscope. Their study revealed several neurons histochemically stained for ACHE, numerous axonal fibers and some synaptic boutons in contact with predominantly dendritic shafts. Biochemical measurements of the levels of ChAT in the dentate gyrus also indicate a supragranular band of intense cholinergic expression (Fonnum, 1970). The first detailed light and electron microscopic studies using immunostaining for ChAT on the cholinergic innervation of the dentate gyrus were performed in the middle 1980s (Frotscher and Leranth, 1985, 1986; Wainer et al., 1984; Clarke, 1985; Leranth and Frotscher, 1987). The light microscopic analyses revealed a dense plexus of ChAT-immunoreactive fibers in the dentate gyrus forming a supragranular band at the interface between the granule cell layer and molecular layer. Very little staining was in the granule cell layer and only a few fibers were located in the hilar region. In addition, ChAT-immunoreactive neurons could also be observed in the dentate hilus (Clarke, 1985; Frotscher et al., 1986). These ChAT-positive cells are rare and non-principal neurons. They are relatively small with round or ovoid perikarya, which give rise to thin spine-free dendrites and are very similar to ChAT-immunoreactive cells in the neocortex of the same animals but were quite different from cholinergic neurons in the basal forebrain, medial septal nucleus, and neostriatum, which were larger and more intensely immunostained. Electron-microscopic analysis of ChAT-containing cells in the hippocampus and fascia dentata revealed that their afferent synaptic contacts are mainly the asymmetric type, and are located on their cell bodies and smooth proximal dendrites. The nuclei of the immunoreactive cells exhibited deep indentations, which are a characteristic of non-pyramidal neurons (Frotscher et al., 1986).

At the electron microscopic level, the vast majority of ChAT-positive synaptic boutons contacts dendritic shafts and forms predominantly symmetric synaptic contacts. In addition, some asymmetric synapses,
11% of the total number of ChAT synapses (Clarke, 1985), could also be observed on dendritic spines. There is a possibility that axons forming symmetric and asymmetric contacts originate from different population of cholinergic neurons; intrinsic and extrinsic. However, it seems unlikely since, lesions of the septohippocampal pathway cause an almost complete removal of cholinergic markers in the dentate gyrus (Mellgren and Srebro, 1973). In the dentate gyrus, the majority of the postsynaptic targets of cholinergic boutons are the granule cells and the ChAT-immunoreactive boutons form both axodendritic and axospinous synapses with these neurons. All of the axospinous synapses observed are asymmetric (Fig. 1). In addition,
5–10% of all postsynaptic elements of cholinergic axons show ultrastructural features of interneurons. A combined electron microscopic, Golgi impregnation and ChAT immunostaining study has shown symmetric synaptic contacts between ChAT-containing axon terminals and non-spiny, smooth dendritic shafts, characteristic for interneurons (Frotscher and Leranth, 1986; Fig. 2). This 5–10% appears to correspond to the proportion of interneurons in the neuropil, suggesting that the interneurons are contacted in a ‘‘quasirandom’’ fashion. Direct evidence that interneurons are indeed innervated by cholinergic fibers is provided by an electron microscopic doubleimmunostaining study that showed ChAT-positive axon terminals forming symmetric synaptic contacts with hilar GAD (glutamic acid decarboxylase)- and SOM (somatostatin)-containing neurons (Leranth and Frotscher, 1987). In addition to the cholinergic innervation of dentate granule cells and interneurons, MSDB cholinergic cells innervate mossy cells densely. A correlated light and electron microscopic doubleimmunostaining study has demonstrated numerous axosomatic synaptic contacts between ChATcontaining axon terminals and calcitonin generelated peptide (CGRP)-immunoreactive neurons in the dentate hilar area (Deller et al., 1999). CGRP

67

Fig. 1. Low (Panel a) and high power (Panel b) electron micrographs (received from Dr. Michael Frotscher) show the result of a combined Golgi impregnation and ChAT immunostaining experiment. On Panel a, Golgi-impregnated (gold-toned) spiny granule cell dendrites are seen. Arrow on the same panel points at a ChAT immunoreactive bouton contacting the spine head of the dendrite (D). Panel b shows the asymmetric synaptic contact (arrow) between the two profiles. Bar scales ¼ 1 mm.

of these neurons could modulate granule cell excitability throughout large portions of the dentate gyrus.

MSDB GABAergic innervation of the dentate gyrus

Fig. 2. Electron micrograph taken from the molecular layer of the dentate gyrus immunostained for ChAT. A ChAT-immunoreactive axon terminal forms asymmetric synaptic contact with a non-spiny dendritic shaft (D). The lack of dendritic spines indicates that this dendrite is the process of an interneuron. Asterisks label axon terminals forming asymmetric synapses with the same dendrite. Bar scale ¼ 1 mm.

is an accepted marker for mossy cells (Freund et al., 1997). Since mossy cells project for a long distance along the longitudinal axis of the hippocampus (Ribak et al., 1985), the cholinergic innervation

It has been shown that when the perforant path is stimulated with a brief test pulse, it evokes excitatory postsynaptic potentials (EPSPs) in the granule cell dendrites, which could be recorded collectively in the dentate hilus as a positive-going field potential. At greater stimulus intensities, a negative-going component is superimposed on the positive-going field potential, as the result of the synchronous activation of many granule cells. This component is referred to as a population spike (PS). When a conditioning pulse is applied to the medial septum just prior to a perforant path test pulse, the EPSP remains unchanged, but the PS is greater than that expected from the sum of the potentials evoked by either stimulus alone (Alvarez-Leefmans and Gardner-Medwin, 1975;

68

Fantie and Goddard, 1982). Thus, the septohippocampal input appears to facilitate the discharge of granule cells. An interesting and unexpected finding of these studies was that stimulation of the medial septum alone, at a site that produced the facilitation of the PS in the dentate gyrus, did not necessarily evoke a field potential in the dentate gyrus. This finding, together with the failure of medial septal stimulation to affect the perforant path-evoked field potential, upon which the PS is superimposed suggests that the facilitation does not result from summation of excitatory input to the granule cells. Two mechanisms that have been proposed by Buzsaki (1984) to account for these effects are: (1) medial septal input acts directly on granule cells to facilitate activation (Robinson and Racine, 1984), and (2) the medial septum acts indirectly, via the inhibitory interneurons responsible for feed-forward and feed-back inhibition (Kandel et al., 1961; Anderson and Eccles, 1962; Mosko et al., 1973; Buzsaki and Eidelburg, 1982). Anatomical studies of the septo-dentate termination patterns show that the innervation of the supragranular and molecular layers are considerably less dense than the innervation of the dentate hilus and subgranular layer (Lynch et al., 1978; Chandler and Crutcher, 1983). Although MSDB fibers in the supragranular and molecular layers are likely to terminate on granule cells, fibers in the subgranular and hilar region appear to terminate on cells having the morphological characteristics of interneurons (Rose and Schubert, 1977; Chandler and Crutcher, 1983). If SH fibers form excitatory connections onto inhibitory interneurons, their activation might synchronize granule cell responses, resulting in a larger population spike from a subsequent perforant path input (Buzsaki, 1984). Alternatively, if they form inhibitory connections onto the interneurons, septal activation might reduce both tonic and feed-forward inhibition of the granule cells, allowing a larger number of these neurons to be activated by the perforant path input, again resulting in a larger population spike. To determine the validity of these hypotheses, Bilkey and Goddard (1985) performed a very elegant study. They have shown that a conditioning pulse to the medial septum, although eliciting no field potential of its own,

facilitated the granule cell population spike evoked by perforant path stimulation, and infusion into the dentate hilus of a GABAA receptor antagonist, picrotoxin, blocked the facilitation. This observation suggests that the facilitatory effect of MSDB stimulation is mediated through an inhibitory connection from the MSDB onto inhibitory interneurons in the dentate gyrus, and that this connection may utilize the neurotransmitter GABA. Indeed, it has been demonstrated in a series of concomitant morphological studies that SH (including the septo-dentate) GABAergic terminals always terminate on GABAergic interneurons, in the rat (Freund and Antal, 1988; Gulyas et al., 1990) and monkey hippocampus (Gulyas et al., 1991). Experiments using double immunostaining for PHAL and markers for different subsets of interneurons have revealed that all examined subpopulations of hippocampal GABAergic interneurons, including those co-expressing parvalbumin, calbindin, SOM, neuropeptide Y, cholecystokinin, and vasoactive intestinal polypeptide, receive input from GABAergic septohippocampal afferents (Freund and Antal, 1988; Gulyas et al., 1990; Miettinen and Freund, 1992a, b; Acsady et al., 1993). The typical innervation pattern is multiple, ‘‘climbing fiber-like’’ contacts on the soma, and proximal and distal dendrites of the postsynaptic interneurons (Fig. 3). All these contacts have proved to be symmetrical synapses. The area where most of the interneurons appear to receive septal GABAergic input is the CA3 subfield, particularly strata oriens and pyramidale (Rose and Schubert, 1977; Nyakas et al., 1987; Freund and Antal, 1988; Gaykema et al., 1990). The dentate hilus is also heavily innervated, but in the CA1 region, a relatively smaller proportion of the interneurons receive multiple synaptic inputs. No systematic differences have been observed in the termination pattern of septal afferents between the dorsal and ventral hippocampus.

Interactions between the septohippocampal cholinergic and GABAergic systems In order to better understand the function of the septohippocampal projection system that contains

69

Fig. 3. Light micrograph (kindly provided by Dr. Attila Gulyas) shows the result of a combined anterograde tracing and calretinin immunostaining study, in the dentate gyrus. The anterograde tracer, biotinylated dextran amine (BDA) was injected into the medial septum diagonal band. Large boutons of BDA-containing axons form multiple, basket-like, putative synaptic contacts with the soma of calretinin-immunoreactive (labeled with a brown diaminobenzidine reaction product) neurons. One of these cells (arrows) located in the supragranular layer (SgL) the other is at the border between the granule cell layer (GcL) and dentate hilar area (H). Bar scale ¼ 50 mm. (See Color Plate 4.3 in color plate section.)

a cholinergic and a GABAergic component, the anatomical and functional interaction between these two systems must be briefly discussed, at the level of the MSDB. It has to be noted that intrinsic cholinergic mechanisms operating within the MSDB are also critical for learning and memory. Infusion of muscarinic agonists directly into the MSDB elicit continuous hippocampal theta rhythm (Monmaur and Breton, 1991; Lawson and Bland, 1993) and facilitate learning and memory-related behaviors both in young (Givens and Olton, 1990) and aged rats (Markowska et al., 1995). Also, intraseptal infusions of muscarinic agonists can alleviate systemic scopolamine-induced amnesia, suggesting that the MSDB is a critical locus for the mnemonic effects of muscarinic drugs (Givens and Olton, 1995). In general, it is assumed that improvements in MSDB-related learning and memory tasks occur as a result of an increase in hippocampal

acetylcholine (ACh) release (Monmaur and Breton, 1991; Givens and Olton, 1994, 1995; Apartis et al., 1998; Bland and Oddie, 1998; Dickinson-Anson et al., 1998). As such, it has been presumed that the memory-enhancing effects of intraseptal administration of muscarinic agonists occurs due to increased firing of MSDB cholinergic neurons (Markowska et al., 1995; Givens and Sarter, 1997). These assumptions have been based on earlier studies that reported a muscarinic receptor-mediated increase in firing of MSDB neurons (Dutar et al., 1983; Lamour et al., 1984), which lead to the hypothesis that Ach, via muscarinic receptors, has a positive feed-back effect on MSDB cholinergic neurons (Dutar et al., 1983; Lamour et al., 1984; see review in, Wu et al., 2004). However, a recent combined electrophysiological and morphological study (Wu et al., 2004), using a novel fluorescent labeling technique to selectively visualize live septohippocampal cholinergic neurons has demonstrated that administration of muscarinic agonists to the MSDB do not excite septo-hippocampal cholinergic neurons, instead they inhibit a subpopulation of them. In contrast, septo-hippocampal GABAergic neurons (visualized by retrograde tracing and parvalbumin immunostaining techniques) are profoundly excited by muscarine administration into the MSDB. Thus the cognition enhancing effects of muscarinic drugs in the MSDB cannot be attributed to an increase in hippocampal ACh release. Instead, disinhibitory mechanisms (Freund and Antal, 1988), due to increased impulse flow in the septo-hippocampal GABAergic pathway, may underlie the cognitionenhancing effects of muscarinic agonists. In support of this view is the morphological observation that axon collaterals of the septo-hippocampal cholinergic neurons heavily innervate MSDB septo-hippocampal GABAergic neurons (Leranth and Frotscher, 1989; Fig. 4). These connections could be involved in the aforementioned process.

Supramamillo-dentate connections It has been shown that prestimulation of supramammillary (SUM) neurons significantly

70

Fig. 4. Electron micrograph shows the result of a double immunostaining experiment for choline acetyltransferase (ChAT) and glutamic acid decarboxylase (GAD), in the rat medial septum diagonal band of Broca. Immunoreactivity for ChAT and GAD was visualized with two contrasting immunomarker, diaminobenzidine reaction and ferritin labeling, respectively. A ChAT-immunoreactive bouton forms asymmetric synaptic contact (arrowheads) with a GAD immunoreactive dendrite. Bar scale ¼ 1 mm.

enhances perforant path-elicited population spikes in the fascia dentata (Mizumori et al., 1989), an effect that could be mimicked by glutamate injections to the lateral supramammillary area (Carre and Harley, 1991). Moreover, Dahl and Winson (1986) speculated about a neuronal control mechanism, a ‘‘gate,’’ mediated by supramammillary afferents that would facilitate information flow in the rat dentate gyrus, in a behavior-dependent manner. Since the mid-1970s, it has been known that a projection exists between the SUM and

hippocampus in the rat (Segal and Landis, 1974). Later, Amaral and Cowan (1980) showed that horseradish peroxidase (HRP) injected into the monkey hippocampus also results in labeled cells in the SUM. Furthermore, it was noted that more cells ipsilateral to the site of injection were labeled than cells in the contralateral SUM and a large number of SUM efferents terminate in the dentate gyrus since, local application of the retrograde tracer, Evans blue to the upper blade of the dentate gyrus produced labeled cell bodies in the SUM (Harley et al., 1983). Labeling was observed throughout the rostrocaudal aspect of the SUM. Therefore, Amaral and Cowan (1980) and Harley et al. (1983) postulated that afferents from the SUM to the hippocampus involve at least as many cells in the SUM as the septal neurons, which give rise to the septohippocampal projection. The specific pathway containing SUM efferents to the hippocampus remains debatable. The medial forebrain bundle (Haglund et al., 1984) and fornix (Veazey et al., 1982) have been suggested as likely candidates. More specifically, Veazey et al. (1982) have postulated that the medial forebrain bundle carries SUM efferents to septal nuclei, whereas the fornix carries projection from the SUM to the hippocampus. At the level of dorsal hippocampus, labeled fibers have been described in the fimbria (Haglund et al., 1984) or the subcallosal fornix (Wiss et al., 1979). Regarding the neurochemical nature of the SUM-hippocampal pathway, studies were able to demonstrate that calretinin-immunoreactive axon terminals in the inner molecular layer of the dentate gyrus (and in the pyramidal layer of CA2) originate in the SUM. Colocalization studies provided evidence that these large projecting neurons contained both calretinin and substance P (SP) but lacked GABA (Nitsch and Leranth, 1993). This observation indicates that the nature of the hypothalamo-hippocampal projection to the dentate is likely to be excitatory. In fact, SP-containing, long projection systems have been shown to exert excitatory actions (Nicoll et al., 1980). Conversely, in the rat, the functional properties of the hypothalamo-hippocampal afferent system have been reported to be at least partially inhibitory (Segal, 1979). The study by Segal (1979) raises a question

71

of how can the physiological action on hippocampal principal neurons of a putatively excitatory pathway become inhibitory. A possible explanation for this is that the supramammillo-hippocampal afferents, in addition to terminating on principal neurons (Magloczky et al., 1994), innervate hippocampal GABAergic interneurons. Indeed, in a study dealing with the innervation of the primate hippocampal formation, thick and varicose SP-immunoreactive axons, forming basket-like structures were identified adjacent to the granule cell layer and in the hilar area of the dentate gyrus and in the molecular layer of the middle portion of CA3 (Nitsch and Leranth, 1994). The location of these basket-like structures outside the principal cell layers indicates that they contact hippocampal non-principal neurons. Furthermore, these SP-containing axons disappear after fimbriafornix transection, indicating that they are of extrinsic origin (Nitsch and Leranth, 1994). A subsequent study on non-human primates (Leranth and Nitsch, 1994), applying correlated light and electron microscopic immunocytochemical, double-labeling technique for SP and parvalbumin and SP and calbindin and subsequent postembedding GABA-immunostaining revealed that this supramammillo-hippocampal afferent system establishes multiple, exclusively asymmetric synapses with three specific subpopulations of non-pyramidal cells: (1) a small portion of parvalbumin-containing basket cells located in or adjacent to the granule cell layer of the dentate gyrus (Fig. 5), which, therefore, inhibit only a subpopulation of granule cells; (2) some of the calbindin-immunoreactive neurons located in the hilar area and in the granule cell layer (Fig. 6); and (3) calbindin-positive cells occurring exclusively in the stratum moleculare of the middle portion of the CA3 subfield. Postembedding immunostaining for GABA revealed that the aforementioned calbindin-containing cells in area CA3 are GABAergic inhibitory neurons. The results of this study indicate that supramammillary afferents, a portion of which is glutamatergic (Kiss et al., 2000) can effectively filter the information flow at different points along the trisynaptic circuit in the monkey hippocampal formation. Dentate granule cells, which are only stimulated by SUM afferents,

Fig. 5. Light micrograph taken from a double immunostained vibratome section of the monkey dentate gyrus. Immunoreactivity for substance P was labeled with a dark-blue Ni-diaminobenzidine reaction, while immunostaining for parvalbumin was visualized by a brown diaminobenzidine reaction. The soma and dendrites of the parvalbumin-containing cell embedded into the granule cell layer (GcL) is contacted by several substance P-immunoreactive axon terminals (arrows). Bar scale ¼ 10 mm. (See Color Plate 4.5 in color plate section.)

will transfer excitatory signals differently than those that are controlled by a feed-forward inhibitory mechanism initiated by these fibers (see Fig. 10, in Leranth and Nitsch, 1994). Catecholaminergic brainstem-dentate connections Serotonergic afferents The serotonergic raphe-hippocampal pathway has a powerful effect on hippocampal electric activity, depression-associated synaptic plasticity, and cognitive behavior. Electrophysiological and

72

Fig. 6. Light (Panel a) and electron micrographs (Panels b, c, d) depicted from the monkey dentate gyrus demonstrate the result of a correlated light and electron microscopic double immunostaining for substance P (labeled by a dark-blue Ni-diaminobenzidine reaction) and calbindin (brown diaminobenzidine chromogen). Panel a shows a calbindin-immunoreactive neuron embedded into the granule cell layer forming putative synaptic contacts (A–D) with substance P-containing axon terminals. Electron microscopic analysis of ultrathin sections cut from the same area shows that boutons A and D form robust asymmetric synaptic contacts (arrowheads on panels c and d) with the soma of this calbindin-containing cell (CB on panel b). Gc-granule cell. Bar scales ¼ panel a, 10 mm; panels b–c, 1 mm.

73

pharmacological studies have shown that the typical effect of serotonin on hippocampal neurons is a hyperpolarization evoked by an increase in K+ conductance, but depolarization and reduction of afterhyperpolarization were also reported (for review, see Freund et al., 1990). The serotonergic innervation of the hippocampus originates largely from the MR and a less numerous projection arises from the dorsal raphe (for review, see Tork, 1990) and includes two types of fibers (Kosofsky and Molliver, 1987). In the dentate gyrus, the most numerous are the thin axons with small evenly distributed varicosities that are present in the subgranular area. The other fiber type has larger boutons that form clusters along the hilar border of the stratum granulosum and contact secondary dendritic branches of two cell types: GABAergic, pyramidal-shaped neurons located subjacent to the granule cell layer, and fusiform cells in the hilus. Both types of these GABAergic cells seem to contain calbindin, but none of them are immunoreactive for parvalbumin (Freund et al., 1990; Halasy et al., 1992). Based on these morphological data, it has been suggested that the two types of serotonergic axons have different mechanisms of action in the hippocampus. Axons with small varicosities (originating in the dorsal raphe; Kosofsky and Molliver, 1987) release serotonin at non-synaptic sites, diffusely, and target cells having 5-HT1-2 receptors, to exert a slow, tonic, G-protein-mediated action. The other type of serotonergic axons with large boutons (originating in the median raphe; Kosofsky and Molliver, 1987) always form synaptic contacts with GABAergic interneurons that have 5-HT3 receptors (Halasy et al., 1992). Thus, stimulation of median raphe serotonin neurons could result in fast excitation of these GABAergic cells and an enhanced GABAA receptor-mediated inhibition of granule cells, a mechanism described in the CA1 area (Roppert and Guy, 1991). An important question is whether there is convergence of the MSDB and MR hippocampal inputs on the same hippocampal interneurons. A very elegant morphological study of Miettinen and Freund (1992a) has demonstrated that parvalbumin-containing interneurons are innervated by MSDB GABAergic afferents, but are avoided

by axons originating in the median raphe. On the other hand, calbindin- and, to a smaller extent, cholecystokinin-containing interneurons are targets for both pathways. In some cases the same individual calbindin- or cholecystokinincontaining neurons received multiple contacts from afferents of both MSDB and median raphe origin. Thus, these observations indicate that different subcortical nuclei modulate largely different inhibitory circuits. However, considering the occasional convergence of the two subcortical nuclei not only onto the same type, but also onto the same individual interneurons, the authors proposed that a particular inhibitory function, most probably feed-forward inhibition in the distal dendritic region, is under the control of both pathways. It has to be noted that the MR serotonin system could also effect the hippocampus via an indirect route. The MR serotonergic neurons heavily innervate the MSDB (Leranth and Vertes, 1999) and exert a robust stimulatory effect on parvalbumincontaining septohippocampal GABAergic neurons (Alreja, 1996). These GABAergic septal cells, in turn, selectively innervate hippocampal basket and chandelier cells (Freund and Antal, 1988), which are known to have powerful inhibitory effects on the output sector (soma and axon hillock) of principal neurons (Freund and Antal, 1988). Thus, serotonergic stimulation of the septohippocampal GABA system results in a disinhibition of principal cells. The situation is more complex, because a population of MR serotonergic neurons project to both the hippocampus and MSDB (e.g., McKenna and Vertes, 2001).

Noradrenergic and dopaminergic afferents to the dentate gyrus Noradrenergic fibers in the hippocampus originate from the locus ceruleus. Noradrenaline in the dentate gyrus promotes and permits long-term perforant path potentiation. Furthermore, phasic locus ceruleus activation produces a delayed protein synthesis-dependent long-term potentiation of synaptic plasticity, suggesting a selective role in long-term memory and increases in the perforant

74

path-evoked population spike (see Chapter 10). Potentiation of the perforant path population spike by noradrenaline could be the result of increased granule cell excitability or reduced inhibition from interneurons of granule cells, or both. However, unit recording in the dentate gyrus shows that exogenous noradrenaline inhibits granule cells and excites presumed inhibitory interneurons (Brown et al., 2005). On the other hand, activation of a2- or b-adrenoceptors excites both the interneurons and granule cells (Brown et al., 2005). Increased inhibitory interneuron activity and inhibition of granule cells is inconsistent with the enhanced granule cell responsiveness and enhanced plasticity in the dentate gyrus reported with population recording. Thus, a critical question concerns the action of noradrenaline on the dentate gyrus interneurons. In the hippocampal formation, noradrenergic innervation is particularly dense in areas receiving mossy fiber inputs, including the hilus of the dentate gyrus and stratum lucidum of the CA3 (e.g., Moudy et al., 1993). In these two areas, the majority of noradrenergic varicosities do not make conventional synaptic contacts. However, those that form synapses terminate on dendritic shafts and somata of GABAergic interneurons forming symmetric membrane specializations (Frotscher and Leranth, 1988; Milner and Bacon, 1989). It should be noted, however, that asymmetric synaptic contacts between tyrosine hydroxylase-containing (presumably noradrenergic) axon terminals and dendritic spines were also observed, in the stratum lucidum of CA3 (Frotscher and Leranth, 1988). Symmetric (presumably inhibitory) synapses support earlier physiological studies suggesting that noradrenaline disinhibits hippocampal pyramidal neurons by decreasing the excitability of GABAergic interneurons (e.g., Madison and Nicoll, 1988). In contrast to the very dense noradrenergic innervation of the dentate gyrus, this structure receives only a minor and diffusely distributed dopaminergic projection that arises mainly from the ventral tegmental area (Swanson, 1982). In spite of the well-established effects of dopamine in hippocampus-related mnemonic functions, little is known about the mechanisms of these effects,

which are likely to reside within the hippocampus. Most of our knowledge derives from receptor studies, characterizing the role of D1–D5 receptors. The results of these studies indicate that in the dentate gyrus, via the aforementioned dopamine receptor subtypes, dopamine is involved in depotentiation and serves to maintain synaptic facilitation in recently potentiated pathways. The net effect helps consolidate information storage (e.g., in Manahan-Voughan and Kulla, 2003).

Commissural connections of the dentate gyrus The majority of commissural fibers occupy the inner molecular layer of the dentate gyrus (e.g., Blackstad, 1956). However, some commissural fibers were described that do not follow the ‘‘classical’’ pattern of fiber lamination and terminate in the outer molecular layer (Deller et al., 1995, 1996a, b; Deller, 1998). The main postsynaptic targets of the commissural projection are the granule cell dendrites (e.g., Frotscher and Zimmer, 1983b; Seress and Ribak 1984), but they also innervate interneurons (Frotscher and Zimmer, 1983a; Seress and Ribak, 1984), which were shown to contain GAD (Frotscher et al., 1984). These interneurons do not represent a homogeneous cell population since, they could be distinguished by their different neuropeptide content, e.g., vasoactive intestinal polypeptide (Leranth and Frotscher, 1983), neuropeptide Y (Deller and Leranth, 1990), and parvalbumin (Deller et al., 1994). This arrangement, that commissural fibers terminate on both principal and inhibitory interneurons in the dentate gyrus, represents the anatomical basis of the feed-forward inhibitory mechanism elicited by stimulation of the commissural system (Buzsaki and Eidelburg, 1981; Buzsaki, 1984). Retrograde tracing experiments revealed that the majority of the cells of origin of the commissural system located in the hilus of the fascia dentata (e.g., Hjorth-Simonsen and Laurberg, 1977; Berger et al., 1980) and several groups of neurons have been identified that contribute to this fiber system. The majority of commissural fibers are axon collaterals of the glutamate-containing

75

mossy cells that are also considered as associational/commissural neurons (Ribak et al., 1985; Scharfman and Schwartzkroin, 1988; Scharfman, 1992, 1995; Soriano and Frotscher, 1994). The second major population of cells projecting to the contralateral hippocampus includes different types of GABAergic interneurons. The first hint of a possible contribution of GABAergic interneurons to the commissural projection was published by Seress and Ribak (1983). They showed that in the hilar area, more than 60% of neurons were GAD positive, whereas
80% of hilar neurons project commissurally, suggesting that at least some of the hilar projection must be GABAergic. Later, this proposition was verified by experiments using combination of retrograde tracing and GAD immunochemistry, as well as anterograde labeling and degeneration (Ribak et al., 1986). The GABAergic commissural projection is not homologous. Different populations of commissurally projecting GABAergic cells co-express SOM (Zimmer et al., 1983; Leranth and Frotscher, 1987), neuropeptide Y (Deller and Leranth, 1990), parvalbumin (Goodman and Sloviter, 1992), or other markers (Leranth and Frotscher, 1987; Sloviter and Nilaver, 1987). However, not all neurons of a given cell type have a commissural collateral. While most, if not all mossy cells appear to project bilaterally (e.g., Frotscher et al., 1991; Scharfman, 1992), only 4–5% of the SOM- (Zimmer et al., 1983) and only 2% of the NPY-containing neurons (Deller and Leranth, 1990) project to the contralateral dentate gyrus. In addition to the aforementioned neurons, commissurally projecting CA3c pyramidal neurons were also described (Gottlieb and Cowan, 1973; Laurberg, 1979; Voneida et al., 1981). These CA3c pyramids appear to have commissural collaterals, which arborize in the contralateral hilus, similar to their ipsilateral collaterals (Ishizuka et al., 1990; Li et al., 1994). Based on the existence of a GABAergic commissural projection, Freund and Buzsaki (1996) have suggested that there is a component of direct inhibition in the feed-forward inhibitory response evoked in the dentate gyrus by commissural stimulation (Buzsaki and Eidelburg, 1981; Buzsaki, 1984).

Glutamatergic innervation of the dentate gyrus During the last several decades, visualizing neurons that utilize glutamate as a neurotransmitter has proven to be a difficulty due to the lack of specific glutamatergic markers. As a result, one of the most influential and significant neuronal systems has remained grossly under investigated. At the turn of the millennium, two independent research groups discovered that a protein that has previously been suggested to mediate the Nadependent uptake of inorganic phosphate across the plasma membrane, also transports glutamate into synaptic vesicles (Bellocchio et al., 2000; Takamori et al., 2000). This protein, originally called brain-specific Na+-dependent inorganic phosphate contransporter (BNPI) (Ni et al., 1994), became the first of what are now called vesicular glutamate transporters (VGLUT1), and revolutionized research about the central glutamatergic system. Shortly after this groundbreaking discovery, a second (VGLUT2) and a third (VGLUT3) vesicular glutamate transporter has also been identified and cloned (Aihara et al., 2000; Bai et al., 2001; Fremeau et al., 2002; Gras et al., 2002; Schafer et al., 2002; Takamori et al., 2002). Since then, a vast amount of data has been collected that has led to a better understanding of the brain glutamatergic circuitry, including that in the dentate gyrus.

BNPI/VGLUT1 Distribution of BNPI/VGLUT1 containing fibers and terminals in the rat dentate gyrus has been investigated by several groups (Bellocchio et al., 1998; Fremeau et al., 2001; Kaneko and Fujiyama, 2002; Kaneko et al., 2002). VGLUT1 fiber density is moderate to intense in most regions of the dentate gyrus except the granule cell layer (Kaneko et al., 2002). A more detailed description has come from Bellocchio and colleagues (Bellocchio et al., 1998). The outer two-thirds of the molecular layer label more strongly for BNPI than does the inner one-third. Further examination of the CA3 region under higher magnification revealed coarse granular labeling at the periphery of the pyramidal cell

76

layer, strongly suggestive of mossy fiber synapses. Electron microscopic, immunoperoxidase labeling in stratum lucidum of CA3 showed prominent reaction product in large axon terminals having the morphological characteristics of mossy fiber boutons. In the hilar area of the dentate gyrus, where mossy fiber collaterals also terminate, a few large terminals similar to those in the CA3 region contain BNPI immunoreactivity. Many smaller terminals in the dentate gyrus that form asymmetric (Gray type I) synapses with dendritic spines are also labeled for BNPI. Symmetric synapses do not contain detectable BNPI, supporting a specific role for the protein in excitatory transmission. In addition, many terminals forming asymmetric synapses do not contain detectable BNPI, indicating expression only in a subset of excitatory synapses (Bellocchio et al., 1998). The distribution of BNPI containing boutons in the stratum moleculare suggests that they represent mainly perforant path inputs from the entorhinal cortex. Indeed, hybridization with the BNPI probe resulted in a strong hybridization signal in neuron-enriched regions of the entorhinal cortex (Ni et al., 1995). On the other hand, the BNPI positivity of mossy fiber-like terminals in the hilus and CA3 suggests that they originate from granule cells. Similar to cell bodies elsewhere in the brain, the somata of dentate gyrus granule cells show no detectable BNPI immunoreactivity (Bellocchio et al., 1998). By contrast, a strong BNPI hybridization signal has been observed in the pyramidal neurons of the hippocampus and granule cells of the dentate gyrus, suggesting that these cells synthesize the protein and then transport it to their terminals (Ni et al., 1994, 1995; Fremeau et al., 2001; Herzog et al., 2001). In conclusion, BNPI/VGLUT1 appears to be the primary glutamatergic marker of principal neurons, including dentate granule cells (Varoqui et al., 2002).

DNPI/VGLUT2 Prior to its identification as VGLUT2 (Aihara et al., 2000), this protein has been known as differentiation-associated Na+-dependent inorganic

phosphate cotransporter (DNPI), having similar functions as BNPI (Hisano et al., 2000). In a rather comprehensive study, VGLUT2-positive fibers and terminals in the rat hippocampus have been characterized, and their possible sources of origin defined (Halasy et al., 2004). The highest density of VGLUT2-immunoreactive boutons has been observed in the inner molecular layer, while only scattered VGLUT2 positive fibers have been detected in the other areas of the dentate gyrus, such as the stratum moleculare and the hilus. In general, VGLUT2-immunoreactive boutons establish exclusively asymmetric synapses. Analysis of VGLUT2-positive synaptic boutons revealed that the majority of postsynaptic targets in the dentate gyrus are dendritic spines, followed by dendritic shafts and granule cell somata (Halasy et al., 2004). These findings are consistent with the results of previous studies in the rat (Fremeau et al., 2001; Kaneko and Fujiyama, 2002; Kaneko et al., 2002; Varoqui et al., 2002). Because only low levels of VGLUT2 mRNA expression have been observed within the hippocampus of rats (Hisano et al., 2000; Fremeau et al., 2001; Herzog et al., 2001), and the granule cell layer of the dentate gyrus shows no signal for DNPI by in situ hybridization (Fremeau et al., 2001), dentate VGLUT2 positive boutons are likely to be of extrahippocampal origin. It has to be noted, however, that VGLUT2 mRNA in mice is concentrated in pyramidal neurons of the hippocampus (Bai et al., 2001), indicating the potential for considerable species differences in the hippocampal expression of VGLUT2. In rats, Fremeau and colleagues suggest a hypothalamic origin for the supragranular VGLUT2-positive fiber cluster because cells in the hypothalamus that project to this layer strongly express DNPI mRNA (Fremeau et al., 2001). In addition to the hypothalamus, VGLUT2 has been found in most septal neurons (Hajszan et al., 2004), and several of these neurons are also positive for GABAergic or cholinergic markers (Danik et al., 2005). However, deafferentation studies (fimbria-fornix transection and entorhinal cortex ablation) led to no significant differences in either the density or distribution pattern of VGLUT2-positive boutons in the dentate gyrus (Halasy et al., 2004), suggesting

77

that the majority of dental VGLUT2 boutons are of intrahippocampal origin. Indeed, light microscopic observation of the hippocampus of colchicine-treated rats revealed a large number of VGLUT2-immunoreactive cell bodies. In the dentate gyrus, the hilar mossy cells represent the most heavily labeled population, while a small proportion of granule cells in the subgranular zone are also VGLUT2-positive (Halasy et al., 2004). It is well known that mossy cells in the dentate gyrus project to the inner molecular layer (Frotscher et al., 1991). Thus, VGLUT2-containing hilar mossy cells may be the source of the VGLUT2 fiber network in the dentate gyrus.

VGLUT3 VGLUT3 is the only known vesicular glutamate transporter that began its ‘‘career’’ without previous history as inorganic phosphate cotransporter. It has been identified in a direct search for additional VGLUT molecules, and cloned (Fremeau et al., 2002; Schafer et al., 2002; Takamori et al., 2002). VGLUT3-immunoreactive terminals surround the granule cell layer of the rat dentate gyrus. VGLUT3 labeling is intense in the molecular layer, corresponding to the proximal part of the granule cell dendrites, and along the hilar border with the granule cell layer, corresponding to the zone where granule cell axons emerge. A dense VGLUT3 network also surrounds the soma of granule cells (Fremeau et al., 2002; Gras et al., 2002; Herzog et al., 2004). A similar distribution pattern has been observed in mice (Schafer et al., 2002). In rats, immunoparticles for VGLUT3 accumulate over vesicle clusters in terminals making classical asymmetrical synapses in the hippocampus (Gras et al., 2002). However, a large number of VGLUT3-positive terminals also form symmetrical synapses (Fremeau et al., 2002; Gras et al., 2002). Further, the labeled terminals make contact with the shaft of proximal dendrites, a location characteristic of inhibitory synapses. However, only a subset of symmetric synapses in the hippocampus appears to stain for VGLUT3 (Fremeau et al., 2002).

Regarding the source of dentate VGLUT3 fibers, one possibility is an intrahippocampal origin. Although moderate levels of VGLUT3 mRNA expression have been observed in the principal cells of pyramidal and dentate granule cell layers in the rat (Fremeau et al., 2002), the distribution of VGLUT3-immunoreactive fibers resembles that observed with markers of GABA terminals. Indeed, the VGLUT3 gene is expressed in scattered interneurons in the hilus of the dentate gyrus (Fremeau et al., 2002; Gras et al., 2002; Herzog et al., 2004). A similar distribution pattern of VGLUT3 mRNA has been demonstrated in mice (Schafer et al., 2002). Immunoreactivity for VGLUT3 is undetectable in pyramidal and dentate granule cells (Fremeau et al., 2002; Somogyi et al., 2004). More importantly, a partial colocalization of VGLUT3 and the GABA marker glutamic acid decarboxylase (GAD) has been observed in the hilar interneurons of the dentate gyrus by means of both in situ hybridization (Herzog et al., 2004) and immunohistochemistry (Fremeau et al., 2002). In addition, colocalization of VGLUT3 with vesicular inhibitory amino acid transporter, a marker of GABAergic neurons and boutons revealed numerous double-labeled nerve endings in the perisomatic terminals that contact the soma of granule cells (Herzog et al., 2004). More specifically, Somogyi and colleagues (Somogyi et al., 2004) have shown that all VGLUT3-positive somata are immunoreactive for cholecystokinin, a marker of a subpopulation of basket cells, and none express markers for other interneuron types in the hippocampus. Boutons expressing VGLUT3, cholecystokinin and GAD are most abundant in the cell layers of the hippocampus (Somogyi et al., 2004). Another source of dentate VGLUT3 fibers may be the subcortical neurons, such as the septohippocampal neurons, the mesolimbic dopaminergic, and raphe serotonergic cells. However, the data obtained so far are controversial. In the rat, the substantia nigra pars compacta and ventral tegmental area contain moderate levels of VGLUT3 mRNA, suggesting expression by dopamine neurons (Fremeau et al., 2002). By contrast, no VGLUT3 mRNA expression has been detected in the substantia nigra by Herzog and colleagues

78

(Herzog et al., 2004), and lack of VGLUT3 expressing neurons has also been observed in the septum and the vertical limb of the diagonal band in the same study. Furthermore, a very high density of VGLUT3 mRNA has been found in the dorsal raphe and MR nuclei (Fremeau et al., 2002; Herzog et al., 2004). All of the serotonin transporter-positive neurons also express VGLUT3 in the dorsal raphe and MR. Interestingly, numerous neurons from the dorsal raphe express VGLUT3 but no serotonin transporter (Gras et al., 2002). However, double-staining for VGLUT3 and the catecholamine marker tyrosine hydroxylase or the plasma membrane serotonin transporter revealed no colocalization in the axons of the dentate gyrus (Fremeau et al., 2002). On the other hand, Somogyi and colleagues (Somogyi et al., 2004) have found a population of VGLUT3 boutons in the hippocampus that are negative for GAD but are labeled for vesicular monoamine transporter type 2 (VMAT2), plasmalemmal serotonin transporter or serotonin, but no colocalization has been found in terminals containing VGLUT3 and vesicular acetylcholine transporter. In mice, the highest VGLUT3 mRNA expression levels and highest density of VGLUT3 mRNA-expressing cells have been observed in the raphe nuclei. In contrast to its mRNA, VGLUT3 immunoreactivity is undetectable in the cell bodies of raphe nuclei in vivo (Schafer et al., 2002). In the hippocampus, VGLUT3 immunoreactivity is absent from cholinergic synapses. Confocal double-immunofluorescence revealed the presence of VGLUT3 in a subpopulation of both thick and thin VMAT2positive varicose fibers in the hippocampus, as well as the absence of VGLUT3 from tyrosine hydroxylase positive terminals. Thus, the hippocampal VGLUT3/VMAT2 system most likely stems from the dorsal raphe and MR nuclei and represents a novel neuronal projection subsystem encoding both glutamatergic and serotonergic neurotransmission in the mouse (Schafer et al., 2002).

Conclusion One common feature of the extrinsic dentate afferent systems is that they originate from a relatively

small number of neurons. However, the majority of these afferents are able to exert a powerful control over the electric activity of the hippocampus. In some of the afferent systems, this efficacy is due to the fact that the majority of the extrinsic afferents terminate on a relatively small, but specific populations of neurons. These target cells include different types of GABAergic interneurons, which, in turn have the ability to control a large number of principal cells (Buzsaki, 1984; Freund and Buzsaki, 1996). For example, The MSDB GABAergic neurons seem to innervate only the GABAergic chandelier and basket cells that have a great influence (disinhibition) on the output sector, soma and axon hillock, of granule cells (Freund and Antal, 1988). Serotonergic fibers originating in the MR, also selectively innervate a specific population of calbindin-containing GABA cells that terminate on the dendritic shafts of principal cells. Thus, in contrast to the disinhibitory function of the MSDBhippocampal GABA system, activation of the MR-hippocampal serotonergic pathway could result in inhibition of the input of principal neurons (Freund et al., 1990). The termination pattern of SUM afferents also seems to be very specific. The overwhelming majority of fibers originating in the SUM form robust asymmetric synaptic contacts with the primary dendrites of granule cells (Leranth and Nitsch, 1994; Magloczky et al., 1994; Nitsch and Leranth, 1996) and, in addition, terminate on specifically positioned GABA interneurons, in a way that they could effectively filter/contrast the signal flow in the trisynaptic circuit (Leranth and Nitsch, 1994). The functional importance of these three major extrahippocampal afferent systems is highlighted by recent observations related to the mnemonic functions of the hippocampus that are associated with synaptoplastic effects of gonadal hormones. Local estrogen administration into the MSDB (Laˆm and Leranth, 2003), MR (Prange-Kiel et al., 2004), and SUM (Leranth and Shanabrough, 2001) of ovariectomized rats all resulted in a robust increase in the density of spine synapses in the hippocampus. In contrast, transection of the fimbria fornix, which contains the hippocampal projection of these subcortical structures, prevents (Leranth et al., 2000) the known effects (Gould et al., 1990).

79

Acknowledgments This work was supported by NIH grants MH060858 and NS042644 to C.L. and MH074021 to T.H., as well as by a Hungarian National Office for Research and Technology grant RET-08/04. References Acsady, L., Halasy, K. and Freund, T.F. (1993) Calretinin is present in non-pyramidal cells of the rat hippocampus–III. Their inputs from the median raphe and medial septal nuclei. Neuroscience, 52: 829–841. Aihara, Y., Mashima, H., Onda, H., Hisano, S., Kasuya, H., Hori, T., Yamada, S., Tomura, H., Yamada, Y., Inoue, I., Kojima, I. and Takeda, J. (2000) Molecular cloning of a novel brain-type Na(+)-dependent inorganic phosphate cotransporter. J. Neurochem., 74(6): 2622–2625. Alreja, M. (1996) Excitatory actions of serotonin on GABAergic neurons of the medial septum diagonal band of Broca. Synapse, 22: 15–27. Alvarez-Leefmans, F.J. and Gardner-Medwin, A.R. (1975) Influences of septum on the hippocampal dental area which are unaccompanied by field potentials. J. Physiol. (Lond.), 249: 14–16. Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182: 851–914. Amaral, D.G. and Cowan, W.M. (1980) Subcortical afferents to the hippocampal formation in the monkey. J. Comp. Neurol., 189: 573–591. Amaral, D.G. and Witter, M.P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31: 571–591. Amaral, D.G. and Witter, M.P. (1995) Hippocampal formation. In: Paxinos G. (Ed.), The rat nervous system (2nd ed.). Academic Press, New York, pp. 443–494. Anderson, P. and Eccles, J.C. (1962) Inhibitory phasing of neuronal discharge. Nature (London), 196: 645–647. Apartis, E., Poindessous-Jazat, F.R., Lamour, Y.A. and Bassant, M.H. (1998) Loss of rhythmically bursting neurons in rat medial septum following selective lesion of septohippocampal cholinergic system. J. Neurophysiol., 79: 1633–1642. Bai, L., Xu, H., Collins, J.F. and Ghishan, F.K. (2001) Molecular and functional analysis of a novel neuronal vesicular glutamate transporter. J. Biol. Chem., 276(39): 36764–36769. Bellocchio, E.E., Hu, H., Pohorille, A., Chan, J., Pickel, V.M. and Edwards, R.H. (1998) The localization of the brain-specific inorganic phosphate transporter suggests a specific presynaptic role in glutamatergic transmission. J. Neurosci., 18(21): 8648–8659. Bellocchio, E.E., Reimer, R.J., Fremeau, R.T.J. and Edwards, R.H. (2000) Uptake of glutamate into synaptic vesicles by an

inorganic phosphate transporter. Science, 289(5481): 957–960. Benzi, G. and Moretti, A. (1998) Is there a rationale for the use of acetylcholinesterase inhibitors in the therapy of Alzheimer’s disease? Eur. J. Pharmacol., 346: 1–13. Berger, T.W., Semple-Rowland, S. and Basset, J. (1980) Hippocampal polymorph neurons are the cells of origin for ipsilateral association and commissural afferents to the dentate gyrus. Brain Res., 215: 329–336. Bilkey, D.K. and Goddard, G.V. (1985) Medial septal facilitation of hippocampal granule cell activity is mediated by inhibition of inhibitory interneurons. Brain Res., 361: 99–106. Blackstad, T.W. (1956) Commissural connections of the hippocampal region of the rat, with special reference to their mode of termination. J. Comp. Neurol., 105: 417–537. Bland, B.H. and Oddie, S.D. (1998) Anatomical, electrophysiological and pharmacological studies of ascending brainstem hippocampal synchronizing pathways. Neurosci. Biobehav. Rev., 22: 259–273. Brito, G.N. and Brito, L.S. (1990) Septohippocampal system and the prelimbic sector of frontal cortex: a neuropsychological battery analysis in the rat. Behav. Brain Res., 36: 127–146. Brown, R.A., Walling, S.G., Milway, J.S. and Harley, C.W. (2005) Locus ceruleus activation suppresses feedforward interneurons and reduces beta-gamma electroencephalogram frequencies while it enhances theta frequencies in rat dentate gyrus. J. Neurosci., 25: 1985–1991. Buckmaster, P.S., Wentzel, H.J., Kunkel, D.D. and Schwartzkroin, P.A. (1996) Axon arbors and synaptic connections of hippocampal mossy cells in the rat in vivo. J. Comp. Neurol., 366: 270–292. Buzsaki, G. (1984) Feed-forward inhibition in the hippocampal formation. Progr. Neurobiol., 22: 131–153. Buzsaki, G. and Eidelburg, E. (1981) Commissural projection to the dentate gyrus of the rat: evidence for feed-forward inhibition. Brain Res., 230: 346–350. Buzsaki, G. and Eidelburg, E. (1982) Direct afferent excitation and long-term potentiation of hippocampal interneurons. J. Neurophysiol., 48: 597–607. Carre, G.P. and Harley, C.W. (1991) Population spike facilitation in the dentate gyrus following glutamate administration to the lateral supramammillary nucleus. Brain Res., 568: 307–310. Chandler, J.P. and Crutcher, K.A. (1983) The septohippocampal projection in the rat: an electron microscopic horseradish peroxidase study. Neuroscience, 10: 685–696. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus. J. Comp. Neurol., 246: 435–458. Clarke, D.J. (1985) Cholinergic innervation of the rat dentate gyrus: an immunocytochemical and electron microscopical study. Brain Res., 360: 349–354. Dahl, D. and Winson, J. (1986) Influence of neurons of the parafascicular region on neuronal transmission from perforant path through dentate gyrus. Brain Res., 377: 211–219.

80 Danik, M., Cassoly, E., Manseau, F., Sotty, F., Mouginot, D. and Williams, S. (2005) Frequent coexpression of the vesicular glutamate transporter 1 and 2 genes, as well as coexpression with genes for choline acetyltransferase or glutamic acid decarboxylase in neurons of rat brain. J. Neurosci. Res., 81(4): 506–521. Deller, T. (1998) The anatomical organization of the rat fascia dentata: new aspects of laminar organization as revealed by anterograde tracing with phaseolus vulgaris-leucoagglutinin. Anat. Embryol., 197: 89–103. Deller, T., Katona, I., Cozzari, C., Frotscher, M. and Freund, T.F. (1999) Cholinergic innervation of mossy cells in the rat fascia dentata. Hippocampus, 9: 314–320. Deller, T. and Leranth, C. (1990) Synaptic connections of NPYimmunoreactive neurons in the rat hilar area. J. Comp. Neurol., 300: 433–447. Deller, T., Martinez, A., Nitsch, R. and Frotscher, M. (1996a) A novel entorhinal projection to the rat dentate gyrus: direct innervation of proximal dendrites and granule cell bodies of granule cells and GABAergic neurons. J. Neurosci., 16: 3322–3333. Deller, T., Nitsch, R. and Frotscher, M. (1994) Associational and commissural afferents of parvalbumin-immunoreactive neurons in the rat hippocampus: a combined immunocytochemical and PHAL study. J. Comp. Neurol., 350: 612–622. Deller, T., Nitsch, R. and Frotscher, M. (1995) Phaseolus vulgaris- Leucoagglutinin (PHAL) tracing of commissural fibers to the rat fascia dentata: evidence for a previously unknown commissural projection to the outer molecular layer. J. Comp. Neurol., 352: 55–68. Deller, T., Nitsch, R. and Frotscher, M. (1996b) Heterogeneity of the commissural projection to the rat dentate gyrus: a Phaseolus vulgaris- Leucoagglutinin tracing study. Neuroscience, 75: 111–121. Dickinson-Anson, H., Aubert, I., Gage, F.H. and Fisher, L.J. (1998) Hippocampal grafts of acetylcholine-producing cells are sufficient to improve behavioral performance following a unilateral fimbria-fornix lesion. Neuroscience, 84: 771–781. Dunnett, S.B., Low, W.C., Iversen, S.D., Stenevi, U. and Bjorklund, A. (1982) Septal transplants restore maze learning in rats with fornix-fimbria lesions. Brain Res., 251: 335–348. Dutar, P., Lamour, Y. and Jobert, A. (1983) Acetylcholine excites identified septo-hippocampal neurons in the rat. Neurosci. Lett., 43: 43–47. Fantie, B.D. and Goddard, G.V. (1982) Septal modulation of the population spike in the fascia dentata produced by perforant path stimulation in the rat. Brain Res., 252: 227–237. Fonnum, F. (1970) Topographical and subcellular localization of choline acetyltransferase in rat hippocampal formation. J. Neurochem., 17: 1029–1037. Fremeau, R.T., Burman, J., Qureshi, T., Tran, C.H., Proctor, J., Johnson, J., Zhang, H., Sulzer, D., Copenhagen, D.R., Storm-Mathisen, J., Reimer, R.J., Chaudhry, F.A. and Edwards, R.H. (2002) The identification of vesicular glutamate transporter 3 suggests novel modes of signaling by glutamate. Proc. Natl. Acad. Sci. U.S.A., 99: 14488–14493.

Fremeau, R.T., Troyer, M.D., Pahner, I., Nygaard, G.O., Tran, C.H., Reimer, R.J., Bellocchio, E.E., Fortin, D., Storm-Mathisen, J. and Edwards, R.H. (2001) The expression of vesicular glutamate transporters defines two classes of excitatory synapse. Neuron, 31: 247–260. Freund, T.F. (1989) GABAergic septohippocampal neurons contain parvalbumin. Brain Res., 478: 375–381. Freund, T.F. and Antal, M. (1988) GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature, 336: 170–173. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Freund, T.F., Gulyas, A.I., Acsady, L., Gorcs, T. and Toth, K. (1990) Serotonergic control of the hippocampus via local inhibitory interneurons. Proc. Natl. Acad. Sci. U.S.A., 87: 8501–8505. Freund, T.F., Hajos, N., Acsady, L., Gorcs, T.J. and Katona, I. (1997) Mossy cells of the rat dentate gyrus are immunoreactive for calcitonin gene-related peptide (CGRP). Eur. J. Neurosci., 9: 1815–1830. Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of dentate gyrus granule cells. J. Neurophysiol., 51: 195–209. Frotscher, M. and Leranth, C. (1985) Cholinergic innervation of the rat hippocampus as revealed by choline acetyltransferase immunocytochemistry: A combined light and electron microscopic study. J. Comp. Neurol., 239: 237–246. Frotscher, M. and Leranth, C. (1986) The cholinergic innervation of the rat fascia dentata: identification of target structures on granule cells by combining choline acetyltransferase immunocytochemistry and Golgi impregnation. J. Comp. Neurol., 243: 58–70. Frotscher, M. and Leranth, C. (1988) Catecholaminergic innervation of pyramidal and GABAergic nonpyramidal neurons in the rat hippocampus. Double label immunostaining with antibodies against tyrosine hydroxylase and glutamate decarboxylase. Histochemistry, 88: 313–319. Frotscher, M., Leranth, C., Lubbers, K. and Oertel, W.H. (1984) Commissural afferents innervate glutamate decarboxylase immunoreactive nonpyramidal neurons in the guinea pig hippocampus. Neurosci. Lett., 46: 137–143. Frotscher, M., Schlander, M. and Leranth, C. (1986) Cholinergic neurons in the hippocampus. A combined light- and electron-microscopic immunocytochemical study in the rat. Cell Tissue Res., 246: 293–301. Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E. (1991) The mossy cells of the fascia dentata: a comparative study of their fine structure and synaptic connections in rodents and primates. J. Comp. Neurol., 312: 145–163. Frotscher, M. and Zimmer, J. (1983a) Commissural fibers terminate on nonpyramidal neurons in the guinea pig hippocampus-a combined Golgi/EM degeneration study. Brain Res., 265: 289–293. Frotscher, M. and Zimmer, J. (1983b) Lesion-induced mossy fibers to the inner molecular layer of the rat fascia dentata: identification of postsynaptic granule cells by the Golgi-EM technique. J. Comp. Neurol., 336: 170–173.

81 Gaykema, R.P., Luiten, P.G., Nyakas, C. and Traber, J. (1990) Cortical projection patterns of the medial septum-diagonal band complex. J. Comp. Neurol., 293: 103–124. Givens, B.S. and Olton, D.S. (1990) Cholinergic and GABAergic modulation of medial septal area: effect on working memory. Behav. Neurosci., 104: 849–855. Givens, B.S. and Olton, D.S. (1994) Local modulation of basal forebrain: effects on working and reference memory. J. Neurosci., 14: 3578–3587. Givens, B.S. and Olton, D.S. (1995) Bidirectional modulation of scopolamine-induced working memory impairments by muscarinic activation of the medial septal area. Neurobiol. Learn. Mem., 63: 269–276. Givens, B. and Sarter, M. (1997) Modulation of cognitive processes by transsynaptic activation of the basal forebrain. Behav. Brain Res., 84: 1–22. Goodman, J.H. and Sloviter, R.S. (1992) Evidence for commissurally projecting parvalbumin-immunoreactive basket cells in the dentate gyrus of the rat. Hippocampus, 2: 13–22. Gottlieb, D.I. and Cowan, W.M. (1973) Autoradiographic studies of the commissural and ipsilateral association connections of the hippocampus and dentate gyrus of the rat. I. The commissural connections. J. Comp. Neurol., 149: 393–422. Gould, E., Woolley, C.S., Frankfurt, M. and McEwen, B.S. (1990) Gonadal steroids regulate dendritic spine density in hippocampal pyramidal cells in adulthood. J. Neurosci., 10: 1286–1291. Gras, C., Herzog, E., Bellenchi, G.C., Bernard, V., Ravassard, P., Pohl, M., Gasnier, B., Giros, B. and El Mestikawy, S. (2002) A third vesicular glutamate transporter expressed by cholinergic and serotoninergic neurons. J. Neurosci., 22: 5442–5451. Gulyas, A.I., Gorcs, T.J. and Freund, T.F. (1990) Innervation of different peptide-containing neurons in the hippocampus by GABAergic septal afferents. Neuroscience, 37: 31–44. Gulyas, A.I., Seress, L., Toth, K., Acsady, L., Antal, M. and Freund, T.F. (1991) Septal GABAergic neurons innervate inhibitory interneurons in the hippocampus of the macaque monkey. Neuroscience, 41: 381–390. Haglund, L., Swanson, L.W. and Kohler, C. (1984) The projection of the supramammillary nucleus to the hippocampal formation: an immunohistochemical and anterograde transport study with the lectin PHA-L in the rat. J. Comp. Neurol., 229: 171–185. Hajszan, T., Alreja, M. and Leranth, C. (2004) Intrinsic vesicular glutamate transporter 2-immunoreactive input to septohippocampal parvalbumin-containing neurons: novel glutamatergic local circuit cells. Hippocampus, 14: 499–509. Halasy, K., Hajszan, T., Kovacs, E.G., Lam, T.T. and Leranth, C. (2004) Distribution and origin of vesicular glutamate transporter 2-immunoreactive fibers in the rat hippocampus. Hippocampus, 14: 908–918. Haglund, L., Swanson, L.W. and Kohler, C. (1984) The projection of the supramammillary nucleus to the hippocampal formation: an immunohistochemical and anterograde

transport study with the lectin PHA-L in the rat. J. Comp. Neurol., 229: 171–185. Harley, C.W., Lacaille, J.-C. and Galway, M. (1983) Hypothalamic afferents to the dorsal dentate gyrus contain acetylcholinesterase. Brain Res., 270: 335–339. Herzog, E., Bellenchi, G.C., Gras, C., Bernard, V., Ravassard, P., Bedet, C., Gasnier, B., Giros, B. and El Mestikawy, S. (2001) The existence of a second vesicular glutamate transporter specifies subpopulations of glutamatergic neurons. J. Neurosci., 21: 1–6. Herzog, E., Gilchrist, J., Gras, C., Muzerelle, A., Ravassard, P., Giros, B., Gaspar, P. and El Mestikawy, S. (2004) Localization of VGLUT3, the vesicular glutamate transporter type 3, in the rat brain. Neuroscience, 123: 983–1002. Hisano, S., Hoshi, K., Ikeda, Y., Maruyama, D., Kanemoto, M., Ichijo, H., Kojima, I., Takeda, J. and Nogami, H. (2000) Regional expression of a gene encoding a neuron-specific Na(+)-dependent inorganic phosphate cotransporter (DNPI) in the rat forebrain. Brain Res. Mol. Brain Res., 83: 34–43. Hjorth-Simonsen, A. and Laurberg, S. (1977) Commissural connections of the fascia dentata in the rat. J. Comp. Neurol., 174: 591–606. Hounsgaard, J. (1978) Presynaptic inhibitory action of acetylcholine in area CA1 of the hippocampus. Exp. Neurol., 62: 787–797. Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization of intrahippocampal projections originating from CA3 pyramidal cells in the rat. J. Comp. Neurol., 295: 580–623. Kandel, E.R., Spencer, W.A. and Brinley, F.J. (1961) Electrophysiology of hippocampal neurons. I. Sequential invasion and synaptic organization. J. Neurophysiol., 24: 225–242. Kaneko, T. and Fujiyama, F. (2002) Complementary distribution of vesicular glutamate transporters in the central nervous system. Neurosci. Res., 42: 243–250. Kaneko, T., Fujiyama, F. and Hioki, H. (2002) Immunohistochemical localization of candidates for vesicular glutamate transporters in the rat brain. J. Comp. Neurol., 444(1): 39–62. Kiss, J., Csaki, A., Bokor, H., Shanabrough, M. and Leranth, C. (2000) The supramammillo-hippocampal and supramammillo-septal glutamatergic/aspartatergic projection in the rat: a combined [3H]D-aspartate autoradiographic and immunohistochemical study. Neuroscience, 97: 657–669. Kiss, J., Patel, A.J. and Freund, T.F. (1990) Distribution of septohippocampal neurons containing parvalbumin or choline acetyltransferase in the rat brain. J. Comp. Neurol., 298: 362–372. Kosofsky, B.E. and Molliver, M.E. (1987) The serotonergic innervation of cerebral cortex: different classes of axon terminals arise from dorsal and medial raphe nuclei. Synapse, 1: 153–168. Laˆm, T.T. and Leranth, C. (2003) Role of the medial septum diagonal band of Broca cholinergic neurons in estrogeninduced spine synapse formation on hippocampal CA1 pyramidal cells of female rats. Eur. J. Neurosci., 17: 1997–2005.

82 Lamour, Y., Dutar, P. and Jobert, A. (1984) Septo-hippocampal and other medial septum diagonal band neurons: electrophysiological and pharmacological properties. Brain Res., 309: 227–239. Laurberg, S. (1979) Commissural and intrinsic connections of the rat hippocampus. J. Comp. Neurol., 184: 685–708. Laurberg, S. and Sorensen, K.E. (1981) Associational and commissural collaterals of neurons in the hippocampal formation (hilus fasciae dentatae and subfield CA3). Brain Res., 212: 287–300. Lawson, V.H. and Bland, B.H. (1993) The role of the septohippocampal pathway in the regulation of hippocampal field activity and behavior: analysis by the intraseptal microinfusion of carbachol, atropine, and procaine. Exp. Neurol., 120: 132–144. Leranth, C. and Frotscher, M. (1983) Commissural afferents to the rat hippocampus terminate on vasoactive intestinal polypeptide-like immunoreactive non-pyramidal neurons: An EM immunocytochemical degeneration study. Brain Res., 276: 357–361. Leranth, C. and Frotscher, M. (1987) Cholinergic innervation of hippocampal GAD- and somatostatin-immunoreactive commissural neurons. J. Comp. Neurol., 261: 33–47. Leranth, C. and Frotscher, M. (1989) The organization of the septal region in the rat brain: cholinergic-GABAergic interconnections and the termination of hippocampo-septal fibers. J. Comp. Neurol., 289: 304–314. Leranth, C. and Nitsch, R. (1994) Morphological evidence that hypothalamic substance P-containing afferents are capable of filtering the signal flow in the monkey hippocampal formation. J. Neurosci., 14: 4079–4094. Leranth, C. and Shanabrough, M. (2001) Supramammillary area mediates subcortical estrogenic action on hippocampal synaptic plasticity. Exp. Neurol., 167: 445–450. Leranth, C., Shanabrough, M. and Horvath, T.L. (2000) Hormonal regulation of hippocampal spine synapse density involves subcortical mediation. Neuroscience, 101: 349–356. Leranth, C. and Vertes, R.P. (1999) Median raphe serotonergic innervation of medial septum/diagonal band of Broca (MSDB) parvalbumin-containing neurons: possible involvement of the MSDB in the desynchronization of hippocampal EEG. J. Comp. Neurol., 410: 586–598. Lewis, P.R. and Shute, C.C.D. (1967) The cholinergic limbic system: projections to hippocampal formation, medial cortex nuclei of the ascending cholinergic reticular system and the subfornical organ and supra-optic crest. Brain, 90: 521–537. Li, X.-G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The hippocampal CA3 network: an in vivo intracellular labeling study. J. Comp. Neurol., 339: 181–208. Lopes de Silva, F.H., Witter, M.P., Boeijinga, P.H. and Lohman, A.H.M. (1990) Anatomic organization and physiology of the limbic cortex. Physiol. Rev., 70: 453–511. Lorente de No, R. (1934) Studies on the structure of the cerebral cortex II. Continuation of the study of the ammonic system. J. Psychol. Neurol., 46: 113–177. Lynch, G., Rose, G. and Gall, C. (1978) Anatomical and functional aspects of the septo-hippocampal projections.

In: Elliot K. and Whelan J. (Eds.), Functions of the Hippocampal System. Elsevier, Amsterdam, pp. 5–24. Madison, D.V. and Nicoll, R.A. (1988) Norepinephrine decreases synaptic inhibition in the rat hippocampus. Brain Res., 442: 131–138. Magloczky, Z., Acsady, L. and Freund, T.F. (1994) Principal cells are the postsynaptic targets of supramammillary afferents in the hippocampus of the rat. Hippocampus, 4: 322–334. Manahan-Voughan, D. and Kulla, A. (2003) Regulation of depotentiation and long-term potentiation in the dentate gyrus of freely moving rats by dopamine D2-like receptors. Cereb. Cortex, 13: 123–135. Markowska, A.L., Olton, D.S. and Givens, B. (1995) Cholinergic manipulations in the medial septal area: age-related effects on working memory and hippocampal electrophysiology. J. Neurosci., 15: 2063–2073. McKenna, J.T. and Vertes, R.P. (2001) Collateral projections from the median raphe nucleus to the medial septum and hippocampus. Brain Res. Bull., 54: 619–630. Mellgren, S.I. and Srebro, B. (1973) Changes in acetylcholinesterase and distribution of degenerating fibres in the hippocampal region after septal lesion in the rat. Brain Res., 52: 19–35. Miettinen, R. and Freund, T.F. (1992a) Convergence and segregation of septal and median raphe inputs onto different subsets of hippocampal inhibitory interneurons. Brain Res., 594: 263–272. Miettinen, R. and Freund, T.F. (1992b) Neuropeptide-Y-containing interneurons in the hippocampus receive synaptic input from median raphe and GABAergic septal afferents. Neuropeptides, 22: 185–193. Milner, T.A. and Bacon, C.E. (1989) GABAergic neurons in the rat hippocampal formation: ultrastructure and synaptic relationships with catecholaminergic terminals. J. Neurosci., 10: 3410–3427. Mizumori, S.J.Y., McNaughton, B.L. and Barnes, C.A. (1989) A comparison of supramammillary and medial septal influences on hippocampal field potentials and single-unit activity. J. Neurophysiol., 61: 15–31. Monmaur, P. and Breton, P. (1991) Elicitation of hippocampal theta by intraseptal carbachol injection in freely moving rats. Brain Res., 544: 150–155. Mosko, S., Lynch, G. and Cotman, C.W. (1973) The distribution of septal projections to the hippocampus in the rat. J. Comp. Neurol., 52: 163–174. Moudy, A.M., Kunkel, D.D. and Schwartzkroin, P.A. (1993) Development of dopamine-beta-hydroxylase-positive fber innervation of the rat hippocampus. Synapse, 15: 307–318. Ni, B., Rosteck, P.R.J., Nadi, N.S. and Paul, S.M. (1994) Cloning and expression, of a cDNA encoding a brain-specific Na(+)-dependent inorganic phosphate cotransporter. Proc. Natl. Acad. Sci. U.S.A., 91(12): 5607–5611. Ni, B., Wu, X., Yan, G.M., Wang, J. and Paul, S.M. (1995) Regional expression and cellular localization of the Na(+)dependent inorganic phosphate cotransporter of rat brain. J. Neurosci., 15(8): 5789–5799.

83 Nicoll, R.A., Schenker, C. and Leeman, S.E. (1980) Substance P as a transmitter candidate. Ann. Rev. Neurosci., 2: 227–268. Nitsch, R. and Leranth, C. (1993) Calretinin immunoreactivity in the monkey hippocampal formation. II: Intrinsic GABAergic and hypothalamic non-GABAergic systems. An experimental tracing and coexistence study. Neuroscience, 55: 797–812. Nitsch, R. and Leranth, C. (1994) Substance P-containing hypothalamic afferents to the monkey hippocampus. An immunocytochemical, tract tracing, and coexistence study. Exp. Brain Res., 101: 231–240. Nitsch, R. and Leranth, C. (1996) GABAergic neurons in the rat dentate gyrus are innervated by subcortical calretinincontaining afferents. J. Comp. Neurol., 364: 425–438. Nyakas, C., Luiten, P.G.M., Spencer, D.G. and Traber, J. (1987) Detailed projection patterns of septa1 and diagonal band efferents to the hippocampus in the rat with emphasis on innervation of CAI and dentate gyrus. Brain Res. Bull., 18: 533–545. Prange-Kiel, J., Rune, G.M. and Leranth, C. (2004) Median raphe mediates estrogenic effects to the hippocampus in female rats. Eur. J. Neurosci., 19: 309–317. Ramon y Cajal, S. (1893) Estructura del asfa de Ammon y fascia dentata. Ann. SOC Esp. Hist. Nat., 22. Ramon y Cajal, S. (1911) Histologie de systeme nerveux de I’Homme et des vertebres tomme II. Maloine, Paris. Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy fibers in rats and gerbils form synapses with the somata and proximal dendrites of basket cells in the dentate gyrus. Hippocampus, 1: 355–364. Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The development, ultrastructure and synaptic connections of the mossy cells of the dentate gyrus. J. Neurocytol., 14: 835–857. Ribak, C.E., Seress, L., Peterson, G.M., Serooggy, K.B., Fallon, J.H. and Schmued, L.C. (1986) A GABAergic inhibitory component within the hippocampal commissural pathway. J. Neurosci., 6: 3492–3498. Robinson, G.B. and Racine, R.J. (1984) Interactions between septodentate and perforant-path inputs to the dentate gyrus. Soc. Neurosci. Abstr., 10: 79. Roppert, N. and Guy, N. (1991) Serotonin facilitates GABAergic transmission in the CA1 region of the rat hippocampus in vitro. J. Physiol. (London), 441: 121–136. Rose, G. and Schubert, P. (1977) Release and transfer of [3H]adenosine derivatives in the cholinergic septa1 system. Brain Res., 121: 353–357. Schafer, M.K., Varoqui, H., Defamie, N., Weihe, E. and Erickson, J.D. (2002) Molecular cloning and functional identification of mouse vesicular glutamate transporter 3 and its expression in subsets of novel excitatory neurons. J. Biol. Chem., 277(52): 50734–50748. Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the molecular layer have lower thresholds for synaptic activation by perforant path than granule cells. J. Neurosci., 11: 1660–1673.

Scharfman, H.E. (1992) Differentiation of rat dentate neurons by morphology and electrophysiology in hippocampal slices: granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells. In: Ribak, C.E., Gall, C.M. and Mody, I. (Eds.), The Dentate Gyrus and its Role in Seizures: Epilepsy Research Supplements. Elsevier, Amsterdam, Vol. 7, pp. 93–109. Scharfman, H.E. (1995) Electrophysiological evidence that dentate hilar mossy cells are excitatory and innervate both granule cells and interneurons. J. Neurophysiol., 74: 179–194. Scharfman, H.E. (1999) The role of nonprincipal cells in dentate gyrus excitability and its relevance to animal models of epilepsy and temporal lobe epilepsy. Adv. Neurol., 79: 805–820. Scharfman, H.E. and Schwartzkroin, P.A. (1988) Electrophysiology of morphologically identified mossy cells of the dentate hilus recorded in guinea pig hippocampal slices. J. Neurosci., 8: 3812–3821. Segal, M. (1979) A potent inhibitory monosynaptic hypothalamo-hippocampal connection. Brain Res., 162: 137–141. Segal, M. and Landis, S. (1974) Afferents to the hippocampus of the rat studied with the method of retrograde transport of horseradish peroxidase. Brain Res., 78: 1–15. Seress, L. (1988) Interspecies comparison of the hippocampal formation shows increased emphasis on the regio superior in the Ammon’s horn of the human brain. J. Hirnforsch., 29: 335–340. Seress, L. and Ribak, C.E. (1983) GABAergic cells in the dentate gyrus appear to be local circuit and projection neurons. Exp. Brain Res., 50: 173–182. Seress, L. and Ribak, C.E. (1984) Direct commissural connections to basket cells of the hippocampal dentate gyrus: anatomical evidence for feed-forward inhibition. J. Neurocytol., 13: 215–225. Shute, C.C.D. and Lewis, P.R. (1966) Electron microscopy of cholinergic terminals and acetylcholinesterase-containing neurons in the hippocampal formation of the rat. Z. Zellforsch. Mikrosk. Anat., 69: 334–343. Sloviter, R.S. and Nilaver, G. (1987) Immunocytochemical localization of GABA-, cholecystokinin, vasoactive intestinal polypeptide, and somatostatin-like immunoreactivity in the area dentata and hippocampus of the rat. J. Comp. Neurol., 256: 42–60. Soltesz, I. and Mody, I. (1994) Patch-clamp recordings reveal powerful GABAergic inhibition in dentate hilar neurons. J. Neurosci., 14: 2365–2376. Somogyi, J., Baude, A., Omori, Y., Shimizu, H., El Mestikawy, S., Fukaya, M., Shigemoto, R., Watanabe, M. and Somogyi, P. (2004) GABAergic basket cells expressing cholecystokinin contain vesicular glutamate transporter type 3 (VGLUT3) in their synaptic terminals in hippocampus and isocortex of the rat. Eur. J. Neurosci., 19: 552–569. Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat fascia dentata are glutamate-immunoreactive. Hippocampus, 4: 65–70. Storm-Mathisen, J. and Blackstad, T.W. (1964) Cholinesterase in the hippocampal region. Distribution and relation to architectonics and afferent systems. Acta Anat., 56: 216–253.

84 Swanson, L.W. (1982) The projection of the ventral tegmental area and adjacent regions: a combined fluorescent and retrograde tracer and immunofluorescence study in the rat. Brain Res. Bull., 9: 321–353. Takamori, S., Malherbe, P., Broger, C. and Jahn, R. (2002) Molecular cloning and functional characterization of human vesicular glutamate transporter 3. EMBO Rep., 3: 798–803. Takamori, S., Rhee, J.S., Rosenmund, C. and Jahn, R. (2000) Identification of a vesicular glutamate transporter that defines a glutamatergic phenotype in neurons. Nature, 407: 189–194. Tork, I. (1990) Anatomy of the serotonergic system. Ann. N.Y. Acad. Sci. U.S.A., 600: 9–34. Varoqui, H., Schafer, M.K., Zhu, H., Weihe, E. and Erickson, J.D. (2002) Identification of the differentiation-associated Na+/PI transporter as a novel vesicular glutamate transporter expressed in a distinct set of glutamatergic synapses. J. Neurosci., 22: 142–155. Veazey, R.B., Amaral, D.G. and Cowan, W.M. (1982) The morphology and connections of the posterior hypothalamus in the cynomolgus monkey (Macaca fascicularis). II. Efferent connections. J. Comp. Neurol., 207: 135–156. Voneida, T.J., Vardaris, R.M., Fish, S.E. and Reiheld, C.T. (1981) The origin of the hippocampal commissure in the rat. Anat. Rec., 201: 91–103.

Wainer, B.H., Bolam, J.P., Freund, T.F., Henderson, Z., Totterdell, S. and Smith, A.D. (1984) Cholinergic synapses in the rat brain: a correlated light and electron microscopic immunohistochemical study employing a monoclonal antibody against choline acetyltransferase. Brain Res., 308: 69–76. Wheal, H.V. and Miller, J.J. (1980) Pharmacological identification of acetylcholine and glutamate excitatory systems in the dentate gyrus of the rat. Brain Res., 182: 145–155. Whitehouse, P.J., Price, D.L., Struble, R.G., Clark, A.W., Coyle, J.T. and Delon, M.R. (1982) Alzheimer’s disease and senile dementia: loss of neurons in the basal forebrain. Science, 215: 1237–1239. Wiss, J.M., Swanson, L.W. and Cowan, W.M. (1979) Evidence for an input to the molecular layer and the stratum granulosum of the dentate gyrus from the supramammillary region of the hypothalamus. Anat. Embryol., 156: 165–176. Wu, M., Hajszan, T., Xu, C., Leranth, C. and Alreja, M. (2004) Group I metabotropic glutamate receptor activation produces a direct excitation of identified septohippocampal cholinergic neurons. J. Neurophysiol., 92: 1216–1225. Zimmer, J., Laurberg, S. and Sunde, N. (1983) Neuroanatomical aspects of normal and transplanted hippocampal tissue. In: Seifert W. (Ed.), Neurobiology of the Hippocampus. Academic Press, New York, pp. 39–64.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 5

The dentate mossy fibers: structural organization, development and plasticity Morten Blaabjerg and Jens Zimmer Anatomy and Neurobiology, Institute of Medical Biology, University of Southern Denmark, Winslowparken 21, DK-5000 Odense C, Denmark

Abstract: Hippocampal mossy fibers are the axons of the dentate granule cells and project to hippocampal CA3 pyramidal cells and mossy cells of the dentate hilus (CA4) as well as a number of interneurons in the two areas. Besides their role in hippocampal function, studies of which are still evolving and taking interesting turns, the mossy fibers display a number of unique features with regard to axonal projections, terminal structures and synaptic contacts, development and variations among species and strains, as well as to normal occurring and lesion-induced plasticity and neural transplantation. These features are the topic of this review, which will use the mossy fiber system of the rat as basis and reference in its aim to provide an up-to-date, yet historically based guide to students in the field. Keywords: granule cells; hippocampus; CA3; CA1; dentate hilus; lesion-induced sprouting; neural transplantation collaterals of the MFs in the dentate hilus also contact several types of neurons there, deep in the hilus as well as immediately below the granule cell layer. The fine, filopodial processes, originally identified only to extend from the classical, large MF terminals, have recently been shown to make separate and specific synaptic contacts with interneurons, both within the dentate hilus and in CA3, thereby adding new functional dimensions to the dentate MF system and its role in hippocampal function (Frotscher et al., 1994, 2006; Acsa´dy et al., 1998; Henze et al., 2000). The dentate MFs were first described by Golgi (1886) and Sala (1891), and were named mossy fibers by Ramo´n y Cajal (1893, 1911), based on the similarity of their large, irregular terminals with the moss-like terminal varicosities of the cerebellar MFs, which he had described earlier. Additional contributions, based on Golgi-impregnated

Introduction to the dentate mossy fiber system The dentate mossy fibers (MFs) are generally known to be the second link in the classical trisynaptic pathway from the entorhinal cortex to the hippocampal subfield CA1, including: (1) the entorhinal perforant path projection to dentate granule cells; (2) the dentate MF projection to CA3 pyramidal cells; and (3) the CA3 Schaffer collateral projection to CA1 (Andersen et al., 1969). In line with this, MFs arise from dentate granule cells, located in the tightly packed cell layer of the fascia dentata or dentate gyrus, and project their main axons through the dentate hilus (CA4) and to the proximal parts of the apical (and basal) dendrites of CA3 pyramidal cells. The extensive network of Corresponding author. Tel.: (+45) 6550 3801; Fax: (+45) 6550 6321; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63005-2

85

86

material, were made by Schaffer (1882) and Koelliker (1896) and later Lorente de No´ (1934), a student of Ramo´n y Cajal, who added further details about the synaptic nature of contacts between the giant MF terminals and the complex spines emerging from the proximal dendrites of large pyramidal cells of regio inferior (CA3) and the so-called mossy cells of the dentate hilus (CA4). The dentate MFs display various differences in number, trajectory and termination among species and strains and individuals within a given species. Here we present a guide to the structural organization of the MF system. For the sake of simplicity, we will use the common Wistar and Sprague-Dawley laboratory rats as the standard to which other species are compared.

1991; Frotscher et al., 1994; Acsa´dy et al., 1998). One rather unique and widely used method for ‘‘bulk visualization’’ of MF terminals, and thereby the MF terminal projection, is the histochemical Timm sulfide silver method (Timm, 1958; Haug, 1967; Danscher, 1981), which takes advantage of a very high concentration of chelatable zinc in the MF terminals (see Danscher et al., 1985). After precipitation by sulfide, the chelatable zinc can be visualized by physical development and the stained elements analyzed by both light and electron microscopy (Danscher and Zimmer, 1978; Laurberg and Zimmer, 1981). After Timm staining, the MF projection and its individual terminals stand out distinctly black, in contrast to the other unstained to lightly-stained (yellowish to brown) laminarspecific terminal fields in the hippocampus and fascia dentata (Figs. 2–4).

Visualization of dentate mossy fibers and terminals The axonal projection and termination of dentate MF have been visualized by several techniques, including classical Golgi-impregnation (Schaffer, 1882; Golgi, 1886; Sala, 1891; Ramo´n y Cajal, 1893, 1911; Koelliker, 1896; Lorente de No´, 1934; Amaral, 1978, 1979; Duffy and Rakic, 1983; Soriano et al., 1983); axonal tract tracing methods, like silver staining of lesion-induced anterograde axonal degeneration (Blackstad et al., 1970; Gaarskjaer, 1978a) and axonal transport of extraand intracellularly injected tracers (Lynch et al., 1973; Swanson et al., 1978; Gaarskjaer, 1981; West et al., 1982; Claiborne et al., 1986; Acsa´dy et al., 1998; Vida and Frotscher, 2000); histochemical and immunocytochemical staining for proteins like calbindin (Lim et al., 1997; Keuker et al., 2003) and neuropeptides like enkephalin, dynorphin, cholecystokinin (CCK), neuropeptide Y (NPY) and substance P (Fitzpatrick and Johnson, 1981; Stengaard-Pedersen et al., 1983; Gall, 1984, 1988; Zimmer et al., 1988a, b; Holm et al., 1992, 1993); and ultrastructural analysis by transmission electron microscopy alone (Fig. 1) (Blackstad and Kjaerheim, 1961; Amaral, 1979) or in combination with the other methods mentioned above (see Amaral, 1979; Frotscher and Zimmer, 1983; Frotscher, 1985, 1989; Frotscher and Seress,

Trajectory and termination of mossy fibers in CA3 (regio inferior) Within the hippocampal region, there are about one million granule cells in the rat fascia dentata (West et al., 1991), about 10 million granule cells in the domestic pig (Holm and West, 1994) and about 15 million granule cells in man (West and Gundersen, 1990). Each cell gives rise to a thin, unmyelinated MF (diameter 0.1–0.7 mm) (Gaarskjaer, 1978a; Amaral, 1979; Claiborne et al., 1986; Acsa´dy et al., 1998), which projects into the subjacent dentate hilus. During its passage through the hilus, each MF gives off a number of branching collaterals, which contact various hilar and immediately subgranularly positioned neurons, whereafter the main axons, grouped in bundles without intervening glia (Fig. 1) (Blackstad and Kjaerheim, 1961), continue toward the pyramidal cell layer of regio inferior (CA3). Within CA3, the main MFs do not branch or give off collaterals as they traject in a laminar fashion along the pyramidal cell layer across the transverse axis of the hippocampus, before bending to take a more longitudinal course in temporal direction when approaching CA1 (Gaarskjaer, 1978b, 1986; Acsa´dy et al., 1998).

87

Fig. 1. Electron microscopy (EM) (A, B) of large MF terminals (mft) from the suprapyramidal MF layer in CA3 of the adult rat hippocampus. The large terminals and terminal profiles form asymmetric synaptic contacts with a number of dendritic spines (sp). Slender extension from these (spinules, s) invade the terminals. Micrographs were kindly provided by M. Frotscher. Abbreviations: a, unmyelinated MFs; d, dendrites; mft, MF terminal; s, small dendritis spine extensions, spinules, invading MF terminals; sp, dendritic spines. Scale bars: 2 mm (A), 1 mm (B).

Supra- and infrapyramidal bands of hippocampal mossy fibers and terminals As they approach the part of the hippocampal pyramidal cell layer is closest to the hilus (CA3c), the MFs form, in classical terms, a ‘‘suprapyramidal band’’, passing along the most proximal parts of the apical dendrites of CA3 pyramidal cells, and an ‘‘infrapyramidal band’’, contacting corresponding proximal parts of the basal dendrites (Figs. 2–4A). Due to its homogeneous transparency in unstained sections, the suprapyramidal MF layer is also referred to as stratum lucidum (Ganser, 1882). The infrapyramidal MF projection is especially subject to variation among species and strains (Fig. 2 and below). By studying the MF system in so-called ‘‘extended hippocampi’’ of Wistar rats, i.e., after having stretched out the longitudinal septotemporal bend of dissected

hippocampi, Gaarskjaer (1978b) found that the suprapyramidal MF terminals in CA3c actually were located within the rather dispersed pyramidal cell layer, and that only a few infrapyramidal MF terminals were present in the corresponding area of CA3 basal dendrites. After horseradish peroxidase-labeling of MFs following intra- and extracellular injections in different parts of the granule cell layer of acute hippocampal slices from young adult Sprague-Dawley rats, Claiborne et al. (1986) found that MFs from granule cells located at different supra- to infrapyramidal locations along the transverse axis of the dentate granule cell layer projected into MF bundles both above and within the pyramidal cell layer in CA3c. MF bundles within the deep part and below the CA3c pyramidal cell layer originated from granule cells located in the infrapyramidal (free or outer) blade of the dentate. Granule cells from different

88

Fig. 2. Timm stained hippocampal sections from rat (A), guinea pig (B), cat (C), European hedgehog (D), dog (E) and man (F), illustrating the distribution of intensely stained, black MF terminals in the different species, as well as other terminal fields and general organization. Note species specific characteristics, like the ‘‘end bulb’’ at the terminal part of the MF layer in guinea pig (black asterisk, B) and the layering of the dentate hilus with MF terminal-free, intermediate or plexiforme sublayer in the same species (white asterisk, B), and MF projections into CA1 in some cats (arrows, C) and European hedgehog (D). Except for the black MF terminal staining, the neuropil of the human hippocampus is virtually unstained, due to the use of special Timm staining procedures for non-perfused tissue (Danscher and Zimmer, 1978). Abbreviations: CA1, hippocampal regio superior or subfield CA1; CA3, hippocampal regio inferior or subfield CA3; FD, fascia dentata (or denate gyrus); g, dentate granule cell layer; h, dentate hilus or CA4; m, dentate molecular layer; mf, MF layer; Sub, subiculum. Scale bars: 500 mm (See Color Plate 5.2 in color plate section.)

89

Fig. 3. Timm stained mouse hippocampal sections, illustrating strain differences in MFs terminal projections between BALB7C (A) and C57B mice (B) and Reeler mutants (C). BALB/c mice have no infrapyramidal MFs, but a short intrapyramidal bundle (black asterisk, A), compared to long infrapyramidal bundles in C57B mice (black asterisk, B). In Reeler mutant mice, granule and pyramidal cell layers are disorganized, resulting in a corresponding distribution of MF terminals in the confluent granule cell layer/hilar area and CA3. However, MFs still avoid CA1, similar to normal mice. Abbreviations: g, dentate granule cell layer; h, dentate hilus (CA4); m, dentate molecular layer. Scale bar: 500 mm (See Color Plate 5.3 in color plate section.)

locations along the transverse axis of the granule cell layer all sent MFs to the distal parts of the suprapyramidal MF layer in CA3a, adjacent to CA2/CA1 (Lorente de No´, 1934). Granule cells located near the tip of the suprapyramidal (hidden or inner) part of the granule cell layer projected directly into the suprapyramidal MF layer, distal to the hilar border, after passing through the intervening part of startum radiatum, as also shown in Ramo´n y Cajal’s figure 478 from a Golgi-stained guinea pig hippocampus (Ramo´n y Cajal, 1911). Having considered the presence and composition of supra-, intra- and infrapyramidal MF bundles in primarily Wistar and Sprague-Dawley rats, it should be noted that considerable differences exist between various strains of rats (Dimitrieva et al., 1993) and mice. BALB/cJ mice have no infrapyramidal MF bundles, but a single intrapyramidal bundle (Fig. 3A), whereas SM/J mice have a distinct infrapyramidal, but no intrapyramidal MF bundle (Barber et al., 1974). These mouse strain differences have been linked to different patterns of pyramidal cell generation (SM/J ¼ inside-out; BALB/cJ ¼ outside-in) by

Vaughn et al. (1977), but might also be induced by strain differences in thyroid hormone levels, which affect the formation of dentate granule cells and the size of the MF projection (Lauder and Mugnaini, 1980; Fredens, 1981; Lipp et al., 1984; Madeira et al., 1988), or differences in the ratio between the number of granule cells and target CA3 pyramidal cells (Gaarskjaer, 1978a). The difference in size of the granule cell layer between the BALB/cJ and C57 mice is seen by comparing Fig. 3A and B. Systemic application of the cell proliferation inhibitor methylazoxymethanolacetate (MAM) to pregnant rats during the late fetal period, when the majority of hippocampal pyramidal cells are formed, induces larger than normal infrapyramidal MF projections from the virtually normal sized fascia dentata into a smaller than normal CA3. In contrast, infrapyramidal MFs are almost absent, when MAM is administered to early postnatal pups during the time when most dentate granule cells develop. This is likely to be due to the fact that MAM reduces the number of the granule cells relative to the virtually normal numbers of CA3 pyramidal cells (unpublished observations).

90

Fig. 4. Timm stained MF terminals (black) in stratum lucidum of CA3 (mf, A) and dentate hilus (B) in semi thin plastic section from adult rat. (A) In CA3, the densely stained MF terminals in the suprapyramidal MF layer (mf) are usually large and contact CA3 pyramidal cell dendrites, but there are more homogeneous, lighter stained MF terminals in similar locations. A small group of MF terminals (imf) are present below the pyramidal cell layer (p). Lighter, Timm stained elements in stratum oriens (o) and stratum radiatum (r) are terminals of the commissural/associational projections to these layers. (B) In the hilus, a large mossy cell body (mc) is seen in contact with MF terminals, which are of variable size and also contacting other dendritic structures. Scale bars: 25 mm (A); 15 mm (B).

Both the width of the suprapyramidal layer and the extent of the infrapyramidal MF layer in CA3 vary along the longitudinal (septotemporal) axis of the rat hippocampus (Gaarskjaer, 1978a, b). At septal levels, MF terminals form an infrapyramidal layer, which extends along the deep side of the CA3 pyramidal cell layer to the CA2-CA1 transition as an almost equal counterpart to the suprapyramidal MF layer. Many MF terminals are also located within the pyramidal cell layer itself, which at these levels is wide and loosely packed. Moving temporally, the intrapyramidal and infrapyramidal MF terminals disappear first from the distal (CA3a) and intermediate (CA3b) parts of CA3. In parallel, the suprapyramidal terminal zone becomes relatively thin. The reduction in the size

of MF terminal fields corresponded to a reduction in the ratio between the number of granule cells and CA3 pyramidal cells moving from septal to more posterior and temporal levels (Gaarskjaer, 1978a). Toward the temporal end of the hippocampus, the border between the suprapyramidal MF and CA3 pyramidal cell layer becomes ‘‘fuzzy,’’ and a small portion of intrapyramidal MFs reappear distally (Gaarskjaer, 1978a, b). Transverse and longitudinal trajectories of mossy fibers in CA3 While projecting along the more proximal parts of CA3 (CA3c and adjacent parts of CA3b), rat MFs deviate only slightly in temporal direction relative

91

to the level of origin (Blackstad et al., 1970; Gaarskjaer 1978b, 1981; Claiborne et al., 1986; Amaral and Witter, 1989; Tamamaki and Nojyo, 1991), remaining within a septotemporal span or lamina of 1.1 mm (Acsa´dy et al., 1998) or less (Gaarskjaer, 1981). As the MFs enter more distal parts of CA3 (distal CA3b and CA3a) facing CA1, the fibers do, however, bend in the temporal direction to project for additional 1–2 mm along the longitudinal axis of the hippocampus (Gaarskjaer, 1978, 1981; Amaral and Witter, 1989). MFs originating at septal levels and from the tip of the suprapyramidal granule cell layer at other levels display the longest temporal descent (Gaarskjaer, 1978b, 1981; Swanson et al., 1978; Tamamaki and Nojyo, 1991; Acsa´dy et al., 1998). After 3dimensional reconstruction of seven biocytininjected granule cells from dorsal levels of the fascia dentata in adult Sprague-Dawley rats, Acsa´dy et al. (1998) measured the MF axon in CA3 to be in avarage 3.24 mm long, of which 0.80 mm was in CA3c, 1.05 mm in CA3b and 1.39 mm in CA3a. In guinea pigs, a rather abrupt turn of MFs followed by a steep, longitudinal descent in temporal direction takes place just close to and within the so-called ‘‘end bulb’’, which is a characteristic bulge on the very distal part of the suprapyramidal MF layer in this species, as demonstrated very clearly by Timm staining (Fig. 2B).

Mossy fibers projections to CA1 (regio superior) Based on Golgi-impregnated brain sections from rabbit, guinea pig, macaque and rodents, Ramo´n y Cajal (1893) and Lorente de No´ (1934) described the large, irregular MF terminals as ‘‘boutons en passant’’ which contacted the proximal dendrites of the large pyramidal cells of regio inferior (CA3), and characterized them by the presence of ‘‘thorny excrescences’’ or complex spines. In particular, Lorente de No´ (1934) noted that MFs did not proceed to contact the smaller pyramidal cells in regio superior (CA1). Located between the two subfields is, in most species, a narrow transitional area or ‘‘Mischzone’’ (Doinikow, 1908), which Lorente de No´ (1934) defined as a separate

subfield, named CA2, and characterized by a mixed content of small CA1-like pyramidal cells and larger CA3-like pyramidal cells, which, however, were devoid of complex spines and MF input (see also Tamamaki et al., 1988; Barthesaghi and Ravasi, 1999). The lack of MF projection to CA1 was confirmed by lesion-induced anterograde axonal tracing (rat, Blackstad et al., 1970), anterograde axonal transport of injected tracers, like horseradish peroxidase (rat, Lynch et al., 1973; West et al., 1982; Claiborne et al., 1986), isotope labelled amino acids (rat, Swanson et al., 1978) and biocytin (man, Lim et al., 1997; rat, Acsa´dy et al., 1998). The same conclusion was drawn from visualization of the MF terminal projection by Timm staining in rat (Fig. 2A) (Haug, 1974), several strains of mice (Fig. 3) (Barber et al., 1974; Fredens, 1981), guinea pig (Fig. 2B) (Blackstad, 1963; Geneser-Jensen et al., 1974), rabbit (HjorthSimonsen, 1977), dog (Fig. 2E), domestic pig (Holm and Geneser., 1991b) and man (Fig. 2F) (Danscher and Zimmer, 1978; Cassell and Brown, 1984). The continuation of tufts of Timm stained, thread-like elements with minor terminal-like swellings from the suprapyramidal MF layer of CA3 into CA2 in several species both confirms that CA2 acts as a transitional zone and that MFs do not reach into CA1. A dogmatic generalization of the absence of an MF projection to CA1 to all species, was, however, proven to be wrong by the discovery that MFs occasionally project into CA1 in some domestic cats (Fig. 2C) (Laurberg and Zimmer, 1980) and is normal in European hedgehogs (Fig. 2D) (Gaarskjaer et al., 1982; West et al., 1984). In the occasional cat with MFs in CA1, the projection was variable, being most extensive or only present at septodorsal levels (Fig. 2C). The MF projection to CA1 in European hedgehogs was present at midposterior (Fig. 2D) to temporal levels only, or along the entire septotemporal extent of the hippocampus. By demonstrating an overlap of MF terminallike, Timm stained elements and pyramidal cells with small cell nuclei corresponding to CA1 pyramidal cells, Gaarskjaer (1986) later provided evidence that the most temporal levels of the rat CA1 also normally receive a MF projection, a

92

phenomenon which might therefore appear to be rather common among species. Distinct guidance or stop cues for MF innervation of CA1 have not been identified so far. Certainly an increased number of granule cells relative to target CA3 pyramidal cells at septal levels in rodents does not by itself cause MFs to enter CA1, even that this is paralleled by a more extensive infrapyramidal termination in CA3 (see above), and also may be a causal factor behind the more extensive MF projections into CA1 in some cats at septodorsal levels (Fig. 2C) (Laurberg and Zimmer, 1980). Occasional observations of MF projections to CA1 in rats with early postnatal lesions of the CA3 to CA2/CA1 transition area suggest that cells in this area act as a barrier (Fig. 5; see also Cook and Crutcher, 1985). Clearly cells of the CA2 subfield have been shown to differ from the surrounding CA3 and CA1 subfields both with regard to neuronal birthdate (Angevine, 1965; Bayer, 1980) and gene expression profile (Lein et al., 2005).

Mossy fiber connections in the dentate hilus (CA4) The cytoarchitectural organization of the dentate hilus displays significant differences among species, ranging from a laminar organization in the guinea pigs (Geneser-Jensen et al., 1974) and pigs (Sus scrofu domesticus) (Holm and Geneser, 1991) to an almost confluent mixture of cell types in rodents (Blackstad, 1956; Amaral, 1978). Observations of the hilar distribution of MF collaterals and terminals in one species may accordingly not apply to another species, as exemplified by laminar distribution of large and small Timm stained MF terminals in the dentate hilus of guinea pig (Fig. 2B) compared to the confluent distribution in the dentate hilus of rat and European hedgehog (Fig. 2A and D). Another obstacle to comparisons of dentate hilar cell and neuropil layers between species is the use of different nomenclatures. As one of the first illustrations of the extensive MF projections within the dentate hilus, (Koelliker, 1896, figure 786; also reproduced by Gaarskjaer, 1986) depicted two Golgi-impregnated granule cells from a three-day-old cat. The main

Fig. 5. Timm stained section from dorsoposterior level of young adult, rat hippocampus. When newborn, the rat received a mechanical lesion of the CA3–CA1 transition (les) as well a transection of the entorhinal perforant path to the fascia dentata. The CA3–CA1 lesion induced an aberrant ingrowth of MFs into CA1, illustrated by presence of small, Timm stained MF terminals (asterisk) along the CA1 pyramidal cell layer. The removal of the perforant path projection to the outer molecular layer (m) induced spread of associational-commissural projections from inner part of the layer, and induction of supragranular MF terminals (arrow) (see Zimmer, 1973, 1974). Abbreviations: g, granule cell layer. Scale bar: 500 mm (See Color Plate 5.5 in color plate section.)

MF axons from the two cells give off several extensively branching collaterals, starting at the level of the layer of polymorphic cells and further along their trajectory toward the hippocampal pyramidal cell layer. Primarily based on observations in Golgi-stained sections of one-week- to one-monthold rabbits, guinea pigs and mice, Ramo´n y Cajal (1893, 1911) described and illustrated the emission

93

of 4–8 collaterals from main MF axons just below the granule cell layer in what he called the plexiform layer (or outer part of the polymorphic cell layer), as well as deeper into the hilus close to the hippocampal pyramidal layer from where they returned back into the hilus. Referring primarily to Ramo´n y Cajal’s well-known schematic diagram of hippocampal connections (Figure 479 in Ramo´n y Cajal, 1911), the MF collateral projections in the (rodent) hilus were, however, long taken to be of limited extension, until Soriano et al. (1983), also using Golgi-impregnation, described extensive MF collateral projections engaging major parts of the rat dentate hilus. This spurred Claiborne et al. (1986) to trace MFs and collaterals from single and small group of granule cells by intracellular and small extracellular injection of horseradish peroxidase in acute hippocampal slices from young adult Sprague-Dawley rats. Later Acsa´dy et al. (1998) combined in vivo intracellular biocytin injections of single granule cells in the dorsal fascia dentata of adult Sprague-Dawley rats with EM immunocytochemistry and 3D reconsruction to provide the most comprehensive and detailed single study of the trajectory and synaptic connections of single MF main axons and collaterals so far, including description of postsynaptic target cells. According to the combined observations of Claiborne et al. (1986) and Acsa´dy et al. (1998) in the rat, each MF main axon emits between 5 and 12 collaterals, less than 0.2 mm thick, at various distances along its course through the hilus. Including secondary and tertiary branches, the total length of the collateral plexus formed by one MF amounts to approx. 2.3 mm (Claiborne et al., 1986). In the transverse plane, collaterals from a single MF cover up to one third of the total hilar extent nearest the location of the parent granule cell, with a total transverse extent of 0.4–0.7 mm, depending on the location of the parent granule cell. Along the longitudinal axis of the dentate hilus, the spread of collaterals never seems to exceed 0.6 mm, with 90% of the collateral network being within a 0.4 mm high lamina (Acsa´dy et al., 1998). Regarding relations to the granule cell layer, individual collaterals running close to the deep border of the granule cell layer have been described in

all species examined, presumably contacting the pyramidal-shaped basket cells in the limiting subzone (cf. Ribak and Peterson, 1991). Among the studied collateral networks only one collateral was reported to enter the granule cell layer and none to reach the dentate molecular layer. This contrasts to the observations of several such MF collateral terminals in Timm staining in normal rats (Fig. 6A) and guinea pigs (Laurberg and Zimmer, 1981; Sloviter et al., 2006), as dealt with below.

Timm staining and species differences, including dentate hilus ‘‘Bulk’’ visualization of the zinc-rich MF terminals by the Timm method, which intensely stains both the characteristic large terminals of the main fibers in CA3 and the corresponding large terminals as well as the smaller MF collateral terminals in the dentate hilus, have been performed for several species, including rat (Figs. 2A and 4A, B) (Haug, 1974; Gaarskjaer, 1978b; Laurberg and Zimmer 1981; Dimitrieva et al., 1993), several strains of mice (Fig. 3) (Barber et al., 1974; Fredens, 1981, including C57/BL/6J, DBA/2J, NMRI, BALB/c/ A, C3H/Tif, A/J, AKR/A mouse strains), guinea pig (Fig. 2B) (Blackstad, 1963; Geneser-Jensen et al., 1974), rabbit (Hjorth-Simonsen, 1977), European hedgehog (Fig. 2D) (Gaarskjaer et al., 1982; West et al., 1984), cat (Fig. 2C) (Laurberg and Zimmer, 1980), dog (Fig. 2E), domestic pig (Holm and Geneser, 1991b), and man (Fig. 2F) (Danscher and Zimmer, 1978; Cassell and Brown, 1984). At low magnification, the Timm staining pattern reveals both the general similarities and the differences between the species in the both the hippocampal and the hilar MF projection. Within the dentate hilus one difference among species is a more distinct layering of hilar neurons (not illustrated) and Timm stained MF terminals in guinea pig (Fig. 2B), pig (Holm and Geneser, 1991b) and in part man (Fig. 2F), as compared to other species like rat (Fig. 2A). The mentioned species are thus characterized by having a virtually MF terminal free zone (corresponding to Ramo´n y Cajal’s (1911) intermediate or plexiforme sublayer), which is located between a narrow irregular

94

Fig. 6. A: Timm stained MF (collateral) terminals, found in normal adult rats to accompany dendrites (arrows), arising from neurons in the limiting subzone just below the dentate granule cell layer or in the dentate hilus, and extending into and sometimes through the dentate granule cell layer (g) into the, commissural-associational terminal zone (c/a). B: Timm stained, dense supragranular MF terminal projection (sgr) found normally at temporal levels of the cat fascia dentata as well as in other species. h, dentate hilus. Scale bar: 50 mm (A); 100 mm (B). (See Color Plate 5.6 in color plate section.)

band of MF collateral terminal staining just below the granule cell layer (Ramo´n y Cajal’s limiting subzone) and a more deeply located, densely stained layer containing large and small MF terminals (Ramo´n y Cajal’s deep polymorphic cell layer).

Intra- and supragranular mossy fiber collaterals — a normal feature and subject to plasticity Focusing on the limiting subzone, it is possible in all species at high magnification to observe small, presumably MF collateral terminals extending along dendrites from the hilus into the granule cell layer and sometimes as far as the molecular layer. In the rat, these intra- and supragranular Timm-positive terminals are most frequent at the dentate crest (Fig. 6A) and toward the tip of the

infrapyramidal blade (Haug, 1974; Laurberg and Zimmer, 1981; Sloviter et al., 2006), but become more widespread and numerous at temporal levels (Gaarskjaer, 1978a). Intra- and supragranular MF terminals can be very prominent at temporal levels, as shown in Fig. 6B for the cat. Aberrant growth of supragranular MF terminals into the commissural-associational terminal zone, which normally cover the inner approximately one third of the dentate molecular layer, can be induced: (a) by sufficiently strong denervation of the dentate molecular layer in developing and adult animals (Fig. 5) (Zimmer, 1973, 1974; Laurberg and Zimmer, 1981; Frotscher and Zimmer, 1983), including excitotoxic loss of hilar mossy cells (see below, Sloviter et al., 2006); (b) in intracerebral rat dentate allografts and mouse dentate xenografts (Sunde and Zimmer, 1981; Jensen et al., 1984; Frotscher and Zimmer, 1986;

95

Fig. 7. (A) Timm stained section from midposterior level of 5-day-old rat, demonstrating a distinct Timm stained MF layer in CA3 and the hilus at this age (arrows). (B, C) Parallel sections of hippocampal slice culture, derived from 7-day-old rat and grown for 3 weeks. Timm stain reveals the distribution of aberrant supragranular and normal CA3 MFs (arrow and mf, respectively, in B) and thionine cell staining to visualize general cellular organization (C). Abbreviations: g, granule cell layer. Scale bars: 500 mm (See Color Plate 5.7 in color plate section.)

Zimmer et al., 1988a, b), depending on the density of afferent host brain fiber innervation; (c) in developing hippocampal tissue slices grown as organotypic explant cultures (Fig. 7B) (Zimmer and Gahwiler, 1984); and (d) in animals with pilocarpine- or kainic acid-induced seizures (Sutula et al., 1988; Mello et al., 1993; Frotscher et al., 2006; Sloviter et al., 2006) and in epileptic patients (Sutula et al., 1989; Blu¨mcke et al., 2000). This appears to be a response to loss of dentate hilar mossy cells normally projecting to the dentate molecular layer. Transection of the main MF projection to CA3 will not by itself induce excessive supragranular MF collateral projection, excluding an axonal pruning or compensatory sprouting effect (Laurberg and Zimmer, 1980, 1981).

Structure of synaptic mossy fiber connections in the dentate hilus and CA3 As outlined by Ramo´n y Cajal (1911), refined by Lorente de No´ (1934) and demonstrated by EM (Blackstad and Kjaerheim, 1961), typical MF synapses are made by the large boutons en passant in contact with the complex spines (thorny excrescences) found on the proximal dendrites of the CA3 pyramidal cells (Fig. 1AB) and hilar mossy cells, as well as by MF collateral terminals contacting different types of neurons in the dentate hilus. Claiborne et al. (1986) and Acsa´dy et al.

(1998) extended the historical perspective by showing that each MF makes approx. 120–150 excitatory synaptic connections with predominantly inhibitory interneurons via axon collaterals in the dentate hilus, 7–12 excitatory, large terminal synaptic contacts with hilar mossy cells and 11–18 corresponding contacts with proximal pyramidal cell dendrites during the transverse MF trajectory in CA3, plus an additional number of terminal contacts made by the more distal and longitudinally running parts of the MFs. These two studies and a number of related ones (see Frotscher et al., 2006 and references therein) have changed the classical view of dentate-CA3 MF connectivity from primarily being a direct synaptic interaction between the excitatory, large MF terminals and the CA3 pyramidal cells and hilar mossy cells, into a pathway with a prominent indirect inhibitory effect not only in the dentate hilus, but also in CA3, apparently acting to refine the activation of CA3 pyramidal cells. The findings were not entirely surprising, however, because short, thick diverging processes and the thin, up to 30 mm-long filamentous or filopodial processes had been noted by Ramo´n y Cajal (1911) as extending from the angles of the large, irregularly shaped MF terminals in the dentate hilus and CA3. These extensions appear to constitute a separate, major presynaptic MF element, contacting defined types of inhibitory GABAergic and peptidergic interneurons in the dentate hilus

96

and the CA3 (Amaral, 1979; Claiborne et al., 1986; Frotscher et al., 1994; Acsa´dy et al., 1998; Henze et al., 2000; Vida and Frotscher, 2000, and references therein). Based on the data provided by Acsa´dy et al., 1998, the synaptic contacts made by a single MF can accordingly be grouped into contacts with hilar mossy cells (made by as few as 7–12 large terminals), CA3 pyramidal cells (11–18 large terminals), hilar interneurons (as many as 120–150 small terminal contacts) and interneurons in CA3 (40–50 filopodial contacts). Regarding the ultrastructural appearance of the MF synapses, the well known, classical type is made by the 4–10 mm large boutons en passant, or in several cases attached to the main MF axon by a short thin extension (Fig. 1) (Ramo´n y Cajal, 1911; Blackstad and Kjaerheim, 1961; Amaral and Dent, 1981). In the hilus these terminals contact the hilar mossy cells. In CA3 they occur at intervals of approx. 135 mm along the main MF, which has been estimated to be every 6th or 7th of the CA3 pyramidal cell that it meets along its trajectory (Claiborne et al., 1986). The contact between the giant MF terminals and the ‘‘thorny excresences’’ on CA3 pyramidal cells and hilar mossy cells are structurally intimate and complex, with spinules projecting from the excresences into the terminals (sp, Fig. 1) (Blackstad and Kjaerheim, 1961; Amaral and Dent, 1981; Ramo´n y Cajal, 1911). The terminals themselves are anchored to the dendritic shaft of CA3 pyramidal cells and to each other by puncta adhaerentia. Within the giant MF terminals (Fig. 1), three different types of vesicles have been described. The majority are small clear vesicles (
40 nm) containing glutamate (Storm-Mathisen, 1981), but the boutons also contain large dense-core vesicles containing neuropeptides such as dynorphin (McGinty et al., 1983), enkephalin (Commons and Milner, 1996), cholecystokinin (Chandy et al., 1995), neuropeptide Y (Gall et al., 1990) and neurokinin-B (Schwarzer et al., 1995) and large clear vesicles (
200 nm; Amaral and Dent, 1981; Chicurel and Harris, 1992; Henze et al., 2000). Besides glutamate and neuropeptides, MF boutons also contain the neuromodulator ATP/ adenosine (Terrian et al., 1989) and zinc (see

section about Timm staining, above). Interestingly, dentate granule cells and MFs also express both glutamic acid decarboxylase 67 and GABA, particularly after seizures (Sandler and Smith, 1991; Sloviter et al., 1996). Each large terminal synaptic complex includes up to 35 synaptic membrane specializations or active zones (Fig. 1) as reported by Chicurel and Harris (1992), which together with the high density and composition of presynaptic sodium and calcium ion channels provides each terminal with a large excitatory drive on the postsynaptic CA3 pyramidal cells (see review by Bischofberger et al., 2006). While the large MF terminals normally appear to exclusively make synaptic contact with CA3 pyramidal cells and hilar mossy cells (Acsa´dy et al., 1998), other presynaptic MF terminals contact interneurons in hilus and CA3. These presynaptic elements are smaller, do not have multiple active zones, and are located en passant or at the end of short or long thin filopodia extending from the giant MF boutons (Acsa´dy et al., 1998), often up to nine per giant bouton (Amaral, 1979; Amaral and Dent, 1981; Frotscher, 1985). In developing rats, some filopodia may extend as far as 30 mm from the giant boutons in the hilus, but shorten to approximately 12 mm in adult rats (Amaral, 1979).

New emerging structural plasticity of individual mossy fiber terminals Based on the structural data on individual MF axonal and terminal organization provided by the studies cited above, and advanced imaging techniques employed on in vivo hippocampal tissue and organotypic hippocampal slice cultures derived from transgenic mice expressing membrane targeted green fluorescent protein (GFP) in few neurons (Thy1-mGFPs ), Caroni and coworkers have engaged in detailed studies on activity dependent plasticity of individual MF termination in CA3 (De Paola et al., 2006; Galimberti et al., 2006). Results obtained so far both in vivo and in long term slice cultures have shown that one large MF terminal per MF (and only one) typically is larger

97

than the others, and that this one, besides of having elongated along its postsynaptic CA3 pyramidal cell dendrite in complex structural contact with complex spines, also can have side branches with other large, satellite MF terminals, contacting the same and neighboring pyramidal cells, thereby forming an local and interconnected, large MF terminal cluster or complex. The formation of individual larger MF terminals and new satellite large MF terminals was found to be activitydependent, regulated by age and housing in enriched environments. With increasing age single MF terminals typically attained larger size by elongating along the postsynaptic CA3 pyramidal cell dendrite, at the same time expressing more presynaptic release sites and engaging in more complex structural relations with the increasingly complex postsynaptic spines developing during the same period. Placing both young adult and older adult mice in enriched environments for a period of time on the other hand made the individual large MF terminals respond by formation of large MF terminal satellites. The mechanisms behind the MF plasticity elicited by select large MF terminals in a focal manner, as demonstrated by Caroni and coworkers, and its role in dentate-hippocampal functional interaction are not known at present. It is, however, tempting to believe that the focal expression and strengthening of synaptic contacts include a postsynaptic feedback from those CA3 pyramidal cells, which at the time of induction are activated by a net excitatory effect of the direct stimulatory and indirect inhibitory MF input.

Development of the dentate mossy fiber system Dentate neurogenesis in brief The dentate MF system has a unique developmental profile relative to other central nervous projections, due to the late start and thereafter extended neurogenesis of the parent dentate granule population, lasting the entire lifetime (Altman and Das, 1965; Angevine, 1965; Bayer, 1980; Eckenhoff and Rakic, 1988; Kuhn et al., 1996; Eriksson et al., 1998; Gould et al., 1998; Dawirs et al., 2000; Gage, 2002; Guidi et al., 2005). In rodents, two third or

more of the granule cells form after birth, and predominantly during the first two postnatal weeks, with ongoing, although declining formation of new granule cells into adulthood and old age (Bayer, 1982; Cameron et al., 1993; Kuhn et al., 1996). Neurogenesis is regulated by genetic (Kempermann et al., 1997, 2006) and epigenetic factors, including normal activity dependent ones and seizures (Gould et al., 1998; Kempermann et al., 1998; Derrick et al., 2000; Gage, 2000; Blu¨mcke et al., 2001; Song et al., 2002; Lehmann et al., 2005; Thom et al., 2005; McCloskey et al., 2006, and references therein). Before dealing with the ontogenic development of the MF projection to the dentate hilus and CA3, it should be kept in mind that late-forming dentate granule cells also have been shown to extend MFs into the hilus and CA3 (Stanfield and Trice, 1988; Hastings and Gould, 1999; van Praag et al., 2002) as well as becoming functionally integrated, as exemplified by the studies of Jessberger and Kempermann (2003), Santarelli et al. (2003), Bruel-Jungerman et al. (2005) and Raineteau et al. (2006).

Development of MF axonal and synaptic connections The developmental growth of MFs into the dentate hilus and CA3 with accompanying terminal and synaptic maturation have been studied in Wistar and Long Evans rats, mice and rabbits by Timm staining or modifications thereof (Zimmer and Haug, 1978; Gaarskjaer, 1985; SanchezAndres et al., 1997; Slomianka and Geneser, 1997), supplemented by retrograde axonal tracing (Gaarskjaer, 1985), Golgi impregnation and EM (Amaral and Dent, 1981) or just by EM (Stirling and Bliss, 1978). In rats, the first sign of MF development by Timm staining appears at the day of birth, where they appear in the hilus subjacent to the early developed parts of the suprapyramidal granule cell layer, and just above the proximal, CA3c pyramidal cell layer (at temporal and mid septo-temporal levels). Development at septal levels seems to be relatively delayed. Within the next 3–5 days, a full-length Timm-stained suprapyramidal MF layer develops, corresponding to

98

the formation of more granule cells and a widening and infrapyramidal (medial) extension of cell layer (Fig. 7A). At this time, Golgi staining revealed dentate granule cells with a relatively mature appearance, especially in the suprapyramidal blade, with MF axons extending into the hilus and sending collaterals to form an infragranular plexus (Amaral and Dent, 1981). By postnatal day 3, bundles of MFs were present above the CA3 pyramidal cell layer. By retrograde axonal tracing following microinjection of the fluorescent dye True Blue in CA3, covering different proximodistal distances from the hilus, Gaarskjaer (1985) demonstrated a rapid elongation of MF along the transverse axis of CA3 during the very first postnatal week, with some retrograde labeling of granule cells in the suprapyramidal layer from the most distal CA1 close part of CA3 (as early as postnatal day 1). The rate of MF outgrowth during the first days was estimated to be at least 0.2 mm at day 1, and nearly 1 mm at day 4. Retrograde tracing showed a continued growth along the septotemporal axis with a septotemporal descent relative to the parent cell location of about 0.5 mm at day 1, reaching nearly 1 mm at day 4 and adult length of 1.6 mm during the next 3–4 weeks. After 10– 12 days postnatal growth appeared to slow. Ultrastructurally, MF boutons identified at day 1 contained both clear and dense core vesicles and formed both symmetrical and asymmetrical synaptic contacts with the apical dendrites of CA3 pyramidal cells, despite the fact that spines were not yet present (Amaral and Dent, 1981). During the following week, the terminals developed a more mature and complex morphology, size and synaptic vesicle content and density. Spines on the proximal CA3 pyramidal cell dendrites started to develop about one week postnatal, but a mature, classic, MF terminal and its intimate relations with ‘‘thorny excrescences’’ developed at three weeks postnatal. Thereafter, both the size of the MF terminals, and the density of Timm staining, continued to increase. It should be noted that the MF system is not mature in two-week-old rats, as demonstrated by Amaral and Dent (1981), who found the mean area and perimeter of the MF terminal profiles to increase substantially during this period, together with a three-fold increase in

the average number of symmetrical membrane junctions (punctae adhaerentia) to seven per terminal profile at day 21. Asymmetrical synapses at the same time reached the adult level of an average number of nine per terminal profile, counted in two dimensions only. It should also be recognized that, after the average number of invaginating spinules per terminal profile increased from two at day 21 to five in more than one-yearold rats. As illustrated above, the rat MF system is subject to considerable development during the first three postnatal weeks and even thereafter, and this should of course be taken into consideration in experimental and other studies of this system. Specific points to note may be that MF synapse formation precedes the development of thorny excrecenses on CA3 pyramidal cells. Moreover, asymmetrical synapses seem to develop at a steady pace, unrelated to the dramatic structural consolidation expressed by a prominent increase in the non-synaptic puncta adhaerentia from day 14 to day 21.

Experimental studies on structural and connective plasticity of dentate mossy fibers, including dentate transplants The parent dentate granule cells, the presynaptic MFs and their terminals and the postsynaptic target cells in the dentate hilus and the hippocampal CA3 subfield are each and interdependently affected by genetic and epigenetic factors and conditions, the effects of which are expressed as direct structural and connective abnormalities, deviations from normal or natural variations among species, strains and individuals. New observations on activity dependent plasticity of individual MF terminals and their postsynaptic CA3 pyramidal cells have already been dealt with above, just as species (and mouse strain) differences have been presented. This section is therefore devoted to other experimental manipulations, ranging from breeding studies and effects of hormonal administration to studies of lesion-induced effects and neural transplantation.

99

Genetics and breeding experiments Several studies have investigated the MF system in mice and rats, selectively bred on the basis of behavioral criteria such as low and high cognitive performance levels in defined tests. These studies have demonstrated significant differences in the organization and size of the MF projection to CA3 (see Bernasconi-Guastalla et al., 1994; Lipp and Wolfer, 1998; and references therein). Using such a heritable difference between two mouse strains (CH3 and CPB-K) as a starting point, Nek et al. (1993) demonstrated that MF terminals in the strain with fewer MFs to CA3 were larger and established more spine synapses per terminal than in the strain with more MFs. Accordingly, the fiber density seemed to have a regulatory effect on the morphology and synaptic contacts of the terminals. Numerous mutants and gene-modified mice with neurological functional and structural defects exist, and have been and are currently screened for effects on cognitive hippocampus-related function and developmental changes in cellular and connective organization. Only the mutant Reeler mouse will be mentioned here, because the mutation clearly affects the organization of the dentate and hippocampal cell layers and as such has involved experimental studies of the hippocampus (Fo¨rster et al., 2006). As seen in Fig. 3C, the MF system is present, but appears disorganized in Reeler mice. This is, however, not due to change in cellular specificity in termination, but a consequence of the disordered location of cells.

Hormonal effects on MF development and cognitive function Lauder and Mugnaini (1980) and Lipp et al. (1984) early on demonstrated the sensitivity of the rodent MF system to early postnatal hyperthyreoidism, inducing an increase in particular infrapyramidal MFs. Interestingly treatment of mice from a strain (DBA/2) virtually devoid of infrapyramidal MFs (correlating with poor radial-maze learning) with daily injections of 2 mg L-thyroxine early postnatally, both induced a prominent infrapyramidal

MF projection and a significant improvement in radial-maze learning (Schwegler et al., 1991). Stress and glucocorticoid stress hormones also effect the MF system (McEwen, 1999), and most interestingly such induced changes seems to be reversible by behavioral training (Sandi et al., 2003).

Lesion-induced MF sprouting and rerouting Several studies referred to above have included interference with the development or integrity of dentate granule cells, hilar cells and CA3 pyramidal cells resulting in changes in the size and distribution of MFs, like for example induction of supragranular MFs following seizures and denervation of the dentate molecular layer (Fig. 5), changes in the size of infrapyramidal projections to CA3 or ingrowths into CA1 (Fig. 5). Only a few studies have addressed the reaction of MFs to transection during development or in the mature brain. In two studies on rat MFs by Laurberg and Zimmer (1980, 1981), supplemented by a subsequent EM study (Frotscher and Zimmer, 1983) and observations in organotypic slice cultures (Zimmer and Gahwiler, 1984, 1987), it was demonstrated that: (1) the main MF projection to CA3, with an decrease in effect over the first 3 postnatal weeks, can be rerouted along the septotemporal axis in response to lesion-induced removal or blockade of access to CA3 pyramidal cells at normal levels of trajectory; (2) the rerouted MFs terminate in expanded suprapyramidal and infrapyramidal terminal zones in accessible adjacent levels; (3) loss of CA3 target cells not by itself elicits an axonal pruning effect with compensatory sprouting of MFs or collaterals into the dentate molecular layer; and (4) lesion-induced sprouting of MF terminals into the dentate molecular layer could be induced in both postnatal developing and young adult rats in a graded manner depending on the density of denervation of the dentate molecular layer. As possible regrowth of severed or newly formed MFs across lesions in CA3 were difficult to prove in vivo, given the possibility for fibers to circumvent the lesion and then invade the normal target area, transections of the MF layer in hippocampal slice cultures and coculturing of dentate

100

and CA3 slice cultures derived from 7-day-old rats were employed. Under both conditions regrowth of MFs into distal CA3 required an intimate contact between the proximal transected MF layer and the distal, normal target zone, contrasting a vivid growth of MFs onto the substratum of glial cells surrounding the cultures.

Neural grafting, testing MF growth and plasticity Intracerebral grafting of tissue blocks of developing fascia dentata from newborn (or late fetal) rats and mice have been used to identify factors regulating the growth and specificity in the formation of nerve connections in the brain in general and to and from the dentate gyrus specifically (Sunde and Zimmer, 1981, 1983; Frotscher and Zimmer, 1986; Zimmer et al., 1988a, b; Tønder et al., 1990). In contrast to extensive growth of monoaminergic and cholinergic graft fibers into host brain and vice

versa at all recipient brain ages, glutamatergic projections, including those of the hippocampal region, are clearly dependent of recipient age with regard to exchange of fibers between the always developing graft tissue and the host brain. In accordance with these general observations, MFs were only seen to grow from dentate grafts into the recipient brain and from the host brain fascia dentata into grafts under certain conditions. More over, except for a minor projection of MFs growing from dentate grafts into the surrounding CA1 area (Sunde and Zimmer, 1981), ingrowth and termination of MFs only occurred into the normal CA3 target area. Figure 8 illustrates one experimental setup, where a small block of neonatal rat fascia dentata had been grafted into the dentate area (Fig. 8B) or next to CA3 (Fig. 8C) in other newborn rats with the purpose of examining whether transplanted dentata granule cells were able to restore the MF projection of the recipient rats, which before the

Fig. 8. (A) Timm stained sections of hippocampus from adult rat, subjected to hippocampal x-irradiation as newborn, stop odentate granule cell formation and lead to few MF terminals. (B) Timm stained section from the contralateral hemisphere of the rat shown in A, depicting a well-integrated dentate transplant (tpl), derived from a small block of fascia dentata, grafted just after the x-irradiation. From the graft, an apparently normal, laminar-specific MF projection (mf) developed, connecting dentate granule cells of the graft (g) with the host CA3. The molecular layer of the graft (m) received a comparably laminar-specific projection of host entorhinal perforant path projections, normalizing the Timm stained laminar appearance of the layer. (C) Slightly displaced dentate transplant (tpl), grafted to x-irradiated newborn hippocampus just after irradiation as part of same experiment (Sunde et al., 1984). Encroaching on the recipient CA3, MFs from the granule cells of this graft has entered adjacent parts of CA3 and project ‘‘downstream’’, but appear to stop at the border with CA1. Surprisingly, no MFs from the graft projected in the ‘‘upstream’’ direction, i.e., toward the host fascia dentata. Scale bar: 500 mm (See Color Plate 5.8 in color plate section.)

101

transplantation had selectively x-irradiated over the hippocampal region, resulting in a severely reduced number of granule cells and MFs for the rest of the animal’s life (Fig. 8A). In the situation with both graft and recipient being neonatal and the graft positioned corresponding to the recipient fascia dentata (Fig. 8B) there was extensive and laminar specific outgrowth of host MFs into the recipient CA3, but no outgrowth ‘‘backward’’ into areas posterior to the graft. Note also in this homotopically placed graft the extensive, laminar specific ingrowth of host entorhinal perforant fibers into the outer parts of the graft molecular layer, illustrated by the presence of normal Timm stained lamination of this layer (m, Fig. 8B.). With placement of the dentate graft in the lateral ventricle, encroaching on the distal parts of CA3

(Fig. 8C), graft MFs entered and innervated the host CA3 in a specific and laminar fashion from the site of entry and ‘‘downstream’’ to CA1. For unknown reasons no graft MFs grew in the opposite of normal direction toward the residual host fascia dentata. Using the presence of the peptide cholecystokinin (CCK) in mouse, but not rat, MF terminals, Zimmer et al. (1988a, b) moreover demonstrated that mouse MFs, originating from a newborn mouse dentate xenograft placed next to the recipient rat fascia dentata, were able to participate in the normal development of the recipient MF system, as illustrated by the presence of mouse xenograft-derived, CCK immunoreactive MF terminals in the adult host rat CA3 (Fig. 9). With proper location of dentate grafts providing direct access

Fig. 9. Innervation of rat MF layer by mouse MF terminals from mouse dentate xenograft, topographically integrated with the recipient rat fascia dentata. (A) Adult rat MF layer stained immunocytochemically for the peptide cholecystokinin (CCK), disclosing some small-size, terminal-like structures with a preferential distribution among the CA3 pyramidal cells (p), and likely to arise from the CCK-immunoreactive neurons (arrow heads). (B) Corresponding level of MF layer in contralateral hippocampus, which early postnatally had received a xenograft of neonatal mouse fascia dentata, encroaching on the rat recipient fascia dentata for some distance along the septotemporal axis. Corresponding to the levels of xenograft integration with the rat host fascia dentata and for some additional distance temporally, mouse MF terminals, distinguished by their normal immunoreactivity for CCK, were present in the host rat MF layer, as documented by their large MF terminal-like size (arrows), and their preferential suprapyramidal position, unlike the smaller elements in the pyramidal cell layer (p) from the CCK-reaction intrinsic neurons (arrowheads). Scale bar: 50 mm.

102

Fig. 10. Adjacent Timm (A) and toluidine blue (B) stained sections from the dorsal hippocampus of adult rat, showing a focal lesion of CA3 containing a small graft of fetal rat CA3 neurons, receiving a Timm stained, MF projection (asterisk, A) from the host fascia dentata. In a two-step procedure, the rat received a localized injection of ibotenic acid, killing all CA3 neurons at the injection site. One week later a cell suspension of late fetal rat CA3 neurons were grafted into the lesion site, where the afferent fibers, here dentate MFs, still remained. Several weeks postgrafting, the CA3 graft (tpl) was found to be well integrated, yet not filling the lesion completely. Part of lesioned CA3 pyramidal cell layer is thus still visible, but densely gliotic (p3). As demonstrated by Timm staining (asterisk, A), host MFs have innervated the fetal CA3 graft. Scale bar: 100 mm.

for graft MFs to enter the recipient MF layer, graft-derived MFs may accordingly innervate the CA3 area of the host hippocampus. Ingrowth of adult recipient brain, glutamatergic fibers into developing graft tissue are normally very restricted, apparently because the glutamatergic fibers, unlike monoaminergic and cholinergic host fibers, are competing with the intrinsic graft fibers, which are in their normal phase of development. One exception to this is when grafts are placed in so called ‘‘axon-sparing’’ lesions, where recipient brain neurons have been killed by ischemia or excitotoxins like ibotenic acid one or more days prior to grafting. Using this paradigm, Tønder et al. (1990) demonstrated that adult recipient brain MFs, primed by preceding ibotenic acid lesion of their normal target CA3 pyramidal cells, indeed were able to innervate grafts of late fetal CA3 pyramidal neurons, placed in the lesioned CA3 area (Fig. 10). Adult MFs with already established connection area can accordingly after preceding axon-sparing lesions of CA3 enter and terminate in fetal grafts of CA3 neurons.

Acknowledgment The authors are grateful for the support received by former and current colleagues during the writing of this review, primarily in terms of help with provision of relevant references for reading and discussions. Prof. Michael Frotscher, Freiburg, Germany is particularly acknowledged for providing the EM pictures used as Fig. 1. Prof. Pico Caroni, Basel, Switzerland is thanked for personal communications on the newly emerging MF terminal plasticity.

References Acsa´dy, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18(9): 3386–3403. Altman, J. and Das, G.D. (1965) Post-natal origin of microneurones in the rat brain. Nature, 207(5000): 953–956. Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182(4 Pt 2): 851–914.

103 Amaral, D.G. (1979) Synaptic extensions from the mossy fibers of the fascia dentata. Anat. Embryol. (Berl.), 155(3): 241–251. Amaral, D.G. and Dent, J.A. (1981) Development of the mossy fibers of the dentate gyrus. I. A light and electron microscopic study of the mossy fibers and their expansions. J. Comp. Neurol., 195(1): 51–86. Amaral, D.G. and Witter, M.P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31(3): 571–591. Andersen, P., Bliss, T.V., Lomo, T., Olsen, L.I. and Skrede, K.K. (1969) Lamellar organization of hippocampal excitatory pathways. Acta Physiol. Scand., 76(1): 4A–5A. Angevine, J.B. Jr., (1965) Time of neuron origin in the hippocampal region. An autoradiographic study in the mouse. Exp. Neurol., Suppl 2: 1–70. Barber, R.P., Vaughn, J.E., Wimer, R.E. and Wimer, C.C. (1974) Genetically-associated variations in the distribution of dentate granule cell synapses upon the pyramidal cell dendrites in mouse hippocampus. J. Comp. Neurol., 156(4): 417–434. Barthesaghi, R. and Ravasi, L. (1999) Pyramidal types in the field CA2 of the guinea pig. Brain Res. Bull., 50: 263–273. Bayer, S.A. (1980) Development of the hippocampal region in the rat. I. Neurogenesis examined with 3H-thymidine autoradiography. J. Comp. Neurol., 190(1): 87–114. Bayer, S.A. (1982) Changes in the total number of dentate granule cells in juvenile and adult rats: a correlated volumetric and 3H-thymidine autoradiographic study. Exp. Brain Res., 46(3): 315–323. Bernasconi-Guastalla, S., Wolfer, D.P. and Lipp, H.P. (1994) Hippocampal mossy fibers and swimming navigation in mice: correlations with size and left-right asymmetries. Hippocampus, 4(1): 53–63. Bischofberger, J., Engel, D., Frotscher, M. and Jonas, P. (2006) Timing and efficacy of transmitter release at mossy fiber synapses in the hippocampal network. Pflugers Arch., 453(3): 361–372. Blackstad, T.W. (1956) Commissural connections of the hippocampal region in the rat, with special reference to their mode of termination. J. Comp. Neurol., 105(3): 417–537. Blackstad, T.W. (1963) Ultrastructural studies on the hippocampal region. Prog. Brain Res., 3: 122–148. Blackstad, T.W., Brink, K., Hem, J. and Jeune, B. (1970) Distribution of hippocampal mossy fibers in the rat. An experimental study with silver impregnation methods. J. Comp. Neurol., 138(4): 433–449. Blackstad, T.W. and Kjaerheim, A. (1961) Special axodendritic synapses in the hippocampal cortex: electron and light microscopic studies on the layer of mossy fibers. J. Comp. Neurol., 117: 133–159. Blu¨mcke, I., Schewe, J.C., Normann, S., Brustle, O., Schramm, J., Elger, C.E. and Wiestler, O.D. (2001) Increase of nestinimmunoreactive neural precursor cells in the dentate gyrus of pediatric patients with early-onset temporal lobe epilepsy. Hippocampus, 11(3): 311–321.

Blu¨mcke, I., Suter, B., Behle, K., Kuhn, R., Schramm, J., Elger, C.E. and Wiestler, O.D. (2000) Loss of hilar mossy cells in Ammon’s horn sclerosis. Epilepsia, 41(Suppl 6): S174–S180. Bruel-Jungerman, E., Laroche, S. and Rampon, C. (2005) New neurons in the dentate gyrus are involved in the expression of enhanced long-term memory following environmental enrichment. Eur. J. Neurosci., 21(2): 513–521. Cameron, H.A., Woolley, C.S., McEwen, B.S. and Gould, E. (1993) Differentiation of newly born neurons and glia in the dentate gyrus of the adult rat. Neuroscience, 56(2): 337–344. Cassell, M.D. and Brown, M.W. (1984) The distribution of Timm’s stain in the nonsulphide-perfused human hippocampal formation. J. Comp. Neurol., 222(3): 461–471. Chandy, J., Pierce, J.P. and Milner, T.A. (1995) Rat hippocampal mossy fibers contain cholecystokinin-like immunoreactivity. Anat. Rec., 243(4): 519–523. Chicurel, M.E. and Harris, K.M. (1992) Three-dimensional analysis of the structure and composition of CA3 branched dendritic spines and their synaptic relationships with mossy fiber boutons in the rat hippocampus. J. Comp. Neurol., 325(2): 169–182. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus. J. Comp. Neurol., 246(4): 435–458. Commons, K.G. and Milner, T.A. (1996) Ultrastructural relationships between leu-enkephalin- and GABA-containing neurons differ within the hippocampal formation. Brain Res., 724(1): 1–15. Cook, T.M. and Crutcher, K.A. (1985) Extensive target cell loss during development results in mossy fibers in the regio superior (CA1) of the rat hippocampal formation. Brain Res., 353(1): 19–30. Danscher, G. (1981) Histochemical demonstration of heavy metals. A revised version of the sulphide silver method suitable for both light and electronmicroscopy. Histochemistry, 71(1): 1–16. Danscher, G., Howell, G., Perez-Clausell, J. and Hertel, N. (1985) The dithizone, Timm’s sulphide silver and the selenium methods demonstrate a chelatable pool of zinc in CNS. A proton activation (PIXE) analysis of carbon tetrachloride extracts from rat brains and spinal cords intravitally treated with dithizone. Histochemistry, 83(5): 419–422. Danscher, G. and Zimmer, J. (1978) An improved Timm sulphide silver method for light and electron microscopic localization of heavy metals in biological tissues. Histochemistry, 55(1): 27–40. Dawirs, R.R., Teuchert-Noodt, G., Hildebrandt, K. and Fei, F. (2000) Granule cell proliferation and axon terminal degradation in the dentate gyrus of gerbils (Meriones unguiculatus) during maturation, adulthood and aging. J. Neural Transm., 107(6): 639–647. De Paola, V., Arber, S. and Caroni, P. (2006) AMPA receptors regulat edynamic equilibrium of presynaptic terminals in mature hippocampal networks. Nat. Neurosci., 6: 491–500. Derrick, B.E., York, A.D. and Martinez Jr., J.L. (2000) Increased granule cell neurogenesis in the adult dentate gyrus

104 following mossy fiber stimulation sufficient to induce longterm potentiation. Brain Res., 857(1–2): 300–307. Dimitrieva, N.I., Gozzo, S., Dmitriev, I.u.S., Lopatina, N.G. and Kassil’, V.G. (1993) Neuroanatomical characteristics of rat strains differing in the ability to form active avoidance conditioned reflexes. Dokl. Akad. Nauk SSSR, 272(5): 1235–1238. Doinikow, B. (1908) Beitrag zur histologie des Ammonshorns. J. Psychol. Neurol. (Leipzig), 13: 166–202. Duffy, C.J. and Rakic, P. (1983) Differentiation of granule cell dendrites in the dentate gyrus of the rhesus monkey: a quantitative Golgi study. J. Comp. Neurol., 214(2): 224–237. Eckenhoff, M.F. and Rakic, P. (1988) Nature and fate of proliferative cells in the hippocampal dentate gyrus during the life span of the rhesus monkey. J. Neurosci., 8(8): 2729–2747. Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M., Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Neurogenesis in the adult human hippocampus. Nat. Med., 4(11): 1313–1317. Fitzpatrick, D. and Johnson, R.P. (1981) Enkephalin-like immunoreactivity in the mossy fiber pathway of the hippocampal formation of the tree shrew (Tupaia glis). Neuroscience, 6(12): 2485–2494. Fo¨rster, E., Jossin, Y., Zhao, X., Frotscher, M. and Goffinet, A.M. (2006) Recent progress in understanding the role of Reelin in the radial neuronal migration, with specific emphasis on the dentate gyrus. Eur. J. Neurosci., 23: 901–909. Fredens, K. (1981) Genetic variation in the histoarchitecture of the hippocampal region of mice. Anat. Embryol. (Berl.), 161(3): 265–281. Frotscher, M. (1985) Mossy fibres form synapses with identified pyramidal basket cells in the CA3 region of the guinea-pig hippocampus: a combined Golgi-electron microscope study. J. Neurocytol., 14(2): 245–259. Frotscher, M. (1989) Mossy fiber synapses on glutamate decarboxylase-immunoreactive neurons: evidence for feed-forward inhibition in the CA3 region of the hippocampus. Exp. Brain Res., 75(2): 441–445. Frotscher, M., Jonas, P. and Sloviter, R.S. (2006) Synapses formed by normal and abnormal hippocampal mossy fibers. Cell Tissue Res., 326(2): 361–367. Frotscher, M. and Seress, L. (1991) Basket cells in the monkey fascia dentata: a Golgi/electron microscopic study. J. Neurocytol., 20(11): 915–928. Frotscher, M., Soriano, E. and Misgeld, U. (1994) Divergence of hippocampal mossy fibers. Synapse, 16(2): 148–160. Frotscher, M. and Zimmer, J. (1983) Lesion-induced mossy fibers to the molecular layer of the rat fascia dentata: identification of postsynaptic granule cells by the Golgi-EM technique. J. Comp. Neurol., 215(3): 299–311. Frotscher, M. and Zimmer, J. (1986) Intracerebral transplants of the rat fascia dentata: a Golgi/electron microscope study of dentate granule cells. J. Comp. Neurol., 246(2): 181–190. Gaarskjaer, F.B. (1978a) Organization of the mossy fiber system of the rat studied in extended hippocampi. II.

Experimental analysis of fiber distribution with silver impregnation methods. J. Comp. Neurol., 178(1): 73–88. Gaarskjaer, F.B. (1978b) Organization of the mossy fiber system of the rat studied in extended hippocampi. I. Terminal area related to number of granule and pyramidal cells. J. Comp. Neurol., 178(1): 49–72. Gaarskjaer, F.B. (1981) The hippocampal mossy fiber system of the rat studied with retrograde tracing techniques. Correlation between topographic organization and neurogenetic gradients. J. Comp. Neurol., 203(4): 717–735. Gaarskjaer, F.B. (1985) The development of the dentate area and the hippocampal mossy fiber projection of the rat. J. Comp. Neurol., 241(2): 154–170. Gaarskjaer, F.B. (1986) The organization and development of the hippocampal mossy fiber system. Brain Res., 396(4): 335–357. Gaarskjaer, F.B., Danscher, G. and West, M.J. (1982) Hippocampal mossy fibers in the regio superior of the European hedgehog. Brain Res., 237(1): 79–90. Gage, F.H. (2000) Mammalian neural stem cells. Science, 287(5457): 1433–1438. Gage, F.H. (2002) Neurogenesis in the adult brain. J. Neurosci., 22(3): 612–613. Galimberti, I., Gogolia, N., Alberi, S., Santos, A.F., Muller, D. and Caroni, P. (2006) Long-term rearrangements of hippocampal mossy fiber terminal connectivity in the adult regulated by experience. Neuron, 50: 749–763. Gall, C. (1984) The distribution of cholecystokinin-like immunoreactivity in the hippocampal formation of the guinea pig: localization in the mossy fibers. Brain Res., 306(1–2): 73–83. Gall, C. (1988) Seizures induce dramatic and distinctly different changes in enkephalin, dynorphin, and CCK immunoreactivities in mouse hippocampal mossy fibers. J. Neurosci., 6: 1852–1862. Gall, C., Lauterborn, J., Isackson, P. and White, J. (1990) Seizures, neuropeptide regulation, and mRNA expression in the hippocampus. Prog. Brain Res., 83: 371–390. Ganser, S. (1882) Vergleichend-anatomische studien u¨ber das gehirn des maulwurfs. Morphol. Jahrb., 7: 591–725. Geneser-Jensen, F.A., Haug, F.M. and Danscher, G. (1974) Distribution of heavy metals in the hippocampal region of the guinea pig. A light microscope study with Timm’s sulfide silver method. Z. Zellforsch. Mikrosk. Anat., 147(4): 441–478. Golgi, C. (1886) Sulla Fina Anatomia Degli Organi Centrali Del Sistema Nervoso. U. Hoepli, Milano, p. 215. Gould, E., Tanapat, P., McEwen, B.S., Flugge, G. and Fuchs, E. (1998) Proliferation of granule cell precursors in the dentate gyrus of adult monkeys is diminished by stress. Proc. Natl. Acad. Sci. U.S.A., 95(6): 3168–3171. Guidi, S., Ciani, E., Severi, S., Contestabile, A. and Bartesaghi, R. (2005) Postnatal neurogenesis in the dentate gyrus of the guinea pig. Hippocampus, 15(3): 285–301. Hastings, N.B. and Gould, E. (1999) Rapid extension of axons into the CA3 region by adult-generated granule cells. J. Comp. Neurol., 413(1): 146–154.

105 Haug, F.M. (1967) Electron microscopical localization of the zinc in hippocampal mossy fibre synapses by a modified sulfide silver procedure. Histochemie, 8(4): 355–368. Haug, F.M. (1974) Light microscopical mapping of the hippocampal region, the pyriform cortex and the corticomedial amygdaloid nuclei of the rat with Timm’s sulphide silver method. I. Area dentata, hippocampus and subiculum. Z. Anat. Entwicklungsgesch., 145(1): 1–27. Henze, D.A., Urban, N.N. and Barrionuevo, G. (2000) The multifarious hippocampal mossy fiber pathway: a review. Neuroscience, 98(3): 407–427. Hjorth-Simonsen, A. (1977) Distribution of commissural afferents to the hippocampus of the rabbit. J. Comp. Neurol., 176(4): 495–513. Holm, I.E. and Geneser, F.A. (1991) Histochemical demonstration of zinc in the hippocampal region of the domestic pig. III. The dentate area. J. Comp. Neurol., 308(3): 409–417. Holm, I.E., Geneser, F.A. and Zimmer, J. (1992) Somatostatinand neuropeptide Y-like immunoreactivity in the dentate area, hippocampus, and subiculum of the domestic pig. J. Comp. Neurol., 322(3): 390–408. Holm, I.E., Geneser, F.A. and Zimmer, J. (1993) Cholecystokinin-, enkephalin-, and substance P-like immunoreactivity in the dentate area, hippocampus, and subiculum of the domestic pig. J. Comp. Neurol., 331(3): 310–325. Holm, I.E. and West, M.J. (1994) Hippocampus of the domestic pig: a stereological study of subdivisional volumes and neuron numbers. Hippocampus, 4(1): 115–125. Jensen, S., Sorensen, T., Møller, A.G. and Zimmer, J. (1984) Intraocular grafts of fresh and freeze-stored rat hippocampal tissue: a comparison of survivability and histological and connective organization. J. Comp. Neurol., 227(4): 559–568. Jessberger, S. and Kempermann, G. (2003) Adult-born hippocampal neurons mature into activity-dependent responsiveness. Eur. J. Neurosci., 18(10): 2707–2712. Kempermann, G., Chesler, E.J., Lu, L., Williams, R.W. and Gage, F.H. (2006) Natural variation and genetic covariance in adult hippocampal neurogenesis. Proc. Natl. Acad. Sci. U.S.A., 103(3): 780–785. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) Genetic influence on neurogenesis in the dentate gyrus of adult mice. Proc. Natl. Acad. Sci. U.S.A., 94(19): 10409–10414. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Experience-induced neurogenesis in the senescent dentate gyrus. J. Neurosci., 18(9): 3206–3212. Keuker, J.I., Rochford, C.D., Witter, M.P. and Fuchs, E. (2003) A cytoarchitectonic study of the hippocampal formation of the tree shrew (Tupaia belangeri). J. Chem. Neuroanat., 26(1): 1–15. Koelliker, A. (1896) Handbuch der Gewebelehre des Menchen 2. Nervensystem des Menchen und der Thiere. Wilhelm Engelmann, Leipzig, p. 874. Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Neurogenesis in the dentate gyrus of the adult rat: age-related decrease of neuronal progenitor proliferation. J. Neurosci., 16(6): 2027–2033.

Lauder, J.M. and Mugnaini, E. (1980) Infrapyramidal mossy fibers in the hippocampus of the hyperthyroid rat. A light and electron microscopic study. Dev. Neurosci., 3(4–6): 248–265. Laurberg, S. and Zimmer, J. (1980) Aberrant hippocampal mossy fibers in cats. Brain Res., 188(2): 555–559. Laurberg, S. and Zimmer, J. (1981) Lesion-induced sprouting of hippocampal mossy fiber collaterals to the fascia dentata in developing and adult rats. J. Comp. Neurol., 200(3): 433–459. Lehmann, K., Butz, M. and Teuchert-Noodt, G. (2005) Offer and demand: proliferation and survival of neurons in the dentate gyrus. Eur. J. Neurosci., 21(12): 3205–3216. Lein, E.S., Callaway, E.M., Albright, T.D. and Gage, F.H. (2005) Redefining the boundaries of the hippocampal CA2 subfield in the mouse using gene expression and 3-dimensional reconstruction. J. Comp. Neurol., 485(1): 1–10. Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997) Connections of the hippocampal formation in humans: I. The mossy fiber pathway. J. Comp. Neurol., 385(3): 325–351. Lipp, H.P., Schwegler, H. and Driscoll, P. (1984) Postnatal modification of hippocampal circuitry alters avoidance learning in adult rats. Science, 225(4657): 80–82. Lipp, H.P. and Wolfer, D.P. (1998) Genetically modified mice and cognition. Curr. Opin. Neurobiol., 8(2): 272–280. Lorente de No´, R. (1934) Studies on the structure of the cerebral cortex II. Continuation of the study of the ammonic system. J. Psychol. Neurol. (Leipzig), 46: 113–177. Lynch, G., Smith, R.L., Mensah, P. and Cotman, C. (1973) Tracing the dentate gyrus mossy fiber system with horseradish peroxidase histochemistry. Exp. Neurol., 40(2): 516–524. Madeira, M.D., Paula-Barbosa, M., Cadete-Leite, A. and Tavares, M.A. (1988) Unbiased estimate of hippocampal granule cell numbers in hypothyroid and in sex-age-matched control rats. J. Hirnforsch., 29(6): 643–650. McCloskey, D.P., Hintz, T.M., Pierce, J.P. and Scharfman, H.E. (2006) Stereological methods reveal the robust size and stability of ectopic hilar granule cells after pilocarpine-induced status epilepticus in the adult rat. Eur. J. Neurosci., 24(8): 2203–2210. McEwen, B.S. (1999) Stress and hippocampal plasticity. Ann. Rev. Neurosci., 22: 105–122. McGinty, J.F., Henriksen, S.J., Goldstein, A., Terenius, L. and Bloom, F.E. (1983) Dynorphin is contained within hippocampal mossy fibers: immunochemical alterations after kainic acid administration and colchicine-induced neurotoxicity. Proc. Natl. Acad. Sci. U.S.A., 80(2): 589–593. Mello, L.E., Cavalheiro, E.A., Tan, A.M., Kupfer, W.R., Pretorius, J.K., Babb, T.L. and Finch, D.M. (1993) Circuit mechanisms of seizures in the pilocarpine model of chronic epilepsy: cell loss and mossy fiber sprouting. Epilepsia, 34(6): 985–995. Nek, N., Schwegler, H., Crusio, WE. and Frotscher, M. (1993) Are the fine-structural characteristics of mouse hippocampal mossy fiber synapses determined by the density of mossy fiber axons? Neurosci. Lett., 158(1): 75–78.

106 van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer, T.D. and Gage, F.H. (2002) Functional neurogenesis in the adult hippocampus. Nature, 415(6875): 1030–1034. Raineteau, O., Hugel, S., Ozen, I., Rietschin, L., Sigrist, M., Arber, S. and Gahwiler, B.H. (2006) Conditional labeling of newborn granule cells to visualize their integration into established circuits in hippocampal slice cultures. Mol. Cell. Neurosci., 32(4): 344–355. Ramo´n y Cajal, S.R. (1893) Estructura del asta de Ammon. Anal. Soc. Esp. Hist. Nat. (Madrid), 22: 53–114. Ramo´n y Cajal, S.R. (1911) Histologie du Syste`me Nerveux de l’Homme et des Verte´bre´s, Vol. II. Maloine, Paris. Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy fibers in rats and gerbils form synapses with the somata and proximal dendrites of basket cells in the dentate gyrus. Hippocampus, 1(4): 355–364. Sala, L. (1891) Zur feineren anatomie des grossen seepherdefusses. Z. Wiss. Zoll., 52: 18–45. Sanchez-Andres, J.V., Palop, J.J., Ramirez, C., Nacher, J., Molowny, A. and Lopez-Gracia, C. (1997) Zinc-positive presynaptic boutons of the rabbit hippocampus during early postnatal development. Dev. Brain Res., 103(2): 171–183. Sandi, C., Davies, H.A., Cordero, M.I., Rodriquez, J.J., Popov, V.I. and Stewart, M.G. (2003) Rapid reversal of stress induced loss of synapses in CA3 of rat hippocampal following water maze training. Eur. J. Neurosci., 17: 2447–2456. Sandler, R. and Smith, A. (1991) Coexistence of GABA and glutamate in mossy fiber terminals of the primate hippocampus: an ultrastructural study. J. Comp. Neurol., 303(2): 177–192. Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F., Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O., Belzung, C. and Hen, R. (2003) Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science, 301(5634): 805–809. Schaffer, K. (1882) Beitrag zur histologie der Ammonshornformation. Arch. Mikr. Anat., 39: 611–632. Schwarzer, C., Williamson, J.M., Lothman, E.W., Vezzani, A. and Sperk, G. (1995) Somatostatin, neuropeptide Y, neurokinin B and cholecystokinin immunoreactivity in two chronic models of temporal lobe epilepsy. Neuroscience, 69(3): 831–845. Schwegler, H., Crusio, W.E., Lipp, H.P., Brust, I. and Mueller, G.G. (1991) Early postnatal hyperthyroidism alters hippocampal circuitry and improves radial-maze learning in adult mice. J. Neurosci., 11(7): 2102–2106. Slomianka, L. and Geneser, F.A. (1997) Postnatal development of zinc-containing cells and neuropil in the hippocampal region of the mouse. Hippocampus, 7(3): 321–340. Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E., Goodman, J.H., Sollas, A.L. and Martin, D.L. (1996) Basalin expression and duction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal dentate gyrus. J. Comp. Neurol., 373(4): 593–618. Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M. (2006) Kainic acid-induced recurrent mossy fiber innervation

of dentate gyrus inhibitory interneurons: possible anatomical substrate of granule cell hyper-inhibition in chronically epileptic rats. J. Comp. Neurol., 494(6): 944–960. Song, H., Stevens, C.F. and Gage, F.H. (2002) Neural stem cells from adult hippocampus develop essential properties of functional CNS neurons. Nat. Neurosci., 5(5): 438–445. Soriano, E., Cobas, A., Lopez, C. and Berbel, P. (1983) Morphology of mossy fiber collaterals in the hilus of the rat. Neurosci. Lett. Suppl., 14: S352. Stanfield, B.B. and Trice, J.E. (1988) Evidence that granule cells generated in the dentate gyrus of adult rats extend axonal projections. Exp. Brain Res., 72(2): 399–406. Stengaard-Pedersen, K., Fredens, K. and Larsson, L.I. (1983) Comparative localization of enkephalin and cholecystokinin immunoreactivities and heavy metals in the hippocampus. Brain Res., 273(1): 81–96. Stirling, R.V. and Bliss, T.V. (1978) Hippocampal mossy fiber development at the ultrastructural level. Prog. Brain Res., 48: 191–198. Storm-Mathisen, J. (1981) Glutamate in hippocampal pathways. Adv. Biochem. Psychopharmacol., 27: 43–55. Sunde, N., Laurberg, S. and Zimmer, J. (1984) Brain grafts can restore irradiation-damaged neuronal connections in newborn rats. Nature, 310(5972): 51–53. Sunde, N. and Zimmer, J. (1981) Transplantation of central nervous tissue. An introduction with results and implications. Acta Neurol. Scand., 63(5): 323–335. Sunde, N.A. and Zimmer, J. (1983) Cellular, histochemical and connective organization of the hippocampus and fascia dentata transplanted to different regions of immature and adult rat brains. Brain Res., 284(2–3): 165–191. Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L. (1989) Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol., 26(3): 321–330. Sutula, T., He, X.X., Cavazos, J. and Scott, G. (1988) Synaptic reorganization in the hippocampus induced by abnormal functional activity. Science, 239(4844): 1147–1150. Swanson, L.W., Wyss, J.M. and Cowan, W.M. (1978) An autoradiographic study of the organization of intrahippocampal association pathways in the rat. J. Comp. Neurol., 181(4): 681–715. Tamamaki, N., Abe, K. and Nojyo, Y. (1988) Three-dimensional analysis of the whole axonal arbors originating from single CA2 pyramidal neurons in the rat hippocampus with the aid of a computer graphic technique. Brain Res., 452(1–2): 255–272. Tamamaki, N. and Nojyo, Y. (1991) Crossing fiber arrays in the rat hippocampus as demonstrated by three-dimensional reconstruction. J. Comp. Neurol., 303(3): 435–442. Terrian, D.M., Hernandez, P.G., Rea, M.A. and Peters, R.I. (1989) ATP release, adenosine formation, and modulation of dynorphin and glutamic acid release by adenosine analogues in rat hippocampal mossy fiber synaptosomes. J. Neurochem., 53(5): 1390–1399. Thom, M., Martinian, L., Williams, G., Stoeber, K. and Sisodiya, S.M. (2005) Cell proliferation and granule cell

107 dispersion in human hippocampal sclerosis. J. Neuropathol. Exp. Neurol., 64(3): 194–201. Timm, F. (1958) Histochemistry of the region of Ammon’s horn. Z. Zellforsch. Mikrosk. Anat., 48(5): 548–555. Tønder, N., Sørensen, T. and Zimmer, J. (1990) Grafting of fetal CA3 neurons to excitotoxic, axon-sparing lesions of the hippocampal CA3 area in adult rats. Prog. Brain Res., 83: 391–409. Vaughn, J.E., Matthews, D.A., Barber, R.P., Wimer, C.C. and Wimer, R.E. (1977) Genetically-associated variations in the development of hippocampal pyramidal neurons may produce differences in mossy fiber connectivity. J. Comp. Neurol., 173(1): 41–52. Vida, I. and Frotscher, M. (2000) A hippocampal interneuron associated with the mossy fiber system. Proc. Natl. Acad. Sci. U.S.A., 97(3): 1275–1280. West, J.R., Van Hoesen, G.W. and Kosel, K.C. (1982) A demonstration of hippocampal mossy fiber axon morphology using the anterograde transport of horseradish peroxidase. Exp. Brain Res., 48(2): 209–216. West, M.J., Gaarskjaer, F.B. and Danscher, G. (1984) The Timm-stained hippocampus of the European hedgehog: a basal mammalian form. J. Comp. Neurol., 226(4): 477–488. West, M.J. and Gundersen, H.J. (1990) Unbiased stereological estimation of the number of neurons in the human hippocampus. J. Comp. Neurol., 296(1): 1–22. West, M.J., Slomianka, L. and Gundersen, H.J. (1991) Unbiased stereological estimation of the total number of

neurons in the subdivisions of the rat hippocampus using the optical fractionator. Anat. Rec., 231(4): 482–497. Zimmer, J. (1973) Changes in the Timm sulfide silver staining pattern of the rat hippocampus and fascia dentata following early postnatal deafferentation. Brain Res., 64: 313–326. Zimmer, J. (1974) Long term synaptic reorganization in rat fascia dentata deafferented at adolescent and adult stages: observations with the Timm method. Brain Res., 76(2): 336–342. Zimmer, J., Finsen, B., Sorensen, T. and Poulsen, P.H. (1988a) Xenografts of mouse hippocampal tissue. Exchange of laminar and neuropeptide specific nerve connections with the host rat brain. Brain Res. Bull., 3: 369–379. Zimmer, J., Finsen, B., Sorensen, T. and Poulsen, P.H. (1988b) Xenografts of mouse hippocampal tissue. Formation of nerve connections between the graft fascia dentata and the host rat brain. Prog. Brain Res., 78: 271–280. Zimmer, J. and Gahwiler, B.H. (1984) Cellular and connective organization of slice cultures of the rat hippocampus and fascia dentata. J. Comp. Neurol., 228(3): 432–446. Zimmer, J. and Gahwiler, B.H. (1987) Growth of hippocampal mossy fibers: a lesion and coculture study of organotypic slice cultures. J. Comp. Neurol., 264(1): 1–13. Zimmer, J. and Haug, F.M. (1978) Laminar differentiation of the hippocampus, fascia dentata and subiculum in developing rats, observed with the Timm sulphide silver method. J. Comp. Neurol., 179(3): 581–617.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 6

Mossy fiber synaptic transmission: communication from the dentate gyrus to area CA3 $

David B. Jaffe1, and Rafael Gutie´rrez2 1 Department of Biology, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249, USA Department of Physiology, Biophysics and Neurosciences, Center for Research and Advanced Studies, Mexico City, Apartado Postal 14-740, Mexico D.F. 07000, Mexico

2

Abstract: Communication between the dentate gyrus (DG) and area CA3 of the hippocampus proper is transmitted via axons of granule cells — the mossy fiber (MF) pathway. In this review we discuss and compare the properties of transmitter release from the MFs onto pyramidal neurons and interneurons. An examination of the anatomical connectivity from DG to CA3 reveals a surprising interplay between excitation and inhibition for this circuit. In this respect it is particularly relevant that the major targets of the MFs are interneurons and that the consequence of MF input into CA3 may be inhibitory or excitatory, conditionally dependent on the frequency of input and modulatory regulation. This is further complicated by the properties of transmitter release from the MFs where a large number of co-localized transmitters, including GABAergic inhibitory transmitter release, and the effects of presynaptic modulation finely tune transmitter release. A picture emerges that extends beyond the hypothesis that the MFs are simply ‘‘detonators’’ of CA3 pyramidal neurons; the properties of synaptic information flow from the DG have more subtle and complex influences on the CA3 network. Keywords: mossy fibers; synaptic transmission; CA3; co-transmitters; plasticity

originally) and Lorente de No both suggested that the large varicosities made synaptic contacts onto CA3 pyramidal neurons (Ramon y Cajal, 1911; Lorente de No, 1934), later to be confirmed with electron microscopy (Blackstad and Kjaerheim, 1961), thereby making this pathway an integral component within the so-called, and overly simplistic, ‘‘tri-synaptic’’ circuit. In this review we will discuss and compare the properties of transmitter release at the most wellcharacterized MF synapses, those in area CA3 onto pyramidal neurons and interneurons. Our discussion will focus on the potential roles of co-localized transmitters and modulators as well as how the properties of synaptic transmission

Neural output from the dentate gyrus (DG) is transmitted via axons of granule cells called the mossy fiber (MF) pathway. The MFs not only project into area CA3 of the hippocampus, but also synapse proximally onto DG basket cells (providing local recurrent inhibition in the dentate) and pyramidal-like neurons in the hilus (Johnston and Amaral, 2004). They are so named because of the large (up to 5 mm diameter) varicosities along the axon. Ramon y Cajal (who named the pathway

$

Both authors contributed equally to this work.

Corresponding author. Tel.: +1 210 458 5843;

Fax: +1 210 458 5658; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63006-4

109

110

from the DG influence the primary target of DG output — the CA3 region of the hippocampus.

MF anatomy: does form follow function? The MFs form three types of synaptic contacts onto its target neurons. First, and most notably, are the large expansions that synapse onto CA3 pyramidal neurons. The large boutons appear at approximately 150 mm intervals (Blackstad et al., 1970) and a single granule cell contacts approximately 15 pyramidal neurons (each terminal synapses onto a single pyramidal neuron). One CA3 pyramidal neuron may receive up to a total of approximately 50 MF inputs only (Claiborne et al., 1986; Amaral et al., 1990). Second, from the large expansions there may extend 2–3 filopodia that make synaptic contacts onto interneurons (Amaral, 1979; Acsady et al., 1998). These so-called filopodia are motile and regulated by glutamatergic neuromodulation (Tashiro et al., 2003). Third, small terminals, resembling boutons of other hippocampal neurons, also contact CA3 interneurons. As a result, it appears that by sheer numbers the major target of the MFs in CA3 is

onto inhibitory interneurons, rather than pyramidal neurons (Fig. 1). The large terminals of the MFs are unique structures in a variety of ways. In addition to the filopodial-like synaptic contacts onto interneurons, mentioned above, the large boutons are filled with a high density of vesicles (Blackstad and Kjaerheim, 1961), express more than one active zone, and encase a complex branched dendritic spine emanating from CA3 pyramidal neurons — generally referred to as a thorny excrescence (Amaral and Dent, 1981; Chicurel and Harris, 1992). Putative synaptic sites on these boutons were originally identified as both asymmetric and symmetric, suggesting differences in their respective properties of transmission (Hamlyn, 1961). Another important anatomical aspect of the MF pathway is that they are generally restricted to a narrow band running within stratum lucidum where the large terminals make contacts onto the most proximal portions of pyramidal neuron apical dendrites (Brown and Johnston, 1983). Additionally, there is a less extensive infrapyramidal MF projection onto the proximal basal dendrites of pyramidal neurons in CA3b and CA3c. The first recordings of MF synaptic transmission were extracellular field potential recordings

Fig. 1. Interneurons are the primary target of DG granule cells. Diagram representing the multiple types of excitatory contacts made by granule cells. Within the dentate/hilar region collaterals of the mossy fibers (MFs) innervate both mossy cells and inhibitory basket cells. In area CA3, the MFs contact pyramidal neurons via large expansions, but also excite interneurons through either filopodial extensions from the large boutons or smaller en passant expansions along MF axons. (See Color Plate 6.1 in color plate section.)

111

made in vivo. Stimulation of the DG triggered a current sink restricted to s. lucidum (Gloor et al., 1963), confirming the anatomical evidence that MF input was restricted to a discrete layer, and supporting the view that the synapse was excitatory. Two conflicting ideas emerge out of the anatomical connectivity from the DG to area CA3, discussed above. On the one hand, the proximal location of the MF large boutons, and the fact that their large terminals are copiously supplied with vesicles, suggests that a single action potential from a granule cell might have a strong excitatory influence on CA3 pyramidal neurons. That is, a single spike invading a single terminal might be capable of triggering an action potential in a CA3 pyramidal neuron. This is the so-called ‘‘detonator synapse’’ hypothesis. The circuitry of area CA3 resembles that of an auto-associative network (Marr, 1971) reflected in the large number of recurrent excitatory connections between CA3 pyramidal neurons (Johnston and Amaral, 2004). The concept of the MFs as detonators led to the hypothesis that these sparse inputs might serve as a ‘‘teacher’’ signal underlying forms of associative learning (McNaughton and Morris, 1987; Rolls, 1989). The presence of a large supply of vesicles suggests a large readily releasable pool (Hallermann et al., 2003), and that the degree of transmitter release could be maintained over time by the large reserve pool of vesicles. The proximal location of these synapses relative to the spike generating zone also would limit any loss of depolarization due to cable filtering (Johnston and Brown, 1983). That individual MF synapses would have large unitary synaptic strength would also be a requirement if the pathway as a whole were to be effective at driving the postsynaptic cell, because — as discussed above — each CA3 pyramidal neuron receives only about 50 mossy inputs (Amaral et al., 1990). On the other hand, the MFs also make direct synaptic contact onto inhibitory interneurons in area CA3 (Fig. 1). Moreover, the ratio of synaptic contacts onto interneurons is approximately fourfold higher than onto pyramidal neurons. Assuming that (i) a single granule cell contacts 15 pyramidal neurons, (ii) each large expansion

contacts directly or via its filopodia up to three interneurons, and (iii) 15 interneurons are innervated by the smaller boutons one-granule cell will contact approximately 60 interneurons compared with only 15 pyramidal neurons (a 4:1 ratio). Therefore, it is conceivable that under certain conditions the firing of granule cells would have a net inhibitory effect on pyramidal neurons in CA3 (Bragin et al., 1995a, b; Penttonen et al., 1997; Acsady et al., 1998).

Excitatory–inhibitory conductance sequence Yamamoto (1972) was the first to utilize brain slice methods and intracellular recording to study MF synaptic transmission onto CA3 pyramidal neurons. Stimulation of the granule cell layer triggered a biphasic response composed of a small compound excitatory postsynaptic potential (EPSP) followed by a larger, overlapping compound inhibitory postsynaptic potential (IPSP), presumably mediated either by feed-forward/feed-back inhibition or the direct stimulation of inhibitory interneurons (Yamamoto, 1972). The IPSP itself was biphasic, comprised of an initial GABAA receptor and later GABAB receptor-mediated phase (Ogata and Ueno, 1976; Knowles et al., 1984). One of Yamamoto’s most important early observations was that paired stimulation of the fibers at short intervals (20–100 ms) markedly potentiated the second EPSP. While low-frequency stimulation (LFS) rarely triggered spikes, frequency facilitation of the compound EPSP was generally capable of reaching spike threshold, in spite of the overlapping IPSPs. Frequency facilitation occurred over a very wide range of frequencies of stimulation (0.05–1 Hz), and could potentiate responses as much as threefold (Salin et al., 1996). Johnston and Brown (1983) applied voltageclamp methods to the study of MF synaptic transmission, recognizing that the advantageous proximal location of MFs relative to the soma permitted reasonable voltage-control and, in turn, space clamp issues were minimized (Johnston and Brown, 1983; Spruston and Johnston, 1992; Spruston et al., 1993). Measuring synaptic currents

112

under voltage-clamp also allowed better resolution of the mixed excitatory–inhibitory conductance sequence, even though it was recognized that there was still considerable overlap between the excitation and inhibition (Brown and Johnston, 1983; Barrionuevo et al., 1986). Two important observations were made in these early studies. First, for compound postsynaptic currents, the GABAAmediated inhibitory conductance was almost fivefold larger than the excitatory conductance. Second, the onset of the inhibitory conductance appeared to be very close to that of the excitatory response. A disynaptic GABAergic response, from a feed-forward or feed-back circuit, should retard the onset of the inhibitory response relative to the excitatory conductance. It is therefore possible that the paradigms used to stimulate the MF pathway might directly trigger the release of GABA, most likely from inhibitory neurons, onto CA3 pyramidal neurons.

Properties of transmitter release from MF synapses Granule cells discharge action potentials down the MFs at basal rates less than 0.5 Hz (Jung and McNaughton, 1993), though firing rates may reach up to 50 Hz during certain types of behaviors (Skaggs et al., 1996; Wiebe and Staubli, 1999; Henze et al., 2002b) and conduction velocity is approximately 7 m/s, consistent with the MFs being an unmyelinated pathway (Langdon et al., 1993). Upon reaching synaptic terminals, presynaptic action potentials trigger synaptic transmission by eliciting Ca2+ influx through multiple types of voltage-gated Ca2+ channels. Blockade of N-, P-, and R-type Ca2+ channels decreases fast transmitter release, while dihydropyridine-sensitive L-type Ca2+ channels do not (Kamiya et al., 1988; Castillo et al., 1994; Nicoll et al., 1994; Yamamoto et al., 1994; Wu et al., 1998; Miyazaki et al., 2005). That said, L-type channels are present in MF presynaptic terminals and allow Ca2+ entry when MFs are stimulated (Tokunaga et al., 2004). Although L-channels do not appear to participate in fast transmitter release, they may play a role in transmission during repetitive stimulation (Reuter, 1995).

To directly study presynaptic action potentials and their interaction with voltage-gated ion channels, patch-clamp methods have recently been applied to the large boutons of the MFs (Bischofberger et al., 2006). One of the first notable observations was that the action potential, while of short duration during LFS, widens during high-frequency stimulation (HFS) due to the rapid inactivation of a voltage-gated K+ conductance (Geiger and Jonas, 2000). Frequency-dependent spike broadening, and the associated increase in Ca2+ influx, provides for, in part, a mechanism underlying frequency-facilitation (Salin et al., 1996). The density of Na+ channels in the boutons appears relatively large, and may be important for the efficient propagation of action potentials (Engel and Jonas, 2005). The load resulting from the greater surface area of the large boutons could potentially extinguish or slow down spike propagation in these unmyelinated fibers. Computer simulations suggested that the high density and properties of these Na+ channels ensures the reliability of synaptic transmission. Ca2+ in the presynaptic terminal is also affected by other mechanisms. Intracellular Ca2+ stores regulate presynaptic Ca2+ concentration (Liang et al., 2002). Ca2+-induced Ca2+ release may work in concert with spike-mediated Ca2+ influx to raise intraterminal Ca2+ concentration. Interestingly, although Ca2+-induced Ca2+ release accounts for a large proportion of the presynaptic signal (Scott and Rusakov, 2006), depletion of Ca2+ stores has no effect on fast transmitter release. Release of Ca2+ may, however, play a role in frequency facilitation (Lauri et al., 2003). In a recent study, it was reported that EPSPs in dentate granule cells, surprisingly, propagated down hundreds of microns along the MFs and were capable of modulating synaptic transmission (Alle and Geiger, 2006). Furthermore, spikeevoked EPSCs in CA3 pyramidal neurons were enhanced when paired with a presynaptic waveform that mimicked an EPSP in the bouton. The mechanism underling this type of modulation may be related to a similar phenomenon observed at the Calyx of Held (Awatramani et al., 2005), but other studies suggest it may also be due to

113

voltage and not solely due to resting Ca2+ levels (Ruiz et al., 2003).

Quantal nature of transmission at the MF-CA3 pyramidal neuron synapse The strong frequency-dependent facilitation of MF synaptic transmission onto CA3 pyramidal neurons implies that the initial probability of transmitter release is small. If the probability of release were high, either a ceiling effect would limit an increase in release (assuming that there is a large reserve pool of primed vesicles) or, alternatively, one could observe synaptic depression due to the refractory period produced by a lack of primed vesicles. Indeed, quantal analysis of unitary synaptic responses onto CA3 pyramidal neurons is consistent with the hypothesis that the initial release probability is in fact low (von Kitzing et al., 1994; Lawrence et al., 2004). This is in spite of the fact that the number of release sites at a single MF synapse is very high (Chicurel and Harris, 1992), and that release at these synapses is most likely multiquantal (Henze et al., 1997). Indeed, in the study by Jonas et al. (1993) quantal content (the number of quanta released per action potential) ranged from 2–16, with a mean of 8, values similar to that first reported by Yamamoto (Yamamoto et al., 1991). Others have found that the probability of release was much lower, o0.3 (Lawrence et al., 2004). There is also evidence, however, that large synaptic currents arise from the release of single quanta (Henze et al., 2002a). This is in contrast to commissural/associational (C/A) synapses, where the probability of release is higher and, concomitantly, frequency facilitation is weaker (Salin et al., 1996). Mean quantal size at MF-CA3 pyramidal neuron synapses is approximately 150 pS, corresponding to about 17 AMPA receptors. These numbers are consistent with unitary responses (unitary conductance ¼ quantal size
quantal content), which generally have magnitudes of approximately 1 nS. Corresponding unitary EPSP amplitudes can be up to 10 mV (Jonas et al., 1993), in contrast to C/A unitary responses that are typically o1 mV (Debanne et al., 1998; Pavlidis

and Madison, 1999). Similarly large unitary EPSPs have been demonstrated for MF-mossy cell synapses (Scharfman et al., 1990).

Evidence for multiple neurotransmitters/modulators in the granule cells Glutamate is believed to be the primary excitatory neurotransmitter released from the MFs (Crawford and Connor, 1973; Terrian et al., 1988), and MF EPSPs are blocked by glutamate receptor antagonists (Sawada et al., 1983). The neuronal glutamate transporter (EAAC1) is the most abundant uptake mechanism and is selectively enriched in hippocampal principal neurons, including DG granule cells (Rothstein et al., 1994). The presynaptic and postsynaptic actions of glutamate released from the MFs are discussed below. Granule cells also contain and release several neuromodulators of different chemical composition (Fig. 2). The most prominent are those of peptidic nature, which are contained and released from large dense-core vesicles. Dynorphin and enkephalin, as well as their receptors, are present in the MFs in rodents and humans, but there is variation between mammalian species (Gall et al., 1981; McGinty et al., 1983; Chavkin et al., 1985; McLean et al., 1987; Terrian et al., 1988; Houser et al., 1990; Chavkin, 2000). The actions of opioid peptides in the hippocampus are inhibitory, mediated by inhibition of Ca++ currents, in the case of dynorphin, or activation of K+ currents, in the case of enkephalins, and are a consequence of activation of G-protein-coupled receptors (Zieglga¨nsberger et al., 1979). In the rat MF system, activation of m receptors facilitates MF LTP in an indirect fashion. Activation of m receptors depresses GABA release from interneurons, which in turn leads to a failure of GABA to inhibit glutamate release from MFs (Jin and Chavkin, 1999). On the other hand, the m opioid receptor agonist DAMGO inhibits low-frequency stimulated MF responses in Sprague-Dawley rats, as it does in the guinea pig (Salin et al., 1995). Direct actions of dynorphin on MF transmission have been described in the guinea pig, but not in the rat. Indeed the MFs possess k receptors, which, upon

114

Fig. 2. Summary of the pre- and post-synaptic constituents of the different MF terminals and of their plasticity. (A) Schematic representation of the giant MF boutons, which contact CA3 pyramidal cells. These terminals are characterized by low probability of basal release and multiple release sites. (B) Schematic representation of filopodial extensions and en passant contacts, which synapse on to interneurons. These have a high probability of basal release and single release sites. Both types of MF terminals contain several neuromodulators and the neurotransmitters glutamate and GABA. Note that both the presynaptic and postsynaptic sites contain several ionotropic GluRs and metabotropic GluRs, which confer to the MF a high degree of plasticity and the capacity for synaptic integration. (C) Relative expression of the different releasable contents and of the receptors and transporters during development, in the adult (D), and after epileptic activity (E). Arrows and font size indicate relative differences between the three states. (See Color Plate 6.2 in color plate section.)

activation, depress neurotransmitter release. Thus, for example, in the guinea pig high frequency stimulation of the MF causes a transient heterosynaptic inhibition of neighboring MF or perforant path synapses in the dentate, which is mediated by the synaptic release of dynorphin that activates presynaptic k receptors (Weisskopf et al., 1993).

However, in the rat, neither exogenous nor endogenous dynorphin affect MF neurotransmission, which is consistent with the finding that k-receptor binding in this projection is dense in the guinea pig and sparse in the Sprague-Dawley rat (see Salin et al., 1995). Dynorphin significantly inhibits MF responses in the hamster, mouse, and

115

in Long-Evans rats (Salin et al., 1995). Interestingly, hippocampal opioid peptides are upregulated after seizure activity (McGinty et al., 1983; Gall, 1988; Gall et al., 1990), whereas a decrease in k receptors has been observed in epileptic human cells (Jeub et al., 1999). Thus, opioid peptides are localized in the MF, presumably for effective control of neurotransmitter release (Weisskopf et al., 1993; Castillo et al., 1996; Drake et al., 1996; Simmons and Chavkin, 1996). Other peptides present in the MF are neuropeptide Y (NPY), neurokinin B and cholecystokinin (Gall, 1984; Gall et al., 1990; Holm et al., 1993; Tonder et al., 1994; Chandy et al., 1995; Schwarzer and Sperk, 1995; Makiura et al., 1999). Like opioid peptides, these peptides have been shown to exert inhibitory actions. Neuropeptide Y is synthesized, stored in, and released from the MFs, and inhibits MF transmission by a presynaptic mechanism (McQuiston and Colmers, 1996; McCarthy et al., 1998) and their receptors are normally present in the MFs themselves (Widdowson, 1993; Jacques et al., 1997). Seizures increase the release of NPY which in turn, tonically inhibits MF synaptic transmission (Tu et al., 2005). Besides, seizures upregulate NPY Y2 receptors, which appear to mediate the effects on MF (Vezzani and Sperk, 2004), although Y5 receptors could play a role (Marsh et al., 1999; Ho et al., 2000). This upregulation reflects a neuroprotective function of NPY that has been shown in animal models of epilepsy and in the epileptic human (Schwarzer et al., 1998; Patrylo et al., 1999; Furtinger et al., 2001; Vezzani and Sperk, 2004). By contrast, substance P, which is contained in and released from the MFs activates its receptors in the same MF, increases glutamate release (Borhegyi and Leranth, 1997; Liu et al., 1999). The MFs also contain high levels of Zn++, which can be released together with glutamate in an activity-dependent manner (StengaardPedersen et al., 1981; Wenzel et al., 1997; Molnar and Nadler, 2001; Qian and Noebels, 2005). Exogenously applied Zn++ blocks glutamate and GABA receptors (Westbrook and Mayer, 1987; Draguhn et al., 1990; Smart et al., 1994). Indeed, Zn++ occupies a high-affinity binding site on NMDA receptors (Vogt et al., 2000) and can

block GABAA receptors composed by a and b but not g subunits (Draguhn et al., 1990). It has been shown that synaptically released Zn++ from the MF pathway strongly modulates NMDA (Vogt et al., 2000) and GABAA receptors in CA3 (Ruiz et al., 2004). Because MF release, besides glutamate, GABA and Zn++ (Gutierrez, 2005), this synapse is most suitable to study the effects of the synaptically released Zn++ on GABA responses elicited by the same terminals. Indeed, it was found that Zn++ tonically depresses the inhibitory actions of GABA, as it reaches a high concentration in the vicinity of the GABAA receptor on CA3 pyramidal cells. Consequently, chelation of Zn++ relieves this inhibition (Ruiz et al., 2004). This is particularly important for developmental and pathological processes, in particular, epilepsy. Indeed, a higher input of Zn++ from sprouted MFs after epileptic activity contributes to inhibit reverberating excitatory activity in the DG (Nadler, 2003; Tu et al., 2005). Other interesting signaling molecules present in the MF are brain derived neurotrophic factor (BDNF) and nerve growth factor (NGF) (Gall and Isackson, 1989; Gall and Lauterborn, 1992; Lowenstein and Arsenault, 1996; Conner et al., 1997; Smith et al., 1997). Granule cells contain relatively high concentrations of BDNF and NGF, which can be released in vivo by the MFs and dendrites to affect other MFs or other cells, respectively (Blo¨chl and Thoenen, 1995; Lowenstein and Arsenault, 1996; Conner et al., 1997). Dendritic targeting of BDNF mRNA and its translation to protein can account for the release of this neurotrophin to neighboring granule cell dendrites (Tongiorgi et al., 2004). This suggests that the ligand–receptor interaction occurs by means of an autocrine/paracrine mechanism. On the other hand, anterograde transport of BDNF (Conner et al., 1997) would effectively provide a mechanism to release BDNF from MF terminals, where it can increase neurotransmitter release (Altar and DiStefano, 1998; Elmer et al., 1998). As with most signaling molecules in the granule cells, BDNF expression is increased after limbic seizures (Isackson et al., 1991). In the MF, a consequence of this would be the activation of TrkB receptors located in the same fibers (He et al., 2002; Scharfman,

116

2005). TrkB activation is thought to promote Ca++ influx into terminals (Berninger et al., 1993) and this is the mechanism that leads to enhanced neurotransmitter release. Eventually, this produces an enhancement of excitability that can lead to the generation of epileptiform activity (Thoenen, 1995; Huang and Reichardt, 2001; Scharfman et al., 2002). Importantly, it has been shown that BDNF can exert a rapid depolarization in hippocampal neurons by promoting Na+ influx through TTXinsensitive channels (Kafitz et al., 1999; Rose et al., 2003). Another important issue is that BDNF may be packaged preferentially in the large MF boutons, although it does reside in the small terminals also, so it can potentially modulate both excitation and feed-forward inhibition in CA3 (Danzer and McNamara, 2004). Besides a possible involvement of hippocampal BDNF in excitability and synaptic plasticity (Thoenen, 1995, 2000), neurotrophins have a differentiating effect on interneurons in the hippocampus (Marty et al., 1996a, b) and cortex (Marty et al., 1997; Patz et al., 2003), by regulating the expression of GABAergic markers. Neuronal activity is the main activator of GAD expression by neurotrophins, differentially modulating transcription and translation in a context-dependent manner (Patz et al., 2003). Interestingly, this is the mechanism shown to underlie the maturation of the GABAergic phenotype in hippocampal interneurons (Marty et al., 1996a, b) and, as recently shown, the expression of all GABAergic markers in granule cells (Gomez-Lira et al., 2005). Granule cells also contain high levels of adenosine (Braas et al., 1986) and its A1 receptor (Rivkees et al., 1995; Swanson et al., 1995). It has been suggested that endogenous tonic activation of A1 receptors underlies the low basal probability of neurotransmitter release from the MFs (Moore et al., 2003; also see Kukley et al., 2005). Their activation produces a Gi/o protein-dependent sustained inhibition of neurotransmitter release (Moore et al., 2003). Other set of receptors present in the MF terminals, and which activation controls neurotransmitter release, are the ATP receptors P2X (Armstrong et al., 2002) and a adrenergic receptors (Scanziani et al., 1993).

Finally, the inhibitory amino acid GABA is probably the most important and controversial addition to the list of modulators and transmitters that are co-localized in the granule cells. Sandler and Smith (1991) provided the first data that suggested the presence GABA in the ‘‘glutamatergic’’ MF in monkey and human hippocampi (Sandler and Smith, 1991). They found GABA immunoreactivity in MF terminals that made asymmetric synaptic contacts with spines arising from large dendrites of CA3 pyramidal cells. They also showed the colocalization of GABA and glutamate within the same terminals with electron microscopy. From this anatomical evidence, and the idea that GAD was not present in the MF, they concluded that GABA had to be either incorporated from the extracellular space or originated from an alternative route of synthesis such as g-hydroxybutyrate. Second, they suggested that at least a component of the inhibitory synaptic potentials evoked in pyramidal cells by DG stimulation had to be of MF origin. In this way, GABA released by the MF could modulate the normal MF glutamatergic responses. On the other hand, studies carried out on MF synaptosomes provided the first neurochemical evidence showing that MF terminals contained and released GABA (Terrian et al., 1988; Taupin et al., 1994a, b). These authors, in agreement with Sandler and Smith (1991), suggested that the amino acid could be synthesized in the granule cells from a route different from the GADdependent pathway. Evidence clarifying the presence and origin of GABA in the MF was provided by Sloviter et al. (1996). These authors conclusively demonstrated that immunoreactivity for the amino acid and its synthesizing enzyme, GAD, was normally present in the MFs of rats, monkeys, and humans (Sloviter et al., 1996). Therefore, if the granule cells had the necessary enzyme for the synthesis of GABA and GABA itself, the granule cells indeed synthesized GABA that could probably be used as a neurotransmitter. Complementing and extending the initial findings of Sandler and Smith (1991) and Sloviter et al. (1996), Bergersen et al. (2003) recently confirmed with immunogold the coexistence of glutamate and GABA in MF synapses,

117

which also contained the respective receptors in apposition to the presynaptic terminal (Bergersen et al., 2003). This study demonstrated that both amino acids coexist in all MF terminals examined and that they have a close spatial relation to synaptic vesicles. Indeed, both GABA and glutamate were shown to be located at a distance that suggests their presence inside vesicles and in the release zones. The estimated concentration of GABA, however, was much lower than that of glutamate within the MF terminals and even lower than that of inhibitory types of terminals (Bergersen et al., 2003). Interestingly, it was demonstrated that provoking seizures by stimulation of the perforant pathway for 24 upregulated the content of both isoforms of GAD and GABA in the MFs, whereas no changes were observed in area CA1 (Sloviter et al., 1996). Other authors (Schwarzer and Sperk, 1995; Lehmann et al., 1996) showed that GAD67 and its mRNA (but not GAD65) were transiently upregulated after seizures provoked by kainic acid (KA) or by the kindling method in the rat. On the other hand, the upregulation of GAD67 and its mRNA in granule cells was further confirmed with other seizure- and epilepsy-induction methods that use chemical convulsants or electrical stimulation (Ding et al., 1998; Makiura et al., 1999; Szabo et al., 2000; Ramirez and Gutierrez, 2001). It was also established that GAD67 could be upregulated in an activity-dependent manner, in the absence of epileptiform activity (Ramirez and Gutierrez, 2001; Gutierrez, 2002). Recently, a direct link between the presence of seizures and the concomitant upregulation of GAD and endogenous GABA was found in MF terminals of epileptic rats (GomezLira et al., 2002). An analysis of the expression of GAD67 at different ages has revealed that GAD67 (but not GAD65 or GAD65 RNA) is also developmentally regulated in the MFs (Maqueda et al., 2003). Indeed, it was shown that GAD67 is expressed in the MFs early in life and then downregulated by days 23–24, after completion of development (Gutierrez, 2003; Maqueda et al., 2003). Likewise, GABA-immunoreactive cells with characteristics of granule cells are found in the stratum granulare of the DG of developing rats but not of adults.

Contrary to the data showing the upregulation of both isoforms of GAD by seizures (Sloviter et al., 1996), several reports have shown that GAD67 (and its mRNA) and not GAD65 is regulated in granule cells by increased activity (Schwarzer and Sperk, 1995; Szabo et al., 2000; Ramirez and Gutierrez, 2001; Maqueda et al., 2003), Ca++ entry and BDNF activation (GomezLira et al., 2005). It has been proposed that the cellular distribution of both enzymes differed and this could be reflected in distinctive functions within neurons (Erlander and Tobin, 1991), i.e., GAD65 may be in synaptic terminals, while GAD67 is present in terminals, somata and dendrites (Kaufman et al., 1991). This suggests that one isoform synthesizes a metabolic pool of GABA (in the soma) and the other the releasable pool (in the terminals). Despite the possible differences in GAD65- and GAD67-originated GABA, there is a strong correspondence of the expression of GAD67 to that of GABA present within the granule cells and their MFs (Gomez-Lira et al., 2002). All these studies suggest that GABA synthesis in the MFs is a means to counteract the enhanced excitability caused by epileptic activity.

Presynaptic modulation As mentioned above, a number of neuromodulators control transmitter release from the MFs. Glutamate and GABA are also important presynaptic modulators of MF synaptic transmission transmission. In particular, one of the unique properties of MF synaptic transmission, in contrast to transmitter release from recurrent synapses onto CA3 pyramidal neurons, is its sensitivity to metabotropic glutamate receptor (mGluR) agonists; mGluR agonists depress MF synaptic transmission (Manzoni et al., 1995; Yoshino et al., 1996). Because of this response, sensitivity to mGluR agonists is a widely used assay to determine if a synaptic response is of MF origin. There are species-specific differences with respect to presynaptic mGluRs. MF glutamatergic neurotransmission in the rat is sensitive to the group II mGluR agonist, DCG-IV, but not to the group III mGluR agonist, L-AP4 (Lanthorn et al., 1984;

118

Kamiya et al., 1996; Maccaferri et al., 1998). The opposite is true for the guinea pig where L-AP4 strongly depresses MF transmission (Lanthorn et al., 1984; Manzoni et al., 1995; Tong et al., 1996; Min et al., 1998). Interestingly, the granule cells and MF of the rat DG express groups II/III mGluR mRNA (Ohishi et al., 1993; Ohishi et al., 1995) and mGluR2,4,7 (Bradley et al., 1996; Shigemoto et al., 1997; Lie et al., 2000). Electron microscopy has revealed immunolabeling for the group III mGluR predominantly in presynaptic active zones of asymmetrical and symmetrical synapses, whereas mGluR-II immunolabeling was found in preterminal rather than terminal portions of axons (Shigemoto et al., 1996, 1997). Although it has been well established that activation of group II mGluR almost completely depresses MF glutamatergic transmission in the rat, several reports show that MF GABAergic transmission is strongly inhibited through activation of group III mGluR (with L-AP4) both in the guinea pig and the rat (Gutierrez, 2000; Walker et al., 2001; Gutierrez, 2002, 2003; Romo-Parra et al., 2003; Kasyanov et al., 2004; Safiulina et al., 2006). This pharmacological profile is not only consistent with neurotransmission of MF origin but it also demonstrates that the modulation of MF GABAergic transmission differs from MF glutamatergic transmission. This physiological evidence gives a functional significance to the anatomical findings showing that both types of receptors are present in rat granule cells with a distinct localization and function (Ohishi et al., 1995; Shigemoto et al., 1996, 1997). These data, together with the finding that activation of mGluR produces a downregulation of the exocytotic machinery (Kamiya and Ozawa, 1999), suggest that group III mGluR are associated with mechanisms of GABA release, and also that there is presynaptic segregation of mGluR receptors according to the class of neurotransmitter to be released. Kainic acid has long been used as an epileptogenic agent and was recognized that it binds to the MF with high affinity (Monaghan and Cotman, 1982; Represa et al., 1987). It is now clear that the several KA receptors that are present in the MF (Wisden and Seeburg, 1993; Darstein et al., 2003)

modulate plasticity at this synapse in a complex manner (Schmitz et al., 2000). Presynaptic GABAB receptors effectively inhibit glutamate release from the MFs (Thompson and Gahwiler, 1992), and it has recently shown that ionotropic GABAA receptors also inhibit glutamate (Ruiz et al., 2003) and GABA release (Trevin˜o and Gutierrez, 2005) and control the excitability of this pathway (Kullmann et al., 2005; Trevin˜o and Gutierrez, 2005). Other subcortical modulators also control the release of transmitter from the MFs. Muscarinic acetylcholine receptors inhibit transmitter release (Williams and Johnston, 1988, 1990) while nicotinic receptor activation raises presynaptic Ca2+ levels and thereby may enhance transmitter release (Gray et al., 1996; also see Vogt and Regehr, 2001).

Co-localization of plasma membrane transporters of glutamate and GABA Glutamate transport is the major mechanism controlling extracellular glutamate levels, preventing excitotoxicity, and averting neural damage associated with hyperexcitability. As mentioned above, the neuronal glutamate transporter (EAAC1) is expressed in granule cells and, surprisingly, in a number of GABAergic neurons (Rothstein et al., 1994; He et al., 2002; Sepkuty et al., 2002). Therefore, it has been suggested that besides controlling extracellular glutamate levels, its function is linked to GABA metabolism, because capture of glutamate is essential for GABA synthesis (Sepkuty et al., 2002). It is not surprising that seizures and epilepsy have direct consequences on EAAC1 expression and function, as has been demonstrated (Gorter et al., 2002; Zhang et al., 2004). On the other hand, it has been proposed that the presence of the membrane transporter of GABA, GAT-1, is restricted to neurons that synthesize and release GABA, and glial cells (Iversen and Kelly, 1975; Radian et al., 1990; Ribak et al., 1996). Some GABAergic cells, immunocytochemically characterized by the presence of GAD67 or GABA, do not contain or contain traces of GAT-1, but not vice versa (Rattray and Priestley, 1993). However, it has been found that GAT-1 is also localized in

119

the glutamatergic granule cells (Frahm et al., 2000; Sperk et al., 2003), and it controls GABA uptake to the MF terminals (Gomez-Lira et al., 2002) but not GABA release at this synapse (Vivar and Gutierrez, 2005; Safiulina et al., 2006). Co-localization of the vesicular transporters for glutamate (VGlut-1) and GABA (VGAT) Because glutamate is a general metabolic substrate and serves as the precursor of inhibitory transmitter GABA, glutamate immunoreactivity is not specific to glutamatergic neurons. Therefore, the detection of glutamate vesicular transporter(s) has been used to establish the glutamatergic phenotype of neurons. As expected for glutamatergic neurons, the MF terminals of the granule cells contain the glutamate vesicular transporter VGlut-1 (Bellocchio et al., 1998; Kaneko et al., 2002). In accordance with immunohistological data showing that the granule cells express GAD and GABA, it was found that they express VGAT mRNA in an activity-dependent manner (Lamas et al., 2001; Gutierrez, 2003; Gomez-Lira et al., 2005). However, these cells do not contain the transporter protein (Chaudhry et al., 1998; Sperk et al., 2003; also see Safiulina et al., 2006). Thus, the lack of detection of VGAT indicates that its expression is too low to be revealed with immunohistochemistry, or the existence of a yet unidentified transporter. Recent experiments show that glutamate and GABA are in close relation to vesicles in MFs terminals (Bergersen et al., 2003) strongly suggesting that VGAT or another related transporter protein must be present in these terminals (Chaudhry et al., 1998; Gomez-Lira et al., 2005). Postsynaptic responses of CA3 pyramidal neurons to MF glutamatergic input As mentioned above, glutamate is believed to be the primary excitatory neurotransmitter released from the MFs (Crawford and Connor, 1973; Terrian et al., 1988). The primary ionotropic glutamate receptors mediating the fast synaptic response at the MF synapse are a-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid (AMPA)-type receptors

(Lanthorn et al., 1984; Neuman et al., 1988; Ito and Sugiyama, 1991; Jonas et al., 1993). Voltageclamp analysis of unitary and compound excitatory postsynaptic currents (EPSC) onto CA3 pyramidal neurons finds the equilibrium potential close to 0 mV, consistent with ionotropic glutamate receptors (Brown and Johnston, 1983). Based on the fast rise times and decay kinetics of the current, the responses are consistent with the properties of AMPA receptors and therefore reflects an electrotonically close synapse (Williams and Johnston, 1991; Jonas et al., 1993). Early studies using radio-labeled NMDA found that the MF terminal field contains a low density of NMDA receptors relative to other regions of the hippocampal formation (Monaghan and Cotman, 1985). Although lower than at other synaptic sites, glutamate release from the MFs activates a small, but measurable, NMDA component that, as expected, exhibits voltage-dependence and slower kinetics, which are characteristics of NMDA receptor-mediated responses (Jonas et al., 1993; Weisskopf and Nicoll, 1995). In contrast to NMDA receptors, the MF terminal field contains a sizable density of highaffinity sites for KA binding (Foster et al., 1981; Monaghan and Cotman, 1982) and KA channels are expressed in CA3 pyramidal neurons (Egebjerg et al., 1991; Werner et al., 1991). Focal application of kainate into the MF terminal field induces a strong depolarization in CA3 pyramidal neurons (Sawada et al., 1988). Postsynaptic activation of kainate receptors requires repetitive stimulation due to their slow kinetics (Castillo et al., 1997; Bortolotto et al., 2003). This, combined with the strong frequency-dependent facilitation of release, suggests that KA receptors contribute to the activity-dependent enhancement of transmission from granule cells to CA3 pyramidal neurons discussed above. Because of the arguments above that MFs release GABA and it acts as a classical neurotransmitter, one would expect that GABA receptors would be located in apposition to MF terminals. In support of this prediction, there is evidence that GABAA receptors cluster with glutamatergic receptors in pyramidal cells in apposition to both glutamatergic and GABAergic terminals

120

(Rao et al., 2000). It was later found, with immunogold detection, that AMPA and GABAA receptors also colocalize at MF synapses in the hilar region (Bergersen et al., 2003). Finally, glutamate released from MFs triggers the release of Ca++ from intracellular stores of CA3 pyramidal cells via mGluR receptor activation (Miller et al., 1996; Jaffe and Brown, 1997). Tetanic stimulation of the MFs triggers waves of spike-independent increases in Ca++ concentration in the proximal apical dendrites of CA3 pyramidal neurons mediated by the activation of type I mGluR receptors (Kapur et al., 2001). Single presynaptic spikes appear to be insufficient to elicit Ca++ release (Reid et al., 2001). Frequency-dependent mGluR release of Ca++ from intracellular stores may be important for longlasting changes in MF synaptic efficacy, discussed below (Yeckel et al., 1999; Wang et al., 2004).

Are MF synapses onto CA3 pyramidal neurons detonators? It is not surprising that LFS of the MFs fails to routinely trigger spikes, given many of the points made above. For example, the major target of the MFs in area CA3 involves the activation of feedforward inhibitory circuits. Glutamate is coreleased with GABA, and a wide-array of other neuromodulators, which potentially inhibit MF transmission. CA3 pyramidal cells also have a threshold that is 10–15 mV from resting potential (Podlogar and Dietrich, 2006). There is a large amplitude unitary EPSP (Jonas et al., 1993), so the MFs may not relay synapses with a high safety factor. An elegant demonstration of the ‘‘conditional’’ nature of MF excitation of CA3 pyramidal neurons was shown by Henze et al. (2002b). With low frequency stimulation (o40 Hz), the likelihood that a CA3 pyramidal neuron would discharge following a granule cell was low. Only at higher frequencies did CA3 pyramidal neurons follow granule cell firing. Interestingly, this may not be the case for all MF targets. Scharfman and colleagues showed faithful following of granule cell firing when the targets of MFs were hilar mossy cells or hilar interneurons (Scharfman et al.,

1990), a difference that may contribute to hilar cell vulnerability to excitotoxicity. The MF-to-CA3 synapse can therefore be considered as ‘‘conditional detonators’’ in two domains. The first, as described above, is the temporal domain where presynaptic facilitation, the slow membrane time constant of CA3 pyramidal neurons (Spruston and Johnston, 1992), and possibly postsynaptic amplification by KA receptors combine to boost EPSP amplitude and fire the cell. The second domain reflects the spatial summation of inputs concomitantly with the MFs. For example, if a MF EPSP occurred coincidently with a perforant path EPSP, the summation of these two would then be suprathreshold (Urban and Barrionuevo, 1998). Although all synapses could be considered as conditional detonators given this definition, the mean strength of unitary MF EPSPs endows them with a greater ability to move the membrane potential closer to threshold.

The granule cells simultaneously release glutamate and GABA: electrophysiological evidence Indirect but compelling evidence has accumulated over the last years of the co-release of glutamate and GABA from the MFs. The first electrophysiological evidence of GABAergic transmission from the MFs to CA3 agreed with the immunohistochemical observations showing that seizures transiently upregulated the expression of GAD65 and GAD67 (Schwarzer and Sperk, 1995; Sloviter et al., 1996). Indeed, it was shown that stimulation of the MFs produced monosynaptic GABA-mediated transmission in pyramidal cells of CA3 in kindled epileptic but not in control healthy rats (Gutierrez, 2000; Gutierrez and Heinemann, 2001). As mentioned above, feed-forward inhibition results from activation of local inhibitory interneurons by MF glutamatergic, excitatory synapses (Crawford and Connor, 1973; Acsady et al.), which in turn inhibit pyramidal cells (Yamamoto, 1972; Brown and Johnston, 1983; Buzsaki, 1984). Thus, activation of the DG leads to monosynaptic excitation and disynaptic inhibition on CA3 pyramidal neurons that, with intracellular

121

recordings, are observed as an EPSP/IPSP sequence. Both these conductances are typically blocked in the presence of blockers of glutamatergic transmission. In kindled epileptic animals, however, a bicuculline-sensitive IPSP is still elicited in the presence of the glutamate receptor blockers. This IPSP had the same latency as the control EPSP and could be inhibited by activation of metabotropic glutamate receptors (mGluR). Expression of the monosynaptic IPSP was transient because it could be observed 24–48 h after the last kindled seizure but was not present if the experiment was carried out a month after the last seizure. The kindled epileptic state is not necessary for this monosynaptic IPSP, because a single seizure (Gutierrez, 2000) or repeated LTP-like stimulation could elicit the response (Gutierrez, 2002). Moreover, the monosynaptic IPSP produced by the activation of the DG persisted when perfusing a medium with a low [Ca++] or the GABAB agonist, baclofen. These manipulations depressed the DG-evoked IPSP amplitude without altering its onset latency or slope. This unequivocally established that the IPSP was a monosynaptic response because, under these conditions, the ability to recruit local interneurons quickly enough to trigger such a short-latency IPSP is unlikely. The possibility of recruiting inhibitory interneurons by activating electrical synapses was also discarded (Gutierrez, 2000). Independently, Walker et al. (2001) demonstrated that monosynaptic GABAergic responses could be normally evoked in CA3 pyramidal cells by MF activation in slices of young guinea pigs (Walker et al., 2001). They showed that these MF GABAergic responses had the same pharmacological and plastic properties as those reported for glutamatergic MF transmission, i.e., there was strong frequency-dependent potentiation, and they were sensitive to an antagonist of the metabotropic glutamate receptor that is expressed preferentially by MFs, L-AP4. They were able to show that minimal stimulation of the MF evoked glutamate only, GABA only, and compound glutamate-GABA currents in pyramidal cells. This indicated that both responses, glutamatergic and GABAergic, had a common origin. In addition, both neurotransmitters could be released synchronously or asynchronously,

discarding the possibility that they are packaged in single vesicles. Moreover, the evidence of MF GABA transmission onto both pyramidal cells (Gutierrez, 2000; Gutierrez and Heinemann, 2001; Walker et al., 2001), and interneurons of CA3 (Romo-Parra et al., 2003; Safiulina et al., 2006) and onto the mossy cells in the hilar region (Bergersen et al., 2003) excludes the possible segregation of neurotransmitters according to the target cell. It has been shown that the putative release of GABA from the MF also occurs during development, when the granule cells express all the markers of the GABAergic phenotype, and provoke GABA receptor-mediated responses in pyramidal cells and interneurons of CA3. This occurs during the first weeks of age, after which the GABAergic phenotype shuts off in a clear-cut manner. Supporting this notion, recent work by Gutie´rrez et al. (2003), Kasyanov et al. (2004) and Safiulina et al. (2006) have shown that GABA is the main neurotransmitter released from MF terminals during the first postnatal days (Kasyanov et al., 2004; Safiulina et al., 2006). Thus, MFs contain two different sets of low- and high-threshold fibers that release GABA and GABA plus glutamate, respectively. The first would disappear with maturation, whereas the second would persist longer or would reappear in pathological conditions, such as in epilepsy (Gutierrez, 2003; Safiulina et al., 2006). It is possible that signaling by both, synaptic and nonsynaptic release of GABA, may play a crucial role in tuning hippocampal network during postnatal development (Gutierrez, 2003, 2005; Safiulina et al., 2006). While the molecular mechanisms that shut off the GABAergic phenotype at maturity have not been disclosed, upregulation of the GABAergic phenotype in the adult depends on protein synthesis, Ca++ entry, and activation of TrkB receptors by BDNF (Gutierrez, 2002; Romo-Parra et al., 2003; Gomez-Lira et al., 2005).

Long-term plasticity at the MF CA3 synapse Like other excitatory synapses in the hippocampal formation, the MF synapse expresses

122

LTP in response to a brief episode of HFS (Yamamoto et al., 1980). As described above, the MF terminal field contains a lower density of NMDA receptors compared with other areas of the hippocampus. This observation motivated Harris and Cotman (1986) to test whether LTP at the MF synapse was dependent on NMDA receptors. They found that in the presence of NMDA receptor antagonists, HFS of the MFs was still capable of triggering LTP (Harris and Cotman, 1986). Although there is agreement that Ca++ is necessary for the induction process, the mechanism underlying the induction of NMDA-independent LTP has been a source of controversy (Nicoll and Schmitz, 2005). Briefly, many experiments point to an induction mechanism that is solely presynaptic. Here the working hypothesis is that HFS triggers an increase in presynaptic Ca++ that, via calmodulin, activates adenylyl cyclase (Zalutsky and Nicoll, 1992; Weisskopf et al., 1994; Villacres et al., 1998). An alternative presynaptic mechanism is that presynaptic KA receptor activation leads to Ca++ influx that triggers release of Ca++ from intracellular stores (Contractor et al., 2001; Lauri et al., 2001; Bortolotto et al., 2003; Lauri et al., 2003), although this remains controversial (Breustedt and Schmitz, 2004). Alternatively, there is evidence for a postsynaptic component to the induction mechanism (Williams and Johnston, 1989; Jaffe and Johnston, 1990). Here, an interaction between Ca++ entry through voltage-gated Ca++ channels and the activation of mGluRs takes place (Kapur et al., 1998; Yeckel et al., 1999; Kapur et al., 2001; Wang et al., 2004). It may be that depending on stimulus conditions, induction may have either a presynaptic or postsynaptic locus (Urban and Barrionuevo, 1996). In contrast to the induction process, there is little controversy regarding the locus for the expression of MF LTP. Numerous studies indicate that synaptic potentiation is mediated by a presynaptic change in transmitter release (Zalutsky and Nicoll, 1990; Yamamoto et al., 1992; Zalutsky and Nicoll, 1992; Xiang et al., 1994; Reid et al., 2004). If induction of MF LTP has a postsynaptic locus, then expression must require a retrograde communication. One possible mechanism involves postsynaptic ephrinB receptors and presynaptic

B-ephrins reverse signaling (Contractor et al., 2002; Armstrong et al., 2006). ‘‘What goes up must go down’’ is an apt phrase to describe MF synaptic plasticity. LFS triggers long-term depression (LTD) of MF synaptic inputs onto CA3 pyramidal neurons (Kobayashi et al., 1996; Yokoi et al., 1996) where presynaptic activation of mGluR2 receptors depresses cAMP levels, countering the expression of LTP.

CA3-interneuron synapses As discussed above, the majority of MF synaptic contacts are onto GABAergic interneurons (Acsady et al., 1998) of which there are a wide variety of subtypes, typically characterized based on a combination of parameters including cell body location, axonal projection, morphology, colocalized peptides, and calcium binding protein content (Parra et al., 1998). Of particular interest are a class of bipolar interneurons whose dendrites primarily reside along and within s. lucidum (Spruston et al., 1997; Vida and Frotscher, 2000), in contrast to other interneurons where the dendritic tree projects orthogonally against the MFs that receive most of their innervation from other pathways. The studies discussed below were made primarily from this specific subtype of interneuron. Like the pyramidal cell synapse, MF transmission onto interneurons is reduced by the activation of presynaptic group II mGluR receptors (Maccaferri et al., 1998; Toth et al., 2000); Pelkey et al., 2005), although other types of mGluRs may play a role in synaptic plasticity at this synapse (discussed below). In contrast to the pyramidal neuron synapse, at MF-interneuron synapses AMPA receptor expression is heterogeneous. Glutamate released from the MFs may activate receptors with or without GluR2 subunits, thereby forming receptors with either Ca++ permeable or impermeable subunits (Toth and McBain, 1998; Toth et al., 2000). Also different from the pyramidal neuron MF synapse is the frequency response of the interneuron, which is highly variable. The response can range from a weak frequency-dependent facilitation (much less pronounced than at the pyramidal

123

neuron synapse) to synaptic depression (Toth et al., 2000). Quantal transmission at interneuron synapses is also different than at the pyramidal neuron synapse. As expected from anatomy, these synapses express a low number of release sites (Acsady et al., 1998). The probability of release at these sites, however, appears higher than at the pyramidal neuron synapse (Lawrence et al., 2004). This difference in probability of release is also consistent with differences frequency facilitation observed between the two synapses (i.e., a low initial probability of release predicts large facilitation). Interestingly, quantal size at interneuron synapses is within the same ranges, if not higher, than for pyramidal neuron synapse. As a result, and combined with a more depolarized resting potenial, unitary responses of interneuron synapses are as capable of triggering action potentials as MF inputs onto pyramidal neurons (Lawrence et al., 2004), and this also appears to be the case for MF transmission to hilar interneurons (Scharfman et al., 1990). Long-lasting plasticity at MF-interneuron synapses also differ from their pyramidal neuron counterparts. Most notably, HFS does not elicit LTP at the MF-interneuron synapse (Maccaferri et al., 1998). Rather, HFS triggers LTD at synapses expressing Ca++-permeable AMPA receptors, while NMDA receptors trigger the induction of LTD at synapses containing Ca++impermeant AMPA receptors (Lei and McBain, 2002). Activation of mGluR7 receptors by L-AP4 produces a chemical version of LTD at this synapse that is distinct from pyramidal neuron contacts (Pelkey et al., 2005) and differentially modulates voltage-gated Ca++ channels in the presynaptic filopodia (Pelkey et al., 2006). If HFS triggers LTD at the MF-interneuron synapse, feed-forward inhibition should be depressed onto CA3 pyramidal neurons following HFS. However, fast synaptic inhibition onto CA3 pyramidal neurons is not affected by LTP of the MFs (Griffith et al., 1986) suggesting that a decrease in feed-forward inhibition through interneurons of s. lucidum is compensated by enhanced recurrent inhibition or other forms of synaptic plasticity within the CA3 network.

Communication from DG to CA3: exciting yet inhibiting A picture is emerging that relates the contribution of MF transmission to excitation and inhibition in area CA3, both in terms of the repertoire of neurotransmitters/modulators released from the fibers but also with respect to circuitry that is required to take into account inhibitory neurons. At low frequencies (o0.5 Hz), frequency facilitation of the MF-CA3 pyramidal neuron synapse is weak and the probability of eliciting a spike is low (Henze et al., 2002b). Furthermore, in this frequency domain granule cells are likely to be firing inhibitory interneurons preferentially relative to pyramidal neurons CA3 (Bragin et al., 1995a, b; Penttonen et al., 1997). As such, concomitant extrinsic input from entorhinal cortex (via the perforant path) or via recurrent excitation via the C/A pathway should fail to excite sets of CA3 pyramidal neurons. But as granule cell firing rates increase, such as when an animal traverses a place field, frequency facilitation can then depolarize CA3 pyramidal neurons to spike threshold. Interestingly, after seizures, MFs tonically inhibit b/g field oscillations and spontaneous subthreshold membrane oscillations of CA3 interneurons in the CA3 area through GABA-medated signaling. Coincident stimulation of the MFs at y and b/g frequencies produces a frequency-dependent excitation of interneurons and the inhibition of pyramidal cells. Indeed, these effects are maximal at the b frequency, suggesting a resonance phenomenon (Trevin˜o et al., 2007). Because of the properties of synaptic plasticity within the auto-associative circuit of CA3, pattern separation, pattern completion, and the encoding of new information will then be dependent on the precise timing of granule cell and pyramidal neuron firing (Kohonen, 1977; Chattarji et al., 1989; Zalutsky and Nicoll, 1990; August and Levy, 1999; Nakazawa et al., 2002; Rolls and Kesner, 2006). References Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403.

124 Alle, H. and Geiger, J.R. (2006) Combined analog and action potential coding in hippocampal mossy fibers. Science, 311: 1290–1293. Altar, C.A. and DiStefano, P.S. (1998) Neurotrophin trafficking by anterograde transport. Trends Neurosci., 21: 433–437. Amaral, D.G. (1979) Synaptic extensions from the mossy fibers of the fascia dentata. Anat. Embryol. (Berl.), 155: 241–251. Amaral, D.G. and Dent, J.A. (1981) Development of the mossy fibers of the DG: I. A light and electron microscopic study of the mossy fibers and their expansions. J. Comp. Neurol., 195: 51–86. Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons, numbers and the hippocampal network. Prog. Brain Res., 83: 1–11. Armstrong, J.N., Brust, T.B., Lewis, R.G. and MacVicar, B.A. (2002) Activation of presynaptic P2
7-like receptors depresses mossy fiber-CA3 synaptic transmission through p38 mitogen-activated protein kinase. J. Neurosci., 22: 5938–5945. Armstrong, J.N., Saganich, M.J., Xu, N.J., Henkemeyer, M., Heinemann, S.F. and Contractor, A. (2006) B-ephrin reverse signaling is required for NMDA-independent long-term potentiation of mossy fibers in the hippocampus. J. Neurosci., 26: 3474–3481. August, D.A. and Levy, W.B. (1999) Temporal sequence compression by an integrate-and-fire model of hippocampal area CA3. J. Comput. Neurosci., 6: 71–90. Awatramani, G.B., Price, G.D. and Trussell, L.O. (2005) Modulation of transmitter release by presynaptic resting potential and background calcium levels. Neuron, 48: 109–121. Barrionuevo, G., Kelso, S.R., Johnston, D. and Brown, T.H. (1986) Conductance mechanism responsible for long-term potentiation in monosynaptic and isolated excitatory synaptic inputs to hippocampus. J. Neurophysiol., 55: 540–550. Bellocchio, E.E., Hu, H., Pohorille, A., Chan, J., Pickel, V.M. and Edwards, R.H. (1998) The localization of the brainspecific inorganic phosphate transporter suggests a specific presynaptic role in glutamatergic transmission. J. Neurosci., 18: 8648–8659. Bergersen, L., Ruiz, A., Bjaalie, J.G., Kullmann, D.M. and Gundersen, V. (2003) GABA and GABAA receptors at hippocampal mossy fibre synapses. Eur. J. Neurosci., 18: 931–941. Berninger, B., Garcia, D.E., Inagaki, N., Hahnel, C. and Lindholm, D. (1993) BDNF and NT-3 induce intracellular Ca++ elevation in hippocampal neurones. Neuroreport, 4: 1303–1306. Bischofberger, J., Engel, D., Frotscher, M. and Jonas, P. (2006) Timing and efficacy of transmitter release at mossy fiber synapses in the hippocampal network. Pflugers Arch., 453: 361–372. Blackstad, T.W., Brink, K., Hem, J. and Jeune, B. (1970) Distribution of hippocampal mossy fibers in the rat. An experimental study with silver impregnation methods. J. Comp. Neurol., 138: 433–449. Blackstad, T.W. and Kjaerheim, A. (1961) Special axodendritic synapses in the hippocampal cortex: electron and

light microscopic studies on the layer of mossy fibers. J. Comp. Neurol., 117: 133–159. Blo¨chl, A. and Thoenen, H. (1995) Characterization of nerve growth factor (NGF) release from hippocampal neurons: evidence for a constitutive and an unconventional sodiumdependent regulated pathway. Eur. J. Neurosci., 7: 1220–1228. Borhegyi, Z. and Leranth, C. (1997) Substance P innervation of the rat hippocampal formation. J. Comp. Neurol., 384: 41–58. Bortolotto, Z.A., Lauri, S., Isaac, J.T. and Collingridge, G.L. (2003) Kainate receptors and the induction of mossy fibre long-term potentiation. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358: 657–666. Braas, K.M., Newby, A.C., Wilson, V.S. and Snyder, S.H. (1986) Adenosine-containing neurons in the brain localized by immunocytochemistry. J. Neurosci., 6: 1952–1961. Bradley, S.R., Levey, A.I., Hersch, S.M. and Conn, P.J. (1996) Immunocytochemical localization of group III metabotropic glutamate receptors in the hippocampus with subtype-specific antibodies. J. Neurosci., 16: 2044–2056. Bragin, A., Jando, G., Nadasdy, Z., Hetke, J., Wise, K. and Buzsaki, G. (1995a) Gamma (40–100 Hz) oscillation in the hippocampus of the behaving rat. J. Neurosci., 15: 47–60. Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M. and Buzsaki, G. (1995b) Dentate EEG spikes and associated interneuronal population bursts in the hippocampal hilar region of the rat. J. Neurophysiol., 73: 1691–1705. Breustedt, J. and Schmitz, D. (2004) Assessing the role of GLUK5 and GLUK6 at hippocampal mossy fiber synapses. J. Neurosci., 24: 10093–10098. Brown, T.H. and Johnston, D. (1983) Voltage-clamp analysis of mossy fiber synaptic input to hippocampal neurons. J. Neurophysiol., 50: 487–507. Buzsaki, G. (1984) Feed-forward inhibition in the hippocampal formation. Prog. Neurobiol., 22: 131–153. Castillo, P.E., Malenka, R.C. and Nicoll, R.A. (1997) Kainate receptors mediate a slow postsynaptic current in hippocampal CA3 neurons. Nature, 388: 182–186. Castillo, P.E., Salin, P.A., Weisskopf, M.G. and Nicoll, R.A. (1996) Characterizing the site and mode of action of dynorphin at hippocampal mossy fiber synapses in the guinea pig. J. Neurosci., 16: 5942–5950. Castillo, P.E., Weisskopf, M.G. and Nicoll, R.A. (1994) The role of Ca2+ channels in hippocampal mossy fiber synaptic transmission and long-term potentiation. Neuron, 12: 261–269. Chandy, J., Pierce, J.P. and Milner, T.A. (1995) Rat hippocampal mossy fibers contain cholecystokinin-like immunoreactivity. Anat. Rec., 243: 519–523. Chattarji, S., Stanton, P.K. and Sejnowski, T.J. (1989) Commissural synapses, but not mossy fiber synapses, in hippocampal field CA3 exhibit associative long-term potentiation and depression. Brain Res., 495: 145–150. Chaudhry, F.A., Reimer, R.J., Bellocchio, E.E., Danbolt, N.C., Osen, K.K., Edwards, R.H. and Storm-Mathisen, J. (1998)

125 The vesicular GABA transporter, VGAT, localizes to synaptic vesicles in sets of glycinergic as well as GABAergic neurons. J. Neurosci., 18: 9733–9750. Chavkin, C. (2000) Dynorphins are endogenous opioid peptides released from granule cells to act neurohumorly and inhibit excitatory neurotransmission in the hippocampus. Prog. Brain Res., 125: 363–367. Chavkin, C., Shoemaker, W.J., McGinty, J.F., Bayon, A. and Bloom, F.E. (1985) Characterization of the prodynorphin and proenkephalin neuropeptide systems in rat hippocampus. J. Neurosci., 5: 808–816. Chicurel, M.E. and Harris, K.M. (1992) Three-dimensional analysis of the structure and composition of CA3 branched dendritic spines and their synaptic relationships with mossy fiber boutons in the rat hippocampus. J. Comp. Neurol., 325: 169–182. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat DG. J. Comp. Neurol., 246: 435–458. Conner, J.M., Lauterborn, J.C., Yan, Q., Gall, C.M. and Varon, S. (1997) Distribution of brain-derived neurotrophic factor (BDNF) protein and mRNA in the normal adult rat CNS: evidence for anterograde axonal transport. J. Neurosci., 17: 2295–2313. Contractor, A., Rogers, C., Maron, C., Henkemeyer, M., Swanson, G.T. and Heinemann, S.F. (2002) Trans-synaptic Eph receptor-ephrin signaling in hippocampal mossy fiber LTP. Science, 296: 1864–1869. Contractor, A., Swanson, G. and Heinemann, S.F. (2001) Kainate receptors are involved in short- and long-term plasticity at mossy fiber synapses in the hippocampus. Neuron, 29: 209–216. Crawford, I.L. and Connor, J.D. (1973) Localization and release of glutamic acid in relation to the hippocampal mossy fibre pathway. Nature, 244: 442–443. Danzer, S.C. and McNamara, J.O. (2004) Localization of brain-derived neurotrophic factor to distinct terminals of mossy fiber axons implies regulation of both excitation and feedforward inhibition of CA3 pyramidal cells. J. Neurosci., 24: 11346–11355. Darstein, M., Petralia, R.S., Swanson, G.T., Wenthold, R.J. and Heinemann, S.F. (2003) Distribution of kainate receptor subunits at hippocampal mossy fiber synapses. J. Neurosci., 23: 8013–8019. Debanne, D., Gahwiler, B.H. and Thompson, S.M. (1998) Long-term synaptic plasticity between pairs of individual CA3 pyramidal cells in rat hippocampal slice cultures. J. Physiol., 507(Pt 1): 237–247. Ding, R., Asada, H. and Obata, K. (1998) Changes in extracellular glutamate and GABA levels in the hippocampal CA3 and CA1 areas and the induction of glutamic acid decarboxylase-67 in dentate granule cells of rats treated with kainic acid. Brain Res., 800: 105–113. Draguhn, A., Verdorn, T.A., Ewert, M., Seeburg, P.H. and Sakmann, B. (1990) Functional and molecular distinction between recombinant rat GABAA receptor subtypes by Zn2+. Neuron, 5: 781–788.

Drake, C.T., Patterson, T.A., Simmons, M.L., Chavkin, C. and Milner, T.A. (1996) Kappa opioid receptor-like immunoreactivity in guinea pig brain: ultrastructural localization in presynaptic terminals in hippocampal formation. J. Comp. Neurol., 370: 377–395. Egebjerg, J., Bettler, B., Hermans-Borgmeyer, I. and Heinemann, S. (1991) Cloning of a cDNA for a glutamate receptor subunit activated by kainate but not AMPA. Nature, 351: 745–748. Elmer, E., Kokaia, Z., Kokaia, M., Carnahan, J., Nawa, H. and Lindvall, O. (1998) Dynamic changes of brain-derived neurotrophic factor protein levels in the rat forebrain after single and recurring kindling-induced seizures. Neuroscience, 83: 351–362. Engel, D. and Jonas, P. (2005) Presynaptic action potential amplification by voltage-gated Na+ channels in hippocampal mossy fiber boutons. Neuron, 45: 405–417. Erlander, M.G. and Tobin, A.J. (1991) The structural and functional heterogeneity of glutamic acid decarboxylase: a review. Neurochem. Res., 16: 215–226. Foster, A.C., Mena, E.E., Monaghan, D.T. and Cotman, C.W. (1981) Synaptic localization of kainic acid binding sites. Nature, 289: 73–75. Frahm, C., Engel, D., Piechotta, A., Heinemann, U. and Draguhn, A. (2000) Presence of gamma-aminobutyric acid transporter mRNA in interneurons and principal cells of rat hippocampus. Neurosci. Lett., 288: 175–178. Furtinger, S., Pirker, S., Czech, T., Baumgartner, C., Ransmayr, G. and Sperk, G. (2001) Plasticity of Y1 and Y2 receptors and neuropeptide Y fibers in patients with temporal lobe epilepsy. J. Neurosci., 21: 5804–5812. Gall, C. (1984) The distribution of cholecystokinin-like immunoreactivity in the hippocampal formation of the guinea pig: localization in the mossy fibers. Brain Res., 306: 73–83. Gall, C. (1988) Seizures induce dramatic and distinctly different changes in enkephalin, dynorphin, and CCK immunoreactivities in mouse hippocampal mossy fibers. J. Neurosci., 8: 1852–1862. Gall, C., Brecha, N., Karten, H.J. and Chang, K.J. (1981) Localization of enkephalin-like immunoreactivity to identified axonal and neuronal populations of the rat hippocampus. J. Comp. Neurol., 198: 335–350. Gall, C. and Lauterborn, J. (1992) The DG: a model system for studies of neurotrophin regulation. Epilepsy Res. Suppl., 7: 171–185. Gall, C., Lauterborn, J., Isackson, P. and White, J. (1990) Seizures, neuropeptide regulation, and mRNA expression in the hippocampus. Prog. Brain Res., 83: 371–390. Gall, C.M. and Isackson, P.J. (1989) Limbic seizures increase neuronal production of messenger RNA for nerve growth factor. Science, 245: 758–761. Geiger, J.R. and Jonas, P. (2000) Dynamic control of presynaptic Ca(2+) inflow by fast-inactivating K(+) channels in hippocampal mossy fiber boutons. Neuron, 28: 927–939. Gloor, P., Vera, C.L. and Sperti, L. (1963) Electrophysiological Studies of Hippocampal Neurons. I. Configuration and

126 laminar analysis of the ‘‘resting’’ potential gradient, of the main-transient response to perforant path, fimbrial and mossy fiber volleys and of ‘‘spontaneous’’ activity. Electroencephalogr. Clin. Neurophysiol., 15: 353–378. Gomez-Lira, G., Lamas, M., Romo-Parra, H. and Gutierrez, R. (2005) Programmed and induced phenotype of the hippocampal granule cells. J. Neurosci., 25: 6939–6946. Gomez-Lira, G., Trillo, E., Ramirez, M., Asai, M., Sitges, M. and Gutierrez, R. (2002) The expression of GABA in mossy fiber synaptosomes coincides with the seizure-induced expression of GABAergic transmission in the mossy fiber synapse. Exp. Neurol., 177: 276–283. Gorter, J.A., Van Vliet, E.A., Proper, E.A., De Graan, P.N., Ghijsen, W.E., Lopes Da Silva, F.H. and Aronica, E. (2002) Glutamate transporters alterations in the reorganizing DG are associated with progressive seizure activity in chronic epileptic rats. J. Comp. Neurol., 442: 365–377. Gray, R., Rajan, A.S., Radcliffe, K.A., Yakehiro, M. and Dani, J.A. (1996) Hippocampal synaptic transmission enhanced by low concentrations of nicotine. Nature, 383: 713–716. Griffith, W.H., Brown, T.H. and Johnston, D. (1986) Voltageclamp analysis of synaptic inhibition during long-term potentiation in hippocampus. J. Neurophysiol., 55: 767–775. Gutierrez, R. (2000) Seizures induce simultaneous GABAergic and glutamatergic transmission in the DG-CA3 system. J. Neurophysiol., 84: 3088–3090. Gutierrez, R. (2002) Activity-dependent expression of simultaneous glutamatergic and GABAergic neurotransmission from the mossy fibers in vitro. J. Neurophysiol., 87: 2562–2570. Gutierrez, R. (2003) The GABAergic phenotype of the ‘‘glutamatergic’’ granule cells of the DG. Prog. Neurobiol., 71: 337–358. Gutierrez, R. (2005) The dual glutamatergic-GABAergic phenotype of hippocampal granule cells. Trends Neurosci., 28: 297–303. Gutierrez, R. and Heinemann, U. (2001) Kindling induces transient fast inhibition in the DG-CA3 projection. Eur. J. Neurosci., 13: 1371–1379. Gutierrez, R., Romo-Parra, H., Maqueda, J., Vivar, C., Ramirez, M., Morales, M.A. and Lamas, M. (2003) Plasticity of the GABAergic phenotype of the ‘‘glutamatergic’’ granule cells of the rat DG. J. Neurosci., 23: 5594–5598. Hallermann, S., Pawlu, C., Jonas, P. and Heckmann, M. (2003) A large pool of releasable vesicles in a cortical glutamatergic synapse. Proc. Natl. Acad. Sci. U.S.A., 100: 8975–8980. Hamlyn, L.H. (1961) Electron microscopy of mossy fibre endings in Ammon’s horn. Nature, 190: 645–646. Harris, E.W. and Cotman, C.W. (1986) Long-term potentiation of guinea pig mossy fiber responses is not blocked by Nmethyl D-aspartate antagonists. Neurosci. Lett., 70: 132–137. He, X.P., Minichiello, L., Klein, R. and McNamara, J.O. (2002) Immunohistochemical evidence of seizure-induced activation of trkB receptors in the mossy fiber pathway of adult mouse hippocampus. J. Neurosci., 22: 7502–7508. Henze, D.A., Card, J.P., Barrionuevo, G. and Ben-Ari, Y. (1997) Large amplitude miniature excitatory postsynaptic

currents in hippocampal CA3 pyramidal neurons are of mossy fiber origin. J. Neurophysiol., 77: 1075–1086. Henze, D.A., McMahon, D.B., Harris, K.M. and Barrionuevo, G. (2002a) Giant miniature EPSCs at the hippocampal mossy fiber to CA3 pyramidal cell synapse are monoquantal. J. Neurophysiol., 87: 15–29. Henze, D.A., Wittner, L. and Buzsaki, G. (2002b) Single granule cells reliably discharge targets in the hippocampal CA3 network in vivo. Nat. Neurosci., 5: 790–795. Ho, M.W., Beck-Sickinger, A.G. and Colmers, W.F. (2000) Neuropeptide Y(5) receptors reduce synaptic excitation in proximal subiculum, but not epileptiform activity in rat hippocampal slices. J. Neurophysiol., 83: 723–734. Holm, I.E., Geneser, F.A. and Zimmer, J. (1993) Cholecystokinin-, enkephalin-, and substance P-like immunoreactivity in the dentate area, hippocampus, and subiculum of the domestic pig. J. Comp. Neurol., 331: 310–325. Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O., Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered patterns of dynorphin immunoreactivity suggest mossy fiber reorganization in human hippocampal epilepsy. J. Neurosci., 10: 267–282. Huang, E.J. and Reichardt, L.F. (2001) Neurotrophins: roles in neuronal development and function. Annu. Rev. Neurosci., 24: 677–736. Isackson, P.J., Huntsman, M.M., Murray, K.D. and Gall, C.M. (1991) BDNF mRNA expression is increased in adult rat forebrain after limbic seizures: temporal patterns of induction distinct from NGF. Neuron, 6: 937–948. Ito, I. and Sugiyama, H. (1991) Roles of glutamate receptors in long-term potentiation at hippocampal mossy fiber synapses. Neuroreport, 2: 333–336. Iversen, L.L. and Kelly, J.S. (1975) Uptake and metabolism of gamma-aminobutyric acid by neurones and glial cells. Biochem. Pharmacol., 24: 933–938. Jacques, D., Dumont, Y., Fournier, A. and Quirion, R. (1997) Characterization of neuropeptide Y receptor subtypes in the normal human brain, including the hypothalamus. Neuroscience, 79: 129–148. Jaffe, D. and Johnston, D. (1990) Induction of long-term potentiation at hippocampal mossy-fiber synapses follows a Hebbian rule. J. Neurophysiol., 64: 948–960. Jaffe, D.B. and Brown, T.H. (1997) Calcium dynamics in thorny excrescences of CA3 pyramidal neurons. J. Neurophysiol., 78: 10–18. Jeub, M., Lie, A., Blumcke, I., Elger, C.E. and Beck, H. (1999) Loss of dynorphin-mediated inhibition of voltage-dependent Ca2+ currents in hippocampal granule cells isolated from epilepsy patients is associated with mossy fiber sprouting. Neuroscience, 94: 465–471. Jin, W. and Chavkin, C. (1999) Mu opioids enhance mossy fiber synaptic transmission indirectly by reducing GABAB receptor activation. Brain Res., 821: 286–293. Johnston, D. and Amaral, D.G. (2004) Hippocampus. In: Shepherd G.M. (Ed.), The Synaptic Organization of the Brain. Oxford University Press, New York.

127 Johnston, D. and Brown, T.H. (1983) Interpretation of voltageclamp measurements in hippocampal neurons. J. Neurophysiol., 50: 464–486. Jonas, P., Major, G. and Sakmann, B. (1993) Quantal components of unitary EPSCs at the mossy fibre synapse on CA3 pyramidal cells of rat hippocampus. J. Physiol., 472: 615–663. Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of unit activity in the hippocampal granular layer. Hippocampus, 3: 165–182. Kafitz, K.W., Rose, C.R., Thoenen, H. and Konnerth, A. (1999) Neurotrophin-evoked rapid excitation through TrkB receptors. Nature, 401: 918–921. Kamiya, H. and Ozawa, S. (1999) Dual mechanism for presynaptic modulation by axonal metabotropic glutamate receptor at the mouse mossy fibre-CA3 synapse. J. Physiol., 518(Pt 2): 497–506. Kamiya, H., Sawada, S. and Yamamoto, C. (1988) Synthetic omega-conotoxin blocks synaptic transmission in the hippocampus in vitro. Neurosci. Lett., 91: 84–88. Kamiya, H., Shinozaki, H. and Yamamoto, C. (1996) Activation of metabotropic glutamate receptor type 2/3 suppresses transmission at rat hippocampal mossy fibre synapses. J. Physiol., 493(Pt 2): 447–455. Kaneko, T., Fujiyama, F. and Hioki, H. (2002) Immunohistochemical localization of candidates for vesicular glutamate transporters in the rat brain. J. Comp. Neurol., 444: 39–62. Kapur, A., Yeckel, M. and Johnston, D. (2001) Hippocampal mossy fiber activity evokes Ca2+ release in CA3 pyramidal neurons via a metabotropic glutamate receptor pathway. Neuroscience, 107: 59–69. Kapur, A., Yeckel, M.F., Gray, R. and Johnston, D. (1998) L-type calcium channels are required for one form of hippocampal mossy fiber LTP. J. Neurophysiol., 79: 2181–2190. Kasyanov, A.M., Safiulina, V.F., Voronin, L.L. and Cherubini, E. (2004) GABA-mediated giant depolarizing potentials as coincidence detectors for enhancing synaptic efficacy in the developing hippocampus. Proc. Natl. Acad. Sci. U.S.A., 101: 3967–3972. Kaufman, D.L., Houser, C.R. and Tobin, A.J. (1991) Two forms of the gamma-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions. J. Neurochem., 56: 720–723. von Kitzing, E., Jonas, P. and Sakmann, B. (1994) Quantal analysis of excitatory postsynaptic currents at the hippocampal mossy fiber-CA3 pyramidal cell synapse. Adv. Second Messenger Phosphoprotein Res., 29: 235–260. Knowles, W.D., Schneiderman, J.H., Wheal, H.V., Stafstrom, C.E. and Schwartzkroin, P.A. (1984) Hyperpolarizing potentials in guinea pig hippocampal CA3 neurons. Cell Mol. Neurobiol., 4: 207–230. Kobayashi, K., Manabe, T. and Takahashi, T. (1996) Presynaptic long-term depression at the hippocampal mossy fiberCA3 synapse. Science, 273: 648–650. Kohonen, T. (1977) Associative Memory: A System Theoretic Approach. Springer-Verlag, New York.

Kukley, M., Schwan, M., Fredholm, B.B. and Dietrich, D. (2005) The role of extracellular adenosine in regulating mossy fiber synaptic plasticity. J. Neurosci., 25: 2832–2837. Kullmann, D.M., Ruiz, A., Rusakov, D.M., Scott, R., Semyanov, A. and Walker, M.C. (2005) Presynaptic, extrasynaptic and axonal GABAA receptors in the CNS: where and why? Prog. Biophys. Mol. Biol., 87: 33–46. Lamas, M., Gomez-Lira, G. and Gutierrez, R. (2001) Vesicular GABA transporter mRNA expression in the DG and in mossy fiber synaptosomes. Brain Res. Mol. Brain Res., 93: 209–214. Langdon, R.B., Johnson, J.W. and Barrionuevo, G. (1993) Asynchrony of mossy fibre inputs and excitatory postsynaptic currents in rat hippocampus. J. Physiol., 472: 157–176. Lanthorn, T.H., Ganong, A.H. and Cotman, C.W. (1984) 2-Amino-4-phosphonobutyrate selectively blocks mossy fiber-CA3 responses in guinea pig but not rat hippocampus. Brain Res., 290: 174–178. Lauri, S.E., Bortolotto, Z.A., Bleakman, D., Ornstein, P.L., Lodge, D., Isaac, J.T. and Collingridge, G.L. (2001) A critical role of a facilitatory presynaptic kainate receptor in mossy fiber LTP. Neuron, 32: 697–709. Lauri, S.E., Bortolotto, Z.A., Nistico, R., Bleakman, D., Ornstein, P.L., Lodge, D., Isaac, J.T. and Collingridge, G.L. (2003) A role for Ca2+ stores in kainate receptordependent synaptic facilitation and LTP at mossy fiber synapses in the hippocampus. Neuron, 39: 327–341. Lawrence, J.J., Grinspan, Z.M. and McBain, C.J. (2004) Quantal transmission at mossy fibre targets in the CA3 region of the rat hippocampus. J. Physiol., 554: 175–193. Lehmann, H., Ebert, U. and Loscher, W. (1996) Immunocytochemical localization of GABA immunoreactivity in dentate granule cells of normal and kindled rats. Neurosci. Lett., 212: 41–44. Lei, S. and McBain, C.J. (2002) Distinct NMDA receptors provide differential modes of transmission at mossy fiberinterneuron synapses. Neuron, 33: 921–933. Liang, Y., Yuan, L.L., Johnston, D. and Gray, R. (2002) Calcium signaling at single mossy fiber presynaptic terminals in the rat hippocampus. J. Neurophysiol., 87: 1132–1137. Lie, A.A., Becker, A., Behle, K., Beck, H., Malitschek, B., Conn, P.J., Kuhn, R., Nitsch, R., Plaschke, M., Schramm, J., Elger, C.E., Wiestler, O.D. and Blumcke, I. (2000) Up-regulation of the metabotropic glutamate receptor mGluR4 in hippocampal neurons with reduced seizure vulnerability. Ann. Neurol., 47: 26–35. Liu, H., Mazarati, A.M., Katsumori, H., Sankar, R. and Wasterlain, C.G. (1999) Substance P is expressed in hippocampal principal neurons during status epilepticus and plays a critical role in the maintenance of status epilepticus. Proc. Natl. Acad. Sci. U.S.A., 96: 5286–5291. Lorente de No, E. (1934) Studies on the structure of the cerebral cortex II. Continuation of the study of the ammonic system. J. Psychol. Neurol., 46: 113–177. Lowenstein, D.H. and Arsenault, L. (1996) The effects of growth factors on the survival and differentiation of cultured DG neurons. J. Neurosci., 16: 1759–1769.

128 Maccaferri, G., Toth, K. and McBain, C.J. (1998) Targetspecific expression of presynaptic mossy fiber plasticity. Science, 279: 1368–1370. Makiura, Y., Suzuki, F., Chevalier, E. and Onteniente, B. (1999) Excitatory granule cells of the DG exhibit a double inhibitory neurochemical content after intrahippocampal administration of kainate in adult mice. Exp. Neurol., 159: 73–83. Manzoni, O.J., Castillo, P.E. and Nicoll, R.A. (1995) Pharmacology of metabotropic glutamate receptors at the mossy fiber synapses of the guinea pig hippocampus. Neuropharmacology, 34: 965–971. Maqueda, J., Ramirez, M., Lamas, M. and Gutierrez, R. (2003) Glutamic acid decarboxylase (GAD)67, but not GAD65, is constitutively expressed during development and transiently overexpressed by activity in the granule cells of the rat. Neurosci. Lett., 353: 69–71. Marr, D. (1971) Simple memory: a theory of archicortex. Philos. Trans. R. Soc. Lond., 262: 23–81. Marsh, D.J., Baraban, S.C., Hollopeter, G. and Palmiter, R.D. (1999) Role of the Y5 neuropeptide Y receptor in limbic seizures. Proc. Natl. Acad. Sci. U.S.A., 96: 13518–13523. Marty, S., Berninger, B., Carroll, P. and Thoenen, H. (1996a) GABAergic stimulation regulates the phenotype of hippocampal interneurons through the regulation of brain-derived neurotrophic factor. Neuron, 16: 565–570. Marty, S., Berzaghi Mda, P. and Berninger, B. (1997) Neurotrophins and activity-dependent plasticity of cortical interneurons. Trends Neurosci., 20: 198–202. Marty, S., Carroll, P., Cellerino, A., Castren, E., Staiger, V., Thoenen, H. and Lindholm, D. (1996b) Brain-derived neurotrophic factor promotes the differentiation of various hippocampal nonpyramidal neurons, including Cajal-Retzius cells, in organotypic slice cultures. J. Neurosci., 16: 675–687. McCarthy, J.B., Walker, M., Pierce, J., Camp, P. and White, J.D. (1998) Biosynthesis and metabolism of native and oxidized neuropeptide Y in the hippocampal mossy fiber system. J. Neurochem., 70: 1950–1963. McGinty, J.F., Henriksen, S.J., Goldstein, A., Terenius, L. and Bloom, F.E. (1983) Dynorphin is contained within hippocampal mossy fibers: immunochemical alterations after kainic acid administration and colchicine-induced neurotoxicity. Proc. Natl. Acad. Sci. U.S.A., 80: 589–593. McLean, S., Rothman, R.B., Jacobson, A.E., Rice, K.C. and Herkenham, M. (1987) Distribution of opiate receptor subtypes and enkephalin and dynorphin immunoreactivity in the hippocampus of squirrel, guinea pig, rat, and hamster. J. Comp. Neurol., 255: 497–510. McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal synaptic enhancement and information storage within a distributed memory system. Trends Neurosci., 10: 408–415. McQuiston, A.R. and Colmers, W.F. (1996) Neuropeptide Y2 receptors inhibit the frequency of spontaneous but not miniature EPSCs in CA3 pyramidal cells of rat hippocampus. J. Neurophysiol., 76: 3159–3168. Miller, L.D., Petrozzino, J.J., Golarai, G. and Connor, J.A. (1996) Ca2+ release from intracellular stores induced by

afferent stimulation of CA3 pyramidal neurons in hippocampal slices. J. Neurophysiol., 76: 554–562. Min, M.Y., Rusakov, D.A. and Kullmann, D.M. (1998) Activation of AMPA, kainate, and metabotropic receptors at hippocampal mossy fiber synapses: role of glutamate diffusion. Neuron, 21: 561–570. Miyazaki, K., Ishizuka, T. and Yawo, H. (2005) Synapse-tosynapse variation of calcium channel subtype contributions in large mossy fiber terminals of mouse hippocampus. Neuroscience, 136: 1003–1014. Molnar, P. and Nadler, J.V. (2001) Synaptically-released zinc inhibits N-methyl-D-aspartate receptor activation at recurrent mossy fiber synapses. Brain Res., 910: 205–207. Monaghan, D.T. and Cotman, C.W. (1982) The distribution of [3H]kainic acid binding sites in rat CNS as determined by autoradiography. Brain Res., 252: 91–100. Monaghan, D.T. and Cotman, C.W. (1985) Distribution of N-methyl-D-aspartate-sensitive L-[3H]glutamate-binding sites in rat brain. J. Neurosci., 5: 2909–2919. Moore, K.A., Nicoll, R.A. and Schmitz, D. (2003) Adenosine gates synaptic plasticity at hippocampal mossy fiber synapses. Proc. Natl. Acad. Sci. U.S.A., 100: 14397–14402. Nadler, J.V. (2003) The recurrent mossy fiber pathway of the epileptic brain. Neurochem. Res., 28: 1649–1658. Nakazawa, K., Quirk, M.C., Chitwood, R.A., Watanabe, M., Yeckel, M.F., Sun, L.D., Kato, A., Carr, C.A., Johnston, D., Wilson, M.A. and Tonegawa, S. (2002) Requirement for hippocampal CA3 NMDA receptors in associative memory recall. Science, 297: 211–218. Neuman, R.S., Ben-Ari, Y., Gho, M. and Cherubini, E. (1988) Blockade of excitatory synaptic transmission by 6-cyano-7nitroquinoxaline-2,3-dione (CNQX) in the hippocampus in vitro. Neurosci. Lett., 92: 64–68. Nicoll, R.A., Castillo, P.E. and Weisskopf, M.G. (1994) The role of Ca2+ in transmitter release and long-term potentiation at hippocampal mossy fiber synapses. Adv. Second Messenger Phosphoprotein Res., 29: 497–505. Nicoll, R.A. and Schmitz, D. (2005) Synaptic plasticity at hippocampal mossy fibre synapses. Nat. Rev. Neurosci., 6: 863–876. Ogata, N. and Ueno, S. (1976) Mode of activation of pyramidal neurons by mossy fiber stimulation in thin hippocampal slices in vitro. Exp. Neurol., 53: 567–584. Ohishi, H., Akazawa, C., Shigemoto, R., Nakanishi, S. and Mizuno, N. (1995) Distributions of the mRNAs for L-2amino-4-phosphonobutyrate-sensitive metabotropic glutamate receptors, mGluR4 and mGluR7, in the rat brain. J. Comp. Neurol., 360: 555–570. Ohishi, H., Shigemoto, R., Nakanishi, S. and Mizuno, N. (1993) Distribution of the mRNA for a metabotropic glutamate receptor (mGluR3) in the rat brain: an in situ hybridization study. J. Comp. Neurol., 335: 252–266. Parra, P., Gulyas, A.I. and Miles, R. (1998) How many subtypes of inhibitory cells in the hippocampus? Neuron, 20: 983–993. Patrylo, P.R., van den Pol, A.N., Spencer, D.D. and Williamson, A. (1999) NPY inhibits glutamatergic excitation in the epileptic human DG. J. Neurophysiol., 82: 478–483.

129 Patz, S., Wirth, M.J., Gorba, T., Klostermann, O. and Wahle, P. (2003) Neuronal activity and neurotrophic factors regulate GAD-65/67 mRNA and protein expression in organotypic cultures of rat visual cortex. Eur. J. Neurosci., 18: 1–12. Pavlidis, P. and Madison, D.V. (1999) Synaptic transmission in pair recordings from CA3 pyramidal cells in organotypic culture. J. Neurophysiol., 81: 2787–2797. Pelkey, K.A., Lavezzari, G., Racca, C., Roche, K.W. and McBain, C.J. (2005) mGluR7 is a metaplastic switch controlling bidirectional plasticity of feedforward inhibition. Neuron, 46: 89–102. Pelkey, K.A., Topolnik, L., Lacaille, J.C. and McBain, C.J. (2006) Compartmentalized Ca2+ channel regulation at divergent mossy-fiber release sites underlies target celldependent plasticity. Neuron, 52: 497–510. Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki, G. (1997) Feed-forward and feed-back activation of the DG in vivo during dentate spikes and sharp wave bursts. Hippocampus, 7: 437–450. Podlogar, M. and Dietrich, D. (2006) Firing pattern of rat hippocampal neurons: a perforated patch clamp study. Brain Res., 1085: 95–101. Qian, J. and Noebels, J.L. (2005) Visualization of transmitter release with zinc fluorescence detection at the mouse hippocampal mossy fibre synapse. J. Physiol., 566: 747–758. Radian, R., Ottersen, O.P., Storm-Mathisen, J., Castel, M. and Kanner, B.I. (1990) Immunocytochemical localization of the GABA transporter in rat brain. J. Neurosci., 10: 1319–1330. Ramirez, M. and Gutierrez, R. (2001) Activity-dependent expression of GAD67 in the granule cells of the rat hippocampus. Brain Res., 917: 139–146. Ramon y Cajal, S. (1911) Histology of the nervous system of man and vertebrates. Translated from Spanish to French by L. Azoulay, from the French to English by N. Swanson and L.W. Swanson (1995) Oxford University Press, New York. Rao, A., Cha, E.M. and Craig, A.M. (2000) Mismatched appositions of presynaptic and postsynaptic components in isolated hippocampal neurons. J. Neurosci., 20: 8344–8353. Rattray, M. and Priestley, J.V. (1993) Differential expression of GABA transporter-1 messenger RNA in subpopulations of GABA neurones. Neurosci. Lett., 156: 163–166. Reid, C.A., Dixon, D.B., Takahashi, M., Bliss, T.V. and Fine, A. (2004) Optical quantal analysis indicates that long-term potentiation at single hippocampal mossy fiber synapses is expressed through increased release probability, recruitment of new release sites, and activation of silent synapses. J. Neurosci., 24: 3618–3626. Reid, C.A., Fabian-Fine, R. and Fine, A. (2001) Postsynaptic calcium transients evoked by activation of individual hippocampal mossy fiber synapses. J. Neurosci., 21: 2206–2214. Represa, A., Tremblay, E. and Ben-Ari, Y. (1987) Kainate binding sites in the hippocampal mossy fibers: localization and plasticity. Neuroscience, 20: 739–748. Reuter, H. (1995) Measurements of exocytosis from single presynaptic nerve terminals reveal heterogeneous inhibition by Ca(2+)-channel blockers. Neuron, 14: 773–779.

Ribak, C.E., Tong, W.M. and Brecha, N.C. (1996) GABA plasma membrane transporters, GAT-1 and GAT-3, display different distributions in the rat hippocampus. J. Comp. Neurol., 367: 595–606. Rivkees, S.A., Price, S.L. and Zhou, F.C. (1995) Immunohistochemical detection of A1 adenosine receptors in rat brain with emphasis on localization in the hippocampal formation, cerebral cortex, cerebellum, and basal ganglia. Brain Res., 677: 193–203. Rolls, E.T. (1989) Functions of neuronal networks in the hippocampus and neocortex in memory. In: Byrne J.H. and Berry W.O. (Eds.), Neural Models of Plasticity. Academic Press, San Diego, CA. Rolls, E.T. and Kesner, R.P. (2006) A computational theory of hippocampal function, and empirical tests of the theory. Prog. Neurobiol., 79: 1–48. Romo-Parra, H., Vivar, C., Maqueda, J., Morales, M.A. and Gutierrez, R. (2003) Activity-dependent induction of multitransmitter signaling onto pyramidal cells and interneurons of hippocampal area CA3. J. Neurophysiol., 89: 3155–3167. Rose, C.R., Blum, R., Pichler, B., Lepier, A., Kafitz, K.W. and Konnerth, A. (2003) Truncated TrkB-T1 mediates neurotrophin-evoked calcium signalling in glia cells. Nature, 426: 74–78. Rothstein, J.D., Martin, L., Levey, A.I., Dykes-Hoberg, M., Jin, L., Wu, D., Nash, N. and Kuncl, R.W. (1994) Localization of neuronal and glial glutamate transporters. Neuron, 13: 713–725. Ruiz, A., Fabian-Fine, R., Scott, R., Walker, M.C., Rusakov, D.A. and Kullmann, D.M. (2003) GABAA receptors at hippocampal mossy fibers. Neuron, 39: 961–973. Ruiz, A., Walker, M.C., Fabian-Fine, R. and Kullmann, D.M. (2004) Endogenous zinc inhibits GABA(A) receptors in a hippocampal pathway. J. Neurophysiol., 91: 1091–1096. Safiulina, V.F., Fattorini, G., Conti, F. and Cherubini, E. (2006) GABAergic signaling at mossy fiber synapses in neonatal rat hippocampus. J. Neurosci., 26: 597–608. Salin, P.A., Scanziani, M., Malenka, R.C. and Nicoll, R.A. (1996) Distinct short-term plasticity at two excitatory synapses in the hippocampus. Proc. Natl. Acad. Sci. U.S.A., 93: 13304–13309. Salin, P.A., Weisskopf, M.G. and Nicoll, R.A. (1995) A comparison of the role of dynorphin in the hippocampal mossy fiber pathway in guinea pig and rat. J. Neurosci., 15: 6939–6945. Sandler, R. and Smith, A.D. (1991) Coexistence of GABA and glutamate in mossy fiber terminals of the primate hippocampus: an ultrastructural study. J. Comp. Neurol., 303: 177–192. Sawada, S., Higashima, M. and Yamamoto, C. (1988) Kainic acid induces long-lasting depolarizations in hippocampal neurons only when applied to stratum lucidum. Exp. Brain Res., 72: 135–140. Sawada, S., Takada, S. and Yamamoto, C. (1983) Selective activation of synapses near the tip of drug-ejecting microelectrode, and effects of antagonists of excitatory amino acids in the hippocampus. Brain Res., 267: 156–160.

130 Scanziani, M., Gahwiler, B.H. and Thompson, S.M. (1993) Presynaptic inhibition of excitatory synaptic transmission mediated by alpha adrenergic receptors in area CA3 of the rat hippocampus in vitro. J. Neurosci., 13: 5393–5401. Scharfman, H.E., Goodman, J.H., Sollas, A.L. and Croll, S.D. (2002) Spontaneous limbic seizures after intrahippocampal infusion of brain-derived neurotrophic factor. Exp. Neurol., 174: 201–214. Scharfman, H.E. (2005) Bain-derived neurotrophic factor and epilepsy-A missing link? Epilepsy Curr., 5:83–88. Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A. (1990) Synaptic connections of dentate granule cells and hilar neurons: results of paired intracellular recordings and intracellular horseradish peroxidase injections. Neuroscience, 37: 693–707. Schmitz, D., Frerking, M. and Nicoll, R.A. (2000) Synaptic activation of presynaptic kainate receptors on hippocampal mossy fiber synapses. Neuron, 27: 327–338. Schwarzer, C., Kofler, N. and Sperk, G. (1998) Up-regulation of neuropeptide Y-Y2 receptors in an animal model of temporal lobe epilepsy. Mol. Pharmacol., 53: 6–13. Schwarzer, C. and Sperk, G. (1995) Hippocampal granule cells express glutamic acid decarboxylase-67 after limbic seizures in the rat. Neuroscience, 69: 705–709. Scott, R. and Rusakov, D.A. (2006) Main determinants of presynaptic Ca2+ dynamics at individual mossy fiber-CA3 pyramidal cell synapses. J. Neurosci., 26: 7071–7081. Sepkuty, J.P., Cohen, A.S., Eccles, C., Rafiq, A., Behar, K., Ganel, R., Coulter, D.A. and Rothstein, J.D. (2002) A neuronal glutamate transporter contributes to neurotransmitter GABA synthesis and epilepsy. J. Neurosci., 22: 6372–6379. Shigemoto, R., Kinoshita, A., Wada, E., Nomura, S., Ohishi, H., Takada, M., Flor, P.J., Neki, A., Abe, T., Nakanishi, S. and Mizuno, N. (1997) Differential presynaptic localization of metabotropic glutamate receptor subtypes in the rat hippocampus. J. Neurosci., 17: 7503–7522. Shigemoto, R., Kulik, A., Roberts, J.D., Ohishi, H., Nusser, Z., Kaneko, T. and Somogyi, P. (1996) Target-cell-specific concentration of a metabotropic glutamate receptor in the presynaptic active zone. Nature, 381: 523–525. Simmons, M.L. and Chavkin, C. (1996) k-Opioid receptor activation of a dendrotoxin-sensitive potassium channel mediates presynaptic inhibition of mossy fiber neurotransmitter release. Mol. Pharmacol., 50: 80–85. Skaggs, W.E., McNaughton, B.L., Wilson, M.A. and Barnes, C.A. (1996) Theta phase precession in hippocampal neuronal populations and the compression of temporal sequences. Hippocampus, 6: 149–172. Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E., Goodman, J.H., Sollas, A.L. and Martin, D.L. (1996) Basal expression and induction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal DG. J. Comp. Neurol., 373: 593–618. Smart, T.G., Xie, X. and Krishek, B.J. (1994) Modulation of inhibitory and excitatory amino acid receptor ion channels by zinc. Prog. Neurobiol., 42: 393–441.

Smith, M.A., Zhang, L.X., Lyons, W.E. and Mamounas, L.A. (1997) Anterograde transport of endogenous brain-derived neurotrophic factor in hippocampal mossy fibers. Neuroreport, 8: 1829–1834. Sperk, G., Schwarzer, C., Heilman, J., Furtinger, S., Reimer, R.J., Edwards, R.H. and Nelson, N. (2003) Expression of plasma membrane GABA transporters but not of the vesicular GABA transporter in dentate granule cells after kainic acid seizures. Hippocampus, 13: 806–815. Spruston, N., Jaffe, D.B., Williams, S.H. and Johnston, D. (1993) Voltage- and space-clamp errors associated with the measurement of electrotonically remote synaptic events. J. Neurophysiol., 70: 781–802. Spruston, N. and Johnston, D. (1992) Perforated patch-clamp analysis of the passive membrane properties of three classes of hippocampal neurons. J. Neurophysiol., 67: 508–529. Spruston, N., Lubke, J. and Frotscher, M. (1997) Interneurons in the stratum lucidum of the rat hippocampus: an anatomical and electrophysiological characterization. J. Comp. Neurol., 385: 427–440. Stengaard-Pedersen, K., Fredens, K. and Larsson, L.I. (1981) Enkephalin and zinc in the hippocampal mossy fiber system. Brain Res., 212: 230–233. Swanson, T.H., Drazba, J.A. and Rivkees, S.A. (1995) Adenosine A1 receptors are located predominantly on axons in the rat hippocampal formation. J. Comp. Neurol., 363: 517–531. Szabo, G., Kartarova, Z., Hoertnagl, B., Somogyi, R. and Sperk, G. (2000) Differential regulation of adult and embryonic glutamate decarboxylases in rat dentate granule cells after kainate-induced limbic seizures. Neuroscience, 100: 287–295. Tashiro, A., Dunaevsky, A., Blazeski, R., Mason, C.A. and Yuste, R. (2003) Bidirectional regulation of hippocampal mossy fiber filopodial motility by kainate receptors: a twostep model of synaptogenesis. Neuron, 38: 773–784. Taupin, P., Ben-Ari, Y. and Roisin, M.P. (1994a) Subcellular fractionation on Percoll gradient of mossy fiber synaptosomes: evoked release of glutamate, GABA, aspartate and glutamate decarboxylase activity in control and degranulated rat hippocampus. Brain Res., 644: 313–321. Taupin, P., Zini, S., Cesselin, F., Ben-Ari, Y. and Roisin, M.P. (1994b) Subcellular fractionation on Percoll gradient of mossy fiber synaptosomes: morphological and biochemical characterization in control and degranulated rat hippocampus. J. Neurochem., 62: 1586–1595. Terrian, D.M., Johnston, D., Claiborne, B.J., Ansah-Yiadom, R., Strittmatter, W.J. and Rea, M.A. (1988) Glutamate and dynorphin release from a subcellular fraction enriched in hippocampal mossy fiber synaptosomes. Brain Res. Bull., 21: 343–351. Thoenen, H. (1995) Neurotrophins and neuronal plasticity. Science, 270: 593–598. Thoenen, H. (2000) Neurotrophins and activity-dependent plasticity. Prog. Brain Res., 128: 183–191. Thompson, S.M. and Gahwiler, B.H. (1992) Comparison of the actions of baclofen at pre- and postsynaptic receptors in the rat hippocampus in vitro. J. Physiol., 451: 329–345.

131 Tokunaga, T., Miyazaki, K., Koseki, M., Mobarakeh, J.I., Ishizuka, T. and Yawo, H. (2004) Pharmacological dissection of calcium channel subtype-related components of strontium inflow in large mossy fiber boutons of mouse hippocampus. Hippocampus, 14: 570–585. Tonder, N., Kragh, J., Finsen, B.R., Bolwig, T.G. and Zimmer, J. (1994) Kindling induces transient changes in neuronal expression of somatostatin, neuropeptide Y, and calbindin in adult rat hippocampus and fascia dentata. Epilepsia, 35: 1299–1308. Tong, G., Malenka, R.C. and Nicoll, R.A. (1996) Long-term potentiation in cultures of single hippocampal granule cells: a presynaptic form of plasticity. Neuron, 16: 1147–1157. Tongiorgi, E., Armellin, M., Giulianini, P.G., Bregola, G., Zucchini, S., Paradiso, B., Steward, O., Cattaneo, A. and Simonato, M. (2004) Brain-derived neurotrophic factor mRNA and protein are targeted to discrete dendritic laminas by events that trigger epileptogenesis. J. Neurosci., 24: 6842–6852. Toth, K. and McBain, C.J. (1998) Afferent-specific innervation of two distinct AMPA receptor subtypes on single hippocampal interneurons. Nat. Neurosci., 1: 572–578. Toth, K., Suares, G., Lawrence, J.J., Philips-Tansey, E. and McBain, C.J. (2000) Differential mechanisms of transmission at three types of mossy fiber synapse. J. Neurosci., 20: 8279–8289. Trevin˜o, M. and Gutierrez, R. (2005) The GABAergic projection of the DG to hippocampal area CA3 of the rat: preand postsynaptic actions after seizures. J. Physiol., 567: 939–949. Trevin˜o, M., Vivar, C. and Gutierrez, R. (2007) b/g Oscillatory activity in the CA3 hippocampal area is depressed by aberrant GABAergic transmission from the DG after seizures. J. Neurosci., 27: 251–259. Tu, B., Timofeeva, O., Jiao, Y. and Nadler, J.V. (2005) Spontaneous release of neuropeptide Y tonically inhibits recurrent mossy fiber synaptic transmission in epileptic brain. J. Neurosci., 25: 1718–1729. Urban, N.N. and Barrionuevo, G. (1996) Induction of hebbian and non-hebbian mossy fiber long-term potentiation by distinct patterns of high-frequency stimulation. J. Neurosci., 16: 4293–4299. Urban, N.N. and Barrionuevo, G. (1998) Active summation of excitatory postsynaptic potentials in hippocampal CA3 pyramidal neurons. Proc. Natl. Acad. Sci. U.S.A., 95: 11450–11455. Vezzani, A. and Sperk, G. (2004) Overexpression of NPY and Y2 receptors in epileptic brain tissue: an endogenous neuroprotective mechanism in temporal lobe epilepsy? Neuropeptides, 38: 245–252. Vida, I. and Frotscher, M. (2000) A hippocampal interneuron associated with the mossy fiber system. Proc. Natl. Acad. Sci. U.S.A., 97: 1275–1280. Villacres, E.C., Wong, S.T., Chavkin, C. and Storm, D.R. (1998) Type I adenylyl cyclase mutant mice have impaired mossy fiber long-term potentiation. J. Neurosci., 18: 3186–3194.

Vivar, C. and Gutierrez, R. (2005) Blockade of the membranal GABA transporter potentiates GABAergic responses evoked in pyramidal cells by mossy fiber activation after seizures. Hippocampus, 15: 281–284. Vogt, K., Mellor, J., Tong, G. and Nicoll, R. (2000) The actions of synaptically released zinc at hippocampal mossy fiber synapses. Neuron, 26: 187–196. Vogt, K.E. and Regehr, W.G. (2001) Cholinergic modulation of excitatory synaptic transmission in the CA3 area of the hippocampus. J. Neurosci., 21: 75–83. Walker, M.C., Ruiz, A. and Kullmann, D.M. (2001) Monosynaptic GABAergic signaling from dentate to CA3 with a pharmacological and physiological profile typical of mossy fiber synapses. Neuron, 29: 703–715. Wang, J., Yeckel, M.F., Johnston, D. and Zucker, R.S. (2004) Photolysis of postsynaptic caged Ca2+ can potentiate and depress mossy fiber synaptic responses in rat hippocampal CA3 pyramidal neurons. J. Neurophysiol., 91: 1596–1607. Weisskopf, M.G., Castillo, P.E., Zalutsky, R.A. and Nicoll, R.A. (1994) Mediation of hippocampal mossy fiber long-term potentiation by cyclic AMP. Science, 265: 1878–1882. Weisskopf, M.G. and Nicoll, R.A. (1995) Presynaptic changes during mossy fibre LTP revealed by NMDA receptor-mediated synaptic responses. Nature, 376: 256–259. Weisskopf, M.G., Zalutsky, R.A. and Nicoll, R.A. (1993) The opioid peptide dynorphin mediates heterosynaptic depression of hippocampal mossy fibre synapses and modulates longterm potentiation. Nature, 365: 188. Wenzel, H.J., Cole, T.B., Born, D.E., Schwartzkroin, P.A. and Palmiter, R.D. (1997) Ultrastructural localization of zinc transporter-3 (ZnT-3) to synaptic vesicle membranes within mossy fiber boutons in the hippocampus of mouse and monkey. Proc. Natl. Acad. Sci. U.S.A., 94: 12676–12681. Werner, P., Voigt, M., Keinanen, K., Wisden, W. and Seeburg, P.H. (1991) Cloning of a putative high-affinity kainate receptor expressed predominantly in hippocampal CA3 cells. Nature, 351: 742–744. Westbrook, G.L. and Mayer, M.L. (1987) Micromolar concentrations of Zn2+ antagonize NMDA and GABA responses of hippocampal neurons. Nature, 328: 640–643. Widdowson, P.S. (1993) Quantitative receptor autoradiography demonstrates a differential distribution of neuropeptide-Y Y1 and Y2 receptor subtypes in human and rat brain. Brain Res., 631: 27–38. Wiebe, S.P. and Staubli, U.V. (1999) Dynamic filtering of recognition memory codes in the hippocampus. J. Neurosci., 19: 10562–10574. Williams, S. and Johnston, D. (1988) Muscarinic depression of long-term potentiation in CA3 hippocampal neurons. Science, 242: 84–87. Williams, S. and Johnston, D. (1989) Long-term potentiation of hippocampal mossy fiber synapses is blocked by postsynaptic injection of calcium chelators. Neuron, 3: 583–588. Williams, S. and Johnston, D. (1990) Muscarinic depression of synaptic transmission at the hippocampal mossy fiber synapse. J. Neurophysiol., 64: 1089–1097.

132 Williams, S.H. and Johnston, D. (1991) Kinetic properties of two anatomically distinct excitatory synapses in hippocampal CA3 pyramidal neurons. J. Neurophysiol., 66: 1010–1020. Wisden, W. and Seeburg, P.H. (1993) A complex mosaic of high-affinity kainate receptors in rat brain. J. Neurosci., 13: 3582–3598. Wu, L.G., Borst, J.G. and Sakmann, B. (1998) R-type Ca2+ currents evoke transmitter release at a rat central synapse. Proc. Natl. Acad. Sci. U.S.A., 95: 4720–4725. Xiang, Z., Greenwood, A.C., Kairiss, E.W. and Brown, T.H. (1994) Quantal mechanism of long-term potentiation in hippocampal mossy-fiber synapses. J. Neurophysiol., 71: 2552–2556. Yamamoto, C. (1972) Activation of hippocampal neurons by mossy fiber stimulation in thin brain sections in vitro. Exp. Brain Res., 14: 423–435. Yamamoto, C., Higashima, M., Sawada, S. and Kamiya, H. (1991) Quantal components of the synaptic potential induced in hippocampal neurons by activation of granule cells, and the effect of 2-amino-4-phosphonobutyric acid. Hippocampus, 1: 93–106. Yamamoto, C., Matsumoto, K. and Takagi, M. (1980) Potentiation of excitatory postsynaptic potentials during and after repetitive stimulation in thin hippocampal sections. Exp. Brain Res., 38: 469–477. Yamamoto, C., Sawada, S. and Kamiya, H. (1992) Enhancement of postsynaptic responsiveness during long-term potentiation of mossy fiber synapses in guinea pig hippocampus. Neurosci. Lett., 138: 111–114.

Yamamoto, C., Sawada, S. and Ohno-Shosaku, T. (1994) Suppression of hippocampal synaptic transmission by the spider toxin omega-agatoxin-IV-A. Brain Res., 634: 349–352. Yeckel, M.F., Kapur, A. and Johnston, D. (1999) Multiple forms of LTP in hippocampal CA3 neurons use a common postsynaptic mechanism. Nat. Neurosci., 2: 625–633. Yokoi, M., Kobayashi, K., Manabe, T., Takahashi, T., Sakaguchi, I., Katsuura, G., Shigemoto, R., Ohishi, H., Nomura, S., Nakamura, K., Nakao, K., Katsuki, M. and Nakanishi, S. (1996) Impairment of hippocampal mossy fiber LTD in mice lacking mGluR2. Science, 273: 645–647. Yoshino, M., Sawada, S., Yamamoto, C. and Kamiya, H. (1996) A metabotropic glutamate receptor agonist DCG-IV suppresses synaptic transmission at mossy fiber pathway of the guinea pig hippocampus. Neurosci. Lett., 207: 70–72. Zalutsky, R.A. and Nicoll, R.A. (1990) Comparison of two forms of long-term potentiation in single hippocampal neurons. Science, 248: 1619–1624. Zalutsky, R.A. and Nicoll, R.A. (1992) Mossy fiber long-term potentiation shows specificity but no apparent cooperativity. Neurosci. Lett., 138: 193–197. Zhang, G., Raol, Y.S., Hsu, F.C. and Brooks-Kayal, A.R. (2004) Long-term alterations in glutamate receptor and transporter expression following early-life seizures are associated with increased seizure susceptibility. J. Neurochem., 88: 91–101. Zieglga¨nsberger, W., French, E.D., Siggins, G.R. and Bloom, F.E. (1979) Opioid peptides may excite hippocampal pyramidal neurons by inhibiting adjacent inhibitory interneurons. Science, 205: 415–417.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 7

Development of cell and fiber layers in the dentate gyrus Michael Frotscher, Shanting Zhao and Eckart Fo¨rster Institute of Anatomy and Cell Biology, University of Freiburg, Albertstr. 17, D-79104 Freiburg, Germany

Abstract: This chapter deals with the laminated organization of the dentate gyrus, particularly with the molecular signals controlling its development. First, sites of granule cell generation, their modes and routes of migration are described. This is followed by an analysis of the molecular determinants governing the formation of a tightly packed granule cell layer that is normal in rodents and primates. Reelin, a protein of the extracellular matrix, plays an important role for the proper migration and lamination of the granule cells during development and for the maintenance of a laminated dentate gyrus in adulthood. Granule cell positioning is crucial for the laminated termination of commissural/associational fibers to the dentate gyrus, suggesting that the granule cells carry positional signals for these fibers. In contrast, not signals of the target cells but molecules of the extracellular matrix, such as hyaluronan, underlie the layer-specific termination of fibers from the entorhinal cortex. The molecular determinants controlling axonal pathfinding and target recognition of the profusely terminating cholinergic and GABAergic subcortical afferents still need to be elucidated. Keywords: granule cell; reelin; entorhinal fibers; commissural/associational fibers; cholinergic fibers et al., 2006; Christie and Cameron, 2006; Tashiro et al., 2006; Zhao et al., 2006). It has been a focus of these recent studies how postnatal granule cell formation can be modified by external stimuli and how the newly generated granule cells become integrated into the mature network of the dentate gyrus. In this chapter, we will not directly deal with the very interesting phenomenon of postnatal neurogenesis in the dentate gyrus and the factors that can modify this process, and the reader is referred to the articles cited above. Here, we will address fundamental issues of dentate gyrus development, such as where granule cells are generated, how they migrate to reach their destination, how the formation of a compact granule cell layer is controlled, and how major afferents to the dentate gyrus are instructed to form distinct, laminated

Introduction When compared to other brain regions, which largely develop prenatally, the dentate gyrus is characterized by an ongoing postnatal development. This enduring neurogenesis of dentate granule cells was first described in the early studies by Altman and coworkers (Altman, 1966; Altman and Das, 1966; Altman and Bayer, 1975; Bayer, 1980; Bayer and Altman, 1987) and has recently been further elaborated (e.g., Kempermann et al., 1997, 1998, 2003, 2004, review; Eriksson et al., 1998; Markakis and Gage, 1999; van Praag et al., 2002, 2005; Schmidt-Hieber et al., 2004; Lie et al., 2005; Aimone Corresponding author. Tel.: +(49) 761 203-5056; Fax: +(49) 761 203-5054; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63007-6

133

134

termination zones for the contact with defined segments of the granule cell’s dendritic arbor. It is likely that many of the signals involved in these processes apply to both early-generated granule cells and granule cells generated in the adult organism. As far as the development of afferent projections to the granule cells is concerned, we will deal with the projection from the entorhinal cortex to the dentate gyrus, commissural/ associational (C/A) projections, and subcortical projections, i.e., the cholinergic and GABAergic projections from the medial septum to the dentate gyrus.

Sites of granule cell generation Like the pyramidal cells of the hippocampus proper, the first granule cells are generated in the ventricular zone. From there they migrate to form the suprapyramidal blade of the dentate gyrus, which develops before the infrapyramidal blade. Later on in development, the secondary proliferation zone emerges within the future hilus. Obviously, this region contains stem cells from the ventricular zone that have retained their proliferative capacity. Cells derived from this secondary proliferation zone form the late developing infrapyramidal blade of the dentate gyrus and are the source of all ongoing postnatal granule cell neurogenesis. How do early- and late-generated granule cells reach their final destinations in the suprapyramidal and infrapyramidal blades of the dentate gyrus, respectively? No definitive answer to this question is possible, since real-time microscopy studies of newly generated, migrating granule cells are not yet available. From studies in the neocortex, we know that there are two major modes of neuronal migration, nuclear translocation and glia-guided neuronal migration (Nadarajah and Parnavelas, 2002). Movement of the nucleus within early-generated processes is generally assumed to be an early form of neuronal migration, allowing for a migration over short distances in the yet small, immature embryonic brain. Later on, when large distances need to be bridged, gliaguided migration becomes the dominant form of

cell movement. It is not clear which of the two forms predominates in the dentate gyrus and whether the temporal sequence assumed for neocortical neurons (Nadarajah and Parnavelas, 2002) holds true for dentate granule cells. It is known, however, that a radial glial scaffold forms in the dentate gyrus that persists long into the postnatal period, suggesting that radial glia-guided migration takes place in the dentate gyrus despite the relatively short distance that granule cells have to migrate from the secondary proliferation zone to their destination in the granule cell layer. Studies in mutants with defects in the reelin signaling cascade provide further evidence for a role of the radial glial scaffold in the proper migration of dentate granule cells. In these mutants, the radial glial scaffold is altered to a varying extent, and this is accompanied by migration defects of the granule cells (Fo¨rster et al., 2002, 2006a, b; Frotscher et al., 2003; Weiss et al., 2003).

Formation of a compact granule cell layer The lamination of neuronal cell bodies and fibers is a characteristic feature of cortical organization. In the dentate gyrus, where the granule cells form a densely packed cell layer, this lamination is particularly striking. Moreover, the granule cells show a clear, bipolar orientation with their dendrites extending into the molecular layer and their axons, the mossy fibers, invading the hilar region. This characteristic organization of the dentate gyrus and its predominant neurons, the granule cells, has been recognized since the first Golgi studies of the hippocampus (Fig. 1; Golgi, 1886; Koelliker, 1896; Ramo´n y Cajal, 1911). What are the molecular signals that control this well-ordered arrangement of the granule cells? Much of our knowledge about the formation of granule cell lamination derives from the study of various mouse mutants. In reeler mice lacking the extracellular matrix protein reelin, the granule cells do not form a compact cell layer but are scattered all over the dentate gyrus. Moreover, their dendrites have lost the uniform orientation toward the hippocampal fissure and extend to various directions (Stanfield and Cowan, 1979;

135

Fig. 1. Original drawing of the dentate gyrus and hippocampus proper by Camillo Golgi (Golgi, 1886). While Golgi was aware of the laminated, bipolar arrangement of the granule cells, he did not correctly draw the course of the mossy fibers. As is known from numerous more recent tracer studies, granule cell axons mainly run in stratum lucidum of CA3 and impinge on proximal dendritic segments of pyramidal neurons. (See Color Plate 7.1 in color plate section.)

Drakew et al., 2002). Granule cell axons do not form the normal compact projection to stratum lucidum of CA3 and appear defasciculated (Drakew et al., 2002). However, like in wildtype mice, mossy fibers in the reeler mutant respect the border with CA1. As mentioned above, the radial glial scaffold is also altered in mutants with defects in the reelin signaling pathway. How can one explain these multiple effects of reelin on the radial glial scaffold, granule cell migration, and granule cell orientation? An explanation is easier after a brief description of the reelin signaling cascade. Reelin is synthesized and secreted by early-generated Cajal-Retzius cells populating the marginal zone of the cortex, which in the dentate gyrus corresponds to the outer molecular layer (Del Rio et al., 1997). Thus, reelin

is not ubiquitously present in the tissue but at specific sites. It came as a surprise that lipoprotein receptors, the very low-density lipoprotein receptor (VLDLR) and the apolipoprotein E receptor 2 (ApoER2) are receptors for reelin. Double-knockouts lacking these two lipoprotein receptors show a reeler-like phenotype (Trommsdorff et al., 1999). Binding of reelin to its lipoprotein receptors results in the phosphorylation of disabled 1 (dab1), an adapter protein interacting with the intracellular domains of ApoER2 and VLDLR, respectively. The reelin signaling cascade eventually involves cytoskeletal proteins such as tau. For details and a survey of recent literature, the reader is referred to relevant review articles (Tissir and Goffinet, 2003; Fo¨rster et al., 2006a, b).

136

Recent studies have provided evidence for a dual role of reelin in the development of the dentate gyrus (Zhao et al., 2004; Fo¨rster et al., 2006a, b). It has been suggested that reelin is a positional signal for the directed growth of radial glial fibers in the dentate gyrus (Fo¨rster et al., 2002; Zhao et al., 2004). In the absence of reelin, its lipoprotein receptors or dab1, no regular radial glial scaffold is formed (Weiss et al., 2003). Moreover, glial fibrillary acid protein (GFAP)-positive glial cells of the hippocampus were found to express molecules of the reelin signaling cascade and showed a preference for reelin in a stripe choice assay (Fo¨rster et al., 2002; Frotscher et al., 2003). It was concluded that the migration defects in mutants of the reelin pathway were due, at least in part, to a malformation of the radial glial scaffold (Frotscher et al., 2003). A second effect of reelin appears to be on neurons directly. In the reeler mutant and in ApoER2 / / VLDLR / mutants, the marginal zone is densely populated by neurons, whereas it is almost cell-free in wildtype animals. Thus, reelin was regarded as a stop signal, preventing migrating neurons from invading the marginal zone. Sanada et al. (2004) have recently shown that phosphorylation of dab1 at

Fig. 2. Rescue of granule cell lamination in a slice culture of reeler dentate gyrus is achieved by coculturing with a wildtype culture providing a reelin-containing marginal zone. Two reeler cultures (rl / 1 and rl / 2) are cocultured with a rat hippocampal slice. A compact cell layer (arrow) has only formed in rl / 1, which was cocultured next to the outer molecular layer of the rat dentate gyrus containing reelin-synthesizing CajalRetzius cells. In rl / 2, which was cultured next to the stratum oriens of CA1, the reeler-specific loose distribution of neurons in the dentate gyrus is retained (arrowhead). Dashed lines represent borders between cultures. CA1, CA3, hippocampal regions CA1 and CA3; DG, dentate gyrus; g, granule cell layer. Scale bar: 200 mm (from Zhao et al., 2004, with permission). (See Color Plate 7.2 in color plate section.)

tyrosine 220 and 232 leads to a detachment of the migrating neuron from the radial fiber, thus terminating the migration process. By coculturing slices of reeler hippocampus with wildtype slices, such that a reelin-containing marginal zone was proximal to the reeler dentate gyrus, Zhao et al. (2004) were able to show that a compact granule cell layer could be rescued in the reeler dentate gyrus, leaving the molecular layer free of neurons. No rescue was achieved when a reeler slice was cocultured with stratum oriens of CA1, which did not contain reelinsynthesizing Cajal-Retzius cells (Fig. 2). These authors also showed that a regularly oriented radial glial scaffold could be rescued under these in vitro conditions. It remains to be determined whether reelin’s effect on radial glial cells or direct effect on neurons, or both, is crucial for the formation of a compact granule cell layer. Formation of afferent fiber lamination in the dentate gyrus With the exception of some subcortical afferents, major projections to the dentate gyrus show a preference for distinct layers. Fibers originating from neurons in layer 2 of the entorhinal cortex (entorhinal fibers) are known to terminate in the outer two-thirds of the molecular layer (Blackstad, 1958; Steward and Scoville, 1976; Amaral and Witter, 1995). C/A fibers, derived from hilar mossy cells projecting ipsilaterally and contralaterally, terminate in the inner molecular layer (Blackstad 1956; Deller et al., 1995), and there is no overlap of these two major projections to the dentate gyrus. In the following paragraphs, we will summarize what is known about the development of these projections and the factors determining their laminar specificities. Entorhinal afferents Fibers from the entorhinal cortex are among the first afferents to reach the dentate gyrus in the rodent (at E18/E19; Super and Soriano, 1994; Ceranik et al., 1999). This early arrival implies that their regional specificity and layer-specific termination in the outer two-thirds of molecular layer

137

cannot be controlled by their definitive target cells, the granule cells, which are mainly generated after birth (see above). How do the entorhinal axons find their long way from the entorhinal cortex across the hippocampal fissure and the subiculum? Recent studies have indicated that early-generated Cajal-Retzius cells located in the marginal zone give rise to an early projection to the entorhinal cortex, thereby providing a template for the later outgrowth of entorhinal fibers. In fact, intracellular labeling studies as well as retrograde and anterograde tracing substantiated this early dentate-entorhinal projection of Cajal-Retzius cells (Ceranik et al., 1999). Moreover, Cajal-Retzius cells seem to serve as transient targets of early arriving entorhinal fibers in the absence of the granule cells. Electron microscopic studies showed that entorhinal axons established synapses with reelin-immunoreactive Cajal-Retzius cells in the outer molecular layer, and selective removal of Cajal-Retzius cells prevented the ingrowth of entorhinal afferents (Del Rio et al., 1997; Frotscher et al., 2001). Collectively, these findings provide evidence for a role of Cajal-Retzius cells as a template for outgrowing entorhinal axons. However, reelin, secreted by Cajal-Retzius cells and forming a component of the extracellular matrix in the outer molecular layer, does not seem to be required for the layer-specific ingrowth of entorhinal axons. In the reeler mutant lacking reelin, entorhinal axons find their correct laminar position (Del Rio et al., 1997; Zhao et al., 2003). Recent studies have provided evidence that other molecules of the extracellular matrix such as hyaluronan and molecules bound to it play an important role in the laminar specificity of entorhinal fibers. Treating slice cultures of hippocampus with hyaluronidase, which degrades hyaluronan, abolishes the layer-specific termination of entorhinal fibers but not of C/A axons (Zhao et al., 2003).

Commissural/associational fibers C/A fibers represent ipsilateral and contralateral projections of hilar mossy cells (Ribak et al., 1985; Frotscher et al., 1991; Frotscher, 1992). C/A fibers

arrive in their termination zone, the inner molecular layer, after the fibers from the entorhinal cortex, i.e., at P2. This implies that many granule cells have already developed proximal dendrites for the contact with C/A fibers, and no transient target neurons, as in the case of the entorhinal projection, are required. In fact, the laminated projection of C/A fibers mirrors the laminated arrangement of the granule cells in wildtype animals. Accordingly, in mutants with defects in the reelin signaling pathway showing migration defects of the granule cells and no compact granular layer, the projection of C/A fibers is diffuse and not laminated (Deller et al., 1999; Gebhardt et al., 2002; Zhao et al., 2003). A diffuse termination of C/A fibers is most pronounced in the reeler mutant and in double-knockouts lacking ApoER2 and VLDLR, but is less prominent in single receptor mutants which only show minor migration defects of the granule cells. Zhao et al. (2003) used slice cocultures to study the development of the commissural projection to the dentate gyrus. By coculturing slices of reeler hippocampus and wildtype hippocampus and tracing the ‘‘commissural’’ projection developing under these in vitro conditions, they showed that the loose distribution of C/ A fibers is not a cell-autonomous effect of the reeler mutation on the C/A projection. C/A fibers originating from a reeler hippocampal culture terminated with correct laminar specificity in the inner molecular layer in a cocultured wildtype slice, whereas C/A fibers from a wildtype culture projecting to a reeler culture terminated profusely, thus reflecting the scattered distribution of the granule cells in the reeler culture (Fig. 3). Finally, rescue of granule cell lamination in a reeler culture by coculturing to a wildtype slice (see above) also rescued the laminated termination of C/A fibers (Zhao et al., 2004; Fo¨rster et al., 2006a, b). Together these data indicate that the granule cells carry positional signals for the laminated termination of C/A fibers. This contrasts to the guidance cues relevant for the proper termination of entorhinal fibers. As described, they are guided to their termination zone by transient pioneer neurons (Cajal-Retzius cells) and molecules of the extracellular matrix. Thus, as one would expect, in the reeler mutant with a scattered distribution of

138

the granule cells, the C/A fibers terminate diffusely, but the entorhino-dentate fibers terminate with laminar specificity in the outer molecular layer forming a sharp border (Zhao et al., 2003). Collectively, these findings show that different

types of molecular signal control laminar specificity of commissural and entorhinal fibers to the dentate gyrus, molecules associated with the cell membranes of the granule cells and extracellular matrix molecules, respectively. In light of these results, the temporal sequence of fiber ingrowth during development does not seem to determine the layered termination of entorhinal and commissural afferents as was hypothesized by Bayer and Altman (1987). This attractive temporal hypothesis was challenged by Frotscher and Heimrich (1993) when they used sequential cocultures of hippocampus with its afferent regions. Reversing the sequence of arrival of entorhinal and commissural fibers did not change their layerspecific termination in the dentate gyrus. Cholinergic and GABAergic afferents from the medial septum While both the entorhinal fibers and C/A fibers are characterized by their segregated, layer-specific termination in the dentate gyrus, no such strictly laminar termination is observed with subcortical Fig. 3. Cocultures of two hippocampal slices to allow for the formation of ‘‘commissural’’ projections, traced by anterogradely transported biocytin injected into the hilar region of one of the two hippocampal slices (sites of biocytin injections labeled by asterisks). (A) Coculturing of two wildtype slices results in the formation of a compact ‘‘commissural’’ projection in the inner molecular layer (black arrow). Open arrow labels ipsilateral mossy fiber projection in stratum lucidum of CA3. CA1, CA3, hippocampal regions CA1 and CA3; g, granule cell layer. (B) Coculture of wildtype (wt) hippocampus and reeler hippocampus. Biocytin-labeled ‘‘commissural’’ fibers from the wildtype culture terminate profusely in the reeler culture, thus reflecting the scattered distribution of their target neurons, the granule cells. Open arrow labels the mossy fiber projection in the wildtype culture. DG, dentate gyrus of the reeler culture with scattered granule cells; P, pyramidal layer in the wildtype culture; P1, P2, double pyramidal layer in the reeler culture. (C) ‘‘Commissural’’ fibers from a reeler culture form a compact projection in the inner molecular layer of the wildtype culture (black arrow), indicating that reeler commissural fibers project to the inner molecular layer as is normal, provided that their target cells, the granule cells, form a tightly packed layer. g, granule cell layer. Scale bar: 100 mm (applies to A–C) (modified from Zhao et al., 2003, with permission; copyright by the Society for Neuroscience). (See Color Plate 7.3 in color plate section.)

139

afferents, such as the cholinergic and GABAergic fibers from the medial septum and catecholaminergic brain stem fibers (the latter not dealt with in this chapter). The large cholinergic neurons of the medial septum/diagonal band (MS/ DB) complex project via the fimbria-fornix to the hippocampus where they terminate profusely in all hippocampal layers (Frotscher and Leranth, 1985). A concentration of cholinergic (choline acetyltransferase-immunoreactive) fibers is regularly observed in the cell body layers, in the dentate gyrus it is mainly in the subgranular zone. The fine, varicose cholinergic fibers establish synaptic contacts with a variety of postsynaptic structures including cell bodies, dendritic shafts, and spines (Frotscher and Leranth, 1986). The GABAergic septohippocampal fibers, not showing a laminated termination pattern like the cholinergic fibers, specifically form synapses with hippocampal GABAergic interneurons, thus serving a disinhibitory function (Freund and Antal, 1988). The cholinergic afferents that terminate on both the hippocampal principal cells and GABAergic interneurons do not have such a cellspecific termination. Little is known about the molecular signals controlling the development of the septohippocampal projection and the different target cell specificities of cholinergic and GABAergic septohippocampal fibers. Sema3C and its receptor neuropilin 2 have been suggested to play a role in the guidance of septal fibers to the hippocampus (Chedotal et al., 1998; Steup et al., 2000; Skutella and Nitsch, 2001). Cholinergic fibers express the p75 neurotrophin receptor (p75NTR), and nerve growth factor (NGF), one of the ligands for this receptor, is expressed by GABAergic interneurons in the hippocampus. Since GABAergic hippocampal neurons project to the septum, they may thus provide both a guiding scaffold and a neurotrophic source for the ingrowing cholinergic fibers. The diffuse termination of septohippocampal fibers would then result from the scattered distribution of GABAergic interneurons in the hippocampus and dentate gyrus. That hippocampal neurons projecting to the septum are involved in the guidance of septohippocampal fibers is supported by tracer studies on the development of these two

projections. The hippocampo-septal projection develops early, clearly before the septohippocampal projection (Linke et al., 1995) with hippocampal fibers reaching the septum around E16. Septal fibers arrive in the hippocampus 2–3 days later. Moreover, it has been shown that the septal fibers grow along hippocampo-septal axons (Linke and Frotscher, 1993). The determinants of the different target cell specificities of cholinergic and GABAergic septohippocampal fibers remain to be elucidated.

Functional considerations The functional significance of the laminated organization of the dentate gyrus is not known. In temporal lobe epilepsy associated with Ammon’s horn sclerosis, the lamination of the dentate gyrus is disrupted, and the granule cells do not form a tightly packed cell layer (granule cell dispersion, see Houser, 1990). These structural changes may indicate that a laminated dentate gyrus with a clear segregation of the input site (granule cell dendrites in the molecular layer) from the output site (granule cell axons in the hilus) is required for the proper function of this brain region. This hypothesis implies that structural alterations, for instance migration defects during development, underlie the development of the epileptic disorder. Recent studies of animal models of epilepsy have suggested a different scenario. Unilateral injections of the glutamate agonist kainate into the hippocampus of mice induced status epilepticus and, with some delay, a characteristic granule cell dispersion (Bouilleret et al., 1999; Heinrich et al., 2006), thus suggesting that the dispersion of granule cells is a result of seizure activity rather than being its source. How to explain, then, the loss of granule cell lamination following seizure activity? Haas et al. (2002) showed that the number of reelin-synthesizing neurons in tissue samples of hippocampus removed from epileptic patients for therapeutical reasons was dramatically decreased when compared to control samples. A similar loss of reelin-synthesizing cells was observed following unilateral kainate injection into the hippocampus of adult mice (Heinrich et al., 2006). Granule cell

140

dispersion thus seems to be accompanied by a loss of reelin-expressing cells both in human temporal lobe epilepsy and in an animal model of the disease. During development, reelin is required for the formation of a tightly packed granule cell layer. The findings from epileptic tissue indicate that reelin is also required for the maintenance of a compact granule cell layer during adulthood. Cajal-Retzius cells synthesizing reelin are very sensitive to glutamate agonists (Del Rio et al., 1997). Hence, epileptic seizures associated with excessive glutamate release may affect Cajal-Retzius cells, resulting in decreased reelin expression. In turn, loss of reelin, a stop signal for neurons, may allow the granule cells to leave their layer, resulting in granule cell dispersion. A role for reelin in the maintenance of a compact granule cell layer in adult animals was confirmed by chronic unilateral infusion of CR-50, a reelin-blocking antibody, into the hippocampus of adult, naı¨ ve animals. On the injection side, but not on the control side, granule cell dispersion developed. There was no granule cell dispersion when the reelin-blocking CR-50 antibody was replaced by nonspecific IgG (Heinrich et al., 2006). These data, as well as other recent studies showing that sprouted mossy fibers in epileptic animals may exert an inhibitory rather than the commonly assumed excitatory effect on hippocampal principal cells (Sloviter et al., 2006), indicate that structural changes in the hippocampus in epilepsy are likely to be the consequence of the disease rather than its cause. However, the functional significance of a strictly laminated dentate gyrus, first noticed by Golgi and his contemporaries, still requires explanation. Acknowledgments Supported by the German Research Foundation (SFB 505, TR-3) and Jung Foundation for Science and Research. References Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for adult neurogenesis in the encoding of time in new memories. Nat. Neurosci., 9: 723–727.

Altman, J. (1966) Autoradiographic and histological studies of postnatal neurogenesis. II. A longitudinal investigation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in infant rats, with special reference to postnatal neurogenesis in some brain regions. J. Comp. Neurol., 128: 431–474. Altman, J. and Bayer, S.A. (1975) Postnatal development of the hippocampal dentate gyrus under normal and experimental conditions. In: Isaacson R.L. and Pribram K.H. (Eds.), The Hippocampus, Vol. 2. Plenum Press, New York, pp. 95–122. Altman, J. and Das, G.D. (1966) Autoradiographic and histological studies of postnatal neurogenesis. I. A longitudinal investigation of the kinetics, migration and transformation of cells incorporating tritiated thymidine in neonate rats, with special reference to postnatal neurogenesis in some brain regions. J. Comp. Neurol., 126: 337–389. Amaral, D.G. and Witter, M.P. (1995) The hippocampal formation. In: Paxinos G. (Ed.), The Rat Nervous System (2nd Ed). Academic Press, New York, pp. 443–494. Bayer, S.A. (1980) Development of the hippocampal region in the rat. I. Neurogenesis examined with 3H-thymidine autoradiography. J. Comp. Neurol., 190: 87–114. Bayer, S.A. and Altman, J. (1987) Directions in neurogenetic gradients and patterns of anatomical connections in the telencephalon. Prog. Neurobiol., 29: 57–106. Blackstad, T.W. (1956) Commissural connections of the hippocampal region in the rat, with special reference to their mode of termination. J. Comp. Neurol., 105: 417–537. Blackstad, T.W. (1958) On the termination of some afferents to the hippocampus and fascia dentata: an experimental study in the rat. Acta Anat. (Basel), 35: 202–214. Bouilleret, V., Ridoux, V., Depaulis, A., Marescaux, C., Nehlig, A. and Le Gal La Salle, G. (1999) Recurrent seizures and hippocampal sclerosis following intrahippocampal kainate injection in adult mice: electroencephalography, histopathology and synaptic reorganization similar to mesial temporal lobe epilepsy. Neuroscience, 89: 717–729. Ceranik, K., Deng, J., Heimrich, B., Lu¨bke, J., Zhao, S., Fo¨rster, E. and Frotscher, M. (1999) Hippocampal CajalRetzius cells project to the entorhinal cortex: retrograde tracing and intracellular labelling studies. Eur. J. Neurosci., 11: 4278–4290. Chedotal, A., del Rio, J.A., Ruiz, M., He, Z., Borrell, V., de Castro, F., Ezan, F., Goodman, C.S., Tessier-Lavigne, M., Sotelo, C. and Soriano, E. (1998) Semaphorins III and IV repel hippocampal axons via two distinct receptors. Development, 125: 4313–4323. Christie, B.R. and Cameron, H.A. (2006) Neurogenesis in the adult hippocampus. Hippocampus, 16: 199–207. Del Rio, J.A., Heimrich, B., Borrell, V., Fo¨rster, E., Drakew, A., Alcantara, S., Nakajima, K., Miyata, T., Ogawa, M., Mikoshiba, K., Derer, P., Frotscher, M. and Soriano, E. (1997) A role for Cajal-Retzius cells and reelin in the development of hippocampal connections. Nature, 385: 70–74. Deller, T., Drakew, A. and Frotscher, M. (1999) Different primary target cells are important for fiber lamination

141 in the fascia dentata: a lesson from reeler mutant mice. Exp. Neurol., 156: 239–253. Deller, T., Nitsch, R. and Frotscher, M. (1995) Phaseolus vulgaris-Leucoagglutinin tracing of commissural fibers to the rat dentate gyrus: evidence for a previously unknown commissural projection to the outer molecular layer. J. Comp. Neurol., 352: 55–68. Drakew, A., Deller, T., Heimrich, B., Gebhardt, C., Del Turco, D., Tielsch, A., Fo¨rster, E., Herz, J. and Frotscher, M. (2002) Dentate granule cells in reeler mutants and VLDLR and ApoER2 knockout mice. Exp. Neurol., 176: 12–24. Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M., Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Neurogenesis in the adult human hippocampus. Nat. Med., 4: 1313–1317. Fo¨rster, E., Jossin, Y., Zhao, S., Chai, X., Frotscher, M. and Goffinet, A.M. (2006a) Recent progress in understanding the role of Reelin in radial neuronal migration, with specific emphasis on the dentate gyrus. Eur. J. Neurosci., 23: 901–909. Fo¨rster, E., Tielsch, A., Saum, B., Weiss, K.H., Johanssen, C., Graus-Porta, D., Mu¨ller, U. and Frotscher, M. (2002) Reelin, Disabled 1, and beta1 integrins are required for the formation of the radial glial scaffold in the hippocampus. Proc. Natl. Acad. Sci. U.S.A., 99: 13178–13183. Fo¨rster, E., Zhao, S. and Frotscher, M. (2006b) Laminating the hippocampus. Nat. Rev. Neurosci., 7: 259–267. Freund, T.F. and Antal, M. (1988) GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature, 336: 170–173. Frotscher, M. (1992) Application of the Golgi/electron microscopy technique for cell identification in immunocytochemical, retrograde labeling, and developmental studies of hippocampal neurons. Microsc. Res. Technol., 23: 306–323. Frotscher, M., Haas, C.A. and Fo¨rster, E. (2003) Reelin controls granule cell migration in the dentate gyrus by acting on the radial glial scaffold. Cereb. Cortex, 13: 634–640. Frotscher, M. and Heimrich, B. (1993) Formation of layerspecific fiber projections to the hippocampus in vitro. Proc. Natl. Acad. Sci. U.S.A., 90: 10400–10403. Frotscher, M. and Leranth, C. (1985) Cholinergic innervation of the rat hippocampus as revealed by choline acetyltransferase immunocytochemistry: a combined light and electron microscopic study. J. Comp. Neurol., 239: 237–246. Frotscher, M. and Leranth, C. (1986) The cholinergic innervation of the rat fascia dentata: identification of target structures on granule cells by combining choline acetyltransferase immunocytochemistry and Golgi impregnation. J. Comp. Neurol., 243: 58–70. Frotscher, M., Seress, L., Abraham, H. and Heimrich, B. (2001) Early generated Cajal-Retzius cells have different functions in cortical development. In: Miyan J.A., Thorndyke M., Beesley P.W. and Bannister C.M. (Eds.), Brain Stem Cells. BIOS Scientific Publishers Ltd., Oxford, pp. 43–49. Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E. (1991) The mossy cells of the fascia dentata: a comparative

study of their fine structure and synaptic connections in rodents and primates. J. Comp. Neurol., 312: 145–163. Gebhardt, C., Del Turco, D., Drakew, A., Tielsch, A., Herz, J., Frotscher, M. and Deller, T. (2002) Abnormal positioning of granule cells alters afferent fiber distribution in the mouse fascia dentata: morphologic evidence from reeler, apolipoprotein E receptor 2-, and very low density lipoprotein receptor knockout mice. J. Comp. Neurol., 445: 278–292. Golgi, C. (1886) Sulla fina anatomica degli organi centrali del sistema nervoso. Hoepli, Milan. Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G., Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for reelin in the development of granule cell dispersion in temporal lobe epilepsy. J. Neurosci., 22: 5797–5802. Heinrich, C., Nitta, N., Flubacher, A., Mu¨ller, M., Fahrner, A., Kirsch, M., Freiman, T., Suzuki, F., Depaulis, A., Frotscher, M. and Haas, C.A. (2006) Reelin deficiency and displacement of mature neurons, but not neurogenesis, underlie the formation of granule cell dispersion in the epileptic hippocampus. J. Neurosci., 26: 4701–4713. Houser, C.R. (1990) Granule cell dispersion in the dentate gyrus of humans with temporal lobe epilepsy. Brain Res., 535: 195–204. Kempermann, G., Gast, D., Kronenberg, G., Yamaguchi, M. and Gage, F.H. (2003) Early determination and long-term persistence of adult-generated new neurons in the hippocampus of mice. Development, 130: 391–399. Kempermann, G., Jessberger, S., Steiner, B. and Kronenberg, G. (2004) Milestones of neuronal development in the adult hippocampus. Trends Neurosci., 27: 447–452. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) More hippocampal neurons in adult mice living in an enriched environment. Nature, 386: 493–495. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Experience-induced neurogenesis in the senescent dentate gyrus. J. Neurosci., 18: 3206–3212. Koelliker, A.v. (1896) Handbuch der Gewebelehre des Menschen. Nervensystem des Menschen und der Thiere (6th Ed). Engelmann, Leipzig, Germany. Lie, D.C., Colamarino, S.A., Song, H.J., Desire, L., Mira, H., Consiglio, A., Lein, E.S., Jessberger, S., Lansford, H., Dearie, A.R. and Gage, F.H. (2005) Wnt signalling regulates adult hippocampal neurogenesis. Nature, 437: 1370–1375. Linke, R. and Frotscher, M. (1993) Development of the rat septohippocampal projection: tracing with DiI and electron microscopy of identified growth cones. J. Comp. Neurol., 332: 69–88. Linke, R., Pabst, T. and Frotscher, M. (1995) Development of the hippocamposeptal projection in the rat. J. Comp. Neurol., 351: 602–616. Markakis, E.A. and Gage, F.H. (1999) Adult-generated neurons in the dentate gyrus send axonal projections to field CA3 and are surrounded by synaptic vesicles. J. Comp. Neurol., 406: 449–460. Nadarajah, B. and Parnavelas, J.G. (2002) Modes of neuronal migration in the developing cerebral cortex. Nat. Rev. Neurosci., 3: 423–432.

142 van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer, T.D. and Gage, F.H. (2002) Functional neurogenesis in the adult hippocampus. Nature, 415: 1030–1034. van Praag, H., Shubert, T., Zhao, C. and Gage, F.H. (2005) Exercise enhances learning and hippocampal neurogenesis in aged mice. J. Neurosci., 25: 8680–8685. Ramo´n y Cajal, S. (1911) Histologie du Syste`me Nerveux de l’Homme et des Verte´bre´s, Vol. 2. Maloine, Paris. Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The development, ultrastructure and synaptic connections of the mossy cells of the dentate gyrus. J. Neurocytol., 14: 835–857. Sanada, K., Gupta, A. and Tsai, L.H. (2004) Disabled-1regulated adhesion of migrating neurons to radial glial fiber contributes to neuronal positioning during early corticogenesis. Neuron, 42: 197–211. Schmidt-Hieber, C., Jonas, P. and Bischofberger, J. (2004) Enhanced synaptic plasticity in newly generated granule cells of the adult hippocampus. Nature, 429: 184–187. Skutella, T. and Nitsch, R. (2001) New molecules for hippocampal development. Trends Neurosci., 24: 107–113. Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M. (2006) Kainic acid-induced recurrent mossy fiber innervation of dentate gyrus inhibitory interneurons: possible anatomical substrate of granule cell hyperinhibition in chronically epileptic rats. J. Comp. Neurol., 494: 944–960. Stanfield, B.B. and Cowan, W.M. (1979) The morphology of the hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol., 185: 393–422. Steup, A., Lohrum, M., Hamscho, N., Savaskan, N.E., Ninnemann, O., Nitsch, R., Fujisawa, H., Pu¨schel, A.W. and Skutella, T. (2000) Sema3C and netrin-1 differentially affect axon growth in the hippocampal formation. Mol. Cell Neurosci., 15: 141–155.

Steward, O. and Scoville, S.A. (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol., 169: 347–370. Super, H. and Soriano, E. (1994) The organization of the embryonic and early postnatal murine hippocampus. II. Development of entorhinal, commissural and septal connections studied with the lipophilic tracer DiI. J. Comp. Neurol., 344: 101–120. Tashiro, A., Sandler, V.M., Toni, N., Zhao, C. and Gage, F.H. (2006) NMDA-receptor-mediated, cell-specific integration of new neurons in adult dentate gyrus. Nature, 442: 929–933. Tissir, F. and Goffinet, A.M. (2003) Reelin and brain development. Nat. Rev. Neurosci., 4: 496–505. Trommsdorff, M., Gotthardt, M., Hiesberger, T., Shelton, J., Stockinger, W., Nimpf, J., Hammer, R.E., Richardson, J.E. and Herz, J. (1999) Reeler/disabled-like disruption of neuronal migration in knockout mice lacking the VLDL receptor and ApoE receptor 2. Cell, 97: 689–701. Weiss, K.H., Johanssen, C., Tielsch, A., Herz, J., Deller, T., Frotscher, M. and Fo¨rster, E. (2003) Malformation of the radial glial scaffold in the dentate gyrus of reeler mice, scrambler mice, and ApoER2/VLDLR-deficient mice. J. Comp. Neurol., 460: 56–65. Zhao, C., Teng, E.M., Summers Jr., R.G.,, Ming, G.L. and Gage, F.H. (2006) Distinct morphological stages of dentate granule neuron maturation in the adult mouse hippocampus. J. Neurosci., 26: 3–11. Zhao, S., Chai, X., Fo¨rster, E. and Frotscher, M. (2004) Reelin is a positional signal for the lamination of dentate granule cells. Development, 131: 5117–5125. Zhao, S., Fo¨rster, E., Chai, X. and Frotscher, M. (2003) Different signals control laminar specificity of commissural and entorhinal fibers to the dentate gyrus. J. Neurosci., 23: 7351–7357.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 8

Genetic regulation of dentate gyrus morphogenesis Guangnan Li and Samuel J. Pleasure Department of Neurology, Programs in Neuroscience, Developmental Biology, University of California at San Francisco, 1550 4th Street, Rock Hall Room 448c, San Francisco, CA 94158, USA

Abstract: The dentate gyrus is one of the small number of forebrain areas that have continued adult neurogenesis. During development the dentate gyrus acquires the capacity for neurogenesis by generating a new neurogenic stem cell niche at the border between the hilus and dentate granule cell layer. This is in distinction to the other prominent zone of continued neurogenesis in the subventricular zone where neurons are born in a structure directly descended from the mid-gestation subventricular zone. The ability to generate this newly formed dentate neurogenic niche is controlled by the action of a number of genes during prenatal and early postnatal development that regulate the fate, survival, migration, expansion, and differentiation of the cellular components of the dentate neurogenic niche. In this review, we provide an updated framework discussing the molecular steps and genes involved in these early stages of dentate gyrus formation. We previously described a molecular framework for dentate gyrus morphogenesis that can be associated with specific gene defects (Li, G., Pleasure, S.J. (2005). Dev. Neurosci., 27, 93–99), and here we add additional recently described molecular players and discuss this framework. the most significant mouse mutants with defects in dentate development.

Neurogenesis in the adult dentate gyrus requires a specialized stem cell niche in the subgranular zone at the boundary between the hilus and dentate granule cell layer. It is now clear that this niche includes quiescent multipotent stem cells, transitamplifying precursor cells, and immature neurons, and that the transitions between these stages of neuronal maturation are regulated by signaling molecules from families of proteins also powerfully involved in dentate development (Pozniak and Pleasure, 2006). We believe that the development of this neurogenic stem cell niche is first established in prenatal and early postnatal life. In this review we will first provide an overview of the neuroanatomy of dentate development and then turn to consideration of the phenotypes of some of

The neuroanatomy of the developing dentate gyrus Studies using classic neuroanatomic labeling methods revealed specialized aspects of dentate gyrus morphogenesis related to the development of the postnatal neurogenic niche (Nowakowski and Rakic, 1979, 1981; Eckenhoff and Rakic, 1984; Altman and Bayer, 1990a, b). The first granule neurons are born and the dentate precursor pool expands in a specialized region of subventricular zone that first becomes apparent adjacent to the dentate notch at mid-gestation. Quickly after that, the earliest dentate granule cells are born, migrate to the nascent dentate gyrus, and provide some of the scaffolding for the developing dentate granule cell layers. Contemporaneously, dividing precursor

Corresponding author. Tel.: þ 1 (415)502-5683; Fax: þ 1 (415)502-4335; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63008-8

143

144

cells destined to settle in the hilus and subgranular zone migrate from the dentate notch in a subpial route toward the dentate anlage (in developmental terms the anlage is the site of an incipient structure). En route and in the dentate anlage these cells continue to divide and produce additional granule neurons locally. The proliferative matrix in the hilus beneath the condensing dentate granule cell layers was termed the tertiary matrix by (Altman and Bayer, 1990a, b) and is the site of a massive production of granule neurons in the first week of postnatal life. During the first postnatal week, the hilus develops a radial glial matrix with radially oriented fibers traversing the dentate granule cell layer with their cell bodies in the subgranular zone and their endfeet attached to the pial surface. By P10, the tertiary matrix has reorganized so that most precursor cells are dividing in the subgranular zone at the border between the condensing granule cell layer and the now identifiable hilus. Granule cells continue to be born in this specialized niche, although in decreasing numbers, throughout adulthood. Our current growing understanding of the molecular steps of dentate development (Li and Pleasure, 2005) is founded on this strong neuroanatomic underpinning (see Fig. 1 for an overview of this process and some of the molecules involved).

The cortical hem in initial development of the dentate gyrus and hippocampus The initial formation of the hippocampus and dentate gyrus shares features with the development of the cortex as a whole since the hippocampal formation occupies the caudomedial — most area of the cortex and there is no evidence for a strict compartmental boundary between the hippocampus and neocortex. Recent studies have shown that gradients of transcription factor expression are critically important in specifying the cortex and driving development of four pallial compartments (Wilson and Rubenstein, 2000; Sur and Rubenstein, 2005). The hippocampus is derived from the most medial pallial domain, adjacent to the region forming the neocortex. Numerous recent studies demonstrate that the gradients of these transcription factors are controlled by diffusible cues

released from localized signaling centers (Grove et al., 1998; Fukuchi-Shimogori and Grove, 2001; Garel et al., 2003; Shimogori et al., 2004; Storm et al., 2006). The cortical hem is the most caudomedial of these signaling centers and occupies a small zone of neuroepithelium adjacent to the region of neuroepithelium where the dentate gyrus arises (Fig. 1). The cortical hem releases Wnt and BMP family members and a variety of axon guidance molecules that are candidates for regulating the development of the dentate gyrus and hippocampus at early stages (Furuta et al., 1997; Grove et al., 1998; Bagri et al., 2002). In addition, mice with early deletion of the cortical hem had dramatic loss of cortical volume and no identifiable hippocampus, implying that factors from the hem are necessary for cortical expansion (Monuki et al., 2001; Yoshida et al., 2006). Interestingly, the cortical hem was also recently shown to be the source of most Cajal-Retzius neurons that are generated in a wave of neurogenesis very early in cortical development and then migrate tangentially to cover the cerebral cortex in the most superficial cortical layer, called the marginal zone (Takiguchi-Hayashi et al., 2004; Bielle et al., 2005; Yoshida et al., 2006; Zhao et al., 2006). Thus, many Cajal-Retzius cells, critical regulators of cortical migration via the release of reelin, are actually the earliest born neurons from the medial pallium and are produced from the neuroepithelium immediately adjacent to the region of neuroepithelium that will generate the dentate granule neurons. Recent studies have now shown that the tangential migration of these neurons is controlled by SDF1 produced by the meninges. and this allows these cells to maintain their superficial cortical position (Borrell and Marin, 2006; Paredes et al., 2006). Interestingly the same secreted chemokine (SDF-1) later controls the migration of cells to the dentate gyrus (Bagri et al., 2002; Lu et al., 2002).

Transcription factors pattern the cortex and hippocampus Mice with mutations in transcription factors normally expressed at their highest levels in the most caudomedial cortex have been found to have

Wnts

DG

Lhx2/FoxG1/Emx2/Lef1

Hip

Secreted Factors

C

D

Lhx5

Bmps

Hem CP

B

p73

A

Transcription Factors

precursors

granule cells Cajal-Retzius cells hippocampus fissure meninges

dentate notch

E11.5

fimbria

E13.5

E17.5

P4

Fig. 1. Schematic diagram showing the anatomic events and some of the genes involved in dentate morphogenesis. (A) At E11.5 the cortical hem (Hem) is adjacent to the dentate neuroepithelium (DG) and the development of both are regulated by Wnts, while the hem and choroid plexus (CP) are also regulated by BMPs. The expression domains for Wnts and BMPs are shown with arrows to the left of the neuroepithelium. The dentate and hippocampal neuroepithelia (Hip) are regulated by the function of a number of transcription factors (Foxg1, Lhx2, Emx2, and Lef1) discussed in the text, while hem development is controlled by p73 and Lhx5. The expression domains for transcription factors are shown to the right of the neuroepithelium and their extent indicated with arrows. (B) At E13.5 the initial migration of precursors is underway from the dentate notch to the forming dentate gyrus, this migration may be targeted in part by SDF1 expressed in the meninges. By this time, Cajal-Retzius cells, derived from the hem, are already present in the marginal zone and their position is maintained by SDF1 signaling. (C) By E17.5 precursor cells and granule cells have begun to mix, forming the tertiary matrix in the nascent hilus and there is continued dentate precursor migration along the subpial migratory course. (D) By P4 the radial reorganization of the dentate is largely accomplished with condensing of the granule cell layers and positioning of precursors in the subgranular zone.

145

146

major defects in hippocampal development. Emx2 is a homeobox gene expressed in a high caudomedial to low anterolateral gradient, and mice with Emx2 mutations have reduced size of caudal cortical structures like the hippocampus and visual cortex with compensatory increases in anterior cortical structures (Bishop et al., 2003; Muzio and Mallamaci, 2003), while mice with overexpression of Emx2 have the reciprocal phenotype, with expansion of caudal cortical structures at the expense of anterior structures (Hamasaki et al., 2004). Consistent with their loss of posterior structures, Emx2 null mice also lack a clearly distinguishable dentate gyrus (the dentate gyrus forms from almost the most embryologically caudal portion of the cortical neuroepithelium) and have reductions in the production of Cajal-Retzius cells from the cortical hem (the most caudal region of cortical neuroepithelium) (Pellegrini et al., 1996; Yoshida et al., 1997; Mallamaci et al., 2000; Shinozaki et al., 2002). Another interesting patterning molecule is the homeobox gene Lhx5, which in the cortex is expressed solely in the cortical hem and Cajal-Retzius cells derived from the hem. Mouse mutants for Lhx5 lack hem development entirely and have dramatic secondary hippocampal defects due to the loss of hem-derived patterning signals (Zhao et al., 1999), as well as defects in hippocampal dependent learning (Paylor et al., 2001). In addition, the transcription factor p73, which is selectively expressed in the cortical hem and CajalRetzius cells derived from it also appear to have major hippocampal and cortical proliferation defects due at least in large part due to disruption of normal cortical hem development, although this phenotype is less severe than the Lhx5 phenotype with the most severe defects being dramatic hypoplasia of the lower blade of the dentate gyrus (Yang et al., 2000; Meyer et al., 2004). Loss of function mutations of transcription factors expressed in the hem lead to hem defects (and secondary hippocampal defects, e.g. the Lhx5 mutant mice), but in contrast, mutation of transcription factors normally excluded from the cortical hem may result in phenotypes that seem to be almost the opposite. Foxg1 is a forkhead transcription factor and Lhx2 a homeobox gene, both of which are normally excluded from the hem, and mutation of

either of these leads to cortical hem expansion (Porter et al., 1997; Bulchand et al., 2001; Monuki et al., 2001; Hanashima et al., 2004; Muzio and Mallamaci, 2005; Zhao et al., 2006). Amazingly, in the case of Foxg1, mutant mice have a massive expansion of all medial cortical structures so that essentially all residual cortical structures have molecular markers consistent with a hippocampal origin and there is no discernible neocortex at all (Muzio and Mallamaci, 2005).

The role of hem-derived signals in hippocampal and dentate development The studies discussed above indicate the central importance of the cortical hem in controlling hippocampal and dentate development at the earliest stages, but what are the factors that the hem makes? In addition to producing Cajal-Retzius cells, the cortical hem is a source of multiple secreted Wnt proteins. The Wnts are a large family (19 ligands in mice) of secreted glycoprotein molecules that act potently during neural development (Ciani and Salinas, 2005). The Wnt signaling pathway is very complex and significant discussion of it is beyond the scope of this review, but most of the hippocampal phenotypes of Wnt signaling mutants or Wnt ligands thus far have been most clearly associated with defects in the so-called ‘‘canonical’’ Wnt signaling pathway. Both of the most studied Wnt signaling pathways involve Wnt ligand binding to one of the cognate receptor types, the Frizzled genes, but after this the two pathways diverge substantially. The canonical pathway requires an additional receptor from the LRP family (LRP6) to stabilize beta-catenin, which is then transported to the nucleus where it acts with the family of Tcf/Lef transcription factors to drive alterations in gene expression. The non-canonical pathway has some transcriptional output but is also involved in local cytoskeletal rearrangements and axon guidance decisions and also utilizes newly recognized receptors from the Ryk family in conjunction with Frizzled receptors. During early cortical development the earliest Wnt ligand expressed in the cortical hem is Wnt3a (Grove et al., 1998) and Wnt3a mutants entirely

147

lack discernible hippocampal morphogenesis, most prominently due to a failure to expand neural precursors in the entire medial pallium (Lee et al., 2000). Mutants in one of the downstream transcription factors, Lef1, which is expressed exclusively in the medial pallium in the cortex, almost completely lack dentate granule neurons (Galceran et al., 2000; Zhou et al., 2004). Interestingly, mice with mutations in the canonical signaling pathway co-receptor LRP6, which are hypomorphic for multiple Wnt signaling pathways during development, have a very similar phenotype to that of the Lef1 mutants (Zhou et al., 2004). The detailed analysis of both the Lef1 and LRP6 mutant phenotypes is most consistent with a relatively specific deficit in expansion of early dentate precursors leading to a failure to populate the dentate with stem cells, and also a related defect in the prenatal radial glial scaffolding (Zhou et al., 2004). In addition to the production of Wnts, the cortical hem also produces members of the extended BMP family, although they are produced in higher quantities by the choroid plexus epithelium adjacent to the hem (Furuta et al., 1997; Grove et al., 1998; Shimogori et al., 2004). Previous studies using BMP ligand mutants demonstrated that there are defects in the development of the medial cortical wall, but early studies did not fully evaluate cortical development in these mutants (Furuta et al., 1997). These studies were quite difficult though, because of the large number of BMPs expressed at the cortical midline. However, a recent analysis of conditional mutants, lacking all BMP receptor signaling during cortical development after E10, showed that BMP signaling was particularly important in development of the choroid plexus (Hebert et al., 2002). However, previous studies showing that BMP signaling is important earlier in regulating Foxg1 expression (Furuta et al., 1997; Monuki et al., 2001) do demonstrate that earlier in cortical development as the cortex is initially patterned, there is likely to be an important role for BMPs. This may not have been apparent in the analysis of the BMP receptor conditional mice because the Cre-driver line used was Foxg1 and this does not start expressing until the telencephalon is already partially specified (Hebert et al., 2002). The important early role of BMPs in cortical patterning was confirmed in a recent study

demonstrating that Twisted Gastrulation mutants (combined with BMP4) have no dorsomedial telencephalic structures (Zakin and De Robertis, 2004). Since Twisted Gastrulation is an important extracellular mediator that enhances BMP function (thus loss of Twisted Gastrulation would mimic loss of function for BMP signaling), this establishes a critical early role for BMPs in the developing medial pallium and hippocampus, whether other later roles for the BMPs can be distinguished remains to be seen.

Reforming the neurogenic niche in the dentate gyrus Once the dentate neuroepithelium is specified by the actions of the signals discussed above, the next step is the process of moving a cohort of neurogenic stem cells with continued capacity for expansion to the developing dentate gyrus. Previous studies have shown that dentate precursors, along with earlyborn granule neurons, migrate in a subpial stream at the boundary between the fimbria and the meninges during mid-gestation (Nowakowski and Rakic, 1979; Altman and Bayer, 1990b; Bagri et al., 2002). Genes that control either migration of these cells directly or the organization of the radial glial network in this region disrupt morphogenesis of the dentate. One recent example of a relatively specific medial cortical radial glial defect comes from conditional FGFR1 mutants. These mice have diminished precursor proliferation in the dentate and defects in the radial glial scaffolding projecting to the dentate (Ohkubo et al., 2004). Even more interesting are the results of a recent in depth analysis of these mice showing a defect in radial glial precursor somal translocation in the entire medial cortex (Smith et al., 2006). Since one of the main distinctions between dentate development and that of the rest of the cortex is that stem cell/precursors must relocate to the developing dentate from their original location in the neuroepithelium, this implies that these mice may actually have a failure to properly seed the dentate gyrus with appropriate numbers of dentate precursors. This is somewhat similar to the canonical Wnt signaling dentate defects seen from loss of Lef1 and LRP6, which have proliferative failure of dentate

148

precursors and defects in the radial glial network as well, but in these Wnt mutants it is not possible to exclude earlier patterning defects (Galceran et al., 2000; Zhou et al., 2004). We also recently described a different Wnt signaling mechanism leading to loss of some dentate precursors in the Frizzled9 mutant mice, which have increased apoptotic cell death of putative dentate precursors as they migrate prenatally, leading to an adult defect in granule cell numbers (Zhao et al., 2005). This is a quite different phenotype than in the Lef1 and LRP6 mutants, and we are continuing to address whether this phenotype is due to a defect in non-canonical signaling or compensation by other Frizzleds that uncovers a later developmental role for canonical Wnt signaling in other aspects of dentate development. The subpial migratory route used by the dentate precursors and immature granule cells to reach the forming dentate gyrus is also a region with very strong expression of the chemokine ligand SDF1, which may regulate migration of these cells (Bagri et al., 2002; Lu et al., 2002). Indeed, we and others demonstrated that SDF1 serves as a direct chemoattractant for cells migrating in the subpial migratory route and that the dentate gyrus is depopulated of precursors without SDF-1, either because of a migratory deficit or because of a proliferative failure of the precursors (Bagri et al., 2002; Lu et al., 2002). Interestingly, mice with defects in integrin receptors or signaling systems downstream of integrins in the migrating cells (e.g. focal adhesion kinase — Fak) also have defects in the organization of the subpial migratory pathway frequently resulting in major defects of the lower blade of the dentate (Forster et al., 2002; Beggs et al., 2003; Niewmierzycka et al., 2005).

Mutants with defects in dentate granule cell differentiation During dentate gyrus development, from the earliest embryonic stages to adulthood, the final step is cell type-specific differentiation into neurons of the appropriate phenotype with the appropriate specialized markers [in the case of granule cells the most specific marker is the divergent homeobox gene Prox1 — (Pleasure et al., 2000)]. Recent work

has shown that this complex process depends on intersecting networks of interacting transcription factors regulated by combinations of extracellular signaling cues. It has been difficult thus far to determine the specific roles of these signaling pathways in granule cell differentiation because of their pleiotrophic effects during earlier development and at different stages of dentate morphogenesis. Mice with mutations in NeuroD, a basic helixloop-helix transcription factor, have quite selective defects in the differentiation of dentate granule neurons without affecting other hippocampal neurons (Miyata et al., 1999; Liu et al., 2000). This selectivity is likely to be due to the expression of Math2 and NeuroD2 (two closely related family members) in essentially all other cortical regions (Pleasure et al., 2000). In NeuroD mutants there are still small numbers of neurons produced that eventually express Prox1 at low levels, and these cells still project axons toward CA3 (the appropriate target) and express Calbindin (another marker of mature granule neurons) (Liu et al., 2000). This indicates that there are likely to be other regulatory genes involved in specifying granule neuron fate and differentiation in addition to NeuroD.

Regulation of dentate granule cell layer density The normal granule cell layer of mammals is a discrete, compact layer with few granule cells located outside its narrow boundaries. Reelin mutants (and mice with mutations in Reelin receptors — VLDLR and ApoER3 — or downstream signaling molecules — Disabled) have a dispersed granule cell layer, and a major defect in proliferative capacity of dentate progenitors, but do not appear to have a primary migratory defect because the earliest granule cells find the dentate gyrus just as they would under normal conditions (Forster et al., 2002; Frotscher et al., 2003; Zhao et al., 2004). The abnormalities in the Reelin mutant may be due to direct effects of Reelin on granule neurons and their organization, but it is also likely that the mechanisms are related to a failure to properly organize the radially organized radial glial network in the early postnatal dentate gyrus. This is because there are also direct effects of Reelin on the radial glial

Fig. 2. NeuroD mutants have radial glial scaffolding defects in the postnatal dentate. We examined the organization of the GFAP+ radially oriented dentate precursors usually residing in the subgranular zone in NeuroD mutant mice. In control mice, Prox1 clearly marks the dentate granule cell layer with the adjacent subgranular zone that serves as the focus of radially oriented GFAP+ fibers. In contrast, in NeuroD mutant mice there is only a small cap of Prox1+ granule cells in the dentate of P21 mutants and there are essentially no GFAP+ processes present in the mutant dentate. (See Color Plate 8.2 in color plate section.)

149

150

cells themselves (Frotscher et al., 2003). Interestingly, recent studies of intractable epilepsy patients (Haas et al., 2002) and animal models of epilepsy (Heinrich et al., 2006) showed that dentate granule cell layer dispersion is associated with loss of Reelinexpressing cells in the dentate marginal zone, implying a crucial role for Reelin in maintaining dentate architecture in adulthood. However, since there was no correlation with severity or duration of epilepsy comparing this group to the other patients with pharmacologically tractable or other forms of temporal lobe epilepsy one can not rule out the role of preexisting developmental anomalies. Thus, in some subsets of epilepsy patients, granule cell dispersion in temporal lobe epilepsy could be due to a neurodevelopmental defect that led to a loss of Reelin, or due to abnormalities in Cajal-Retzius cell genesis or distribution (Haas et al., 2002). We recently also noted that the residual granule cells in NeuroD mutant mice are diffusely distributed as compared to the residual granule cells found in Lef1 mutants. This implies that the few granule cells made when NeuroD is missing either have a cell autonomous defect in compaction or that their might be a non-autonomous defect similar to the radial glial phenotype in Reelin mice. In other words, newborn granule cells lacking NeuroD may fail to properly organize the forming dentate radial glial scaffolding and secondarily cause a granule cell dispersion phenotype. This would presumably be non-autonomous because radial glial cells do not express NeuroD (whereas they do express Reelin receptors) and might point to a role for newborn granule cells in organizing their own lamination. If this is true, we would predict that NeuroD mice would display radial glial defects even though these cells never express NeuroD. While it is difficult to resolve this issue without more specifically expressed markers and conditional alleles, we have found that NeuroD mice do have radial glial scaffolding defects (Fig. 2).

Continued use of developmental signaling systems in the adult dentate gyrus Recent studies have begun to establish that cues regulating dentate development continue to function

in adult neurogenesis (Pozniak and Pleasure, 2006). Sonic Hedgehog (Shh) is a crucial ligand in the establishment of the dentate neurogenic niche during development (Machold et al., 2003) and clearly acts directly on migratory precursors in the subpial migration pathway (Ahn and Joyner, 2005). It has now become apparent that Shh regulates the proliferation of rapidly dividing adult dentate precursors and quiescent stem cells as well (Lai et al., 2003; Ahn and Joyner, 2005). In addition, Wnts have now been shown to regulate the proliferation of committed dentate precursors and Wnt signaling is required for adult neurogenesis (Lie et al., 2005). It is likely that over the next few years roles will become established for most, if not all, developmental modulators of dentate development in adult neurogenesis as well.

Acknowledgments The authors would like to acknowledge the other members of the Pleasure Lab for helpful discussions. This work was funded by K02 MH074958, R01 MH066084 and P50 NS35902 to S.J.P.

References Ahn, S. and Joyner, A.L. (2005) In vivo analysis of quiescent adult neural stem cells responding to Sonic hedgehog. Nature, 437: 894–897. Altman, J. and Bayer, S.A. (1990a) Migration and distribution of two populations of hippocampal granule cell precursors during the perinatal and postnatal periods. J. Comp. Neurol., 301: 365–381. Altman, J. and Bayer, S.A. (1990b) Mosaic organization of the hippocampal neuroepithelium and the multiple germinal sources of dentate granule cells. J. Comp. Neurol., 301: 325–342. Bagri, A., Gurney, T., He, X., Zou, Y.R., Littman, D.R., Tessier-Lavigne, M. and Pleasure, S.J. (2002) The chemokine SDF1 regulates migration of dentate granule cells. Development, 129: 4249–4260. Beggs, H.E., Schahin-Reed, D., Zang, K., Goebbels, S., Nave, K.A., Gorski, J., Jones, K.R., Sretavan, D. and Reichardt, L.F. (2003) FAK deficiency in cells contributing to the basal lamina results in cortical abnormalities resembling congenital muscular dystrophies. Neuron, 40: 501–514. Bielle, F., Griveau, A., Narboux-Neme, N., Vigneau, S., Sigrist, M., Arber, S., Wassef, M. and Pierani, A. (2005) Multiple origins of Cajal-Retzius cells at the borders of the developing pallium. Nat. Neurosci., 8: 1002–1012.

151 Bishop, K.M., Garel, S., Nakagawa, Y., Rubenstein, J.L. and O’leary, D.D. (2003) Emx1 and Emx2 cooperate to regulate cortical size, lamination, neuronal differentiation, development of cortical efferents, and thalamocortical pathfinding. J. Comp. Neurol., 457: 345–360. Borrell, V. and Marin, O. (2006) Meninges control tangential migration of hem-derived Cajal-Retzius cells via CXCL12/ CXCR4 signaling. Nat. Neurosci., 9: 1284–1293. Bulchand, S., Grove, E.A., Porter, F.D. and Tole, S. (2001) LIM-homeodomain gene Lhx2 regulates the formation of the cortical hem. Mech. Dev., 100: 165–175. Ciani, L. and Salinas, P.C. (2005) WNTs in the vertebrate nervous system: from patterning to neuronal connectivity. Nat. Rev. Neurosci., 6: 351–362. Eckenhoff, M.F. and Rakic, P. (1984) Radial organization of the hippocampal dentate gyrus: a Golgi, ultrastructural, and immunocytochemical analysis in the developing rhesus monkey. J. Comp. Neurol., 223: 1–21. Forster, E., Tielsch, A., Saum, B., Weiss, K.H., Johanssen, C., Graus-Porta, D., Muller, U. and Frotscher, M. (2002) Reelin, Disabled 1, and beta 1 integrins are required for the formation of the radial glial scaffold in the hippocampus. Proc. Natl. Acad. Sci. U.S.A., 99: 13178–13183. Frotscher, M., Haas, C.A. and Forster, E. (2003) Reelin controls granule cell migration in the dentate gyrus by acting on the radial glial scaffold. Cereb. Cortex, 13: 634–640. Fukuchi-Shimogori, T. and Grove, E.A. (2001) Neocortex patterning by the secreted signaling molecule FGF8. Science, 294: 1071–1074. Furuta, Y., Piston, D.W. and Hogan, B.L. (1997) Bone morphogenetic proteins (BMPs) as regulators of dorsal forebrain development. Development, 124: 2203–2212. Galceran, J., Miyashita-Lin, E.M., Devaney, E., Rubenstein, J.L. and Grosschedl, R. (2000) Hippocampus development and generation of dentate gyrus granule cells is regulated by LEF1. Development, 127: 469–482. Garel, S., Huffman, K.J. and Rubenstein, J.L. (2003) Molecular regionalization of the neocortex is disrupted in Fgf8 hypomorphic mutants. Development, 130: 1903–1914. Grove, E.A., Tole, S., Limon, J., Yip, L. and Ragsdale, C.W. (1998) 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: 2315–2325. Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G., Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for reelin in the development of granule cell dispersion in temporal lobe epilepsy. J. Neurosci., 22: 5797–5802. Hamasaki, T., Leingartner, A., Ringstedt, T. and O’leary, D.D. (2004) EMX2 regulates sizes and positioning of the primary sensory and motor areas in neocortex by direct specification of cortical progenitors. Neuron, 43: 359–372. Hanashima, C., Li, S.C., Shen, L., Lai, E. and Fishell, G. (2004) Foxg1 suppresses early cortical cell fate. Science, 303: 56–59. Hebert, J.M., Mishina, Y. and Mcconnell, S.K. (2002) BMP signaling is required locally to pattern the dorsal telencephalic midline. Neuron, 35: 1029–1041.

Heinrich, C., Nitta, N., Flubacher, A., Muller, M., Fahrner, A., Kirsch, M., Freiman, T., Suzuki, F., Depaulis, A., Frotscher, M. and Haas, C.A. (2006) Reelin deficiency and displacement of mature neurons, but not neurogenesis, underlie the formation of granule cell dispersion in the epileptic hippocampus. J. Neurosci., 26: 4701–4713. Lai, K., Kaspar, B.K., Gage, F.H. and Schaffer, D.V. (2003) Sonic hedgehog regulates adult neural progenitor proliferation in vitro and in vivo. Nat. Neurosci., 6: 21–27. Lee, S.M., Tole, S., Grove, E. and Mcmahon, A.P. (2000) A local Wnt-3a signal is required for development of the mammalian hippocampus. Development, 127: 457–467. Li, G. and Pleasure, S.J. (2005) Morphogenesis of the dentate gyrus: what we are learning from mouse mutants. Dev. Neurosci., 27: 93–99. Lie, D.C., Colamarino, S.A., Song, H.J., Desire, L., Mira, H., Consiglio, A., Lein, E.S., Jessberger, S., Lansford, H., Dearie, A.R. and Gage, F.H. (2005) Wnt signalling regulates adult hippocampal neurogenesis. Nature, 437: 1370–1375. Liu, M., Pleasure, S.J., Collins, A.E., Noebels, J.L., Naya, F.J., Tsai, M.J. and Lowenstein, D.H. (2000) Loss of BETA2/ NeuroD leads to malformation of the dentate gyrus and epilepsy. Proc. Natl. Acad. Sci. U.S.A., 97: 865–870. Lu, M., Grove, E.A. and Miller, R.J. (2002) Abnormal development of the hippocampal dentate gyrus in mice lacking the CXCR4 chemokine receptor. Proc. Natl. Acad. Sci. U.S.A., 99: 7090–7095. Machold, R., Hayashi, S., Rutlin, M., Muzumdar, M.D., Nery, S., Corbin, J.G., Gritli-Linde, A., Dellovade, T., Porter, J.A., Rubin, L.L., Dudek, H., Mcmahon, A.P. and Fishell, G. (2003) Sonic hedgehog is required for progenitor cell maintenance in telencephalic stem cell niches. Neuron, 39: 937–950. Mallamaci, A., Mercurio, S., Muzio, L., Cecchi, C., Pardini, C.L., Gruss, P. and Boncinelli, E. (2000) The lack of Emx2 causes impairment of Reelin signaling and defects of neuronal migration in the developing cerebral cortex. J. Neurosci., 20: 1109–1118. Meyer, G., Cabrera Socorro, A., Perez Garcia, C.G., Martinez Millan, L., Walker, N. and Caput, D. (2004) Developmental roles of p73 in Cajal-Retzius cells and cortical patterning. J. Neurosci., 24: 9878–9887. Miyata, T., Maeda, T. and Lee, J.E. (1999) NeuroD is required for differentiation of the granule cells in the cerebellum and hippocampus. Genes Dev., 13: 1647–1652. Monuki, E.S., Porter, F.D. and Walsh, C.A. (2001) Patterning of the dorsal telencephalon and cerebral cortex by a roof plate-Lhx2 pathway. Neuron, 32: 591–604. Muzio, L. and Mallamaci, A. (2003) Emx1, emx2 and pax6 in specification, regionalization and arealization of the cerebral cortex. Cereb. Cortex, 13: 641–647. Muzio, L. and Mallamaci, A. (2005) Foxg1 confines Cajal-Retzius neuronogenesis and hippocampal morphogenesis to the dorsomedial pallium. J. Neurosci., 25: 4435–4441. Niewmierzycka, A., Mills, J., St-Arnaud, R., Dedhar, S. and Reichardt, L.F. (2005) Integrin-linked kinase deletion from mouse cortex results in cortical lamination defects resembling cobblestone lissencephaly. J. Neurosci., 25: 7022–7031.

152 Nowakowski, R.S. and Rakic, P. (1979) The mode of migration of neurons to the hippocampus: a Golgi and electron microscopic analysis in foetal rhesus monkey. J. Neurocytol., 8: 697–718. Nowakowski, R.S. and Rakic, P. (1981) The site of origin and route and rate of migration of neurons to the hippocampal region of the rhesus monkey. J. Comp. Neurol., 196: 129–154. Ohkubo, Y., Uchida, A.O., Shin, D., Partanen, J. and Vaccarino, F.M. (2004) Fibroblast growth factor receptor 1 is required for the proliferation of hippocampal progenitor cells and for hippocampal growth in mouse. J. Neurosci., 24: 6057–6069. Paredes, M.F., Li, G., Berger, O., Baraban, S.C. and Pleasure, S.J. (2006) Stromal-derived factor-1 (CXCL12) regulates laminar position of Cajal-Retzius cells in normal and dysplastic brains. J. Neurosci., 26: 9404–9412. Paylor, R., Zhao, Y., Libbey, M., Westphal, H. and Crawley, J.N. (2001) Learning impairments and motor dysfunctions in adult Lhx5-deficient mice displaying hippocampal disorganization. Physiol. Behav., 73: 781–792. Pellegrini, M., Mansouri, A., Simeone, A., Boncinelli, E. and Gruss, P. (1996) Dentate gyrus formation requires Emx2. Development, 122: 3893–3898. Pleasure, S.J., Collins, A.E. and Lowenstein, D.H. (2000) Unique expression patterns of cell fate molecules delineate sequential stages of dentate gyrus development. J. Neurosci., 20: 6095–6105. Porter, F.D., Drago, J., Xu, Y., Cheema, S.S., Wassif, C., Huang, S.P., Lee, E., Grinberg, A., Massalas, J.S., Bodine, D., Alt, F. and Westphal, H. (1997) Lhx2, a LIM homeobox gene, is required for eye, forebrain, and definitive erythrocyte development. Development, 124: 2935–2944. Pozniak, C.D. and Pleasure, S.J. (2006) A tale of two signals: Wnt and Hedgehog in dentate neurogenesis. Sci. STKE, 319: pe5. Shimogori, T., Banuchi, V., Ng, H.Y., Strauss, J.B. and Grove, E.A. (2004) Embryonic signaling centers expressing BMP, WNT and FGF proteins interact to pattern the cerebral cortex. Development, 131: 5639–5647. Shinozaki, K., Miyagi, T., Yoshida, M., Miyata, T., Ogawa, M., Aizawa, S. and Suda, Y. (2002) Absence of Cajal-Retzius cells and subplate neurons associated with defects of tangential cell migration from ganglionic eminence in Emx1/2 double mutant cerebral cortex. Development, 129: 3479–3492. Smith, K.M., Ohkubo, Y., Maragnoli, M.E., Rasin, M.R., Schwartz, M.L., Sestan, N. and Vaccarino, F.M. (2006) Midline radial glia translocation and corpus callosum formation require FGF signaling. Nat. Neurosci., 9: 787–797.

Storm, E.E., Garel, S., Borello, U., Hebert, J.M., Martinez, S., Mcconnell, S.K., Martin, G.R. and Rubenstein, J.L. (2006) Dose-dependent functions of Fgf8 in regulating telencephalic patterning centers. Development, 133: 1831–1844. Sur, M. and Rubenstein, J.L. (2005) Patterning and plasticity of the cerebral cortex. Science, 310: 805–810. Takiguchi-Hayashi, K., Sekiguchi, M., Ashigaki, S., Takamatsu, M., Hasegawa, H., Suzuki-Migishima, R., Yokoyama, M., Nakanishi, S. and Tanabe, Y. (2004) Generation of reelin-positive marginal zone cells from the caudomedial wall of telencephalic vesicles. J. Neurosci., 24: 2286–2295. Wilson, S.W. and Rubenstein, J.L. (2000) Induction and dorsoventral patterning of the telencephalon. Neuron, 28: 641–651. Yang, A., Walker, N., Bronson, R., Kaghad, M., Oosterwegel, M., Bonnin, J., Vagner, C., Bonnet, H., Dikkes, P., Sharpe, A., Mckeon, F. and Caput, D. (2000) p73-deficient mice have neurological, pheromonal and inflammatory defects but lack spontaneous tumours. Nature, 404: 99–103. Yoshida, M., Assimacopoulos, S., Jones, K.R. and Grove, E.A. (2006) Massive loss of Cajal-Retzius cells does not disrupt neocortical layer order. Development, 133: 537–545. Yoshida, M., Suda, Y., Matsuo, I., Miyamoto, N., Takeda, N., Kuratani, S. and Aizawa, S. (1997) Emx1 and Emx2 functions in development of dorsal telencephalon. Development, 124: 101–111. Zakin, L. and De Robertis, E.M. (2004) Inactivation of mouse Twisted gastrulation reveals its role in promoting Bmp4 activity during forebrain development. Development, 131: 413–424. Zhao, C., Aviles, C., Abel, R.A., Almli, C.R., Mcquillen, P. and Pleasure, S.J. (2005) Hippocampal and visuospatial learning defects in mice with a deletion of frizzled 9, a gene in the Williams syndrome deletion interval. Development, 132: 2917–2927. Zhao, C., Guan, W. and Pleasure, S.J. (2006) A transgenic marker mouse line labels Cajal-Retzius cells from the cortical hem and thalamocortical axons. Brain Res., 1077: 48–53. Zhao, S., Chai, X., Forster, E. and Frotscher, M. (2004) Reelin is a positional signal for the lamination of dentate granule cells. Development, 131: 5117–5125. Zhao, Y., Sheng, H.Z., Amini, R., Grinberg, A., Lee, E., Huang, S., Taira, M. and Westphal, H. (1999) Control of hippocampal morphogenesis and neuronal differentiation by the LIM homeobox gene Lhx5. Science, 284: 1155–1158. Zhou, C.J., Zhao, C. and Pleasure, S.J. (2004) Wnt signaling mutants have decreased dentate granule cell production and radial glial scaffolding abnormalities. J. Neurosci., 24: 121–126.

SECTION II

Cellular Analyses

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 9

Ultrastructure and synaptic connectivity of cell types in the adult rat dentate gyrus Charles E. Ribak and Lee A. Shapiro Department of Anatomy & Neurobiology, University of California at Irvine, Irvine, CA 92697-1275, USA

Abstract: The rat hippocampal dentate gyrus is an extensively studied structural component of the limbic system. It is the first station in the classical tri-synaptic circuit of the hippocampus in that its major input arises from the entorhinal cortex via the perforant pathway. The second part of this circuit arises from the projection cells of the dentate gyrus, the granule cells, which send their axons to the pyramidal cells of CA3. Within the dentate gyrus, there also is an extensive inhibitory network of cells that are involved in synchronizing the rhythmic firing of the granule cells. This chapter provides a review of the ultrastructural features and synaptic connectivity of both projection cells and local circuit neurons in the dentate gyrus. Keywords: granule cells; basket cells; mossy cells; chandelier cells; commissural cells; fusiform cells; GABAergic interneurons 1963; Seress and Ribak, 1985; Lubbers and Frotscher, 1987). The nucleus is usually round or slightly oval and has heterochromatin clumping against its nuclear envelope. It is uncommon to observe any infoldings of the nucleus of granule cells. Processes arise from the somata of granule cells and can be easily distinguished by the poles from which they arise. The axons of the granule cells, the well-described mossy fibers, arise from the hilar pole of the granule cell body (Ramon y Cajal, 1911), although one report indicated that the axon can arise from the apical dendrite (Yan et al., 2001). At their origin, they display the typical features of an axon initial segment with few organelles, bundles of microtubules, a subaxolemmal density and infrequent axon initial segment synapses (Steward and Ribak, 1986). It is known that mossy fibers arborize in the hilus and stratum lucidum of CA3. In addition, mossy fibers can be found either intragranular or supragranular. The intragranular fibers are often observed orthogonal

Granule cells: the principal cells of the dentate gyrus The principal cells of the dentate gyrus are the granule cells, and they are mainly found in the granule cell layer, which is up to
7–8 cells thick (Ramon y Cajal, 1911; Lorente de No´, 1934; Laatsch and Cowan, 1966). Granule cells are reported to exist infrequently in the two other parts of the dentate gyrus, the molecular layer and hilus (Amaral, 1978; Gaarskjaer and Laurberg, 1983; Marti-Subirana et al., 1986; Dashtipour et al., 2001). It should be noted that the vast majority of the cell bodies in the granule cell layer are granule cells, but not all of the cell bodies are of this type (Fig. 1). In the adult rat, granule cells (Fig. 1) display a round nucleus
10–12 mm in diameter, a thin shell of perikaryal cytoplasm, and both symmetric and asymmetric axosomatic synapses (Blackstad, Corresponding author. Tel.: +1 949-824-5494 & +1 949-8244558; Fax: +1 949-824-8549; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63009-X

155

156

H

A NN GC D

GC

A

GL

Fig. 1. An electron micrograph obtained from the part of the granule cell layer (GL) adjacent to the hilus (H) illustrates the ultrastructural features of granule cells (GC), radial glial cells (A) and a doublecortin-immunolabeled newborn neuron (NN). The mature granule cells (GC) located within the GL display a round or oval nucleus and a thin shell of perikaryal cytoplasm. The cell body of the doublecortin-immunolabeled cell (NN) is much smaller than that of the granule cells (GC). Note that a doublecortin-labeled apical process (white arrow) is adjacent to the radial process of an astrocyte (A), identified as a radial glial cell based on other characteristics. Another astrocyte (A) with similar ultrastructural features is shown at the lower part of this image within the GL where a proximal apical dendrite (D) from another mature granule cell apposes it. Many myelinated axons are present in the neuropil of the hilus (H) and the lower part of the GL. Scale bar ¼ 2 mm.

to the granule cell layer where they form asymmetric synapses with the apical dendrites and somata of basket cells (Ribak and Peterson, 1991). The supragranular fibers are located in the inner molecular layer and are found to sprout there after seizures, but are infrequent in the normal adult rat (Nadler et al., 1980; Buckmaster et al., 2002). These supragranular mossy fibers target the dendrites of granule cells to contribute to recurrent excitatory circuits after seizures. In addition, the

hilar basal dendrites observed following seizures are also postsynaptic to mossy fibers (Ribak et al., 2000). Thus, seizures can alter the morphology and connectivity of granule cells. The apical dendrites of granule cells extend radially through the granule cell layer (Fig. 1) and then arborize in the molecular layer and their tips eventually grow to reach the hippocampal fissure. These dendrites arise as a thick process with organelles that are typical of the perikaryal

157

cytoplasm (Fig. 1), including cisternae of the granular endoplasmic reticulum and Golgi complex (Blackstad, 1963; Laatsch and Cowan, 1966). The apical dendrites also display spines that are more common in the molecular layer, where perforant pathway axons target their outer two-thirds portion. In addition, there are several types of interneurons that synapse on the apical dendrites of granule cells in the molecular layer and the granule cell layer. These interneurons will be described in detail later in this chapter. In light of the relatively recent discovery of functional neurogenesis of granule cells in the adult rat dentate gyrus, it is necessary to describe the ultrastructural features of this population of newborn neurons. The newborn granule cells have been described using doublecortin (DCX)immunolabeled preparations and electron microscopy (Shapiro et al., 2005). These newborn granule cells are initially found in the subgranular zone with no dendritic processes and are cradled by an astrocyte, considered to be a radial glial cell (Shapiro et al., 2005). The diameter of these DCX-labeled cells is only
6 mm, which is smaller than that of the mature granule cells in the layer (Fig. 1). Also, these newborn cells have only a thin shell of perikaryal cytoplasm with few organelles. As the DCX-labeled newborn neuron migrates toward the granule cell layer, it extends an apical dendritic process along the radial process of the radial glial cells (Fig. 1). At this stage, the DCXlabeled cell body has a larger diameter and a thicker shell of perikaryal cytoplasm, relative to the DCX-labeled cells that lack processes in the subgranular zone. It should be noted that the dendrites in the molecular layer are much thinner than the apical dendrites of the mature granule cells (Shapiro et al., 2007). Although basal dendrites are not normally present on mature granule cells in the adult rat dentate gyrus, they are a transient feature of newborn granule cells in the young and adult rat (Seress and Pokorny, 1981; Ribak and Seress, 1990; Ribak et al., 2004), and persist following seizures (Ribak et al., 2000). It is pertinent to note that in human and non-human primates, basal dendrites are a persistent feature of some mature granule cells in the adult dentate gyrus (Seress and Mrzljak, 1987).

Radial glial cells in the adult dentate gyrus: the mothers of adult born granule cells Of the many types of progenitor cells in the adult rat dentate gyrus, the radial glial cells appear to have an intimate relationship with the newborn granule cells. These radial glial cells are immunoreactive for GFAP, are found in the hilus, subgranular zone, in the granule cell layer, at the border between the granule cell layer and the molecular layer and in the molecular layer (Kosaka and Hama, 1986). In this relationship, the radial glial cell and its non-radial watery cytoplasmic processes have been shown to encompass most of the newborn granule cell at the border between the subgranular zone and the granule cell layer (Fig. 1). These glial processes contain only sparse organelles and bundles of intermediate glial filaments (Shapiro et al., 2005). The region of the newborn neuron that is not surrounded by the radial glial cell is shown to give rise to an apical process that grows toward the granule cell layer. In this relationship, the radial process of the radial glial cell provides a scaffold for the apical dendrite from the newborn neuron to grow through the granule cell layer and into the molecular layer (Shapiro et al., 2005).

Basket cells: a plexus of inhibitory axons in the granule cell layer There are five distinct types of basket cells that have been described in the rat dentate gyrus (Ribak and Seress, 1983). These basket cells (Figs. 2 and 3) are located in the granule cell layer or the hilus, within 50 mm of the base of the granule cell layer and may express glutamate decarboxylase, GABA, or the calcium-binding protein, parvalbumin (Ribak et al., 1978, 1990; Seress and Ribak, 1983; Lubbers and Frotscher, 1987; Nitsch et al., 1990; Halasy and Somogyi, 1993). The shapes of these cells can be pyramidal (Fig. 2), horizontal, fusiform or multipolar. The ultrastructural features of all five types of basket cells are very similar (Ribak and Anderson, 1980; Ribak and Seress, 1983). They have a somal size of 20–30 mm with large nuclei containing euchromatin and

158

A

B G

G

G

Fig. 2. (A) shows the cell body and two proximal basal dendrites (arrows) of a pyramidal basket cell that was processed by the Golgi/ electron microscopy (EM) method and was previously identified at the light microscopic level. Note the gold particles subjacent to its plasma membrane. Only a small part of its nucleus (N) is shown but its perikaryal cytoplasm is packed with many organelles. Note the relatively larger cell body of the basket cell as compared to that of the neighboring granule cells (G). Cell bodies of two oligodendroglia (O) flank this basket cell at the border between the granule cell layer and the hilus (H). (B) is another basket cell processed with the Golgi/EM method to show the characteristic features of the nuclei of these cells. These features include nuclear infoldings (black arrows), a prominent nucleolus and intranuclear filaments (arrowhead). Like the basket cell in (A), this cell also has gold particles (white arrow) beneath its cell membrane while the surrounding granule cells (G) are not labeled. Scale bars ¼ 2 mm in A and 3 mm in B. (Reprinted with permission from Ribak and Seress, 1983.)

159

intranuclear rods and sheets, and displaying extensive nuclear infoldings (Figs. 2 and 3). Their thick shell of perikaryal cytoplasm has abundant cisternae of granular endoplasmic reticulum, which aggregate to form Nissl bodies, a well-developed Golgi complex, numerous mitochondria, free ribosomes and lysosomes. Golgi-electron microscopic preparations revealed details about the basket cell axons, showing that they have an extensive arborization concentrated mostly in the granule cell layer, and to a lesser extent at the border with the molecular layer. The axon terminals of basket cells contain numerous mitochondria and a sparse number of pleomorphic synaptic vesicles. These terminals were shown to form symmetric synapses, most commonly with the somata and proximal apical dendrites of granule cells (Ribak and Seress, 1983). In rare instances the basket cell axon terminal was observed to form a synapse with a spine (Ribak and Seress, 1983). It is important to emphasize that through these connections, these basket cells provide feedback inhibition to the granule cells of the dentate gyrus. In contrast, basket cells have not been observed to synapse on axon initial segments in adult rats, however, examples of this were documented in young (postnatal 10–16 days) rats (Seress and Ribak, 1990). The dendrites of basket cells are aspinous or sparsely spinous and are found in all layers of the dentate gyrus (Ribak and Seress, 1983). The aspinous dendrites have a mixture of asymmetric and symmetric synapses on their surfaces. In the hilus, the basal dendrites of pyramidal and fusiform basket cells have the greatest concentration of axon terminals apposed to their surfaces, and most of these axon terminals contain round synaptic vesicles and form asymmetric synapses. Some of the axon terminals that synapse on the basal dendrites of pyramidal basket cells were identified as mossy fiber terminals (Ribak and Seress, 1983). The basket cells receive afferents from at least two other fiber systems besides the mossy fibers from granule cells. They include the commissural axons from the contralateral dentate gyrus that terminate in the inner molecular layer and beneath the granule cell layer, and perforant path fibers that terminate in the outer molecular layer (Seress and

Ribak, 1984; Zipp et al., 1989). In addition to forming synapses on granule cell apical dendrites, these latter afferents also form synapses onto basket cell apical dendrites. These connections were suggested to function as a feed-forward inhibitory circuit for granule cells (Seress and Ribak, 1984; Zipp et al., 1989; Kneisler and Dingledine, 1995a). Other studies have shown that CA3 and the septum provide excitatory and inhibitory input, respectively, to basket cells (Gulyas et al., 1990; Scharfman, 1994; Kneisler and Dingledine, 1995b). Therefore, the GABAergic inhibitory basket cells mediate both feedback and feed-forward inhibition in the dentate gyrus (Kneisler and Dingledine, 1995a).

Mossy cells: the cells with great convergence of mossy fibers The mossy cells represent a distinct population of neurons that dominate the landscape of the dentate gyrus hilar region (Ribak et al., 1985; Frotscher et al., 1991). The mossy cells are characterized by large triangular or multipolar somata (25–30 mm diameter). These somata contain a round nucleus without infoldings, a thick shell of perikaryal cytoplasm and abundant organelles (Fig. 4). Arising from these somata are three or four thick primary dendrites (Fig. 4), which bifurcate to form an elaborate dendritic plexus that is mostly restricted to the hilus. These cells have complex spines known as thorny excrescences on their somata and proximal dendrites (Fig. 4), while their distal dendrites have typical spines. The majority of axon terminals forming synapses onto these dendrites arise from mossy fibers of granule cells (Fig. 4). Therefore, this massive granule cell input provides a large convergence to mossy cells, and this input was shown to be excitatory to mossy cells (Scharfman et al., 1990). The axons of these mossy cells bifurcate and the major projection is to the distant portions of the inner molecular layer (Soltesz et al., 1993). The axon terminals of mossy cells were observed to make asymmetric synapses onto postsynaptic targets in the hilus and molecular layer of the dentate gyrus (Ribak et al., 1985; Scharfman et al., 1990) and showed immunoreactivity primarily for

160

Fig. 3. Illustrations of a pyramidal basket cell that was immunolabeled for parvalbumin. (A) shows a semi-thin section of this cell in the lower part of the granule cell layer (GL) next to the hilus (H). Its apical dendrite branches in the inner molecular layer (ML). The small punctate labeling in the GL probably represents the basket cell axons that outline the cell bodies of unlabeled granule cells. The electron micrograph in (B) shows the soma of this parvalbumin-immunolabeled pyramidal basket cell with the typical features of this cell type. These include nuclear infoldings (arrowheads), intranuclear rod (open arrow), axosomatic synapses (closed arrow) and a thick perikaryal cytoplasm. Note the size difference between this soma and the somata of the adjacent granule cells (G). (C) is an enlargement of the axosomatic symmetric synapses (arrows) denoted in (B) with an arrow. The axon terminals (T) forming these synapses with the parvalbumin-immunolabeled basket cell body have pleomorphic vesicles and several mitochondria. (D) is another example of axosomatic synapses with the same basket cell body but these are located at the transition with its apical dendrite and these terminals (T) form asymmetric synapses (arrows). Note that this part of the soma displays more immunolabeling. Scale bar in A ¼ 10 mm in A, 2.5 mm in B, 0.1 mm in C and D. (Reprinted with permission from Ribak et al., 1990.)

161

Fig. 4. (A) shows the cell body and two thick proximal dendrites (D) of a mossy cell that was processed by the Golgi/EM method and was previously identified at the light microscopic level. Most of its cell body is occupied by a large round nucleus (N). A large somal spine (arrow) is found to the right of the cell body. Two dendrites (D) arise from the cell body. (B) shows a portion of a dendrite (D) from this mossy cell. Note the numerous gold-labeled spines arising from this dendrite. Many of these spines are postsynaptic to large axon terminals (M) that resemble mossy fiber tufts. (C) is an enlargement of a mossy fiber (M and outlined by dashed line) that forms a synapse (arrow) with a gold-labeled dendrite of this mossy cell and with one of its gold-labeled spines (S). Scale bars ¼ 4 mm in A, 1.5 mm in B and 0.5 mm in C. (Reprinted with permission from Ribak et al., 1985.)

glutamate, but never for GABA (Soriano and Frotscher, 1994). Within the hilus, glutamate-positive mossy cell axon terminals targeted GABApositive dendritic shafts of hilar interneurons and GABA-negative dendritic spines. Mossy cell axons in the inner molecular layer form asymmetric synapses with dendritic spines associated with GABA-negative, granule cell dendrites. Thus, excitatory (glutamatergic) mossy cell terminals contact GABAergic interneurons and nonGABAergic neurons in the hilar region and GABA-negative dendrites of granule cells in the molecular layer. This pattern of connectivity is consistent with the hypothesis that mossy cells provide excitatory feedback to granule cells in a dentate gyrus associational network and also activate local hilar inhibitory elements (Wenzel et al., 1997).

Fusiform cells (spiny and aspiny) in the hilus The fusiform cells of the rat dentate gyrus are located in the hilus, within 100 mm of the base of the granule cell layer (Fig. 5). These cells are one of the most prevalent cell types in the subgranular zone of the dentate gyrus (Ribak and Seress, 1988). The fusiform cells are bipolar, with oval cell bodies, and their dendrites typically run parallel to the granule cell layer, and within the subgranular zone (Fig. 5). The dendrites of these cells are either spiny or sparsely spiny. The spiny fusiform cell receives the majority of its input from the axon collaterals of the mossy fiber axons from the granule cells. Interestingly, the spines are not only observed on the dendrites, but are also observed on the soma (Fig. 5). The spiny fusiform cell has a soma that is similar in appearance to the mossy

162

Fig. 5. Illustrations of two types of fusiform cells that were processed by the Golgi/EM method and were previously identified at the light microscopic level. (A) is an electron micrograph of a sparsely spiny fusiform cell body and one of its proximal dendrites (D). The nucleus (N) displays a prominent nucleolus and numerous infoldings (large arrows). The perikaryal cytoplasm contains many cisternae of granular endoplasmic reticulum (ER) that are organized into stacks or Nissl bodies. A few small somal spines (small arrows) are also shown. (B) is a light micrograph of the same sparsely spiny fusiform cell shown in (A) to illustrate its flattened soma (large arrow) and its dendrites that run parallel to the granule cell layer (small arrows). (C) shows a spiny fusiform cell body with many gold-labeled somal spines (arrows). The nucleus (N) in this section occupies less area than would the nucleus obtained from a section through the center of the soma. The cisternae of the granular endoplasmic reticulum (ER) do not form parallel stacks. Electron-dense lysosomes as well as mitochondria are randomly distributed in the perikaryal cytoplasm. A small capillary (C) is also shown in the upper left corner of the micrograph. (D) is an enlargement of the two somal spines (S) indicated by the top arrow in (C). Large axon terminals packed with agranular round vesicles form asymmetric synapses (large arrows) with the gold-labeled spines. In contrast, the somal surface is contacted by a small terminal that appears to form a symmetric synapse (small arrow). Scale bars ¼ 20 mm in (A), 2 mm in (B), 2 mm in (C) and 0.5 mm in (D). (Reprinted with permission from Ribak and Seress, 1988.)

cell because it displays Nissl bodies and little or no nuclear infolding (Fig. 5). In contrast, the fusiform cells with sparsely spiny dendrites have somata with nuclear infoldings (Fig. 5) like that of basket cells and few Nissl bodies. It should also be noted that this latter cell type displays a variety of

axodendritic synapses, suggesting a more diverse synaptic input than that of the spiny fusiform cell. Several studies have demonstrated the presence of somatostatin immunolabeling in hilar cells with the morphology of fusiform cells (Bakst et al., 1986; Milner and Bacon, 1989; Leranth et al.,

163

1990). These studies provided a valuable insight into the axonal projections of the fusiform cell. It was shown that the somatostatin-labeled axons formed a dense plexus in the outer molecular layer where they made symmetric synapses onto the granule cell dendrites (Bakst et al., 1986; Milner and Bacon, 1989; Leranth et al., 1990). In addition, somatostatin-labeled axon terminals in the hilus were also reported (Milner and Bacon, 1989). It should be noted that more than 90% of the hilar neurons expressing somatostatin are GABAergic (Esclapez and Houser, 1995) and that
50% of them express neuropeptide Y (Deller and Leranth, 1990).

Chandelier cells: a GABAergic cell-type targeting axon initial segments of granule cells Chandelier cells are observed throughout the cerebral cortex, but were first described in the hippocampus by Kosaka (1980) and Somogyi et al. (1983a, b), and later in the dentate gyrus by Soriano and Frotscher (1989). All chandelier cells share two features that are unique among inhibitory interneurons; their axons form rows of boutons and they make symmetric synapses exclusively on axon initial segments. The chandelier cells of the dentate gyrus have their somata located within or immediately adjacent to the granule cell layer. The cell bodies and dendrites of these neurons exhibit several of the characteristic ultrastructural features of non-granule cells, including a large perikaryal cytoplasm, nuclear infoldings, intranuclear inclusions and a large number of synapses on the soma and aspiny dendrites. Within the dentate gyrus, chandelier cell axons densely innervate the granule cell layer with radially oriented terminal rows, and also form an extensive plexus in the hilus. The dendrites of the chandelier cells extend radially through the molecular layer and can reach as far as the hippocampal fissure (Soriano and Frotscher, 1989; Soriano et al., 1990; Han et al., 1993; Buhl et al., 1994). Occasionally, basal dendrites from chandelier cells cross the granule cell layer toward the hilus (Soriano et al., 1990). There are also three classes of chandelier cells that have their somata in

the hilus. The three chandelier cell types observed in the hilus projected to granule cells at the same septo–temporal level where their cell bodies were located. It should also be noted that the chandelier cells exhibit immunolabeling for glutamate decarboxylase, GABA-transporter-1 and parvalbumin (Ribak et al., 1990, 1996; Soriano et al., 1990). Thus, the dentate gyrus chandelier cell provides a spatially selective innervation of granule cells and complements that of the basket cell that provides GABAergic inhibition to cell bodies and proximal dendrites.

Commissural neurons: communication and synchronization across hemispheres involving excitatory and inhibitory neurons Within the dentate gyrus, there are at least two classes of cells that send projections to the contralateral dentate gyrus. Using retrograde labeling with horseradish peroxidase (HRP), Seroogy et al. (1983) showed different electron microscopic features for two classes of labeled commissural neuron. The first type consisted of cells with somata that exhibited round or oval nuclei with no intranuclear inclusions and had exclusively symmetric synapses on their somata. The main dendrites of those neurons were thick and tapering. This type had features that resembled the morphology of the mossy cell. A subsequent study using combined retrograde transport of HRP with Golgi/electron microscopy (EM) confirmed this suggestion (Frotscher, 1992). The second type of labeled neuronal soma had infolded nuclei containing intranuclear rods or sheets, displayed both symmetric and asymmetric synapses on its soma and had dendrites that were less thick and generally aspinous. This type of commissurally labeled cell had features that were similar to the dentate gyrus basket cell, a local circuit neuron associated with GABAergic inhibition, as described above. Another line of evidence supported the possibility that GABAergic neurons had commissural projections. In a quantitative study of GABAergic neurons in the hilus of the dentate gyrus, Seress and Ribak (1983) showed that 60% of the hilar neurons are GAD-positive. Because previous

164

studies indicated that 80% of hilar neurons give rise to associational and commissural pathways, many GABAergic neurons in the hilus were suggested to be projection neurons, and a subsequent combined tracer and immunofluorescence study showed several double-labeled GABAergic neurons in the hilus contralateral to the injection site (Ribak et al., 1986). Another confirmation of this conclusion came with the elegant intracellular labeling study of Han et al. (1993). In that study, the axons of several classes of GABAergic interneurons were mapped in the rat dentate gyrus, including one that had its axon terminals distributed in the commissural and associational pathway termination field. Because they named the cell using both its cell body location and axon terminal field, and the cell body of these neurons resided in the hilus, this cell type was referred to as the hilus/ commissural-associational pathways (HICAP) cell (Han et al., 1993). Therefore, the commissural projection in the rat includes both excitatory (mossy cells) and inhibitory (HICAP cells) neurons. Other GABAergic interneuron types Two other GABAergic neuron types were observed in the dentate gyrus and were described simultaneously with the HICAP cell (Han et al., 1993). One of these is another type of hilar cell that has an axon, which ramifies in the perforant path terminal field in the outer two-thirds of the molecular layer. This hilar cell with axon terminals distributed in the perforant path termination field was named the hilus/perforant pathway (HIPP) cell. Electron microscopy of this cell type revealed sparsely spiny dendrites that were covered with many synaptic boutons on both their shafts and their spines but no details were provided about the HIPP cell’s somal features (Han et al., 1993). It should be noted that this cell type has its axon terminal field in the same location as the somatostatin-labeled cell described above in the fusiform cell section. That cell type was also GABAergic (Esclapez and Houser, 1995). The other GABAergic cell type described by Han et al. (1993) had its cell body in the molecular

layer and its dendritic and axonal domains were confined to the perforant path terminal zone. This cell type was referred to as the molecular layer/ perforant pathway (MOPP) cell and its axon made symmetric synapses exclusively onto dendritic shafts, 60% of which were shown to emit spines (Halasy and Somogyi, 1993). The smooth dendrites of the MOPP cell were also restricted to the outer two-thirds of the molecular layer, where they received both GABA-negative and GABA-positive synaptic inputs. Similar to the HIPP cells, ultrastructural details about the MOPP cell’s soma were not included in the description.

Conclusion There have been many studies on the cell types in the dentate gyrus and their synaptic connections. Because the dentate gyrus is the first station in the classical tri-synaptic circuit of the hippocampus, understanding its anatomical organization is essential to understand the function of this structure. Here we have outlined the ultrastructural features and synaptic connections of the principal cell type (the granule cells) and the projection and local circuit neurons that make up the dentate gyrus.

Acknowledgments The authors are grateful to the following collaborators who contributed significantly to one or several studies in this review: Drs. David Amaral, Nich Brecha, Khashayar Dashtipour, James Fallon, Michael Frotscher, Mathew Korn, J. Victor Nadler, Robert Nitsch, Andre Obenaus, Maxine Okazaki, Gary Peterson, Kihachi Saito, Larry Schmued, Laszlo Seress, Kim Seroogy, Igor Spigelman, Oswald Steward, Winnie Tong, James Vaughn and Xiao-Xin Yan. This review is dedicated to the mentor of C.E.R., Alan Peters, who perfected the Golgi/EM method for the analysis of short axonal connections. We acknowledge the support from NIH grant R01-NS38331 (to C.E.R.) and NIH training grant T32-NS45540 (for L.A.S.).

165

References Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 195: 851–914. Bakst, I., Avendano, C., Morrison, J.H. and Amaral, D.G. (1986) An experimental analysis of the origins of somatostatin-like immunoreactivity in the dentate gyrus of the rat. J. Neurosci., 6: 1452–1462. Blackstad, T. (1963) Ultrastructural studies on the hippocampal region. Prog. Brain Res., 3: 122–148. Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002) Axon sprouting in a model of temporal lobe epilepsy creates a predominantly excitatory feedback circuit. J. Neurosci., 22: 6650–6658. Buhl, E.H., Han, Z.S., Lorinczi, Z., Stezhka, V.V., Karnup, S.V. and Somogyi, P. (1994) Physiological properties of anatomically identified axo-axonic cells in the rat hippocampus. J. Neurophysiol., 71: 1289–1307. Dashtipour, K., Tran, P.H., Okazaki, M.M., Nadler, J.V. and Ribak, C.E. (2001) Ultrastructural features and synaptic connections of hilar ectopic granule cells in the rat dentate gyrus are different from those of granule cells in the granule cell layer. Brain Res., 890: 261–271. Deller, T. and Leranth, C. (1990) Synaptic connections of neuropeptide Y (NPY) immunoreactive neurons in the hilar area of the rat hippocampus. J. Comp. Neurol., 300: 433–447. Esclapez, M. and Houser, C.R. (1995) Somatostatin neurons are a subpopulation of GABA neurons in the rat dentate gyrus: evidence from colocalization of pre-prosomatostatin and glutamate decarboxylase messenger RNAs. Neuroscience, 64: 339–355. Frotscher, M. (1992) Application of the Golgi/electron microscopy technique for cell identification in immunocytochemical, retrograde labeling, and developmental studies of hippocampal neurons. Microsc. Res. Tech., 23: 306–323. Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E. (1991) The mossy cells of the fascia dentata: a comparative study of their fine structure and synaptic connections in rodents and primates. J. Comp. Neurol., 312: 145–163. Gaarskjaer, F.B. and Laurberg, S. (1983) Ectopic granule cells of hilus fasciae dentatae projecting to the ipsilateral regio inferior of the rat hippocampus. Brain Res., 274: 11–16. Gulyas, A.I., Gorcs, T.J. and Freund, T.F. (1990) Innervation of different peptide-containing neurons in the hippocampus by GABAergic septal afferents. Neuroscience, 37: 31–44. Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple GABAergic innervation of granule cells in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 411–429. Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 395–410. Kneisler, T.B. and Dingledine, R. (1995a) Spontaneous and synaptic input from granule cells and the perforant path to

dentate basket cells in the rat hippocampus. Hippocampus, 5: 151–164. Kneisler, T.B. and Dingledine, R. (1995b) Synaptic input from CA3 pyramidal cells to dentate basket cells in rat hippocampus. J. Physiol., 487: 125–146. Kosaka, T. (1980) The axon initial segment as a synaptic site: ultrastructure and synaptology of the initial segment of the pyramidal cell in the rat hippocampus (CA3 region). J. Neurocytol., 9: 861–882. Kosaka, T. and Hama, K. (1986) Three-dimensional structure of astrocytes in the rat dentate gyrus. J. Comp. Neurol., 249: 242–260. Laatsch, R.H. and Cowan, W.M. (1966) Electron microscopic studies of the dentate gyrus of the rat. I. Normal structure with special reference to synaptic organization. J. Comp. Neurol., 128: 359–396. Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Afferent and efferent synaptic connections of somatostatin-immunoreactive neurons in the rat fascia dentata. J. Comp. Neurol., 295: 111–122. Lorente de No´, R. (1934) Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system. J. Psychol. Neurol. Lpz., 46: 113–177. Lubbers, K. and Frotscher, M. (1987) Fine structure and synaptic connections of identified neurons in the rat fascia dentata. Anat. Embryol. (Berl.), 177: 1–14. Marti-Subirana, A., Soriano, E. and Garcia-Verdugo, J.M. (1986) Morphological aspects of the ectopic granule-like cellular populations in the albino rat hippocampal formation: a Golgi study. J. Anat., 144: 31–47. Milner, T.A. and Bacon, C.E. (1989) Ultrastructural localization of somatostatin-like immunoreactivity in the rat dentate gyrus. J. Comp. Neurol., 290: 544–560. Nadler, J.V., Perry, B.W. and Cotman, C.W. (1980) Selective reinnervation of hippocampal area CA1 and the fascia dentata after destruction of CA3-CA4 afferents with kainic acid. Brain Res., 182: 1–9. Nitsch, R., Soriano, E. and Frotscher, M. (1990) The parvalbumin-containing nonpyramidal neurons in the rat hippocampus. Anat. Embryol. (Berl.), 181: 413–425. Ramon y Cajal, S. (1911) Histologie du Systeme Nerveux de l’Homme et des Vertebres, Vol. 2. Maloine, Paris. Ribak, C.E. and Anderson, L. (1980) Ultrastructure of the pyramidal basket cells in the dentate gyrus of the rat. J. Comp. Neurol., 192: 903–916. Ribak, C.E., Korn, M.J., Shan, Z. and Obenaus, A. (2004) Dendritic growth cones and recurrent basal dendrites are typical features of newly generated dentate granule cells in the adult hippocampus. Brain Res., 1000: 195–199. Ribak, C.E., Nitsch, R. and Seress, L. (1990) Proportion of parvalbumin-positive basket cells in the GABAergic innervation of pyramidal and granule cells of the rat hippocampal formation. J. Comp. Neurol., 300: 449–461. Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy fibers in rats and gerbils form synapses with the somata and proximal dendrites of basket cells in the dentate gyrus. Hippocampus, 1: 355–364.

166 Ribak, C.E. and Seress, L. (1983) Five types of basket cell in the hippocampal dentate gyrus: a combined Golgi and electron microscopic study. J. Neurocytol., 12: 577–597. Ribak, C.E. and Seress, L. (1988) A Golgi-electron microscopic study of fusiform neurons in the hilar region of the dentate gyrus. J. Comp. Neurol., 271: 67–78. Ribak, C.E. and Seress, L. (1990) Postnatal development of the light and electron microscopic features of basket cells in the hippocampal dentate gyrus of the rat. Anat. Embryol. (Berl.), 181: 547–565. Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The development, ultrastructure and synaptic connections of the mossy cells of the dentate gyrus. J. Neurocytol., 14: 835–857. Ribak, C.E., Seress, L., Peterson, G.M., Seroogy, K.B., Fallon, J.H. and Schmued, L.C. (1986) A GABAergic inhibitory component within the hippocampal commissural pathway. J. Neurosci., 6: 3492–3498. Ribak, C.E., Tong, W.M. and Brecha, N.C. (1996) GABA plasma membrane transporters, GAT-1 and GAT-3, display different distributions in the rat hippocampus. J. Comp. Neurol., 367: 595–606. Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and Nadler, J.V. (2000) Status epilepticus-induced hilar basal dendrites on rodent granule cells contribute to recurrent excitatory circuitry. J. Comp. Neurol., 428: 240–253. Ribak, C.E., Vaughn, J.E. and Saito, K. (1978) Immunocytochemical localization of glutamic acid decarboxylase in neuronal somata following colchicine inhibition of axonal transport. Brain Res., 140: 315–332. Scharfman, H.E. (1994) Synchronization of area CA3 hippocampal pyramidal cells and non-granule cells of the dentate gyrus in bicuculline-treated rat hippocampal slices. Neuroscience, 59: 245–257. Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A. (1990) Synaptic connections of dentate granule cells and hilar neurons: results of paired intracellular recordings and intracellular horseradish peroxidase injections. Neuroscience, 37: 693–707. Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule cells are normal features of the fetal and adult dentate gyrus of both monkey and human hippocampal formations. Brain Res., 405: 169–174. Seress, L. and Pokorny, J. (1981) Structure of the granular layer of the rat dentate gyrus. J. Anat., 133: 181–195. Seress, L. and Ribak, C.E. (1983) GABAergic cells in the dentate gyrus appear to be local circuit and projection neurons. Exp. Brain Res., 50: 173–182. Seress, L. and Ribak, C.E. (1984) Direct commissural connections to the basket cells of the hippocampal dentate gyrus: anatomical evidence for feed-forward inhibition. J. Neurocytol., 13: 215–225. Seress, L. and Ribak, C.E. (1985) A substantial number of asymmetric axosomatic synapses is a characteristic of the granule cells of the hippocampal dentate gyrus. Neurosci. Lett., 56: 21–26.

Seress, L. and Ribak, C.E. (1990) The synaptic connections of basket cell axons in the developing rat hippocampal formation. Exp. Brain Res., 81: 500–508. Seroogy, K.B., Seress, L. and Ribak, C.E. (1983) Ultrastructure of commissural neurons of the hilar region in the hippocampal dentate gyrus. Exp. Neurol., 82: 594–608. Shapiro, L.A., Korn, M.J., Shan, Z. and Ribak, C.E. (2005) GFAP-expressing radial glia-like cell bodies are involved in a one-to-one relationship with doublecortin-immunolabeled newborn neurons in the adult dentate gyrus. Brain Res., 1040: 81–91. Shapiro, L.A., Upadhyaya, P. and Ribak, C.E. (2007) Spatiotemporal profile of dendritic outgrowth from newly born granule cells in the adult rat dentate gyrus. Brain Res., 1149: 30–37. Soltesz, I., Bourassa, J. and Deschenes, M. (1993) The behavior of mossy cells of the rat dentate gyrus during theta oscillations in vivo. Neuroscience, 57: 555–564. Somogyi, P., Nunzi, M.G., Gorio, A. and Smith, A.D. (1983a) A new type of specific interneuron in the monkey hippocampus forming synapses exclusively with the axon initial segments of pyramidal cells. Brain Res., 259: 137–142. Somogyi, P., Smith, A.D., Nunzi, M.G., Gorio, A., Takagi, H. and Wu, J.Y. (1983b) Glutamate decarboxylase immunoreactivity in the hippocampus of the cat: distribution of immunoreactive synaptic terminals with special reference to the axon initial segment of pyramidal neurons. J. Neurosci., 3: 1450–1468. Soriano, E. and Frotscher, M. (1989) A GABAergic axo-axonic cell in the fascia dentata controls the main excitatory hippocampal pathway. Brain Res., 503: 170–174. Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat fascia dentata are glutamate-immunoreactive. Hippocampus, 4: 65–69. Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic chandelier cells in the rat fascia dentata: Golgi-electron microscopy and immunocytochemical studies. J. Comp. Neurol., 293: 1–25. Steward, O. and Ribak, C.E. (1986) Polyribosomes associated with synaptic specializations on axon initial segments: localization of protein-synthetic machinery at inhibitory synapses. J. Neurosci., 6: 3079–3085. Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E. and Schwartzkroin, P.A. (1997) Ultrastructural localization of neurotransmitter immunoreactivity in mossy cell axons and their synaptic targets in the rat dentate gyrus. Hippocampus, 7: 559–570. Yan, X.X., Spigelman, I., Tran, P.H. and Ribak, C.E. (2001) Atypical features of rat dentate granule cells: recurrent basal dendrites and apical axons. Anat. Embryol. (Berl.), 203: 203–209. Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989) Entorhinal fibers form synaptic contacts on parvalbuminimmunoreactive neurons in the rat fascia dentata. Brain Res., 495: 161–166.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 10

Morphological development and maturation of granule neuron dendrites in the rat dentate gyrus Omid Rahimi1 and Brenda J. Claiborne2, 1

Department of Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, TX 78229, USA 2 Department of Biology, University of Texas at San Antonio, One UTSA Circle, San Antonio, TX 78249, USA

Abstract: The first granule neurons in the dentate gyrus are born during late embryogenesis in the rodent, and the primary period of granule cell neurogenesis continues into the second postnatal week. On the day of birth in the rat, the oldest granule neurons are visible in the suprapyramidal blade and exhibit rudimentary dendrites extending into the molecular layer. Here we describe the morphological development of the dendritic trees between birth and day 14, and we then review the process of dendritic remodeling that occurs after the end of the second week. Data indicate that the first adult-like granule neurons are present on day 7, and, furthermore, physiological recordings demonstrate that some granule neurons are functional at this time. Taken together, these results suggest that the dentate gyrus may be incorporated into the hippocampal circuit as early as the end of the first week. The dendritic trees of the granule neurons, however, continue to increase in size until day 14. After that time, the dendritic trees of the oldest granule neurons are sculpted and refined. Some dendrites elongate while others are lost, resulting in a conservation of total dendritic length. We end this chapter with a review of the quantitative aspects of granule cell dendrites in the adult rat and a discussion of the relationship between the morphology of a granule neuron and the location of its cell body within stratum granulosum and along the transverse axis of the dentate gyrus. Keywords: dendritic trees; spines; filopodia; neonates; hippocampus

forms synapses on pyramidal neurons in the CA3 region of the hippocampus proper. The apical dendrites of the granule neurons bifurcate as they traverse the molecular layer, and the vast majority of terminal branches reach the top of the layer in the adult. The dendritic trees of most granule neurons are elliptical, and all dendrites of granule neurons in the adult dentate gyrus are covered with spines. The primary period of granule cell neurogenesis occurs over a two- to three-week period in the rodent, beginning in late embryogenesis and continuing through the second postnatal week. In the rat, although a few granule neurons are born as early as

Introduction Granule neurons are the principal cell type in the dentate gyrus, and their cell bodies are located in stratum granulosum of the suprapyramidal and infrapyramidal blades. Dendrites extend from the apical pole of the granule cell body into the overlying molecular layer, and the axon, or mossy fiber, exits from the basal pole. The axon gives rise to collateral branches in the hilar region and then Corresponding author. Tel.: +1 210 458 5487; Fax: 1 210 458 5669; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63010-6

167

168

embryonic day 14, over 80% are born after the birth of the animal (which occurs at about embryonic day 21) and neurogenesis peaks near the end of the first week of life (Bayer and Altman, 1974; Schlessinger et al., 1975). It is worth noting that the granule neurons are the last cells to be generated in the hippocampal formation, and it is well known that granule cell neurogenesis continues into adulthood (Altman and Das, 1965; Kaplan and Hinds, 1977). Here we focus on granule neurons that are generated in the neonatal rat. From recent evidence in the mouse, it appears that adult-generated granule neurons progress through a similar set of stages as they develop and mature (Zhao et al., 2006). Any description of the development and maturation of the granule neurons must take into account the temporal and spatial gradients of granule cell neurogenesis (Schlessinger et al., 1975; Cowan et al., 1980, 1981). The earliest born granule neurons form stratum granulosum in the septal portion of the dentate gyrus, and neurons that are generated later form the more temporal portions of the dentate gyrus. This gradient is referred to as the septotemporal gradient. A second gradient exists along the transverse axis of the dentate gyrus and is of considerable importance for developmental and morphological studies. As the first granule neurons are born, they form the cell-body layer at the tip of the suprapyramidal blade. As additional neurons are generated, they form the cell-body layer in the middle of the suprapyramidal blade and then the portions of stratum granulosum closest to the crest region. This gradient continues as more neurons are generated such that the youngest neurons make up stratum granulosum in the infrapyramidal blade. A third gradient exists within stratum granulosum. The neurons that are born first move into their final position at the top of the cell-body layer near the molecular layer, and the younger neurons move into position beneath them such that they are located in the bottom portion of stratum granulosum near the hilar border. This developmental pattern is in contrast to the ‘‘inside-out’’ pattern found in other areas of the mammalian cerebral cortex in which the later-generated neurons move through the earlier-generated cells to occupy positions at the top of the cell-body layer.

Thus the oldest granule neurons are most likely to be found at the top of stratum granulosum near the distal tip of the suprapyramidal blade at the septal pole, whereas the youngest neurons are located predominantly in the infrapyramidal blade near the temporal pole and in the deeper portions of stratum granulosum along the entire extent of the transverse axis of the dentate gyrus. In the rat, the suprapyramidal blade begins to form in late embryogenesis as the first granule neurons are born — it is visible as a separate structure on the day of birth and consists of a cell-body layer and a relatively thin molecular layer (Cowan et al., 1980). The molecular layer increases greatly in width over the first several weeks. It is less than 100 mm at day 4 and increases to just over 200 mm at day 14; it averages approximately 300 mm in width in young adult rats (Loy et al., 1977; Claiborne et al., 1990; Rihn and Claiborne, 1990). The infrapyramidal blade is barely visible on the day of birth and grows more slowly than the suprapyramidal blade during the first week, increasing from approximately 45 mm in width on day 4 to approximately 110 mm on day 10; it measures between 205 and 240 mm in young adult rats (Loy et al., 1977; Claiborne et al., 1990). Here we describe the development and maturation of the dendritic trees of the granule neurons in the rat, and we review the quantitative data on dendritic morphology in the adult. We consider the developmental period to encompass the time from the birth of the animal through day 14. By day 14, the oldest granule neurons have assumed their adult form and size. From day 14 to 60, however, the neurons go through a period of maturation during which the dendritic tree is sculpted and refined and the density of spines continues to increase. The dendritic trees of granule neurons appear to be mature by day 60: unpublished data from our lab indicate that they do not undergo any quantitative changes between 60 and 180 days.

Development of granule neuron dendrites Because of the prolonged time-course of granule cell neurogenesis in neonatal rats, a wide range of dendritic morphologies are observed on any one

169

day during the first and second postnatal weeks (Fricke, 1975; Seress and Pokorny, 1981; Wenzel et al., 1981; Lu¨bbers and Frotscher, 1988; Liu et al., 1996, 2000; Ye et al., 2000; Jones et al., 2003). It is possible, however, to identify distinct stages in the development of granule cell dendritic trees, either by comparing the various neuronal structures present on a single day or by examining the progression of dendritic morphologies over the course of the neonatal period. As an example of the first approach, Lu¨bbers and Frotscher (1988) identified several stages of developing granule neurons in a 5-day-old rat. Neurons in the earliest stages of development had only rudimentary dendritic trees, each consisting of two or more primary apical dendrites exiting from the cell body, along with one or more basal branches. The primary apical dendrites gave rise to higher order branches that were relatively short and thin and that exhibited varicosities and the occasional growth cone. Spines were not present on these immature cells. In contrast, the most mature granule neurons on day 5 had a more elaborate apical tree with longer branches that exhibited numerous varicosities as well as occasional growth cones and filopodia. A few spines were present on the apical dendrites. Based on data from Lu¨bbers and Frotscher (1988) and a number of other investigators, here we describe a sequence of stages that characterize the morphological development of granule neuron dendrites in the young rat. A variety of techniques have been used to examine developing granule neurons during the first few weeks of life, including Golgi impregnation (Fricke, 1975; Duffy and Teyler, 1978a, b; Wenzel et al., 1981; Lu¨bbers and Frotscher, 1988; Zafirov et al., 1994), intracellular injection (Liu et al., 1996, 2000; Ye et al., 2000), and retrograde labeling (Jones et al., 2003). Fricke (1975) characterized the dendritic structures of Golgi-stained granule neurons in tissue from rats at postnatal days 1, 2, 4, 8, 12, and 20, as well as in tissue from adult rats over the age of 60 days. Other labs using Golgi-impregnated material have examined a similar range of ages. Duffy and Teyler (1978a, b) analyzed granule neurons in sections from rats at 7, 14, 30, 60, and 210 days of age, Wenzel et al. (1981) determined morphologies at a

variety of ages between days 0 and 180, and Zafirov et al. (1994) described granule neurons at days 5, 10, 15, and 20. Trommer and colleagues (Liu et al., 1996, 2000; Ye et al., 2000) injected granule neurons in slices from rats between the ages of 5 and 32 days with biocytin, whereas Jones et al. (2003) used retrograde labeling with DiI to analyze the dendritic structures of granule neurons in rats between the ages of 2 and 9 days. The latter group focused on the oldest granule neurons — those that were located near the tip of the suprapyramidal blade and in the top portion of stratum granulosum. On the day of birth and on postnatal day 1 in the rat, granule neurons are visible in the suprapyramidal blade of Golgi-stained tissue and exhibit rudimentary trees. Some have only a few short, stubby branches, whereas others have longer and more numerous dendrites (Fricke, 1975; Wenzel et al., 1981). The dendrites are smooth and growth cones are not typically observed. Wenzel et al. (1981) considered granule neurons on the day of birth to be in the first stage of development and labeled them as primitive or early neuroblasts. As noted by Fricke (1975), it is a bit surprising that all of the stained granule neurons at this age have such sparse dendritic trees — although most granule neurons are likely to be only a few days old at this time, a few neurons are born as early as embryonic day 14 and are already 7 days old on the day of birth. On postnatal days 2 and 3, the oldest granule neurons, in the suprapyramidal blade, have one or more primary apical dendrites and varying numbers of shorter, higher-order branches (Fig. 1A and B; Jones et al., 2003). Granule cells located at the very top of stratum granulosum have several primary dendrites whereas those located a bit deeper in the layer are more likely to have only one thick primary dendrite emerging from the cell body, as described previously for granule neurons in adult rodents (Fricke, 1975; Desmond and Levy, 1982; Green and Juraska, 1985; Claiborne et al., 1990). In the developing rat, the primary dendrites of both the superficial and the deep granule cells tend to branch at or near the top of stratum granulosum, and the diameters of the dendrites change abruptly at branch points. The majority of branches are still relatively smooth at this age, with only occasional varicosities or growth cones.

170

Fig. 1. Photomontage (A) and drawings (B and C) made from serial micrographs of DiI-labeled granule neurons from a 3day-old (A, B) and a 6-day-old rat (C). Note the presence of immature features on dendrites at this age. Solid-curved arrow, growth cone; solid-straight arrow, varicosity; open-straight arrow, filopodia; long-filled arrow, abrupt diameter change; open arrowheads, continuation of axon; a, axon; c, axon collateral. The montage and the drawings are shown at the same magnification. Scale bar ¼ 50 mm. (Adapted with permission from Jones et al., 2003.)

A few dendrites, however, exhibit numerous growth cones, varicosities, and filopodia. Most investigators report the absence of dendritic spines at these early ages, although Wenzel et al. (1981) noted ‘‘spine-resembling’’ structures at day 0 and a few spines at day 2. Basal dendrites are common on neurons at days 2 and 3. They tend to be

thinner than apical dendrites but exhibit the same immature features, including growth cones, varicosities, and filopodia. Basal dendrites are considered to be an immature feature of granule neurons in the rodent — they are not commonly found on adult granule neurons (Seress and Pokorny, 1981; Lu¨bbers and Frotscher, 1988). As development proceeds, the most mature granule neurons exhibit more extensive dendritic trees with longer branches and a full complement of immature features (Fricke, 1975; Duffy and Teyler, 1978a, b; Wenzel et al., 1981; Lu¨bbers and Frotscher, 1988; Jones et al., 2003). On day 4, dendrites have larger varicosities and an increased number of filopodia as compared to neurons in younger animals (Jones et al., 2003). Diameter changes are abrupt at branch points, and most dendrites terminate before reaching the top of the molecular layer. Basal dendrites are present on the majority of cells. Although Jones et al. (2003) did not report any spines on granule neuron dendrites at day 4, Fricke (1975) noted the occasional spine on Golgi-stained dendrites at this age. On day 5, the oldest granule neurons are characterized by numerous apical branches with many filopodia and varicosities and several basal branches (Fig. 2A; Wenzel et al., 1981; Lu¨bbers and Frotscher, 1988; Zafirov et al., 1994; Jones et al., 2003). Abrupt diameter changes and growth cones are present but are less numerous than on day 4. Spines are scattered throughout the dendritic trees of most of the older neurons on day 5, although they are found in clusters on a few dendrites on a small number of cells (Seress and Pokorny, 1981; Wenzel et al., 1981; Lu¨bbers and Frotscher, 1988; Zafirov et al., 1994; Jones et al., 2003). Wenzel et al. (1981) characterized granule neurons between 5 and 8 days of age as intermediate neuroblasts and considered this period to be the second stage of granule neuron development. The most mature granule neurons observed on day 6 differ somewhat from those in the younger animals (Fig. 1C; Jones et al., 2003). Growth cones and varicosities are smaller and less numerous, whereas the number of filopodia is greater (Fig. 2B). Abrupt diameter changes and basal dendrites are less frequent. Most dendrites on these neurons reach the hippocampal fissure at the top of the

171

Fig. 2. Stacked series of individual images taken with a confocal laser-scanning microscope. Dendrites are from a 5-dayold rat (A) and a 6-day-old rat (B). The dendrite shown in B is from one of the most mature neurons seen on day 6. Although many spines are present, immature features are still observed on day 6. Open arrow, filopodia; filled arrow, spines. Scale bar ¼ 10 mm. (Adapted with permission from Jones et al., 2003.)

molecular layer, and spines are present in varying densities and on varying numbers of dendritic branches. Jones et al. (2003) observed the first adult-like granule neurons on day 7 (Fig. 3). These neurons were considered to be adult-like because they exhibited the three primary characteristics of adult granule neurons: they were devoid of immature features except for occasional varicosities or filopodia, the vast majority of the dendrites reached the top of the molecular layer, and all dendrites were covered with spines. The cell bodies of the adult-like granule neurons were located towards the top of the granule cell layer in the suprapyramidal blade and were most likely between 7 and 10 days old at this time. Other investigators had suggested that adultlike granule neurons were present at about this time.

Fricke (1975) illustrated a granule neuron with numerous spines from an 8-day-old rat. Wenzel et al. (1981) noted that granule neurons in 10-day-old rats resembled adult neurons, although they reported the presence of growth cones and varicosities and suggested that such features were indicative of continued dendritic growth. They considered granule cells at this age to be in the third stage of development and labeled them maturing or young neurons. Zafirov et al. (1994) reported that some granule neurons display extensive dendritic trees with spines and without immature features on day 10. Similarly, Liu et al. (2000) described granule neurons with at least some adult-like features on day 7 and illustrated an adult-like granule cell from a 12-day-old rat. Thus it is now clear that a small proportion of the oldest granule neurons exhibit adult-like morphological features by the end of the first postnatal week in the rat. These data, in combination with results from other studies, suggest that the dentate gyrus may be functional at this early age. For example, our in vivo studies demonstrated that longterm potentiation (LTP) and long-term depression (LTD) can be elicited at medial perforant path synapses onto the granule neurons at day 7 (O’Boyle et al., 2004), confirming earlier in vitro studies (Duffy and Teyler, 1978b; Trommer et al., 1995). It is also worth noting that the granule cell afferents are in their approximate adult locations at this time, and that the dendritic trees of hilar interneurons are well developed (Cowan et al., 1980; Seay-Lowe and Claiborne, 1992). Furthermore, the axons (or mossy fibers) of the oldest granule neurons reach region CA3 of the hippocampus proper on the day of birth or soon thereafter (Minkwitz, 1976; Stirling and Bliss, 1978; Jones et al., 2003), and granule cell stimulation can induce LTD in CA3 pyramidal neurons by day 7 (Battistin and Cherubini, 1994). Taken together, these data suggest that the traditional view of the dentate gyrus as a ‘‘late developing’’ structure may be incorrect; at least some of the granule neurons are capable of functioning within the hippocampal circuit by the end of the first postnatal week, at about the same time that adult-like properties are first observed in the pyramidal neurons of the hippocampus proper.

172

Fig. 3. Drawing made from serial micrographs of a DiI-labeled granule neuron from a 7-day-old rat. Note the absence of immature features and the prevalence of spines. A varicosity is visible on the axon (short, filled arrow); open arrowhead, continuation of axon. Scale bar ¼ 50 mm. (Adapted with permission from Jones et al., 2003.)

Although granule neurons with adult-like features are present by the end of the first week, it is important to note that granule neuron dendrites continue to elongate during the second postnatal week (Fricke, 1975; Duffy and Teyler, 1978a, b; Seress and Pokorny, 1981; Wenzel et al., 1981; Zafirov et al., 1994). Using one of the first computer-microscope systems designed for quantitative, three-dimensional morphological studies of single neurons, Fricke (1975) analyzed the dendritic trees of developing Golgi-impregnated granule neurons over the first postnatal month in the rat. He found that the total dendritic length (defined as the sum of the lengths of all individual dendritic segments) of a granule neuron increased between day 8 and day 12, from an average of approximately 500 mm on day 8 to an average of approximately 1000 mm on day 12, with a wide variability at both ages. Duffy and Teyler (1978a, b) also reported that granule neurons increased dramatically in size between day 7 and day 14, and Zafirov et al. (1994) showed that total dendritic lengths increased from approximately 300 mm at day 5 to

approximately 465 mm on day 10, with no significant increase between days 10 and 15. As granule neuron dendrites continue to elongate after day 7, spine densities and synaptic contacts also increase. As noted above, spines are observed on granule neurons on days 4 and 5 in the rat, and in some cases, spines have been reported as early as days 2 or 3. Jones et al. (2003) considered spines to be protrusions that were less than 2 mm in length (whereas longer protrusions were considered to be filopodia) and noted that dendritic spines on neonatal granule neurons ranged in shape from short, stubby protrusions, either with or without heads, to those with long, thin necks ending in definitive spine heads (Desmond and Levy, 1985; Trommald and Hulleberg, 1997). Counts of spines on DiI-labeled dendrites located in the middle of the molecular layer revealed an average of 0.40 spines/mm in 5-day-old rats and 0.57 spines/mm in 6-day-old animals (Jones et al., 2003). Densities increased to 0.81 spines/mm by day 7, although they were still far below the values reported for dendrites in the same

173

region in the adult rat (1.66 spines/mm; Desmond and Levy, 1985). Zafirov et al. (1994) also demonstrated that spine densities increased after day 5. These densities, however, were slightly lower than those reported by Jones et al. (2003), perhaps because they calculated spine densities across the entire dendritic tree. Based on measurements of the total dendritic length and the total number of spines per cell, they found that spine densities increased from day 5 through day 15, with 0.11 spines/mm on day 5, 0.47 spines/mm on day 10, and 0.67 spines/mm on day 15. Duffy and Teyler (1978a, b) also reported that the number of spines per granule neuron more than doubled between day 7 and day 14, increasing from approximately 100 spines/cell on day 7 to approximately 255 spines/cell on day 14. Data from the above studies are substantiated by the detailed analysis of spine densities on dendrites of various orders over the course of granule neuron development and maturation done by Wenzel et al. (1981). For example, they reported that densities on 4th order branches increased from 0.05 spines/mm on day 5 to 0.12 spines/mm on day 8 and to 0.18 spines/mm on day 10. Electron microscope studies demonstrate that only a few synapses are present in the molecular layer of the suprapyramidal blade on postnatal day 1 (0.4 synapses/100 mm2; Cowan et al., 1980). There is, however, a dramatic increase in synapses over the first 10 days (Cowan et al., 1980). On days 4 and 5, there are approximately 2–3 synapses/ 100 mm2 and, by day 10, densities have increased to 11.0 synapses/100 mm2 (Crain et al., 1973; Cowan et al., 1980). In the molecular layer of the infrapyramidal blade, synapses are present on day 5 (0.8 synapses/100 mm2) and synaptic density increases about sixfold by day 10 (6.1 synapses/ 100 mm2; Cowan et al., 1980). In addition to this considerable increase in synapse densities, the molecular layer also increases in volume approximately fourfold between days 5 and 10, resulting in approximately a 16-fold increase in absolute synapse number. It is also worth noting that a high percentage of synapses are found on dendritic shafts during the neonatal period (Cowan et al., 1981). On day 5, about 50% of synapses are onto dendritic shafts whereas on day 10, approximately

40% are onto shafts. The percentage of shaft synapses continues to decline into adulthood, and by day 41, only approximately 10% of synapses are found on dendritic shafts with the remainder contacting dendritic spines. In addition to synapses in the molecular layer, synaptic contacts are found on granule cell bodies in the neonatal rat. A fair proportion is symmetric and stain for glutamate-decarboxylase (Lubbers and Frotscher, 1988; Seress et al., 1989). In summary, during the first week of life, granule neuron dendrites undergo a sequence of morphological changes. Immature features appear and then regress, dendrites elongate, and spines and synapses develop. By day 7, the oldest granule neurons exhibit adult-like characteristics immature features have regressed, the vast majority of dendrites reach the top of the molecular layer, and all dendrites are covered with spines. During the second postnatal week, dendrites continue to elongate and spine and synaptic densities increase dramatically. By day 14, a considerable number of granule neurons have attained adult-like characteristics. It also appears that the large en passant boutons of the mossy fibers have attained an adult-like shape and complexity by day 14, even as the number of boutons continue to increase into young adulthood (Stirling and Bliss, 1978; Amaral and Dent, 1981).

Maturation of granule neuron dendritic trees While granule neurons in 14-day-old rats qualitatively resemble adult cells, quantitative studies suggest that their dendrites continue to change as the animal matures. Early work indicated that the dendritic tree might increase in size after the second week. Fricke (1975) reported that the total dendritic lengths of Golgi-stained granule neurons increased after day 12, reaching adult values of a little over 1500 mm on day 20. Duffy and Teyler (1978a, b) also reported an increase in granule cell dendritic length between days 14 and 30. In addition, they found a slight but not statistically significant decrease between days 30 and 60 and another slight increase at day 210 — the average length at 210 days was approximately the same as that at day 30. It is not clear from the methods, however, whether

174

the reported lengths reflect measurements of the sum of all dendritic branch lengths or the length of the longest dendrite. Duffy and Teyler (1978a, b) state that they measured ‘‘from mid-soma to the most distal dendritic process’’, and the reported values are quite similar to the width of the molecular layer (Rihn and Claiborne, 1990). Thus, these lengths may reflect the distance from the soma to the most distal dendritic tips. Liu et al. (2000), using intracellular labeling techniques, also found that granule cells enlarged after day 14 although the increases in total dendritic length were not statistically significant. Whereas the studies described above suggest that dendritic trees increase in size during the third and fourth weeks, Wenzel et al. (1981) reported that granule cell dendritic length and the degree of branching reached adult proportions at day 15. To resolve this issue, Rihn and Claiborne (1990) used intracellular labeling techniques and three-dimensional analyses to quantify the dendritic trees of only the oldest granule neurons in the suprapyramidal blade between days 14 and 60. The somata of these neurons were located in the top portion of the granule cell layer, between the tip and the middle of the suprapyramidal blade. Neurons from rats of five age groups were examined: the first group included rats between the ages of 14 and 19 days and the oldest group included animals between 50 and 60 days. The dendritic trees from rats in each age group exhibited similar structures, having between 1 and 4 primary apical dendrites that were covered with spines (Fig. 4). Several qualitative differences, however, were apparent between neurons in the youngest rats and those in the oldest animals. For example, neurons in the youngest rats had thicker dendrites in the proximal and middle thirds of the molecular layer, and they exhibited a higher number of short, terminal segments in the distal third of the layer. Analyses of these maturing granule neurons demonstrated that both dendritic growth and regression occurred between days 14 and 60, leading to a conservation of total dendritic length during this period (Rihn and Claiborne, 1990). The molecular layer of the suprapyramidal blade increased by approximately 50% (from an average of 205 to 305 mm) between days 14 and 60, and the vast

majority of dendrites reached the top of the molecular layer in animals of all ages, suggesting that individual dendrites elongated as the animal matured. The number of dendritic segments, however, decreased from an average of 36 segments in the 14- to 19-day-old rats to an average of 28 segments in the 50- to 60-day-old animals, with the majority of the decrease occurring by day 29. (Segments were defined as a length of dendrite between an origin and a branch point, between two branch points, or between a branch point and a distal termination point.) As a consequence of this concurrent branch elongation and loss, the average total dendritic length did not change significantly between days 14 and 60 (3086 and 3417 mm, respectively). A similar process of dendritic loss with no change in total dendritic length has been documented for nonpyramidal neurons in the rat visual cortex (Parnavelas and Uylings, 1980). Thus, although granule cells have attained their adult dendritic lengths by postnatal day 14, dendritic remodeling occurs between days 14 and 60 with the elongation of some branches and the loss of others. In addition, the number of spines and synapses increase (Duffy and Teyler, 1978a, b; Cowan et al., 1981; Wenzel et al., 1981; Zafirov et al., 1994). Duffy and Teyler (1978a, b) showed a slight increase in the number of spines per cell between days 14 and 30. Values remained stable between days 30 and 60 and then increased again between days 60 and 210. Seress and Pokorny (1981) noted that spine densities increased between days 10 and 25, and Zafirov et al. (1994) reported that spine densities (calculated on the basis of total spines per cell divided by the total dendritic length) increased from 0.67 spines/mm at day 15 to 1.16 spines/mm at day 20. Wenzel et al. (1981) found that spine densities increased dramatically on dendrites in almost all branch orders between days 15 and 30. For example, densities increased from approximately 0.2 to 0.59 spines/mm on 4th order branches and from 0.14 to 0.56 spines/mm on 6th order branches. Furthermore, they found that densities also continued to increase slightly after day 30; when all branch orders were considered together, densities increased from approximately 0.5 to approximately 0.6 spines/mm between days 30 and 180. Synaptic density in the molecular layer

175

Fig. 4. Camera lucida drawing of a granule neuron from a 53-day-old rat. Note the significant increase in size as compared to the adult-like neuron from a 7-day-old rat shown in Fig. 3. Scale bar ¼ 50 mm.

also increased with maturation (Cowan et al., 1980). In the molecular layer of the suprapyramidal blade, densities increased from 11 synapses/ 100 mm2 on day 10 to 36 synapses/100 mm2 on day 21. Only a slight increase was seen after day 21; densities were 37 synapses/100 mm2 on day 41. Similarly, the percentage of spine synapses increased considerably from day 10 to 21, but did not exhibit much of a change between days 21 and 41. Spine synapses comprise 47% of all synapses on day 10, 84% on day 21, and 88% on day 41. In summary, the total dendritic lengths of the oldest granule neurons in the suprapyramidal blade appear to be established by day 14. Quantitative changes in the dendritic tree do occur after this time, however; some dendrites elongate while others are lost, thus leading to a conservation of total length. Interestingly, there is one other report of regression in the hippocampal formation during

maturation. The axons or mossy fibers of the granule neurons exhibit filopodial-like extensions that elongate during the first two postnatal weeks, reach a peak in length on day 14, and then decrease to adult lengths by day 28 (Amaral, 1979). The time course of their growth and retraction is remarkably similar to the time course of dendritic branch growth and regression in the oldest granule neurons — whether the two processes are governed by the same or similar mechanisms is not yet known. In this regard, preliminary data from our lab indicate that dendritic branch loss may be affected by incoming neuronal activity. Specifically, we found that blockade of N-methyl-D-aspartate (NMDA) glutamate receptors between days 14 and 24 did not affect dendritic growth, but did result in a decrease in dendritic branch loss in the oldest granule neurons (Blake and Claiborne, 1995). Because branches continued to grow

176

and fewer branches were lost, the granule neurons in the treated animals had greater total dendritic lengths than did those in the controls. It is worth noting that the glutamate released from synapses on granule cell dendrites in the middle third of the molecular layer binds to NMDA receptors, and that the afferents to this region arise from the entorhinal cortex. The entorhinal cortex, in turn, receives inputs from the visual, auditory, and somatosensory cortices, and a number of significant events occur in the maturation of these sensory systems soon after day 14 in the rat. For example, the eyes open approximately at day 15 and adult sensitivities to sound are reached between days 16 and 18 (Tilney, 1933; Crowley and Hepp-Reymond, 1966). Therefore, it is not surprising that granule cell remodeling begins shortly after day 14 and may be governed by afferent inputs from the entorhinal cortex.

Granule neuron dendrites in the adult rat Quantitative data on the dendritic trees of adult granule neurons are based on two- and three-dimensional measurements of both Golgi-impregnated and intracellularly labeled cells. Fricke (1975) was the first to make use of a computermicroscope system to analyze Golgi-stained neurons in three dimensions, whereas Desmond and Levy (1982) developed a novel probabilistic method for quantifying dendritic trees of granule neurons from Golgi-stained tissue in two dimensions. They corrected for cut dendrites at the edge of sections by estimating their branching and termination patterns based on the patterns observed for intact dendrites. Corrected values were compared with data from tracings of six neurons (two from each region) that were followed through serial sections. Only granule neurons located in the middle 80% of the longitudinal extent of the hippocampus and with a minimum of cut dendrites in the proximal third of the molecular layer were included. Claiborne et al. (1990) and Rihn and Claiborne (1990) applied the computational techniques first used by Fricke (1975) to quantify the dendritic trees of intracellularly labeled granule neurons in relatively thick transverse slices (400 mm) of the

hippocampal formation. They analyzed only those neurons that were completely stained, had no cut dendrites in the proximal portion of the molecular layer, and had fewer than two-severed dendrites in the distal two-thirds of the layer. Here we review the quantitative parameters of adult granule neurons reported by the above investigators and by others employing similar techniques. We first discuss the available data on the entire population of granule neurons, and we then review the relationship between the morphology of a granule neuron and the location of its cell body within stratum granulosum and within the two blades of the dentate gyrus. The cell body of an adult granule neuron in the rat is approximately 10 mm wide and approximately 19 mm long (Claiborne et al., 1990). Between 1 and 5 primary apical dendrites exit from the apical pole of the cell body and bifurcate relatively close to the soma (Fig. 5). Fricke (1975) noted that most branching occurs within 100 mm of the soma and within the vicinity of the boundary between stratum granulosum and stratum moleculare. Desmond and Levy (1982) found that the majority of primary segments branch within the granule cell layer and first ninth of the molecular layer. Furthermore, they reported that nearly all first-order branch points occur within 50 mm of the cell body and within the first 30 mm of the molecular layer. Primary dendrites bifurcate into higher order segments, and up to eighth-order branches have been reported, with an average maximum branch order of 5.7 (Claiborne et al., 1990). Granule neurons have a total of approximately 30 dendritic segments, with reported numbers ranging from 22 to 40 (Fricke, 1975; Seress and Pokorny, 1981; Desmond and Levy, 1982; Green and Juraska, 1985; Claiborne et al., 1990; Rihn and Claiborne, 1990). Based on Golgi impregnations, Fricke (1975) reported an average total dendritic length of 1602 mm (range from 773 to 2445 mm) for adult granule neurons, whereas Seress and Pokorny (1981) reported an average of 2405 mm, and Green and Juraska (1985) found total dendritic lengths ranging from 1161 to 1279 mm. After correcting for cut dendrites, Desmond and Levy (1982) calculated an average total dendritic length of 3662 mm

177

Fig. 5. Computer-generated plots of three-dimensional reconstructions of the dendritic trees of granule cells located at various positions along the transverse axis of the dentate gyrus. Each neuron is from a different animal, and the total dendritic length of each dendritic tree is indicated. Note that the total dendritic lengths of the trees in the suprapyramidal blade are greater than those of the trees in the infrapyramidal blade. CA3, field CA3 of the hippocampus proper; GL, granule cell layer. Scale bar ¼ 100 mm. (Adapted with permission from Claiborne et al., 1990.)

for Golgi-stained neurons. Claiborne and colleagues reported similar averages of 3221 and 3417 mm, respectively, for intracellularly labeled granule neurons analyzed in two different studies (Fig. 5; Claiborne et al., 1990; Rihn and Claiborne, 1990). Desmond and Levy (1982) reported that approximately 12% of the total length was found within stratum granulosum, approximately 25% within the proximal third of the molecular layer, and the remaining portion was restricted to the distal two-thirds. Similarly, data from Claiborne et al. (1990) indicated that 30% of the total length was found in the granule cell layer and proximal third of the molecular layer, 30% in the middle third and 40% in the distal third. These results suggest that more dendritic length is available for synapses from the entorhinal afferents that

terminate in the distal two-thirds of the layer than for commissural and associational afferents that make contacts in the inner third of the layer. The majority of synapses onto the granule neuron dendrites occur on spines in the adult. As discussed above, a number of groups have shown that spine and synaptic densities increase throughout granule neuron development and maturation, and their reported values for densities in maturing and young adult rats are reviewed above (Duffy and Teyler, 1978a, b; Cowan et al., 1981; Seress and Pokorny, 1981; Wenzel et al., 1981; Zafirov et al., 1994). Other investigators have reported spine densities only for granule neurons in young adult rats. Fricke (1975) noted that spine densities were low on dendrites of adult granule neurons within the first 20 mm of the soma and within 50 mm of dendritic

178

terminations. He found a maximum of approximately 1.3 spines/mm on dendrites in the remaining portion of the dendritic tree. Desmond and Levy (1985) reported that spines were infrequent on dendrites within stratum granulosum and that there were three major types of spines on adult granule cell dendrites in the molecular layer: stubby, mushroom-shaped, and thin. They calculated spine densities for each type in the proximal, middle, and distal thirds of the molecular layer of the two blades. In both blades and in all three portions of the layer, densities were highest for the thin spines and were lowest for mushroom-shaped protrusions. When all spine types were included and corrections made for obscured spines, densities averaged 1.6 spines/mm on dendrites of neurons in the suprapyramidal blade and 1.3 spines/mm on those in the infrapyramidal blade. These values are similar to those reported by Fricke (1975) but are much greater than those reported by Wenzel et al. (1981; see above) perhaps because different criteria were used for the inclusion of spines (Desmond and Levy, 1985). In summary, the dendritic trees of granule neurons in young adult rats are composed of approximately 30 spine-covered segments and have total dendritic lengths of approximately 3400 mm. It is not surprising that these total lengths are much less than those of the relatively large pyramidal neurons in the hippocampus proper. Reports of the total dendritic lengths of pyramidal neurons in region CA1 vary from an average of 10,800 to 17,400 mm (Ishizuka et al., 1995; Mainen et al., 1996; Pyapali and Turner, 1996; Pyapali et al., 1998; Megias et al., 2001), whereas the total lengths for pyramidal neurons in region CA3 range from a low of 7000 mm for pyramidal neurons in region CA3c to a high of 19,800 mm for those in region CA3a (Ishizuka et al., 1995; Turner et al., 1995; Gonzales et al., 2001). It is also not surprising that the granule neurons are the most electrotonically compact of the three major classes of hippocampal neurons (Carnevale et al., 1997). Given the gradients of granule cell neurogenesis, it is of interest that several parameters of granule neuron dendrites are correlated with the location of the parent cell body in the granule cell layer — both along the transverse axis of the dentate gyrus

and within the depth of the granule cell layer. Data demonstrate that the dendritic trees of neurons in the suprapyramidal blade are larger than those in the infrapyramidal blade (Fig. 5). Fricke (1975) reported that suprapyramidal neurons have greater total dendritic lengths than do those in the infrapyramidal blade (1674 mm vs. 1482 mm), and Desmond and Levy (1982) confirmed this result for Golgi-stained granule neurons that were traced through multiple sections (3107 mm for suprapyramidal neurons vs. 2078 mm for infrapyramidal blade cells). Similarly, Claiborne et al. (1990) showed that suprapyramidal granule neurons have greater total dendritic lengths (3478 mm vs. 2793 mm), more dendritic segments (31 vs. 27), and wider dendritic spreads in the transverse plane of the hippocampal formation (347 mm vs. 288 mm) than do the infrapyramidal blade neurons. These results demonstrate that suprapyramidal granule neurons have more dendritic length available for afferent contacts than do neurons in the opposite blade. In addition, Desmond and Levy (1985) reported that spine density is greater on dendrites of suprapyramidal neurons (1.6 spines/mm vs. 1.3 spines/mm), and this result, combined with the increased length of the suprapyramidal neurons, suggests that neurons in this blade receive many more synaptic contacts than do neurons in the infrapyramidal blade. Overall, these data indicate that the dendritic trees of granule neurons in the suprapyramidal blade are likely to be larger than those in the infrapyramidal blade and to receive more afferent contacts. It is of considerable interest that the axon trajectories of the granule neurons also vary with granule cell location. Axons of granule neurons located toward the tip of the suprapyramidal blade traverse stratum radiatum of region CA3 before entering stratum lucidum, avoiding both the hilar region and the proximal part of the pyramidal cell layer, whereas axons of neurons in the crest and the infrapyramidal blade travel through the hilar region and contact pyramidal neurons in the proximal portion of field CA3 (Claiborne et al., 1986). It is likely that such morphological distinctions in the granule neurons, in combination with physiological differences (Scharfman et al., 2002; Chawla et al., 2005), may lead to functional differences

179

between the two blades. This idea is best exemplified by recent results showing that induction of the activity-regulated, immediate early gene Arc following behavioral stimulation occurs in the suprapyramidal blade, but not in the infrapyramidal blade (Chawla et al., 2005). Not only do some morphological parameters vary according to the location of the cell body along the transverse axis of the dentate gyrus, several dendritic parameters also differ according to the depth of the parent cell body in the granule cell layer. Neurons with somata located at the top of the layer tend to have multiple primary branches whereas those with cell bodies located deeper in the layer tend to have only one primary dendrite (Fricke, 1975; Seress and Pokorny, 1981; Desmond and Levy, 1982; Green and Juraska, 1985; Claiborne et al., 1990). Granule neurons in the superficial half of the granule cell layer also exhibit nearly twice the density of axosomatic synapses as do cells in the bottom half of the layer (Lee et al., 1982). In addition, Green and Juraska (1985) showed that Golgi-stained granule neurons with somata in the superficial third of the granule cell layer had greater maximum widths in the transverse plane of the dentate gyrus, more branches in orders 1 through 3, and fewer branches in orders 4 through 6. These neurons also had greater total dendritic lengths (1279 vs. 1161) than did deeper neurons; 80% of the neurons in their sample were located in the suprapyramidal blade. In contrast, Claiborne et al. (1990) did not find a statistically significant difference between the average total lengths of neurons with somata in the superficial half of the granule cell layer and of those with somata in the bottom half of the layer in either blade. When only neurons in the suprapyramidal blade were analyzed, however, they found that superficial granule neurons had more primary dendrites (2.4 vs. 1.5) and larger transverse spreads (378 mm vs. 293 mm) than deeper neurons. In addition, 42% of the total length of superficial cells was in the distal third of the layer as compared to 37% of the length of the deeper cells. For neurons in the infrapyramidal blade, the only significant difference was that superficial cells had larger transverse spreads than did deeper neurons (311 mm vs. 244 mm).

Thus, a number of structural differences exist between the dendritic trees of neurons located in the superficial portion of stratum granulosum and those located in the deeper aspects of the layer. At least one of the differences may have functional consequences. Carnevale et al. (1997) demonstrated that the single primary dendrites of deeper neurons attenuated voltage signals spreading from the soma out to the dendrites, thereby reducing the effect of somatic events, including action potentials, on molecular layer synapses. Given that adult-generated granule neurons migrate into the lower portion of the granule cell layer, it will be of interest to determine whether other functional differences correlate with the depth of granule cell bodies in stratum granulosum in adult animals.

Summary Granule neurons in the dentate gyrus of the rat undergo periods of development and maturation before reaching their final adult form and size. In the rat, the primary period of neurogenesis begins during late embryogenesis and continues over the first two weeks of life. The oldest granule neurons occupy the suprapyramidal blade whereas later generated cells form the infrapyramidal blade. On the day of birth, granule neurons in the suprapyramidal blade exhibit a few sparse dendrites. Rudimentary trees appear over the next few days and by day 4, dendritic branching is quite extensive. Immature features are abundant at this time and include abrupt diameter changes, varicosities, filopodia, growth cones, and basilar dendrites. On days 5 and 6, elaborate dendritic trees are present, there is a reduction in the frequency of immature features, and some spines are visible on the most mature cells. On day 7, the first few adult-like trees are present on granule neurons in the suprapyramidal blade. Their dendrites reach the top of the molecular layer and no longer exhibit immature features except for an occasional filopodium or varicosity. All of the branches are covered with spines. Physiological recordings demonstrate that some granule neurons are functional at this time, suggesting that the dentate gyrus may be incorporated into the hippocampal circuit by the end of the first week.

180

After day 7, the dendritic trees increase in size, and, by day 14, numerous adult-like granule cells are present. The oldest granule neurons then undergo a process of refinement. Many dendrites continue to elongate during this period, and other branches are lost; the simultaneous growth and regression results in a conservation of total dendritic length after day 14. Spine densities, however, continue to increase. Adult granule neurons in the rat have approximately 30 dendritic segments and total dendritic lengths of approximately 3400 mm. A variety of morphological features are correlated with the location of the granule cell body, both along the transverse axis of the dentate gyrus and within the depth of the granule cell layer. Recent evidence suggests that there are functional distinctions between the two blades of the dentate gyrus; it will be of considerable interest to determine whether differences in granule cell morphologies underlie these distinctions.

Acknowledgment This work was supported by NIH grant #GM08194 to BJC.

References Altman, J. and Das, G.D. (1965) Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol., 124(3): 319–335. Amaral, D.G. (1979) Synaptic extensions from the mossy fibers of the fascia dentata. Anat. Embryol. (Berl.), 155: 241–251. Amaral, D.G. and Dent, J.A. (1981) Development of the mossy fibers of the dentate gyrus: I. A light and electron microscopic study of the mossy fibers and their expansions. J. Comp. Neurol., 195: 51–86. Battistin, T. and Cherubini, E. (1994) Developmental shift from long-term depression to long-term potentiation at the mossy fibre synapses in the rat hippocampus. Eur. J. Neurosci., 6(11): 1750–1755. Bayer, S.A. and Altman, J. (1974) Hippocampal development in the rat: cytogenesis and morphogenesis examined with autoradiography and low-level X-irradiation. J. Comp. Neurol., 158(1): 55–79. Blake, N.M.J. and Claiborne, B.J. (1995) Possible role of the Nmethyl-D-aspartate (NMDA) receptor in the remodeling of dendritic arbors during development. Soc. Neurosci. Abstr., 21: 1292.

Carnevale, N.T., Tsai, K.Y., Claiborne, B.J. and Brown, T.H. (1997) Comparative electrotonic analysis of three classes of rat hippocampal neurons. J. Neurophysiol., 78(2): 703–720. Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P., Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton, B.L. and Barnes, C.A. (2005) Sparse, environmentally selective expression of Arc RNA in the upper blade of the rodent fascia dentata by brief spatial experience. Hippocampus, 15(5): 579–586. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus. J. Comp. Neurol., 246: 435–458. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1990) Quantitative, three-dimensional analysis of granule cell dendrites in the rat dentate gyrus. J. Comp. Neurol., 302: 206–219. Cowan, W.M., Stanfield, B.B. and Amaral, D.G. (1981) Further observations on the development of the dentate gyrus. In: Cowan W.M. (Ed.), Studies in Developmental Neurobiology: Essays in Honor of Viktor Hamburger. Oxford University Press, New York, pp. 395–435. Cowan, W.M., Stanfield, B.B. and Kishi, K. (1980) The development of the dentate gyrus. Curr. Top. Dev. Biol., 15: 103–157. Crain, B., Cotman, C., Taylor, D. and Lynch, G. (1973) A quantitative electron microscopic study of synaptogenesis in the dentate gyrus of the rat. Brain Res., 63: 195–204. Crowley, D.E. and Hepp-Reymond, M.C. (1966) Development of cochlear function in the ear of the infant rat. J. Comp. Physiol., 62: 427–432. Desmond, N.L. and Levy, W.B. (1982) A quantitative anatomical study of the granule cell dendritic fields of the rat dentate gyrus using a novel probabilistic method. J. Comp. Neurol., 212: 131–145. Desmond, N.L. and Levy, W.B. (1985) Granule cell dendritic spine density in the rat hippocampus varies with spine shape and location. Neurosci. Lett., 54: 219–224. Duffy, C.J. and Teyler, T.J. (1978a) Development of habituation in the dentate gyrus of rat: physiology and anatomy. Brain Res. Bull., 3: 305–310. Duffy, C.J. and Teyler, T.J. (1978b) Development of potentiation in the dentate gyrus of rat: physiology and anatomy. Brain Res. Bull., 3: 425–430. Fricke, R.A. (1975) Studies on the morphology and development of the hippocampus and dentate gyrus. Ph.D. Dissertation, Washington University, St. Louis, MO. Gonzales, R.B., DeLeon Galvan, C.J., Rangel, Y.M. and Claiborne, B.J. (2001) Distribution of thorny excrescences on CA3 pyramidal neurons in the rat hippocampus. J. Comp. Neurol., 430(3): 357–368. Green, E.J. and Juraska, J.M. (1985) The dendritic morphology of hippocampal dentate granule cells varies with their position in the granule cell layer: a quantitative Golgi study. Exp. Brain Res., 59: 582–586. Ishizuka, N., Cowan, W.M. and Amaral, D.G. (1995) A quantitative analysis of the dendritic organization of pyramidal cells in the rat hippocampus. J. Comp. Neurol., 362: 17–45.

181 Jones, S.P., Rahimi, O., O’Boyle, M.P., Diaz, D.L. and Claiborne, B.J. (2003) Maturation of granule cell dendrites after mossy fiber arrival in hippocampal field CA3. Hippocampus, 13: 413–427. Kaplan, M.S. and Hinds, J.W. (1977) Neurogenesis in the adult rat: electron microscopic analysis of light radioautographs. Science, 197: 1092–1094. Lee, K.S., Gerbrandt, L. and Lynch, G. (1982) Axo-somatic synapses in the normal and x-irradiated dentate gyrus: factors affecting the density of afferent innervation. Brain Res., 249: 51–56. Liu, X., Tilwalli, S., Ye, G., Lio, P.A., Pasternak, J.F. and Trommer, B.L. (2000) Morphologic and electrophysiologic maturation in developing dentate gyrus granule cells. Brain Res., 856: 202–212. Liu, Y.B., Lio, P.A., Pasternak, J.F. and Trommer, B.L. (1996) Developmental changes in membrane properties and postsynaptic currents of granule cells in rat dentate gyrus. J. Neurophysiol., 76(2): 1074–1088. Loy, R., Lynch, G. and Cotman, C.W. (1977) Development of afferent lamination in the fascia dentata of the rat. Brain Res., 121: 229–243. Lu¨bbers, K. and Frotscher, M. (1988) Differentiation of granule cells in relation to GABAergic neurons in the rat fascia dentata. Combined Golgi/EM and immunocytochemical studies. Anat. Embryol. (Berl.), 178: 119–127. Mainen, Z.F., Carnevale, N.T., Zador, A.M., Claiborne, B.J. and Brown, T.H. (1996) Electrotonic architecture of hippocampal CA1 pyramidal neurons based on three-dimensional reconstructions. J. Neurophysiol., 76(3): 1904–1923. Megias, M., Emri, Z., Freund, T.F. and Gulyas, A.I. (2001) Total number and distribution of inhibitory and excitatory synapses on hippocampal CA1 pyramidal cells. Neuroscience, 102: 527–540. Minkwitz, H.G. (1976) Zur Entwicklung der Neuronenstruktur des Hippocampus wa¨hrend der pa¨r- und postnatalen Ontogenese der Albinoratte. J. Hirnforsch., 17: 213–231. O’Boyle, M.P., Do, V., Derrick, B.E. and Claiborne, B.J. (2004) In vivo recordings of long-term potentiation and long-term depression in the dentate gyrus of the neonatal rat. J. Neurophysiol., 91: 613–622. Parnavelas, J.G. and Uylings, H.B. (1980) The growth of nonpyramidal neurons in the visual cortex of the rat: a morphometric study. Brain Res., 193(2): 373–382. Pyapali, G.K., Sik, A., Penttonen, M., Buzsaki, G. and Turner, D.A. (1998) Dendritic properties of hippocampal CA1 pyramidal neurons in the rat: intracellular staining in vivo and in vitro. J. Comp. Neurol., 391: 335–352. Pyapali, G.K. and Turner, D.A. (1996) Increased dendritic extent in hippocampal CA1 neurons from aged F344 rats. Neurobiol. Aging, 17(4): 601–611. Rihn, L.L. and Claiborne, B.J. (1990) Dendritic growth and regression in rat dentate granule cells during late postnatal development. Brain Res. Dev. Brain Res., 54: 115–124.

Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and Goodman, J.H. (2002) Structural and functional asymmetry in the normal and epileptic rat dentate gyrus. J. Comp. Neurol., 454: 424–439. Schlessinger, A.R., Cowan, W.M. and Gottlieb, D.I. (1975) An autoradiographic study of the time of origin and the pattern of granule cell migration in the dentate gyrus of the rat. J. Comp. Neurol., 159(2): 149–175. Seay-Lowe, S.L. and Claiborne, B.J. (1992) Morphology of intracellularly labeled interneurons in the dentate gyrus of the immature rat. J. Comp. Neurol., 324: 23–36. Seress, L., Frotscher, M. and Ribak, C.E. (1989) Local circuit neurons in both the dentate gyrus and Ammon’s horn establish synaptic connections with principal neurons in five day old rats: a morphological basis for inhibition in early development. Exp. Brain Res., 78: 1–9. Seress, L. and Pokorny, J. (1981) Structure of the granular layer of the rat dentate gyrus. A light microscopic and Golgi study. J. Anat., 133(2): 181–195. Stirling, R.V. and Bliss, T.V. (1978) Hippocampal mossy fiber development at the ultrastructural level. Prog. Brain Res., 48: 191–198. Tilney, F. (1933) Behavior in its relation to the development of the brain. II. Correlation between the development of the brain and behavior in the albino rat from embryonic states to maturity. Bull. Neurol. Inst. N.Y., 3: 252–358. Trommald, M. and Hulleberg, G. (1997) Dimensions and density of dendritic spines from rat dentate granule cells based on reconstructions from serial electron micrographs. J. Comp. Neurol., 377(1): 15–28. Trommer, B.L., Kennelly, J.J., Colley, P.A., Overstreet, L.S., Slater, N.T. and Pasternak, J.F. (1995) AP5 blocks LTP in developing rat dentate gyrus and unmasks LTD. Exp. Neurol., 131: 83–92. Turner, D.A., Li, X.G., Pyapali, G.K., Ylinen, A. and Buzsaki, G. (1995) Morphometric and electrical properties of reconstructed hippocampal CA3 neurons recorded in vivo. J. Comp. Neurol., 356: 580–594. Wenzel, J., Stender, G. and Duwe, G. (1981) Zur Entwicklung der Neuronenstruktur der Fascia dentata bei der Ratte. Neurohistologish-morphometrische, ultrastrukturelle and experimentelle Untersuchungen. J. Hirnforsch., 22: 629–683. Ye, G.L., Liu, X.S., Pasternak, J.F. and Trommer, B.L. (2000) Maturation of glutamatergic neurotransmission in dentate gyrus granule cells. Brain Res. Dev. Brain Res., 124: 33–42. Zafirov, S., Heimrich, B. and Frotscher, M. (1994) Dendritic development of dentate granule cells in the absence of their specific extrinsic afferents. J. Comp. Neurol., 345: 472–480. Zhao, C., Teng, E.M., Summers Jr., R.G., Ming, G.L. and Gage, F.H. (2006) Distinct morphological stages of dentate granule neuron maturation in the adult mouse hippocampus. J. Neurosci., 26(1): 3–11.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 11

Physiological studies of human dentate granule cells Anne Williamson1, and Peter R. Patrylo2 1

Department of Neurosurgery, Yale University School of Medicine, New Haven, CT 06518, USA 2 Department of Physiology, Southern Illinois University, Carbondale, IL 62901, USA

Abstract: The availability of human hippocampi obtained through surgery (usually for treatment of temporal lobe epilepsy) has allowed us to investigate the properties of the human dentate in a way that cannot be done with other brain regions. The dentate has been the primary focus of these studies because of its relative preservation in all patient specimens. Moreover, there is extensive synaptic reorganization of numerous neurotransmitter systems in this the fascia dentate (dentate gyrus and the hilus) in humans with specific forms of TLE. These changes are not evident in tissue from patients with seizure that begin outside the hippocampus, and, as a result, this tissue provides an invaluable resource for comparisons. Physiological data using both slices and acutely dissociated cells demonstrate that the granule cells have membrane properties similar to those of rodents although there are specific changes that appear to be associated with seizures. Similarly, in the non-sclerotic hippocampi, the synaptic properties are similar to those reported in rodents. There are also a number of parallels between the findings in humans and in status animal models of temporal lobe epilepsy. This review will cover analyses of membrane properties as well as of glutamatergic, GABAergic, and neuromodulatory systems. Thus, while there are a number of issues that invariably arise with studies of pathological human tissue, this tissue is ideally suited to verify and refine animal models of temporal lobe epilepsy. In addition, one can argue that human tissue provides the only resource to evaluate the ways that granule cells recorded from laboratory animals approximate human granule cell physiology. Keywords: temporal lobe epilepsy; hippocampal slice; mossy fiber sprouting; neuropeptides; inhibition the work of Schwartzkroin and Prince (1976), and extending to the present day. The majority of these studies have used physiological methods, using both slices and acutely dissociated cells. More recent studies have begun to use molecular biological and imaging approaches to expand our understanding of the human dentate gyrus. This review will focus on the physiology of the human dentate gyrus with special attention to the changes observed in patients with the most common pathology associated with temporal lobe epilepsy, medial temporal lobe sclerosis (MTS) (de Lanerolle et al., 2003).

Introduction We have a greater understanding of the physiological properties of the dentate gyrus than any other region of the human brain because of the availability of tissue resected from patients with pharmacoresistant temporal lobe epilepsy (TLE). Numerous groups have studied resected human hippocampi over the past 20 years, starting with

Corresponding author. Tel.: +1 203 785 5327; Fax: +1 203 737 -2159; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63011-8

183

184

While there are a number of issues associated with studies of human material, we argue that use of human tissue is useful because it provides virtually the only resource for comparison to rodent tissue. In addition, tissue from patients with intractable TLE provides a critical foundation for the development of accurate animal models of TLE. The major issues to consider are variability of the patient population and the lack of true controls. Other, more minor concerns, include variability in the surgical resection of the tissue and hence tissue viability, and the fact that the area of the hippocampus that is resected may be limited. The vast majority of hippocampi that are available for physiological study are resected for treatment of medically intractable seizures and thus are not from normal brain. However, within the population of TLE patients, there are distinct groups that can provide useful comparisons. While rapidly obtained autopsy material can serve as a nonpathological control for anatomical studies, true control material for physiological studies is extremely rare. The bulk of available tissue has the pathological features of MTS which include profound neuronal loss with the trend being CA1> hilusCA3>CA2DG. In severely sclerotic hippocampi, up to 80% of neurons are lost; in a representative series of patients, the mean neuronal loss (for all hippocampal regions) was 64% normalized to autopsy controls (Kim et al., 1990; de Lanerolle et al., 2003). There is a concomitant reactive gliosis in MTS hippocampi; however, the role of reactive astrocytes in epileptic tissue is only now being investigated (Binder and Steinhauser, 2006). The best available comparison material for typical MTS patients is those patients with no evidence of a lesion and where there is no atrophy detectable on MRI. There is no consensus on the terminology for this patient group. For this discussion, these patients will be referred to as paradoxical TLE (PTLE). This group of patients provides the most reasonable comparison material for the MTS tissue for several reasons. First, the confounds of previous seizure experience and exposure to anticonvulsants can be nullified with this group. Second, there is little region-specific cell

loss and no evidence for synaptic reorganization. See Cohen-Gadol et al. (2005) and de Lanerolle et al. (2003) for detailed descriptions of this pathology. Figure 1A shows quantitative cell counts for autopsy controls, MTS and PTLE material in several hippocampal regions. Note that there is no significant difference between the cell counts for the PTLE group and the non-pathological autopsy controls in all four regions sampled. A second comparison population includes hippocampi from patients with extrahippocampal lesions. In these cases, the hippocampus is removed to limit seizure recurrence. This tissue is referred to as Mass-Associated TLE (MaTLE). The epileptogenic lesions include gliomas (usually oligodendrogliomas or low grade astrocytomas), cavernous malformations, and cysts (Beaumont and Whittle, 2000). We have noted no physiological differences in tissue from these two nonsclerotic populations and therefore will group them. For clarity, we will refer to these as comparison hippocampi. Similarly, granule cells from MTS or comparison hippocampi will be referred to as MTS or comparison granule cells.

Membrane properties of human dentate granule cells There are several published reports, which describe the membrane properties of human dentate granule cells relative to rodents, conducted in slices. Based on routinely obtained physiological measures, human granule cells can be successfully recorded in slices and are comparable in their resting membrane potentials and input resistances to those of rodents (Isokawa et al., 1991; Williamson et al., 1993, 1999; Dietrich et al., 1999a). Thus, any differences between rodent and human tissue cannot be attributed solely to slice ‘‘health.’’ However, there are differences in the firing properties of these cells that may be a feature of human tissue. Williamson et al. (1993) showed that, using sharp electrode recordings, there was reduced spike frequency accommodation between guinea pig and human granule cells from nonsclerotic hippocampi. No other changes were noted between the rodent and comparison tissue.

185

Fig. 1. Characteristics of epileptic human dentate. A compares the cell loss in different pathologies between the dentate granule cells, CA fields and the hilus. Note that the dentate is less affected than the CA1 region in the MTS tissue, and that there is little difference between the autopsy controls and the comparison material. Data courtesy of Dr. Jung H. Kim. B shows the presence of graded bursting activity in response to molecular layer stimulation (stimulus strength in mA). The arrows show examples of spontaneous excitatory activity, which is another common physiological finding in MTS tissue that is rare in comparison material. The resting membrane potential in the comparison granule cell was 58 mV and 55 in the MTS tissue. Spontaneous events are likely to be glutamatergic, not GABAergic, because the RMP is depolarized to ECl.

186

These data suggest that there may be some differences in intracellular calcium dynamics between the two species, but could also reflect changes due to anticonvulsant drug use. A similar study was carried out by Dietrich et al. (1999) using MTS tissue. In this study, two populations of granule cells were described that could be discriminated by the characteristics of the fast afterhyperpolarization (AHP) and the loss of a medium duration AHP. There were predictable differences in spike frequency adaptation between the two groups of cells that correlated with the extent of the AHPs. The majority of the cells that had shorter AHPs also demonstrated synaptic hyperexcitability (see below); however, 21% had physiological features comparable to rodents (and primates (St John et al., 1997)) indicating that some of the granule cells in sclerotic hippocampi retain normal membrane and synaptic properties. Other biophysical properties of human hippocampal neurons have been studied using acutely dissociated cells. This preparation is advantageous because it allows high quality voltage clamp recordings; however, the possible effect of the enzyme treatment on extracellular proteins is a potential limitation. Using this approach provides the best means to address the properties of Na+, Ca2+, and K+ channels in human granule cells. Na+ currents Several studies have examined the properties of Na+ currents in acutely isolated granule cells from epileptic patients. It is notable that there were very large currents in these cells compared to cells from rat tissue or from non-focal neocortical tissue. The other noteable feature of the Na+ currents was that there was a component of the current that had a slow recovery from inactivation (Reckziegel et al., 1998). In a subsequent study of granule cells from epileptic patients, it was shown that there was a limited effect of the anticonvulsant carbemazepine on these channels (Reckziegel et al., 1999). Comparable studies have not been conducted in non-epileptic cells, so it is difficult to

determine if these differences in Na+ channel properties reflect treatment with anticonvulsants, many of which act on Na+ channels. The changes in the Na+ channel transcripts that have been reported in human granule cells may account for some of these physiological changes (Lombardo et al., 1996). K+ currents With few exceptions, there are few differences between the K+ channels in human granule cells and those in rats. These currents could be separated into at least four types by their kinetic and pharmacological properties. These include at least one voltage-dependent current similar to those observed in mammalian hippocampal neurons, and two Ca2+-dependent K+ currents that most probably correspond to SK- and BK-type currents. The sole significant difference was that a classical A-type current could be detected in some patients with MTS but not in comparison material (Beck et al., 1996). Ca2+ currents Beck et al. (1998, 1999) and Nagerl et al. (2000) demonstrated that granule cells express N-, L-, and T-type Ca2+ channels based on their pharmacological sensitivity. Similar to studies in epileptic rats, the only significant change in these properties was an increase in the T-type Ca channel density. This increased density of a low threshold Ca2+ current may lead to greater Ca2+ entry during synaptic activity, and thus modulate a variety of systems, including GABAA receptormediated inhibition (Isokawa, 1998). HCN currents Finally, the ability to generate rhythmic activity in neurons can be regulated by the hyperpolarization-activated cyclic nucleotide gated current (HCN). While a detailed analysis of this current has not been performed in human tissue, there is evidence for enhanced HCN mRNA in human

187

dentate granule cells from sclerotic hippocampi. This finding parallels that seen in the rat (Bender et al., 2003). The consequences of excess HCN could lead to granule cells with physiological properties closer to CA1 pyramidal cells, i.e., more depolarized membrane potentials, and cells that are more likely to have rebound activity following recovery from a hyperpolarization. Moreover, in conjunction with the enhanced T channel density described above, granule cells from MTS hippocampi will be more prone to oscillatory activity, analogous to thalamic relay cells. While prominent oscillations have not been reported in human granule cells, subtle changes in ion channels may allow for enhanced neural synchronization. Synaptic properties Excitatory synaptic activity A consistent observation in granule cells from patients with MTS is that bursts, i.e., a depolarizing envelope with three or more action potentials riding on it, can be evoked by stimulation in the outer molecular layer. This stimulus will primarily activate the perforant path, but will also activate any recurrent mossy fibers. It is difficult to selectively activate the medial vs. lateral perforant path in the human, so we refer below to the site of stimulation, rather than a specific fiber branch of the perforant pathway. The percentage of granule cells that display the stimulus-induced bursts varies widely between patients, as well as in published reports (Franck et al., 1995; Williamson et al., 1995a; Cohen-Gadol et al., 2005). In our sharp electrode recordings from 190 cells (75 patients), we noted bursting in 70/102 (68.6%) cells from 41 patients with MTS. In contrast, this type of activity was occurred in only 15/72 (20.5%) cells recorded from comparison hippocampi. However, it is important to note that we never saw greater than a three action potential burst in the comparison material while four or more were routinely observed in the MTS tissue. Thus, in a large sample, bursting is a characteristic of hippocampi from patients with MTS, but is not exclusively found in this group of patients.

Examples of this graded bursting activity are shown in Figs. 1B and 2A. The evoked bursting activity in slices is dependent on NMDA receptor activation (Isokawa and Levesque, 1991; Franck et al., 1995; Cohen-Gadol et al., 2005). In our studies, we noted a 53.6% and 47.5% decrease in the area of the evoked response in MTS granule cells following application of the NMDA receptor antagonists amino phosphonovaleric acid (APV; 30 mM) or Zn2+ (200 mM), respectively. In contrast, there was only a 6.4% change after APV and 18.5% decrease following Zn2+in the comparison tissue (Williamson and Spencer, 1995). These data are consistent with voltage clamp recordings in MTS slices (Isokawa and Levesque, 1991) showing an APV-sensitive component of the evoked EPSC. In addition, studies using acutely dissociated granule cells described prolonged NMDA channel openings in cells from MTS patients. In this study it was noteable that calcineurin did not modulate the open channel conductance in these cells, similar to kindled rat granule cells (Lieberman and Mody, 1999). This may be related to the loss of calbindin D28 K in these cells (Magloczky et al., 1997).

Role of mossy fiber sprouting in granule cell excitability Robust recurrent mossy fiber sprouting into the inner third of the molecular layer in MTS hippocampi was described by a number of authors in the early 1980s using a variety of techniques, including Timm stain and dynorphin immunoreactivity (de Lanerolle et al., 1989; Sutula et al., 1989; Houser et al., 1990). This is not seen in the comparison tissue or in autopsy material. However, defining the role of the sprouted mossy fibers in epileptogenesis continues to remain controversial (e.g., Patrylo and Dudek, 1998; Wuarin and Dudek, 2001; Harvey and Sloviter, 2005; Frotscher et al., 2006; Sloviter et al., 2006). We have used a variety of approaches to investigate the possible effect of mossy fiber sprouting on human granule cell excitability. As in rat models of TLE with mossy fiber sprouting, robust

188

Fig. 2. Granule cell hyperexcitability is associated with mossy fiber sprouting. A shows representative traces comparing the response to molecular layer (orthodromic) and hilar (antidromic) stimulation in a patient with MTS. Note that there is a greater response to maximal hilar stimulation than to stimulation of outer molecular layer. Moreover, note that hilar stimulation produced a synaptic response prior to an antidromic spike. These data support the hypothesis that mossy fibers are primarily excitatory in resected tissue. B shows Timm staining from the region of the dentate where the cell shown in A was recorded. Note the robust supergranular labeling indicating the presence of sprouted fibers. (C1) Stimulation protocols have the problem of activating multiple fiber systems. Therefore, in the presence of 30 mM bicuculline, we applied a bolus of glutamate to the granule cell layer at a distance from the recorded cell and were able to evoke a long lasting barrage of EPSPs. The glutamate was applied at a distance from the recorded cell and thus we were not observing direct glutamate effects. C2 shows that a similar response was not observed when the glutamate was applied to the hilus. (See Color Plate 11.2 in color plate section.)

synaptic events can be evoked from granule cells with hilar (antidromic) stimulation in MTS hippocampi, and this is likely to represent the antidromic activation of recurrent mossy fibers,

which then form functional excitatory contacts onto granule cells (Masukawa et al., 1992). An example comparing responses to orthodromic and antidromic stimulation in the same cell, recorded

189

from sclerotic hippocampus, is shown in Fig. 2A. This hippocampus was characterized by robust mossy fiber sprouting, as shown in Fig. 2B. Note that progressively longer and more complex events occurred following the antidromic spike, consistent with the hypothesis that there is a complex network of recurrent fibers that can easily activate a given granule cell. A second approach was glutamate application, using microdrops, to the granule cells located at a distance from the recorded neuron. Presumably the granule cells would be interconnected by recurrent circuits in the MTS tissue, and the GABAA receptor antagonist bicuculline (30 mM) was added to ensure that inhibition would not mask the effects of recurrent excitatory circuits. As shown in Fig. 2C, the microdrop application was followed by a transient increase in the frequency of spontaneous activity (presumably EPSPs). This type of response was not observed when a glutamate was applied to the hilus. These data demonstrate that there are robust polysynaptic connections made by the recurrent mossy fibers onto granule cells in MTS hippocampi. We were unable to evoke synaptic activity with hilar stimulation or with glutamate microdrops applied to the granule cell layer in the dentate of comparison hippocampi, supporting the hypothesis that this activity is mediated by synaptic reorganization. Moreover, they also indicate that the excitatory mossy cells of the hilus are not able to induce this type of graded excitation, presumably because of the longitudinal orientation of their axonal arbors. However, little is known about mossy cells in the human dentate. An important caveat to the results described above is that effects of hilar stimulation, electrical or chemical, could have been due to activation of basal dendrites of granule cells. Primates are known to have basal dendrites (see chapter by L. Seress in this volume). Approximately 40% of human granule cells from epileptic tissue have been reported to have basal dendrites, and they extend into the hilus, where they could be activated by either electrical or chemical stimulation (von Campe et al., 1997). Basal dendrites do not exist normally in the rat, and are greatly increased in

animal model of MTS (Ribak et al., 2000; Dashtipour et al., 2002). However, the data shown in Fig. 2C suggest that activation of recurrent mossy fibers is more effective in synchronizing granule cell activity than activating granule cell basal dendrites. An additional clue to the role of mossy fiber sprouting in generating hyperexcitability is provided by a distinct group of patients that have cell loss predominantly in CA1 with less damage in the dentate gyrus, hilus and CA3, and there is little to no mossy fiber sprouting (as determined by dynorphin immunohistochemistry; (de Lanerolle et al., 1997, 2003). In this group, the ability to generate bursts with perforant path stimulation was comparable to that seen in typical MTS patients; however, there was significantly less asynchronous synaptic activity and fewer spontaneous baseline EPSPs (de Lanerolle et al., 1997). Therefore, only specific aspects of the observed hyperexcitability can be attributed to robust sprouting. We cannot rule out a possible contribution of minimal sprouting that was not detected by the immunostaining, however. In addition to sprouting, a common feature of the MTS dentate is granule cell dispersion (Houser, 1990). The mechanisms for this dispersion have included deficiencies in Reelin (Haas et al., 2002; Heinrich et al., 2006) leading to a broad cell body layer. However, it appears that the wider granule cell layer is linked with MTS, because the width of the granule cell layers are correlated with the degree of sprouting. However, total cell loss is the greatest predictor of the density of sprouting (El Bahh et al., 1999). The enhanced width of the granule cell layer in MTS hippocampi may reflect an increase in dentate granule cell neurogenesis, which occurs after seizures in laboratory animals, especially animal models of TLE (Scharfman, 2004). Neural progenitors have been described in the human dentate gyrus (Roy et al., 2000) and thus it is likely that different populations of granule cells are found in the human hippocampus that are of different ages. How the physiological differences in cells, related to their age and degree of incorporation into the hippocampal circuitry, has not been addressed in human material.

190

GABAergic inhibition The possible role of altered inhibition in epileptic tissue has been extensively discussed in the literature (Bernard et al., 1998, 2000). A significant problem in studying GABAergic inhibition in the dentate gyrus is the enormous complexity of the interneuronal systems found within the hilus and in the dentate gyrus, with different populations of interneurons subserving different functions and contacting distinct regions of the granule cells (Freund and Buzsaki, 1996). There are numerous changes in interneuronal networks in the MTS hippocampus compared to either comparison tissue or autopsy controls (de Lanerolle et al., 1989, 1992; Magloczky et al., 2000; Wittner et al., 2002). The primary changes include a loss of hilar interneurons that co-localize specific peptides and GABA. Those cells immunoreactive for neuropeptide Y (NPY) and somatostatin (SST) appear to be especially vulnerable (de Lanerolle et al., 1992; Mathern et al., 1995), while cells immunoreactive for parvalbumin and calbindin are preserved (Magloczky et al., 2000). However, despite these findings, there is no significant loss of immunoreactivity for glutamic acid decarboxylase, the GABA synthetic enzyme (Babb et al., 1989; Mathern et al., 1995). Adding to the complexity of describing the changes seen in inhibitory systems are the alterations in fiber pathways. This includes sprouting of NPY, SS, and SP systems (de Lanerolle et al., 1992; Mathern et al., 1995). Many of these changes were recently reviewed by Magloczky and Freund (2005). Physiologically, several groups have reported a decrease in evoked GABAA receptor-mediated inhibition in MTS hippocampi relative to comparison material (Isokawa, 1998; Williamson et al., 1999; Fig. 3). The degree of reduction is similar to that seen in the pilocarpine rat model (Kobayashi and Buckmaster, 2003). As with the membrane properties, no significant differences in polysynaptically IPSP/C conductances have been reported in parallel studies between control rats and the comparison material in humans (Isokawa, 1996; Williamson et al., 1999). It was interesting that both polysynaptically and monosynaptically evoked GABAA receptor-mediated events were

reduced to the same degree, suggesting that a decreased afferent input onto the remaining interneurons does not explain the decreased inhibitory efficacy. The decreased evoked responses may be to due the loss the SS and NPY immunoreactive interneurons that target the dendritic arbor rather than the cell body, because there was no change in the density of parvalbumin-immunoreactive terminals at the cell body of granule cells in tissue removed from patients with MTS compared to controls (Seress et al., 1993). Finally, the decrease in evoked inhibition is not due to changes in GABA receptor density or responsiveness. The studies by Shumate et al. (1998) and Gibbs et al. (1996) clearly show that acutely dissociated cells have an enhanced response to applied GABA and that the pharmacology of the GABA responses demonstrate an increased sensitivity to Zn2+, suggesting changes in the g subunit composition of these receptors. These pharmacological experiments are consistent with the mRNA studies from single granule cells showing a significant increase in the GABA a5 receptor subunit (Brooks-Kayal et al., 1999). Later studies revealed that there is significant diversity in the patterns of GABAA receptors expressed by different granule cells in pediatric patients with intractable TLE (Porter et al., 2005). Thus, the specific characteristics of GABA receptor-mediated responses may be quite heterogeneous between patients. Taken together, these data indicate that, like the changes in excitatory systems, there is a complex interplay between changes in pre- and postsynaptic elements. The decrease in evoked (polysynaptic and monosynaptic) IPSPs could reflect presynaptic modulation of GABA release, or changes in the amount of GABA available for release despite increases in postsynaptic receptor density.

Synaptic plasticity Several forms of synaptic plasticity are expressed in the dentate gyrus and have been intensively studied in rodents. These include long-term potentiation (LTP), induced by brief periods of high frequency stimulation, and long-term depression (LTD), induced by low frequency stimuli. While

191

Fig. 3. Decreased poly- and monosynaptic GABAA receptor-mediated inhibition in MTS hippocampi. We noted significant decreases in molecular layer-evoked inhibitory responses. A shows events delivered in normal medium; the stimulus intensity was set at twice that needed to evoke a single action potential. The d in A and B indicate the points at which the fast and slow IPSPs were measured. B shows monosynaptic responses evoked in presence of APV and CNQX (30 mM each). The stimulus intensity was set to the greatest that did not evoke an antidromic action potential. C and D shows grouped IPSP conductance data for polysynaptically and monosynaptically-evoked IPSPs, respectively. Note the profound and significant decrease in the IPSP conductance in the MTS tissue relative to comparison material. po0.05; po0.001.

the link between learning and memory and these forms of synaptic plasticity remains controversial, the availability of human material provides an additional avenue to examine this issue, because memory deficits are a common clinical feature of MTS (Helmstaedter et al., 2003). Both LTP and LTD can be reliably obtained in slices from epileptic human temporal neocortex with pharmacology comparable to that seen in rodents (Chen et al., 1996). However, both forms of synaptic plasticity are impaired in patients with MTS relative to patients without this pathology (Beck et al., 2000). Moreover, as shown in Fig. 4, there is a direct correlation between the degree of LTP impairment and a measure of neuropsychological function, the selective reminding task

(SRT) (Bell et al., 2005). For these studies, 200 Hz stimuli were delivered in a theta burst pattern and the change in the response was measured 30 min following the tetani. This value was then correlated with each patient’s SRT z score on the selective reminding task, a measure of lateralized hippocampal function. Taken together, these data support the hypothesis that altered synaptic plasticity is associated with the cognitive dysfunction in MTS patients.

Neurotransmitter transport Another aspect of neurotransmission that has been investigated in the human dentate is GABA

192

Fig. 4. Altered synaptic plasticity in the hippocampi of patients with MTS. A shows examples of field PSPs recorded in the molecular layer of a MTS patient. Note that there was pronounced depression of the evoked response following five trains of 100 Hz stimuli (100 stimuli/train delivered at 30 s intervals). B1 shows the group data for seven patients and B2 shows the cumulative probability plot demonstrating that depression was the predominant response. C shows that those patients that did not express LTD with this stimulation protocol had normal SRT z scores (control is set to 0) while those that did show LTD had markedly lower SRT scores.

uptake, commonly assessed using focal GABA application (Williamson et al., 1995b; Patrylo et al., 2001). These studies showed prolonged responses to GABA in the MTS dentate, which was not evident in the CA2 region (Telfeian et al., 1999). Moreover, there was a blunted effect of NO-711, a selective GAT-1 inhibitor, in the MTS tissue compared to comparison material. Finally, the responses to NPA, a GAT inhibitor that induces heterotransport (Honmou et al., 1995), were also reduced in MTS hippocampi. These data were interpreted as impaired GABA uptake. Parallel findings were made in the kainic acid rat model of TLE (Patrylo et al., 2001); however, see Frahm et al. (2003). GAT-1 is found both on neurons and on astrocytes, so these data do not address the role of gliosis in the regulation of GABA uptake. These physiological data are somewhat consistent with anatomical data showing a loss of GAT-1 around granule cell somata (where the GABA was applied), but an increase in the dendrites (Mathern et al., 1999; Lee et al., 2006).

There are a number of possible consequences of altered GABA transport, which could affect the overall excitability of the dentate. For example, granule cells with the GABAA d receptor subunit could be involved in tonic inhibition (Peng et al., 2004). Moreover, a slowed GABA clearance will also allow for binding to GABAB receptors, many of which are presynaptic, and can limit neurotransmitter release (Sperk et al., 2004). Conversely, there may be a reduced efficacy of GABA in this tissue due to GABA receptor desensitization. Finally, under conditions of high interneuronal activity, a reduction in neuronal uptake could limit the availability of GABA for repackaging and release. Which of these possibilities dominates under normal conditions has yet to be resolved. Glutamate transport is somewhat more complex. The bulk of glutamate uptake is mediated by glial transporters (Danbolt, 2001) and there is no real consensus about changes in MTS vs. nonMTS hippocampi (see Binder and Steinhauser, 2006 for review). For example, there is little evidence for changes in glial transporter density

193

when studied with Western blot (e.g., Eid et al., 2004), however, immunohistochemical studies show decreases in EAAT-2 (Mathern et al., 1999; Proper et al., 2002). However, there are no published studies measuring glial glutamate transport in human astrocytes that can shed light on these conflicting anatomical findings. Neuromodulation There is now anatomical and physiological evidence for multiple neuromodulatory systems in the human dentate gyrus. Three systems will be described here, with the understanding that there are numerous other modulators that may be important in regulating the excitatory tone of the human dentate. Metabotropic glutamate receptors (mGluRs) Several lines of evidence suggest that excitability in epileptic human tissue is constrained by metabotropic glutamate receptors. First, extracellular glutamate levels are very high in MTS hippocampi (Cavus et al., 2005) so that there is ample transmitter available to bind mGluRs. Second, Dietrich et al. (1999b) demonstrated that there is a reduced efficacy of L-AP4 (a type 2 mGluR agonist) in MTS tissue vs. comparison material. Similar findings were reported by Patrylo et al. (1999). One possibility is that there is little additional effect of mGluR agonists because the receptors are saturated. As shown in Fig. 5, there is relatively little effect of L-AP4 compared to other presynaptic modulators, such as NPY. The anatomic distribution of mGluR5 has been studied in the human dentate and a significant upregulation of this postsynaptic receptor has been observed by several investigators (Tang et al., 2001; Notenboom et al., 2006). However, the specific physiological effects of mGluR5 receptors in human tissue have not been investigated. Neuropeptides As in rodents, there is tremendous diversity within the interneuronal population in terms of

neuropeptide expression and numerous changes in peptide distributions have been observed as described above. Fewer physiological studies have been performed relative to the anatomical ones. The best studied peptide is NPY which, as in rodent studies, is a potent modulator of excitatory activity (rodent, Patrylo et al., 1999). While there is a loss of NPY-immunopositive cells in MTS hippocampi, there is evidence for axonal sprouting of NPY-immunoreactive (IR) fibers that cover the entire molecular layer, instead of being limited to the inner third of the molecular layer (de Lanerolle et al., 1989). When NPY was exogenously applied to human granule cells, there were no changes in the membrane properties, but a significant decrease in the evoked response. It was notable that NPY had a greater effect in cells with greater excitability: there was a 60% decrease in the evoked response when there were three or more evoked action potentials, but only a 34% decrease in cells where only one action potential could be evoked (Patrylo et al., 1999). This is consistent with the observations that NPY receptors are expressed on mossy fibers, but not on the perforant path inputs (Klapstein and Colmers, 1993), suggesting that there is a greater effect in those hippocampi with robust synaptic reorganization. Another peptide that is reorganized in MTS hippocampi is somatostatin. The SS-IR hilar interneurons appear to be highly susceptible to seizures in both human MTS and animal models (Robbins et al., 1991; Buckmaster and Schwartzkroin, 1995; Vezzani and Hoyer, 1999). As with NPY, the primary effects of SS are on excitatory, rather than inhibitory neurotransmission (Tallent and Siggins, 1997). SS receptors are expressed at specific synapses such as associational-comissural fibers, but not on perforant path fibers. As shown in Fig. 5A, there was only a modest effect of SS on orthodromic and evoked hyperexcitability. The ability of SS to reduce hyperexcitability was much less than that of NPY (Fig. 5B). However, this may reflect the downregulation of postsynaptic SS receptors in the outer dendritic regions MTS hippocampi (Csaba et al., 2005). These are two examples of the physiological effects numerous peptides that are known to exist

194

Fig. 5. Modulation of granule cell synaptic response by the neuropeptides NPY and somatostatin and glutamate receptor antagonists. A shows that NPY (1 mM) dramatically reduced the orthodromic response in an MTS case. In contrast, somatostatin (SS) had a less robust effect on the evoked response. Grouped data showing the percent change for NPY and SS are shown in C1. The effect of NPY was significantly greater than that of SS (po0.05). C2 shows the percent change in area mediated by glutamate receptor antagonists APV (30 mM) and the mGluR group 2 antagonist L-AP4 (100 mM) in MTS tissue.

in hilar interneurons and which exhibit some reorganization. Much more work will need to be done to determine how these changes in both fiber and receptor distributions will affect the overall synaptic output. It is also important to note that there are significant differences between humans and rodents in many of these systems, and thus data from rodent models may not accurately predict the situation in the human (Magloczky et al., 2000).

Metabolism A consistent feature of epileptic tissue is hypometabolism. This has been assessed in vivo using a variety of methods including glucose or O2 use with PET and SPECT, as well as phosphocreatine

and NAA (a neuronal mitochondrial marker) with MRS (Casse et al., 2002; Van Paesschen, 2004a, b; Pan et al., 2005). However, none of these in vivo approaches allows for a regional assessment of metabolism. The cellular basis for this hypometabolism remains somewhat unclear; however, several recent studies have begun to examine the issue. In the elegant studies of Kann et al. (2005), electrical stimulation was used to metabolically activate the tissue and the changes in the levels of NAD(P)H autofluorescence, as well as in mitochondrial potential, were studied. These experiments demonstrated that there was a pronounced drop in NADH fluorescence and a smaller overshoot during stimulation in the dentate as well as in other hippocampal areas. Despite this effect, fluorescence imaging of mitochondrial potentials

195

indicated that individual mitochondria exhibited normal voltage changes to these stimuli. Overall, these results were interpreted to indicate reduced mitochondrial function, because small overshoots are associated with decreased efficacy of mitochondrial NAD(P)H+ reduction. An alternate explanation is that these changes reflect the altered neuronal-glial metabolic coupling know to exist in human TLE (Petroff et al., 2002). However, it is important to note that the dentate is less metabolically impaired in MTS than other hippocampal regions, due in part to the differences in neuronal loss between the CA fields and the dentate (Kunz et al., 2000). Despite these imaging studies, there have been few studies on the interaction between metabolism and neuronal function in either animals or humans. Our initial studies in this field demonstrated that there were strong correlations between the levels of phosphocreatine and the extent of neuronal bursting as well as with the IPSP conductance. For both variables, lower PCr was associated with greater excitability (Williamson et al., 2005). Finally, together with the mitochondrial imaging data, these results suggest that the intimate relationship between excitability and energetics may be altered in specific ways in MTS hippocampi.

Conclusions While physiological/functional studies of human tissue are restricted to surgical specimens, the varieties in pathology allow for some conclusions to be made about the changes that are specific to MTS. These include robust synaptic reorganization that promotes hyperexcitability, and a decrease in the efficacy of evoked GABA inhibition. The alterations in GABAergic and neuromodulatory systems are quite complex and the specific consequences of these changes will be difficult to separate. However, because there are distinct differences between these systems in humans and in animal models (Magloczky and Freund, 2005), these studies will require human material. The availability of viable human specimens allows us the unique opportunity to study the physiological changes associated with TLE and provide insight

into other pathologies with hippocampal involvement, such as Alzheimer’s disease.

References Babb, T.L., Pretorius, J.K., Kupfer, W.R. and Crandall, P.H. (1989) Glutamate decarboxylase-immunoreactive neurons are preserved in human epileptic hippocampus. J. Neurosci., 9: 2562–2574. Beaumont, A. and Whittle, I.R. (2000) The pathogenesis of tumour associated epilepsy. Acta Neurochir. (Wien.), 142: 1–15. Beck, H., Blumcke, I., Kral, T., Clusmann, H., Schramm, J., Wiestler, O.D., Heinemann, U. and Elger, C.E. (1996) Properties of a delayed rectifier potassium current in dentate granule cells isolated from the hippocampus of patients with chronic temporal lobe epilepsy. Epilepsia, 37: 892–901. Beck, H., Steffens, R., Elger, C.E. and Heinemann, U. (1998) Voltage-dependent Ca2+ currents in epilepsy. Epilepsy Res, 32: 321–332. Beck, H., Steffens, R., Heinemann, U. and Elger, C.E. (1999) Ca(2+)-dependent inactivation of high-threshold Ca(2+) currents in hippocampal granule cells of patients with chronic temporal lobe epilepsy. J Neurophysiol, 82: 946–954. Bell, B.D., Fine, J., Dow, C., Seidenberg, M. and Hermann, B.P. (2005) Temporal lobe epilepsy and the selective reminding test: the conventional 30-minute delay suffices. Psychol. Assess., 17: 103–109. Bender, R.A., Soleymani, S.V., Brewster, A.L., Nguyen, S.T., Beck, H., Mathern, G.W. and Baram, T.Z. (2003) Enhanced expression of a specific hyperpolarization-activated cyclic nucleotide-gated cation channel (HCN) in surviving dentate gyrus granule cells of human and experimental epileptic hippocampus. J. Neurosci., 23: 6826–6836. Bernard, C., Cossart, R., Hirsch, J.C., Esclapez, M. and Ben-Ari, Y. (2000) What is GABAergic inhibition? How is it modified in epilepsy? Epilepsia, 41(Suppl 6): S90–S95. Bernard, C., Esclapez, M., Hirsch, J.C. and Ben-Ari, Y. (1998) Interneurones are not so dormant in temporal lobe epilepsy: a critical reappraisal of the dormant basket cell hypothesis. Epilepsy Res., 32: 93–103. Binder, D.K. and Steinhauser, C. (2006) Functional changes in astroglial cells in epilepsy. Glia, 54: 358–368. Brooks-Kayal, A.R., Shumate, M.D., Jin, H., Lin, D.D., Rikhter, T.Y., Holloway, K.L. and Coulter, D.A. (1999) Human neuronal gamma-aminobutyric acid(A) receptors: coordinated subunit mRNA expression and functional correlates in individual dentate granule cells. J. Neurosci., 19: 8312–8318. Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons and inhibition in the dentate gyrus of the rat in vivo. J. Neurosci., 15: 774–789. Casse, R., Rowe, C.C., Newton, M., Berlangieri, S.U. and Scott, A.M. (2002) Positron emission tomography and epilepsy. Mol. Imag. Biol., 4: 338–351.

196 Cavus, I., Kasoff, W.S., Cassaday, M.P., Jacob, R., Gueorguieva, R., Sherwin, R.S., Krystal, J.H., Spencer, D.D. and Abi-Saab, W.M. (2005) Extracellular metabolites in the cortex and hippocampus of epileptic patients. Ann. Neurol., 57: 226–235. Chen, W.R., Lee, S., Kato, K., Spencer, D.D., Shepherd, G.M. and Williamson, A. (1996) Long-term modifications of synaptic efficacy in the human inferior and middle temporal cortex. Proc. Natl. Acad. Sci. U.S.A., 93: 8011–8015. Cohen-Gadol, A.A., Bradley, C.C., Williamson, A., Kim, J.H., Westerveld, M., Duckrow, R.B. and Spencer, D.D. (2005) Normal magnetic resonance imaging and medial temporal lobe epilepsy: the clinical syndrome of paradoxical temporal lobe epilepsy. J. Neurosurg., 102: 902–909. Csaba, Z., Pirker, S., Lelouvier, B., Simon, A., Videau, C., Epelbaum, J., Czech, T., Baumgartner, C., Sperk, G. and Dournaud, P. (2005) Somatostatin receptor type 2 undergoes plastic changes in the human epileptic dentate gyrus. J. Neuropathol. Exp. Neurol., 64: 956–969. Danbolt, N.C. (2001) Glutamate uptake. Prog. Neurobiol., 65: 1–105. Dashtipour, K., Yan, X.X., Dinh, T.T., Okazaki, M.M., Nadler, J.V. and Ribak, C.E. (2002) Quantitative and morphological analysis of dentate granule cells with recurrent basal dendrites from normal and epileptic rats. Hippocampus, 12: 235–244. Dietrich, D., Clusmann, H., Kral, T., Steinhauser, C., Blumcke, I., Heinemann, U. and Schramm, J. (1999a) Two electrophysiologically distinct types of granule cells in epileptic human hippocampus. Neuroscience, 90: 1197–1206. Dietrich, D., Kral, T., Clusmann, H., Friedl, M. and Schramm, J. (1999b) Reduced function of L-AP4-sensitive metabotropic glutamate receptors in human epileptic sclerotic hippocampus. Eur. J. Neurosci., 11: 1109–1113. Eid, T., Thomas, M.J., Spencer, D.D., Runden-Pran, E., Lai, J.C., Malthankar, G.V., Kim, J.H., Danbolt, N.C., Ottersen, O.P. and de Lanerolle, N.C. (2004) Loss of glutamine synthetase in the human epileptogenic hippocampus: possible mechanism for raised extracellular glutamate in mesial temporal lobe epilepsy. Lancet, 363: 28–37. El Bahh, B., Lespinet, V., Lurton, D., Coussemacq, M., Le Gal La Salle, G. and Rougier, A. (1999) Correlations between granule cell dispersion, mossy fiber sprouting, and hippocampal cell loss in temporal lobe epilepsy. Epilepsia, 40: 1393–1401. Frahm, C., Stief, F., Zuschratter, W. and Draguhn, A. (2003) Unaltered control of extracellular GABA-concentration through GAT-1 in the hippocampus of rats after pilocarpine-induced status epilepticus. Epilepsy Res., 52: 243–252. Franck, J.E., Pokorny, J., Kunkel, D.D. and Schwartzkroin, P.A. (1995) Physiologic and morphologic characteristics of granule cell circuitry in human epileptic hippocampus. Epilepsia, 36: 543–558. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470.

Frotscher, M., Jonas, P. and Sloviter, R.S. (2006) Synapses formed by normal and abnormal hippocampal mossy fibers. Cell Tissue Res, 326: 361–367. Gibbs III, J.W., Zhang, Y.F., Kao, C.Q., Holloway, K.L., Oh, K.S. and Coulter, D.A. (1996) Characterization of GABAA receptor function in human temporal cortical neurons. J. Neurophysiol., 75: 1458–1471. Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G., Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for reelin in the development of granule cell dispersion in temporal lobe epilepsy. J. Neurosci., 22: 5797–5802. Harvey, B.D. and Sloviter, R.S. (2005) Hippocampal granule cell activity and c-Fos expression during spontaneous seizures in awake, chronically epileptic, pilocarpine-treated rats: implications for hippocampal epileptogenesis. J. Comp. Neurol., 488: 442–463. Heinrich, C., Nitta, N., Flubacher, A., Muller, M., Fahrner, A., Kirsch, M., Freiman, T., Suzuki, F., Depaulis, A., Frotscher, M. and Haas, C.A. (2006) Reelin deficiency and displacement of mature neurons, but not neurogenesis, underlie the formation of granule cell dispersion in the epileptic hippocampus. J. Neurosci., 26: 4701–4713. Helmstaedter, C., Kurthen, M., Lux, S., Reuber, M. and Elger, C.E. (2003) Chronic epilepsy and cognition: a longitudinal study in temporal lobe epilepsy. Ann. Neurol., 54: 425–432. Honmou, O., Kocsis, J.D. and Richerson, G.B. (1995) Gabapentin potentiates the conductance increase induced by nipecotic acid in CA1 pyramidal neurons in vitro. Epilepsy Res., 20: 193–202. Houser, C.R. (1990) Granule cell dispersion in the dentate gyrus of humans with temporal lobe epilepsy. Brain Res., 535: 195–204. Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O., Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered patterns of dynorphin immunoreactivity suggest mossy fiber reorganization in human hippocampal epilepsy. J. Neurosci., 10: 267–282. Isokawa, M. (1996) Decrement of GABAA receptor-mediated inhibitory postsynaptic currents in dentate granule cells in epileptic hippocampus. J. Neurophysiol., 75: 1901–1908. Isokawa, M. (1998) Modulation of GABAA receptor-mediated inhibition by postsynaptic calcium in epileptic hippocampal neurons. Brain Res., 810: 241–250. Isokawa, M., Avanzini, G., Finch, D.M., Babb, T.L. and Levesque, M.F. (1991) Physiologic properties of human dentate granule cells in slices prepared from epileptic patients. Epilepsy Res., 9: 242–250. Isokawa, M. and Levesque, M.F. (1991) Increased NMDA responses and dendritic degeneration in human epileptic hippocampal neurons in slices. Neurosci. Lett., 132: 212–216. Kann, O., Kovacs, R., Njunting, M., Behrens, C.J., Otahal, J., Lehmann, T.N., Gabriel, S. and Heinemann, U. (2005) Metabolic dysfunction during neuronal activation in the ex vivo hippocampus from chronic epileptic rats and humans. Brain, 128: 2396–2407.

197 Kim, J.H., Guimaraes, P.O., Shen, M.Y., Masukawa, L.M. and Spencer, D.D. (1990) Hippocampal neuronal density in temporal lobe epilepsy with and without gliomas. Acta Neuropathol. (Berl.), 80: 41–45. Klapstein, G.J. and Colmers, W.F. (1993) On the sites of presynaptic inhibition by neuropeptide Y in rat hippocampus in vitro. Hippocampus, 3: 103–111. Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition of dentate granule cells in a model of temporal lobe epilepsy. J. Neurosci., 23: 2440–2452. Kunz, W.S., Kudin, A.P., Vielhaber, S., Blumcke, I., Zuschratter, W., Schramm, J., Beck, H. and Elger, C.E. (2000) Mitochondrial complex I deficiency in the epileptic focus of patients with temporal lobe epilepsy. Ann. Neurol., 48: 766–773. de Lanerolle, N.C., Brines, M., Williamson, A., Kim, J.H. and Spencer, D.D. (1992) Neurotransmitters and their receptors in human temporal lobe epilepsy. Epilepsy Res. Suppl., 7: 235–250. de Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer, D.D. (1989) Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy. Brain Res., 495: 387–395. de Lanerolle, N.C., Kim, J.H., Williamson, A., Spencer, S.S., Zaveri, H.P., Eid, T. and Spencer, D.D. (2003) A retrospective analysis of hippocampal pathology in human temporal lobe epilepsy: evidence for distinctive patient subcategories. Epilepsia, 44: 677–687. de Lanerolle, N.C., Williamson, A., Meredith, C., Kim, J.H., Tabuteau, H., Spencer, D.D. and Brines, M.L. (1997) Dynorphin and the kappa 1 ligand [3H]U69,593 binding in the human epileptogenic hippocampus. Epilepsy Res., 28: 189–205. Lee, T.S., Bjornsen, L.P., Paz, C., Kim, J.H., Spencer, S.S., Spencer, D.D., Eid, T. and de Lanerolle, N.C. (2006) GAT1 and GAT3 expression are differently localized in the human epileptogenic hippocampus. Acta Neuropathol. (Berl.), 111: 351–363. Lieberman, D.N. and Mody, I. (1999) Properties of single NMDA receptor channels in human dentate gyrus granule cells. J. Physiol., 518(Pt 1): 55–70. Lombardo, A.J., Kuzniecky, R., Powers, R.E. and Brown, G.B. (1996) Altered brain sodium channel transcript levels in human epilepsy. Brain Res. Mol. Brain Res., 35: 84–90. Magloczky, Z. and Freund, T.F. (2005) Impaired and repaired inhibitory circuits in the epileptic human hippocampus. Trends Neurosci., 28: 334–340. Magloczky, Z., Halasz, P., Vajda, J., Czirjak, S. and Freund, T.F. (1997) Loss of Calbindin-D28K immunoreactivity from dentate granule cells in human temporal lobe epilepsy. Neuroscience, 76: 377–385. Magloczky, Z., Wittner, L., Borhegyi, Z., Halasz, P., Vajda, J., Czirjak, S. and Freund, T.F. (2000) Changes in the distribution and connectivity of interneurons in the epileptic human dentate gyrus. Neuroscience, 96: 7–25. Masukawa, L.M., Uruno, K., Sperling, M., O’connor, M.J. and Burdette, L.J. (1992) The functional relationship between

antidromically evoked field responses of the dentate gyrus and mossy fiber reorganization in temporal lobe epileptic patients. Brain Res., 579: 119–127. Mathern, G.W., Babb, T.L., Pretorius, J.K. and Leite, J.P. (1995) Reactive synaptogenesis and neuron densities for neuropeptide Y, somatostatin, and glutamate decarboxylase immunoreactivity in the epileptogenic human fascia dentata. J. Neurosci., 15: 3990–4004. Mathern, G.W., Mendoza, D., Lozada, A., Pretorius, J.K., Dehnes, Y., Danbolt, N.C., Nelson, N., Leite, J.P., Chimelli, L., Born, D.E., Sakamoto, A.C., Assirati, J.A., Fried, I., Peacock, W.J., Ojemann, G.A. and Adelson, P.D. (1999) Hippocampal GABA and glutamate transporter immunoreactivity in patients with temporal lobe epilepsy. Neurology, 52: 453–472. Nagerl, U.V., Mody, I., Jeub, M., Lie, A.A., Elger, C.E. and Beck, H. (2000) Surviving granule cells of the sclerotic human hippocampus have reduced Ca(2+) influx because of a loss of calbindin-D(28k) in temporal lobe epilepsy. J Neurosci., 20: 1831–1836. Notenboom, R.G., Hampson, D.R., Jansen, G.H., Van Rijen, P.C., Van Veelen, C.W., Van Nieuwenhuizen, O. and De Graan, P.N. (2006) Up-regulation of hippocampal metabotropic glutamate receptor 5 in temporal lobe epilepsy patients. Brain, 129: 96–107. Pan, J.W., Kim, J.H., Cohen-Gadol, A., Pan, C., Spencer, D.D. and Hetherington, H.P. (2005) Regional energetic dysfunction in hippocampal epilepsy. Acta Neurol. Scand., 111: 218–224. Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking of new glutamatergic pathways in the dentate gyrus of hippocampal slices from kainate-induced epileptic rats. J. Neurophysiol., 79: 418–429. Patrylo, P.R., Spencer, D.D. and Williamson, A. (2001) GABA uptake and heterotransport are impaired in the dentate gyrus of epileptic rats and humans with temporal lobe sclerosis. J. Neurophysiol., 85: 1533–1542. Patrylo, P.R., Van Den Pol, A.N., Spencer, D.D. and Williamson, A. (1999) NPY inhibits glutamatergic excitation in the epileptic human dentate gyrus. J. Neurophysiol., 82: 478–483. Peng, Z., Huang, C.S., Stell, B.M., Mody, I. and Houser, C.R. (2004) Altered expression of the delta subunit of the GABAA receptor in a mouse model of temporal lobe epilepsy. J. Neurosci., 24: 8629–8639. Petroff, O.A., Errante, L.D., Rothman, D.L., Kim, J.H. and Spencer, D.D. (2002) Glutamate-glutamine cycling in the epileptic human hippocampus. Epilepsia, 43: 703–710. Porter, B.E., Zhang, G., Celix, J., Hsu, F.C., Raol, Y.H., Telfeian, A., Gallagher, P.R., Coulter, D.A. and Brooks-Kayal, A.R. (2005) Heterogeneous GABAA receptor subunit expression in pediatric epilepsy patients. Neurobiol. Dis., 18: 484–491. Proper, E.A., Hoogland, G., Kappen, S.M., Jansen, G.H., Rensen, M.G., Schrama, L.H., Van Veelen, C.W., Van Rijen,

198 P.C., Van Nieuwenhuizen, O., Gispen, W.H. and De Graan, P.N. (2002) Distribution of glutamate transporters in the hippocampus of patients with pharmaco-resistant temporal lobe epilepsy. Brain, 125: 32–43. Reckziegel, G., Beck, H., Schramm, J., Elger, C.E. and Urban, B.W. (1998) Electrophysiological characterization of Na+ currents in acutely isolated human hippocampal dentate granule cells. J. Physiol., 509(Pt 1): 139–150. Reckziegel, G., Beck, H., Schramm, J., Urban, B.W. and Elger, C.E. (1999) Carbamazepine effects on Na+ currents in human dentate granule cells from epileptogenic tissue. Epilepsia, 40: 401–407. Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and Nadler, J.V. (2000) Status epilepticus-induced hilar basal dendrites on rodent granule cells contribute to recurrent excitatory circuitry. J. Comp. Neurol., 428: 240–253. Robbins, R.J., Brines, M.L., Kim, J.H., Adrian, T., de Lanerolle, N., Welsh, S. and Spencer, D.D. (1991) A selective loss of somatostatin in the hippocampus of patients with temporal lobe epilepsy. Ann. Neurol., 29: 325–332. Roy, N.S., Wang, S., Jiang, L., Kang, J., Benraiss, A., Harrison-Restelli, C., Fraser, R.A., Couldwell, W.T., Kawaguchi, A., Okano, H., Nedergaard, M. and Goldman, S.A. (2000) In vitro neurogenesis by progenitor cells isolated from the adult human hippocampus. Nat. Med., 6: 271–277. Scharfman, H.E. (2004) Functional implications of seizure-induced neurogenesis. Adv. Exp. Med. Biol., 548: 192–212. Schwartzkroin, P.A. and Prince, D.A. (1976) Microphysiology of human cerebral cortex studied in vitro. Brain Res., 115: 497–500. Seress, L., Gulyas, A.I., Ferrer, I., Tunon, T., Soriano, E. and Freund, T.F. (1993) Distribution, morphological features, and synaptic connections of parvalbumin- and calbindin D28k-immunoreactive neurons in the human hippocampal formation. J. Comp. Neurol., 337: 208–230. Shumate, M.D., Lin, D.D., Gibbs III, J.W., Holloway, K.L. and Coulter, D.A. (1998) GABA(A) receptor function in epileptic human dentate granule cells: comparison to epileptic and control rat. Epilepsy Res., 32: 114–128. Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M. (2006) Kainic acid-induced recurrent mossy fiber innervation of dentate gyrus inhibitory interneurons: possible anatomical substrate of granule cell hyper-inhibition in chronically epileptic rats. J. Comp. Neurol., 494: 944–960. Sperk, G., Furtinger, S., Schwarzer, C. and Pirker, S. (2004) GABA and its receptors in epilepsy. Adv. Exp. Med. Biol., 548: 92–103. St John, J.L., Rosene, D.L. and Luebke, J.I. (1997) Morphology and electrophysiology of dentate granule cells in the rhesus monkey: comparison with the rat. J. Comp. Neurol., 387: 136–147. Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L. (1989) Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol., 26: 321–330.

Tallent, M.K. and Siggins, G.R. (1997) Somatostatin depresses excitatory but not inhibitory neurotransmission in rat CA1 hippocampus. J. Neurophysiol., 78: 3008–3018. Tang, F.R., Lee, W.L. and Yeo, T.T. (2001) Expression of the group I metabotropic glutamate receptor in the hippocampus of patients with mesial temporal lobe epilepsy. J. Neurocytol., 30: 403–411. Telfeian, A.E., Spencer, D.D. and Williamson, A. (1999) Lack of correlation between neuronal hyperexcitability and electrocorticographic responsiveness in epileptogenic human neocortex. J. Neurosurg., 90: 939–945. Van Paesschen, W. (2004a) Ictal SPECT. Epilepsia, 45(Suppl 4): 35–40. Van Paesschen, W. (2004b) Qualitative and quantitative imaging of the hippocampus in mesial temporal lobe epilepsy with hippocampal sclerosis. Neuroimag. Clin. North Am., 14: 373–400. Vezzani, A. and Hoyer, D. (1999) Brain somatostatin: a candidate inhibitory role in seizures and epileptogenesis. Eur. J. Neurosci., 11: 3767–3776. Von Campe, G., Spencer, D.D. and de Lanerolle, N.C. (1997) Morphology of dentate granule cells in the human epileptogenic hippocampus. Hippocampus, 7: 472–488. Williamson, A., Patrylo, P.R., Pan, J., Spencer, D.D. and Hetherington, H. (2005) Correlations between granule cell physiology and bioenergetics in human temporal lobe epilepsy. Brain, 128: 1199–1208. Williamson, A., Patrylo, P.R. and Spencer, D.D. (1999) Decrease in inhibition in dentate granule cells from patients with medial temporal lobe epilepsy. Ann. Neurol., 45: 92–99. Williamson, A. and Spencer, D. (1995) Zinc reduces dentate granule cell hyperexcitability in epileptic humans. Neuroreport, 6: 1562–1564. Williamson, A., Spencer, D.D. and Shepherd, G.M. (1993) Comparison between the membrane and synaptic properties of human and rodent dentate granule cells. Brain Res., 622: 194–202. Williamson, A., Spencer, S.S. and Spencer, D.D. (1995a) Depth electrode studies and intracellular dentate granule cell recordings in temporal lobe epilepsy. Ann. Neurol., 38: 778–787. Williamson, A., Telfeian, A.E. and Spencer, D.D. (1995b) Prolonged GABA responses in dentate granule cells in slices isolated from patients with temporal lobe sclerosis. J. Neurophysiol., 74: 378–387. Wittner, L., Eross, L., Szabo, Z., Toth, S., Czirjak, S., Halasz, P., Freund, T.F. and Magloczky, Z.S. (2002) Synaptic reorganization of calbindin-positive neurons in the human hippocampal CA1 region in temporal lobe epilepsy. Neuroscience, 115: 961–978. Wuarin, J.P. and Dudek, F.E. (2001) Excitatory synaptic input to granule cells increases with time after kainate treatment. J. Neurophysiol., 85: 1067–1077.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 12

Hilar mossy cells: functional identification and activity in vivo Darrell A. Henze1 and Gyo¨rgy Buzsa´ki2, 1 Merck Research Laboratories, West Point, PA 19486, USA Center for Molecular and Behavioral Neuroscience, Rutgers, The State University of New Jersey, Newark, NJ, USA

2

Abstract: Network oscillations are proposed to provide the framework for the ongoing neural computations of the brain. Thus, an important aspect of understanding the functional roles of various cell classes in the brain is to understand the relationship of cellular activity to the ongoing oscillations. While many studies have characterized the firing properties of cells in the hippocampal network including granule cells, pyramidal cells and interneurons, information about the activity of dentate mossy cells in the intact brain is scant. Here we review the currently available information and describe biophysical properties and network-related firing patterns of mossy cells in vivo. These new observations will assist in the extracellular identification of this unique cell type and help elucidate their functional role in behaving animals. Keywords: mossy cell; hilus; network patterns; sharp wave; gamma; slow oscillation Mossy cells are perhaps the most underinvestigated neurons in the hippocampal formation. Part of the scarcity of information can be explained by the low number of MCs (approximately 10,000 in the rat) (Amaral et al., 1990). Unlike other principal (excitatory) cell types of the hippocampus, MCs do not form recognizable layers with densely packed somata. Instead, they are scattered in the hilar region under the granule cell layer, making their in vivo accessibility for physiological studies difficult. Most of what we know about the functions of MCs comes from the pioneering in vitro studies of Scharfman (Scharfman and Schwartzkroin, 1988; Scharfman et al., 1990, 2001; Scharfman, 1991, 1992a, b, 1994a–c, 1995), in vivo studies of Schwartzkroin et al. (Buckmaster et al., 1992, 1993, 1996; Buckmaster and Schwartzkroin 1995; Wenzel et al., 1997) and pathoanatomical studies of epilepsy and ischemia by Sloviter et al. (Sloviter, 1989, 1991a, 1994; Sloviter et al., 1991,

Introduction The granule cell is the most numerous cell type of the dentate gyrus (1 million in one hemisphere in the rat, (Amaral et al., 1990), serves to integrate entorhinal input and sends its messages to three major target cell populations. These three cell groups include mossy cells (MCs) of the hilus, CA3 pyramidal cells and interneurons of both dentate and CA3 regions. Although the MCs are the least numerous cell type in the hilar region, their unique properties suggest that they are likely to be a very important cell type of the dentate region. This chapter examines the properties of the MCs and their possible role in overall hippocampal function.

Corresponding author. Tel.: +1 (973) 353-1080 ext. 3131; Fax: +1 (973) 353-1820; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63012-X

199

200

2003; Goodman et al., 1993; Zappone and Sloviter, 2004). Behavioral correlates of MCs are unknown due to the lack of criteria for the reliable identification of MCs with extracellular methods. A major goal of this chapter is to provide an overview of the available knowledge about the firing patterns and biophysical properties of anatomically identified MCs and their network-related behavior, and discuss how this information can facilitate studies on MCs in the intact, behaving animal.

Mossy cell anatomy Mossy cells of the hilus were first recognized by Cajal (1911) and Lorente de No (1934) for their dendrites covered with large spines. These cells were later given the name ‘‘mossy cells’’ by Amaral (1978) due to the ‘‘mossy’’ appearance of their large spines (see Fig. 1A). It has also been demonstrated that all MCs selectively stain for GluR2/ 3 receptors as opposed to other cells of the hilus (Petralia and Wenthold, 1992; Leranth et al., 1996; Fujise and Kosaka, 1999). Interestingly, there appears to be differences in the peptide content of the mouse and hamster MC population, with ventral

MCs containing calretinin (Murakawa and Kosaka, 2001) which is largely absent in MCs of the dorsal hilus. This is in contrast to the rat where no MCs contain calretinin. In addition, calcitonin gene related peptide has been reported to selectively label MCs in the rat (Freund et al., 1997). Because these peptide markers are conspicuously absent in the pyramidal cells of the hippocampus proper, their differential staining can be taken as clear justification of separating MCs of the dentate gyrus and pyramidal neurons of the Ammon’s horn. The MCs are usually multipolar and have tapering dendrites that largely remain restricted to the hilus proper, although occasional dendrites are observed in the molecular layer of the dentate gyrus in both rats (Scharfman, 1991), and more often in primates (Frotscher et al., 1991; Buckmaster and Amaral, 2001). These dendrites in the molecular layer can receive direct input from the entorhinal cortex bypassing the granule cells. This variability in direct EC input is likely to be important for physiological function. All CA3 pyramidal cells, including those with mostly horizontal dendrites residing in zone 3 of Amaral (1978), send at least one dendritic branch to the stratum lacunosum-moleculare and therefore B

A

Ca3 PC Layer

C

20 mV 0.6 nA 50 ms

DG granule cell layer

Fig. 1. Hilar mossy cell and associated basic properties. (A) Biocytin labeled mossy cell. Note the large spines covering the soma and proximal dendrites (arrows). Example current step evoked traces from the mossy cell shown in A. (B) Responses to a series of hyperpolarizing steps from a resting potential of
57 mV. Note the robust ‘‘sag’’ observed for the larger hyperpolarizing steps and the long charging curve for the smallest hyperpolarizing step. For this cell, the in vivo input resistance was 52 MO. (C) The response of this cell to a strong depolarizing step (1.0 nA). Note that there is only very weak accommodation observed.

201

receive inputs from layer 2 neurons of the entorhinal cortex. In contrast, MCs without dendrites in the molecular layer generally receive only indirect information from the entorhinal input relayed by the granule cells (but see Kohler, 1985; Deller, 1998). The proximal dendrites and somata of MCs are covered by large ‘‘thorny excrescences’’ (Fig. 1). While the thorny excrescences at first seem similar to those of the proximal apical dendrite of CA3 pyramidal cells, they are qualitatively different in that they resemble ‘‘clusters of spheres’’ (Amaral, 1978) where as CA3 excrescences appear more irregular and thorny in their shape (Chicurel and Harris, 1992). As in CA3 pyramidal cells, the thorny excrescences of MCs receive synaptic input from the mossy fibers of the granule cells (Amaral, 1978; Murakawa and Kosaka, 2001). Recent findings suggest that the mossy fiber input to CA3 is a mixed glutamatergic/GABAergic input for the first three weeks of development in the rat or the young guinea pig (Walker et al., 2001). Under normal conditions in the adult, there is no detectable GABAergic transmission. However, mossy fiber GABAergic transmission is restored in the adult after periods of strong repetitive stimulation or hyperexcitability such as epilepsy (see Gutierrez, 2003, for review). There are also significant GADpositive, inhibitory inputs to the somatic region of MCs suggesting a strong perisomatic inhibitory input to these cells (Acsady et al., 2000; Murakawa and Kosaka, 2001). In fact, the perisomatic inhibition to MCs is very strong (15–40 times more synapses per soma) compared to the inhibitory cells of the hilus. The perisomatic inhibitory innervation of MCs is primarily from terminals that contain parvalbumin or cholecystokinin (CCK) (Acsady et al., 2000), similar to the pyramidal cells (Freund and Buzsa´ki, 1996). There is also a perisomatic and proximal dendritic input from cholinergic fibers (Deller et al., 1999) which may serve to provide excitatory tone to the MCs during oscillations such as theta. The more distal, or so-called peripheral dendrites, are innervated by a variety of inputs including more mossy fiber inputs onto regular small spines as well as other anatomically uncharacterized inputs making asymmetric and symmetric synapses (Frotscher et al., 1991).

It is possible that at least some of the input to the more peripheral dendrites is from CA3 pyramidal cells which extend axons back into the hilus and granule cell layer (Li et al., 1994). However, direct anatomical evidence for a CA3-mossy cell communication is still lacking (but see below for discussion of functional data). The dendrites of MCs are extensive, extending several hundred micrometers in both medio-lateral and dorso-caudal directions, suggesting a largely cylindrical dendritic tree arborizing largely in the subgranular zone. The large span of the dendritic arbor suggests that MCs are innervated by spatially distributed granule cells. The 100:1 ratio of granule cells to MCs predicts that a typical MC receives inputs from as many as a hundred granule cells. Given the relative sparse distribution of mossy fiber terminals in the hilus, it is also likely that each granule cell innervates only one or two MCs with large mossy fiber boutons. This low divergence and convergence should be contrasted to the widespread reciprocal innervation of granule cells by the MCs (see below). The MCs make glutamatergic (Soriano and Frotscher, 1994; Wenzel et al., 1997) asymmetric synapses with both excitatory and inhibitory postsynaptic targets (Frotscher et al., 1991; Buckmaster et al., 1996). In general, a single MC’s synaptic targets can be divided into three classes: local hilar targets; longitudinal targets (located >1 mm either septally or temporally); and contralateral targets. These three target classes have been observed in both rodents and primates. Excitatory innervation of CA3 pyramidal cells by the MCs has been repeatedly suggested (Buckmaster et al., 1996; Buckmaster and Amaral, 2001) but conclusive anatomical evidence for this alleged connection is missing. The main local hilar targets are mostly aspiny dendrites of interneurons and potentially dendrites of other MCs, although anatomical proof for mutual MC communication is lacking. CA3 recurrent collaterals also innervate aspiny interneurons in the hilus (Buckmaster et al., 1996) but conspicuously avoid spiny interneurons of which there are many in the hilus (Wittner et al., 2006). The major bulk of the axon cloud of MCs (>90% of ipsilateral synaptic contacts) target dentate

202

granule cell dendrites in the inner third of the molecular layer of the dentate gyrus rostral and caudal to the soma and dendrites of the parent MC (Buckmaster et al., 1992, 1996; Wenzel et al., 1997). This longitudinal arrangement is of major physiological consequence because it eliminates the possibility of a local granule cell–mossy cell–granule cell recurrent excitatory loop. Granule cells, which drive discharge of the MCs, do not receive excitatory information about the firing status of their targets. Instead, the output of the activated MC is distributed over the septaltemporal extent of the dentate. In these target areas, MCs innervate both granule cells and interneurons, giving rise to potentially interesting physiological scenarios. First, a granule cell discharging a target MC can silence competing granule cells over large territories via the feed-forward excitation of interneurons by the MC. However, if this were the sole function of such widespread axon arborization, there would be no need for MC excitatory innervation of granule cells. Through the latter numerically dominant excitatory pathway, the possibility exists that MCs can synchronize spatially distinct granule cells in the septotemporal axis, although the functional efficacy of the mossy cell–granule cell synapse is not well characterized (but see Scharfman, 1995). From the latter perspective, the main function of MCs would be to integrate functions of large numbers of active granules cells in the septotemporal axis of the hippocampus. The contralateral cellular targets of MCs are largely undefined. However, in addition to innervating granule cells, at least hilar neuropeptide Y (NPY) containing cells may receive input from contralateral MCs (Deller and Leranth, 1990). It will be interesting to learn the extent and identity of the contralateral targets and how they compare to the predominant granule cell targets ipsilaterally.

Functional cellular connectivity of MCs The anatomical connectivity described above provides a prospective framework for understanding the functional role of the MCs in normal

hippocampal function. This can be summarized most simply as primarily providing distributed excitatory feedback to dentate granule cells and secondarily providing excitatory drive to local inhibitory interneurons of the hilus. The functional importance of the contralateral projection is harder to predict since the anatomical targets are not yet characterized. A body of work by Scharfman and colleagues (Scharfman and Schwartzkroin, 1988; Scharfman et al., 1990, 2001; Scharfman, 1991, 1992a, b, 1994a–c, 1995) has provided functional data to support the predictions of the anatomical connectivity to and from MCs. For example, MCs that have dendrites that extend into the DG molecular layer have a lower threshold to fire in response to perforant path stimulation (Scharfman, 1991). A challenging series of studies in ventral slices used paired recordings from anatomically confirmed hilar, dentate and CA3 pyramidal cells to investigate the functional connectivity in this region. Paired recording of granule cells or CA3 pyramidal cells with MCs showed that single action potentials in either GCs or PCs can evoke EPSPs in MCs (Scharfman, 1994b). Another paired recording study demonstrated that MCs monosynaptically excite both granule cells and inhibitory interneurons. In addition, evidence was observed for polysynaptic inhibition of GCs in response to MC activity (Scharfman, 1995). The hypothesized physiological function of the low convergence and divergence of granule cells onto CA3 pyramidal cells is to disperse (‘‘orthogonalize’’) the entorhinal information onto the large recurrent system of CA3 neurons during the encoding of memories and provide multiple but sparse representation (Treves and Rolls, 1992, 1994). In the retrieval process, the auto-associative CA3 recurrent system, in turn, can recover the whole memory representation from partial or fragmental information (‘‘pattern completion’’) (McNaughton and Morris, 1987; Kanerva, 1988). Given this model of memory encoding and retrieval, we can ask what role might MCs play in this system? Although the anatomical connectivity between granule cells and MCs is similar to the granule cell–CA3 connections, given the small number of MCs and the lack of their reciprocal

203

excitation with one another makes it unlikely that MCs are part of the pattern completion mechanism. Instead, they may assist in increasing the sparseness of the memory representation in the recurrent CA3 system by the hypothesized feed forward suppression of granule cells in the septotemporal axis or by allowing a sparse but coordinated transfer of information from different segments of the entorhinal cortex onto the CA3 system. This may be an important combinatorial mechanism because the metric of spatial representation increases in the dorso-caudal axis of the entorhinal cortex with corresponding projections to different segments of the septotemporal axis of the hippocampus (Hafting et al., 2005). The synchronizing mechanism of granule cells by the MCs in the longitudinal axis would secure the simultaneous but distributed representation of the different sampling metric of the entorhinal cortex in the associative CA3 network. This hypothesis differs from an alternative proposed by Buckmaster and Schwartzkroin (1994) dubbed the ‘‘granule cell association’’ hypothesis. In that hypothesis, the MCs provide the necessary links to form associative connections within the dentate network analogous to the associational collaterals of pyramidal cells in the CA3 region.

Cellular properties of MCs The basic cellular properties of MCs are quite distinctive from other cell types in the hilar region. The MCs tend to have higher input resistance and strong inward rectification in response to hyperpolarization. Figure 1B shows a typical MC from a set of 10 cells we have recorded in vivo in response to a series of hyperpolarizing steps. Each anatomically verified MC recorded in urethane anesthetized rats had action potentials that crossed 0 mV, a mean resting membrane potential of 5871.8 mV, and a mean input resistance of 6178.6 MO (means7SEM). The resting potential is similar to that of CA3 (63 mV; n ¼ 84) and CA1 (65 mV; n ¼ 280) pyramidal cells but different from the more hyperpolarized granule cells (74 mV; n ¼ 41). Of the principal cell types, MCs had the largest input resistance

(CA3 ¼ 53.8 mO; n ¼ 8; CA1 ¼ 48.4 MO (Henze and Buzsaki, 2001)). The smallest hyperpolarizing step applied (0.2 nA) resulted in a hyperpolarization of
22 mV and the time constant is best fit with a double exponential with tau1 ¼ 7.5 ms and tau2 ¼ 250 ms. A strong inward rectification can be seen with a step injection of 0.4 nA that results ultimately in a maximum hyperpolarization of 35 mV. The hyperpolarizations of the membrane potential by small step currents were best fit by a double exponential in seven of the nine cells where it could be measured; the mean time constant values were 15.9 and 188.2 ms. The remaining two MCs were best fit with single exponentials with time-constants of 32 and 26 ms. Figure 1C shows that the MCs typically do not show burst firing in response to depolarizing steps (1.0 nA) and only show weak accommodation for the duration of the step depolarization. This behavior can be contrasted to the typical burst pattern and strong spike accommodation in response to current steps in CA3 pyramidal cells (personal observations, Bilkey and Schwartzkroin, 1990; Buckmaster et al., 1993; Scharfman, 1993b). The background synaptic activity in MCs has been reported to be quite high, both in vivo and in vitro (Strowbridge et al., 1992; Scharfman, 1993a).We also have observed high background activity that included some very large events (>10 mV; Fig. 2). It is likely that these giant PSPs with fast rise times arise from the mossy fiber synaptic inputs to the MCs reflecting either synchronous multivesicular release from a complex MF bouton or perhaps the release of large individual quanta as has been reported in CA3 pyramidal cells (e.g. see Henze et al., 2002). Although the magnitude of the giant PSPs is quite variable, it is unlikely that it reflects varying convergence of activity from multiple granule cells. First, the convergence of granule cells onto MCs is low (
100). Second, the density of mossy boutons in the hilus is quite low. Third, intracellularly labeled neighboring granule cells never showed spatially clustered boutons that would otherwise suggest common targets (Acsady et al., 2000). One feature of MC activity that we have observed that has not been previously described

204

A 10 mV 0.1 s B

2 mV 0.01 s

Fig. 2. (A) Intracellular recording of a hilar mossy cell resting at 55 mV (solid line). Notice the large spontaneous postsynaptic potentials that occur from the baseline that sometimes lead to action potentials. (B) Higher resolution depiction of two large EPSPs from the box in (A).

is their relatively high background firing rate (e.g. see Figs. 3, 5, and 9C). Our findings under urethane anesthesia are partly consistent with Scharfman (1993a), who showed that there is often a high rate of firing but it waxes and wanes in vitro. However, it is not consistent with other studies in vivo (Soltesz et al., 1993) under ketamine/ xylazine where rates of less than 1 Hz were reported. The source of this high firing rate is not well understood because granule cells under urethane anesthesia are typically hyperpolarized and fire at a low rate. One hypothesis is that spontaneous release from MF boutons can lead to MC action potentials (Henze et al., 2002). If the high firing pattern is confirmed by investigations in the freely behaving animal, it can be contrasted with the more clustered firing patterns of CA3 pyramidal cells. Figure 3 illustrates the autocorrelograms calculated from our MC recordings. These autocorrelograms should facilitate the comparison of the spike dynamics of MCs under anesthesia and in the drug-free animal in future studies. Another essential piece of information in the process of physiological identification of MCs is the extracellular features and shape of the extracellular action potential. To this end, we have recorded from a single MC using simultaneous intracellular and extracellular electrodes (Fig. 4). The waveform we have observed is somewhat

unusual in that the extracellular unit waveform has a rounded trough that has not been observed in other hippocampal unit waveforms. It is possible that the rounded (wide) pattern derives from the summed extracellular currents from the soma and the thick dendrites of MCs. In support of this interpretation, the extracellular spikes were recorded by three shanks of the silicon probe, a total distance of 300 mm. We would suggest that the horizontally wide current distribution of MC spikes may be used as a distinguishing feature in extracellular recordings from the granule cells and other nearby smaller size interneurons.

MC cellular activity in a network Although there is a relative lack of information about the physiological role of MCs under normal conditions, their role has been a frequently discussed topic of debate in the epilepsy literature. This is because MCs are often observed to be reduced in post-mortem tissue taken from people who have had temporal lobe epilepsy (TLE) (Sloviter et al., 1991). Sloviter and colleagues proposed the so-called ‘‘dormant basket cell’’ hypothesis of epilepsy (Sloviter, 1991b). The dormant basket cell hypothesis holds that the importance of the excitatory tone provided by the MCs is to provide excitation of hilar inhibitory interneurons which in turn then provide strong inhibition of granule cells. When MCs are lost due to neurodegeneration associated with TLE, the inhibitory basket cells lose their tonic excitatory drive resulting in a net disinhibition of dentate gyrus granule cells (Sloviter, 1994; Sloviter et al., 2003). A contrasting hypothesis has been called the ‘‘irritable mossy cell’’ hypothesis as proposed by Soltesz and colleagues. In this view, it is not the loss of MCs that leads to a net excitation of granule cells, instead it is the remaining MCs that have higher firing rates and provide uncontrolled excitatory feedback to the dentate gyrus granule cells thus exacerbating the epileptic process (Santhakumar et al., 2000; Ratzliff et al., 2002). Both of the ‘‘dormant basket cell’’ and ‘‘irritable mossy cell’’ hypotheses have their attractions. Recent studies by both camps have provided

205

F

A

-0.10

-0.05

0.00

0.05

0.10

-0.10

-0.05

0.00

0.05

0.10

-0.10

-0.05

0.00

0.05

0.10

-0.10

-0.05

0.00

0.05

0.10

-0.10

-0.05

0.00

0.05

0.10

-0.10

-0.05

0.00

0.05

0.10

G

B

-0.10

-0.05

0.00

0.05

0.10

H

C

-0.10

-0.05

0.00

0.05

0.10

I

D

-0.10

-0.05

0.00

0.05

0.10

J

E

-0.10

-0.05

0.00

0.05

0.10

Fig. 3. Hilar mossy cells show a variety of firing patterns as assessed by autocorrellograms (A–J). The autocorrelogram for each of the 10 MCs in our dataset was calculated from spontaneous spiking from the resting potential (no injected current). The majority of the cells show a pattern that is more reminiscent of that seen for repetitively firing interneurons than the more bursty firing of CA1 or CA3 pyramidal cells (e.g. Csicsvari et al., 1998).

206

Spontaneous

Evoked 20 mV 0.5 mS

IC EC

50 microV 0.5ms

40 microV 0.5ms

Fig. 4. Extracellular waveforms of mossy cells during spontaneously and evoked spiking. A recording was obtained where the intracellularly recorded mossy cell was also observed on the extracellular electrode. The extended shape of the mossy cell dendrites allowed the extracellular signal to be observed on three shanks of the extracellular silicon probe (150 mm between shanks). The extracellular electrode track is indicated by the white arrows. The soma of the mossy cell is indicated by the yellow arrow. There was a difference in the shape of the action potentials, both intracellular and extracellular, suggesting differences in the site of spike initiation for spontaneous and evoked spikes. (See Color Plate 12.4 in Color Plate Section.)

evidence to support the respective theories. Three days after kainic acid induced status epilepticus, there is reduction in inhibition and thus increase in excitability of the dentate gyrus that correlates with the degree of MC loss (Zappone and Sloviter, 2004). However, Ratzliff et al. (2004) showed that in the hippocampal slice, if they acutely removed MCs via manual destruction, there was a net reduction in the excitability of the dentate gyrus presumably due to the loss of direct excitatory input from MCs. In all likelihood, both hypotheses are at least partially correct and strict interpretation of these studies is confounded by the technical challenges of studying the MCs in vivo under nonpathological conditions. The intrinsic firing rates of MCs in the intact unanesthetized animal is not

known, and both low (Soltesz et al., 1993) and high firing rates (Figs. 3, 5, and 9C) have been observed under anesthesia. Bulk stimulation of fiber pathways such as the perforant path input to the dentate gyrus is inherently not physiological in that the precise synchrony of synaptic input that results very rarely, if ever, happens in natural processing. As such, the ratio of synchronously activated excitatory to inhibitory inputs may be very different than that observed during normal ongoing hippocampal function. Although the hilus also contains a large variety of interneurons (Amaral, 1978; Mizumori et al., 1990; Halasy and Somogyi, 1993; Buhl et al., 1994; Sik et al., 1997), the contribution of these interneurons to the rich variety of dentate area network

207

patterns is not known. Our overriding hypothesis is that the main goal of both MCs and hilar interneurons is to control network patterns during various behaviors. This rationale has been successfully applied to characterize CA1 interneurons (Csicsvari et al., 1999; Klausberger et al., 2003, 2004, 2005), where the initial network patternbased classification (theta phase and sharp waves) in freely behaving rats was followed by juxtacellular labeling of similarly characterized interneurons under anesthesia. We suggest that a similar approach in the dentate gyrus can be equally fruitful. Here we have preliminary yet more advanced knowledge of the network contribution of MCs under anesthesia than in the behaving animals. In the dentate area, several distinct oscillatory patterns are present. Network oscillations are known to involve a rhythmic pattern of periods of active inhibition that are counterbalanced by periods of permissive excitability (e.g. theta in CA1) (Buzsaki, 2002). Perhaps the unique role of the MCs is to provide active excitation to granule cells in the periods between the rhythmic inhibitory inputs driven by interneurons. Although all the various patterns are mediated by a limited set of excitatory pathways, it is expected that MCs and the various interneuron groups may differentially participate in these network patterns because the dynamics of activation can differentially affect neurons with different properties. The following patterns can be used to characterize the network contribution of MCs and contrast them to granule cells, CA3 pyramidal cells and hilar area interneuron types. 1. In the exploring rat and during REM sleep, large amplitude theta oscillations are present in the dentate gyrus. Theta waves in the dentate region are coherent with but phase-shifted by approximately 2701 relative to the theta oscillation in the CA1 pyramidal layer (Buzsaki et al., 1983). 2. Concurrent with theta, another prominent pattern in the dentate area is the gamma frequency oscillation (40–100 Hz). The power of gamma oscillations is strongly phase modulated by the slower theta rhythm and both theta and gamma oscillations in the dentate

gyrus depend mainly on inputs from the entorhinal cortex (Bragin et al., 1995a; Penttonen et al., 1998). In addition to the classical gamma frequency, slower (12–40 Hz; beta) oscillations are also observed, often with a larger amplitude in the hilus. It is not clear whether the slow oscillation is simply a slower version of gamma or comprises a physiologically distinct rhythm. 3. The largest amplitude dentate gyrus event is the ‘‘dentate spike’’. This is a short duration (o60 ms), large amplitude (>0.5–2.5 mV) field potential characterized by synchronous discharge of granule cells and interneurons and suppression of CA3 pyramidal cells (Bragin et al., 1995b; Penttonen et al., 1997). Two types of dentate spikes have been distinguished. The first type is a short burst of gamma oscillation consisting of 2–5 waves, one of which of excessively high amplitude with large sinks in the outer third of the dentate molecular layer. The second type, observed in a subset of animals, has a somewhat different voltage vs. depth profile with a large sink located in the middle third of the dentate molecular layer. Our unpublished observations suggest that type 2 dentate spikes occur when thalamocortical high voltage spindles (Buzsaki et al., 1988) invade the dentate area. 4. Neocortical slow oscillations (Steriade et al., 1990) also exert an impact on the firing patterns of the hippocampus, likely by way of the entorhinal cortex (Isomura et al., 2006; Wolansky et al., 2006). During the UP state of slow oscillations, gamma power in the dentate gyrus and spiking of neurons increases dramatically. In contrast, dentate gamma activity decreases during the DOWN state (corresponding to delta waves of deep sleep in the neocortex) but CA3 pyramidal cells may increase their firing rates and generate gamma oscillations (Isomura et al., 2006) 5. Sharp waves-ripple complexes (SPW) are truly self-organized endogenous hippocampal events that occur during slow-wave sleep, immobility and consummatory behaviors

208

(Buzsaki et al., 1983). They arise in the CA3 recurrent system and can spread to the CA1 region and the dentate region. The transient excitation of CA1 neurons gives rise to a short-lived fast oscillation (‘‘ripple’’) (O’Keefe and Nadel, 1978; Buzsa´ki et al., 1992; Ylinen et al., 1995). No ripples are associated with SPW in the dentate area but putative interneurons and, occasionally, granule cells are depolarized and discharge in synchrony with CA1 ripples, although the typical response is hyperpolarization (Penttonen et al., 1997). We submit that these five unique population patterns can be used in future studies to functionally classify dentate region neurons in the freely behaving animal. In turn, these same patterns can be used in anesthetized rats where they can be labeled by intracellular or juxtacellular methods. The network-based classification should be combined with the spike dynamics and wave-shape features of extracellular spikes. Fortunately, the repertoire of rat hippocampal network activity patterns under urethane anesthesia largely reflects what is observed in the drug-free rat. Using this approach we have been able to observe how MCs behave in relation to some of these naturally occurring network patterns. Theta It has been previously reported that MC membrane potential shows rhythmic oscillations in the theta band that are phase-locked to the extracellular theta oscillation in the contralateral hippocampus (Soltesz et al., 1993). However, as noted above, this study reported a very low (o1 Hz) basal firing rate for the MCs. Nevertheless, our observations support the involvement of MCs in theta oscillations. Individual MCs can either be depolarized (8 of 10) or hyperpolarized (2 of 10) by a transition from slow wave sleep to theta evoked by a tail pinch (Fig. 5). This behavior is similar to pyramidal cells (Kamondi et al., 1998). In addition, the membrane potential of MCs shows a co-variation with the extracellular field, with peak depolarization and discharge slightly after the

A 1 0.2 mV 10 mV 0.5 s

2

B 1 0.1 mV 10 mV 0.5 s

2

Fig. 5. (A) Example of mossy cell inhibition by transition to theta rhythm evoked by tail pinch in urethane anesthetized rat. Extracellular recording (A1) from area CA1 recorded simultaneously with a hilar mossy cell (A2). A tail pinch was applied at the arrow and the mossy cell responded by slowing its firing rate. (B) Example of mossy cell excitation by transition to theta rhythm evoked by tail pinch in urethane anesthetized rat. Extracellular recording (B1) from area CA3 recorded simultaneously with (B2) a hilar mossy cell that was excited by tail pinch (arrow) evoked theta.

peak of the locally derived theta oscillation (Fig. 6) and coherent with the discharge of some interneuron types. We predict that MCs in the behaving rat will keep a similar relationship to local hilar/CA3c theta oscillations. Beta/gamma oscillations The power of gamma frequency oscillation in the hilus is phase modulated by the slower theta (Bragin et al., 1995a). This gamma frequency

209

Fig. 6. (A) Continuous theta/delta power ratio from an extracellular electrode located in the hilus. The increases in theta power at
45 and 200 s was induced via tail pinch. (B) rastergrams of isolated units recorded from three extracellular tetrodes in the hilus/area CA3c. (C) Intracellular membrane potential recorded from a mossy cell during this same recording epoch while passing hyperpolarizing current to maintain the resting potential near 80 mV. Notice that during the periods of theta power increase, the mossy cell experiences a depolarization from the 80 mV holding potential. (D) Two examples of average mossy cell membrane potential time aligned to the two ‘‘theta on’’ putative interneuronal units indicated by the start symbols in (C). The upper trace is the average membrane potential of the MC. The lower trace is the average wide-band extracellular trace. The autocorrelogram and waveform for the isolated IN units are also shown as insets.

210

A

Gamma cycle Theta oscillation

20µV

EC unit for MC 200 ms

B

1mV 20 ms 25µV 1ms

Fig. 7. (A) Average of extracellular hilar field potential aligned to the peaks of the intracellular MC spontaneous action potentials recorded over a 200 sec recording period. This is the same case as in Fig. 4 where the extracellular electrode picked up the same MC as was recorded intracellularly. The extracellular unit correlate of the intracellular AP is indicated. Notice that the MC fires on the end of the gamma cycle waveform and there is an overall theta oscillation present that is time aligned to the intracellular spike. (B) Red traces: MC membrane potential averaged and time aligned on the spike times of isolated interneurons recorded in the hilus/CA3c. Left column: IN unit autocorrelogram, black traces: average IN unit waveforms. Note similar phase relationship of the interneuron-timed intracellular gamma frequency oscillation.

modulation is also evident in the relationship between the local field and fast spiking interneurons and their effect on the MC (Fig. 7). When the action potentials of MCs are used as reference for averaging local field potentials, phaselocking of MC spikes to the gamma oscillation, superimposed on the peak of theta waves becomes evident (Fig. 7A). Furthermore, when nearby fast-spiking putative interneurons are used as the reference,

membrane oscillation of MCs in the gamma frequency band becomes evident (Fig. 7B). Similar phase-locked behavior has been also observed in the beta oscillation frequency range as well. Slow oscillations Slow oscillations arise in neocortical networks (Steriade et al., 1993a–c) and spread to the

211

Fig. 8. (A) CA3 unit activity and MC membrane potential fluctuations. Mossy cells show synaptic activity correlated with unit activity in CA3 (2/2 cells recorded with extracellular electrode placed in CA3c. Upper traces show extracellular wideband recording middle trace bandpass filtered between 800 Hz and 3 kHz. Lower traces show intracellular membrane potential. (B) Cross-correlation between spontaneous MC spikes recorded intracellularly and all extracellular units combined. Lower panels are shuffled controls. (C) Hilar/ CA3c population bursts correlate with spontaneous mossy cell spikes. Hilar/CA3c bursts were defined as five unit spikes with less than 20 ms inter-spike interval. The spikes could originate from any unit. Shuffled correlograms are shown for controls. (D) Crosscorrelation between spontaneous mossy cell spikes and all hilar/CA3 units combined for a second extracellular/intracellular recording pair.

entorhinal cortex and subiculum from where they can invade the dentate gyrus as well (Isomura et al., 2006; Wolansky et al., 2006). At the single cell level, most cortical neurons show a bimodal distribution of the membrane potential and these UP and DOWN states alternate relatively rhythmically at 0.5–2 Hz. Slow oscillations are most prominent under anesthesia but are also present in deep stages of slow wave sleep (Achermann and Borbely, 1997), where the transient delta waves correspond to the DOWN (or silent) state. At the

transition of the entorhinal DOWN–UP state, the surge of excitation induces a strong discharge of granule cells and also gamma frequency oscillations (Isomura et al., 2006). Often one of these gamma waves becomes excessively large and has been referred to as the dentate spike (Bragin et al., 1995b). In this case, the MC firing is likely timed by feed forward excitation from dentate gyrus granule cells that are driven by the entorhinal input. The surge of activity in the input from entorhinal cortex is also reflected by the sudden

212

Fig. 9. CA1 sharpwave/ripples are associated with inhibition of mossy cells. (A) Example of a CA1 sharpwave/ripple complex recorded extracellularly in the pyramidal layer of CA1. (B) Intracellular spiking activity of mossy cell is suppressed during SPW. (C, D) Temporal expansion of region in box in A and B. Mossy cell membrane response during SPW/ripples has a reversal near 60 mV. (E, F) Average extracellular SPW and mossy cell membrane potential from a hyperpolarized baseline (E; n ¼ 12 SPWs) and a more depolarized baseline (F; n ¼ 11 SPWs). (G) Plot of the peak change in mossy cell membrane potential during CA1 SPWs from 15 recording epochs at different membrane potentials from four mossy cells (different symbols). The best fit linear regression line is shown (R ¼ 0.57621, Po0.025).

213

increase of population firing in the dentate. We exploited the high levels of CA3 activity that occur during the slow oscillatory pattern for the examination of the behavior of MCs. The transition from DOWN to UP state was defined when a group of dentate/hilar/CA3 neurons fired a burst of activity. This analysis revealed that MCs become periodically active and silent according to the level of activity in the entorhinal–hippocampal network (Fig. 8).

CA1 sharp waves Although SPW arises in the CA3–CA1 regions, the recurrent axon collaterals of CA3 can directly excite hilar interneurons and even granule cells (Li et al., 1994; Penttonen et al., 1997). Typically, feed-forward inhibition prevails by this mechanism as reflected by the SPW-locked hyperpolarization of granule cells. However, this occasional failure of inhibition and/or a concerted activation of a single granule cell by the recurrent CA3 axons can robustly discharge granule cells during SPWs (Penttonen et al., 1997). Our observations, to date, suggest that SPWs are associated with net inhibition of MCs. Figure 9A–D show the temporal relationship between a spontaneous SPW recorded in area CA1 (Fig. 9A, B) while recording from a MC in the hilus (Fig. 9C, D). As can be appreciated from this example, the ongoing spontaneous firing of the MC is inhibited in the period overlapping with and following the SPW/ripple complex. This inhibitory effect is probably mediated via activation of chloride flux through GABAA receptors since the change in MC membrane potential associated with a CA1 SPW has a reversal potential near 60 mV similar to GABAA reversal potentials in vivo (Fig. 9E–G). The potential source of this inhibition is the increased SPW-related firing of hilar interneurons driven either directly by the recurrent CA3 collaterals or the rarely discharging granule cells (Penttonen et al., 1997). It appears that, at least under anesthesia, the strong inhibition can prevent MCs from discharging in response to their granule cell inputs during SPWs. However, the situation might be quite different in the drug-free animal and it is

expected that future studies will clarify whether MCs can become active participants in SPW events. It is worth noting here that in slices treated with a GABAA receptor antagonist, or slices from an epileptic rat, CA3 and MCs are engaged in population bursts while granule cells remain hyperpolarized, suggesting that CA3 pyramidal cells may directly excite MCs (Scharfman, 1994a; Scharfman et al., 2001). The failure of MCs to discharge during SPWs in the intact brain would imply that the hypothesized SPW-mediated consolidation of synaptic circuits in the CA3–CA1 networks (Buzsa´ki, 1989) can proceed independent of the modification of the synapses established by the MCs. Conclusion Although MCs are well-known and critical components of the dentate circuitry, their physiological function and exact involvement in various hyperexcitable phenomena has remained elusive. A major technical problem is the lack of reliable physiological criteria that may be used in extracellular recordings in freely behaving animal for the positive identification of MCs. Our intracellular characterization of some of their biophysical features and network-related behavior are the first steps in this direction. Abbreviations CCK EPSP GABA GAD MC NPY

cholecystokinin excitatory postsynaptic potential gamma-aminobutyric acid glutamic acid decarboxylase mossy cell neuropeptide Y

References Achermann, P. and Borbely, A.A. (1997) Low-frequency (o1 Hz) oscillations in the human sleep electroencephalogram. Neuroscience, 81: 213–222. Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G. and Freund, T.F. (2000) Unusual target selectivity of

214 perisomatic inhibitory cells in the hilar region of the rat hippocampus. J. Neurosci., 20: 6907–6919. Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182: 851–914. Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons, numbers and the hippocampal network. Prog. Brain Res., 83(1–11): 1–11. Bilkey, D.K. and Schwartzkroin, P.A. (1990) Variation in electrophysiology and morphology of hippocampal CA3 pyramidal cells. Brain Res., 514: 77–83. Bragin, A., Jando, G., Nadasdy, Z., Hetke, J., Wise, K. and Buzsaki, G. (1995a) Gamma (40–100 Hz) oscillation in the hippocampus of the behaving rat. J. Neurosci., 15: 47–60. Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M. and Buzsaki, G. (1995b) Dentate EEG spikes and associated interneuronal population bursts in the hippocampal hilar region of the rat. J. Neurophysiol., 73: 1691–1705. Buckmaster, P.S. and Amaral, D.G. (2001) Intracellular recording and labeling of mossy cells and proximal CA3 pyramidal cells in macaque monkeys. J. Comp. Neurol., 430: 264–281. Buckmaster, P.S. and Schwartzkroin, P.A. (1994) Hippocampal mossy cell function: a speculative view. Hippocampus, 4: 393–402. Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons and inhibition in the dentate gyrus of the rat in vivo. J. Neurosci., 15: 774–789. Buckmaster, P.S., Strowbridge, B.W., Kunkel, D.D., Schmiege, D.L. and Schwartzkroin, P.A. (1992) Mossy cell axonal projections to the dentate gyrus molecular layer in the rat hippocampal slice. Hippocampus, 2: 349–362. Buckmaster, P.S., Strowbridge, B.W. and Schwartzkroin, P.A. (1993) A comparison of rat hippocampal mossy cells and CA3c pyramidal cells. J. Neurophysiol., 70: 1281–1299. Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwartzkroin, P.A. (1996) Axon arbors and synaptic connections of hippocampal mossy cells in the rat in vivo. J. Comp. Neurol., 366: 271–292. Buhl, E.H., Han, Z.S., Lorinczi, Z., Stezhka, V.V., Karnup, S.V. and Somogyi, P. (1994) Physiological properties of anatomically identified axo-axonic cells in the rat hippocampus. J. Neurophysiol., 71: 1289–1307. Buzsa´ki, G. (1989) Two-stage model of memory trace formation: a role for ‘‘noisy’’ brain states. Neuroscience, 31: 551–570. Buzsaki, G. (2002) Theta oscillations in the hippocampus. Neuron, 33: 325–340. Buzsaki, G., Bickford, R.G., Armstrong, D.M., Ponomareff, G., Chen, K.S., Ruiz, R., Thal, L.J. and Gage, F.H. (1988) Electric activity in the neocortex of freely moving young and aged rats. Neuroscience, 26: 735–744. Buzsa´ki, G., Horvath, Z., Urioste, R., Hetke, J. and Wise, K. (1992) High-frequency network oscillation in the hippocampus. Science, 256: 1025–1027.

Buzsaki, G., Leung, L.W. and Vanderwolf, C.H. (1983) Cellular bases of hippocampal EEG in the behaving rat. Brain Res., 287: 139–171. Cajal, S.R. (1911) Histologie du Systeme Nerveux de l’Homme et des Vertebres. Oxford University Press, New York. Chicurel, M.E. and Harris, K.M. (1992) Three-dimensional analysis of the structure and composition of CA3 branched dendritic spines and their synaptic relationships with mossy fiber boutons in the rat hippocampus. J. Comp. Neurol., 325: 169–182. Csicsvari, J., Hirase, H., Czurko, A. and Buzsaki, G. (1998) Reliability and state dependence of pyramidal cell-interneuron synapses in the hippocampus: an ensemble approach in the behaving rat. Neuron, 21: 179–189. Csicsvari, J., Hirase, H., Czurko, A., Mamiya, A. and Buzsa´ki, G. (1999) Oscillatory coupling of hippocampal pyramidal cells and interneurons in the behaving rat. J. Neurosci., 19: 274–287. Deller, T. (1998) The anatomical organization of the rat fascia dentata: new aspects of laminar organization as revealed by anterograde tracing with phaseolus vulgaris-luecoagglutinin (PHAL). Anat. Embryol. (Berl.), 197: 89–103. Deller, T., Katona, I., Cozzari, C., Frotscher, M. and Freund, T.F. (1999) Cholinergic innervation of mossy cells in the rat fascia dentata. Hippocampus, 9: 314–320. Deller, T. and Leranth, C. (1990) Synaptic connections of neuropeptide Y (NPY) immunoreactive neurons in the hilar area of the rat hippocampus. J. Comp. Neurol., 300: 433–447. Freund, T.F. and Buzsa´ki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Freund, T.F., Hajos, N., Acsady, L., Gorcs, T.J. and Katona, I. (1997) Mossy cells of the rat dentate gyrus are immunoreactive for calcitonin gene-related peptide (CGRP). Eur. J. Neurosci., 9: 1815–1830. Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E. (1991) The mossy cells of the fascia dentata: a comparative study of their fine structure and synaptic connections in rodents and primates. J. Comp. Neurol., 312: 145–163. Fujise, N. and Kosaka, T. (1999) Mossy cells in the mouse dentate gyrus: identification in the dorsal hilus and their distribution along the dorsoventral axis. Brain Res., 816: 500–511. Goodman, J.H., Wasterlain, C.G., Massarweh, W.F., Dean, E., Sollas, A.L. and Sloviter, R.S. (1993) Calbindin-D28k immunoreactivity and selective vulnerability to ischemia in the dentate gyrus of the developing rat. Brain Res., 606: 309–314. Gutierrez, R. (2003) The GABAergic phenotype of the ‘‘glutamatergic’’ granule cells of the dentate gyrus. Prog. Neurobiol., 71: 337–358. Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser, E.I. (2005) Microstructure of a spatial map in the entorhinal cortex. Nature, 436: 801–806. Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple GABAergic innervation of granule cells in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 411–429.

215 Henze, D.A. and Buzsaki, G. (2001) Action potential threshold of hippocampal pyramidal cells in vivo is increased by recent spiking activity. Neuroscience, 105: 121–130. Henze, D.A., McMahon, D.B., Harris, K.M. and Barrionuevo, G. (2002) Giant miniature EPSCs at the hippocampal mossy fiber to CA3 pyramidal cell synapse are monoquantal. J. Neurophysiol., 87: 15–29. Isomura, Y., Sirota, A., Montgomery, S., Mizuseki, K., O¨zen, S., Henze, D.A. and Buzsa´ki, G. (2006) Integration and Segregation of Activity in Entorhinal-hippocampal Subregions by Neocortical Slow Oscillations. Neuron, 52: 871–882. Kamondi, A., Acsady, L., Wang, X.J. and Buzsa´ki, G. (1998) Theta oscillations in somata and dendrites of hippocampal pyramidal cells in vivo: activity-dependent phase-precession of action potentials. Hippocampus, 8: 244–261. Kanerva, P. (1988) Sparse Distributed Memory. Bradford/MIT Press, Cambridge, MA. Klausberger, T., Magill, P.J., Marton, L.F., Roberts, J.D., Cobden, P.M., Buzsaki, G. and Somogyi, P. (2003) Brainstate- and cell-type-specific firing of hippocampal interneurons in vivo. Nature, 421: 844–848. Klausberger, T., Marton, L.F., Baude, A., Roberts, J.D., Magill, P.J. and Somogyi, P. (2004) Spike timing of dendritetargeting bistratified cells during hippocampal network oscillations in vivo. Nat. Neurosci., 7: 41–47. Klausberger, T., Marton, L.F., O’Neill, J., Huck, J.H., Dalezios, Y., Fuentealba, P., Suen, W.Y., Papp, E., Kaneko, T., Watanabe, M., Csicsvari, J. and Somogyi, P. (2005) Complementary roles of cholecystokinin- and parvalbuminexpressing GABAergic neurons in hippocampal network oscillations. J. Neurosci., 25: 9782–9793. Kohler, C. (1985) A projection from the deep layers of the entorhinal area to the hippocampal formation in the rat brain. Neurosci. Lett., 56: 13–19. Leranth, C., Szeidemann, Z., Hsu, M. and Buzsaki, G. (1996) AMPA receptors in the rat and primate hippocampus: a possible absence of GluR2/3 subunits in most interneurons. Neuroscience, 70: 631–652. Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The hippocampal CA3 network: an in vivo intracellular labeling study. J. Comp. Neurol., 339: 181–208. Lorente de No, R. (1934) Studies on the structure of the cerebral cortex. II. Continuation of the study of the ammonic system. J. Psychol. Neurol. (Leipzig), 46: 113–177. McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal synaptic enhancement and information storage within a distributed memory system. TINS, 10(10): 408–415. Mizumori, S.J., Barnes, C.A. and McNaughton, B.L. (1990) Behavioral correlates of theta-on and theta-off cells recorded from hippocampal formation of mature young and aged rats. Exp. Brain Res., 80: 365–373. Murakawa, R. and Kosaka, T. (2001) Structural features of mossy cells in the hamster dentate gyrus, with special reference to somatic thorny excrescences. J. Comp. Neurol., 429: 113–126. O’Keefe, J. and Nadel, L. (1978) Hippocampus as a cognitive map. Clarendon, Oxford.

Penttonen, M., Kamondi, A., Acsady, L. and Buzsaki, G. (1998) Gamma frequency oscillation in the hippocampus of the rat: intracellular analysis in vivo. Eur. J. Neurosci., 10: 718–728. Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki, G. (1997) Feed-forward and feed-back activation of the dentate gyrus in vivo during dentate spikes and sharp wave bursts. Hippocampus, 7: 437–450. Petralia, R.S. and Wenthold, R.J. (1992) Light and electron immunocytochemical localization of AMPA-selective glutamate receptors in the rat brain. J. Comp. Neurol., 318: 329–354. Ratzliff, A.H., Howard, A.L., Santhakumar, V., Osapay, I. and Soltesz, I. (2004) Rapid deletion of mossy cells does not result in a hyperexcitable dentate gyrus: implications for epileptogenesis. J. Neurosci., 24: 2259–2269. Ratzliff, A.H., Santhakumar, V., Howard, A. and Soltesz, I. (2002) Mossy cells in epilepsy: rigor mortis or vigor mortis? Trends Neurosci., 25: 140–144. Santhakumar, V., Bender, R., Frotscher, M., Ross, S.T., Hollrigel, G.S., Toth, Z. and Soltesz, I. (2000) Granule cell hyperexcitability in the early post-traumatic rat dentate gyrus: the ‘irritable mossy cell’ hypothesis. J. Physiol., 524(Pt 1): 117–134. Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the molecular layer have lower thresholds for synaptic activation by perforant path than granule cells. J. Neurosci., 11: 1660–1673. Scharfman, H.E. (1992a) Blockade of excitation reveals inhibition of dentate spiny hilar neurons recorded in rat hippocampal slices. J. Neurophysiol., 68: 978–984. Scharfman, H.E. (1992b) Differentiation of rat dentate neurons by morphology and electrophysiology in hippocampal slices: granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells. Epilepsy Res. Suppl., 7: 93–109. Scharfman, H.E. (1993a) Characteristics of spontaneous and evoked EPSPs recorded from dentate spiny hilar cells in rat hippocampal slices. J. Neurophysiol., 70: 742–757. Scharfman, H.E. (1993b) Spiny neurons of area CA3c in rat hippocampal slices have similar electrophysiological characteristics and synaptic responses despite morphological variation. Hippocampus, 3: 9–28. Scharfman, H.E. (1994a) EPSPs of dentate gyrus granule cells during epileptiform bursts of dentate hilar ‘‘mossy’’ cells and area CA3 pyramidal cells in disinhibited rat hippocampal slices. J. Neurosci., 14: 6041–6057. Scharfman, H.E. (1994b) Evidence from simultaneous intracellular recordings in rat hippocampal slices that area CA3 pyramidal cells innervate dentate hilar mossy cells. J. Neurophysiol., 72: 2167–2180. Scharfman, H.E. (1994c) Synchronization of area CA3 hippocampal pyramidal cells and non-granule cells of the dentate gyrus in bicuculline-treated rat hippocampal slices. Neuroscience, 59: 245–257. Scharfman, H.E. (1995) Electrophysiological evidence that dentate hilar mossy cells are excitatory and innervate both granule cells and interneurons. J. Neurophysiol., 74: 179–194.

216 Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A. (1990) Synaptic connections of dentate granule cells and hilar neurons: results of paired intracellular recordings and intracellular horseradish peroxidase injections. Neuroscience, 37: 693–707. Scharfman, H.E. and Schwartzkroin, P.A. (1988) Electrophysiology of morphologically identified mossy cells of the dentate hilus recorded in guinea pig hippocampal slices. J. Neurosci., 8: 3812–3821. Scharfman, H.E., Smith, K.L., Goodman, J.H. and Sollas, A.L. (2001) Survival of dentate hilar mossy cells after pilocarpineinduced seizures and their synchronized burst discharges with area CA3 pyramidal cells. Neuroscience, 104: 741–759. Sik, A., Penttonen, M. and Buzsa´ki, G. (1997) Interneurons in the hippocampal dentate gyrus: an in vivo intracellular study. Eur. J. Neurosci., 9: 573–588. Sloviter, R.S. (1989) Calcium-binding protein (calbindin-D28k) and parvalbumin immunocytochemistry: localization in the rat hippocampus with specific reference to the selective vulnerability of hippocampal neurons to seizure activity. J. Comp. Neurol., 280: 183–196. Sloviter, R.S. (1991a) Feedforward and feedback inhibition of hippocampal principal cell activity evoked by perforant path stimulation: GABA-mediated mechanisms that regulate excitability in vivo. Hippocampus, 1: 31–40. Sloviter, R.S. (1991b) Permanently altered hippocampal structure, excitability, and inhibition after experimental status epilepticus in the rat: the ‘‘dormant basket cell’’ hypothesis and its possible relevance to temporal lobe epilepsy. Hippocampus, 1: 41–66. Sloviter, R.S. (1994) The functional organization of the hippocampal dentate gyrus and its relevance to the pathogenesis of temporal lobe epilepsy. Ann. Neurol., 35: 640–654. Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D. (1991) Calcium-binding protein (calbindin-D28k) and parvalbumin immunocytochemistry in the normal and epileptic human hippocampus. J. Comp. Neurol., 308: 381–396. Sloviter, R.S., Zappone, C.A., Harvey, B.D., Bumanglag, A.V., Bender, R.A. and Frotscher, M. (2003) ‘‘Dormant basket cell’’ hypothesis revisited: relative vulnerabilities of dentate gyrus mossy cells and inhibitory interneurons after hippocampal status epilepticus in the rat. J. Comp. Neurol., 459: 44–76. Soltesz, I., Bourassa, J. and Deschenes, M. (1993) The behavior of mossy cells of the rat dentate gyrus during theta oscillations in vivo. Neuroscience, 57: 555–564. Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat fascia dentata are glutamate-immunoreactive. Hippocampus, 4: 65–69.

Steriade, M., Contreras, D., Curro, D.R. and Nunez, A. (1993a) The slow (o1 Hz) oscillation in reticular thalamic and thalamocortical neurons: scenario of sleep rhythm generation in interacting thalamic and neocortical networks. J. Neurosci., 13: 3284–3299. Steriade, M., Gloor, P., Llinas, R.R., Lopes de Silva, F.H. and Mesulam, M.M. (1990) Report of IFCN committee on basic mechanisms. Basic mechanisms of cerebral rhythmic activities. Electroencephalogr. Clin. Neurophysiol., 76: 481–508. Steriade, M., Nunez, A. and Amzica, F. (1993b) A novel slow (o1 Hz) oscillation of neocortical neurons in vivo: depolarizing and hyperpolarizing components. J. Neurosci., 13: 3252–3265. Steriade, M., Nunez, A. and Amzica, F. (1993c) Intracellular analysis of relations between the slow (o1 Hz) neocortical oscillation and other sleep rhythms of the electroencephalogram. J. Neurosci., 13: 3266–3283. Strowbridge, B.W., Buckmaster, P.S. and Schwartzkroin, P.A. (1992) Potentiation of spontaneous synaptic activity in rat mossy cells. Neurosci. Lett., 142: 205–210. Treves, A. and Rolls, E.T. (1992) Computational constraints suggest the need for two distinct input systems to the hippocampal CA3 network. Hippocampus, 2: 189–199. Treves, A. and Rolls, E.T. (1994) Computational analysis of the role of the hippocampus in memory. Hippocampus, 4: 374–391. Walker, M.C., Ruiz, A. and Kullmann, D.M. (2001) Monosynaptic GABAergic signaling from dentate to CA3 with a pharmacological and physiological profile typical of mossy fiber synapses. Neuron, 29: 703–715. Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E. and Schwartzkroin, P.A. (1997) Ultrastructural localization of neurotransmitter immunoreactivity in mossy cell axons and their synaptic targets in the rat dentate gyrus. Hippocampus, 7: 559–570. Wittner, L., Henze, D.A., Zaborszky, L. and Buzsaki, G. (2006) Hippocampal CA3 pyramidal cells selectively innervate aspiny interneurons. Eur. J. Neurosci., 24: 1286–1298. Wolansky, T., Clement, E.A., Peters, S.R., Palczak, M.A. and Dickson, C.T. (2006) Hippocampal slow oscillation: a novel EEG state and its coordination with ongoing neocortical activity. J. Neurosci., 26: 6213–6229. Ylinen, A., Bragin, A., Nadasdy, Z., Jando, G., Szabo, I., Sik, A. and Buzsa´ki, G. (1995) Sharp wave-associated high-frequency oscillation (200 Hz) in the intact hippocampus: network and intracellular mechanisms. J. Neurosci., 15: 30–46. Zappone, C.A. and Sloviter, R.S. (2004) Translamellar disinhibition in the rat hippocampal dentate gyrus after seizure-induced degeneration of vulnerable hilar neurons. J. Neurosci., 24: 853–864.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 13

Interneurons of the dentate gyrus: an overview of cell types, terminal fields and neurochemical identity Carolyn R. Houser1,2, 1

Department of Neurobiology and Brain Research Institute, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095, USA Research Service, VA Greater Los Angeles Healthcare System, West Los Angeles, Los Angeles, CA 90073, USA

2

Abstract: Interneurons of the dentate gyrus are a diverse group of neurons that use GABA as their primary neurotransmitter. Morphological studies of these neurons have been challenging since no single neuroanatomical method provides a complete view of these interneurons. However, through the integration of findings obtained from multiple methods, an interesting picture of this complex group of neurons is emerging, and this review focuses on studies in rats and mice. In situ hybridization of mRNAs for the two isoforms of the GABA synthesizing enzyme, glutamate decarboxylase (GAD65 and GAD67), demonstrates the abundance of GABA neurons in the dentate gyrus and their high concentration in the hilus and along the base of the granule cell layer. Likewise, immunohistochemical studies, particularly of GAD65, demonstrate the rich fields of GABA terminals not only around the somata of granule cells but also in the dendritic regions of the molecular layer. This broad group of GABA neurons and their terminals can be subdivided according to their morphological characteristics, including the distribution of their axonal plexus, and their neurochemical identity. Intracellular labeling of single interneurons has been instrumental in demonstrating the extensiveness of their axonal plexus and the relatively specific spatial distribution of their axonal fields. These findings have led to the broad classification of interneurons into those that terminate primarily at perisomatic regions and those that innervate the dendrites of granule cells. The interneurons also can be classified according to their neuropeptide and calcium-binding protein content. These and other molecules contribute to the rich diversity of dentate interneurons and may provide opportunities for selectively regulating specific groups of GABA neurons in the dentate gyrus in order to enhance their function or protect vulnerable neurons from damage. Keywords: GABA; glutamate decarboxylase (GAD); interneurons; hilus; parvalbumin; somatostatin; calretinin early morphological studies identified at least 21 cell types in the dentate hilus alone (Amaral, 1978). Despite such diversity, a unifying feature of the dentate interneurons is their use of g-aminobutyric acid (GABA) as their major neurotransmitter. However, as in other cortical and hippocampal regions, multiple subtypes of GABA

Introduction Interneurons of the dentate gyrus are a particularly fascinating and diverse group of neurons, and Corresponding author. Tel.: +1 (310) 206-1567; Fax: +1 (310) 825-2224; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63013-1

217

218

neurons can be distinguished, and several classification systems have emerged. These include categorization according to the axonal distributions and postsynaptic targets of the interneurons, their neuropeptide or calcium-binding protein content, and their physiological characteristics (Parra et al., 1998; Maccaferri and Lacaille, 2003; Somogyi and Klausberger, 2005; Jinno and Kosaka, 2006). Each system highlights special features of the interneuron population, and, while there is not a perfect overlap between categories in the different systems, some relationships can be established. This review will focus first on the broad distribution of GABA neuron cell bodies and axon terminals within the dentate gyrus, and then consider several major categories of dentate interneurons and their relationships to the overall distribution of GABA neurons and their processes.

Distribution of GABA neurons Demonstrating the entire population of GABA neurons in the dentate gyrus in histological preparations has proven to be an interesting challenge. Immunohistochemical studies utilizing various antisera to GABA and its synthesizing enzyme, glutamate decarboxylase (GAD), have yielded different and, occasionally, contradictory results (Ribak et al., 1978; Sloviter and Nilaver, 1987; Babb et al., 1988; Woodson et al., 1989; Jinno et al., 1998). Although cell bodies in many regions of the hippocampal formation are labeled with these methods, several groups of neurons in the dentate gyrus, particularly those in the hilus, are often inconsistently labeled. Alternate methods of identifying the cell bodies of GABA neurons in the dentate gyrus are thus needed. Currently, one of the most effective methods for demonstrating the entire population of GABA neurons in the dentate gyrus is in situ hybridization of either GAD65 or GAD67 mRNA (Houser et al., 2000, for review). Both forms of GAD can synthesize GABA, but they have functional differences that include different interactions with the cofactor pyridoxal phosphate (Erlander et al., 1991; Kaufman et al., 1991; Martin et al., 1991; Esclapez et al., 1994). However, the mRNAs for

both isoforms appear to be present in essentially all GABAergic interneurons of the dentate gyrus (Houser and Esclapez, 1994). Thus clear visualization of the entire population of GABA neuron somata is possible, and the abundance of these neurons within the relatively small region of the dentate gyrus is striking (Fig. 1). Such in situ hybridization methods for identifying GABA neurons have the additional advantage of labeling the cell bodies of these GABA neurons without occlusion by axon terminals. GAD mRNA-labeled neurons are distributed in all three relatively distinct regions of the dentate gyrus. First, GAD mRNA-labeled neurons are highly concentrated in the hilus, and their distribution often helps define the hilar region (Fig. 1A, B). While labeled interneurons are abundant in the deep parts of the hilus, between the tip of the hilus and CA3, they are also numerous in the hilar regions that are located beneath the granule cell layer (Fig. 1A, C). Second, labeled neurons associated with the granule cell layer are frequently aligned along the base of this layer and may be either highly concentrated or more dispersed along this zone (Fig. 1A–C). Some GAD mRNA-labeled cell bodies are also located within the granule cell layer and along its outer border. Finally, GABA neurons are distributed diffusely within the molecular layer, and a few are found along the hippocampal fissure at the outer border of the molecular layer (Fig. 1A, C). These basic patterns of GABA neurons are present in both the dorsal (septal) and ventral (temporal) regions of the dentate gyrus (Fig. 1B). The numbers of GAD mRNA-labeled neurons in all three regions (hilus, granule cell layer and molecular layer) generally appear to be larger than those identified with other labeling methods, and the numerous labeled neurons in the hilus are particularly notable. Their abundance could give the impression that most hilar neurons are GABA neurons. However, non-GABA neurons are also present in the hilus (Fig. 2A), in approximately equal numbers, and the vast majority of these are mossy cells that use glutamate as their transmitter. Double labeling for GAD mRNA and a general neuronal marker, such as the neuron-specific nuclear protein NeuN, demonstrates the extensive

219

Fig. 1. In situ hybridization of GAD65 in the rat dentate gyrus. (A) In a coronal section through the rostral dentate gyrus, labeled neurons are numerous in the deep regions of the hilus (H), between the hilar tip and CA3, and in the region beneath the granule cell layer (arrow). Labeled neurons are aligned along the base of the granule cell layer (G) and are dispersed in the molecular layer (M). (B) In a coronal section through a more caudal level of the dentate gyrus, labeled neurons are concentrated in the dorsal and ventral hilus and are numerous along the base of the granule cell layer (G). (C) At higher magnification, labeled neurons in the hilus (H) can be distinguished from those at the base of the granule cell layer (G). Labeled neurons are evident at all laminar levels of the molecular layer (M).

intermingling of GABA and non-GABA neurons throughout the hilus (Fig. 2A). (Mossy cells are generally considered to be principal cells and are discussed elsewhere in this volume; see reviews by Seress, Henze and Buszaki, and Scharfman.) Identifying the relatively large numbers of GAD mRNA-labeled neurons in the dentate gyrus provides a base on which to consider subpopulations of GABA neurons. Visualizing the entire group of GABA neurons in the region also makes it clear that only limited numbers of dentate GABA

neurons are being identified as the different subgroups of GABA neurons are labeled. As one example, NADPH-diaphorase labels exclusively GABAergic interneurons in the dentate gyrus, and the labeled neurons are relatively abundant (Valtschanoff et al., 1993; Czeh et al., 2005). Yet, even though NADPH-diaphorase-labeled neurons are present in all regions of the dentate gyrus and are not restricted to a specific subtype of interneuron, they represent only a portion of the GABA neurons in each region (Fig. 2B).

220

Fig. 2. Double labeling of GAD65 mRNA and either NeuN (A) or NADPH-diaphorase (B) in the rat dentate gyrus. (A) GAD65 mRNA-labeled neurons (black) are numerous throughout the dentate hilus (H) where they are intermingled with NeuN single-labeled (non-GABA) neurons (red; examples at arrows). Some GAD65 mRNA-labeled neurons are present in the granule cell layer (G) but are most abundant at the base of this layer. Most neurons in the molecular layer (M) are labeled for GAD65 mRNA. Granule cells and CA3 pyramidal cells are single-labeled for NeuN. (B) NADPH-diaphorase (dark blue) is present in only subgroups of the GAD mRNA-labeled neurons (red) in the hilus (H), granule cell layer (G) and molecular layer (M) (examples at arrows). See Plate 13.2 in Color Plate Section.

While it is expected that many peptides and calcium-binding proteins will define specific subgroups of GABA neurons, as will be discussed in later sections, it is less clear why immunohistochemical labeling of GABA, as well as GAD65 and GAD67 proteins, does not label all GABA neurons as reliably as the in situ hybridization labeling of GAD mRNAs. One possibility is that the actual levels of GAD and GABA in the cell bodies of GABA neurons vary, and that this

influences the extent of cell body labeling. The varying somal content could be related to physiological conditions, including the levels of activity of the neurons, as well as transport of GAD and GABA from the soma to axon terminals. As additional antibodies that recognize GAD and GABA are obtained and new labeling methods are developed, the most effective methods for labeling GABA neurons are likely to change. However, currently in situ hybridization of either

221

GAD65 or GAD67 mRNA yields the most complete and reliable labeling of cell bodies of the entire population of GABA neurons in the dentate gyrus. Immunohistochemical localization of the GAD proteins generally provides labeling of some but not all cell bodies, but these immunohistochemical methods are essential for demonstrating the extensive plexus of GABAergic terminals within the dentate gyrus. The following descriptions of immunohistochemical labeling will consider GABA neurons in both rat and mouse. The most detailed information on the distributions and connections of dentate interneurons has been obtained from rat, but there is considerable interest in the characteristics of these neurons in mice, due to the rapid increase in studies of genetically modified animals. Thus the illustrations of immunolabeling will be from mouse (C57/BL6) primarily. Current findings suggest, however, that most morphological and biochemical aspects of dentate interneurons are quite similar in rat and mouse (personal observations; Ma´tya´s et al., 2004).

Axonal arborizations and postsynaptic targets of GABA neurons GABAergic terminals are abundant throughout the dentate gyrus, but reach their highest densities in the outer third of the molecular layer (Fig. 3A). Slightly lower concentrations of GABAergic terminals are evident in the inner two-thirds of this layer. These high concentrations of GABAergic axon terminals in the dendritic regions of the granule cells are consistent with electron microscopic evidence that, in quantitative terms, GABAergic synapses in dendritic regions significantly outnumber those in somatic regions of the dentate gyrus (75% vs. 25%) (Halasy and Somogyi, 1993a). Within the granule cell layer, GABAergic terminals surround the granule cell somata, and a narrow band of increased labeling is present at the outer border of the layer (Fig. 3A). The lowest concentration of GABAergic terminals is found in the hilus where labeled terminals are dispersed throughout the neuropil. While the cell bodies of some hilar neurons are surrounded by GABAergic terminals, others exhibit little perisomatic labeling.

Several subgroups of GABA neurons contribute to the rich fields of GABAergic terminals in the dentate gyrus, and many of these interneurons have extensive axonal arborizations not only within a slice but also along the septotemporal length of the dentate gyrus (Han et al., 1993; Buckmaster and Schwartzkroin, 1995; Sik et al., 1997). Indeed the axonal patterns and postsynaptic targets of the interneurons prove to be the most distinguishing morphological features of the different subclasses of interneurons. The current classification of dentate interneurons is based primarily on the studies of Somogyi and colleagues in which interneurons were labeled intracellularly, and their axonal plexus and targets were identified with light and electron microscopic methods (Halasy and Somogyi, 1993b; Han et al., 1993). Interneurons in the dentate gyrus, as in other cortical regions, can be divided into those that form synapses at either perisomatic or dendritic locations. Interneurons with synapses at perisomatic sites can be divided further into those that synapse on either the cell bodies and proximal dendrites (basket cells) or the axon initial segments (axoaxonic cells) of granule cells. The cell bodies of many basket cells are located along the base of the granule cell layer but are also found within the granule cell layer and along its outer border. While basket cells have a variety of shapes and dendritic patterns, the pyramidal basket cells are the most readily identified because of their large cell bodies and long apical dendrites (Ribak and Seress, 1983). But regardless of the dendritic patterns, the axons are concentrated in the granule cell layer, where each basket cell forms synaptic contacts with the cell bodies of multiple granule cells (Han et al., 1993). The distributions of the axon terminals of basket cells on granule cell somata are reflected in the numerous GAD-labeled terminals throughout the granule cell layer (Fig. 3A). In addition, basket cells form synaptic contacts with other GABA neurons, including other basket cells (Freund and Buzsa´ki, 1996). A second class of GABA neurons with perisomatic targets is the axoaxonic cell. These interneurons are amazingly selective in their targets and, in the dentate gyrus, form synapses exclusively on axon initial segments of granule cells

222

Fig. 3. GAD65-labeled axon terminals in the mouse dentate gyrus (A) and hypothetical cell types of origin (B). (A & B) Dense terminal fields in the outer molecular layer (Mo) originate from hilar (H) and molecular layer GABA neurons that project to the perforant path zone (HIPP and MOPP cells, respectively). Moderately dense terminal fields in the inner molecular layer (Mi) originate from hilar interneurons that innervate the commissural-association projection region (HICAP cells). GABAergic terminals in the granule cell layer (G) orginate from basket and axoaxonic cells that provide perisomatic innervation.

(Soriano and Frotscher, 1989; Soriano et al., 1990; Han et al., 1993; DeFelipe, 1999). Their axons branch extensively, and each branch forms a series of synaptic contacts along an axon initial segment. The numerous descending branches resemble the candles or lights on a chandelier, and this led to their early identification as chandelier cells in the cerebral cortex (Szenta´gothai, 1974; Somogyi, 1977). The cell bodies of these neurons are often located near the granule cell layer, including locations along the outer border of the layer, but can also be found within the hilus. Because of the proximal location of the axon terminals of basket cells and axoaxonic cells, these neurons can exert powerful control over the output of granule cells and potentially contribute to synchronization of

granule cells through their contacts with multiple granule cells (Miles et al., 1996). The interneurons that terminate on granule cell dendrites, and form the GABAergic plexus in the molecular layer, can be divided into multiple classes based on the location of their cell bodies and the laminar distribution of their axon terminals in the molecular layer (Halasy and Somogyi, 1993b; Han et al., 1993). Interestingly, the cell bodies of neurons that innervate granule cell dendrites in the molecular layer are located in both the hilus and the molecular layer itself. GABA neurons with their cell bodies in the hilus and axon projections to the outer two-thirds of the molecular layer are identified as hilar perforant pathassociated (HIPP) cells. Likewise, neurons with

223

cell bodies in the hilus that project to the inner third of the molecular layer are identified as hilar commissural-association pathway-related (HICAP) cells. Because of the location of the cell bodies of these neurons in the hilus, they receive considerable input from the granule cells and are thus positioned to provide feedback inhibition to granule cells near the location of either the excitatory input from the perforant path or commissural-association neurons (primarily mossy cells) in the hilus (Han et al., 1993). In contrast, GABA neurons with cell bodies in the molecular layer are likely to receive excitatory input directly from the perforant path or, potentially, commissural-association fibers and thus could provide feed-forward inhibition to granule cell dendrites (Freund and Buzsa´ki, 1996). Molecular layer neurons that innervate the outer molecular layer have been named molecular layer perforant path-associated (MOPP) cells (Halasy and Somogyi, 1993b). Influences of these neurons on granule cells could be critical for regulating the response of the granule cells to their major excitatory inputs. As descriptions of specific cell types are obtained, it is possible to relate the general locations of the interneurons and their terminal fields to the broad distribution of axon terminals in the dentate gyrus (Fig. 3). Many of the cell bodies along the base of the granule cell layer, and within the layer, are likely to be basket cells and axoaxonic cells that contribute to the GABAergic terminals within the granule cell layer (Fig. 3). In contrast, many GABA neurons in the hilus are dendrite-innervating neurons and are a major source of the GABAergic innervation of the molecular layer. The dense innervation in the outer molecular layer is derived in part from the HIPP cells whereas the more moderate densities of terminals in the inner molecular region include axonal projections from the HICAP cells (Fig. 3). However, it is not possible to distinguish different classes of hilar neurons, such as the HICAP and HIPP cells, on the basis of their locations in the hilus or the patterns of their proximal dendrites. Most GABA neurons in the molecular layer presumably contribute to the GABAergic plexus within the layer, with the highest concentration of

their terminals in the perforant path zone (MOPP cells). Two other subgroups of GABA neurons in the dentate gyrus have unique patterns of axonal terminations but cannot be readily distinguished by their cell body locations, dendritic morphology, or laminar distributions of their axons. The first is a group of interneurons that selectively contact other interneurons. Many of these GABA neurons are located in the hilus and contain the calcium-binding proteins calretinin and calbindin in the rat (Gulya´s et al., 1996). However, the chemical identity of interneuron-specific GABA neurons in the mouse is not yet known. Another subgroup of interneurons in the hilus projects beyond the dentate to the medial septum. The hippocampo-septal neurons have been described most extensively in CA1 and CA3 of the rat (Alonso and Kohler, 1982; Toth and Freund, 1992). However, they are also found in the hilus, and many express somatostatin (Zappone and Sloviter, 2001; Jinno and Kosaka, 2002a; Gulya´s et al., 2003). In the septal region, these neurons form synapses with parvalbumin (PV)-containing GABA neurons that, in turn, project back to the hippocampus and dentate gyrus and innervate other GABA neurons (Freund and Antal, 1988; Acsa´dy et al., 2000b). This recurrent GABAergic circuit is viewed as a potential regulator of hippocampal theta activity although the precise mechanism of action remains to be determined (Dragoi et al., 1999).

Neurochemical identity Interneurons of the dentate gyrus can also be classified according to their content of specific calcium-binding proteins and neuropeptides, and a number of distinct groups of interneurons have been identified, even though overlap is found among some subgroups. By studying the cell bodies and axon terminals of these chemically defined interneurons, it has been possible to ‘‘dissect’’ the broad, complex pattern of GABAergic cell bodies and terminals and relate several of the groups to the morphologically identified cell types discussed previously.

224

Parvalbumin PV-containing interneurons have a restricted distribution of their axon terminals and form synapses predominantly on cell bodies and axon initial segments of granule cells, consistent with their identity as basket and axo-axonic cells (Ribak et al., 1990). PV-labeled cell bodies are located primarily near the granule cell layer and are most prominent at the base of the granule cell layer (Fig. 4A). However, some PV-labeled interneurons are also located near the junction of the granule cell and molecular layers (Fig. 4A, C), and smaller numbers of PV-labeled neurons are found in the hilus and molecular layer. Many labeled neurons along the base of the granule cell layer resemble pyramidal basket cells, with labeled apical dendrites that often branch and ascend through the full extent of the molecular layer and basal

dendrites that extend in the subgranular region (Fig. 4A). The PV-labeled neurons receive substantial input from axon collaterals of the mossy fibers through synaptic contacts on their cell bodies and proximal dendrites within the granule cell layer and also on their basal dendrites (Blasco-Iba´nez et al., 2000). These interneurons can thus provide strong feedback inhibition of the granule cells. In addition, the ascending dendrites of many PV-labeled interneurons are positioned to receive input from the major afferents to the dentate gyrus, including the entorhinal cortex (Zipp et al., 1989). Thus the PV-labeled interneurons can also provide feed-forward inhibition to the granule cells, in response to excitatory inputs to their dendrites in the molecular layer. Although PV-containing neurons are considered to be a prominent class of interneurons in the

Fig. 4. Parvalbumin (PV)-labeled neurons in the mouse dentate gyrus. (A) Cell bodies of PV-labeled neurons are located at the base of the granule cell layer (G), and their apical dendrites ascend through the granule cell layer and often branch as they extend through the molecular layer (M). Basal dendrites are evident along the inner border of the granule cell layer (arrows). (B) A large PV-labeled interneuron has the characteristics of a pyramidal basket cell and is located at the base of the granule cell layer (G). (C) A PV-labeled multipolar interneuron in the inner molecular layer extends several dendrites (arrows) into the molecular layer (M). The axon is likely to descend and contribute to the axonal plexus in the granule cell layer (G). (D) PV-labeled terminals are concentrated in the granule cell layer where they surround the cell bodies and axon initial segments of granule cells.

225

dentate gyrus, the number of labeled cell bodies in the dentate is often relatively low (Ribak et al., 1990). The percentage of GAD67-labeled interneurons that express PV in the granule cell layer is estimated to be 20–25% in the mouse (Jinno and Kosaka, 2002b). This contrasts with the considerably higher 40% of interneurons that express PV in the pyramidal cell layer of CA3 and CA1 of the hippocampus (Jinno and Kosaka, 2002b). The limited number of PV-labeled cell bodies could reflect generally low numbers of PV-containing dentate interneurons, or, possibly, a low level of the protein in the cell bodies despite the presence of PV in the axon terminals. Such differential intracellular labeling has been suggested in some experimental and pathological conditions, including human temporal lobe epilepsy, in which cell body labeling of PV is substantially decreased despite persistent terminal labeling (Sloviter et al., 1991; Scotti et al., 1997; Wittner et al., 2001). The precise function of PV in the interneurons remains uncertain, but the presence of this calcium-binding protein could be a physiological adaptation related to the high levels of activity of this group of neurons and their identification as fast-spiking neurons in several brain regions (Freund and Buzsa´ki, 1996). Interestingly, PVlabeled interneurons express particularly high levels of cytochrome c in the rat, and this, along with high numbers of mitochondrial profiles in their axon terminals, is consistent with high levels of metabolic activity in this group of GABA neurons (Gulya´s et al., 2006).

Cholecystokinin The peptide cholecystokinin (CCK) is also located in interneurons near the granule cell layer, and these neurons constitute a second group of GABAergic basket cells. There appears to be little overlap between the CCK- and PV-labeled neurons in the hippocampus, and this suggests that there are different groups of basket cells with potentially different but complementary functions, as has been demonstrated for basket cells in CA1 (Klausberger et al., 2005). Relatively few CCK-labeled cell bodies are found in the dentate gyrus, but their terminal fields

resemble the patterns of PV-containing terminals within the granule cell layer, including a slightly increased density of fibers and terminals along the outer border of the granule cell layer. In contrast to the PV-containing group, CCK-labeled neurons are restricted to the basket cell population and apparently do not contribute to the axoaxonic subgroup, at least in the rat.

Somatostatin Somatostatin neurons constitute one of the largest chemically defined subgroups of GABA neurons in the dentate gyrus, and virtually all somatostatin neurons of this region are located within the hilus (Fig. 5A). Furthermore, approximately 55% of hilar GABA neurons express somatostatin (Esclapez and Houser, 1995; Buckmaster and Jongen-Reˆlo, 1999; Jinno and Kosaka, 2003). In both mouse and rat, somatostatin-containing neurons are more abundant, and constitute a slightly higher percentage of total GABA neurons in the ventral than in the dorsal dentate gyrus (Buckmaster et al., 1994; Esclapez and Houser, 1995; Jinno and Kosaka, 2003). While somatostatin-labeled cell bodies are confined primarily to the hilus in the dentate gyrus, their terminals fields are most prominent in the outer molecular layer where the perforant path fibers terminate (Bakst et al., 1986; Katona et al., 1999). Thus many of the somatostatin neurons in the hilus are considered to be HIPP cells, based on the location of their cell bodies and axon terminals. The dendrites of somatostatin neurons extend for considerable distances across the hilus but generally remain within the region (Fig. 5B). Although the proximal dendrites of many of the somatostatin interneurons are smooth and sparsely spinous, as is characteristic of many GABAergic interneurons (Ribak, 1978), the more distal dendrites have a considerable number of spines (Fig. 5C). These extensive dendrites and their spines are likely targets of the granule cell mossy fibers, and the mossy fiber contacts on GABA neurons outnumber those on mossy cells by a ratio of approximately 5:1 (Acsa´dy et al., 1998). The somatostatin neurons, in turn, form a moderately dense

226

Fig. 5. Somatostatin-containing neurons in the ventral dentate gyrus of the mouse. (A) Numerous somatostatin-immunoreactive cell bodies are present in the hilus (H), and a labeled axonal plexus is evident in the outer molecular layer (M). Little labeling is found in the granule cell layer (G) except for axons from hilar somatostatin neurons that project through this layer. (B) Localization of green fluorescent protein (GFP) in somatostatin neurons of a transgenic mouse reveals the extensive dendrites of these neurons. Several dendrites extend across large extents of the hilus (arrows). (C) GFP-labeled dendrites extend from a multipolar hilar neuron, and small spines can be detected on some dendritic segments (arrows).

axonal plexus in the outer two-thirds of the molecular layer (Bakst et al., 1986). The somatostatin terminals synapse primarily with granule cell dendrites and are often located adjacent to asymmetric excitatory synapses (Katona et al., 1999). Somatostatin neurons are thus ideally positioned to modulate the influence of the perforant path on dentate granule cells in response to ongoing granule cell activity (see chapter by Tallent in this volume). Some hilar somatostatin neurons project beyond the dentate gyrus to the medial septum. In both mouse and rat, a high proportion of the hippocampo-septal neurons in the hilus express somatostatin (Zappone and Sloviter, 2001; Jinno and Kosaka, 2002a). Furthermore, retrograde labeling studies in the mouse indicate that a surprisingly high percentage (44%) of somatostatin-labeled neurons in the hilus project to the septal region (Jinno and Kosaka, 2002a). Currently, it is unclear whether separate groups of GABA neurons project to either the dentate molecular layer or the medial septum, or whether single somatostatin neurons project to both regions. Regardless, the

high percentage of hilar somatostatin neurons that innervate the septum suggests that somatostatin as well as GABA could have important functional roles in the hippocampal-septal circuitry. Somatostatin neurons in the dentate gyrus are highly vulnerable to excitatory and ischemic damage, and loss of these neurons is commonly observed in models of temporal lobe epilepsy and forebrain ischemia (e.g. Johansen et al., 1987; Sloviter, 1987; Buckmaster and Dudek, 1997). Characteristics of these neurons that may contribute to their vulnerability include the high levels of excitatory mossy fiber contacts (Acsa´dy et al., 1998) and the relatively low GABAergic innervation on the soma and proximal dendrites of these interneurons (Acsa´dy et al., 2000a).

Calretinin Most classes of GABAergic interneurons in the dentate gyrus appear to be quite similar in their patterns of neuropeptide and calcium-binding

227

Fig. 6. Comparisons of calretinin labeling in the dorsal dentate gyrus of mouse (A) and rat (B). (A) Calretinin labeling in the mouse includes darkly-labeled (arrows) and moderately-labeled (arrowheads) neurons in the hilus (H) that correspond to GABAergic interneurons and mossy cells, respectively. In addition, small labeled cell bodies at the base of the granule cell layer (G) are likely to be newly-generated granule cells. A dense band of labeled axon terminals (asterisk) from the mossy cells is prominent in the inner molecular layer (M). (B) In the rat, darkly-labeled neurons (arrows) in the hilus (H) are presumed to be GABA neurons. A narrow band of labeled terminals is evident along the outer border of the granule cell layer (G), and these may originate in the supramammillary region.

protein expression in rats and mice. However, several species differences are found in the neurons that express the calcium-binding protein calretinin (Liu et al., 1996; Blasco-Iba´n˜ez and Freund, 1997). Most notably, in the mouse, calretinin is expressed in mossy cells of the hilus, and such mossy cells are particularly numerous in the ventral hilus (BlascoIba´n˜ez and Freund, 1997; Fujise et al., 1998). Calretinin is present in axon terminals of the mossy cells, and this creates a dense band of labeling in the inner third of the molecular layer (the major terminal field of the mossy cells) in the mouse (Fig. 6A). In contrast, labeling is primarily restricted to GABAergic interneurons of the dentate gyrus in the rat, and only a narrow band of calretinin labeling is present in the supragranular region in the rat (Fig. 6B). These terminals are not from mossy cells but may originate from the supramammillary region and are considered to be glutamatergic (Maglo´czky et al., 1994; Kiss et al., 2000). Despite calretinin labeling of the mossy cells in the mouse, some GABAergic interneurons that coexpress calretinin can be detected (Gulya´s et al., 1996). The GABA neurons are located primarily in the hilus, near the granule cell layer, and sometimes can be distinguished by their darker labeling compared to that of mossy cells in the region

(Fig. 6A). Another difference between the two species is the lack of spiny calretinin-containing neurons in the mouse (Blasco-Iba´n˜ez and Freund, 1997; Fujise et al., 1998; Ma´tya´s et al., 2004). Such neurons are distinctive hilar neurons in the rat and belong to the subgroup of GABA neurons that contacts primarily other interneurons (Gulya´s et al., 1996). Finally, in the mouse, newborn granule cells along the base of the granule cell layer express calretinin (Fig. 6A) (Liu et al., 1996; Blasco-Iba´n˜ez and Freund, 1997). Likewise, CajalRetzius cells that are presumed to use glutamate as their neurotransmitter are labeled for calretinin in the mouse. Thus, calretinin labeling of cell bodies in the dentate gyrus is largely restricted to GABA neurons in the rat but is present in several groups of glutamatergic, as well as some GABAergic, neurons, in the mouse.

Neuropeptide Y Within the dentate gyrus, neuropeptide Y (NPY)labeled cell bodies are located primarily in the hilus (Fig. 7) and constitute one of the larger groups of hilar GABA neurons. It has been estimated that NPY-containing neurons account for

228

Fig. 7. NPY-labeled neurons in the dorsal (A) and ventral (B) hilus of the mouse dentate gyrus. (A & B) Labeled neurons are largely confined to the hilus (H) at all levels of the dentate gyrus. Labeled neurons are relatively numerous in the hilus but are sparse in the granule cell (G) and molecular (M) layers as well as the pyramidal cell layer of CA3.

approximately 57–82% of GAD67-labeled neurons in the dorsal and ventral hilus, respectively (Jinno and Kosaka, 2003). These relatively high percentages suggest that there is considerable colocalization of NPY and somatostatin in GABA neurons in the hilus (Kohler et al., 1987; Deller and Leranth, 1990). Small numbers of NPY-labeled cell bodies are also located at the base of the granule cell layer and in the molecular layer. The axon terminal labeling of NPY-expressing neurons is generally limited with current immunohistochemical methods, but some terminals are evident in the outer molecular layer, suggesting that many of the NPY-containing neurons are HIPP cells and resemble somatostatin neurons. Some NPYlabeled axon terminals are also present in the hilus, and many of their targets are likely to be mossy cells (Deller and Leranth, 1990; Acsa´dy et al., 2000a).

Further characterization of dentate interneurons The presence of specific neuropeptides and calcium-binding proteins in selected interneurons has been extremely useful for identifying major classes of GABA neurons in the hippocampal formation. Yet, in some regions, the number of these neurons appears small when compared to the total number of GABA neuron cell bodies and terminals.

GABA neurons in two locations within the dentate gyrus appear to be particularly underrepresented by the currently available markers for subgroups of GABA neurons. These include the numerous GABA neurons in the hilar or polymorph region that is located below the base of the granule cell layer and GABA neurons in the molecular layer. Other molecules however call attention to some of these GABA neurons. For example, virtually all of the molecular layer interneurons express high levels of the a1 subunit of the GABAA receptor, as do many, but not all, groups of interneurons in the dentate gyrus (Fig. 8) (Gao and Fritschy, 1994; Esclapez et al., 1996; Sperk et al., 1997; Peng et al., 2004). In addition, these molecular layer interneurons express the d subunit of the GABAA receptor and, consistent with such expression, exhibit tonic inhibition (Glykys et al., 2007). Activation of GABAA receptors on the cell bodies of GABA neurons could be an additional mechanism through which the activity of GABA neurons is modulated or limited. Inhibition of even small numbers of GABA neurons that have extensive axonal projections, such as those in the molecular layer, could reduce inhibition and thus enhance excitation of the granule cells. Numerous other peptides and proteins are localized in GABA neurons in the dentate gyrus, and their identities and distributions highlight additional, unique characteristics of these interneurons.

229

1992; Buckmaster and Dudek, 1997). Continued characterization of the vulnerable groups of GABA neurons could eventually lead to specific neuroprotective strategies. Other molecules are localized primarily in interneurons but are found in several subtypes of GABA neurons in the hippocampal formation. These includes Reelin (Pesold et al., 1998) and nerve growth factor (Lauterborn et al., 1993; Pascual et al., 1998). The preferential localization of these particular substances in interneurons suggests roles for these GABA neurons in basic developmental processes that extend beyond their essential roles in regulating information processing, rhythm generation and excitability of the hippocampal formation. Conclusions Fig. 8. GABAA receptor a1 subunit labeling in the mouse dentate gyrus. The a1 subunit is highly expressed on the surface of cell bodies and dendritic processes of interneurons in the granule cell (G) and molecular (M) layers (examples at arrows). Labeled neurons at the base of the granule cell layer closely resemble parvalbumin-labeled pyramidal basket cells (compare with Fig. 4B).

Some expression patterns mirror the major subdivisions of interneurons, based on their innervation of perisomatic or dendritic regions, for which PV and somatostatin serve as prototypic markers, respectively. For example, mu-opioid receptors are preferentially localized on PV-labeled neurons whereas delta-opioid receptors are more abundant on somatostatin neurons (Commons and Milner, 1996; Stumm et al., 2004; Drake and Milner, 2006). Likewise, the a1 subunit of the GABAA receptor is highly expressed on PV-labeled neurons, as well as other interneurons in the dentate gyrus, but shows very little or no expression on somatostatin neurons in the hilus (Gao and Fritschy, 1994; Esclapez et al., 1996). Differences between these major groups of dentate interneurons are also reflected in their vulnerability to ischemic and seizure-induced damage. Somatostatin interneurons are some of the most vulnerable neurons in the dentate gyrus, whereas PV-containing neurons are more resistant to damage in some models (Sloviter, 1987; Sperk et al.,

Despite years of study, GABA neurons in the dentate gyrus continue to hold surprises and challenges. No single morphological method provides an adequate view of the richness and complexity of these neurons. Thus the findings obtained from different methods are pieces of a puzzle that require assembly. Key features of the emerging picture are the extensive and yet highly defined axonal plexus of single interneurons in the dentate gyrus; the rich repertoire of molecules that each interneuron possesses; and the remarkable functional regulation of levels of GABA, GAD and many neuropeptides in these interneurons. While such complexity can be daunting, it also provides opportunities for discovering new ways to enhance normal GABAergic function and possibly protects these critical neurons from damage in pathological conditions.

Acknowledgments I thank Christine Huang and Siroun Tahtakran for excellent histologic and photographic work, and in situ hybridization labeling, respectively, and gratefully acknowledge the numerous contributions of Drs. Monique Esclapez, Zechun Peng and Nianhui Zhang to our studies of GABA neurons in the hippocampus. These studies were supported

230

by Veterans Administration Medical Research Funds and National Institutes of Health grants NS046524 and NS051311 to C.R.H.

References Acsa´dy, L., Kamondi, A., Sik, A., Freund, T. and Buzsa´ki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Acsa´dy, L., Katona, I., Martinez-Guijarro, F.J., Buzsa´ki, G. and Freund, T.F. (2000a) Unusual target selectivity of perisomatic inhibitory cells in the hilar region of the rat hippocampus. J. Neurosci., 20: 6907–6919. Acsa´dy, L., Pascual, M., Rocamora, N., Soriano, E. and Freund, T.F. (2000b) Nerve growth factor but not neurotrophin-3 is synthesized by hippocampal GABAergic neurons that project to the medial septum. Neuroscience, 98: 23–31. Alonso, A. and Kohler, C. (1982) Evidence for separate projections of hippocampal pyramidal and non-pyramidal neurons to different parts of the septum in the rat brain. Neurosci. Lett., 31: 209–214. Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182: 851–914. Babb, T.L., Pretorius, J.K., Kupfer, W.R. and Brown, W.J. (1988) Distribution of glutamate-decarboxylase-immunoreactive neurons and synapses in the rat and monkey hippocampus: light and electron microscopy. J. Comp. Neurol., 278: 121–138. Bakst, I., Avendano, C., Morrison, J.H. and Amaral, D.G. (1986) An experimental analysis of the origins of somatostatin-like immunoreactivity in the dentate gyrus of the rat. J. Neurosci., 6: 1452–1462. Blasco-Iba´n˜ez, J.M. and Freund, T.F. (1997) Distribution, ultrastructure, and connectivity of calretinin-immunoreactive mossy cells of the mouse dentate gyrus. Hippocampus, 7: 307–320. Blasco-Iba´nez, J.M., Martı´ nez-Guijarro, F.J. and Freund, T.F. (2000) Recurrent mossy fibers preferentially innervate parvalbumin-immunoreactive interneurons in the granule cell layer of the rat dentate gyrus. Neuroreport, 11: 3219–3225. Buckmaster, P.S. and Dudek, F.E. (1997) Neuron loss, granule cell axon reorganization, and functional changes in the dentate gyrus of epileptic kainate-treated rats. J. Comp. Neurol., 385: 385–404. Buckmaster, P.S. and Jongen-Reˆlo, A.L. (1999) Highly specific neuron loss preserves lateral inhibitory circuits in the dentate gyrus of kainate-induced epileptic rats. J. Neurosci., 19: 9519–9529. Buckmaster, P.S., Kunkel, D.D., Robbins, R.J. and Schwartzkroin, P.A. (1994) Somatostatin-immunoreactivity in the hippocampus of mouse, rat, guinea pig, and rabbit. Hippocampus, 4: 167–180.

Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons and inhibition in the dentate gyrus of the rat in vivo. J. Neurosci., 15(1): 774–789. Commons, K.G. and Milner, T.A. (1996) Cellular and subcellular localization of delta opioid receptor immunoreactivity in the rat dentate gyrus. Brain Res., 738: 181–195. Czeh, B., Hajnal, A. and Seress, L. (2005) NADPH-diaphorase positive neurons of the rat hippocampal formation: regional distribution, total number and colocalization with calcium binding proteins. Prague Med. Rep., 106: 261–274. DeFelipe, J. (1999) Chandelier cells and epilepsy. Brain, 122(Pt 10): 1807–1822. Deller, T. and Leranth, C. (1990) Synaptic connections of neuropeptide Y (NPY) immunoreactive neurons in the hilar area of the rat hippocampus. J. Comp. Neurol., 300: 433–447. Dragoi, G., Carpi, D., Recce, M., Csicsvari, J. and Buzsaki, G. (1999) Interactions between hippocampus and medial septum during sharp waves and theta oscillation in the behaving rat. J. Neurosci., 19: 6191–6199. Drake, C.T. and Milner, T.A. (2006) Mu opioid receptors are extensively co-localized with parvalbumin, but not somatostatin, in the dentate gyrus. Neurosci. Lett., 403: 176–180. Erlander, M.G., Tillakaratne, N.J.K., Feldblum, S., Patel, N. and Tobin, A.J. (1991) Two genes encode distinct glutamate decarboxylases. Neuron, 7: 91–100. Esclapez, M., Chang, D.K. and Houser, C.R. (1996) Subpopulations of GABA neurons in the dentate gyrus express high levels of the a1 subunit of the GABAA receptor. Hippocampus, 6: 225–238. Esclapez, M. and Houser, C.R. (1995) Somatostatin neurons are a subpopulation of GABA neurons in the rat dentate gyrus: evidence from co-localization of pre-prosomatostatin and glutamate decarboxylase mRNAs. Neuroscience, 64: 339–355. Esclapez, M., Tillakaratne, N.J.K., Kaufman, D.L., Tobin, A.J. and Houser, C.R. (1994) Comparative localization of two forms of glutamic acid decarboxylase and their mRNAs in rat brain supports the concept of functional differences between the forms. J. Neurosci., 14: 1834–1855. Freund, T.F. and Antal, M. (1988) GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature, 336: 170–173. Freund, T.F. and Buzsa´ki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Fujise, N., Liu, Y., Hori, N. and Kosaka, T. (1998) Distribution of calretinin immunoreactivity in the mouse dentate gyrus: II. Mossy cells, with special reference to their dorsoventral difference in calretinin immunoreactivity. Neuroscience, 82: 181–200. Gao, B. and Fritschy, J.M. (1994) Selective allocation of GABAA receptors containing the a1 subunit to neurochemically distinct subpopulations of rat hippocampal interneurons. Eur. J. Neurosci., 6: 837–853. Glykys, J., Peng, Z., Chandra, D., Homanics, G.E., Houser, C.R. and Mody, I. (2007) A new naturally occurring GABAA receptor subunit partnership with high sensitivity to ethanol. Nat. Neurosci., 10: 40–48.

231 Gulya´s, A.I., Hajos, N., Katona, I. and Freund, T.F. (2003) Interneurons are the local targets of hippocampal inhibitory cells which project to the medial septum. Eur. J. Neurosci., 17: 1861–1872. Gulya´s, A.I., Buzsa´ki, G., Freund, T.F. and Hirase, H. (2006) Populations of hippocampal inhibitory neurons express different levels of cytochrome c. Eur. J. Neurosci., 23: 2581–2594. Gulya´s, A.I., Ha´jos, N. and Freund, T.F. (1996) Interneurons containing calretinin are specialized to control other interneurons in the rat hippocampus. J. Neurosci., 16(10): 3397–3411. Halasy, K. and Somogyi, P. (1993a) Distribution of GABAergic synapses and their targets in the dentate gyrus of rat: a quantitative immunoelectron microscopic analysis. J. Hirnforsch, 34: 299–308. Halasy, K. and Somogyi, P. (1993b) Subdivision in the multiple GABAergic innervation of granule cells in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 411–429. Han, Z.-S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 395–410. Houser, C.R. and Esclapez, M. (1994) Localization of mRNAs encoding two forms of glutamic acid decarboxylase in the rat hippocampal formation. Hippocampus, 5: 530–545. Houser, C.R., Esclapez, M. and Zhang, N. (2000) GABA Neurons of the hippocampal formation: localization, vulnerability and plasticity. In: Martin D.L. and Olsen R.W. (Eds.), GABA in the Nervous System: The View at Fifty Years. Lippincott Williams & Wilkins, Philadelphia, PA, pp. 337–355. Jinno, S., Aika, Y., Fukuda, T. and Kosaka, T. (1998) Quantitative analysis of GABAergic neurons in the mouse hippocampus, with optical disector using confocal laser scanning microscope. Brain Res., 814(1–2): 55–70. Jinno, S. and Kosaka, T. (2002a) Immunocytochemical characterization of hippocamposeptal projecting GABAergic nonprincipal neurons in the mouse brain: a retrograde labeling study. Brain Res., 945: 219–231. Jinno, S. and Kosaka, T. (2002b) Patterns of expression of calcium binding proteins and neuronal nitric oxide synthase in different populations of hippocampal GABAergic neurons in mice. J. Comp. Neurol., 449: 1–25. Jinno, S. and Kosaka, T. (2003) Patterns of expression of neuropeptides in GABAergic nonprincipal neurons in the mouse hippocampus: quantitative analysis with optical disector. J. Comp. Neurol., 461: 333–349. Jinno, S. and Kosaka, T. (2006) Cellular architecture of the mouse hippocampus: a quantitative aspect of chemically defined GABAergic neurons with stereology. Neurosci. Res., 56: 229–245. Johansen, F.F., Zimmer, J. and Diemer, N.H. (1987) Early loss of somatostatin neurons in dentate hilus after cerebral ischemia in the rat precedes CA-1 pyramidal cell loss. Acta Neuropathol. (Berl.), 73: 110–114.

Katona, I., Acsa´dy, L. and Freund, T. (1999) Postsynaptic targets of somatostatin-immunoreactive interneurons in the rat hippocampus. Neuroscience, 88: 37–55. Kaufman, D.L., Houser, C.R. and Tobin, A.J. (1991) Two forms of the gamma-aminobutyric acid synthetic enzyme glutamate decarboxylase have distinct intraneuronal distributions and cofactor interactions. J. Neurochem., 56: 720–723. Kiss, J., Csaki, A., Bokor, H., Shanabrough, M. and Leranth, C. (2000) The supramammillo-hippocampal and supramammillo-septal glutamatergic/aspartatergic projections in the rat: a combined [3H]D-aspartate autoradiographic and immunohistochemical study. Neuroscience, 97: 657–669. Klausberger, T., Marton, L.F., O’Neill, J., Huck, J.H., Dalezios, Y., Fuentealba, P., Suen, W.Y., Papp, E., Kaneko, T., Watanabe, M., Csicsvari, J. and Somogyi, P. (2005) Complementary roles of cholecystokinin and parvalbumin-expressing GABAergic neurons in hippocampal network oscillations. J. Neurosci., 25: 9782–9793. Kohler, C., Eriksson, L.G., Davies, S. and Chan-Palay, V. (1987) Co-localization of neuropeptide tyrosine and somatostatin immunoreactivity in neurons of individual subfields of the rat hippocampal region. Neurosci. Lett., 78: 1–6. Lauterborn, J.C., Tran, T.M., Isackson, P.J. and Gall, C.M. (1993) Nerve growth factor mRNA is expressed by GABAergic neurons in rat hippocampus. Neuroreport, 5: 273–276. Liu, Y., Fujise, N. and Kosaka, T. (1996) Distribution of calretinin immunoreactivity in the mouse dentate gyrus. I. General description. Exp. Brain Res., 108: 389–403. Maccaferri, G. and Lacaille, J.C. (2003) Interneuron diversity series: hippocampal interneuron classifications — making things as simple as possible, not simpler. Trends Neurosci., 26: 564–571. Maglo´czky, Z., Acsa´dy, L. and Freund, T. (1994) Principal cells are the postsynaptic targets of supramammillary afferents in the hippocampus of the rat. Hippocampus, 4: 322–334. Martin, D.L., Martin, S.B., Wu, S.J. and Espina, N. (1991) Cofactor interactions and the regulation of glutamate decarboxylase activity. Neurochem. Res., 16: 243–249. Ma´tya´s, F., Freund, T.F. and Gulya´s, A.I. (2004) Immunocytochemically defined interneuron populations in the hippocampus of mouse strains used in transgenic technology. Hippocampus, 14: 460–481. Miles, R., To´th, K., Gulya´s, A.I., Ha´jos, N. and Freund, T.F. (1996) Differences between somatic and dendritic inhibition in the hippocampus. Neuron, 16: 815–823. Parra, P., Gulyas, A.I. and Miles, R. (1998) How many subtypes of inhibitory cells in the hippocampus? Neuron, 20: 983–993. Pascual, M., Rocamora, N., Acsa´dy, L., Freund, T.F. and Soriano, E. (1998) Expression of nerve growth factor and neurotrophin-3 mRNAs in hippocampal interneurons: morphological characterization, levels of expression, and colocalization of nerve growth factor and neurotrophin-3. J. Comp. Neurol., 395: 73–90.

232 Peng, Z., Huang, C.S., Stell, B.M., Mody, I. and Houser, C.R. (2004) Altered expression of the d subunit of the GABAA receptor in a mouse model of temporal lobe epilepsy. J. Neurosci., 24: 8629–8639. Pesold, C., Impagnatiello, F., Pisu, M.G., Uzunov, D.P., Costa, E., Guidotti, A. and Caruncho, H.J. (1998) Reelin is preferentially expressed in neurons synthesizing gamma-aminobutyric acid in cortex and hippocampus of adult rats. Proc. Natl. Acad. Sci. U.S.A., 95: 3221–3226. Ribak, C.E. (1978) Aspinous and sparsely-spinous stellate neurons in the visual cortex of rats contain glutamic acid decarboxylase. J. Neurocytol., 7: 461–478. Ribak, C.E., Nitsch, R. and Seress, L. (1990) Proportion of parvalbumin-positive basket cells in the GABAergic innervation of pyramidal and granule cells of the rat hippocampal formation. J. Comp. Neurol., 300: 449–461. Ribak, C.E. and Seress, L. (1983) Five types of basket cell in the hippocampal dentate gyrus: a combined Golgi and electron microscopic study. J. Neurocytol., 12: 577–597. Ribak, C.E., Vaughn, J.E. and Saito, K. (1978) Immunocytochemical localization of glutamic acid decarboxylase in neuronal somata following colchicine inhibition of axonal transport. Brain Res., 140: 315–332. Scotti, A.L., Bollag, O., Kalt, G. and Nitsch, C. (1997) Loss of perikaryal parvalbumin immunoreactivity from surviving GABAergic neurons in the CA1 field of epileptic gerbils. Hippocampus, 7: 524–535. Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in the hippocampal dentate gyrus: an in vivo intracellular study. Eur. J. Neurosci., 9: 573–588. Sloviter, R.S. (1987) Decreased hippocampal inhibition and a selective loss of interneurons in experimental epilepsy. Science, 235: 73–76. Sloviter, R.S. and Nilaver, G. (1987) Immunocytochemical localization of GABA-, cholecystokinin-, vasoactive intestinal polypeptide-, and somatostatin-like immunoreactivity in the area dentata and hippocampus of the rat. J. Comp. Neurol., 256: 42–60. Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D. (1991) Calcium-binding protein (calbindin-D28K) and parvalbumin immunocytochemistry in the normal and epileptic human hippocampus. J. Comp. Neurol., 308: 381–396. Somogyi, P. (1977) A specific ‘axo-axonal’ interneuron in the visual cortex of the rat. Brain Res., 136: 345–350. Somogyi, P. and Klausberger, T. (2005) Defined types of cortical interneurone structure space and spike timing in the hippocampus. J. Physiol., 562: 9–26.

Soriano, E. and Frotscher, M. (1989) A GABAergic axo-axonic cell in the fascia dentata controls the main excitatory hippocampal pathway. Brain Res., 503: 170–174. Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic chandelier cells in the rat fascia dentata: Golgi-electron microscopy and immunocytochemical studies. J. Comp. Neurol., 293: 1–25. Sperk, G., Marksteiner, J., Gruber, B., Bellmann, R., Mahata, M. and Ortler, M. (1992) Functional changes in neuropeptide Y- and somatostatin-containing neurons induced by limbic seizures in the rat. Neuroscience, 50: 831–846. Sperk, G., Schwarzer, C., Tsunashima, K., Fuchs, J. and Sieghart, W. (1997) GABAA receptor subunits in the rat hippocampus I: immunocytochemical distribution of 13 subunits. Neuroscience, 80: 987–1000. Stumm, R.K., Zhou, C., Schulz, S. and Hollt, V. (2004) Neuronal types expressing mu- and delta-opioid receptor mRNA in the rat hippocampal formation. J. Comp. Neurol., 469: 107–118. Szenta´gothai, J. (1974) Conceptual models of neural organization. Neurosci. Res. Prog. Bull., 12: 307–510. Toth, K. and Freund, T.F. (1992) Calbindin D28k-containing nonpyramidal cells in the rat hippocampus: their immunoreactivity for GABA and projection to the medial septum. Neuroscience, 49: 793–805. Valtschanoff, J.G., Weinberg, R.J., Kharazia, V.N., Nakane, M. and Schmidt, H.H. (1993) Neurons in rat hippocampus that synthesize nitric oxide. J. Comp. Neurol., 331: 111–121. Wittner, L., Maglo´czky, Z., Borhegyi, Z., Hala´sz, P., To´th, S., Ero¨ss, L., Szabo´, Z. and Freund, T.F. (2001) Preservation of perisomatic inhibitory input of granule cells in the epileptic human dentate gyrus. Neuroscience, 108: 587–600. Woodson, W., Nitecka, L. and Ben-Ari, Y. (1989) Organization of the GABAergic system in the rat hippocampal formation: a quantitative immunocytochemical study. J. Comp. Neurol., 280: 254–271. Zappone, C.A. and Sloviter, R.S. (2001) Commissurally projecting inhibitory interneurons of the rat hippocampal dentate gyrus: a colocalization study of neuronal markers and the retrograde tracer Fluoro-gold. J. Comp. Neurol., 441: 324–344. Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989) Entorhinal fibers form synaptic contacts on parvalbuminimmunoreactive neurons in the rat fascia dentata. Brain Res., 495: 161–166.

SECTION III

Neurotransmitters and Neuromodulators

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 14

Functional regulation of the dentate gyrus by GABA-mediated inhibition Douglas A. Coulter1,2, and Gregory C. Carlson1 1

2

The Children’s Hospital of Philadelphia, Abramson Pediatric Research Center, Room 410D, 3516 Civic Center Boulevard, Philadelphia, PA 19104-4318, USA Departments of Pediatrics, Neurology, and Neuroscience, University of Pennsylvania School of Medicine, Philadelphia, PA, USA

Abstract: Dentate granule cells are characterized by their low levels of excitability, an important aspect of hippocampal function, which distinguishes them from other principal cells of the hippocampus. This low excitability derives in large part from the degree and nature of GABAergic inhibition evident in the dentate gyrus. Granule cells express a unique and complex assortment of GABAA receptor subunits, found in few areas of the brain. Associated with this receptor complexity, granule cells are endowed with both synaptic and extrasynaptic GABAA receptors with distinctive properties. In particular, extrasynaptic GABAA receptors in granule cells exhibit high affinity for GABA and do not desensitize. This results in activation of a tonic current by ambient levels of GABA present in the extracellular space. This tonic current contributes significantly to the circuit properties of the dentate gyrus. Both synaptic and extrasynaptic GABAA receptors exhibit profound dysregulation in animal models of temporal lobe epilepsy, which may contribute to the hippocampal hyperexcitability that defines this disorder. Keywords: hippocampus; dentate gyrus; GABA receptor; inhibition; epilepsy 1976; Wilson and McNaughton, 1993). This ‘sparse coding’ of dentate granule cells is theorized to be important in information processing and memory formation of the hippocampus (McNaughton and Morris, 1987). It also reflects a general trend evident in most studies of dentate granule cell excitability: these cells exhibit a fundamental reluctance to fire, particularly synchronously in network bursts. This is due, in part, to a combination of intrinsic factors (including hyperpolarized resting membrane potential, lack of conductances which permit phasic firing or ‘‘bursting’’, and marked spike frequency adaptation), but is principally a consequence of the powerful feedforward and feedback GABAergic inhibition characteristic of dentate gyrus circuit function.

Introduction As a part of its role in memory formation, the hippocampus encodes a cognitive map of the space in which an animal (or human) navigates. During environmental exploration, studies recording place fields in rat hippocampal neurons have typically demonstrated that information coding of spatial position (i.e. degree of firing and firing patterns) is specific but sparse in dentate granule cells (Jung and McNaughton, 1993; Chawla et al., 2005) compared to dorsal hippocampal pyramidal cells (O’Keefe,

Corresponding author. Tel.: +1 215-590-1937; Fax: +1 215-590-4121; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63014-3

235

236

In addition to its important role in cognitive processing, the low excitability of the dentate gyrus may serve to filter or ‘gate’ synchronous excitatory activity in entorhinal cortex, preventing this type of activity from hyperactivating and damaging the relatively fragile hippocampal structures downstream, and from triggering seizure activity (Heinemann et al., 1992; Lothman et al., 1992). Like sparse coding, this gating function of the dentate gyrus is also due to powerful feedforward and feedback GABAergic inhibition, as well as intrinsic properties of granule cells. Furthermore, the filter function of the dentate gyrus may be compromised in animals with epilepsy, or in animals in the process of developing epilepsy, and this loss of function may reflect alterations in GABAA receptor expression and function in dentate granule cells. The intent of the present chapter is to discuss dentate gyrus circuit excitability in context of the GABA receptors, which granule cells express, and also to extend this discussion to how alterations in expression and function of inhibitory synaptic receptors in epileptic animals may disrupt normal operation of the dentate gyrus.

Heterogeneous composition and function of GABAA receptors in the CNS GABAA receptors are members of the cysteineloop ligand-gated ion channel family, which, like other members of this family, are pentameric assemblies of subunits that surround a central ion selective pore. The properties of the pore in GABA receptors make it primarily permeable to chloride and bicarbonate ions. The net result of activation of GABA receptors is therefore flow of chloride into cells down its concentration gradient, concomitant hyperpolarization of the membrane potential, and a large conductance increase. Both of these events (hyperpolarization and activation of a shunting conductance) usually result in a decreased propensity to fire action potentials in a neuron following GABA receptor activation, i.e. inhibition. To date, 19 GABAA receptor subunits have been cloned from the mammalian CNS, encompassing 8 families (a1–6, b1–3, g1–3, d, e, y, p, and r1–3). Additional complexity is conferred on

possible receptor composition by alternative splicing, which is found in several subunit families (reviewed in Farrant and Nusser, 2005, and in references therein). Combinatorial calculations based on random assembly of pentameric receptors would result in a million or more possible GABAA receptors. However, this complexity is simplified by the fact that there appears to be a preferred receptor stoichiometry (usually two a, two b, and a g subunit as the prevalent receptor, with the g subunit replaced by d, e, or p subunit in rare subsets of receptors). In GABAA receptors isolated from brain, there may be as few as 10–20 preferred receptor subunit configurations (Wisden et al. 1992; McKernan and Whiting, 1996; Pirker et al. 2000). This heterogeneity in subunit composition of GABAA receptors has significant functional consequences. The subunit composition of a GABAA receptor dictates the kinetic properties, cellular and subcellular localization, and pharmacology of the receptor when expressed in neurons.

GABAA receptors expressed by dentate granule cells In situ hybridization (Wisden et al., 1992), immunocytochemical (Sperk et al., 1997; Pirker et al., 2000), and single-cell combined antisense RNA amplification profiling-functional studies (Brooks-Kayal et al., 1998) have demonstrated that granule cells express as many as 10 or more GABAA receptor subunits, even when only a single cell is measured (Brooks-Kayal et al., 1998). Of 19 possible subunits, technological and experimental issues have limited profiling to 13 subunits or less in most cases. In these studies, granule cells express five of six a subunits (with no expression of a6, and a3 found in some studies, and not in others), three of three b subunits, one of three g subunits (only g2), as well as d subunits (Wisden et al., 1992; Sperk et al., 1997; Brooks-Kayal et al., 1998; Pirker et al., 2000). Pharmacological studies examining the actions of subunit selective agonists and antagonists on synaptic responses in dentate granule cells have provided some insight into which of the above described set of 10 subunits may comprise

237

subsynaptic GABAA receptors. In normal, control rodents, zolpidem (a hypnotic that preferentially augments responses in a1-containing GABAA receptors markedly enhances inhibitory synaptic responses in dentate granule cells, while zinc has little or no antagonist effect, even at very high concentrations (Buhl et al., 1996; Cohen et al., 2003). This suggests that, in addition to g2 subunits required to anchor GABAA receptors in synapses (Essrich et al., 1998) and an undefined set of b subunits, a1 subunits are highly expressed in synaptic GABAA receptors in dentate granule cells, since GABAA receptors containing these subunits respond preferentially to zolpidem, and are very resistant to blockade by zinc. Other a subunits, such as a2 and a4 may also make a contribution to synaptic responses. The degree of contribution of these other subunits has not as yet been defined experimentally in granule cells. In addition, this ensemble of synaptic GABAA receptors changes significantly in animals with epilepsy, which will be discussed in detail below.

Tonic and phasic inhibition in dentate granule cells In addition to synaptic GABAA receptors, which demonstrate a significant contribution of a1 subunits, dentate granule cells express a relatively unique set of GABAA receptors comprised of a4bd subunits. Receptors with this composition are only found to any significant extent in one other region of the brain, the thalamus, where they are expressed in thalamocortical relay neurons (McKernan and Whiting, 1996). These two brain regions, the dentate gyrus and the thalamus, share a functional requirement to fire action potentials in discrete, extremely small populations of neurons, and to resist synchronous firing during information transfer. These requirements are evident as ‘sparse coding’ properties discussed above for the dentate gyrus, and as somatosensory and visual representations evident in thalamic projections to cortex, where synchronous firing would disrupt information transfer. Given the shared circuit requirements in these two brain regions, and the selective expression of a4bd GABA receptors in only these regions, this suggests that receptors of this

composition may play a critical role in discrete coding of information, and in opposing synchronous firing. What can the properties of a4bd receptors tell us about how they may accomplish this role in defining circuit function in the dentate gyrus? Receptors lacking the g2 subunit are not clustered in synapses (Essrich et al., 1998), and so a4bd receptors are primarily localized extra- or perisynaptically (Wei et al., 2003; Sun et al., 2004). Receptors of this composition are endowed with a set of pharmacological properties which make them perfectly suited to constitute extrasynaptic receptors: they have extremely high affinity for GABA (in the nM compared to the low mM range for synaptic receptors), and they exhibit little or no desensitization (Stell and Mody, 2002; Mtchedlishvili and Kapur, 2006; reviewed in Farrant and Nusser, 2005). Therefore, these receptors can respond to ambient or synaptic spillover levels of GABA, and can maintain a tonic current for sustained periods without desensitizing, both of which define tonic GABAA receptors (Farrant and Nusser, 2005). Gene targeting studies in mice support this hypothesis, because it was found that deletion of either the a4 or d subunit significantly reduced tonic current as well as accelerated decay of IPSCs in dentate granule cells, both of which were accompanied by alterations in neuronal excitability (Spiegelman et al., 2003; Wei et al., 2003; Chandra et al., 2006). Therefore, tonic and spillovermediated GABAA currents derived from activity of a4bd-containing receptors are a significant and prevalent feature of dentate granule cell neurophysiological characteristics, and are important contributors to the function of the dentate gyrus.

The ‘gatekeeper’ function of the dentate gyrus is maintained by GABAergic inhibition Assessing the role of GABAergic inhibition in regulating function of the dentate gyrus requires simultaneous recording in afferents to the dentate gyrus, the dentate gyrus itself, and efferents of the dentate gyrus during synaptic activation. This is necessary to ascertain how the dentate gyrus

238

circuitry may filter and constrain the amplitude and duration of afferent inputs as it passes information on to area CA3 of the hippocampus. In addition, these recordings need to be conducted under conditions where the extracellular milieu can be easily manipulated. Several studies have been conducted examining filtering or ‘gatekeeper’ function of the dentate gyrus under control and pathophysiological conditions (animal models of epilepsy). Behr et al. (1998), using multiple field potential recordings in entorhinal cortex, dentate gyrus, and area CA3, found that, in control animals, epileptiform bursts originating in entorhinal cortex did not propagate through the dentate gyrus to activate CA3 with a similar burst. Rather, the input was filtered, and only moderate activity was recorded in CA3. This contrasted with kindled animals, where the entorhinal cortical epileptiform burst activity was able to propagate through the dentate gyrus and trigger similar epileptiform bursts in area CA3. No attempt was made in the Behr et al. (1998) study to determine what aspect of dentate gyrus circuit physiology was responsible for the filtering function in control animals, nor what aspect of circuit function was perturbed in kindled animals. We (Carlson et al., 2002a; Ang et al., 2006) have utilized voltage sensitive dye imaging techniques to monitor function of the dentate gyrus in control and epileptic animals, as entorhinal cortical afferents to dentate gyrus are activated. Voltage sensitive dyes, when combined with state-of-the-art CCD cameras, allow monitoring of multiple sites within a brain slice (in our case, 6400 sites), at heretofore unprecedented temporal resolution (1–5 kHz frame rates). This allows circuit activity to be monitored with high temporal and spatial resolution, which can facilitate recording of synaptic integration within multiple circuits. This permits monitoring of afferent activation of the dentate gyrus, processing within the dentate, and propagation of dentate signals to efferent structures to be monitored simultaneously, while recording from individual neurons with a patch electrode (Ang et al., 2006; see Fig. 1). Data derived from these studies clearly illustrate both the ‘gatekeeper’ function of the dentate (lack

of hippocampal activation despite robust, synchronous excitation of the dentate gyrus by entorhinal cortical afferents; Fig. 1A–C; see also Ang et al., 2006). In addition, these studies also demonstrate that this dentate filtering is accomplished by activation of GABAA receptors. Even modest disinhibition (by picrotoxin concentrations sufficient to block 20–25% of synaptic responses) results in collapse of the gate of the dentate gyrus (Fig. 1D–F), a finding which was also evident in earlier imaging studies of dentate gyrus function (Iijima et al., 1996). In the Carlson et al. (2002a) studies, picrotoxin was used in modest concentrations to trigger disinhibition. Picrotoxin is a non-competitive GABAA antagonist, and has equivalent efficacy in blocking both lower affinity, synaptic and higher affinity, extrasynaptic GABAA receptors (Stell and Mody, 2002). In contrast, gabazine, a competitive GABAA antagonist, is much more effective in blocking lower affinity, synaptic GABAA receptors than tonic, extrasynaptic receptors, since it can compete more effectively for the GABA binding site when receptors have relatively low affinity for GABA. When applied in nanomolar concentrations, sufficient to block 70% of synaptic GABA responses, but only a few percent of tonic GABAA receptor responses (Stell and Mody, 2002), gabazine had no effect on gatekeeper function of the dentate gyrus (Carlson et al., 2002b; Carlson and Coulter, unpublished observations). This implicates tonic GABAA receptors as critical mediators of dentate gyrus filter function. This concept of a critical role of tonic GABAA receptors as important regulators of dentate gyrus function was further supported by a recent study, which described ovarian cycle-linked reduction in expression of the d subunit of GABAA receptors in dentate granule cells, together with a reduction in tonic current, which were accompanied by an increase in seizure susceptibility (Maguire et al., 2005). Increased seizure susceptibility could be mimicked by experimental downregulation of the d subunit. This demonstration that reduced expression of a single subunit (d) critical for tonic current in dentate granule cells altered dentate circuit function in whole animals further corroborates the

239

Fig. 1. ‘Gatekeeper’ function of the dentate gyrus is maintained by GABAergic inhibition. Simultaneous voltage sensitive dye (A snapshot taken at the peak of the response, B trace illustrating the VSD response over time), patch clamp (C), and field potential (C) recording of dentate gyrus response to perforant path activation in control ACSF. Note robust activation of dentate gyrus molecular layer (red color in A, corresponding to a 10–15 mV EPSP in B), which does not result in activation of downstream structures (note lack of response in area CA3 in A and B). This lack of CA3 activation is because dentate granule cells do not fire action potentials in response to perforant path activation under these conditions. This is evident in both the patch (C, top trace, the neuron depolarized to Vm of 50 mV) and field potential recording (C, bottom trace), due to powerful feedforward inhibition activated by perforant path stimulation (C, note large IPSP in patch recording). The importance of inhibition in mediating this ‘gatekeeper’ function is illustrated in responses in panels D, E, and F, following perfusion with 5 mM picrotoxin, a non-competitive GABAA receptor antagonist. This concentration blocks 20–25% of inhibition (see inset [located above panel E] depicting an averaged spontaneous IPSC [sIPSC] before and after perfusion with 5 mM picrotoxin). During 25% GABAergic blockade, perforant path activation resulted in powerful activation of both the dentate gyrus and downstream structures (CA3 and hilus; D, E). It also triggered action potential firing in dentate granule cells (see patch and field potential recordings in F, both of which exhibit action potential firing). A grayscale image of the slice, with patch and field potential recording electrode location is depicted in the inset above A. From Carlson and Coulter (unpublished). (See Color Plate 14.1 in color plate section.)

240

important role played by these small amplitude, sustained currents in regulating function of the dentate gyrus.

Alterations in GABAA receptor expression in dentate granule cells, and function of the dentate gyrus in animal models of epilepsy Temporal lobe epilepsy is defined by seizures discharges which activate the temporal lobe, including the hippocampus. Because the dentate gyrus is hypothesized to be a critical checkpoint regulating excitability of the limbic system (Heinemann et al., 1992; Lothman et al., 1992), it has been a focus of multiple studies in animal models of epilepsy, assessing whether cellular, synaptic, and circuit properties are altered in a manner consistent with seizure susceptibility. A primary focus of this work has been GABAA receptor expression and function, as well as inhibitory synaptic function.

Upregulation of synaptic GABAA receptors in granule cells of epileptic animals If levels of GABAergic inhibition are critical in mediating gatekeeper function of the dentate gyrus, and compromised filter function of the dentate is a primary contributor to seizure generation in animals with temporal lobe epilepsy, then one might expect that overall levels of expression of GABAA receptors might be reduced in epileptic animals. In both kindling and post-status epilepticus models of temporal lobe epilepsy, studies examining GABAA receptor function in dentate granule cells have described a paradoxical upregulation in expression of GABAA receptors, both synaptically (Otis et al., 1994; Buhl et al., 1996; Nusser et al., 1998; Cohen et al., 2003), and in whole cell studies (Gibbs et al., 1997; Brooks-Kayal et al., 1998; Mtchedlishvili et al., 2001; Leroy et al., 2004). There is a consistent finding of a virtual doubling in the amplitude of quantal inhibitory synaptic responses, accompanied by a doubling in density of GABAA receptors recorded in whole-cell GABA responses, evident across multiple, divergent models of temporal lobe epilepsy.

Alterations in pharmacology and subunit expression of dentate granule cell GABAA receptors in epileptic animals This upregulation of GABAA receptors appears to be inconsistent with the hypothesis that compromised inhibition in the dentate gyrus might contribute to seizure generation in epileptic animals. However, experimental evidence provides additional clues as to how alterations in inhibition may contribute to seizure susceptibility. Not only is there an upregulation in density of GABAA receptors (discussed above), but the nature (subunit composition) of the receptors themselves is altered. This is evident as large-scale changes in the pharmacological properties of the receptors, and changes in the expression patterns of subunits encoding GABAA receptors in dentate granule cells. Synaptic GABAA receptors exhibit the de novo appearance of sensitivity to zinc blockade (Buhl et al., 1996; Gibbs et al., 1997; Brooks-Kayal et al., 1998; Cohen et al., 2003), as well as loss of sensitivity to benzodiazepine site agonists (Gibbs et al., 1997; Mtchedlishvili et al., 2001; Leroy et al., 2004). Both of these findings (elevated zinc sensitivity, reduced benzodiazepine sensitivity) suggest that dentate granule cell GABAA receptors exhibit altered subunit composition following the development of epilepsy. This has been demonstrated in individual dentate granule cells with this pharmacological phenotype, with a net downregulation in expression of a1 subunits, and an upregulation in expression of a4 subunits (Brooks-Kayal et al., 1998). This a subunit switch, if manifest as an upregulation in a4 subunits in synapses, could explain both the enhanced sensitivity to zinc blockade and diminished sensitivity to benzodiazepine site agonists, particularly the a1-preferring GABA receptor modulator, zolpidem.

Possible consequences of epilepsy-associated altered subunit composition of GABAA receptors: zincinduced collapse of augmented inhibition This altered zinc sensitivity of GABAA receptors in animals with epilepsy may have functional significance. In addition to alterations in inhibitory

241

synaptic responses, the dentate gyrus of animals with epilepsy frequently demonstrates aberrant sprouting of the output axons of granule cells, the mossy fibers. These axons, perhaps in response to loss of targets in the hilus and area CA3, sprout and reinnervate the proximal dendritic tree of other dentate granule cells, creating a reentrant excitatory circuit. In addition to releasing glutamate, these mossy fibers release large quantities of zinc (reviewed in Coulter, 2000). Therefore, in the epileptic brain, there is an apposition of the de novo appearance of a zinc delivery system, sprouted mossy fibers, as well as the de novo appearance of zinc-sensitive synaptic GABAA receptors. The simultaneous manifestation of these two alterations in the epileptic dentate gyrus has led to the hypothesis that, during periods of repetitive afferent activation of the dentate gyrus, which triggers zinc release, there may be a zinc-induced collapse of augmented GABAergic inhibition, facilitating seizure initiation (Buhl et al., 1996; Gibbs et al., 1997; Coulter, 2000). Evidence supporting this hypothesis includes the de novo appearance of zinc-sensitivity to ‘gatekeeper’ function of the dentate gyrus evident in voltage-sensitive dye imaging studies (Carlson et al., 2002a). However, all of the above studies demonstrating compromised circuit function and enhanced zinc responses of GABAA receptors have examined the effects of exogenously applied zinc. For the above hypothesis to be operative, zinc released at excitatory synapses needs to be sufficiently mobile to diffuse to and block neighboring GABAergic synapses. In vitro studies examining the effects of endogenously released zinc have as yet failed to identify effects of zinc release on inhibitory synaptic responses in slices prepared from epileptic animals (Molnar and Nadler, 2001).

Reductions in tonic GABAA current in granule cells of epileptic animals A second finding has recently been described in animal models of temporal lobe epilepsy: reduction in tonic GABAA receptor current accompanied by a downregulation in expression of d subunits (Peng et al., 2004). Given that tonic GABAA receptors

appear critical in ‘gatekeeper’ function of the dentate gyrus, this downregulation in d subunit expression could compromise function of the dentate gyrus, despite the concomitant upregulation of synaptic GABAA receptors. However, recent circuit studies have failed to identify compromised ‘gatekeeper’ function of the dentate gyrus in epileptic animals (Ang et al., 2006), under stimulus parameter conditions where blockade of tonic GABAA receptors induced collapse of the dentate gate in normal animals (Carlson et al., 2002b). This suggests that possible dentate gate compromise mediated by downregulation in expression of d subunits (and reduction in tonic GABAA currents) in epileptic animals may require certain patterns of afferent activation to become operative. Perhaps repetitive activation, as during seizure initiation, will elevate extrasynaptic GABA concentrations and enhance tonic current in normal animals (suppressing seizures), and this major check on excitability will be compromised in animals with epilepsy. Conclusions The dentate gyrus is a structure characterized by low cellular excitability of its output neurons, dentate granule cells. This low excitability is important in cognitive function of the hippocampus, and results predominantly from the high degree of inhibitory synaptic regulation, as well as the unique GABAA receptor properties of these cells, including expression of tonic GABAA receptors rarely expressed in other brain regions. Altered expression and function of GABAA receptors in dentate granule cells is a prominent feature of temporal lobe epilepsy, a disorder characterized by hypersynchronous discharges involving the hippocampus. This association of aberrant GABAA receptor properties in granule cells from hyperexcitable hippocampi further demonstrates the critical role of inhibition in controlling dentate gyrus function. References Ang, C.W., Carlson, G.C. and Coulter, D.A. (2006) Massive and specific dysregulation of direct cortical input to the

242 hippocampus in temporal lobe epilepsy. J. Neurosci., 26: 11850–11856. Behr, J., Lyson, K.J. and Mody, I. (1998) Enhanced propagation of epileptiform activity through the kindled dentate gyrus. J. Neurophysiol., 79: 1726–1732. Brooks-Kayal, A.R., Shumate, M.D., Jin, H., Lin, D.D., Rikhter, T.Y. and Coulter, D.A. (1998) Selective changes in single cell GABAA receptor subunit expression and function in temporal lobe epilepsy. Nat. Med., 4: 1166–1172. Buhl, E.H., Otis, T.S. and Mody, I. (1996) Zinc-induced collapse of augmented inhibition by GABA in a temporal lobe epilepsy model. Science, 271: 369–373. Carlson, G., Lee, C.J. and Coulter, D.A. (2002a) Zinc facilitates hyperexcitability in the hippocampus of epileptic rats. Epilepsia, 43(Suppl 7): 2.031. Carlson, G.C., Lee, C.J. and Coulter, D.A. (2002b) The role of inhibition and dentate gatekeeper function: voltage-sensitive dye imaging study. Soc. Neurosci. Abstr., 28: 145.9. Chandra, D., Jia, F., Peng, Z., Suryanarayanan, A., Werner, D.F., Spigelman, I., Houser, C.R., Olsen, R.W., Harrison, N.L. and Homanics, G.E. (2006) GABAA receptor a4 subunits mediate extrasynaptic inhibition in thalamus and dentate gyrus and the action of gaboxadol. PNAS, 103: 15230–15235. Chawla, K.M., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P., Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton, B.L. and Barnes, C.A. (2005) Sparse, environmentally selective expression of Arc RNA in the upper blade of the rodent fascia dentate by brief spatial experience. Hippocampus, 15: 579–586. Cohen, A.S., Lin, D.D., Quirk, G.L. and Coulter, D.A. (2003) Dentate granule cell GABAA receptors in epileptic hippocampus: enhanced synaptic efficacy and altered pharmacology. Eur. J. Neurosci., 17: 1607–1616. Coulter, D.A. (2000) Mossy fiber zinc and temporal lobe epilepsy: pathological association with altered epileptic GABAA receptors in dentate granule cells. Epilepsia, 41(Suppl 6): S96–S99. Essrich, C., Lorez, M., Benson, J.A., Fritschy, J.M. and Luscher, B. (1998) Postsynaptic clustering of major GABAA receptor subtypes requires the gamma 2 subunit and gephyrin. Nat. Neurosci., 1: 563–571. Farrant, M. and Nusser, Z. (2005) Variations on an inhibitory theme: phasic and tonic activation of GABAA receptors. Nat. Rev. Neurosci., 6: 215–229. Gibbs III, J.W., Shumate, M.D. and Coulter, D.A. (1997) Differential epilepsy-associated alterations in postsynaptic GABAA receptor function in dentate granule and CA1 neurons. J. Neurophysiol., 77: 1924–1938. Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and Zhang, C.L. (1992) The dentate gyrus as a regulated gate for the propagation of epileptiform activity. Epilepsy Res. Suppl., 7: 273–280. Iijima, T., Witter, M.P., Ichikawa, M., Tominage, T., Kajiwara, R. and Matsumoto, G. (1996) Entorhinal-hippocampal interactions revealed by real-time imaging. Science, 272: 1176–1179.

Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of unit activity in the hippocampal granular cell layer. Hippocampus, 3: 165–182. O’Keefe, J. (1976) Place units in the hippocampus of the freely moving rat. Exp. Neurol., 51: 78–109. Leroy, C., Poisbeau, P., Keller, A.F. and Nehlig, A. (2004) Pharmacological plasticity of GABAA receptors at dentate gyrus synapses in a rat model of temporal lobe epilepsy. J. Physiol., 557: 473–487. Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The dentate gyrus as a control point for seizures in the hippocampus and beyond. Epilepsy Res. Suppl., 7: 301–313. Maguire, J.L., Stell, B.M., Rafizadeh, M. and Mody, I. (2005) Ovarian cycle-linked changes in GABA(A) receptors mediating tonic inhibition alter seizure susceptibility and anxiety. Nat. Neurosci., 8(6): 797–804. McKernan, R.M. and Whiting, P.J. (1996) Which GABAAreceptor subtypes really occur in the brain. Trends Neurosci., 19: 139–143. McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal synaptic enhancement and information storage within a distributed memory system. Trends Neurosci., 10: 408–415. Molnar, P. and Nadler, J.V. (2001) Lack of effect of mossy fiber-released zinc on granule cell GABAA receptors in the pilocarpine model of epilepsy. J. Neurophysiol., 85: 1932–1940. Mtchedlishvili, Z., Bertram, E.H. and Kapur, J. (2001) Diminished allopregnanolone enhancement of GABAA receptor currents in a rat model of chronic temporal lobe epilepsy. J. Physiol., 537: 453–465. Mtchedlishvili, Z. and Kapur, J. (2006) High-affinity, slowly desensitizing GABAA receptors mediate tonic inhibition in hippocampal dentate granule cells. Mol. Pharmacol., 69: 564–575. Nusser, Z., Hajos, N., Somogyi, P. and Mody, I. (1998) Increased number of synaptic GABAA receptors underlies potentiation at hippocampal inhibitory synapses. Nature, 395: 172–177. Otis, T.S., De Koninck, Y. and Mody, I. (1994) Lasting potentiation of inhibition is associated with an increased number of gamma-aminobutyric acid type A receptors activated during miniature inhibitory postsynaptic currents. PNAS, 91: 7698–7702. Peng, Z., Huang, C.S., Stell, B.S., Mody, I. and Houser, C.R. (2004) Altered expression of the d subunit of the GABAA receptor in a mouse model of temporal lobe epilepsy. J. Neurosci., 24: 8629–8639. Pirker, S., Schwarzer, C., Wieselthaler, A., Sieghart, W. and Sperk, G. (2000) GABAA receptors: immunocytochemical distribution of 13 subunits in the adult rat brain. Neuroscience, 101: 815–850. Sperk, G., Schwarzer, C., Tsunaskima, K., Fuchs, K. and Sieghart, W. (1997) GABA(A) receptor subunits in the rat hippocampus. I: immunocytochemical distribution of 13 subunits. Neuroscience, 80: 987–1000. Spiegelman, I., Li, Z., Liang, J., Cagetti, E., Sanzadeh, S., Mihalek, R.M., Homanics, G.E. and Olsen, R.W. (2003)

243 Reduced inhibition and sensitivity to neurosteroids in hippocampus of mice lacking the GABA(A) receptor delta subunit. J. Neurophysiol., 90: 903–910. Stell, B.M. and Mody, I. (2002) Receptors with different affinities mediate phasic and tonic GABAA conductances in hippocampal neurons. J. Neurosci., 22(1–5): RC223. Sun, C., Sieghart, W. and Kapur, J. (2004) Distribution of alpha1, alpha4, gamma2, and delta subunits of GABAA receptors in hippocampal dentate granule cells. Brain Res., 1029: 207–216.

Wei, W., Zhang, N., Peng, Z., Houser, C.R. and Mody, I. (2003) Perisynaptic localization of d-containing GABAA receptors and their activation by GABA spillover in the mouse dentate gyrus. J. Neurosci., 23: 10650–10661. Wilson, M.A. and McNaughton, B.L. (1993) Dynamics of the hippocampal ensemble code for space. Science, 261(5124): 1055–1058. Wisden, W., Laurie, D.J., Monyer, H. and Seeburg, P.H. (1992) The distribution of 13 GABAA receptor subunit mRNAs in the rat brain. I. Telencephalon, diencephalon, mesencephalon. J. Neurosci., 12: 1040–1062.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 15

Opioid systems in the dentate gyrus Carrie T. Drake1, Charles Chavkin2 and Teresa A. Milner1, 1

Division of Neurobiology, Department of Neurology and Neuroscience, Weill-Cornell Medical College, 411 East 69th Street, New York, NY 10021, USA 2 Department of Pharmacology, University of Washington, Seattle, WA 98195, USA

Abstract: Opiate drugs alter cognitive performance and influence hippocampal excitability, including longterm potentiation (LTP) and seizure activity. The dentate gyrus (DG) contains two major opioid peptides, enkephalins and dynorphins, which have opposing effects on excitability. Enkephalins preferentially bind to delta- and mu-opioid receptors (DORs and MORs) while dynorphins preferentially bind to kappaopioid receptors (KORs). Opioid receptors can also be activated by exogenous opiate drugs such as the MOR agonist morphine. Enkephalins are contained in the mossy fiber pathway, in the lateral perforant path (PP) and in scattered GABAergic interneurons. MORs and DORs are predominantly in distinct subpopulations of GABAergic interneurons known to inhibit granule cells, and are present at low levels within granule cells. MOR and DOR agonists increase excitability and facilitate LTP in the molecular layer. Anatomical and physiological evidence is consistent with somatodendritic and axon terminal targeting of both MORs and DORs. Dynorphins are in the granule cells, most abundantly in mossy fibers but also in dendrites. KORs have been localized to granule cell mossy fibers, supramammillary afferents to granule cells, and PP terminals. KOR agonists, including endogenous dynorphins, diminish the induction of LTP. Recent evidence indicates that opiates and opioids also modulate other processes in the hippocampal formation, including adult neurogenesis, the actions of gonadal hormones, and development of neonatal transmitter systems. Keywords: enkephalin; dynorphin; opioid receptors; GABA interneuron; disinhibition; neurogenesis; seizure; long-term potentiation

Endogenous opioid peptides also modulate hippocampal excitability (reviewed by Simmons and Chavkin, 1996) and the endogenous hippocampal opioid systems are implicated in learning (Sandin et al., 1998), including that associated with drug use (Nestler, 2001). Recent evidence indicates that opiates and opioids also modulate other processes in the hippocampal formation, including adult neurogenesis, the actions of gonadal hormones, and development of neonatal transmitter systems (Eisch et al., 2000; Slamberova´ et al., 2003; Schindler et al., 2004). The dentate gyrus (DG)

Introduction Application of opiate drugs alters cognitive performance and influences hippocampal excitability, including long-term potentiation (LTP) (Guerra et al., 1987; Bramham, 1992; Wagner and Chavkin, 1992; Pu et al., 2002; Robbins and Everitt, 2002) and seizure activity (Hong et al., 1993).

Corresponding author. Tel.: +1 (212) 570 2900; Fax: +1 (212) 988 3672; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63015-5

245

246

contains several types of opioid peptides, which have varying degrees of receptor selectivity, and opioid receptors, which can be activated by both endogenous opioid peptides and exogenous opiate drugs. Most studies of opioid systems have been conducted on rat, and although other species studied are generally similar, there are some notable species differences.

Opioid peptides The enkephalins and dynorphins are the most abundant opioids in the DG. They are derived from two genes with distinct but overlapping cellular distributions. Proenkephalin-derived opioids include [Met5]-enkephalin, [Leu5]-enkephalin, and other C-terminally extended forms of [Met5]enkephalin (Gubler et al., 1982; Noda et al., 1982). Prodynorphin-derived opioids include [Leu5]-enkephalin, dynorphin A(1–17), dynorphin A(1–8), dynorphin B(1–13), leumorphin [a.k.a. dynorphin B(1–29)], alpha-neo-endorphin, and beta-neo-endorphin (Kakidani et al., 1982). The first five amino acids of dynorphin peptides are identical to the (entire) sequence of [Leu5]enkephalin. At times this has complicated the localization and study of the individual opioids, however, fairly selective agents have been developed. The recently discovered opioid endomorphin is extremely rare in the rat DG (Pierce and Wessendorf, 2000). Another opioid, beta-endorphin (a product of the pro-opiomelanocortin precursor), was reported in the hippocampal formation (Zakarian and Smyth, 1979, 1982) based on radioimmunoassay and immunofluorescence; however later chromatographic work showed that the immunoreactivity did not correspond to authentic beta-endorphin (Chavkin et al., 1985b). Finally, nociceptin/orphanin FQ is a hectadecapeptide similar in structure to dynorphin A (Meunier et al., 1995; Reinscheid et al., 1995). Although it shares structural similarity to opioids, its actions and pharmacology are distinct in that it acts through the naloxone-insensitive ORL1 (a.k.a NOP) receptor (Heinricher, 2003).

Opioid receptors Three major types of opioid receptors have been characterized; MORs, DORs, and KORs (reviewed by Martin, 1983). These receptor types have varying preferences for endogenous opioid peptides and exogenous ligands. MORs have a high affinity for endomorphin and the opiate alkaloid morphine (Martin, 1983; Zadina et al., 1997). MORs and DORs preferentially interact with enkephalins over dynorphins, and DORs have a 10-fold higher affinity than MORs for enkephalins (Corbett et al., 1993). KORs have a high affinity for dynorphins but not enkephalins (Chavkin et al., 1982, 1985b; Corbett et al., 1982) and also bind tightly to the exogenous agent ethylketocyclazocine (Corbett et al., 1993). As described below, there is also evidence that dynorphins, although primarily KOR agonists, have some affinity for MORS and can modulate N-methyl-D-aspartic acid (NMDA) receptors (Chavkin et al., 1985a; Caudle et al., 1994; Caudle and Dubner, 1998). The complexity of opioid receptors goes far beyond three straightforward types. Additional opioid receptor types and subtypes have been proposed based on receptor-binding studies, receptor autoradiography, and behavioral studies. These include mu1 and mu2 (Wolozin and Pasternak, 1981), mu/delta (Rothman et al., 1983, 1987), delta1 and delta2 (Jiang et al., 1991; Mattia et al., 1991), and kappa1, kappa2, and kappa3 (Nock et al., 1988; Zukin et al., 1988; Clark et al., 1989). MOR, DOR, and KOR subtypes may arise from several sources. The MOR, DOR, and KOR have been cloned (Evans et al., 1992; Kieffer et al., 1992; Chen et al., 1993; Minami et al., 1993; Yasuda et al., 1993), and although to date only one gene has been found for each receptor type (Slowe et al., 1999), other genes may exist. Another potential source of subtypes is alternatively spliced opioid receptor mRNA (Zimprich et al., 1995; Pan et al., 1999). In both rat and mouse DG, splice variants derived from the cloned MOR1-gene are found, and have slightly different distributions, implying region-specific mRNA processing (Abbadie et al., 2000). Yet another source of opioid receptor diversity is functional association with other proteins. Like other G protein-coupled receptors,

247

opioid receptors form stable associations with one another (termed ‘‘dimers’’) and with accessory proteins. These interactions may serve to increase pharmacological complexity, to aid in transport or maturation of the receptors, and to regulate signal transduction. Both homodimers (e.g., DOR–DOR) and heterodimers (e.g., KOR–DOR) have been demonstrated (see Bouvier, 2001; Rios et al., 2001). A novel opioid-like receptor, termed ORL1, was identified based on homology to opioid receptors, particularly KORs. This receptor has a high affinity for OFQ/nociceptin, but not opioids (Bunzow et al., 1994). The ORL1 and its endogenous ligand OFQ/nociceptin are present in the DG and modify many of the same functions as do opioids; the history and details of this system are reviewed in detail elsewhere (Mogil and Pasternak, 2001). As we describe in the following review of the anatomy and physiology of opioid systems in the DG, the different opioid systems share at least two features. One, in the hippocampal opioid system there is frequently a mismatch between opioid receptors and endogenous opioid peptides. Mismatches at the regional level (e.g., tens or hundreds of micrometers) raise the question of how much of a given pool of opioid peptides reaches opioid receptors (see McLean et al., 1987) and predict distinct effects of endogenously released opioids compared to exogenous opiates. Mismatches at the ultrastructural level indicate that opioids often are not active at synapses, but instead function through nonsynaptic diffusion to nearby cellular targets (e.g., Drake et al., 1994, 2002). Two, although different opioids have differing effects on excitability, they all share the property of being modulators of fast neurotransmission. Their modulatory influences have regionally selective effects on plasticity and sometimes exert complex or subtle influences on dentate circuits. Furthermore, physiologically significant functional effects may occur in regions with relatively few receptors (Weisskopf et al., 1993).

Anatomical distribution of opioids and opioid receptors The distribution of opioid peptides and their receptors is summarized schematically in Figs. 1

and 2 and discussed in detail below. The present description of opioid peptides is restricted to enkephalins and dynorphins, since the endogenous MOR ligand endomorphin is present only in rare scattered fibers in the rat DG (Pierce and Wessendorf, 2000).

Enkephalins Both [met5]-enkephalin and [leu5]-enkephalin are present in the DG (Hong et al., 1980; Gall et al., 1981). Enkephalins are present in afferents to the DG and in DG neurons in a wide range of species including mice, rats, guinea pigs, hamsters, and humans (Gall, 1988b; Herkenham and McLean, 1988; Rees et al., 1994). Enkephalin is present in a subset of mossy fibers (Gall et al., 1981; Gall, 1988a), and is expressed by
15% of granule cells (Johnston and Morris, 1994). Leu-enkephalin immunoreactivity is present in both types of terminals formed by mossy fibers, that is, small terminals in the hilus and the large complex mossy terminals (Commons and Milner, 1995). Enkephalins also are found in a portion of the lateral perforant path (PP) (Gall et al., 1981; Fredens et al., 1984; Gall, 1988a), particularly the temporal portion of this pathway. At the subcellular level, leu-enkephalin immunoreactivity in the rat is primarily associated with distinctive organelles known as large dense-core vesicles (described in more detail in Opioid Peptide Release) in both the small hilar and large mossy terminals. Additionally, a few interneuron-like cells in the hilus are leu-enkephalin-immunoreactive (-ir) (Gall et al., 1981; Commons and Milner, 1995). Analysis of proenkephalin mRNA expression indicates additional proenkephalin-expressing interneurons in the granule cell layer (GCL) and molecular layer (Stumm et al., 2004). There are some relevant species differences in dentate enkephalins. Compared to the rat, the mouse appears to have sparse enkephalin immunoreactivity in the hilus (Khalid et al., personal observation). In the guinea pig there is a prominent bulge of enkephalin immunoreactivity in stratum lucidum (SLu) at the CA3/CA2 border. Enkephalin-ir fibers can been seen to course along

248

SO

CA1 p

SP

SUB

SR SLM

TAT PP p

SLu

HIL

m g

CA3

DG GCL ML

ENK DYN

Fig. 1. Location of enkephalins and dynorphins in the hippocampal formation. Opioid peptides are found in all regions of the hippocampal formation: the hippocampus proper (fields CA1 and CA3), the dentate gyrus (DG), and the subiculum (SUB). The glutamatergic neurons [pyramidal cells (p), granule cells (g), mossy cells (m)] and their projections are shown. GABAergic interneurons are scattered in all laminae. The perforant path (PP) from entorhinal cortex provides the major excitatory innervation. Subcortical afferents and contralateral hippocampal projections enter via the fimbria/fornix (not shown) adjacent to CA3. Enkephalins (green) are in some granule cells, particularly their mossy fiber terminals in hilus and stratum lucidum (SLu). Mossy fibers extensively innervate hilar mossy cells and CA3 pyramidal cells. Enkephalins are in a portion of the projection from entorhinal cortex; the lateral portion of the perforant path (PP) in the DG and the temporal-ammonic tract (TAT) in CA1. Enkephalins are in a few scattered interneurons, which are relatively abundant on the border of stratum radiatum (SR) and stratum lacunosum-moleculare (SLM). Dynorphins (yellow) are in the granule cells. Dynorphin peptides are most striking in the mossy fiber terminals, and are also in the mossy fiber axons, granule cell somata, and granule cell dendrites. Abbreviations: SO, stratum oriens; SP, stratum pyramidale; GCL, granule cell layer. (See Color Plate 15.1 in color plate section.)

the longitudinal axis of the hippocampal formation at this point and are almost certainly in the thick end-bulb formed by the septotemporal projections of mossy fibers (Tielen et al., 1982; McLean et al., 1987). In the rat DG, the enkephalin-containing PP fibers are restricted to the outer molecular layer, that is, the terminal zone of the lateral PP (Gall et al., 1981). In contrast, two bands of enkephalin immunoreactivity are seen in molecular layer of guinea pig, one in the lateral PP terminal zone and the other in the medial PP terminal zone (Tielen et al., 1982). Additionally, while the rat contains dense enkephalin immunoreactivity in the PP terminal zone, in the guinea pig this immunoreactivity is considerably lighter (McLean et al., 1987). Consistent with

studies described below, this observation indicates that prodynorphin-derived peptides play a larger role in the DG of the guinea pig than the rat. Anatomical data are consistent with the suggestion that enkephalins nonsynaptically modulate inhibitory transmission. In the dentate molecular layer, leu-enkephalin-ir terminals occasionally appose GABA-ir terminals, and in the hilus leuenkephalin-ir terminals often contact gamma amino butyric acid (GABA)-containing perikarya and dendrites (Commons and Milner, 1996b). These data suggest that leu-enkephalin may have an important role in regulating inhibition of GABA-containing neurons. One interesting aspect of enkephalin biology is that the production may be regulated by highly

249 + ENK

Molecular Layer

_

PP

GABA

? ?

+ ACh

SubP Granule Cell Layer

Granule Cell

?

GABA PARV

Hilus

+ ENK/DYN

+ ACh

GABA NPY SOM

MOR

to CA3

+

DOR KOR

Fig. 2. Subcellular locations of opioid peptides and receptors in the dentate gyrus. Enkephalins (green shading) are in: (1) somatodendritic, axon, and terminal portions of granule cells; (2) lateral perforant path axons and terminals in the outer molecular layer; and (3) occasional GABAergic interneurons (not shown). Mu-opioid receptors (MORs; blue) are most frequently in all portions of parvalbumin (PARV)-containing basket cells that innervate granule cell somata and proximal dendrites. MORs are also in cholinergic and GABAergic afferents from the lateral septum/diagonal band. Delta-opioid receptors (DORs; pink) are in many hilar interneurons that contain somatostatin or neuropeptide Y (SOM/NPY) and inhibit the distal dendrites of granule cells. PARV-containing neurons express DOR mRNA. DORs, and to a lesser extent MORs, are occasionally present in granule cell dendrites. Kappa-opioid receptors (KORs) in guinea pigs are in substance P containing afferents to granule cells, in granule cell mossy fibers, and most likely on perforant path terminals in the outer molecular layer. (See Color Plate 15.2 in color plate section.)

region-specific mechanisms. For example, the biosynthesis and posttranslational proteolytic processing of proenkephalin in the mossy fiber system is distinct from other brain regions, suggesting that particular functions of the enkephalins may be elicited depending on which met-enkephalin-containing peptides are produced in a given region or subregion (White et al., 1986).

Dynorphin Dynorphin has a similar distribution in rats, mice, guinea pigs, squirrels, hamsters, and humans (Chavkin et al., 1985b; McLean et al., 1987; Gall, 1988a; Houser et al., 1990). The vast majority of dynorphin in the hippocampal formation is made by the dentate granule cells, and dynorphin

250

immunoreactivity is abundant in the mossy fiber pathway in the hilus and CA3 SLu (McGinty et al., 1983; Khachaturian et al., 1993). In the hilus, both the large mossy terminals and the smaller mossy fiber collateral terminals contain dynorphin immunoreactivity (Pierce et al., 1999), suggesting nonselective release near all types of synaptic targets of these terminals, that is, CA3 pyramidal cells, mossy cells, and interneurons. Released dynorphin is more stable than enkephalin (Wagner et al., 1991) and diffusion may play a large role in the actions of dynorphin. Some dynorphin may be contributed by afferents. For example, in situ hybridization histochemistry in rat identified preprodynorphin mRNA-containing perikarya in hippocampally projecting regions of the septum (Merchenthaler et al., 1997). Also, at the extreme ventral pole of the guinea pig hippocampal formation, a diffuse band of varicose dynorphin-ir processes is found in the molecular layer (Drake et al., 1994). The granule cell dendrites may also be a relevant source of dynorphin. As discussed in more detail below, dynorphin, like enkephalins and other peptides, is stored in dense-core vesicles that are readily visible by electron microscopy (Drake et al., 1994; Pierce et al., 1999). In the molecular layer of the guinea pig, midway along the septotemporal axis, electron microscopic analysis showed that although some dynorphin-immunoreactivity is present in unmyelinated axons, the majority (74%) of dynorphin-labeled dense-core vesicles are found in granule cell dendrites (Drake et al., 1994). Dendritic dynorphin has also been observed in samples from epileptic human DG (Zhang and Houser, 1999). The functional relevance of dynorphin release in guinea pig outer molecular layer was suggested by physiological data showing that the time course of released dynorphin actions was more consistent with local release than with release from hilar mossy fibers (Drake et al., 1994). Release of dendritic dynorphin in the molecular layer involves calcium channel types and mechanisms distinct from axonal dynorphin release (Simmons et al., 1995). There is an interesting compartmentalization of dynorphin within guinea pig granule cell dendrites. The dendritic spines contain more dynorphin

immunoreactivity than do dendritic shafts, and dynorphin-containing dense-core vesicles in spines are usually near the extrasynaptic plasma membrane and seldom near synapses (Drake et al., 1994). In contrast, dynorphin-containing densecore vesicles in dendritic shafts are often in the center of the profile, appearing to be in transit. This localization may be relevant to synaptic plasticity. In the outer two-thirds of the molecular layer, spines receive virtually all of the excitatory input to granule cells. Individual spines in some ways act as discrete functional units, with different calcium dynamics and chemical events than the adjacent dendritic shaft (reviewed by Yuste and Tank, 1996). Such compartmentalization is thought to be critical for synapse-specific plasticity (e.g., LTP) (Yuste and Tank, 1996). Only 26% of dynorphin-labeled dense-core vesicles in the guinea pig molecular layer are in axons or axon terminals; of these, most are in small terminals that may have originated in entorhinal cortex or are rare collaterals of granule cells. Consistent with the latter possibility, the inner molecular layer (which is almost completely devoid of entorhinal afferents) contains the largest proportion of axonal dense-core vesicles (Drake et al., 1994). In the inner molecular layer, dynorphin-containing terminals form asymmetric synapses on granule cell dendrites. These may be mossy fiber collaterals, consistent with physiological evidence for direct granule cell-to-granule cell connections (Terman et al., 2000). Interestingly, at the extreme ventral pole of the guinea pig hippocampal formation, the number of axonal dynorphin-labeled dense-core vesicles was higher than at other levels, with no change in the number of dendritic dense-core vesicles (Drake et al., 1994), suggesting that some innervation differences are present. The most likely possibility seems to be the mossy fiber collaterals, which as described above are more common at the ventral pole. This dorsal–ventral difference in dynorphin distribution has not been further studied, but is intriguing in light of other functional and structural differences in dynorphin systems of the dorsal compared to the ventral hippocampal formation (McDaniel et al., 1990; Pierce et al., 1999).

251

Opioid peptide release Like many other neuropeptides, dynorphin and enkephalins are stored in large dense-core vesicles. These are larger (80–120 nm diameter) than small synaptic vesicles, the storage vesicles for fast transmitters such as GABA or glutamate. Dense-core vesicles are synthesized, transported, and released differently from small synaptic vesicles (reviewed in De Camilli and Jahn, 1990). In particular, dense-core vesicles require a more prolonged, intense stimulation of the neuron to fuse with the membrane and release their contents (Seward et al., 1995). This is linked to differential calcium coupling: dense-core vesicle release is facilitated by a small, widespread elevation of calcium while small synaptic vesicle release requires robust, localized calcium elevation (Verhage et al., 1991). Consistent with this, endogenous opioid release in the hippocampal formation requires specific forms of highintensity stimulation (Wagner et al., 1990, 1991, 1993). Dense-core vesicles often fuse with, and release their contents from, portions of the membrane that are distant from classical synaptic specializations (Thureson-Klein and Klein, 1990). Such nonsynaptic or extrasynaptic sites of release can be the preferential location of dense-core vesicle fusion, as in Aplysia californica (Karhunen et al., 2001). In rat and guinea pig mossy fiber terminals, dense-core vesicles containing enkephalin- or dynorphin-immunoreactivities are distributed along both extrasynaptic and synaptic portions of the membrane (Drake et al., 1994; Commons and Milner, 1995; Pierce et al., 1999). The majority of these dense-core vesicles are extrasynaptic, and changes in the distribution of membrane-associated dense-core vesicles following seizure suggest that dynorphin-containing dense-core vesicles are specifically targeted to particular portions of the terminals (Pierce et al., 1999).

Opioid receptors Delta-opioid receptors (DORs) Functional DOR binding, as indicated by receptor autoradiography, is diffusely distributed throughout

the DG of mice and rats (Goodman et al., 1980; Crain et al., 1986; Gulya et al., 1986; Mansour et al., 1987; McLean et al., 1987; Tempel and Zukin, 1987; Mansour et al., 1988). Immunocytochemical labeling of DORs in the rat (Mansour et al., 1993) and mouse (Bausch et al., 1995a) revealed that DOR-ir neurons and processes are scattered in the DG, in a distribution that overlaps with, but is not restricted to, regions known to contain enkephalin. Interestingly, neurons with intense DOR immunoreactivity (Commons and Milner, 1996a) or high levels of DOR mRNA (Stumm et al., 2004) are found in the central hilus, while neurons with light DOR-labeling are in the GCL. Close examination with electron microscopy (Commons and Milner, 1996a) further identified DOR immunoreactivity in dendritic spines of granule cells, as well as in perikarya and dendrites of granule cells and nongranule cells. In DOR-ir dendrites, DOR labeling is most often affiliated with the plasmalemmal surface near excitatory-type synapses, consistent with a role in modulating excitatory neurotransmission. Additionally, presynaptic modulation by DORs is suggested by DOR immunoreactivity in axon terminals of both mouse (Bausch et al., 1995a) and rat (Commons and Milner, 1996a). DOR-labeled terminals exhibit a range of excitatory and inhibitory-type morphologies and are in all dentate lamina, consistent with a variety of neuronal sources (Commons and Milner, 1996a). However, DOR immunoreactivity is found on the plasmalemmal surface in axon terminals less frequently than in dendrites. Most DOR-labeled perikarya contain GABA (Bausch et al., 1995a; Commons and Milner, 1996a; Stumm et al., 2004). DOR has been identified within several interneuron subtypes, although variations in prevalence have been reported depending on the technique used. For example, with dual-labeling immunocytochemistry 75% of DORir neurons sampled in the hilus contained neuropeptide Y (NPY) immunoreactivity and 39% contained somatostatin (SOM) immunoreactivity (Commons and Milner, 1996a). Conversely, 75% of NPY neurons and 93% of SOM neurons contained DOR immunoreactivity (Commons and Milner, 1996a). This suggests that almost all of the somatostatinergic cells in the hilus, which are a subpopulation of NPY neurons, contain DOR. In

252

this study, few DOR-ir neurons were found in the infragranular zone (home to many of the parvalbumin (PARV)-containing basket cells). On the other hand, using a combination of immunocytochemistry and in situ hybridization, DOR mRNA was observed in more than 90% of the PARVpositive neurons sampled in the GCL and hilus, in
11% of hilar SOM-ir neurons, and in none of the calretinin-ir neurons (Stumm et al., 2004). The reason for the discrepancy in prevalence with the earlier study (Commons and Milner, 1996a) is not clear. Possibilities include translation-induced differences (e.g., low production of the DOR protein in PARV-containing neurons or low levels of SOM production in some of the neurons containing SOM mRNA), or perhaps a dimerization of MORs and DORs (George et al., 2000; Gomes et al., 2000) in PARV-containing neurons that leads to lower immunocytochemical recognition of DOR protein. None of the DOR mRNA-containing neurons in the dentate were observed to contain proenkephalin mRNA (Stumm et al., 2004), suggesting little autoregulation of DOR-bearing interneurons. However, since enkephalins and modest DOR levels have both been observed in granule cells, autocrine regulation of mossy fibers or granule cell dendrites remains a possibility. In summary, DORs are primarily in several interneuron populations that inhibit granule cells, are present at low levels in granule cells, and are absent from the calretinin-containing interneurons that inhibit other interneurons. Since the targets of the identified DOR-containing interneuron populations are relatively well understood (for review see (Freund and Buzsa´ki, 1996), a functional link can be postulated. The presence of DORs in SOM/ NPY-containing interneurons suggests a role in the inhibition of granule cell distal dendrites, and in PARV neurons (if the DOR protein is indeed made by these cells) suggests a role in inhibiting granule cell output.

Mu-opioid receptors (MORs) Receptor autoradiography has shown that MOR binding in the DG is more prominent than DOR

binding, although both are similarly distributed in a diffuse pattern (Mansour et al., 1988). By light microscopy, MOR immunoreactivity is present in scattered neurons in the rat and mouse DG (Arvidsson et al., 1995; Bausch et al., 1995b; Mansour et al., 1995; Ding et al., 1996; Bausch and Chavkin, 1997; Kalyuzhny and Wessendorf, 1997; Drake and Milner, 1999). Dual-labeling immunocytochemistry and ultrastructural morphological characterization showed that MORlabeling is in both the somatodendritic and axonal compartments of GABAergic interneurons (Drake and Milner, 1999). Notably, MOR is common in presumed basket cell terminals that form inhibitory-type synapses with granule cell somata and dendrites (Drake and Milner, 1999). Analysis of the neurochemical content of MOR-labeled cells indicates that MOR mRNA (Stumm et al., 2004) and immunoreactivity (Drake and Milner, 2006) are most frequent in PARV-containing interneurons, which comprise one group of basket cells (see Freund and Buzsa´ki, 1996 for review). MOR immunoreactivity (Drake and Milner, 2006) and mRNA (Stumm et al., 2004) are also a modest number of the SOM-containing hilar interneurons (the ‘‘HIPP’’ neurons that project to the outer molecular layer). In the outer molecular layer, MOR-ir axons and terminals form inhibitory-type synapses with dendrites containing NMDA receptor immunoreactivity (Milner and Drake, 2001). These results are consistent with physiological studies described below, and suggest that opioids could inhibit GABAergic terminals that modulate granule cell dendrites, boosting depolarizing events in granule cells and facilitating the activation of NMDA receptors. Finally, MOR mRNA is in a small proportion (10%) of calretinin-containing interneurons in the GCL (Stumm et al., 2004). This is in contrast to DOR mRNA, which is absent from calretinin-ir neurons (Stumm et al., 2004). Together the above morphological and dual-labeling studies suggest that MORs are most frequent in interneurons specialized to inhibit granule cell output (i.e., PARV-containing basket cells), are in a limited number of interneurons that inhibit granule cell distal dendrites, and are occasionally in interneurons that innervate other interneurons.

253

The subcellular location of MOR immunoreactivity is consistent with extrasynaptic activation by diffusing endogenous opioids. In somata and dendrites of GABAergic neurons, MOR immunoreactivity is often along the plasma membrane but rather distant from synapses (Drake and Milner, 1999). Moreover, in the rat hilus, profiles bearing MORs are not contacted by many leu-enkephalinlabeled profiles, but about one-third of MOR-ir axons, terminals, and dendrites are within 3 mm of leu-enkephalin-ir profiles. An estimation of the effective range of one leu-enkephalin-containing vesicle can be made by calculating how far leuenkephalin released from a single dense-core vesicle would diffuse to reach a concentration that would activate half of the available MORs. By knowing the estimated affinity of leu-enkephalin for MORs, the concentration of leu-enkephalin in a single dense-core vesicle, and the extracellular volume fraction of the hilus, we have calculated that the approximate volume occupied by relevant concentration of peptide from one dense-core vesicle is 35–104 fl, with a radius of 2.03–2.92 mm (for more detail see Drake et al., 2002). Thus, many MORs are likely to be within a functionally relevant distance of released enkephalin (Drake et al., 2002). The lack of synaptic associations suggests that activation of MORs by leu-enkephalin requires volume transmission, and furthermore that dendritic, axonal, and terminal MORs are equally likely to be activated by leu-enkephalin released from mossy fibers. There is also a possibility that some MORs are activated in an autocrine fashion by enkephalins. One group observed that 75% of hilar cells with MOR mRNA also label for proenkephalin mRNA (Stumm et al., 2004). However, these neurons may not translate both MORs and enkephalin, as proenkephalin- and MOR- mRNAs are not seen in the same transmitter-identified hilar populations (Stumm et al., 2004) and an electron microsopic study of the hilus showed little colocalization of immunoreactivities for MOR and leuenkephalin in dendrite or axon terminal profiles (Drake et al., 2002). This suggests that, at least in the hilus, autoregulation of neurotransmitter release by MORs may be minor. Rather, enkephalins released from the mossy fibers or other interneurons appear more likely to activate hilar MORs.

MORs are in a few other interesting locations in the DG. First, MOR immunoreactivity is seen in occasional dendrites of granule cells (Drake and Milner, 1999). This location is consistent with other evidence that MOR agonists can act directly on granule cells (Piguet and North, 1993; Stumm et al., 2004), and contrasts with the hippocampus proper, where MORs are absent from principal cells (Drake and Milner, 1999). Second, MORs are on some afferents to the DG. MOR labeling is in cholinergic septal terminals in the hilus (Kaplan et al., 2004) and in a subset of GABAergic septal afferents (Alreja et al., 2000). MOR-ir terminals also contact the same hilar neuronal targets as do cholinergic terminals (Kaplan et al., 2004). These findings suggest the existence of numerous interactions between opioid systems and septal projections in the DG.

Kappa-opioid receptors (KORs) The rat DG contains little KOR binding and virtually no KOR immunoreactivity (Mansour et al., 1988, 1996; Unterwald et al., 1991). [However, see Maderspach et al., 1991, 1995)] Considerably more KOR binding is found in the guinea pig DG (McLean et al., 1987; Wagner et al., 1991). In guinea pig, KOR binding is most intense in the dentate outer molecular layer, with a thin band in the inner molecular layer (McLean et al., 1987). The outer molecular layer KORs disappear following entorhinal cortex lesions, consistent with a presynaptic localization in PP afferents (Simmons et al., 1994). At least two subtypes of pharmacologically identified KORs (K1 and K2) have been identified in the guinea pig DG (Zukin et al., 1988; Unterwald et al., 1991). Two groups, using different carefully characterized antibodies, have localized KOR immunoreactivity in the guinea pig brain, and found a pattern that largely overlaps with KOR binding but has some notable differences (Arvidsson et al., 1994; Drake et al., 1996). The light microscopic distribution of KOR immunoreactivity overlaps the distribution of KOR binding in the inner molecular layer (Drake et al., 1996), but is much less intense in the outer molecular layer. This initially surprising result

254

suggests that in these latter regions KOR-binding may reflect receptors that are less accessible to the antibodies (due to alternative conformations or associations with other proteins), or are different subtypes. KOR immunoreactivity in the inner molecular layer is near (within 100 mm) large stores of dynorphin, but these immunoreactivities do not overlap, suggesting diffusion must occur if the KORs are to be activated (Drake et al., 1996). The reliance on diffusion is consistent with other strong evidence that released dynorphin can diffuse a sufficient distance to activate dentate KORs (Wagner et al., 1991, 1993; Drake et al., 1994). KOR immunoreactivity in the granule cell and inner molecular layers is restricted to unmyelinated axons and axon terminals. Many KOR immunolabeled terminals colocalize substance P (SubP) and form asymmetric synapses with granule cell perikarya and large unlabeled dendrites concentrated in the dentate inner molecular layer (Drake et al., 1997). Following fornix lesions, KOR- and SubP-ir processes dramatically decrease in the molecular layer. These data suggest a role for KORs in presynaptically modulating supramammillary afferents to granule cells, thus helping to regulate information flow into the DG. KORs in the DG have an interesting subcellular location. There is an association with the inner surfaces of plasma membranes (Drake et al., 1996), as expected for the cytoplasmic portion of functional cell-surface receptors. KOR immunoreactivity is also found in large dense-core vesicles (Drake et al., 1996). This localization suggests the possibility that KORs are incorporated into the presynaptic membrane when the dense-core vesicle’s contents are released. Since strong stimulation is required to release dense-core vesicles (Thureson-Klein and Klein, 1990), the implication is that these KORs may be a reserve pool of receptors that undergo excitation-dependent presentation to the plasma membrane. Such a mechanism is of obvious relevance for a role in LTP or in response to hyperstimulation. A presence in dense-core vesicles may be a common property of KORs, since KORs have been localized to dense-core vesicles in other brain regions (Shuster et al., 1999; Svingos et al., 1999). Moreover, in

hypothalamus, translocation of KORs to the plasma membrane is observed following the stimulation of peptide release (Shuster et al., 1999), providing direct evidence for activity-dependent presentation of KORs to the membrane.

Opioid physiology Mu- and delta-opioid actions in the hippocampus are largely disinhibitory (Zieglga¨nsberger et al., 1979). For example, enkephalin was found to hyperpolarize GABAergic interneurons in the rat hippocampus (Madison and Nicoll, 1988), and MOR and DOR activation blocked GABAergic input and induced disinhibitory effects in the rat DG (Neumaier et al., 1988). Whole-cell voltage clamp recordings of granule cells also showed that mu-receptor activation reduced monosynaptic stimulation-evoked inhibitory post-synaptic currents (IPSCs) (Xie et al., 1992). The reduction in GABAergic tone generally increases the excitability of the hippocampus in ways that affect seizure susceptibility and synaptic plasticity [i.e., LTP and long-term depression (LTD)] (Cohen et al., 1992; Morris and Johnston, 1995). LTP induction by high-frequency stimulation of the PP input to the DG in rat hippocampal slices was facilitated by MOR and DOR activation (Bramham and Sarvey, 1996) and this was found to be caused by disinhibition. The inhibition of GABAergic neurons by MORs and DORs is due to membrane hyperpolarization, and the ionic basis has been characterized by several groups using voltage clamp electrophysiology. Mu opioids activate both inward-rectifying and voltage-gated potassium channels in hippocampal interneurons (Wimpey and Chavkin, 1991). Opioids can also increase the M-type potassium channel conductance in hippocampus (Moore et al., 1994). G protein-coupled inwardly rectifying potassium channels (GIRK or Kir3) can be activated by the MOR agonist DAMGO in slices from mice (Luscher et al., 1997). These multiple hyperpolarizing mechanisms reduce inhibitory input within the DG and indirectly increase granule cell excitability.

255

Kappa opioids can directly affect presynaptic release of excitatory amino acids in the guinea pig DG, leading to inhibition of excitatory transmission. (However, in rat, disinhibition resembling MOR and DOR activation has been reported (Neumaier et al., 1988).) Intracellular recordings of dentate granule cells demonstrated that KOR activation significantly reduced the amplitude of glutamatergic excitatory post-synaptic potentials (EPSPs) while having no direct effects on membrane conductance (Wagner et al., 1992). Providing further evidence for presynaptic inhibition, a KOR-selective agonist inhibited excitatory responses to PP stimulation but not to applied glutamate, and potentiated paired-pulse facilitation (Simmons et al., 1994). Kappa receptor activation blocked excitatory amino acid release from PP and recurrent hilar input to granule cells in the DG, and this inhibition reduced LTP induction of these two pathways (Terman et al., 1994, 2000). Much of the evidence for KOR agonist inhibition of LTP in the DG (and in region CA3) has been reviewed by (Morris and Johnston, 1995). Opioids cause presynaptic inhibition of certain subcortical afferents to the DG. Neurochemical measures of transmitter release showed that the stimulated release of serotonin (Yoshioka et al., 1993), norepinephrine (Jackisch et al., 1984; Matsumoto et al., 1994), and acetylcholine (Jackisch et al., 1986) were each directly inhibited by opioid receptor activation. The ionic mechanisms of these inhibitory presynaptic actions have not been directly defined because it is difficult to electrophysiologically record from hippocampal nerve terminals, but molecular mechanisms that have been suggested include activation of delayed rectifying potassium channels (Wimpey and Chavkin, 1991), opioid inhibition of voltage-sensitive calcium channels by Gbg binding (Herlitze et al., 1996; Ikeda, 1996), or direct inhibition of the vesicular fusion machinery by opioid receptor activation (Scholz and Miller, 1992; Capogna et al., 1996).

Opioids and seizures Seizures modulate, and are modulated by, hippocampal opioid systems. The topic of seizures and

dentate opioids has been reviewed elsewhere (Tortella et al., 1988; Simmons and Chavkin, 1996; Simonato and Romualdi, 1996), and will be discussed only briefly here. Levels of enkephalins and dynorphins in the hippocampal formation are altered in humans with temporal lobe epilepsy (Houser et al., 1990; Rees et al., 1994), as well as in rodent seizure models (Hong et al., 1993). Dentate levels of enkephalin and dynorphin are altered in rodents by seizure-inducing treatments, including electroconvulsive shock, amygdala kindling (Vindrola et al., 1981; McGinty et al., 1986), and administration of excitatory chemical agents (Hong et al., 1980, 1988; Pierce et al., 1999; Pierce and Milner, 2001). The time course and direction of peptide changes varies; in general enkephalin mRNA and immunoreactivity increase within hours of seizure while dynorphin decreases rapidly but then recovers (for reviews, see Hong et al., 1993; Pierce et al., 1999). Opioid receptors in the DG also change following seizures, with both MORs and DORs showing a shift in distribution that is in line with anatomical reorganization of particular neuronal circuits, such as mossy fiber sprouting (Bausch and Chavkin, 1997; Skyers et al., 2003). Electrophysiological responses to MOR and DOR agonists also are altered after such seizure-associated circuit reorganization (Bausch and Chavkin, 1997), consistent with a persistent functional change in the dentate opioid systems. Opioid regulation of excitability within the DG and hippocampus also has profound effects on seizure susceptibility. Iontophoretic application of opioids produced epileptiform increases in excitability (French and Siggins, 1980). Morphinepretreated rats show enhanced susceptibility to amydala kindling and morphological changes in the DG (Rocha et al., 1996). Seizure development is thought to involve MOR activation, since MOR agonists, but not DOR agonists, produce convulsions and wet-dog shakes (Lee et al., 1989; Hong et al., 1993). However, in some reports, MOR agonists administered systemically had anticonvulsant actions that may reflect their inhibitory actions on other brain regions (reviewed by Simmons and Chavkin, 1996). KOR activation has been consistently reported to dampen seizure (reviewed by Tortella et al., 1988). For example,

256

KOR activation in rats reduces excitability of the dentate and reduces pilocarpine-induced seizures (Bausch et al., 1998). This anti-convulsant action is consistent with the inhibition of excitatory neurotransmission in the DG and CA3 produced by exogenous and endogenous KOR agonists.

Opioids and gonadal steroids There is ample precedent for interactions between opioids and gonadal steroids in other brain regions (e.g., hypothalamus). Because of this, and because both opioids (for details see Opioid Physiology) and gonadal steroids (for review see Woolley and Schwartzkroin, 1998) modulate hippocampal plasticity, it has been hypothesized that these two systems may interact in the hippocampal formation. A few studies have examined gonadal steroid effects on opioid receptors. Estradiol treatment has been reported to increase MOR binding in hippocampal homogenates (Piva et al., 1995). Since estrogens do not appear to alter levels of hippocampal MOR mRNA (Quinones-Jenab et al., 1997), the increased MOR binding presumably involves another mechanism, such as unmasking, receptor dimerization, or increased translation efficiency. Evidence also suggests that sex hormones may affect levels of dentate opioid peptides. One group (Roman et al., 2006) used radioimmunoassay to determine whether opioid peptide levels changed across the estrous cycle. They found no significant changes in the levels of dynorphins or proenkephalin-derived peptides in homogenates of the hippocampal formation, however, the ratio of dynorphins to their conversion products (including perhaps leu-enkephalin) fluctuated. This suggests that enzymatic processing of dynorphin peptides changes across the estrous cycle, and may affect the availability of dynorphin and its derivatives. Using anatomical methods and focusing on particular subregions, we recently found evidence that estrogen affects the levels of leu-enkephalin and dynorphin immunoreactivity in the DG and the CA3 SLu (Torres-Reveron et al., 2006a, b). Rats in proestrus (high estrogen) have significantly higher levels of leu-enkephalin in the hilus and in

CA3 compared to rats in diestrus. Rats with 24 h (but not 6 or 72 h) of estrogen replacement showed increased levels of leu-enkephalin. Dynorphinimmunoreactivity was elevated only in the apex of the hilus in proestrus rats, but showed significant increases in the entire hilus and CA3 SLu 24 h after estrogen replacement in ovariectomized rats. These data suggest that during the estrogen peaks of the ovarian cycle, a greater amount of enkephalin is available for release from mossy fibers and perhaps enkephalinergic interneurons, suggesting greater excitability in hippocampal circuits. We also found that immunoreactivity for the estrogen receptor ERb was colocalized with enkephalin- or dynorphin-labeling in mossy fiber terminals and smaller terminals. Thus, estrogen may directly influence enkephalin and/or dynorphin release through activation of estrogen receptors on opioidergic neurons.

Opioids and neurogenesis The DG is one of the few brain regions that undergoes neurogenesis throughout adulthood (see Gould and Gross, 2002). New granule cells, and interneurons (Liu et al., 2003), are generated in the subgranular zone of the hilus by progenitor cells (Seri et al., 2001; Turlejski and Djavadian, 2002). As noted in earlier sections, granule cells express low levels of MORs and DORs (Commons and Milner, 1996a; Drake and Milner, 1999), and contain enkephalin (Gall et al., 1981) and dynorphin (Chavkin, 2000). Some hilar interneurons likewise contain opioid receptors and/or enkephalin (Gall et al., 1981; Commons and Milner, 1996a). Neurogenesis has been linked to certain forms of hippocampal-dependent associative learning (Shors et al., 2001, 2002) and is influenced by many factors; it is enhanced by exercise and antidepressant drugs, inhibited by stress, and fluctuates across the female estrous cycle (Tanapat et al., 1999; Eisch et al., 2000; Gould and Gross, 2002). Since cognitive impairment is one of the problems associated with chronic opiate use, and opiates can impair neurogenesis in the prenatal brain, Eisch et al. (2000) examined whether opiates also may affect neurogenesis in adult rats. Although an

257

acute exposure to morphine did not affect neurogenesis, neurogenesis was inhibited by chronic morphine treatment or heroin self-administration. The mechanism requires opioid-receptor activation, as evidenced by sensitivity to the opioid receptor antagonist naltrexone, and does not involve stress hormones (Eisch et al., 2000). Newly born neurons contain MORs (Eisch and Harburg, 2006), consistent with the possibility that morphine may directly act on newly born neurons. These findings suggest that inhibition of neurogenesis is one mechanism by which opiates may exert long-lasting effects on dentate circuits.

Prenatal morphine exposure produces functional consequences for the hippocampal formation. Although rats that are exposed to morphine prenatally did not have a reduced seizure threshold to MOR agonists, they showed an opioid-receptor dependent decrease in sensitivity to bicuculline-induced seizures (Schindler et al., 2004). Similarly, in slices from adult male rats, prenatal morphine exposure was reported to enhance susceptibility to epileptiform activity in the outer molecular layer, but to shift long-term plasticity of PP-granule cell synapses in favor of LTD (Velı´ sek et al., 2000). Clearly the effects of prenatal morphine are complex, and involve both afferents and intrinsic dentate circuits.

Prenatal morphine and opioid system development Summary Prenatal exposure to opiates has many consequences for the developing brain, including alteration of the dentate opioid system and of phenomena known to be influenced by opioids, such as LTP and seizure threshold. In a study of the DG of adult male rats exposed to morphine prenatally, proenkephalin mRNA and met-enkephalin peptides were decreased, while prodynorphin mRNA and dynorphin B peptides were increased (Schindler et al., 2004). Increases in the density of MOR, but not DOR, binding were reported in some strata of hippocampal subfields in morphine-exposed males (Schindler et al., 2004). In an investigation of the interactions between prenatal morphine exposure and adult hormone status (Slamberova´ et al., 2003), MOR density in the DG was reduced in prenatal, morphineexposed males that received gonadal hormone replacement compared to gonadectomized males. Controls exposed to prenatal saline did not show this difference. There was a lower density of MORs in the DG of females after hormone replacement, but this occurred with either prenatal morphine treatment or with prenatal saline exposure, suggesting that interactions between prenatal morphine and adult hormone status were less relevant in the female. Interestingly, the hippocampus proper was affected differently than the DG, with prenatal exposure to morphine altering the density of MORs in adult female but not adult male rats (Slamberova´ et al., 2003).

Opiate drugs alter cognitive performance and influence hippocampal excitability, including LTP and seizure activity. In the DG, the two major opioid peptides, enkephalins, and dynorphins, have opposing effects on excitability. Enkephalins preferentially bind to DORs and MORs while dynorphins preferentially bind to KORs. Opioid receptors can also be activated by exogenous opiate drugs such as morphine. Enkephalins are contained in the mossy fiber pathway, in the lateral PP and in scattered GABAergic interneurons. MORs and DORs are predominantly in distinct subpopulations of GABAergic interneurons known to inhibit granule cells, and are present at low levels within granule cells. MOR and DOR agonists increase excitability and facilitate LTP in the molecular layer. Anatomical and physiological evidence is consistent with somatodendritic and axon terminal targeting of both MORs and DORs. Dynorphins are most abundant in mossy fibers but also are in the granule cells and their dendrites. KORs have been localized to granule cell mossy fibers, supramammillary afferents to granule cells, and PP terminals. KOR agonists, including endogenous dynorphins, diminish the induction of LTP. Opiates and opioids also modulate other processes in the hippocampal formation, including adult neurogenesis, the actions of gonadal hormones, and development of neonatal transmitter systems.

258

Abbreviations DG DOR EPSPs GABA GCL GIRK ir KOR LTD LTP MOR ORL1 PARV PP SLM SLu SO SOM/NPY SP SubP SR SUB TAT

dentate gyrus delta-opioid receptor excitatory post-synaptic potentials gamma amino butyric acid granule cell layer G protein-coupled inwardly rectifying potassium channels immunoreactive kappa-opioid receptor long-term depression long-term potentiation mu-opioid receptor opioid-like receptor parvalbumin perforant path stratum lacunosum-moleculare stratum lucidum stratum oriens somatostatin/neuropeptide Y stratum pyramidale substance P stratum radiatum subiculum temporal-ammonic tract

Acknowledgments We thank Ms. Katherine Mitterling and Mr. Bradley Graustein for technical assistance. Supported by NIH grants: DA 08259 (T.A.M; C.T.D.; C.C.) and HL 18974 (T.A.M.; CTD) and DA04123 (C.C.).

References Abbadie, C., Pan, Y.X., Drake, C.T. and Pasternak, G.W. (2000) Comparative immunohistochemical distributions of carboxy terminus epitopes from the mu-opioid receptor splice variants MOR-1D, MOR-1 and MOR-1C in the mouse and rat CNS. Neuroscience, 100: 141–153. Alreja, M., Shanabrough, M., Liu, W.M. and Leranth, C. (2000) Opioids suppress IPSCs in neurons of the rat medial septum/diagonal band of Broca: Involvement of m-opioid receptors and septohippocampal GABAergic neurons. J. Neurosci., 20: 1179–1189.

Arvidsson, U., Riedl, M., Chakrabarti, S., Lee, J.H., Nakano, A.H., Dado, R.J., Loh, H.H., Law, P.Y., Wessendorf, M.W. and Elde, R. (1995) Distribution and targeting of a muopioid receptor (MOR1) in brain and spinal cord. J. Neurosci., 15: 3328–3341. Arvidsson, U., Risling, M., Frise´n, J., Piehl, F., Fried, K., Ho¨kfelt, T. and Cullheim, S. (1994) trkC-like immunoreactivity in the primate descending serotoninergic system. Eur. J. Neurosci., 6: 230–236. Bausch, S.B. and Chavkin, C. (1997) Changes in hippocampal circuitry after pilocarpine-induced seizures as revealed by opioid receptor distribution and activation. J. Neurosci., 17: 477–492. Bausch, S.B., Esteb, T.M., Terman, G.W. and Chavkin, C. (1998) Administered and endogenously released kappa opioids decrease pilocarpine-induced seizures and seizureinduced histopathology. J. Pharmacol. Exp. Ther., 284: 1147–1155. Bausch, S.B., Patterson, T.A., Appleyard, S.M. and Chavkin, C. (1995a) Immunocytochemical localization of delta opioid receptors in mouse brain. J. Chem. Neuroanat., 8: 175–189. Bausch, S.B., Patterson, T.A., Ehrengruber, M., Lester, H.A., Davidson, N. and Chavkin, C. (1995b) Colocalization of mu opioid receptors with GIRK1 potassium channels in rat brain: an immunocytochemical study. Receptors Channels, 3: 221–241. Bouvier, M. (2001) Oligomerization of G-protein-coupled transmitter receptors. Nat. Rev. Neurosci., 2: 274–286. Bramham, C.R. (1992) Opioid receptor dependent long-term potentiation: peptidergic regulation of synaptic plasticity in the hippocampus. Neurochem. Int., 20: 441–455. Bramham, C.R. and Sarvey, J.M. (1996) Endogenous activation of m and d-1 opioid receptors is required for longterm potentiation induction in the lateral perforant path: Dependence on GABAergic inhibition. J. Neurosci., 16: 8123–8131. Bunzow, J.R., Saez, C., Mortrud, M., Bouvier, C., Williams, J.T., Low, M. and Grandy, D.K. (1994) Molecular cloning and tissue distribution of a putative member of the rat opioid receptor gene family that is not a mu, delta or kappa opioid receptor type. FEBS Lett., 347: 284–288. Capogna, M., Gahwiler, B.H. and Thompson, S.M. (1996) Presynaptic inhibition of calcium-dependent and -independent release elicited with ionomycin, gadolinium, and alphalatrotoxin in the hippocampus. J. Neurophysiol., 75: 2017–2028. Caudle, R.M., Chavkin, C. and Dubner, R. (1994) Kappa2 opioid receptors inhibit NMDA receptor-mediated synaptic currents in guinea pig CA3 pyramidal cells. J. Neurosci., 14: 5580–5589. Caudle, R.M. and Dubner, R. (1998) Ifenprodil blocks the excitatory effects of the opioid peptide dynorphin 1–17 on NMDA receptor mediated currents in the CA3 region of the guinea pig hippocampus. Neuropeptides, 32: 87–95. Chavkin, C. (2000) Dynorphins are endogenous opioid peptides released from granule cells to act neurohumorly and inhibit

259 excitatory neurotransmission in the hippocampus. Prog. Brain Res., 125: 363–367. Chavkin, C., Henriksen, S.J., Siggins, G.R. and Bloom, F.E. (1985a) Selective inactivation of opioid receptors in rat hippocampus demonstrates that dynorphin-A and -B may act on mu-receptors in the CA1 region. Brain Res., 331: 366–370. Chavkin, C., James, I.F. and Goldstein, A. (1982) Dynorphin is a specific endogenous ligand of the kappa opioid receptor. Science, 215: 413–415. Chavkin, C., Shoemaker, W.J., McGinty, J.F., Bayon, A. and Bloom, F.E. (1985b) Characterization of the prodynorphin and proenkephalin neuropeptide systems in rat hippocampus. J. Neurosci., 5: 808–816. Chen, Y., Mestek, A., Liu, J., Hurley, J.A. and Yu, L. (1993) Molecular cloning and functional expression of a mu-opioid receptor from rat brain. Mol. Pharmacol., 44: 8–12. Clark, J.A., Liu, L., Price, M., Hersh, B., Edelson, M. and Pasternak, G.W. (1989) Kappa opiate receptor multiplicity: evidence for two U50,488-sensitive kappa 1 subtypes and a novel kappa 3 subtype. J. Pharmacol. Exp. Ther., 251: 461–468. Cohen, G.A., Doze, V.A. and Madison, D.V. (1992) Opioid inhibition of GABA release from presynaptic terminals of rat hippocampal interneurons. Neuron, 9: 325–335. Commons, K.G. and Milner, T.A. (1995) Ultrastructural heterogeneity of enkephalin-containing neurons in the rat hippocampal formation. J. Comp. Neurol., 358: 324–342. Commons, K.G. and Milner, T.A. (1996a) Cellular and subcellular localization of delta opiate receptor immunoreactivity in the rat dentate gyrus. Brain Res., 738: 181–195. Commons, K.G. and Milner, T.A. (1996b) The ultrastructural relationships between leu-enkephalin and GABA containing neurons differ within the hippocampal formation. Brain Res., 724: 1–15. Corbett, A.D., Paterson, S.J. and Kosterlitz, H.W. (1993) Selectivity of ligands for opioid receptors. In: Herz A. (Ed.), Opioids I. Springer-Verlag, Berlin, pp. 645–680. Corbett, A.D., Paterson, S.J., McKnight, A.T., Magnan, J. and Kosterlitz, H.W. (1982) Dynorphin and dynorphin are ligands for the kappa-subtype of opiate receptor. Nature, 299: 79–81. Crain, B.J., Chang, K.J. and McNamara, J.O. (1986) Quantitative autoradiographic analysis of mu and delta opioid binding sites in the rat hippocampal formation. J. Comp. Neurol., 246: 170–180. De Camilli, P. and Jahn, R. (1990) Pathways to regulated exocytosis in neurons. Annu. Rev. Physiol., 52: 625–645. Ding, Y.Q., Kaneko, T., Nomura, S. and Mizuno, N. (1996) Immunohistochemical localization of m-opioid receptors in the central nervous system of the rat. J. Comp. Neurol., 367: 375–402. Drake, C.T., Chang, P.C., Harris, J.A. and Milner, T.A. (2002) Neurons with mu opioid receptors interact indirectly with enkephalin-containing neurons in the rat dentate gyrus. Exp. Neurol., 176: 254–261. Drake, C.T., Chavkin, C. and Milner, T.A. (1997) Kappa opioid receptor-like immunoreactivity is present in substance

P-containing subcortical afferents in guinea pig dentate gyrus. Hippocampus, 7: 36–47. Drake, C.T. and Milner, T.A. (1999) Mu opioid receptors are in somatodendritic and axonal compartments of GABAergic neurons in rat hippocampal formation. Brain Res., 849: 203–215. Drake, C.T. and Milner, T.A. (2006) Mu opioid receptors are extensively co-localized with parvalbumin, but not somatostatin, in the dentate gyrus. Neurosci. Lett., 403: 176–180. Drake, C.T., Patterson, T.A., Simmons, M.L., Chavkin, C. and Milner, T.A. (1996) Kappa opioid receptor-like immunoreactivity in guinea pig brain: ultrastructural localization in presynaptic terminals in hippocampal formation. J. Comp. Neurol., 370: 377–395. Drake, C.T., Terman, G.W., Simmons, M.L., Milner, T.A., Kunkel, D.D., Schwartzkroin, P.A. and Chavkin, C. (1994) Dynorphin opioids present in dentate granule cells may function as retrograde inhibitory neurotransmitters. J. Neurosci., 14: 3736–3750. Eisch, A.J., Barrot, M., Schad, C.A., Self, D.W. and Nestler, E.J. (2000) Opiates inhibit neurogenesis in the adult rat hippocampus. Proc. Natl. Acad. Sci. U.S.A., 97: 7579–7584. Eisch, A.J. and Harburg, G.C. (2006) Opiates, psychostimulants, and adult hippocampal neurogenesis: Insights for addiction and stem cell biology. Hippocampus, 16: 271–286. Evans, C.J., Keith, D.E., Morrison, H., Magendzo, K. and Edwards, R.H. (1992) Cloning of a delta opioid receptor by functional expression. Science, 258: 1952–1955. Fredens, K., Stengaard Pedersen, K. and Larsson, L.I. (1984) Localization of enkephalin and cholecystokinin immunoreactivities in the perforant path terminal fields of the rat hippocampal formation. Brain Res., 304: 255–263. French, E.D. and Siggins, G.R. (1980) An iontophoretic survey of opioid peptide actions in the rat limbic system: in search of opiate epileptogenic mechanisms. Regul. Pept., 1: 127–146. Freund, T.F. and Buzsa´ki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Gall, C. (1988a) Seizures induce dramatic and distinctly different changes in enkephalin, dynorphin, and CCK immunoreactivities in mouse hippocampal mossy fibers. J. Neurosci., 8: 1852–1862. Gall, C.M. (1988b) Localization and seizure-induced alterations of opioid peptides and CCK in the hippocampus. Natl. Inst. Drug Abuse Res. Monogr. Ser., 82: 12–32. Gall, C., Brecha, N., Karten, H.J. and Chang, K.-J. (1981) Localization of enkephalin-like immunoreactivity to identified axonal and neuronal populations of the rat hippocampus. J. Comp. Neurol., 198: 335–350. George, S.R., Fan, T., Xie, Z., Tse, R., Tam, V., Varghese, G. and O’Dowd, B.F. (2000) Oligomerization of mu- and delta-opioid receptors, generation of novel functional properties. J. Biol. Chem., 275: 26128–26135. Gomes, I., Jordan, B.A., Gupta, A., Trapaidze, N., Nagy, V. and Devi, L.A. (2000) Heterodimerization of m and d opioid receptors: a role in opiate synergy. J. Neurosci., 20: NIL22–NIL26.

260 Goodman, R.R., Snyder, S.H., Kuhar, M.J. and Young III, W.S. (1980) Differentiation of delta and mu opiate receptor localizations by light microscopic autoradiography. Proc. Natl. Acad. Sci. U.S.A., 77: 6239–6243. Gould, E. and Gross, C.G. (2002) Neurogenesis in adult mammals: some progress and problems. J. Neurosci., 22: 619–623. Gubler, U., Seeburg, P., Hoffman, B.J., Gage, L.P. and Udenfriend, S. (1982) Molecular cloning establishes proenkephalin as precursor of enkephalin-containing peptides. Nature, 295: 206–208. Guerra, D., Sole, A., Cami, J. and Tobena, A. (1987) Neurophysiological performance in opiate addicts after rapid detoxification. Drug Alcohol Depend., 20: 261–270. Gulya, K., Gehlert, D.R., Wamsley, J.K., Mosberg, H., Hruby, V.J. and Yamamura, H.I. (1986) Light microscopic autoradiographic localization of delta opioid receptors in the rat brain using a highly selective bis-penicillamine cyclic enkephalin analog. J. Pharmacol. Exp. Ther., 238: 720–726. Heinricher, M.M. (2003) Orphanin FQ/nociceptin: from neural circuitry to behavior. Life Sci., 73: 813–822. Herkenham, M. and McLean, S. (1988) The anatomical relationship of opioid peptides and opiate receptors in the hippocampi of four rodent species. Natl. Inst. Drug Abuse Res. Monogr. Ser., 82: 33–47. Herlitze, S., Garcia, D.E., Mackie, K., Hille, B., Scheuer, T. and Catterall, W.A. (1996) Modulation of Ca2+ channels by G-protein beta gamma subunits. Nature, 380: 258–262. Hong, J.S., McGinty, J.F., Grimes, L., Kanamatsu, T., Obie, J. and Mitchell, C.L. (1988) Seizure-induced alterations in the metabolism of hippocampal opioid peptides suggest opioid modulation of seizure-related behaviors. Natl. Inst. Drug Abuse Res. Monogr. Ser., 82: 48–66. Hong, J.S., McGinty, J.F., Lee, P.H.K., Xie, C.W. and Mitchell, C.L. (1993) Relationship between hippocampal opioid peptides and seizures. Prog. Neurobiol., 40: 507–528. Hong, J.S., Wood, P.L., Gillin, J.C., Yang, H.Y. and Costa, E. (1980) Changes of hippocampal Met-enkephalin content after recurrent motor seizures. Nature, 285: 231–232. Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O., Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered patterns of dynorphin immunoreactivity suggest mossy fiber reorganization in human hippocampal epilepsy. J. Neurosci., 10: 267–282. Ikeda, S.R. (1996) Voltage-dependent modulation of N-type calcium channels by G-protein beta gamma subunits. Nature, 380: 255–258. Jackisch, R., Geppert, M., Brenner, A.S. and Illes, P. (1986) Presynaptic opioid receptors modulating acetylcholine release in the hippocampus of the rabbit. Naunyn Schmiedebergs Arch. Pharmacol., 332: 156–162. Jackisch, R., Werle, E. and Hertting, G. (1984) Identification of mechanisms involved in the modulation of release of noradrenaline in the hippocampus of the rabbit in vitro. Neuropharmacology, 23: 1363–1371. Jiang, Q., Takemori, A.E., Sultana, M., Portoghese, P.S., Bowen, W.D., Mosberg, H.I. and Porreca, F. (1991)

Differential antagonism of opioid delta antinociception by [D-Ala2,Leu5,Cys6]enkephalin and naltrindole 50 -isothiocyanate: evidence for delta receptor subtypes. J. Pharmacol. Exp. Ther., 257: 1069–1075. Johnston, H.M. and Morris, B.J. (1994) Nitric oxide alters proenkephalin and prodynorphin gene expression in hippocampal granule cells. Neuroscience, 61: 435–439. Kakidani, H., Furutani, Y., Takahashi, H., Noda, M., Morimoto, Y., Hirose, T., Asai, M., Inayama, S., Nakanishi, S. and Numa, S. (1982) Cloning and sequence analysis of cDNA for porcine beta-neo-endorphin/dynorphin precursor. Nature, 298: 245–249. Kalyuzhny, A.E. and Wessendorf, M.W. (1997) CNS GABA neurons express the mu-opioid receptor: immunocytochemical studies. Neuroreport, 20(8): 3367–3372. Kaplan, T.J., Skyers, P.R., Tabori, N.E., Drake, C.T. and Milner, T.A. (2004) Ultrastructural evidence for mu-opioid modulation of cholinergic pathways in rat dentate gyrus. Brain Res., 1019: 28–38. Karhunen, T., Vilim, F.S., Alexeeva, V., Weiss, K.R. and Church, P.J. (2001) Targeting of peptidergic vesicles in cotransmitting terminals. J. Neurosci., 21: RC127. Khachaturian, H., Schaefer, M.K.H. and Lewis, M.E. (1993) Anatomy and function of the endogenous opioid system. In: Herz A. (Ed.), Opioid I. Springer-Verlag, Berlin, pp. 471–497. Kieffer, B.L., Befort, K., Gaveriaux-Ruff, C. and Hirth, C.G. (1992) The delta opioid receptor: isolation of a cDNA by expression cloning and pharmacological characterization. Proc. Natl. Acad. Sci. U.S.A., 89: 12048–12052. Lee, P.H., Obie, J. and Hong, J.S. (1989) Opioids induce convulsions and wet dog shakes in rats: mediation by hippocampal mu, but not delta or kappa opioid receptors. J. Neurosci., 9: 692–697. Liu, S., Wang, J., Zhu, D.Y., Fu, Y., Lukowiak, K. and Lu, Y.M. (2003) Generation of functional inhibitory neurons in the adult rat hippocampus. J. Neurosci., 23: 732–736. Luscher, C., Jan, L.Y., Stoffel, M., Malenka, R.C. and Nicoll, R.A. (1997) G protein-coupled inwardly rectifying K+ channels (GIRKs) mediate postsynaptic but not presynaptic transmitter actions in hippocampal neurons. Neuron, 19: 687–695. Maderspach, K., Nemeth, K., Simon, J., Benyhe, S., Szucs, M. and Wollemann, M. (1991) A monoclonal antibody recognizing kappa—but not mu—and delta-opioid receptors. J. Neurochem., 56: 1897–1904. Maderspach, K., Takacs, J., Niewiadomska, G. and Csillag, A. (1995) Postsynaptic and extrasynaptic localization of kappaopioid receptor in selected brain areas of young rat and chick using an anti-receptor monoclonal antibody. J. Neurocytol., 24: 478–486. Madison, D.V. and Nicoll, R.A. (1988) Enkephalin hyperpolarizes interneurons in the rat hippocampus. J. Physiol., 398: 123–130. Mansour, A., Burke, S., Pavlic, R.J., Akil, H. and Watson, S.J. (1996) Immunohistochemical localization of the cloned kappa 1, receptor in the rat CNS and pituitary. Neuroscience, 71: 671–690.

261 Mansour, A., Fox, C.A., Burke, S., Akil, H. and Watson, S.J. (1995) Immunohistochemical localization of the cloned mu opioid receptor in the rat CNS. J. Chem. Neuroanat., 8: 283–305. Mansour, A., Khachaturian, H., Lewis, M.E., Akil, H. and Watson, S.J. (1987) Autoradiographic differentiation of mu, delta, and kappa opioid receptors in the rat forebrain and midbrain. J. Neurosci., 7: 2445–2464. Mansour, A., Khachaturian, H., Lewis, M.E., Akil, H. and Watson, S.J. (1988) Anatomy of CNS opioid receptors. Trends Neurosci., 11: 308–314. Mansour, A., Thompson, R.C., Akil, H. and Watson, S.J. (1993) Delta opioid receptor mRNA distribution in the brain: comparison to delta receptor binding and proenkephalin mRNA. J. Chem. Neuroanat., 6: 351–362. Martin, W.R. (1983) Pharmacology of opioids. Pharmacol. Rev., 35: 283–323. Matsumoto, M., Yoshioka, M., Togashi, H., Hirokami, M., Tochihara, M., Ikeda, T., Smith, C.B. and Saito, H. (1994) mu-Opioid receptors modulate noradrenaline release from the rat hippocampus as measured by brain microdialysis. Brain Res., 636: 1–8. Mattia, A., Vanderah, T., Mosberg, H.I. and Porreca, F. (1991) Lack of antinociceptive cross-tolerance between [D-Pen2, D-Pen5]enkephalin and [D-Ala2]deltorphin II in mice: evidence for delta receptor subtypes. J. Pharmacol. Exp. Ther., 258: 583–587. McDaniel, K.L., Mundy, W.R. and Tilsom, H.A. (1990) Microinjection of dynorphin into the hippocampal formation impairs spatial learning in rats. Pharmacol. Biochem. Behav., 35: 429–435. McGinty, J.F., Henriksen, S.J., Goldstein, A., Terenius, L. and Bloom, F.E. (1983) Dynorphin is contained within hippocampal mossy fibers: immunochemical alterations after Kainic acid administration and colchicineinduced neurotoxicity. Proc. Natl. Acad. Sci. U.S.A., 80: 589–593. McGinty, J.F., Kanamatsu, T., Obie, J. and Hong, J.S. (1986) Modulation of opioid peptide metabolism by seizures: differentiation of opioid subclasses. Natl. Inst. Drug Abuse Res. Monogr. Ser., 71: 89–101. McLean, S., Rothman, R.B., Jacobson, A.E., Rice, K.C. and Herkenham, M. (1987) Distribution of opiate receptor subtypes and enkephalin and dynorphin immunoreactivity in the hippocampus of squirrel, guinea pig, rat, and hamster. J. Comp. Neurol., 255: 497–510. Merchenthaler, I., Maderdrut, J.L., Cianchetta, P., Shughrue, P.J. and Bronstein, D. (1997) In situ hybridization histochemical localization of prodynorphin messenger RNA in the central nervous system of the rat. J. Comp. Neurol., 384: 211–232. Meunier, J.-C., Mollereau, C., Toll, L., Suaudeau, C., Moisand, C., Alvinerie, P., Butour, J.-L., Guillemot, J.-C., Ferrara, P., Monsarrat, B., Mazarguil, H., Vassart, G., Parmentier, M. and Costentin, J. (1995) Isolation and structure of the endogenous agonist of opioid receptor-like ORL1 receptor. Nature, 377: 532–536.

Milner, T.A. and Drake, C.T. (2001) Ultrastructural evidence for presynaptic m opioid receptor modulation of synaptic plasticity in NMDA-receptor-containing dendrites in the dentate gyrus. Brain Res. Bull., 54: 131–140. Minami, M., Hosoi, Y., Toya, T., Katao, Y., Maekawa, K., Katsumata, S., Yabuuchi, K., Onogi, T. and Satoh, M. (1993) In situ hybridization study of kappa-opioid receptor mRNA in the rat brain. Neurosci. Lett., 162: 161–164. Mogil, J.S. and Pasternak, G.W. (2001) The molecular and behavioral pharmacology of the orphanin FQ/nociceptin peptide and receptor family. Pharmacol. Rev., 53: 381–415. Moore, S.D., Madamba, S.G., Schweitzer, P. and Siggins, G.R. (1994) Voltage-dependent effects of opioid peptides on hippocampal CA3 pyramidal neurons in vitro. J. Neurosci., 14: 809–820. Morris, B.J. and Johnston, H.M. (1995) A role for hippocampal opioids in long-term functional plasticity. Trends Neurosci., 18: 350–355. Nestler, E.J. (2001) Molecular basis of long-term plasticity underlying addiction. Nat. Rev. Neurosci., 2: 119–128. Neumaier, J.F., Mailheau, S. and Chavkin, C. (1988) Opioid receptor-mediated responses in the dentate gyrus and CA1 region of the rat hippocampus. J. Pharmacol. Exp. Ther., 244: 564–570. Nock, B., Rajpara, A., O’Connor, L.H. and Cicero, T.J. (1988) Autoradiography of [3 H]U-69593 binding sites in rat brain: evidence for kappa opioid receptor subtypes. Eur. J. Pharmacol., 154: 27–34. Noda, M., Furutani, Y., Takahashi, H., Toyosato, M., Hirose, T., Inayama, S., Nakanishi, S. and Numa, S. (1982) Cloning and sequence analysis of cDNA for bovine adrenal preproenkephalin. Nature, 295: 202–206. Pan, Y.X., Xu, J., Bolan, E., Abbadie, C., Chang, A., Zuckerman, A., Rossi, G. and Pasternak, G.W. (1999) Identification and characterization of three new alternatively spliced muopioid receptor isoforms. Mol. Pharmacol., 56: 396–403. Pierce, J.P., Kurucz, O. and Milner, T.A. (1999) The morphometry of a peptidergic transmitter system before and after seizure. I. Dynorphin B-like immunoreactivity in the hippocampal mossy fiber system. Hippocampus, 9: 255–276. Pierce, J.P. and Milner, T.A. (2001) Parallel increases in the synaptic and surface areas of mossy fiber terminals following seizure induction. Synapse, 39: 249–256. Pierce, T.L. and Wessendorf, M.W. (2000) Immunocytochemical mapping of endomorphin-2-immunoreactivity in rat brain. J. Chem. Neuroanat., 18: 181–207. Piguet, P. and North, R.A. (1993) Opioid actions at mu and delta receptors in the rat dentate gyrus in vitro. J. Pharmacol. Exp. Ther., 266: 1139–1146. Piva, F., Limonta, P., Dondi, D., Pimpinelli, F., Martini, L. and Maggi, R. (1995) Effects of steroids on the brain opioid system. J. Steroid Biochem. Mol. Biol., 53: 343–348. Pu, L., Bao, G.B., Xu, N.J., Ma, L. and Pei, G. (2002) Hippocampal long-term potentiation is reduced by chronic opiate treatment and can be restored by re-exposure to opiates. J. Neurosci., 22: 1914–1921.

262 Quinones-Jenab, V., Jenab, S., Ogawa, S., Inturrisi, C. and Pfaff, D.W. (1997) Estrogen regulation of mu-opioid receptor mRNA in the forebrain of female rats. Brain Res., 47: 1. Rees, H., Ang, L.C., Shul, D.D., George, D.H., Begley, H. and McConnell, T. (1994) Increase in enkephalin-like immunoreactivity in hippocampi of adults with generalized epilepsy. Brain Res., 652: 113–119. Reinscheid, R.K., Nothacker, H.P., Bourson, A., Ardati, A., Henningsen, R.A., Bunzow, J.R., Grandy, D.K., Langen, H., Monsma Jr., F.J. and Civelli, O. (1995) Orphanin FQ: a neuropeptide that activates an opioid-like G protein-coupled receptor. Science, 270: 792–794. Rios, C.D., Jordan, B.A., Gomes, I. and Devi, L.A. (2001) G-protein-coupled receptor dimerization: modulation of receptor function. Pharmacol. Ther., 92: 71–87. Robbins, T.W. and Everitt, B.J. (2002) Limbic-striatal memory systems and drug addiction. Neurobiol. Learn. Mem., 78: 625–636. Rocha, L., Ackermann, R.F. and Engel Jr., J. (1996) Effects of chronic morphine pretreatment on amygdaloid kindling development, postictal seizure and suppression and benzodiazepine receptor binding in rats. Epilepsy Res., 23: 225–233. Roman, E., Ploj, K., Gustafsson, L., Meyerson, B.J. and Nylander, I. (2006) Variations in opioid peptide levels during the estrous cycle in Sprague-Dawley rats. Neuropeptides, 40: 195–206. Rothman, R.B., Bowen, W.D., Schumacher, U.K. and Pert, C.B. (1983) Effect of beta-FNA on opiate receptor binding: preliminary evidence for two types of mu receptors. Eur. J. Pharmacol., 95: 147–148. Rothman, R.B., Jacobson, A.E., Rice, K.C. and Herkenham, M. (1987) Autoradiographic evidence for two classes of mu opioid binding sites in rat brain using [125I]FK33824. Peptides, 8: 1015–1021. Sandin, J., Nylander, I., Georgieva, J., Scho¨tt, P.A., O¨gren, S.O. and Terenius, L. (1998) Hippocampal dynorphin B injections impair spatial learning in rats: a kappa-opioid receptor-mediated effect. Neuroscience, 85: 375–382. Schindler, C.J., Slamberova, R., Rimanoczy, A., Hnactzuk, O.C., Riley, M.A. and Vathy, I. (2004) Field-specific changes in hippocampal opioid mRNA, peptides, and receptors due to prenatal morphine exposure in adult male rats. Neuroscience, 126: 355–364. Scholz, K.P. and Miller, R.J. (1992) Inhibition of quantal transmitter release in the absence of calcium influx by a G protein-linked adenosine receptor at hippocampal synapses. Neuron, 8: 1139–1150. Seri, B., Garcia-Verdugo, J.M., McEwen, B.S. and AlvarezBuylla, A. (2001) Astrocytes give rise to new neurons in the adult mammalian hippocampus. J. Neurosci., 21: 7153–7160. Seward, E.P., Chernevskaya, N.I. and Nowycky, M.C. (1995) Exocytosis in peptidergic nerve terminals exhibits two calcium-sensitive phases during pulsatile calcium entry. J. Neurosci., 15: 3390–3399. Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and Gould, E. (2001) Neurogenesis in the adult is involved in the formation of trace memories. Nature, 410: 372–376.

Shors, T.J., Townsend, D.A., Zhao, M., Kozorovitskiy, Y. and Gould, E. (2002) Neurogenesis may relate to some but not all types of hippocampal-dependent learning. Hippocampus, 12: 578–584. Shuster, S.J., Riedl, M., Li, X.R., Vulchanova, L. and Elde, R. (1999) Stimulus-dependent translocation of kappa opioid receptors to the plasma membrane. J. Neurosci., 19: 2658–2664. Simmons, M.L. and Chavkin, C. (1996) Endogenous opioid regulation of hippocampal function. Int. Rev. Neurobiol., 39: 145–196. Simmons, M.L., Terman, G.W., Drake, C.T. and Chavkin, C. (1994) Inhibition of glutamate release by presynaptic kappa 1—opioid receptors in the guinea pig dentate gyrus. J. Neurophysiol., 72: 1697–1705. Simmons, M.L., Terman, G.W., Gibbs, S.M. and Chavkin, C. (1995) L-type calcium channels mediate dynorphin neuropeptide release from dendrites but not axons of hippocampal granule cells. Neuron, 14: 1265–1272. Simonato, M. and Romualdi, P. (1996) Dynorphin and epilepsy. Prog. Neurobiol., 50: 557–583. Skyers, P.S., Einheber, S., Pierce, J.P. and Milner, T.A. (2003) Increased mu-opioid receptor labeling is found on inner molecular layer terminals of the dentate gyrus following seizures. Exp. Neurol., 179: 200–209. Slamberova´, R., Rimanoczy, A., Bar, N., Schindler, C.J. and Vathy, I. (2003) Density of mu-opioid receptors in the hippocampus of adult male and female rats is altered by prenatal morphine exposure and gonadal hormone treatment. Hippocampus, 13: 461–471. Slowe, S.J., Simonin, F., Kieffer, B. and Kitchen, I. (1999) Quantitative autoradiography of mu-, delta- and kappa 1 opioid receptors in kappa-opioid receptor knockout mice. Brain Res., 818: 335–345. Stumm, R.K., Zhou, C., Schulz, S. and Hollt, V. (2004) Neuronal types expressing mu- and delta-opioid receptor mRNA in the rat hippocampal formation. J. Comp. Neurol., 469: 107–118. Svingos, A.L., Colago, E.E.O. and Pickel, V.M. (1999) Cellular sites for dynorphin activation of kappa-opioid receptors in the rat nucleus accumbens shell. J. Neurosci., 19: 1804–1813. Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999) Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat. J. Neurosci., 19: 5792–5801. Tempel, A. and Zukin, R.S. (1987) Neuroanatomical patterns of the mu, delta, and kappa opioid receptors of rat brain as determined by quantitative in vitro autoradiography. Proc. Natl. Acad. Sci. U.S.A., 84: 4308–4312. Terman, G.W., Drake, C.T., Simmons, M.L., Milner, T.A. and Chavkin, C. (2000) Opioid modulation of recurrent excitation in the hippocampal dentate gyrus. J. Neurosci., 20: 4379–4388. Terman, G.W., Wagner, J.J. and Chavkin, C. (1994) Kappa opioids inhibit induction of long-term potentiation in the dentate gyrus of the guinea pig hippocampus. J. Neurosci., 14: 4740–4747. Thureson-Klein, A.K. and Klein, R.L. (1990) Exocytosis from neuronal large dense-core vesicles. Int. Rev. Cytol., 121: 67–126.

263 Tielen, A.M., van Leeuwen, F.W. and Lopes da Silva, F.H. (1982) The localization of leucine-enkephalin immunoreactivity within the guinea pig hippocampus. Exp. Brain Res., 48: 288–295. Torres-Reveron, A., Khalid, S., Drake, C.T. and Milner, T.A. (2006a) Enkephalin and dynorphin immunoreactivities increase in the dorsal hippocampus in response to estrogens and colocalize with estrogen receptor beta in mossy fiber terminals. Program No. 724.6 2006. Neuroscience Meeting Planner Atlanta GA: Society for Neuroscience, 2006. Online. Torres-Reveron, A., Khalid, S., Drake, C.T. and Milner, T.A. (2006b) Enkephalin-immunoreactivity colocalizes with estrogen receptor immunoreactivity in terminals and increases in proestrus and estrogen replaced female rats in the dorsal hippocampus. Program No. Tn-26 2006. Poster Session Abstracts St. Paul Minnesota. International Narcotics Research Conference. Tortella, F.C., Echevarria, E., Robles, L., Mosberg, H.I. and Holaday, J.W. (1988) Anticonvulsant effects of mu (DAGO) and delta (DPDPE) enkephalins in rats. Peptides, 9: 1177–1181. Turlejski, K. and Djavadian, R. (2002) Life-long stability of neurons: a century of research on neurogenesis, neuronal death and neuron quantification in adult CNS. Prog. Brain Res., 136: 39–65. Unterwald, E.M., Knapp, C. and Zukin, R.S. (1991) Neuroanatomical localization of kappa 1 and kappa 2 opioid receptors in rat and guinea pig brain. Brain Res., 562: 57–65. Velı´ sek, L., Stanton, P.K., Moshe´, S.L. and Vathy, I. (2000) Prenatal morphine exposure enhances seizure susceptibility but suppresses long-term potentiation in the limbic system of adult male rats. Brain Res., 869: 186–193. Verhage, M., McMahon, H.T., Ghijsen, W.E.J.M., Boomsma, F., Scholten, G., Wiegant, V.M. and Nicholls, D.G. (1991) Differential release of amino acids, neuropeptides, and catecholamines from isolated nerve terminals. Neuron, 6: 517–524. Vindrola, O., Briones, R., Asai, M. and Fernandez Guardiola, A. (1981) Amygdaloid kindling enhances the enkephalin content in rat brain. Neurosci. Lett., 21: 39–44. Wagner, J.J., Caudle, R.M. and Chavkin, C. (1992) Kappaopioids decrease excitatory transmission in the dentate gyrus of the guinea pig hippocampus. J. Neurosci., 12: 132–141. Wagner, J.J., Caudle, R.M., Neumaier, J.F. and Chavkin, C. (1990) Stimulation of endogenous opioid release displaces mu receptor binding in rat hippocampus. Neuroscience, 37: 45–53. Wagner, J.J. and Chavkin, C. (1992) Endogenous opioids and LTP: another view. Neurochem. Int., 20: 457–460. Wagner, J.J., Evans, C.J. and Chavkin, C. (1991) Focal stimulation of the mossy fibers releases endogenous dynorphins that bind kappa 1-opioid receptors in guinea pig hippocampus. J. Neurochem., 57: 333–343. Wagner, J.J., Terman, G.W. and Chavkin, C. (1993) Endogenous dynorphins inhibit excitatory neurotransmission and block LTP induction in the hippocampus. Nature, 363: 451–454.

Weisskopf, M.G., Zalutsky, R.A. and Nicoll, R.A. (1993) The opioid peptide dynorphin mediates heterosynaptic depression of hippocampal mossy fibre synapses and modulates longterm potentiation. Nature, 362: 423–427. White, J.D., Gall, C.M. and McKelvy, J.F. (1986) Proenkephalin is processed in a projection-specific manner in the rat central nervous system. Proc. Natl. Acad. Sci. U.S.A., 83: 7099–7103. Wimpey, T.L. and Chavkin, C. (1991) Opioids activate both an inward rectifier and a novel voltage-gated potassium conductance in the hippocampal formation. Neuron, 6: 281–289. Wolozin, B.L. and Pasternak, G.W. (1981) Classification of multiple morphine and enkephalin binding sites in the central nervous system. Proc. Natl. Acad. Sci. U.S.A., 78: 6181–6185. Woolley, C.S. and Schwartzkroin, P.A. (1998) Hormonal effects on the brain. Epilepsia, 39(Suppl 8): S2–S8. Xie, C.W., Morrisett, R.A. and Lewis, D.V. (1992) Mu opioid receptor-mediated modulation of synaptic currents in dentate granule cells of rat hippocampus. J. Neurophysiol., 68: 1113–1120. Yasuda, K., Raynor, H., Kong, C.D., Breder, J., Takeda, J., Reisine, T. and Bell, G.I. (1993) Cloning and functional comparison of kappa and delta opioid receptors from mouse brain. Proc. Natl. Acad. Sci. U.S.A., 90: 6736–6740. Yoshioka, M., Matsumoto, M., Togashi, H., Smith, C.B. and Saito, H. (1993) Opioid receptor regulation of 5-hydroxytryptamine release from the rat hippocampus measured by in vivo microdialysis. Brain Res., 613: 74–79. Yuste, R. and Tank, D.W. (1996) Dendritic integration in mammalian neurons, a century after Cajal. Neuron, 16: 701–716. Zadina, J.E., Hackler, L., Ge, L.J. and Kastin, A.J. (1997) A potent and selective endogenous agonist for the m-opiate receptor. Nature, 386: 499–502. Zakarian, S. and Smyth, D. (1979) Distribution of active and inactive forms of endorphins in rat pituitary and brain. Proc. Natl. Acad. Sci. U.S.A., 76: 5972–5976. Zakarian, S. and Smyth, D.G. (1982) Distribution of betaendorphin-related peptides in rat pituitary and brain. Biochem. J., 202: 561–571. Zieglga¨nsberger, W., French, E.D., Siggins, G.R. and Bloom, F.E. (1979) Opioid peptides may excite hippocampal pyramidal neurons by inhibiting adjacent inhibitory interneurons. Science, 205: 415–417. Zimprich, A., Simon, T. and Hollt, V. (1995) Cloning and expression of an isoform of the rat mu opioid receptor (rMOR1B) which differs in agonist induced desensitization from rMOR1. FEBS Lett., 359: 142–146. Zukin, R.S., Eghbali, M., Olive, D., Unterwald, E.M. and Tempel, A. (1988) Characterization and visualization of rat and guinea pig brain kappa opioid receptors: evidence for kappa 1 and kappa 2 opioid receptors. Proc. Natl. Acad. Sci. U.S.A., 85: 4061–4065. RU:1 do you mean KORs too? above you mention KORs do opposite things( and see below comment 8. RU:2 could you provide a reference - to a review because yours is the only opioids chapter in the book.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 16

Somatostatin in the dentate gyrus Melanie K. Tallent Department of Pharmacology and Physiology, Drexel University College of Medicine, 245 N. 15 St., Philadelphia, PA 19102, USA

Abstract: The neuropeptide somatostatin (SST) is expressed in a discrete population of interneurons in the dentate gyrus. These interneurons have their soma in the hilus and project to the outer molecular layer onto dendrites of dentate granule cells, adjacent to perforant path input. SST-containing interneurons are very sensitive to excitotoxicty, and thus are vulnerable to a variety of neurological diseases and insults, including epilepsy, Alzheimer’s disease, traumatic brain injury, and ischemia. The SST gene contains a prototypical cyclic AMP response element (CRE) site. Such a regulatory site confers activity-dependence to the gene, such that it is turned on when neuronal activity is high. Thus SST expression is increased by pathological conditions such as seizures and by natural stimulation such as environmental enrichment. SST may play an important role in cognition by modulating the response of neurons to synaptic input. In the dentate, SST and the related peptide cortistatin (CST) reduce the likelihood of generating long-term potentiation, a cellular process involved in learning and memory. Thus these neuropeptides would increase the threshold of input required for acquisition of new memories, increasing ‘‘signal to noise’’ to filter out irrelevant environmental cues. The major mechanism through which SST inhibits LTP is likely through inhibition of voltage-gated Ca2+ channels on dentate granule cell dendrites. Transgenic overexpression of CST in the dentate leads to profound deficits in spatial learning and memory, validating its role in cognitive processing. A reduction of synaptic potentiation by SST and CST in dentate may also contribute to the well-characterized antiepileptic properties of these neuropeptides. Thus SST and CST are important neuromodulators in the dentate gyrus, and disruption of this signaling system may have major impact on hippocampal function. Keywords: somatostatin; cortistatin; neuropeptide; electrophysiology; long-term potentiation; interneuron; cognition; epilepsy Introduction

neurons or HIPP cells), however, its physiological role here is relatively unknown. The selective targeting of SST-containing interneuron terminals to distal dendrites of principal neurons in both dentate and cornu ammonis suggest that this peptide may play a similar role throughout the hippocampus. Interest in the role of SST in the dentate has been heightened by studies showing it is upregulated by mild seizures, but SST-containing interneurons are vulnerable to severe seizures. What

Somatostatin (SST) is a neuropeptide abundantly expressed throughout the brain and periphery. In the dentate gyrus, SST has been extensively used as a chemical marker of a subset of interneurons located in the hilus (hilar perforant path associated Corresponding author. Tel.: +1 215 762 8208; Fax: +1 215 762 2299; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63016-7

265

266

role, if any, the loss of the SSTergic system plays in postseizure excitability in the dentate is still under investigation in our lab and others. Recent evidence suggests that SST may be a critical regulator of synaptic plasticity that underlies learning and memory in the dentate (Baratta et al., 2002). SST was first discovered in 1972 by Vale and colleagues as a hypothalamic inhibitor of growth hormone release from the anterior pituitary (Brazeau et al., 1972). In the periphery, SST generally acts as an inhibitor of hormone release from the gut and pancreas (Reisine, 1995). In the brain, SST also generally has inhibitory actions at the cellular level, but can increase neurotransmitter release in some instances, likely through disinhibitory mechanisms [i.e., inhibition of inhibitory interneurons, Chesselet and Reisine (1983)]. SST distribution in the brain is extensive, with virtually every region showing some localization at some point in development. SST expression in forebrain is prominent, containing perhaps the highest brain levels outside the hypothalamus. More recently, a related peptide from a distinct gene was discovered by de Lecea and colleagues, cortistatin (CST). The distribution of CST is more regionally limited than SST, being largely restricted to cortex and hippocampus (de Lecea et al., 1996). SST activates at least three types of K+ channels, KIR, KM, and KLK (Moore et al., 1988; Velimirovic et al., 1995; Schweitzer et al., 1998). These actions of SST hyperpolarize neurons, moving them away from their threshold for firing. Activation of K+ channels also decreases input resistance of neurons, lessening the depolarizing actions of excitatory synaptic inputs. SST inhibits voltage-sensitive Ca2+ channels in neurons as well (Wang et al., 1990b; Boehm and Betz, 1997). This action could also limit excitability, either through postsynaptic decrease of Ca2+-induced depolarizations and/or signaling, or through presynaptic inhibition of neurotransmitter release. However, increasing evidence suggests that the major mechanism for inhibition of neurotransmitter release by Gi/Go coupled receptors is by direct inhibition of the synaptic protein SNAP-25 by bg subunits (Blackmer et al., 2005; Gerachshenko et al., 2005). SST inhibition of glutamate release in cultured hippocampal neurons was shown to require Gai2,

Gai3, or Gao1 (Straiker et al., 2002). Since Ga subunits show preference in coupling to distinct bg subunits (Robishaw and Berlot, 2004), specificity in Ga coupling likely confers specificity in bg subunits that that interact with SNAP-25. SST and CST mediate their actions via the same family of receptors. Five SST receptors have been cloned, SST1–SST5. These are prototypical Gi/Go G-protein coupled receptors most closely related to the opioid family of receptors. Only SST2 is alternatively spliced, with two described variants, SST2a and SST2b (Schindler et al., 1999). Expression of SST1–SST4 in brain has been confirmed; SST5 expression in brain is still somewhat controversial (Stroh et al., 1999; Schulz et al., 2000). SST2, SST3, and SST4 are all expressed in hippocampus (Dournaud et al., 1996a; Handel et al., 1999; Schreff et al., 2000; Schulz et al., 2000); SST1 expression in forebrain has not been validated using antibodies that have had their specificity confirmed in SST1 receptor knockout mice (Hervieu and Emson, 1998; Schulz et al., 2000). Interesting distinct subcellular localization of the different SST receptor subtypes has been reported on the principal neurons of hippocampus. SST2 expression is largely restricted to soma and proximal dendrites (Dournaud et al., 1996a), SST4 to medial and distal dendrites (Schreff et al., 2000), and SST3 has a pattern of expression unique from any other G protein-coupled receptor in that it is exclusively expressed on neuronal cilia (Handel et al., 1999). Thus, although these three receptors are expressed in the same neurons, there is little spatial overlap (Schreff et al., 2000). Excepting axons (which may contain SST1, Hervieu and Emson, 1998), the principal projecting neurons in hippocampus therefore are virtually coated with SST receptors, suggesting they are exquisitely sensitive to this neuropeptide, and that SST is an extremely important signaling molecule in this region.

SST receptors activate multiple signaling systems Studies in expression systems have shown SST to couple to multiple signaling systems, including adenylyl cyclase, protein kinase C, and MAP

267

kinases. Signaling of these receptors in brain has been less studied, besides their regulation of voltage-dependent ion channels. Most of the studies examining signaling pathways activated by SST receptors have used recombinant receptors in expression systems. Such studies have shown that SST receptors can activate or inhibit a multitude of effectors, most of which are typical for the Gi/Go family of G proteins (Neves et al., 2002). For example, activation of all cloned SST receptors leads to the inhibition of adenylyl cyclase (Moller et al., 2003). Activation of tyrosine phosphates and phospholipase C (PLC) has also been reported for all of the cloned receptors (Lahlou et al., 2004), although one study in which all five human receptors were expressed in the same cell line showed only SST3 and SST5 significantly activated PLC and increased intracellular Ca2+ (Siehler and Hoyer, 1999). Activation of MAP kinases has been observed for the majority of the receptors (Florio et al., 1999), specifically ERK (Lahlou et al., 2003), and p38 (Moller et al., 2003), although inhibition of ERK has also been reported (Todisco et al., 1994; Pola et al., 2003; Lahlou et al., 2004). SST has also been shown to activate PI3 kinase (Lahlou et al., 2003). Several studies have addressed signaling mechanisms of endogenous SST receptors in cell lines and in situ; interest has been raised because of the finding that SST has antiproliferative actions on cancer cells. Since activation of SST receptors inhibits adenylyl cyclase activity, inhibition of cAMP-dependent signaling pathways would be a likely mechanism for the antiproliferative effects of SST. One such molecule is the cAMP response element binding protein (CREB), a transcription factor that regulates protein expression by binding to CRE sites in their promoter region. Indeed, the SST gene itself has such a CRE site (as does SST2, Kimura et al., 2001), suggesting a mechanism of reciprocal regulation (Montminy and Bilezikjian, 1987). SST was shown to inhibit CREB phosphorylation in GH4 somatotrophic cells via inhibition of PKA (Tentler et al., 1997). SST has been shown to regulate other CREB-regulated signaling molecules in several cell lines. In the mouse insulinoma cell line MIN6, that expresses only SST3, SST transiently increases and then decreases the

CREB-regulated immediate early gene c-fos, and MAP kinase activity, an activator of CREB (Yoshitomi et al., 1997). This suggests SST receptor activation can inhibit CREB independently of adenylyl cyclase, although this was not directly assessed in this study. Interestingly, although SST inhibition of c-fos was blocked by pertussis toxin (PTX), activation of c-fos expression was PTX-insensitive, suggesting a non-Gi/Go mechanism (Yoshitomi et al., 1997). In the pituitary adenoma cell line GH3, SST inhibited c-fos mRNA expression through a PTX-sensitive mechanism that appeared to involve inhibition of ERK activity and Elk-1 phosphorylation (Todisco et al., 1994). This same group also found SST inhibited binding of the AP-1 transcription complex to its target DNA, which is necessary for c-fos activity. This action of SST was blocked by PTX and the phosphatase inhibitors sodium orthovanadate and okadaic acid (Todisco et al., 1994, 1995). In the pancreatic cell line AR42J and in human leukemia cell lines, SST also inhibited c-fos expression in a PTX-sensitive fashion (Ishihara et al., 1999; Cowles et al., 2002). Thus most studies suggest that SST inhibits cAMP-mediated gene expression, which is in keeping with its antiproliferative actions and its Gi/Go coupling. However, robust activation of c-fos expression by SST3 has been reported (Yoshitomi et al., 1997). This action may be mediated by activation of MAP kinases via the small GTPase Ras (Florio et al., 1999; Lahlou et al., 2003), which is a prototypical Gimediated signaling pathway (Neves et al., 2002). SST has also been found to stimulate adenylyl cyclase in somatotrophic cells (Ramirez et al., 2002), suggesting atypical G-protein coupling can occur in some cell types. It is interesting that SST receptors inhibit adenylyl cyclase, leading to a decrease in cAMP levels and decreased activity of PKA, while these same receptors can activate the MAP kinase ERK. Thus, although most studies suggest that SST inhibits CREB-mediated transcription, activation has also been reported. These contrasting actions may be receptor subtype and tissue specific, although this issue has not been adequately addressed in situ or in vivo. This complexity in signaling is of course not unique to SST receptors,

268

and simply because SST receptors have been shown that they can couple to a pathway in an artificial system, it does not necessarily follow that they do couple to these pathways to mediate their important physiological effects. Surprisingly few studies have addressed modulation of signaling pathways by SST in the brain, aside from its well-known actions on K+ and Ca2+ channels (Mihara et al., 1987; Inoue et al., 1988; Ikeda and Schofield, 1989; Wang et al., 1990a, b; Meriney et al., 1994; Connor et al., 1997; Sodickson and Bean, 1998). Inhibition of cAMP by SST in membranes from several brain regions has been observed, including cortex and hippocampus (Raynor and Reisine, 1992; Blake, 2001). Inhibition of c-fos mRNA expression has been reported in the trigeminal nucleus and arcuate nucleus (Bereiter, 1997; Dickson et al., 1997; Zheng et al., 1997), however, it is unclear whether this action is through activation of signaling pathways or through inhibition of neuronal firing, a widespread action of SST due to its augmentation of K+ currents.

Decrease in SST levels are found in Alzheimer’s brain Davies et al. (1980) first reported reduced SST levels in postmortem brains of Alzheimer’s disease patients. Since then there have been dozens of studies confirming and expanding on this initial observation. SST-containing interneurons were shown to be selectively associated with neuronal tangles in Alzheimer’s brain, suggesting a mechanism by which these neurons are vulnerable (Roberts et al., 1985). However, studies in rats suggest that lesioning cholinergic neurons also decreases SST neurons in cortex, suggesting the SST reduction is downstream of the well-characterized cholinergic deficits (Zhang et al., 1998). Other studies have examined regional changes in SST reductions, showing a loss of SST-containing neurons in frontal and parahippocampal cortex and hippocampus (Grouselle et al., 1998). A reduction in SST mRNA levels per surviving neuron was also observed in hippocampus but not cortex (Dournaud et al., 1994). A strong negative

correlation was found between levels of SST in cerebrospinal fluid (CSF) and degree of dementia in patients with Alzheimer’s disease (Edvinsson et al., 1993; Minthon et al., 1997). Thus a reduction in the CNS SST system is consistently observed in brains from Alzheimer’s patients; however, it is currently unknown how or whether this loss contributes to Alzheimer’s dementia, as the function of SST in cognition is still unclear. Interestingly, SST receptors appear to be largely preserved in Alzheimer’s disease (Kumar, 2005), suggesting them as potential targets for therapeutics. In mouse models of Alzheimer’s, decreases in brain SST levels have also been observed (Tomidokoro et al., 2000; Ramos et al., 2006), suggesting they could be appropriate models to study the role of SST in this disease. A novel role for SST with direct implications on Alzheimer’s disease has recently been reported. SST was shown to increase degradation of b-amyloid by upregulating the activity of neprilysin, an Ab degrading peptidase (Saito et al., 2005). Interestingly, in SST knockout mice, immunoreactive neprilysin was reduced by 78% in the molecular layer of the dentate gyrus. A corresponding 1.5-fold increase in Ab 42, the b amyloid fragment most closely associated with Alzheimer’s disease and amyloid plaques, was found in brains of SST knockout mice (Saito et al., 2005). Like SST, neprilysin decreases with aging and in Alzheimer’s disease (Yasojima et al., 2001; Iwata et al., 2002). An Alzheimer’s transgenic mouse model (APP23) crossed with a neprilysin knockout mouse showed increased oligomeric forms of Ab at synapses, and cognitive impairment, including deficits in spatial learning and object recognition memory at young ages prior to accumulation of amyloid plaques (Huang et al., 2006). Thus one functional consequence of reduced SST levels in brain may be increased accumulation of Ab, and upregulation of SST could be protective. Changes in CNS SST levels have been reported in other diseases as well, in particular those associated with cognitive dysfunction. Whether these changes influence the disease state is unknown. Reduced CSF SST levels were found in patients with depression (Agren and Lundqvist, 1984) or

269

multiple sclerosis (Vecsei et al., 1990), although regional changes in brain have not been explored in these diseases. In bipolar disorder, transcription analysis using microarrays found that SST was one of eight genes that were significantly correlated with the severity of the disease in postmortem tissue. In this study, SST was upregulated, and was the only gene to show significant association with bipolarism when genetic polymorphisms were analyzed (Nakatani et al., 2006). Increases in CSF SST was also found in the related disorder, mania (Sharma et al., 1995). Elevated SST levels are also detected in basal ganglia of patients with Huntington’s disease (Aronin et al., 1983; Nemeroff et al., 1983). In Parkinson’s disease, a reduction of SST in hippocampus and cortex was found only in patients showing cognitive deficits, suggesting that changes occur in the later stages of this disease (Epelbaum et al., 1983). Reduction in cortical SST is also found in schizophrenia (Gabriel et al., 1996); interestingly, in rat chronic haloperidol treatment increases cortical SST levels (Sakai et al., 1995).

SST is expressed in a subset of hilar interneurons in dentate gyrus In rat dentate gyrus, 16% of GAD-containing interneurons express SST (Kosaka et al., 1988). This neuropeptide has a distinct pattern of expression; it is co-localized with gamma amino butyric acid (GABA) in a specific subset of hilar neurons, the HIPP cells that project to the outer molecular layer. These neurons also have an axonal plexus within the hilus, suggesting they can modulate hilar neurons in addition to their projection to the outer molecular layer, where they presumably influence granule cells (Leranth et al., 1990; Lubke et al., 1998). Some SST neurons have axons that cross the hippocampal fissure into CA1, suggesting interactions of SST terminals from the dentate with SST terminals from the analogs population of SST neurons in CA1, which terminate nearby in the stratum lanunosum moleculare. The dendrites of HIPP cells are spiny and appear to be restricted to the hilus. They express mGluR1 and substance P receptors (Freund and Buzsaki, 1996).

Approximately 30% of SST interneurons coexpress neuropeptide Y. This is a common theme throughout the brain, and co-localization can be as common among SST and neuropeptide Y cells as 100% (see Chronwall et al., 1984; McDonald, 1989; Figueredo-Cardenas et al., 1996). The functional implication of this co-expression has not been examined in the dentate or elsewhere, but since both of these neuropeptides have generally inhibitory actions, they are likely to act synergistically when co-released. In rat and primate, including humans, immunostaining reveals a dense plexus of SST immunoreactivity in the outer molecular layer, in agreement with the projection pattern of HIPP cells (Amaral et al., 1988; Milner and Bacon, 1989; Leranth et al., 1990; Austin and Buckmaster, 2004; Csaba et al., 2005). Similar findings have been reported in other species, including pig (Holm et al., 1992), guinea pig, and rabbit (Buckmaster et al., 1994). In mice, however, this SST staining profile is absent (Buckmaster et al., 1994; Jinno and Kosaka, 2000). This does not appear to be due to a difference in the projection pattern of HIPP cells, since enhanced green fluorescent protein (EGFP) staining is present in the outer molecular layer in mice engineered to express this marker in SST-containing neurons (see below Oliva et al., 2000). Whether this difference in SST density in the outer molecular layer reflects a functional difference of SST between rats and mice is unknown. Electron microscopy has shown that HIPP cell axons terminate on spines of dentate granule cell (DGC) dendrites, forming symmetrical (inhibitory) synapses. Interestingly, asymmetrical (excitatory) synapses are frequently observed on the same spines, suggesting entorhinal afferents and HIPP axons form synapses on the same spines (Milner and Bacon, 1989; Leranth et al., 1990). Thus, based on the anatomical data, one would predict that HIPP interneurons would mediate feedback inhibition of granule cells, with their dendrites remaining in the hilus and receiving input from mossy fibers, and their axons terminating in the outer molecular layer adjacent to excitatory perforant path input onto granule cell dendrites. Definitive proof of this feedback role is lacking,

270

in fact, enhanced feedback inhibition has been consistently observed after seizure-induced death of HIPP neurons (Swanson et al., 1998; Sokal and Large, 2001; Kobayashi and Buckmaster, 2003). HIPP-like neurons have been characterized electrophysiologically in rat (Scharfman, 1992; Mott et al., 1997; Lubke et al., 1998), however, identification in these studies were based on limited morphological analysis and there are some discrepancies in the findings. An interesting model to study HIPP neuron properties has been developed. A subset of interneurons was engineered to express EGFP in a transgenic mouse, the GIN mice (GFP-expressing inhibitory neurons), using the GAD67 promoter (Oliva et al., 2000). SST-containing neurons account for 90–98% of the EGFP expressing cells in hippocampus and cortex. In the hilus, 95% of the EGFP expressing neurons were found to contain SST, although only 16% of the SST-containing neurons expressed EGFP (Oliva et al., 2000). Anatomically, the hilar EGFP interneurons were similar to the HIPP cells described by Freund and Buszaki (1996) in rat, with terminals in the outer two-thirds of the molecular layer. These neurons also were positive for mGluR1, also a characteristic of HIPP cells (Freund and Buzsaki, 1996). These mice have been used to characterize electrophysiological properties of SST interneurons in sensorimotor cortex, which revealed distinct subpopulations of these neurons (Halabisky et al., 2006). These mice provide an excellent model system for a rigorous study of electrophysiological and network properties of SST-containing HIPP cells in mice.

(Vela et al., 2003). In aged primates, a dramatic decline in hippocampal SST mRNA is also observed, although no regional analysis was performed in this study (Hayashi et al., 1997). Interestingly, a significant negative correlation has been found in cognitive decline in aged rats and SST levels (Dournaud et al., 1996b). That a similar negative relationship has been observed between SST levels and dementia in both Parkinson’s (Epelbaum et al., 1983) and Alzheimer’s disease in humans (Davies et al., 1980) suggests an intriguing potential role for SST in cognition. However, in most cases it is unclear whether this decrease in SST expression is due to changes in SST gene or transcript regulation, or due to the death of SST-containing interneurons (see below). If the latter is true, the relationship between SST levels and cognitive decline becomes more indirect. CST expression in the dentate gyrus is postnatally regulated as well, with a pattern of expression that is distinct from SST. At P0 in rat, CST mRNA is apparent in hippocampal stratum oriens but not the dentate. By P10, expression throughout the hippocampus is increasing, and SST-positive neurons appear in the hilus. Expression of CST mRNA in hilar neurons peaks at approximately postnatal day 16 and then declines, with very few labeled neurons present in adult hilus (de Lecea et al., 1997a). Whether CST has a specific function during development is unknown, although developmental differences in its electrophysiological actions have been reported (de Lecea et al., 1997a). Interestingly, in aged mice (24 months), CST levels have increased in hilar interneurons (Winsky-Sommerer et al., 2004).

SST expression in the hippocampus is regulated during development

SST expression is activity-dependent

Expression of SST in hippocampus changes during development. A surge in SST immunoreactivity occurs between days E12 and E15, with expression beginning in the subiculum and progressing ‘‘backward’’ though the hippocampus, so that hilar expression appears last (Rapp and Amaral, 1988). After maturity, a decline in SST levels occurs throughout rat hippocampus, with hilar expression showing the largest decrease in aged animals

The SST gene promoter has a prototypical CRE site that has been studied extensively to understand mechanisms of action of CREB (Montminy and Bilezikjian, 1987; Powers et al., 1989; Montminy et al., 1996). The transcription factor CREB is a major transducer of activity-dependent gene expression and thus mediates transformation of electrical to genetic signaling. CREB mediates its actions by interacting with CRE sites on

271

promoters. Thus, SST mRNA and ultimately peptide levels are upregulated by neuronal activity. SST in hilar interneurons appears to be especially sensitive to upregulation by seizure-like activity. This could be due to dense innervation of these neurons by mossy fibers, the axons of DGCs, because granule cells and other principal neurons of hippocampus are active during many types of seizures, particularly limbic seizures. Studies using c-fos, a downstream target of CREB, indicate that hilar interneurons are ‘‘turned on’’ transcriptionally shortly after DGCs and before CA3 or CA1 pyramidal neurons (Le Gal La Salle, 1988; Peng and Houser, 2005). Thus, it is not surprising that the intense neuronal activation underlying seizures results in upregulation of SST expression. Hippocampal kindling upregulates hilar SST mRNA and peptide once stage II seizures (mild seizures) are reached (Bendotti et al., 1993; Vezzani et al., 1996). This effect was transient, lasting
1 week. Repeated, but not single electroconvulsive shock (ECS)-induced seizures increases SST mRNA in the hilus (Passarelli and Orzi, 1993; Mikkelsen and Woldbye, 2006). This regulation is specific to dentate, as no significant changes were detected in CA3 and CA1 (Passarelli and Orzi, 1993). Another study showed repeated ECS increased the density of SST immunoreactivity in the outer molecular layer as well as in hilar cell bodies. This effect was additive over 36 days of repeated ECS. Interestingly, lesioning the perforant path input to the dentate did not alter ECS-induced upregulation of SST (Kragh et al., 1994). Kainate-induced seizures also induce expression of SST mRNA in the hilus, and elsewhere in hippocampus, beginning as early as 3 h after seizures (Hashimoto and Obata, 1991). Ectopic expression in DGCs and pyramidal cells has also been reported after kainate-induced seizures, suggesting SST expression can be activated in neurons that normally do not express the neuropeptide (Hashimoto and Obata, 1991; Calbet et al., 1999). Expression of CST, with a promoter distinct from SST, is not induced by kainate, demonstrating that these two related peptides are independently regulated (Calbet et al., 1999). Considering that neuropeptides are released when peptidergic neurons are activated at high

frequencies, it seems logical that during seizures, SST release would be robust. This has proven to be the case. In vivo voltammetry showed detectable SST release in hippocampus 48 min following a seizure, with increased release continuing for the 3 h recording period after the seizure (Manfridi et al., 1991). In hippocampal slices, both baseline and K+-induced SST release was significantly higher in kindled rats than in control rats (Vezzani et al., 1992). An in vivo study showed similar findings using microdialysis, that is, an increase in both basal and K+-stimulated release in hippocampus after kindling (Marti et al., 2000). That SST release is enhanced after kindling suggests that upregulation of SST after seizures has functional implications. Interestingly, SST has antiepileptic properties in hippocampus. This was first shown using in vivo models in rats, where SST was injected intrahippocampally or ICV. For example, Vezzani and colleagues demonstrated that SST or SST2 selective ligands could decrease the severity of seizures evoked by pentylenetetrazol (Perez et al., 1995), kainate (Vezzani et al., 2000), or quinolic acid (Vezzani et al., 1991). This group also showed that continuous infusion of anti-SST antibody into the hippocampus decreased the latency, and dramatically reduced the number of kindling stimulations required to reach stage V seizures, suggesting endogenous SST plays an important role in maintaining a seizure-free state. This was confirmed by a different group, who also demonstrated that anti-SST antibody injected into the hippocampus blocked the protective effects of SST, administered ICV, on picrotoxin-kindled rats (Mazarati and Telegdy, 1992). This study supports the hypothesis that hippocampal SST is critical in controlling epileptogenesis and seizures. We used an in vitro model system to examine whether SST could suppress epileptiform activity in the rat hippocampal slice (Tallent and Siggins, 1999). We found a robust effect of SST to decrease epileptiform events generated either by increasing excitability (by removing extracellular Mg++) or by decreasing inhibition (by blocking GABAA receptors). Although this study focused on epileptiform events in CA1 and CA3, our later studies in mouse hippocampal slices showed that SST has

272

robust effects on synaptic potentiation in dentate gyrus (Baratta et al., 2002, discussed below), a mechanism that likely contributes to epileptogenesis in this region (Sutula and Steward, 1986, 1987; Wasterlain et al., 1999). Thus, SST has potential sites of action throughout the hippocampus to control seizure activity. SST expression in the dentate is also regulated by pathological conditions that do not involve seizures. For example, SST expression increases 72 h following traumatic brain injury (Cook et al., 1998). Interestingly, this upregulation of endogenous SST could be protective, since ICV administration of SST or CST protected against ischemia-induced neuronal death (Rauca et al., 1999). Stress also increases expression and release of SST in the hilus (Arancibia et al., 2001).

dentate at E18, validating the presence of functional SST receptors (Thoss et al., 1996). At P5, SST2 and SST3 expression levels remain flat, while SST4 significantly increases, so that mRNA for SST2, SST3, and SST4 are expressed at similar levels. Likewise, an increase in labeling of SST receptors using radioligands is apparent in dentate from E18 to P5 (Thoss et al., 1996). From P5 to adult (no intermediate ages have been reported), SST1 levels have increased so that it is significantly expressed for the first time, SST2 levels modestly increase, SST4 levels modestly decrease, while SST3 levels robustly increase (Thoss et al., 1996). Regulation of SST receptor subtypes during aging has not been reported, although one study showed expression of SST1–SST5 in cortex of aged humans (Kumar, 2005).

Expression of SST receptors in dentate gyrus

Plasticity in expression of SST receptors

Prior to the cloning of the SST receptors, autoradiographic studies showed dense, high-affinity binding of radiolabeled SST analogs throughout the molecular layer, with much less binding in the granule cell layer and diffuse binding in the hilus (Leroux et al., 1993). These data suggested that SST released from terminals could bind to receptors throughout the dendritic layer, if it could diffuse far enough without proteolysis. Characterization of mRNA expression of the different receptor subtypes extended these earlier studies by providing evidence for SST1–SST4 in DGCs (Bruno et al., 1992; Kong et al., 1994). Immunohistochemical detection has confirmed the presence of SST2, SST3, and SST4 on DGCs (Dournaud et al., 1996a; Handel et al., 1999; Schreff et al., 2000). As described above for all hippocampal principal neurons, SST2 is expressed predominately on DGCs soma and proximal dendrites, SST3 on neuronal cilia, and SST4 is almost exclusively dendritic. Developmental changes in SST receptor mRNA expression have been reported (Thoss et al., 1995, 1996). Only SST2 and SST3 mRNA are present in significant amounts at E18, suggesting these two receptors could play a role in development. Robust high-affinity binding of SST ligands is present in

Activity-dependent changes in expression of SST receptors in the dentate following seizures have been observed, but do not appear to involve changes in gene expression. After kainate-induced seizures, no change was observed in mRNA for SST1–SST4 in dentate gyrus, although a decrease for SST3 and SST4 mRNA was found in CA1, corresponding to the death of pyramidal cells in this region (Perez et al., 1995). Kindling also did not cause changes in mRNA for SST receptors in the dentate, although a 40% reduction in SST receptor binding was observed in the molecular layer immediately, and 1 week following, kindlinginduced seizures. Binding returned to normal 30 days following seizures. Decreased SST binding without changes in mRNA expression could have been due to receptor downregulation as a consequence of increased release of SST, although it is unclear why reductions in binding were still apparent 1 week following seizures unless receptor-containing neurons were damaged or were lost (Piwko et al., 1996). Kindling also reduced immunoreactivity for SST2 in the outer molecular layer; this was associated with an increase in SST immunoreactivity, again suggesting downregulation on binding of the receptor to its ligand (Csaba et al., 2004). In dentate gyrus from humans with

273

epilepsy, SST2 receptor is strongly regulated (Csaba et al., 2005). In hippocampal tissue resected from patients with temporal lobe seizures, SST2 expression was decreased in CA1 and CA3, likely reflecting the neuronal loss found in this regions in patients exhibiting hippocampal sclerosis (De Lanerolle et al., 2003). In contrast, in DGCs, which are more resistant to seizure-induced death, SST2 mRNA was strongly upregulated. However, SST2 receptor binding and immunoreactivity showed a strong decrease in the outer molecular layer, suggesting SST2 may be redistributed.

Hilar SST interneurons are vulnerable to excitotoxicity A hallmark of epileptic hippocampus is the death of the SST-containing hilar interneurons. This was first shown by Sloviter (1987) in a rat model after induction of seizures by perforant path stimulation. In hippocampal tissue surgically removed from humans with intractable temporal lobe seizures, a similar pattern of loss was seen (De Lanerolle et al., 1989; Robbins et al., 1991). Of all subsets of GABAergic neurons, this class of interneurons appears to be selectively vulnerable, since other interneuronal markers, such as cholecystokinin and vasoactive intestinal peptide, appear unchanged (Robbins et al., 1991). This vulnerability of SST interneurons to seizureinduced death has since been confirmed in other animal models, including kainate- and pilocarpineinduced seizures (Buckmaster and Dudek, 1997; Kobayashi and Buckmaster, 2003). Interestingly, in electrically kindled rats, a model in which spontaneous seizures do not generally develop, most studies show that the hilar SST neurons are spared (Bendotti et al., 1993; Schwarzer et al., 1996; Vezzani et al., 1996; Simonato et al., 1998). These studies suggest that a functional consequence of this interneuronal loss is the development of spontaneous seizures, and that the hilar SST neurons or SST itself is important in preventing epileptogenesis. However, rate of epileptogenesis is not strictly correlated with degree of hilar neuron loss (e.g., Lahteinen et al., 2003), suggesting other factors are likely to contribute.

Interestingly, changes in the distribution of SSTimmunoreactive axons are also observed in the dentate of hippocampal tissue resected from patients with temporal lobe epilepsy (Mathern et al., 1995; Csaba et al., 2005). In control postmortem tissue or in hippocampus of patients without hippocampal sclerosis, a fine, diffuse network of SST-containing processes was observed in the outer two-thirds of the molecular layer. However, in patients with hippocampal sclerosis, accompanying the dramatic loss of SST-containing hilar neurons was a change in the SST-immunoreactive fibers in the dentate. SST-containing axons in the molecular layer were more numerous, thicker, longer, and more strongly immunoreactive for SST. Large terminal varicosities were also present. Similar, though less dramatic changes were observed in fibers in the hilus (Csaba et al., 2005). These results suggest changes in the remaining SST-containing neurons that are consistent with axon sprouting (Mathern et al., 1995). Thus, SST interneurons are activated by seizures, and SST expression initially increases. This is followed by death of the SST neurons, presumably via excitotoxic mechanisms. Interestingly, this group of neurons is selectively vulnerable to other insults as well. Ischemia, for example, also causes a reduction in SST-containing hilar neurons (Johansen et al., 1987; Bering et al., 1997), as does traumatic brain injury (Lowenstein et al., 1992). Both ethanol ingestion and withdrawal induces the death of these neurons (Andrade et al., 1992). Neurological diseases in which a reduction in hilar SST has been found to include Alzheimer’s disease (Chan-Palay, 1987) and an animal model of HIV dementia (Mitchell et al., 1999). What causes this subset of interneurons to be selectively vulnerable to multiple types of insults? Two major theories have been advanced; that these interneurons are ‘‘overstimulated’’ during hyperexcitability because they receive input from both mossy fibers and perforant path, or that they lack appropriate Ca2+ binding proteins so that they are vulnerable to Ca2+-induced excitotoxicity. The first hypothesis assumes that HIPP interneurons have a unique innervation pattern

274

compared with other, less vulnerable, hilar interneurons. However, the majority of anatomical data suggests that their main excitatory input comes exclusively from mossy fibers of DGCs (see above). In fact, when activity-dependent c-fos activation is measured following seizures, HIPP neurons are activated after DGCs (Le Gal La Salle, 1988; Peng and Houser, 2005), supporting the anatomical data. As for Ca2+ binding proteins, SST containing interneurons in the hilus do not contain parvalbumin, calretinin, or calbindin, although most CA1 and CA3 SST interneurons contain one of these Ca2+ binding proteins (Bouilleret et al., 2000). Thus, it is an appealing hypothesis that the underlying vulnerability of hilar SST interneurons vs. SST interneurons in the rest of the hippocampus is due to their lack of Ca2+ binding proteins. Indeed, a study using Ca2+ binding protein knockout mice suggests that this may be a contributing factor. This study found that in wild type mice, intrahippocampal injection of kainate resulted in a reduction of SST containing interneurons in CA1 to 43 and 29% of control 1 and 30 days after injection, respectively. However, a dramatic decrease in SST-containing interneurons after kainate injection was observed in mice with both parvalbumin and calbindin knocked out, but not with a knockout of parvalbumin alone. In the double knockout, only 12 and 6% of the CA1 SST-containing neurons remained 1 and 30 days after kainate injection, respectively. In mice with both parvalbumin and calretinin knocked out, a larger decrease in SST immunoreactivity in CA1 was observed 30 days but not 1 day following kainate injections (Bouilleret et al., 2000). Although changes in network excitability cannot be ruled out (although no differences in behavioral seizures was observed between the groups), this study suggests that knockout of calbindin in particular leads to increased susceptibility of SST interneurons to seizure-induced death. Thus, the lack of any of these Ca2+ binding proteins in hilar SST interneurons may contribute to their vulnerability to seizures and other insults.

Electrophysiological effects of SST in dentate gyrus On the basis of projection profile of SST-containing HIPP cells, in that their major target is the distal dendrites of DGCs, and that SST-containing terminals are often found adjacent to presumed lateral perforant path synapses, it seems likely that a major role for SST would be to regulate perforant path input onto DGCs. Unlike CA1, and to a lesser extent CA3, the electrophysiological effects of SST in the dentate gyrus have not been well characterized, especially in rat. In rat CA1 pyramidal neurons, the major effects of SST appear to be postsynaptic. SST and CST robustly enhance the M-current (Moore et al., 1988; Schweitzer et al., 1990; de Lecea et al., 1996), a subthreshold noninactivating K+ current that is important in setting the excitatory tone of neurons. SST has similar actions on the M-current in mouse CA1 pyramidal neurons (Qiu and Tallent, unpublished observations). SST and CST also augment another K+ current in these neurons, a voltage-independent leak current (Schweitzer et al., 1998, 2003). By acting on these two K+ currents, application of SST causes a several millivolt hyperpolarization of the resting membrane potential and a decrease in input resistance. Thus, incoming excitatory synaptic input would be less likely to trigger action potentials. SST also has presynaptic actions in rat CA1 pyramidal neurons, inhibiting EPSCs but not IPSCs (Tallent and Siggins, 1997). Interestingly, we have observed no presynaptic effect of SST in mouse CA1 pyramidal neurons, even though in side-by-side studies SST reduced EPSC amplitudes in rat CA1 pyramidal neurons (Qiu and Tallent, unpublished observations). In rat CA3 pyramidal neurons, SST also causes membrane hyperpolarization and reduces associational/commissural EPSCs presynaptically (Tallent and Siggins, 1999). In cultured immature rat hippocampal pyramidal neurons, SST presynaptically inhibits glutamate but not GABA release, and inhibits Ca2+ currents (Boehm and Betz, 1997). In vivo studies by de Lecea and colleagues (de Lecea et al., 1997a) have shown that CST depresses population spike amplitude in DGCs in

275

immature (P15) but not adult rats. Interestingly, CST is significantly expressed in dentate gyrus (in hilar interneurons) only at early stages of development but not in adults (see above; de Lecea et al., 1997a). These results suggest developmental changes in some SST receptor that mediates this CST effect. The developmental regulation of SST receptors in the dentate gyrus has not been extensively examined, although only SST4 mRNA expression is higher in immature than adult dentate gyrus (Thoss et al., 1995). The only detailed characterization of the electrophysiological effects of SST in dentate was done by our group in adult mouse (Baratta et al., 2002). In mouse DGCs, no action of SST on postsynaptic K+ currents was detected, and the peptide did not alter membrane potential. SST also did not affect the firing properties of DGCs. We also examined whether SST modulated excitatory synaptic input onto DGCs. We found that SST did not affect field EPSPs recorded in the outer molecular layer and evoked by low frequency stimulation the lateral perforant path (entorhinal cortical input). However, in spite of these negative findings on pre- and postsynaptic actions, SST robustly depressed long-term potentiation (LTP) at lateral perforant path synapses. Thus, although not acting on low-frequency input, SST depressed synaptic plasticity generated by high frequency input. This finding corroborates the speculated role of SST based on the close association between SST-containing terminals and lateral perforant path input. We further examined the mechanism by which SST inhibited LTP. Since LTP at this synapse is n-methyl-D-aspartate (NMDA)-receptor dependent, we determined whether SST could reduce pharmacologically isolated NMDA receptormediated synaptic responses. SST did not act on NMDA receptor-mediated EPSCs recorded in voltage-clamp, but modestly depressed NMDA receptor-mediated field EPSPs recorded extracellularly. Interestingly, the SST effect was blocked when recordings were made in the presence of oconotoxin, a specific blocker of N-type Ca2+ channels. These results indicated that SST does not directly inhibit NMDA receptors but acts on

Ca2+ channels activated during the NMDAreceptor mediated depolarization. We next recorded Ca2+ spikes generated in DGCs by depolarization. SST inhibited the Ca2+ spikes; this inhibition remained in the presence of the L-type Ca2+ blocker nifedipine and the T-type Ca2+ channel blocker nickel. However, SST did not inhibit Ca2+ spikes generated in the presence of o-contotoxin, indicating that the peptide was acting specifically on N-type Ca2+ channels. In the presence of o-conotoxin, no LTP could be generated at lateral perforant path synapses, suggesting this mechanism alone could account for SST inhibition of LTP (Baratta et al., 2002). We have also demonstrated that transgenic overexpression of CST prevents generation of LTP at lateral perforant path synapses (Tallent et al., 2005a). In adult wild type mice, CST is not expressed in dentate (de Lecea et al., 1997a, b). We generated a CST overexpressing transgenic mouse in which the transgene was driven by the neuronspecific enolase promoter. CST transgene expression was detected in the hippocampus, cortex, and reticular thalamus, with especially high expression in the dentate. These mice had normal baseline synaptic transmission at lateral perforant path synapses, with the exception of reduced pairedpulse potentiation at interstimulus intervals less than 100 ms. However, no LTP could be generated at these synapses, both in vivo and in vitro (Fig. 1). CA1 LTP in the mice was not significantly different from wild type 60 min following induction. A correlated deficit in spatial learning was present in the CST transgenic mice (Fig. 2), suggesting synaptic plasticity in dentate gyrus may be critical for this type of learning, as previous studies had suggested (Kauer et al., 1988; Moser et al., 1998; Nakao et al., 2002). These results and our results with exogenous SST (Baratta et al., 2002) indicate SST receptors may mediate important regulatory control over synaptic plasticity in dentate gyrus. It is interesting that SST appears to have a somewhat different physiological role in dentate gyrus than in CA1 and CA3 hippocampus. In cornu ammonis, high frequency activation of somatostatinergic

276

A 260

B Control

CST Transgenic

220 Baseline

% Control fEPSP slope

220

Post-HFT

0.25 mV 5ms

180

180 140 140

100

100

60

CST

60 -10

0

10

20 30 Time (min)

40

50

60

-10

0

10

20 30 Time (min)

40

50

60

C 260

Baseline

CST/Post-HFT

220

180

140 CST

100

60 -10

0

10

20 30 Time (min)

40

50

60

Fig. 1. Transgenic overexpression or exogenous application of CST prevents induction of LTP at lateral perforant path/dentate granule cell synapses. (A) Mean data from control and CST transgenic slices. Initial slope is plotted over time 10 min prior to 60 min following high frequency trains (double arrows). Reduced short-term potentiation and no long-term potentiation is seen in CST transgenics (open circles) relative to controls (filled triangles). Inset: Representative fEPSPs from a control and a transgenic mouse before (gray) and 60 min following two 1 s 100 Hz trains (black). No potentiation of the fEPSP is seen in slices from dentate of the CST transgenic mouse. Scale bar labels are the same for all subsequent traces. (B) Exogenous CST also reduces LTP. CST (1 mM; open circles) was superfused beginning 7 min prior to HFTs for a total of 8 min (black bar). A small decrease in the fEPSP slope was observed prior to HFTs (maximal inhibition was 1174% 7 min following the beginning of the superfusion). No significant potentiation is present 60 min following the trains when CST was superfused during the induction protocol. Inset: Representative fEPSPs before and 60 min following HFTs with CST superfusion. (C) Exogenous CST does not affect LTP maintenance. 1 mM CST was applied from 30 to 45 min (black bar) following HFTs (open circles). No significant effect on magnitude of LTP was observed. Inset: Representative baseline and 60 min post-train fEPSP from experiment where CST was applied 30–45 min following train. (Reprinted from Tallent et al., 2005a, copyright 2005, with permission from Elsevier.)

277

Fig. 2. CST transgenic mice show a deficit in spatial learning in the Barnes platform maze. In this mouse strain, overexpression of CST was largely restricted to dentate and reticular thalamus. The percent of mice in each group of males (top) and females (bottom) meeting the criterion for learning in this task (three errors or less on seven of eight consecutive trials) across the 40 days of testing. Both male and female CST transgenic mice showed impairment in this task as determined by significantly fewer mice in these groups learning this task. (Reprinted from Tallent et al., 2005a, copyright 2005, with permission from Elsevier.)

interneurons would release SST to hyperpolarize pyramidal neurons, decreasing their likelihood of firing during intense activation. Presynaptic inhibition of glutamate release by SST (in rat but not mouse) would act synergistically to inhibit high frequency activity. Thus, SST has robust effects on epileptiform activity in both CA1 and CA3 hippocampus, in both rats (Tallent and Siggins, 1999) and mice (Qiu et al., 2005). Interestingly, though, SST and CST has less robust actions on synaptic plasticity in CA1, reducing but not blocking LTP (Qiu et al., 2003; Tallent et al., 2005b). In dentate gyrus, SST has no observable actions on membrane potential or firing rate, thus this peptide would not regulate responsivity of DGCs to normal, low-frequency synaptic input. Perhaps this regulatory mechanism, important in pyramidal

neurons, would be less effective mechanism in DGCs, with their already quite hyperpolarized membrane potential (
78 to 80 mV). Certainly regulation of the M-current in particular would be an ineffective mechanism to control output of granule cells, since, unlike in CA1, this current would not be active at the resting membrane potential of DGCs (in fact we do not observe this current in granule cells, even though immunohistochemical studies suggest its component subunits are present: Cooper et al., 2000; Yus-Najera et al., 2003). We also did not observe presynaptic effects of SST or CST in mouse dentate gyrus (Baratta et al., 2002; Tallent et al., 2005b). Thus in spite of the rather dense localization of SST receptors on granule cell dendrites, they do not appear to be functionally important during standard synaptic transmission even when SST or CST is exogenously applied. However, the effect of these peptides on synaptic plasticity in dentate gyrus is profound. No LTP is generated at lateral perforant path synapses when nanomolar concentrations of SST is applied prior to induction. Likewise, transgenic overexpression of CST prevents generation of any perforant path LTP both in vivo and in vitro (Tallent et al., 2005b). Therefore, in spite of the similar projection pattern of somatostatinergic interneurons to distal dendrites of principle neurons, and of the similar localization of SST receptors on principle cell soma and dendrites in both regions, SST may have different physiological roles in the cornu ammonis vs. dentate gyrus. In dentate, release of SST during high frequency events would act to increase the threshold for inducing LTP. This mechanism may be important in increasing the signal to noise properties of synaptic input into the dentate, as well as in preventing invasion of seizures into the hippocampus.

Conclusions SST in the dentate gyrus is expressed primarily in a subset of interneurons of the hilus which coexpress GABA and project to the dendritic layer of DGCs. This places them in a position to modulate granule cell activity, and the striking plasticity of these neurons, as well as their vulnerability,

278

suggests that they could play an important role in normal and pathological conditions. Indeed, SST levels in different brain regions are altered in many diseases associated with cognitive deficits. It is tempting to speculate that these peptides are important in regulation of synaptic changes that underlie learning and memory. Although our studies in CST overexpressing mice support this hypothesis, further characterization of the role of SST and CST in cognitive processing is needed. List of abbreviations CREB CSF CST DGC EGFP GABA HIPP LTP NMDA PTX SST

cAMP response element binding protein cerebrospinal fluid cortistatin dentate granule cell enhanced green fluorescent protein gamma amino butyric acid hilar perforant path associated long-term potentiation n-methyl-D-aspartate pertussis toxin somatostatin

References Agren, H. and Lundqvist, G. (1984) Low levels of somatostatin in human CSF mark depressive episodes. Psychoneuroendocrinology, 9(3): 233–248. Amaral, D.G., Insausti, R. and Campbell, M.J. (1988) Distribution of somatostatin immunoreactivity in the human dentate gyrus. J. Neurosci., 8(9): 3306–3316. Andrade, J.P., Fernando, P.M., Madeira, M.D., Paula-Barbosa, M.M., Cadete-Leite, A. and Zimmer, J. (1992) Effects of chronic alcohol consumption and withdrawal on the somatostatin-immunoreactive neurons of the rat hippocampal dentate hilus. Hippocampus, 2(1): 65–71. Arancibia, S., Payet, O., Givalois, L. and Tapia-Arancibia, L. (2001) Acute stress and dexamethasone rapidly increase hippocampal somatostatin synthesis and release from the dentate gyrus hilus. Hippocampus, 11(4): 469–477. Aronin, N., Cooper, P.E., Lorenz, L.J., Bird, E.D., Sagar, S.M., Leeman, S.E. and Martin, J.B. (1983) Somatostatin is increased in the basal ganglia in Huntington disease. Ann. Neurol., 13(5): 519–526. Austin, J.E. and Buckmaster, P.S. (2004) Recurrent excitation of granule cells with basal dendrites and low interneuron

density and inhibitory postsynaptic current frequency in the dentate gyrus of macaque monkeys. J. Comp. Neurol., 476(3): 205–218. Baratta, M.V., Lamp, T. and Tallent, M.K. (2002) Somatostatin depresses long-term potentiation and Ca2+ signaling in mouse dentate gyrus. J. Neurophysiol., 88(6): 3078–3086. Bendotti, C., Vezzani, A., Tarizzo, G. and Samanin, R. (1993) Increased expression of GAP-43, somatostatin and neuropeptide Y mRNA in the hippocampus during development of hippocampal kindling in rats. Eur. J. Neurosci., 5(10): 1312–1320. Bereiter, D.A. (1997) Morphine and somatostatin analogue reduce c-fos expression in trigeminal subnucleus caudalis produced by corneal stimulation in the rat. Neuroscience, 77(3): 863–874. Bering, R., Draguhn, A., Diemer, N.H. and Johansen, F.F. (1997) Ischemia changes the coexpression of somatostatin and neuropeptide Y in hippocampal interneurons. Exp. Brain. Res., 115(3): 423–429. Blackmer, T., Larsen, E.C., Bartleson, C., Kowalchyk, J.A., Yoon, E.-J., Preininger, A.M., Alford, S., Hamm, H.E. and Martin, T.F.J. (2005) G protein [beta][gamma] directly regulates SNARE protein fusion machinery for secretory granule exocytosis. Nat. Neurosci., 8(4): 421–425. Blake, A.D. (2001) Somatostatin receptor subtype 1 (sst(1)) regulates intracellular 30 ,50 -cyclic adenosine monophosphate accumulation in rat embryonic cortical neurons: evidence with L-797,591, an sst(1)-subtype-selective nonpeptidyl agonist. Neuropharmacology, 40(4): 590–596. Boehm, S. and Betz, H. (1997) Somatostatin inhibits excitatory transmission at rat hippocampal synapses via presynaptic receptors. J. Neurosci., 17(11): 4066–4075. Bouilleret, V., Schwaller, B., Schurmans, S., Celio, M.R. and Fritschy, J.M. (2000) Neurodegenerative and morphogenic changes in a mouse model of temporal lobe epilepsy do not depend on the expression of the calcium-binding proteins parvalbumin, calbindin, or calretinin. Neuroscience, 97(1): 47–58. Brazeau, P., Vale, W., Burgus, R., Ling, N., Rivier, J. and Guillemin, R. (1972) Hypothalamic polypeptide that inhibits the secretion of immunoreactive pituitary growth hormone. Science, 129: 77–79. Bruno, J., Yu, Y., Song, J. and Berlowitz, M. (1992) Molecular cloning and functional expression of a novel brain somatostatin receptor. PNAS, 89: 11151–11155. Buckmaster, P.S. and Dudek, F.E. (1997) Neuron loss, granule cell axon reorganization, and functional changes in the dentate gyrus of epileptic kainate-treated rats. J. Comp. Neurol., 385(3): 385–404. Buckmaster, P.S., Kunkel, D.D., Robbins, R.J. and Schwartzkroin, P.A. (1994) Somatostatin-immunoreactivity in the hippocampus of mouse, rat, guinea pig, and rabbit. Hippocampus, 4(2): 167–180. Calbet, M., Guadano-Ferraz, A., Spier, A.D., Maj, M., Sutcliffe, J.G., Przewlocki, R. and de Lecea, L. (1999) Cortistatin and somatostatin mRNAs are differentially

279 regulated in response to kainate. Brain Res. Mol. Brain Res., 72(1): 55–64. Chan-Palay, V. (1987) Somatostatin immunoreactive neurons in the human hippocampus and cortex shown by immunogold/silver intensification on vibratome sections: coexistence with neuropeptide Y neurons, and effects in Alzheimer-type dementia. J. Comp. Neurol., 260(2): 201–223. Chesselet, M.F. and Reisine, T.D. (1983) Somatostatin regulates dopamine release in rat striatal slices and cat caudate nuclei. J. Neurosci., 3(1): 232–236. Chronwall, B.M., Chase, T.N. and O’donohue, T.L. (1984) Coexistence of neuropeptide Y and somatostatin in rat and human cortical and rat hypothalamic neurons. Neurosci. Lett., 52(3): 213–217. Connor, M., Ingram, S.L. and Christie, M.J. (1997) Cortistatin increase of a potassium conductance in rat locus coeruleus in vitro. Br. J. Pharmacol., 122(8): 1567–1572. Cook, J.L., Marcheselli, V., Alam, J., Deininger, P.L. and Bazan, N.G. (1998) Temporal changes in gene expression following cryogenic rat brain injury. Brain Res. Mol. Brain Res., 55(1): 9–19. Cooper, E.C., Aldape, K.D., Abosch, A., Barbaro, N.M., Berger, M.S., Peacock, W.S., Jan, Y.N. and Jan, L.Y. (2000) Colocalization and coassembly of two human brain M-type potassium channel subunits that are mutated in epilepsy. Proc. Natl. Acad. Sci. U.S.A., 97(9): 4914–4919. Cowles, R.A., Segura, B.J. and Mulholland, M.W. (2002) Regulation of carbachol-induced c-fos mRNA expression in AR42J cells by somatostatin receptor subtypes 1, 2, and 3. Pancreas, 25(3): 239–244. Csaba, Z., Pirker, S., Lelouvier, B., Simon, A., Videau, C., Epelbaum, J., Czech, T., Baumgartner, C., Sperk, G. and Dournaud, P. (2005) Somatostatin receptor type 2 undergoes plastic changes in the human epileptic dentate gyrus. J. Neuropathol. Exp. Neurol., 64(11): 956–969. Csaba, Z., Richichi, C., Bernard, V., Epelbaum, J., Vezzani, A. and Dournaud, P. (2004) Plasticity of somatostatin and somatostatin sst2A receptors in the rat dentate gyrus during kindling epileptogenesis. Eur. J. Neurosci., 19(9): 2531–2538. Davies, P., Katzman, R. and Terry, R.D. (1980) Reduced somatostatin-like immunoreactivity in cerebral cortex from cases of Alzheimer disease and Alzheimer senile dementia. Nature, 288(5788): 279–280. De Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer, D.D. (1989) Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy. Brain Res., 495(2): 387–395. De Lanerolle, N.C., Kim, J.H., Williamson, A., Spencer, S.S., Zaveri, H.P., Eid, T. and Spencer, D.D. (2003) A retrospective analysis of hippocampal pathology in human temporal lobe epilepsy: evidence for distinctive patient subcategories. Epilepsia, 44(5): 677–687. de Lecea, L., Criado, J.R., Prospero-Garcia, O., Gautivik, K.M., Schweitzer, P., Danielson, P.E., Dunlop, C.L.M., Siggins, G.R., Henriksen, S.J. and Sutcliffe, J.G. (1996) A cortical neuropeptide with neuronal depressant and sleepmodulating properties. Nature, 381: 242–245.

de Lecea, L., Del Rio, J.A., Criado, J.R., Alcantara, S., Morales, M., Danielson, P.E., Henriksen, S.J., Soriano, E. and Sutcliffe, J.G. (1997a) Cortistatin is expressed in a distinct subset of cortical interneurons. J. Neurosci., 17(15): 5868–5880. de Lecea, L., Ruiz-Lozano, P., Danielson, P.E., Peelle-Kirley, J., Foye, P.E., Frankel, W.N. and Sutcliffe, J.G. (1997b) Cloning, mRNA expression, and chromosomal mapping of mouse and human preprocortistatin. Genomics, 42(3): 499–506. Dickson, S.L., Viltart, O., Bailey, A.R. and Leng, G. (1997) Attenuation of the growth hormone secretagogue induction of Fos protein in the rat arcuate nucleus by central somatostatin action. Neuroendocrinology, 66(3): 188–194. Dournaud, P., Cervera-Pierot, P., Hirsch, E., Javoy-Agid, F., Kordon, C., Agid, Y. and Epelbaum, J. (1994) Somatostatin messenger RNA-containing neurons in Alzheimer’s disease: an in situ hybridization study in hippocampus, parahippocampal cortex and frontal cortex. Neuroscience, 61(4): 755–764. Dournaud, P., Gu, Y.Z., Schonbrunn, A., Mazella, J., Tannenbaum, G.S. and Beaudet, A. (1996a) Localization of the somatostatin receptor SST2A in rat brain using a specific antipeptide antibody. J. Neurosci., 16(14): 4468–4478. Dournaud, P., Jazat-Poindessous, F., Slama, A., Lamour, Y. and Epelbaum, J. (1996b) Correlations between water maze performance and cortical somatostatin mRNA and highaffinity binding sites during ageing in rats. Eur. J. Neurosci., 8(3): 476–485. Edvinsson, L., Minthon, L., Ekman, R. and Gustafson, L. (1993) Neuropeptides in cerebrospinal fluid of patients with Alzheimer’s disease and dementia with frontotemporal lobe degeneration. Dementia, 4(3–4): 167–171. Epelbaum, J., Ruberg, M., Moyse, E., Javoy-Agid, F., Dubois, B. and Agid, Y. (1983) Somatostatin and dementia in Parkinson’s disease. Brain Res., 278(1–2): 376–379. Figueredo-Cardenas, G., Morello, M., Sancesario, G., Bernardi, G. and Reiner, A. (1996) Colocalization of somatostatin, neuropeptide Y, neuronal nitric oxide synthase and NADPH-diaphorase in striatal interneurons in rats. Brain Res., 735(2): 317–324. Florio, T., Yao, H., Carey, K.D., Dillon, T.J. and Stork, P.J. (1999) Somatostatin activation of mitogen-activated protein kinase via somatostatin receptor 1 (SSTR1). Mol. Endocrinol., 13(1): 24–37. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6(4): 347–470. Gabriel, S.M., Davidson, M., Haroutunian, V., Powchik, P., Bierer, L.M., Purohit, D.P., Perl, D.P. and Davis, K.L. (1996) Neuropeptide deficits in schizophrenia vs. Alzheimer’s disease cerebral cortex. Biol. Psychiatry, 39(2): 82–91. Gerachshenko, T., Blackmer, T., Yoon, E.-J., Bartleson, C., Hamm, H.E. and Alford, S. (2005) G[beta][gamma] acts at the C terminus of SNAP-25 to mediate presynaptic inhibition. Nat. Neurosci., 8(5): 597–605. Grouselle, D., Winsky-Sommerer, R., David, J.P., Delacourte, A., Dournaud, P. and Epelbaum, J. (1998) Loss of

280 somatostatin-like immunoreactivity in the frontal cortex of Alzheimer patients carrying the apolipoprotein epsilon 4 allele. Neurosci. Lett., 255(1): 21–24. Halabisky, B., Shen, F., Huguenard, J.R. and Prince, D.A. (2006) Electrophysiological classification of somatostatinpositive interneurons in mouse sensorimotor cortex. J. Neurophysiol., 96(2): 834–845. Handel, M., Schulz, S., Stanarius, A., Schreff, M., ErdtmannVourliotis, M., Schmidt, H., Wolf, G. and Hollt, V. (1999) Selective targeting of somatostatin receptor 3 to neuronal cilia. Neuroscience, 89(3): 909–926. Hashimoto, T. and Obata, K. (1991) Induction of somatostatin by kainic acid in pyramidal and granule cells of the rat hippocampus. Neurosci. Res., 12(4): 514–527. Hayashi, M., Yamashita, A. and Shimizu, K. (1997) Somatostatin and brain-derived neurotrophic factor mRNA expression in the primate brain: decreased levels of mRNAs during aging. Brain Res., 749(2): 283–289. Hervieu, G. and Emson, P.C. (1998) The localization of somatostatin receptor 1 (sst1) immunoreactivity in the rat brain using an N-terminal specific antibody. Neuroscience, 85(4): 1263–1284. Holm, I.E., Geneser, F.A. and Zimmer, J. (1992) Somatostatinand neuropeptide Y-like immunoreactivity in the dentate area, hippocampus, and subiculum of the domestic pig. J. Comp. Neurol., 322(3): 390–408. Huang, S.M., Mouri, A., Kokubo, H., Nakajima, R., Suemoto, T., Higuchi, M., Staufenbiel, M., Noda, Y., Yamaguchi, H., Nabeshima, T., Saido, T.C. and Iwata, N. (2006) Neprilysinsensitive synapse-associated Abeta oligomers impair neuronal plasticity and cognitive function. J. Biol. Chem., 281(26): 17941–17951. Ikeda, S.R. and Schofield, G.G. (1989) Somatostatin blocks a calcium current in rat sympathetic ganglion neurons. J. Physiol., 409: 221–240. Inoue, M., Nakajima, S. and Nakajima, Y. (1988) Somatostatin induces an inward rectification in rat locus coeruleus neurons through a pertussis toxin sensitive mechanism. J. Physiol., 407: 177–198. Ishihara, S., Hassan, S., Kinoshita, Y., Moriyama, N., Fukuda, R., Maekawa, T., Okada, A. and Chiba, T. (1999) Growth inhibitory effects of somatostatin on human leukemia cell lines mediated by somatostatin receptor subtype 1. Peptides, 20(3): 313–318. Iwata, N., Takaki, Y., Fukami, S., Tsubuki, S. and Saido, T.C. (2002) Region-specific reduction of A beta-degrading endopeptidase, neprilysin, in mouse hippocampus upon aging. J. Neurosci. Res., 70(3): 493–500. Jinno, S. and Kosaka, T. (2000) Colocalization of parvalbumin and somatostatin-like immunoreactivity in the mouse hippocampus: quantitative analysis with optical dissector. J. Comp. Neurol., 428(3): 377–388. Johansen, F.F., Zimmer, J. and Diemer, N.H. (1987) Early loss of somatostatin neurons in dentate hilus after cerebral ischemia in the rat precedes CA-1 pyramidal cell loss. Acta Neuropathol., 73(2): 110–114.

Kauer, J.A., Malenka, R.C. and Nicoll, R.A. (1988) A persistent postsynaptic modification mediates long-term potentiation in the hippocampus. Neuron, 1(10): 911–917. Kimura, N., Tomizawa, S., Arai, K.N. and Osamura, R.Y. (2001) Characterization of 50 -flanking region of rat somatostatin receptor sst2 gene: transcriptional regulatory elements and activation by Pitx1 and estrogen. Endocrinology, 142(4): 1427–1441. Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition of dentate granule cells in a model of temporal lobe epilepsy. J. Neurosci., 23(6): 2440–2452. Kong, H., Depaoli, A.M., Breder, C.D., Yasuda, K., Bell, G.I. and Reisine, T. (1994) Differential expression of messenger RNAs for somatostatin receptor subtypes SSTR1, SSTR2, and SSTR3 in adult rat brain: analysis by RNA blotting and in situ hybridization histochemistry. Neuroscience, 59(1): 175–184. Kosaka, T., Wu, J.-Y. and Benoit, R. (1988) GABAergic neurons containing somatostatin-like immunoreactivity in the rat hippocampus and dentate gyrus. Exp. Brain Res., 71: 388–398. Kragh, J., Tonder, N., Finsen, B.R., Zimmer, J. and Bolwig, T.G. (1994) Repeated electroconvulsive shocks cause transient changes in rat hippocampal somatostatin and neuropeptide Y immunoreactivity and mRNA in situ hybridization signals. Exp. Brain Res., 98(2): 305–313. Kumar, U. (2005) Expression of somatostatin receptor subtypes (SSTR1-5) in Alzheimer’s disease brain: an immunohistochemical analysis. Neuroscience, 134(2): 525–538. Lahlou, H., Guillermet, J., Hortala, M., Vernejoul, F., Pyronnet, S., Bousquet, C. and Susini, C. (2004) Molecular signaling of somatostatin receptors. Ann. N.Y. Acad. Sci., 1014(1): 121–131. Lahlou, H., Saint-Laurent, N., Esteve, J.P., Eychene, A., Pradayrol, L., Pyronnet, S. and Susini, C. (2003) sst2 Somatostatin receptor inhibits cell proliferation through Ras-, Rap1-, and B-Raf-dependent ERK2 activation. J. Biol. Chem., 278(41): 39356–39371. Lahteinen, S., Pitkanen, A., Koponen, E., Saarelainen, T. and Castren, E. (2003) Exacerbated status epilepticus and acute cell loss, but no changes in epileptogenesis, in mice with increased brain-derived neurotrophic factor signaling. Neuroscience, 122(4): 1081–1092. Le Gal La Salle, G. (1988) Long-lasting and sequential increase of c-fos oncoprotein expression in kainic acid-induced status epilepticus. Neurosci. Lett., 88(2): 127–130. Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Afferent and efferent synaptic connections of somatostatinimmunoreactive neurons in the rat fascia dentata. J. Comp. Neurol., 295(1): 111–122. Leroux, P., Weissmann, D., Pujol, J.-F. and Vaudry, H. (1993) Quantitative autoradiography of somatostatin receptors in the rat limbic system. J. Comp. Neurol., 331: 389–401. Lowenstein, D.H., Thomas, M.J., Smith, D.H. and Mcintosh, T.K. (1992) Selective vulnerability of dentate hilar neurons following traumatic brain injury: a potential mechanistic link

281 between head trauma and disorders of the hippocampus. J. Neurosci., 12(12): 4846–4853. Lubke, J., Frotscher, M. and Spruston, N. (1998) Specialized electrophysiological properties of anatomically identified neurons in the hilar region of the rat fascia dentata. J. Neurophysiol., 79(3): 1518–1534. Manfridi, A., Forloni, G.L., Vezzani, A., Fodritto, F. and De Simoni, M.G. (1991) Functional and histological consequences of quinolinic and kainic acid-induced seizures on hippocampal somatostatin neurons. Neuroscience, 41(1): 127–135. Marti, M., Bregola, G., Morari, M., Gemignani, A. and Simonato, M. (2000) Somatostatin release in the hippocampus in the kindling model of epilepsy: a microdialysis study. J. Neurochem., 74(6): 2497–2503. Mathern, G.W., Babb, T.L., Pretorius, J.K. and Leite, J.P. (1995) Reactive synaptogenesis and neuron densities for neuropeptide Y, somatostatin, and glutamate decarboxylase immunoreactivity in the epileptogenic human fascia dentata. J. Neurosci., 15(5 pt 2): 3990–4004. Mazarati, A.M. and Telegdy, G. (1992) Effects of somatostatin and anti-somatostatin serum on picrotoxin-kindled seizures. Neuropharmacology, 31(8): 793–797. Mcdonald, A.J. (1989) Coexistence of somatostatin with neuropeptide Y, but not with cholecystokinin or vasoactive intestinal peptide, in neurons of the rat amygdala. Brain Res., 500(1–2): 37–45. Meriney, S.D., Gray, D.B. and Pilar, G.R. (1994) Somatostatininduced inhibition of neuronal calcium current modulated by cGMP-dependent protein kinase. Nature, 369(6478): 336–339. Mihara, S., North, R.A. and Surprenant, A. (1987) Somatostatin increases an inwardly rectifying potassium conductance in guinea pig submucosa plexus neurones. J. Physiol., 390: 335–355. Mikkelsen, J.D. and Woldbye, D.P. (2006) Accumulated increase in neuropeptide Y and somatostatin gene expression of the rat in response to repeated electroconvulsive stimulation. J. Psychiatr. Res., 40(2): 153–159. Milner, T.A. and Bacon, C.E. (1989) Ultrastructural localization of somatostatin-like immunoreactivity in the rat dentate gyrus. J. Comp. Neurol., 290(4): 544–560. Minthon, L., Edvinsson, L. and Gustafson, L. (1997) Somatostatin and neuropeptide Y in cerebrospinal fluid: correlations with severity of disease and clinical signs in Alzheimer’s disease and frontotemporal dementia. Dement. Geriatr. Cogn. Disord., 8(4): 232–239. Mitchell, T.W., Buckmaster, P.S., Hoover, E.A., Whalen, L.R. and Dudek, F.E. (1999) Neuron loss and axon reorganization in the dentate gyrus of cats infected with the feline immunodeficiency virus. J. Comp. Neurol., 411(4): 563–577. Moller, L.N., Stidsen, C.E., Hartmann, B. and Holst, J.J. (2003) Somatostatin receptors. Biochim. Biophys. Acta, 1616(1): 1–84. Montminy, M., Brindle, P., Arias, J., Ferreri, K. and Armstrong, R. (1996) Regulation of somatostatin gene transcription by cyclic adenosine monophosphate. Metabolism, 45(8 Suppl 1): 4–7.

Montminy, M.R. and Bilezikjian, L.M. (1987) Binding of a nuclear protein to the cyclic-AMP response element of the somatostatin gene. Nature, 328(6126): 175–178. Moore, S.D., Madamba, S.G., Joels, M. and Siggins, G.R. (1988) Somatostatin augments the M-current in hippocampal neurons. Science, 239(4837): 278–280. Moser, E.I., Krobert, K.A., Moser, M.B. and Morris, R.G. (1998) Impaired spatial learning after saturation of long-term potentiation. Science, 281(5385): 2038–2042. Mott, D.D., Turner, D.A., Okazaki, M.M. and Lewis, D.V. (1997) Interneurons of the dentate-hilus border of the rat dentate gyrus: morphological and electrophysiological heterogeneity. J. Neurosci., 17(11): 3990–4005. Nakao, K., Ikegaya, Y., Yamada, M.K., Nishiyama, N. and Matsuki, N. (2002) Hippocampal long-term depression as an index of spatial working memory. Eur. J. Neurosci., 16(5): 970–974. Nakatani, N., Hattori, E., Ohnishi, T., Dean, B., Iwayama, Y., Matsumoto, I., Kato, T., Osumi, N., Higuchi, T., Niwa, S.I. and Yoshikawa, T. (2006) Genome-wide expression analysis detects eight genes with robust alterations specific to bipolar I disorder: relevance to neuronal network perturbation. Hum. Mol. Genet., 15(12): 1949–1962. Nemeroff, C.B., Youngblood, W.W., Manberg, P.J., Prange Jr., A.J. and Kizer, J.S. (1983) Regional brain concentrations of neuropeptides in Huntington’s chorea and schizophrenia. Science, 221(4614): 972–975. Neves, S.R., Ram, P.T. and Iyengar, R. (2002) G protein pathways. Science, 296(5573): 1636–1639. Oliva Jr., A.A., Jiang, M., Lam, T., Smith, K.L. and Swann, J.W. (2000) Novel hippocampal interneuronal subtypes identified using transgenic mice that express green fluorescent protein in GABAergic interneurons. J. Neurosci., 20(9): 3354–3368. Passarelli, F. and Orzi, F. (1993) Somatostatin mRNA in the hippocampal formation following electroconvulsive shock in the rat. Neurosci. Lett., 153(2): 197–201. Peng, Z. and Houser, C.R. (2005) Temporal patterns of fos expression in the dentate gyrus after spontaneous seizures in a mouse model of temporal lobe epilepsy. J. Neurosci., 25(31): 7210–7220. Perez, J., Vezzani, A., Civenni, G., Tutka, P., Rizzi, M., Schuepbach, E. and Hoyer, D. (1995) Functional effects of DPhe-c[Cys-Tyr-D-Trp-Lys-Val-Cys]-Trp-NH2 and differential changes in somatostatin receptor messenger RNAs, binding sites and somatostatin release in kainic acid-treated rats. Neuroscience, 65(4): 1087–1097. Piwko, C., Thoss, V.S., Samanin, R., Hoyer, D. and Vezzani, A. (1996) Status of somatostatin receptor messenger RNAs and binding sites in rat brain during kindling epileptogenesis. Neuroscience, 75(3): 857–868. Pola, S., Cattaneo, M.G. and Vicentini, L.M. (2003) Antimigratory and anti-invasive effect of somatostatin in human neuroblastoma cells: involvement of Rac and MAP kinase activity. J. Biol. Chem., 278(42): 40601–40606.

282 Powers, A.C., Tedeschi, F., Wright, K.E., Chan, J.S. and Habener, J.F. (1989) Somatostatin gene expression in pancreatic islet cells is directed by cell-specific DNA control elements and DNA-binding proteins. J. Biol. Chem., 264(17): 10048–10056. Qiu, C., Suntheimer, R.J. and Tallent, M.K. (2003) Somatostatin depresses CA1 long term potentiation. Abstract Viewer/ Itinerary Planner. Washington, DC: Society for Neuroscience, Program No. 889.1. Qiu, C., Suzuki, C., Zeyda, T., Hochgeschwender, U., de Lecea, L. and Tallent, M.K. (2005) Increased seizure severity and reduced somatostatin effects in hippocampus of somatostatin receptor subtype 4 (SST4) knockout mice. 2005 Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience, Program No. 607.11. Ramirez, J.L., Gracia-Navarro, F., Garcia-Navarro, S., Torronteras, R., Malagon, M.M. and Castano, J.P. (2002) Somatostatin stimulates GH secretion in two porcine somatotrope subpopulations through a cAMP-dependent pathway. Endocrinology, 143(3): 889–897. Ramos, B., Baglietto-Vargas, D., Del Rio, J.C., MorenoGonzalez, I., Santa-Maria, C., Jimenez, S., Caballero, C., Lopez-Tellez, J.F., Khan, Z.U., Ruano, D., Gutierrez, A. and Vitorica, J. (2006) Early neuropathology of somatostatin/NPY GABAergic cells in the hippocampus of a PS1  APP transgenic model of Alzheimer’s disease. Neurobiol. Aging, 27(11): 1658–1672. Rapp, P.R. and Amaral, D.G. (1988) The time of origin of somatostatin-immunoreactive neurons in the rat hippocampal formation. Brain Res., 469(1–2): 231–239. Rauca, C., Schafer, K. and Hollt, V. (1999) Effects of somatostatin, octreotide and cortistatin on ischaemic neuronal damage following permanent middle cerebral artery occlusion in the rat. Naunyn. Schmiedebergs Arch. Pharmacol., 360(6): 633–638. Raynor, K. and Reisine, T. (1992) Differential coupling of somatostatin1 receptors to adenylyl cyclase in the rat striatum vs. the pituitary and other regions of the rat brain. J. Pharmacol. Exp. Ther., 260: 841–848. Reisine, T. (1995) Somatostatin. Cell Mol. Neurobiol., 15(6): 597–614. Robbins, R.J., Brines, M.L., Kim, J.H., Adrian, T., De Lanerolle, N., Welsh, S. and Spencer, D.D. (1991) A selective loss of somatostatin in the hippocampus of patients with temporal lobe epilepsy. Ann. Neurol., 29(3): 325–332. Roberts, G.W., Crow, T.J. and Polak, J.M. (1985) Location of neuronal tangles in somatostatin neurons in Alzheimer’s disease. Nature, 314: 92–94. Robishaw, J.D. and Berlot, C.H. (2004) Translating G protein subunit diversity into functional specificity. Curr. Opin. Cell Biol., 16(2): 206–209. Saito, T., Iwata, N., Tsubuki, S., Takaki, Y., Takano, J., Huang, S.M., Suemoto, T., Higuchi, M. and Saido, T.C. (2005) Somatostatin regulates brain amyloid beta peptide Abeta42 through modulation of proteolytic degradation. Nat. Med., 11(4): 434–439. Sakai, K., Maeda, K., Chihara, K. and Kaneda, H. (1995) Increases in cortical neuropeptide Y and somatostatin

concentrations following haloperidol-depot treatment in rats. Neuropeptides, 29(3): 157–161. Scharfman, H.E. (1992) Differentiation of rat dentate neurons by morphology and electrophysiology in hippocampal slices: granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells. Epilepsy Res. Suppl., 7: 93–109. Schindler, M., Humphrey, P.P., Lohrke, S. and Friauf, E. (1999) Immunohistochemical localization of the somatostatin sst2(b) receptor splice variant in the rat central nervous system. Neuroscience, 90(3): 859–874. Schreff, M., Schulz, S., Handel, M., Keilhoff, G., Braun, H., Pereira, G., Klutzny, M., Schmidt, H., Wolf, G. and Hollt, V. (2000) Distribution, targeting, and internalization of the sst4 somatostatin receptor in rat brain. J. Neurosci., 20(10): 3785–3797. Schulz, S., Handel, M., Schreff, M., Schmidt, H. and Hollt, V. (2000) Localization of five somatostatin receptors in the rat central nervous system using subtype-specific antibodies. J. Physiol. Paris, 94(3–4): 259–264. Schwarzer, C., Sperk, G., Samanin, R., Rizzi, M., Gariboldi, M. and Vezzani, A. (1996) Neuropeptides-immunoreactivity and their mRNA expression in kindling: functional implications for limbic epileptogenesis. Brain Res. Brain Res. Rev., 22(1): 27–50. Schweitzer, P., Madamba, S. and Siggins, G.R. (1990) Arachidonic acid metabolites as mediators of somatostatininduced increase of neuronal M-current. Nature, 346(6283): 464–467. Schweitzer, P., Madamba, S.G. and Siggins, G.R. (1998) Somatostatin increases a voltage-insensitive K+ conductance in rat CA1 hippocampal neurons. J. Neurophysiol., 79(3): 1230–1238. Schweitzer, P., Madamba, S.G. and Siggins, G.R. (2003) The sleep-modulating peptide cortistatin augments the h-current in hippocampal neurons. J. Neurosci., 23(34): 10884–10891. Sharma, R.P., Bissette, G., Janicak, P.G., Davis, J.M. and Nemeroff, C.B. (1995) Elevation of CSF somatostatin concentrations in mania. Am. J. Psychiatry, 152(12): 1807–1809. Siehler, S. and Hoyer, D. (1999) Characterisation of human recombinant somatostatin receptors. 4. Modulation of phospholipase C activity. Naunyn Schmiedebergs Arch. Pharmacol., 360(5): 522–532. Simonato, M., Bregola, G., Beani, L., Vezzani, A., Sala, R., Raiteri, M. and Bonanno, G. (1998) Time- and region-specific variations in somatostatin release following amygdala kindling in the rat. J. Neurochem., 70(1): 252–259. Sloviter, R.S. (1987) Decreased hippocampal inhibition and a selective loss of interneurons in experimental epilepsy. Science, 235(4784): 73–76. Sodickson, D.L. and Bean, B.P. (1998) Neurotransmitter activation of inwardly rectifying potassium current in dissociated hippocampal CA3 neurons: interactions among multiple receptors. J. Neurosci., 18(20): 8153–8162. Sokal, D.M. and Large, C.H. (2001) The effects of GABA(B) receptor activation on spontaneous and evoked activity in the dentate gyrus of kainic acid-treated rats. Neuropharmacology, 40(2): 193–202.

283 Straiker, A.J., Borden, C.R. and Sullivan, J.M. (2002) G-protein alpha subunit isoforms couple differentially to receptors that mediate presynaptic inhibition at rat hippocampal synapses. J. Neurosci., 22(7): 2460–2468. Stroh, T., Kreienkamp, H.J. and Beaudet, A. (1999) Immunohistochemical distribution of the somatostatin receptor subtype 5 in the adult rat brain: predominant expression in the basal forebrain. J. Comp. Neurol., 412(1): 69–82. Sutula, T. and Steward, O. (1986) Quantitative analysis of synaptic potentiation during kindling of the perforant path. J. Neurophysiol., 56(3): 732–746. Sutula, T. and Steward, O. (1987) Facilitation of kindling by prior induction of long-term potentiation in the perforant path. Brain Res., 420(1): 109–117. Swanson, T.H., Sperling, M.R. and O’connor, M.J. (1998) Strong paired pulse depression of dentate granule cells in slices from patients with temporal lobe epilepsy. J. Neural Transm., 105(6–7): 613–625. Tallent, M.K., Fabre, V., Qiu, C., Calbet, M., Lamp, T., Baratta, M.V., Suzuki, C., Levy, C.L., Siggins, G.R. and Henriksen, S.J. (2005a) Cortistatin overexpression in transgenic mice produces deficits in synaptic plasticity and learning. Mol. Cell. Neurosci., 30(3): 465–475. Tallent, M.K., Patterson, C., Hon, B., Qiu, C., Zeyda, T., Hochgeschwender, U. and de Lecea, L. (2005b) Somatostatin receptor subtype 3 is critical for hippocampal-dependent memory: studies in SST3 receptor knockout mice. Abstract Viewer/Itinerary Planner. Washington, DC: Society for Neuroscience, Program No. 607.10. Tallent, M.K. and Siggins, G.R. (1997) Somatostatin depresses excitatory but not inhibitory neurotransmission in rat CA1 hippocampus. J. Neurophysiol., 78(6): 3008–3018. Tallent, M.K. and Siggins, G.R. (1999) Somatostatin acts in CA1 and CA3 to reduce hippocampal epileptiform activity. J. Neurophysiol., 81(4): 1626–1635. Tentler, J.J., Hadcock, J.R. and Gutierrez-Hartmann, A. (1997) Somatostatin acts by inhibiting the cyclic 30 ,50 -adenosine monophosphate (cAMP)/protein kinase A pathway, cAMP response element-binding protein (CREB) phosphorylation, and CREB transcription potency. Mol. Endocrinol., 11(7): 859–866. Thoss, V.S., Duc, D. and Hoyer, D. (1996) Somatostatin receptors in the developing rat brain. Eur. J. Pharmacol., 297(1–2): 145–155. Thoss, V.S., Perez, J., Duc, D. and Hoyer, D. (1995) Embryonic and postnatal mRNA distribution of five somatostatin receptor subtypes in the rat brain. Neuropharmacology, 34(12): 1673–1688. Todisco, A., Campbell, V., Dickinson, C.J., Delvalle, J. and Yamada, T. (1994) Molecular basis for somatostatin action: inhibition of c-fos expression and AP-1 binding. Am. J. Physiol. Gastrointest. Liver Physiol., 267(2): G245–G253. Todisco, A., Seva, C., Takeuchi, Y., Dickinson, C.J. and Yamada, T. (1995) Somatostatin inhibits AP-1 function via multiple protein phosphatases. Am. J. Physiol., 269(1 Pt 1): G160–G166.

Tomidokoro, Y., Harigaya, Y., Matsubara, E., Ikeda, M., Kawarabayashi, T., Okamoto, K. and Shoji, M. (2000) Impaired neurotransmitter systems by Abeta amyloidosis in APPsw transgenic mice overexpressing amyloid beta protein precursor. Neurosci. Lett., 292(3): 155–158. Vecsei, L., Csala, B., Widerlov, E., Ekman, R., Czopf, J. and Palffy, G. (1990) Lumbar cerebrospinal fluid concentrations of somatostatin and neuropeptide Y in multiple sclerosis. Brain Res. Bull., 25(3): 411–413. Vela, J., Gutierrez, A., Vitorica, J. and Ruano, D. (2003) Rat hippocampal GABAergic molecular markers are differentially affected by ageing. J. Neurochem., 85(2): 368–377. Velimirovic, B.M., Koyano, K., Nakajima, S. and Nakajima, Y. (1995) Opposing mechanisms of regulation of a G-protein-coupled inward rectifier K+ channel in rat brain neurons. Proc. Natl. Acad. Sci. U.S.A., 92(5): 1590–1594. Vezzani, A., Monno, A., Rizzi, M., Galli, A., Barrios, M. and Samanin, R. (1992) Somatostatin release is enhanced in the hippocampus of partially and fully kindled rats. Neuroscience, 51(1): 41–46. Vezzani, A., Rizzi, M., Conti, M. and Samanin, R. (2000) Modulatory role of neuropeptides in seizures induced in rats by stimulation of glutamate receptors. J. Nutr., 130(4S Suppl): 1046S–1048S. Vezzani, A., Schwarzer, C., Lothman, E.W., Williamson, J. and Sperk, G. (1996) Functional changes in somatostatin and neuropeptide Y containing neurons in the rat hippocampus in chronic models of limbic seizures. Epilepsy Res., 26(1): 267–279. Vezzani, A., Serafini, R., Stasi, M.A., Vigano, G., Rizzi, M. and Samanin, R. (1991) A peptidase-resistant cyclic octapeptide analogue of somatostatin (SMS 201-995) modulates seizures induced by quinolic acid and kainic acid differently in the rat hippocampus. Neuropharmacology, 30(4): 345–352. Wang, H.-L., Dichter, M.A. and Reisine, T. (1990a) Lack of cross-desensitization of somatostatin-14 and somatostatin-28 receptors coupled to potassium channels in rat neocortical neurons. Mol. Pharm., 38(3): 357–361. Wang, H.-L., Reisine, T. and Dichter, M.A. (1990b) Somatostatin-14 and somatostatin-28 inhibit calcium currents in rat neocortical neurons. Neuroscience, 38(2): 335–342. Wasterlain, C.G., Mazarati, A.M., Shirasaka, Y., Thompson, K.W., Penix, L., Liu, H. and Katsumori, H. (1999) Seizureinduced hippocampal damage and chronic epilepsy: a hebbian theory of epileptogenesis. Adv. Neurol., 79: 829–843. Winsky-Sommerer, R., Spier, A.D., Fabre, V., de Lecea, L. and Criado, J.R. (2004) Overexpression of the human beta-amyloid precursor protein downregulates cortistatin mRNA in PDAPP mice. Brain Res., 1023(1): 157–162. Yoshitomi, H., Fujii, Y., Miyazaki, M., Nakajima, N., Inagaki, N. and Seino, S. (1997) Involvement of MAP kinase and c-fos signaling in the inhibition of cell growth by somatostatin. Am. J. Physiol., 272(5 Pt 1): E769–E774. Yasojima, K., Akiyama, H., Mcgeer, E.G. and Mcgeer, P.L. (2001) Reduced neprilysin in high plaque areas of Alzheimer brain: a possible relationship to deficient degradation of betaamyloid peptide. Neurosci. Lett., 297(2): 97–100.

284 Yus-Najera, E., Munoz, A., Salvador, N., Jensen, B.S., Rasmussen, H.B., Defelipe, J. and Villarroel, A. (2003) Localization of KCNQ5 in the normal and epileptic human temporal neocortex and hippocampal formation. Neuroscience, 120(2): 353–364. Zhang, Z.J., Lappi, D.A., Wrenn, C.C., Milner, T.A. and Wiley, R.G. (1998) Selective lesion of the cholinergic basal forebrain causes a loss of cortical

neuropeptide Y and somatostatin neurons. Brain Res., 800(2): 198–206. Zheng, H., Bailey, A., Jiang, M.H., Honda, K., Chen, H.Y., Trumbauer, M.E., Van Der Ploeg, L.H., Schaeffer, J.M., Leng, G. and Smith, R.G. (1997) Somatostatin receptor subtype 2 knockout mice are refractory to growth hormonenegative feedback on arcuate neurons. Mol. Endocrinol., 11(11): 1709–1717.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 17

Neuropeptide Y in the dentate gyrus Gu¨nther Sperk1,, Trevor Hamilton2 and William F. Colmers2, 2

1 Department of Pharmacology, Medical University Innsbruck, Peter-Mayr-Str. 1a, 6020 Innsbruck, Austria Department of Pharmacology, University of Alberta, 9– 36 Medical Sciences Building, Edmonton, AB, T6G 2H7, Canada

Abstract: Neuropeptide Y (NPY) is contained in at least four types of GABAergic interneurons in the dentate gyrus, many of which also contain somatostatin and give rise to the dense NPY innervation of the dentate outer molecular layer. In humans but not rats, minute amounts of NPY are also normally expressed in dentate granule cells, while seizure activity in rats induces robust NPY expression in granule cells. Y1 and Y2 receptors are the most abundant NPY receptors expressed in the dentate gyrus. Y1 receptors are postsynaptic receptors, primarily located on granule cell dendrites in the molecular layer and some interneurons, while Y2 receptors are presynaptic receptors mediating inhibition of glutamate release, and potentially that of NPY and GABA depending on their presynaptic localization, and may also be expressed on some hilar interneurons. In humans, monkeys and mice, Y2 receptors are also present on mossy fibers, but not in most rat species, though functional evidence suggests their presence. Hilar interneurons containing NPY degenerate in temporal lobe epilepsy and in Alzheimer’s disease and reduced levels of NPY in dentate hilus are associated with depression. By activating Y1 receptors, NPY also exerts powerful neuroproliferative effects on subgranular zone progenitor cells, increasing the number of newly born granule cells in the adult dentate gyrus. Functionally, NPY exerts anticonvulsive actions mediated by Y2 receptors at mossy fiber terminals, but there are no presynaptic responses to NPY at perforant path inputs to dentate granule cells in rats or mice. NPY also has potentially complicated actions on NPY-containing interneurons. Elevated expression of NPY in mossy fibers of the rat, sprouting of NPY interneurons in the human dentate, and over-expression of Y2 receptors in mossy fibers indicate an anticonvulsive role of endogenous NPY in epilepsy. However, the physiological role of NPY in the healthy dentate gyrus remains unclear. Keywords: neuropeptide Y; Y1 receptors; Y2 receptors; epilepsy Neuropeptide Y (NPY) is a 36 amino acid peptide first isolated from porcine brain extracts (Tatemoto et al., 1982). It is so named because it possesses tyrosine residues at both the C- and the N-terminal ends and three other tyrosine residues within its amino acid sequence. It was discovered and

isolated due to its C-terminal amidation common for many neuropeptides. NPY belongs to a larger family of pancreatic peptides that in addition includes peptide YY (PYY) and pancreatic polypeptide (PP). In the periphery, NPY is primarily expressed in sympathetic neurons and in the adrenal gland, and also in neurons of the enteric nervous system including the submucosal ganglion (Lundberg et al., 1983; Sundler et al., 1983). Unlike the other members of the pancreatic PP family, NPY is abundant in the central nervous system

Corresponding author. Tel.: +43-512-9003-71210;

E-mail: [email protected] (G. Sperk) Corresponding author.Tel.: 780-492-1933;

E-mail: [email protected] (W.F. Colmers) DOI: 10.1016/S0079-6123(07)63017-9

285

286

(de Quidt and Emson, 1986; Morris, 1989). It is preferentially located in subpopulations of GABAergic interneurons in the telencephalon (Hendry et al., 1984) in noradrenergic neurons originating from the locus coeruleus (Everitt et al., 1984; Harfstrand et al., 1987; Smialowska, 1988), and in hypothalamic neurons of the arcuate and other nuclei (Smith and Parent, 1986; Elmquist et al., 2005).

Neurons containing NPY in the dentate gyrus and their synaptic contacts NPY-immunoreactive (IR) neurons, all of which express GABA, are abundant in the dentate gyrus of humans (Chan-Palay et al., 1986; Furtinger et al., 2001), monkeys (Smith et al., 1985; Smith and Parent, 1986; Kohler et al., 1986), rodents (Kohler et al., 1986; Haas et al., 1987; Morris, 1989; Deller and Leranth, 1990) and pigs (Holm et al., 1992). NPY-containing perikarya are present in all layers of the fascia dentata. The largest population of hippocampal NPY-IR neurons is contained in the polymorphic region of the dentate gyrus, the dentate hilus (Kohler et al., 1986). In their careful examination of NPY-IR neurons in the rat dentate gyrus, Deller and Leranth (1990) classified four different types of neurons: The majority (>60%, type 2 cells) were identified as medium-sized multipolar and fusiform hilar neurons with dendrites occasionally reaching the outer molecular layer; the next most common (20%, type 3 cells) were pyramidal-shaped cells in the granule cell layer with long apical dendrites reaching the outer molecular layer. Comprising the remaining 20% were large multipolar cells located in the deep hilus (type 1); and small multipolar NPY-IR cells located in the molecular layer (type 4). About half of the hilar NPY-IR neurons (types 1 and 2) are also IR for somatostatin and the projection of these neurons has a high density of somatostatin-IR axon terminals (Kohler et al., 1986; Freund and Buzsaki, 1996). Light and electron microscopic studies show that the majority of NPY-IR terminals are located in the outer molecular layer of the rat dentate gyrus, where they establish symmetric (Gray type 2) synaptic contacts on dendritic shafts and

occasionally on spines of granule cells (Kohler et al., 1986; Deller and Leranth, 1990). While most of these axons pass through the granule cell layer and do not form contacts with somata of granule cells, a moderate number of NPY-IR synapses are also present on dendrites in the inner molecular layer and on the cell body of granule cells. The few synaptic contacts formed are symmetric, consistent with the GABAergic nature of these interneurons. Furthermore, numerous symmetric NPY-IR synapses are found on dendrites and somata of neurons in the hilar area, which support the pharmacological results reported below. Dendrites of some NPY-IR neurons in the hilar area receive input from granule cell axons, the mossy fibers axon collaterals. From transection studies, there is also evidence that NPY-IR dendrites are innervated by perforant pathway axon terminals in the outer molecular layer (Deller and Leranth, 1990). Commissurotomy revealed direct commissural input to NPY-IR dendrites at the border of the inner and outer molecular layer and in the hilus. Although the vast majority of NPY-IR neurons appear to be local circuit neurons, 2% of hilar NPY-IR neurons seem to project to the contralateral hippocampus, because they become retrogradely labeled when horseradish peroxidase (coupled to wheat germ agglutinin) is injected into the contralateral hippocampus (Deller and Leranth, 1990). These authors also report that symmetric NPY-IR synapses are found on the cell bodies and dendrites of hilar neurons. In primates (humans and monkeys) the distribution of NPY-IR neurons is similar to the rat (Chan-Palay et al., 1986; Kohler et al., 1986). The innervation of the outer molecular layer by hilar neurons appears to be, however, more prominent than in the rat. In the monkey, a small portion of NPY-IR hilar interneurons contains parvalbumin (Nitsch and Leranth, 1991). In humans, unlike naı¨ ve rats, faint NPY-IR is also observed in mossy fiber terminals (Chan-Palay et al., 1986; Furtinger et al., 2001). Neuropeptide Y receptors in the fascia dentata Several types of NPY receptors (Y1, Y2, Y4, Y5) have been identified in the brain (Dumont et al.,

287

1992; Parker and Herzog, 1999). In the hippocampus of rats, mice, and humans, Y1 and Y2 receptors are the most abundant. Y1 receptor mRNA is prominently expressed in granule cells and correspond to Y1 receptor binding sites on granule cell dendrites in the molecular layer, where they receive their input from hilar NPY interneurons (Dumont et al., 1998; Gackenheimer et al., 2001). Expression of Y1 receptors on other cells of the dentate gyrus (e.g., hilar interneurons) has not unequivocally been demonstrated so far, although electrophysiological results strongly suggest their presence (below). Interestingly NPY seems to augment proliferation of progenitor cells located in the subgranular zone (SGZ) by activation of Y1 receptors, a mechanism reduced in Y1 receptor knock out mice (Howell et al., 2005). This suggests that Y1 receptors are also present on dentate granule cell (DGC) progenitors. The site of Y2 receptor expression seems to be in part species-dependent (Dumont et al., 1998). In all species investigated so far, Y2 receptor mRNA is highly expressed in pyramidal cells. Whereas in most rat strains, notably in Sprague-Dawley rats, no Y2 message was demonstrated in DGCs, whereas Wistar rats and most mouse strains seem to express Y2 mRNA there (Parker and Herzog, 1999; Wolak et al., 2003; Kishi et al., 2005). Curiously, there is strong evidence that mossy fiber terminals in Sprague-Dawley rats express highly functional Y2 receptors (McQuiston and Colmers, 1996), and some evidence exists for Y2 receptormediated postsynaptic actions in some DGCs (McQuiston et al., 1996). In humans, prominent expression of Y2 mRNA was demonstrated in granule cells, and Y2 receptor binding was found in the terminal areas of mossy fibers in the dentate hilus and stratum lucidum of humans but not rats. Recent studies using antibodies to Y1 and Y2 receptors support the findings using receptor autoradiography (Stanic et al., 2006). In addition, Stanic et al. (2006) provided evidence that Y2 receptors may also be located on hilar interneurons, indicating that these receptors may also mediate inhibition of GABA release from interneurons in addition to suppressing glutamate release at mossy fibers.

Electrophysiological effects of NPY on dentate neurons Mossy fiber terminals As it does in area CA1 (Colmers et al., 1987, 1988; Colmers and Bleakman, 1994), NPY also negatively modulates synaptic transmission at the mossy fiber-to-CA3 synapse. Synaptic excitation, evoked by stimulation of the mossy fiber pathway, is potently and reversibly inhibited by bath application of NPY in a concentration-dependent manner (Klapstein and Colmers, 1993; Guo et al., 2002; El Bahh et al., 2005). The mechanism underlying this NPY-mediated inhibition of glutamate release from mossy fibers was determined by quantal synaptic analysis. Measurements of the frequency and amplitude of spontaneous (sEPSC) and miniature (mEPSC) excitatory postsynaptic currents were made in CA3 pyramidal neurons in the whole-cell patch configuration. Application of NPY and the Y2-preferring agonist [ahx5 24]NPY increased the interval and decreased the amplitude of sEPSCs (Fig. 1) but had no effect on either parameter when mEPSCs were measured in the presence of the voltage-dependent Na+ channel blocker, tetrodotoxin (McQuiston and Colmers, 1996). This indicates that Y2 receptors on mossy fiber terminals inhibit glutamate release and decrease excitatory input into CA3 pyramidal neurons. Subsequent experiments in area CA1 determined that NPY inhibits glutamate release via a suppression of presynaptic calcium influx (Qian et al., 1997). Controversy developed because it was unclear whether the Y5 receptor also played a role in this inhibition (Guo et al., 2002). However, in experiments at the mossy fiber-CA3 and stratum radiatum-CA1 synapse in hippocampal slices, presynaptic inhibitory responses to the application either of the Y2-preferring agonist [ahx5 24]NPY, or the Y5-preferring agonist, AlaAib NPY, were both blocked entirely by pretreatment with the potent and selective Y2 receptor antagonist, BIIE0246 (El Bahh et al., 2005). As this antagonist has no activity at the Y5 receptor (Doods et al., 1999), this indicates that the Y2 receptor is the only one mediating this action of NPY, and that at high concentrations, the

288 a. NPY (1 µM)

50 pA

Control

Washout

500 msec

c.

b.

NPY Control NPY13-36

300 pA

Control

20 msec

d.

e.

[Leu31Pro34] - NPY

PYY

Control

Control

Fig. 1. Y2- and Y1-mediated effects in the dentate gyrus. (a) Spontaneous synaptic currents recorded in CA3 pyramidal neuron. Sweeps in each column represent consecutive traces from neuron in control (left), in the presence of 1 mM NPY (center) and after 60 min washout (right). Note the reduction in frequency of excitatory postsynaptic currents (downward deflections) with NPY, and increase upon washout. Subsequent experiments with selective agonists in this paper demonstrate the Y2 nature of this response (adapted from McQuiston and Colmers, 1996, with permission). (b–e) Ca2+currents in acutely isolated rat dentate granule cells are inhibited by agonists at Y1 receptors. Shown superimposed are current traces evoked in this neuron in control and in the presence of the respective agonist. NPY and PYY are pan-agonists, while Leu31, Pro34 NPY is a Y1/Y5-preferring agonist, and NPY13 36 is a Y2/ Y5-preferring agonist. In 30% of neurons tested, a Y2 receptor agonist also inhibited the Ca2+current, while the Y1 agonist was effective in every neuron tested (adapted from McQuiston et al., 1996, with permission. r 1996, by the Society for Neuroscience).

289

Y5-preferring agonist has some activity at Y2 receptors (El Bahh et al., 2005). This conclusion is further supported by the observation that in hippocampal slices prepared from mutant mice lacking functional Y2 receptors, there is no inhibition of either the mossy fiber or stratum radiatumevoked field excitatory postsynaptic potentials (EPSPs) by NPY or the Y5-preferring agonist AlaAib NPY (El Bahh et al., 2005). The Y2 receptor is therefore responsible for NPY-mediated inhibition of glutamate release from mossy fiber terminals, most likely by suppression of voltage-dependent Ca2+ influx through presynaptic calcium channels.

Actions of NPY at somata and dendrites of dentate granule cells (DGCs) In the dentate molecular layer, the effect of NPY on EPSPs is, if at all detectable, certainly far less pronounced. Stimulation of the perforant path or commissural inputs evokes EPSPs in the molecular layer that are either unaltered (Klapstein and Colmers, 1993), or show very minor inhibition (Bijak and Smialowska, 1995) with bath application of NPY. This apparent absence of NPY actions within the dentate gyrus itself was puzzling, given the significant levels of NPY receptor expression previously reported in this region (Chan-Palay et al., 1986; Kohler et al., 1986), and prompted further investigations. Using patch clamp recording and simultaneous calcium imaging, McQuiston et al. (1996) evoked action potentials in rat DGCs in brain slices. Bath application of NPY did not alter resting Ca2+levels in the soma or dendrites of the DGCs, but did significantly and reversibly decrease the depolarization-induced Ca2+ influx in the soma and dendrites of these same DGCs. Because voltage clamp conditions are less than ideal in neurons with their dendritic trees intact, as occurs in brain slices, mechanistic studies of this postsynaptic NPY action were undertaken in acutely isolated DGCs, which are electrically more tractable. Somatic Ca2+ currents were isolated pharmacologically and NPY and receptor subtype-preferring agonists were applied, as were selective blockers of different

Ca2+channel subtypes. Calcium currents were inhibited by the NPY Y1- and Y5-preferring agonist, [Leu31Pro34]NPY but were much less commonly inhibited by the NPY Y2- (and Y5)preferring agonist, NPY13-36 (Fig. 1). These observations were consistent with the actions of NPY on Ca2+ influx preferentially occurring via Y1 receptor activation. Selective blockade of N-type Ca2+ channels occluded all actions of NPY, consistent with an action of Y1 receptors to selectively suppress current through this subtype of voltage-dependent Ca2+channel (VDCC) (McQuiston et al., 1996). Functionally, the inhibition of calcium currents by NPY was postulated to regulate the release of dynorphin from DGC dendrites, which is mediated by L- and N-type VDCCs (Simmons et al., 1995), but this has not been investigated further. Certainly, the Y1 receptor does mediate the inhibition of N-type VDCCs in DGCs. Given the remarkable amounts of NPY in the dentate, particularly in the molecular layer, it is reasonable to presume that there will be some significant physiological consequences of Y1 receptor activation on the dendrites and soma of DGCs.

NPY effects on hilar interneurons The presence of NPY receptors in the hilus is controversial, with some studies showing no immunoreactivity (Dumont et al., 1998; Gackenheimer et al., 2001) and others demonstrating their existence (Kopp et al., 2002; Paredes et al., 2003). Electrophysiological recordings of hilar interneurons in the mouse hippocampus performed by Paredes et al. (2003) demonstrated the presence of a G-protein-coupled inwardly rectifying potassium current (GIRK) that was activated by NPY in 8/17 dentate hilar interneurons studied. Furthermore, the Y1- and Y5-preferring agonist, [Leu31Pro34]NPY mimicked the effects of NPY, also in about half the neurons tested, consistent with a Y1 receptor-mediated modulation of GIRK currents. This is similar to a Y1 receptor-mediated action described in reticular and ventrobasal nucleus neurons of the thalamus (Sun et al., 2001a, b). Paredes et al. (2003) also used

290

immunocytochemisty to determine that NPYcontaining interneurons of the dentate hilus contain Y1 receptors, consistent with their electrophysiological findings. These observations suggest that a significant subpopulation of hilar interneurons, similar in percentage to that expressing NPY itself, are inhibited by NPY. This inhibition of inhibitory interneurons is a potential explanation for the proconvulsant actions of Y1 agonists in the dentate in vivo (Gariboldi et al., 1998). NPY effects on long-term potentiation (LTP) LTP is reliably evoked with high frequency stimulation of the perforant pathway in the dentate gyrus in vivo. Induction of perforant pathway LTP occurs as a result of DGC dendrites undergoing an NMDA receptor-mediated AMPA receptor insertion that is responsible for the transformation of ‘silent synapses’ to active ones (Bi and Poo, 1998; Lin et al., 2006; Moga et al., 2006). Whittaker et al. (1999) injected NPY intracerebroventricularly (ICV) prior to LTP induction in vivo. This resulted in an inhibition of the induction and maintenance of perforant path LTP. The previously described Y1 receptor-mediated inhibition of dendritic calcium currents (McQuiston et al., 1996) must be considered a potential candidate postsynaptic mechanism for the observed inhibition of LTP. Alternatively, Y1 receptor activation may also alter downstream effectors necessary for LTP induction, such as calcium-calmodulin kinase II, protein kinase C or mitogen activated/extracellular –signal-regulated kinase (Lin et al., 2006). An alternative interpretation is that NPY inhibits perforant pathway glutamate release on to dendrites of DGCs (Whittaker et al., 1999). However, there is scant evidence for any presynaptic actions of NPY on perforant path inputs, as discussed above (Klapstein and Colmers, 1993). Therefore, the possibility exists that NPY may inhibit perforant pathway LTP by reducing postsynaptic Ca2+influx through its documented action at N-type VDCCs, or by inhibiting a downstream signal effector.

NPY effects on synaptosomes Experiments using synaptosomes to examine NPY function have produced interesting results and interpretations. Synaptosomes, a biochemical preparation of isolated synaptic terminals, are commonly used for the analysis of synaptic release. The application of a high extracellular K+ solution stimulates synaptosomal glutamate release, which can be quantitated fluorimetrically. This preparation can also be used to study presynaptic calcium influx, using fluorescent calcium indicators. Preincubation of synaptosomes with the compound of interest allows testing of possible effects on transmitter release. The release of glutamate from synaptosomes prepared from the dentate gyrus is significantly reduced by preincubation of the Y1- and Y5-preferring agonist [Leu31Pro34]NPY and by the Y2- and Y5-preferring agonist NPY13-36 (Silva et al., 2001). These effects occurred both in the presence of normal and very low levels of extracellular Ca2+, indicating that the mechanism is Ca2+-independent. Thus, in synaptosomes prepared from the rat dentate gyrus, Y1 and Y2 receptors appear to mediate suppression of glutamate release. With the recent discovery of the Y5 receptor and Y5 selective agonists, the Y5 receptor has also been investigated in the dentate gyrus. As is the case with Y1 and Y2 agonists, the Y5 agonist NPY (19-23)-(Gly(1),Ser(3),Gln(4),Thr(6),Ala(31), Aib(32),Gln(34))-pancreatic PP also inhibits P/Q-type VDCC-mediated glutamate release from synaptosomes (Silva et al., 2003). Interestingly, despite the significant actions of Y1and Y5-preferring agonists in this preparation, neither the Y1 receptor antagonist BIBP3226 nor the Y5 receptor antagonist L-152,804 are able to block the inhibition mediated by these agonists The Y2 antagonist BIIE0246, on the other hand, potently blocks the inhibition of glutamate release, which has lead to the suggestion that Y2 receptors are functionally coupled to Y1 and Y5 receptors and can override their modulatory effects (Silva et al., 2003). However, an alternate interpretation is suggested by the way the experiments in this paper were conducted: the investigators added the agonist and antagonists at the same time, spun the

291

synaptosomes, then reapplied the mixture when resuspending them. Because the agonist peptides are much larger and ‘‘stickier’’ molecules than are the antagonists, it is possible that more of the agonists remained behind when the supernatant was removed, and preferentially occupied the receptors, giving an impression that the antagonists were ineffective. This would not be the case for the Y2 receptor antagonist, BIIE0246, which exhibits an irreversible form of antagonism upon prolonged exposure to the receptor (Dautzenberg & Neysari, 2005; El Bahh et al., 2005). It might be possible to test this hypothesis by pretreating the synaptosomes with the antagonist alone, then resuspending them after centrifugation in the presence of both antagonist and agonist. Notwithstanding this possibility, it appears that in synaptosomes prepared from the dentate gyrus, Y1, Y2, and Y5 agonists can inhibit glutamate release. Moreover, the most prominent inhibition occurs with the Y2 agonist that likely mediates the inhibitory effect (Silva et al., 2003). These results differ from those from the electrophysiological studies which suggest that NPY-mediated inhibition of glutamate release does not occur in the dentate gyrus (Klapstein and Colmers, 1993; Mcquiston and Colmers, 1996; Mcquiston et al., 1996; El Bahh et al., 2002, 2005). A number of potential differences in methods may explain these differences. First, while the electrophysiological studies represent the evoked or spontaneous activity in defined neurons and synaptic pathways, synaptosome preparations contain all terminals that survive isolation, including inputs to interneurons, terminals from other pathways, etc. in addition to the perforant path inputs studied. Furthermore, mossy fiber terminals, which have multiple release sites (Lawrence and McBain, 2003) appear to respond to Y2 receptor agonists (McQuiston and Colmers, 1996), and although there is some evidence for the presence of Y5 receptors in the dentate gyrus (Guo et al., 2002), this is disputed (El Bahh et al., 2005). However, the presence of Y2 receptors on the terminals of mossy fibers may be sufficient to explain the observed inhibition of glutamate release in synaptosomes.

NPY effects on neurogenesis In the last decade, it has become clear that neural precursor cells in the SGZ of the dentate gyrus supply new granule cells on an ongoing basis (see chapter by Parent, this volume). In the SGZ, neuronal precursors proliferate and continually migrate into the granule cell layer where they mature and become functional granule cells. The specific factors that influence dentate gyrus neurogenesis are, therefore, of significant interest because their impact will modify the memory formation process (Aimone et al., 2006). Exercise, growth factors, environmental enrichment, aging, hormones such as estrogen and prolactin, glutamatergic neurotransmission, adrenal hormones, LTP, lesions, seizure activity, and the presence of NPY are all associated with a regulation of dentate neurogenesis (cf. Parent, this volume). In NPY knockout mice, it was fortuitously observed that proliferation of olfactory precursor cells was impaired, and NPY was seen to promote neurogenesis there, via a Y1 receptor mechanism (Hansel et al., 2001). Howell et al. (2005) examined whether NPY might affect proliferation of neural progenitors in the SGZ by culturing neuroblast precursor cells from the early postnatal dentate gyrus in the absence or presence of NPY in the culture medium. NPY treatment significantly increased the number of neurons and neuroblasts that incorporated Brd-U, consistent with a proliferative action of NPY. The strongly Y1-preferring agonist, F7,P34 NPY mimicked the effect of NPY application in vitro, and the effect of NPY was blocked by application of the Y1 antagonist BIBP3226 in the presence of NPY. The neuroproliferative action of NPY was also abolished by inhibiting the extracellular signal-regulated kinase (ERK)1/2, a subgroup of mitogen activated kinases (Scharfman et al., 2000). Therefore, it is likely that a downstream effect of the Y1 receptor is the activation of ERK1/2. These authors then validated the observations in Y1 receptor knockout mice in vivo. Y1 / mice had a significantly lower SGZ proliferation rate and a decreased number of neurons expressing doublecortin than in wild types. Thus, data from in vitro and in vivo models supports a Y1 receptor-mediated

292

neuroproliferative action of NPY in the hilus of the dentate gyrus. This may also play a role in depression (below).

Neuropeptide Y and its receptors in disease Epilepsy In animal models of epilepsy and in human temporal lobe epilepsy, the NPY system undergoes considerable changes in the dentate gyrus, both in the expression of the peptide and of its receptors (Vezzani et al., 1999). With repeated, mildly convulsive stimuli, such as kindling, expression of NPY and its Y2 receptors become up-regulated in hilar interneurons (Fig. 2) (Gruber et al., 1994; Vezzani et al., 1999), and NPY expression is up-regulated in DGCs (Marksteiner et al., 1990; Sperk et al., 1992; Goodman and Sloviter, 1993; Causing et al., 1996). Viral vectors inducing overexpression of NPY have clear anticonvulsive effects when infused locally into the hippocampus of rats (Richichi et al., 2004). On the other hand, prolonged seizures (e.g., during status epilepticus) cause degeneration of hilar interneurons (Sloviter, 1983; Sperk et al., 1992). However, surviving neurons may still up-regulate NPY expression. In rat models of temporal lobe epilepsy, expression of Y2 receptors is up-regulated simultaneously in granule cells and mossy fibers (Fig. 2) (Schwarzer et al., 1998). Over-expression of NPY and its Y2 receptors seems to have a protective and anticonvulsive role caused by the Y2-mediated presynaptic inhibition of glutamate release from mossy fibers (Klapstein and Colmers, 1993; Greber et al., 1994; El Bahh et al., 2005). Conversely, Y1 receptors located on granule cell dendrites are down-regulated with seizure activity (Kofler et al., 1997). Because Gariboldi et al. (1998) demonstrated that Y1 receptor agonists injected into the dentate gyrus have proconvulsant action in vivo, this change could ameliorate seizure activity or compensate for the increased excitation (see also below). The prominent anticonvulsive action of NPY has been suggested to be mediated by Y5 receptor stimulation (Woldbye et al., 1997). This has been debated, however, especially since

expression of Y5 receptors seems to be negligible, at least in the mouse hippocampus and a crucial role of Y2 receptors has been shown in the anticonvulsive action of NPY (El Bahh et al., 2005; see also below). In hippocampal tissue surgically removed from patients for treatment of temporal lobe epilepsy, enhanced expression of NPY mRNA has been observed in interneurons of the dentate hilus, together with a markedly increased total length of NPY-positive fibers in the inner and outer dentate molecular layer, and the stratum lacunosum-moleculare (Furtinger et al., 2001). Interestingly, the pattern of interneuronal axon sprouting overlapped in part with terminal areas both of normal mossy fibers and the collaterals that sprout into the inner molecular layer in the epileptic brain (mossy fiber sprouting). In dentate gyrus of epilepsy patients, Y2 receptors were up-regulated and Y1 receptors down-regulated in mossy fibers and in the molecular layer, respectively (Furtinger et al., 2001). This indicates that NPY, released from sprouted interneurons could also reach Y2 receptors located on mossy fibers. In epileptic rats, NPY has been shown to have tonic inhibitory actions, mediated via Y2 receptors (Tu et al., 2005). These changes in NPY-related neuronal circuitry may be the basis for the anticonvulsive action of NPY demonstrated ex vivo in hippocampal tissue obtained from patients with temporal lobe epilepsy (Patrylo et al., 1999).

Alzheimer’s disease Chan-Palay et al. (1986) observed a marked reduction in hilar NPY neurons in Alzheimer’s Disease (AD) patients at autopsy. Assuming that the degeneration of the septo-hippocampal pathway may be related to the cognitive impairment in AD patients, it is interesting that lesions of the septo-hippocampal pathway result in an initial increase in NPY-IR in the dentate hilus (Hortnagl et al., 1990; Bayer and Milner, 1993) and a subsequent loss of hilar NPY neurons (Milner et al., 1997). It is possible that the initial increase NPYIR may reflect sustained stimulation of these neurons leading then to their loss through

293

Fig. 2. NPY and its receptors in control and epileptic rats. NPY-IR is present in numerous interneurons of the dentate gyrus that project to the outer molecular layer and have collaterals within the dentate hilus (see the respective labeling in a, representing a high magnification image of NPY-IR in the dorsal dentate gyrus). In epileptic rats (b; 30 days after kainic acid-induced seizures; lower magnification image of the dorsal hippocampus) varying portions of hilar NPY neurons degenerate (note the loss in NPY-IR in the outer molecular layer). After seizures mossy fibers strongly express NPY-IR unlike in naı¨ ve animals. The arrow in b marks sprouted mossy fiber terminals in the inner molecular layer also containing NPY. (c) Y1 receptors are postsynaptic receptors located mainly on dendrites of granule cells (here labeled with the Y1/Y5 ligand 125I-Pro34-PYY) and become reduced after kainic acid-induced seizures (arrow in d; 8 days after application of kainic acid injection). Panels e and f depict brain sections labeled with the Y2/Y5 specific ligand 125 I-PYY(3 36). Note the strong Y2-specific labeling of strata oriens and radiatum in control rats (e). In kainic acid treated rats strong labeling of mossy fibers can be seen representing presynaptic Y2 receptors (arrow in f; 8 days after application of kainic acid injection).

excitotoxic mechanisms (Chan-Palay et al., 1986; Milner et al., 1997).

Depression In ‘‘depressed’’ Flinder’s Sensitive Line (FSL) rats, basal expression of NPY and Y1 mRNA in the dentate gyrus is significantly lower than in control animals (Jimenez-Vasquez et al., 2006). An upregulation of Y1 receptor binding in the dentate gyrus was also observed, most likely due to an increased affinity for and/or decreased internalization of NPY receptors (Jimenez-Vasquez et al., 2006). Electroconvulsive therapy (ECT), a common treatment for depressive disorders, raises NPY mRNA levels in the hilus (Mikkelsen et al.,

1994), and Y1 receptor mRNA expression in DGCs (Madsen et al., 2000) of naı¨ ve rats. In FSL rats, ECT also increases NPY mRNA and Y1 mRNA expression, and reduces Y1 receptor binding in the dentate gyrus (Jimenez-Vasquez et al., 2006). This altered NPY expression was measured after an ECT-induced reduction in depressive behavior, measured with the Porsolt swim test. Interestingly, ECT increases seizure threshold in epileptic rats, which correlates with a reduction in NPY binding sites in the dentate gyrus (Bolwig et al., 1999). Fluoxetine, a serotonin-selective reuptake inhibitor used in the treatment of depression, has also been observed to increase NPY mRNA in the dentate gyrus of Flinders Sensitive Line rats (Caberlotto et al., 1999), although this was not replicated in a different lab using tricyclic

294

antidepressants in Sprague-Dawley rats (Bellmann and Sperk, 1993). Thus, a downregulation of Y1 receptor mRNA is intimately related to the pathophysiology of depression and is altered by treatments that have been efficacious in treating depressive disorders, possibly via an increase in NPY receptor expression. These effects are not only limited to animal models. Depressed humans have a decreased NPY-like immunoreactivity (NPY-LI) in their cerebrospinal fluid compared to non-depressed control subjects (Widerlov et al., 1988; Gjerris et al., 1992). Consistent with the hypothesis that a hypofunction of the NPY system is related to depression, depressed human subjects exhibit increased NPY-LI after ECT (Mathe, 1999). Lastly, it appears that neurogenesis in the dentate gyrus is necessary for the actions of antidepressants (Dranovsky and Hen, 2006). It is therefore tempting to speculate that some of the antidepressant actions of the NPY system may be mediated through neuroproliferative actions. Summary NPY is an extraordinarily abundant peptide in the dentate gyrus, but the roles that it plays remain somewhat obscure. It is clear that the dentate gyrus NPY peptide-receptor system is extremely sensitive to ongoing activity, and is poised to regulate many important aspects of the plasticity of dentate gyrus circuitry. Certainly, the links between NPY and excitability, memory formation and depression are important and warrant intensive study. Abbreviations AD DGC ECT EPSP GABA GIRK ICV LTP

Alzheimer’s disease dentate granule cell electroconvulsive therapy excitatory postsynaptic potential gamma-aminobutyric acid G-protein-coupled inwardly-rectifying potassium current intracerebroventricular long-term potentiation

MEPSC NPY NPY-IR NPY-LI PYY PP SEPSC SGZ VDCC

miniature excitatory postsynaptic current neuropeptide Y neuropeptide Y immunoreactive neuropeptide Y-like immunoreactivity peptide YY pancreatic polypeptide spontaneous excitatory postsynaptic current subgranular zone voltage dependent Ca2+ channel

References Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for adult neurogenesis in the encoding of time in new memories. Nat. Neurosci., 9(6): 723–727. Bayer, L.E. and Milner, T.A. (1993) Transient increases in neuropeptide Y-like immunoreactivity in dentate hilar neurons following fimbria/fornix transection. J. Neurosci. Res., 34(4): 434–441. Bellmann, R. and Sperk, G. (1993) Effects of antidepressant drug treatment on levels of NPY or prepro-NPY-mRNA in the rat brain. Neurochem. Int., 22(2): 183–187. Bi, G.Q. and Poo, M.M. (1998) Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type. J. Neurosci., 18(24): 10464–10472. Bijak, M. and Smialowska, M. (1995) Effects of neuropeptide Y on evoked potentials in the CA1 region and the dentate gyrus of the rat hippocampal slice. Pol. J. Pharmacol., 47(4): 333–338. Bolwig, T.G., Woldbye, D.P. and Mikkelsen, J.D. (1999) Electroconvulsive therapy as an anticonvulsant: a possible role of neuropeptide Y (NPY). J. ECT, 15(1): 93–101. Caberlotto, L., Jimenez, P., Overstreet, D.H., Hurd, Y.L., Mathe, A.A. and Fuxe, K. (1999) Alterations in neuropeptide Y levels and Y1 binding sites in the Flinders Sensitive Line rats, a genetic animal model of depression. Neurosci. Lett., 265(3): 191–194. Causing, C.G., Makus, K., Ma, Y., Miller, F.D. and Colmers, W.F. (1996) Selective upregulation of Ta1 a-tubulin and Neuropeptide Y mRNAs after intermittent excitatory stimulation in adult rat hippocampus in vivo. J. Comp. Neurol., 367(1): 132–146. Chan-Palay, V., Kohler, C., Haesler, U., Lang, W. and Yasargil, G. (1986) Distribution of neurons and axons immunoreactive with antisera against neuropeptide Y in the normal human hippocampus. J. Comp. Neurol., 248(3): 360–375. Colmers, W.F. and Bleakman, D. (1994) Neuropeptide Y effects on the electrical properties of neurons. Trends Neurosci., 17(9): 373–379.

295 Colmers, W.F., Lukowiak, K.D. and Pittman, Q.J. (1987) Presynaptic action of neuropeptide Y in area CA1 of the rat hippocampal slice. J. Physiol., 383: 285–299. Colmers, W.F., Lukowiak, K. and Pittman, Q.J. (1988) Neuropeptide Y action in the rat hippocampal slice: site and mechanism of presynaptic inhibition. J. Neurosci., 8(10): 3827–3837. Dautzenberg, F.M. and Neysari, S. (2005) Irreversible binding kinetics of neuropeptide Y ligands to Y2 but not to Y1 and Y5 receptors. Pharmacology, 75(1): 21–29. Deller, T. and Leranth, C. (1990) Synaptic connections of neuropeptide Y (NPY) immunoreactive neurons in the hilar area of the rat hippocampus. J. Comp. Neurol., 300(3): 433–447. Doods, H., Gaida, W., Wieland, H.A., Dollinger, H., Schnorrenberg, G., Esser, F., Engel, W., Eberlein, W. and Rudolf, K. (1999) BIIE0246: a selective and high affinity neuropeptide Y Y(2) receptor antagonist. Eur. J. Pharmacol., 384(2-3): R3–R5. Dranovsky, A. and Hen, R. (2006) Hippocampal neurogenesis: regulation by stress and antidepressants. Biol. Psychiatry, 59(12): 1136–1143. Dumont, Y., Jacques, D., Bouchard, P. and Quirion, R. (1998) Species differences in the expression and distribution of the neuropeptide Y Y1, Y2, Y4, and Y5 receptors in rodents, guinea pig, and primates brains. J. Comp. Neurol., 402(3): 372–384. Dumont, Y., Martel, J.C., Fournier, A., St-Pierre, S. and Quirion, R. (1992) Neuropeptide Y and neuropeptide Y receptor subtypes in brain and peripheral tissues. Prog. Neurobiol., 38(2): 125–167. El Bahh, B., Balosso, S., Hamilton, T., Herzog, H., Beck-Sickinger, A.G., Sperk, G., Gehlert, D.R., Vezzani, A. and Colmers, W.F. (2005) The anti-epileptic actions of neuropeptide Y in the hippocampus are mediated by Y and not Y receptors. Eur. J. Neurosci., 22(6): 1417–1430. El Bahh, B., Cao, J.Q., Beck-Sickinger, A.G. and Colmers, W.F. (2002) Blockade of neuropeptide Y(2) receptors and suppression of NPY’s anti-epileptic actions in the rat hippocampal slice by BIIE0246. Br. J. Pharmacol., 136(4): 502–509. Elmquist, J.K., Coppari, R., Balthasar, N., Ichinose, M. and Lowell, B.B. (2005) Identifying hypothalamic pathways controlling food intake, body weight, and glucose homeostasis. J. Comp. Neurol., 493(1): 63–71. Everitt, B.J., Hokfelt, T., Terenius, L., Tatemoto, K., Mutt, V. and Goldstein, M. (1984) Differential co-existence of neuropeptide Y (NPY)-like immunoreactivity with catecholamines in the central nervous system of the rat. Neuroscience, 11(2): 443–462. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6(4): 347–470. Furtinger, S., Pirker, S., Czech, T., Baumgartner, C., Ransmayr, G. and Sperk, G. (2001) Plasticity of Y1 and Y2 receptors and neuropeptide Y fibers in patients with temporal lobe epilepsy. J. Neurosci., 21(15): 5804–5812. Gackenheimer, S.L., Schober, D.A. and Gehlert, D.R. (2001) Characterization of neuropeptide Y Y1-like and Y2-like

receptor subtypes in the mouse brain. Peptides, 22(3): 335–341. Gariboldi, M., Conti, M., Cavaleri, D., Samanin, R. and Vezzani, A. (1998) Anticonvulsant properties of BIBP3226, a non-peptide selective antagonist at neuropeptide Y Y1 receptors. Eur. J. Neurosci., 10(2): 757–759. Gjerris, A., Widerlov, E., Werdelin, L. and Ekman, R. (1992) Cerebrospinal fluid concentrations of neuropeptide Y in depressed patients and in controls. J. Psychiatry Neurosci., 17(1): 23–27. Goodman, J.H. and Sloviter, R.S. (1993) Cocaine neurotoxicity and altered neuropeptide Y immunoreactivity in the rat hippocampus; a silver degeneration and immunocytochemical study. Brain Res., 616(1–2): 263–272. Greber, S., Schwarzer, C. and Sperk, G. (1994) Neuropeptide Y inhibits potassium-stimulated glutamate release through Y2 receptors in rat hippocampal slices in vitro. Br. J. Pharmacol., 113(3): 737–740. Gruber, B., Greber, S., Rupp, E. and Sperk, G. (1994) Differential NPY mRNA expression in granule cells and interneurons of the rat dentate gyrus after kainic acid injection. Hippocampus, 4(4): 474–482. Guo, H., Castro, P.A., Palmiter, R.D. and Baraban, S.C. (2002) Y5 receptors mediate neuropeptide Y actions at excitatory synapses in area CA3 of the mouse hippocampus. J. Neurophysiol., 87(1): 558–566. Haas, H.L., Hermann, A., Greene, R.W. and Chan-Palay, V. (1987) Action and location of neuropeptide tyrosine (Y) on hippocampal neurons of the rat in slice preparations. J. Comp. Neurol., 257(2): 208–215. Hansel, D.E., Eipper, B.A. and Ronnett, G.V. (2001) Neuropeptide Y functions as a neuroproliferative factor. Nature, 410(6831): 940–944. Harfstrand, A., Fuxe, K., Terenius, L. and Kalia, M. (1987) Neuropeptide Y-immunoreactive perikarya and nerve terminals in the rat medulla oblongata: relationship to cytoarchitecture and catecholaminergic cell groups. J. Comp. Neurol., 260(1): 20–35. Hendry, S.H., Jones, E.G., DeFelipe, J., Schmechel, D., Brandon, C. and Emson, P.C. (1984) Neuropeptide-containing neurons of the cerebral cortex are also GABAergic. Proc. Natl. Acad. Sci. USA, 81(20): 6526–6530. Holm, I.E., Geneser, F.A. and Zimmer, J. (1992) Somatostatinand neuropeptide Y-like immunoreactivity in the dentate area, hippocampus, and subiculum of the domestic pig. J. Comp. Neurol., 322(2): 390–408. Hortnagl, H., Sperk, G., Sobal, G. and Maas, D. (1990) Cholinergic deficit induced by ethylcholine aziridinium (AF64A) transiently affects somatostatin and neuropeptide Y levels in rat brain. J. Neurochem., 54(5): 1608–1613. Howell, O.W., Doyle, K., Goodman, J.H., Scharfman, H.E., Herzog, H., Pringle, A., Beck-Sickinger, A.G. and Gray, W.P. (2005) Neuropeptide Y stimulates neuronal precursor proliferation in the post-natal and adult dentate gyrus. J. Neurochem., 93(3): 560–570. Jimenez-Vasquez, P.A., Diaz-Cabiale, Z., Caberlotto, L., Bellido, I., Overstreet, D., Fuxe, K. and Mathe, A.A. (2006)

296 Electroconvulsive stimuli selectively affect behavior and neuropeptide Y (NPY) and NPY Y(1) receptor gene expressions in hippocampus and hypothalamus of Flinders Sensitive Line rat model of depression. Eur. Neuropsychopharmacol., 17(4): 298–308. Kishi, T., Aschkenasi, C.J., Choi, B.J., Lopez, M.E., Lee, C.E., Liu, H., Hollenberg, A.N., Friedman, J.M. and Elmquist, J.K. (2005) Neuropeptide Y Y1 receptor mRNA in rodent brain: distribution and colocalization with melanocortin-4 receptor. J. Comp. Neurol., 482(3): 217–243. Klapstein, G.J. and Colmers, W.F. (1993) On the sites of presynaptic inhibition by neuropeptide Y in rat hippocampus in vitro. Hippocampus, 3(1): 103–111. Kofler, N., Kirchmair, E., Schwarzer, C. and Sperk, G. (1997) Altered expression of NPY-Y1 receptors in kainic acid induced epilepsy in rats. Neurosci. Lett., 230(2): 129–132. Kohler, C., Eriksson, L., Davies, S. and Chan-Palay, V. (1986) Neuropeptide Y innervation of the hippocampal region in the rat and monkey brain. J. Comp. Neurol., 244(3): 384–400. Kopp, J., Xu, Z.Q., Zhang, X., Pedrazzini, T., Herzog, H., Kresse, A., Wong, H., Walsh, J.H. and Hokfelt, T. (2002) Expression of the neuropeptide Y Y1 receptor in the CNS of rat and of wild-type and Y1 receptor knock-out mice. Focus on immunohistochemical localization. Neuroscience, 111(3): 443–532. Lawrence, J.J. and McBain, C.J. (2003) Interneuron diversity series: containing the detonation–feedforward inhibition in the CA3 hippocampus. Trends Neurosci., 26(11): 631–640. Lin, Y.W., Yang, H.W., Wang, H.J., Gong, C.L., Chiu, T.H. and Min, M.Y. (2006) Spike-timing-dependent plasticity at resting and conditioned lateral perforant path synapses on granule cells in the dentate gyrus: different roles of N-methylD-aspartate and group I metabotropic glutamate receptors. Eur. J. Neurosci., 23(9): 2362–2374. Lundberg, J.M., Terenius, L., Hokfelt, T. and Goldstein, M. (1983) High levels of neuropeptide Y in peripheral noradrenergic neurons in various mammals including man. Neurosci. Lett., 42(2): 167–172. Madsen, T.M., Greisen, M.H., Nielsen, S.M., Bolwig, T.G. and Mikkelsen, J.D. (2000) Electroconvulsive stimuli enhance both neuropeptide Y receptor Y1 and Y2 messenger RNA expression and levels of binding in the rat hippocampus. Neuroscience, 98(1): 33–39. Mathe, A.A. (1999) Neuropeptides and electroconvulsive treatment. J. ECT, 15(1): 60–75. Marksteiner, J., Ortler, M., Bellmann, R. and Sperk, G. (1990) Neuropeptide Y biosynthesis is markedly induced in mossy fibers during temporal lobe epilepsy of the rat. Neurosci. Lett., 112(2-3): 143–148. McQuiston, A.R. and Colmers, W.F. (1996) Neuropeptide Y2 receptors inhibit the frequency of spontaneous but not miniature EPSCs in CA3 pyramidal cells of rat hippocampus. J. Neurophysiol., 76(5): 3159–3168. McQuiston, A.R., Petrozzino, J.J., Connor, J.A. and Colmers, W.F. (1996) Neuropeptide Y1 receptors inhibit N-type calcium currents and reduce transient calcium increases in rat dentate granule cells. J. Neurosci., 16(4): 1422–1429.

Mikkelsen, J.D., Woldbye, D., Kragh, J., Larsen, P.J. and Bolwig, T.G. (1994) Electroconvulsive shocks increase the expression of neuropeptide Y (NPY) mRNA in the piriform cortex and the dentate gyrus. Brain Res. Mol. Brain Res., 23(4): 317–322. Milner, T.A., Wiley, R.G., Kurucz, O.S., Prince, S.R. and Pierce, J.P. (1997) Selective changes in hippocampal neuropeptide Y neurons following removal of the cholinergic septal inputs. J. Comp. Neurol., 386(1): 46–59. Moga, D.E., Shapiro, M.L. and Morrison, J.H. (2006) Bidirectional redistribution of AMPA but not NMDA receptors after perforant path simulation in the adult rat hippocampus in vivo. Hippocampus, 16(11): 990–1003. Morris, B.J. (1989) Neuronal localisation of neuropeptide Y gene expression in rat brain. J. Comp. Neurol., 290(3): 358–368. Nitsch, R. and Leranth, C. (1991) Neuropeptide Y (NPY)immunoreactive neurons in the primate fascia dentata; occasional coexistence with calcium-binding proteins: a light and electron microscopic study. J. Comp. Neurol., 309(4): 430–444. Paredes, M.F., Greenwood, J. and Baraban, S.C. (2003) Neuropeptide Y modulates a G protein-coupled inwardly rectifying potassium current in the mouse hippocampus. Neurosci. Lett., 340(1): 9–12. Parker, R.M. and Herzog, H. (1999) Regional distribution of Y-receptor subtype mRNAs in rat brain. Eur. J. Neurosci., 11(4): 1431–1448. Patrylo, P.R., van den Pol, A.N., Spencer, D.D. and Williamson, A. (1999) NPY inhibits glutamatergic excitation in the epileptic human dentate gyrus. J. Neurophysiol., 82(1): 478–483. Qian, J., Colmers, W.F. and Saggau, P. (1997) Inhibition of synaptic transmission by neuropeptide Y in rat hippocampal area CA1: modulation of presynaptic Ca2+ entry. J. Neurosci., 17(21): 8169–8177. de Quidt, M.E. and Emson, P.C. (1986) Distribution of neuropeptide Y-like immunoreactivity in the rat central nervous system–II. Immunohistochemical analysis. Neuroscience, 18(3): 545–618. Richichi, C., Lin, E.J., Stefanin, D., Colella, D., Ravizza, T., Grignaschi, G., Veglianese, P., Sperk, G., During, M.J. and Vezzani, A. (2004) Anticonvulsant and antiepileptogenic effects mediated by adeno-associated virus vector neuropeptide Y expression in the rat hippocampus. J. Neurosci., 24(12): 3051–3059. Scharfman, H.E., Goodman, J.H. and Sollas, A.L. (2000) Granule-like neurons at the hilar/CA3 border after status epilepticus and their synchrony with area CA3 pyramidal cells: functional implications of seizure-induced neurogenesis. J. Neurosci., 20(16): 6144–6158. Schwarzer, C., Kofler, N. and Sperk, G. (1998) Up-regulation of neuropeptide Y-Y2 in an animal model of temporal lobe epilepsy. Mol. Pharmacol., 53(1): 6–13. Silva, A.P., Carvalho, A.P., Carvalho, C.M. and Malva, J.O. (2001) Modulation of intracellular calcium changes and glutamate release by neuropeptide Y1 and Y2 receptors in the

297 rat hippocampus: differential effects in CA1, CA3 and dentate gyrus. J. Neurochem., 79(2): 286–296. Silva, A.P., Carvalho, A.P., Carvalho, C.M. and Malva, J.O. (2003) Functional interaction between neuropeptide Y receptors and modulation of calcium channels in the rat hippocampus. Neuropharmacology, 44(2): 282–292. Simmons, M.L., Terman, G.W., Gibbs, S.M. and Chavkin, C. (1995) L-type calcium channels mediate dynorphin neuropeptide release from dendrites but not axons of hippocampal granule cells. Neuron, 14(6): 1265–1272. Sloviter, R.S. (1983) ‘‘Epileptic’’ brain damage in rats induced by sustained electrical stimulation of the perforant path. I. Acute electrophysiological and light microscopic studies. Brain Res. Bull., 10(5): 675–697. Smialowska, M. (1988) Neuropeptide Y immunoreactivity in the locus coeruleus of the rat brain. Neuroscience, 25(1): 123–131. Smith, Y. and Parent, A. (1986) Neuropeptide Y-immunoreactive neurons in the striatum of cat and monkey: morphological characteristics, intrinsic organization and co-localization with somatostatin. Brain Res., 372(2): 241–252. Smith, Y., Parent, A., Kerkerian, L. and Pelletier, G. (1985) Distribution of neuropeptide Y immunoreactivity in the basal forebrain and upper brainstem of the squirrel monkey (Saimiri sciureus). J. Comp. Neurol., 236(1): 71–89. Sperk, G., Marksteiner, J., Gruber, B., Bellmann, R., Mahata, M. and Ortler, M. (1992) Functional changes in neuropeptide Y- and somatostatin-containing neurons induced by limbic seizures in the rat. Neuroscience, 50(4): 831–846. Stanic, D., Brumovsky, P., Fetissov, S., Shuster, S., Herzog, H. and Hokfelt, T. (2006) Characterization of neuropeptide Y2 receptor protein expression in the mouse brain. I. Distribution in cell bodies and nerve terminals. J. Comp. Neurol., 499(3): 357–390. Sun, Q.Q., Akk, G., Huguenard, J.R. and Prince, D.A. (2001a) Differential regulation of GABA release and neuronal excitability mediated by neuropeptide Y1 and Y2 receptors in rat thalamic neurons. J. Physiol., 531(1): 81–94.

Sun, Q.Q., Huguenard, J.R. and Prince, D.A. (2001b) Neuropeptide Y receptors differentially modulate G-protein-activated inwardly rectifying K+ channels and high-voltageactivated Ca2+ channels in rat thalamic neurons. J. Physiol., 531(1): 67–79. Sundler, F., Moghimzadeh, E., Hakanson, R., Ekelund, M. and Emson, P. (1983) Nerve fibers in the gut and pancreas of the rat displaying neuropeptide-Y immunoreactivity. Intrinsic and extrinsic origin. Cell Tissue Res., 230(3): 487–493. Tatemoto, K., Carlquist, M. and Mutt, V. (1982) Neuropeptide Y—a novel brain peptide with structural similarities to peptide YY and pancreatic polypeptide. Nature, 296(5858): 659–660. Tu, B., Timofeeva, O., Jiao, Y. and Nadler, J.V. (2005) Spontaneous release of neuropeptide Y tonically inhibits recurrent mossy fiber synaptic transmission in epileptic brain. J. Neurosci., 25(7): 1718–1729. Vezzani, A., Sperk, G. and Colmers, W.F. (1999) Neuropeptide Y: emerging evidence for a functional role in seizure modulation. Trends Neurosci., 22(1): 25–30. Widerlov, E., Lindstrom, L.H., Wahlestedt, C. and Ekman, R. (1988) Neuropeptide Y and peptide YY as possible cerebrospinal fluid markers for major depression and schizophrenia, respectively. J. Psychiatr. Res., 22(1): 69–79. Whittaker, E., Vereker, E. and Lynch, M.A. (1999) Neuropeptide Y inhibits glutamate release and long-term potentiation in rat dentate gyrus. Brain Res., 827(1-2): 229–233. Wolak, M.L., DeJoseph, M.R., Cator, A.D., Mokashi, A.S., Brownfield, M.S. and Urban, J.H. (2003) Comparative distribution of neuropeptide Y Y1 and Y5 receptors in the rat brain by using immunohistochemistry. J. Comp. Neurol., 464(3): 285–311. Woldbye, D.P., Larsen, P.J., Mikkelsen, J.D., Klemp, K., Madsen, T.M. and Bolwig, T.G. (1997) Powerful inhibition of kainic acid seizures by neuropeptide Y via Y5-like receptors. Nat. Med., 3(7): 761–764.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 18

Norepinephrine and the dentate gyrus Carolyn W. Harley Department of Psychology, Memorial University of Newfoundland, St. John’s, NL, A1B 3X9, Canada

Abstract: Norepinephrine’s role in the dentate gyrus is assessed based on a review of what is known about its innervation and receptor patterns and its functional effects at both cellular and behavioral levels. The data support seven hypotheses: (1) Norepinephrine’s functional actions are primarily mediated by b adrenoceptors and include electrophysiological enhancement of responses to excitatory input and glycogenolytic metabolic support of excitatory synaptic activity. (2) At the cellular level, locus coeruleus burst release of norepinephrine transiently inhibits feedforward interneurons and either excites or inhibits subpopulations of feedback interneurons. Consistent with reduced feedforward inhibition, granule cell firing is transiently increased. Concomitant EEG effects include transient increases in theta power and decreases in beta and gamma power. (3) Norepinephrine selectively promotes the processing of medial perforant path spatial input. This effect is mediated both through short- and long-term potentiation of cell excitability and through delayed potentiation of synaptic input. A critical level of norepinephrine release is required for long-term effects to norepinephrine alone. Norepinephrine release switches early phase frequency-induced long-term potentiation of perforant path input to an enduring late phase form and can reinstate decayed long-term potentiation. Norepinephrine also promotes frequency-induced potentiation of granule cell output at the mossy fiber to CA3 connection. (4) Local increases in norepinephrine accompany glutamate release and release of other neurotransmitters providing a mechanism for norepinephrine enhancement effects independent of locus coeruleus firing. (5) Stimuli, such as novelty and reward and punishment, which activate locus coeruleus neurons, enhance responses to medial perforant path input and engage late phase frequency-induced long-term potentiation through b adrenoceptor activation. (6) Behavioral studies are consistent with the mechanistic evidence for a norepinephrine role in promoting learning and memory and assisting retrieval. (7) The overall profile suggests lower levels of norepinephrine may facilitate pattern completion or memory retrieval while higher levels would recruit global remapping and promote long-term episodic memory. Keywords: locus coeruleus; LTP; LTD; novelty; glycogen metabolism; theta EEG; gamma EEG; global remapping; feedback inhibition; feedforward inhibition; alpha adrenoceptors; beta adrenoceptors; medial perforant path; lateral perforant path; synaptic plasticity DG cells, and reviews NE’s role in promoting long-term plasticity, based primarily on rodent data.

The present chapter examines norepinephrine (NE) innervation and receptor patterns in the dentate gyrus (DG), considers NE’s effects on Corresponding author. Tel.: +1 709 737 7974; Fax: +1 709 737 4000; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63018-0

299

300

NE innervation Blackstad first characterized NE innervation of the rat hippocampus, reporting the densest NE innervation in the hilus of the DG, particularly the subgranular zone (Blackstad et al., 1967). NE fibers were less common in cell body layers, suggesting axodendritic contacts. Within the molecular layer, the dendritic zone of DG granule cells, NE fibers were denser in the middle molecular layer, a target of medial perforant path fibers from the medial entorhinal cortex, as compared to the outer molecular layer, the target of lateral entorhinal fibers. With retrograde tracing, the origin of the DG NE innervation was identified as multipolar and fusiform cells in the dorsal part of the dorsal pontine locus coeruleus (LC) (Haring and Davis, 1985). Electron microscope (EM) images of the NE fibers in DG reveal small granular vesicles, indicating NE content in
0.4–1.2 mm diameter varicosities. Each varicosity contains
20 vesicles, of which 55% can be classified as small granular vesicles. Large granular vesicles (6%) and a few mitochondria are also present (Koda and Bloom, 1977; Milner and Bacon, 1989b). Intervaricosity axons are 0.1–0.15 mm in diameter (Milner and Bacon, 1989b). In 6600 mm2 samples, the hilus has 1 NE bouton (varicosity) for every 400 boutons of other types; while the granule cell layer has 1/500, and the molecular layer has 1/ 4000 NE boutons. Total NE bouton density differs among regions. An average of 1500 and 2000 boutons are located in hilar and molecular layer samples, respectively, while only 500 boutons are present in each granule cell layer sample (Koda and Bloom, 1977). These estimates of NE innervation density may be compared to estimates of 1/8800–1/14,500 boutons for the neocortex (Lapierre et al., 1973). The majority of NE axon contacts in DG end on small dendrites in the subgranular layer where a third make symmetric synapses, a third form asymmetric synapses, and a third type make close associations without specializations. Terminals in the molecular layer have a similar typology. Thirty percent of terminals end at spines and do not have specialized synaptic profiles. On large dendrites

and granule cell somata, symmetric synapses and appositions are typical (Milner and Bacon, 1989b). Biochemically, NE content in the hippocampus is highest in DG (about twice that in CA1) with higher levels in ventral (600 ng/g) compared to dorsal DG (360 ng/g) (Loy et al., 1980) or similar values (500 ng/g) in both regions (Hortnagl et al., 1991) depending on the study. Quantification of varicosities in DG (2.4 million/cubic mm) also suggest twice as many varicosities there as in CA1 (Oleskevich et al., 1989); however when divided by cell number, the ratio of NE varicosities to cells (20–40/1) is smaller in DG than CA1 (180/1). The ratio is lowest in the granule cell layer itself (2–4/ 1), consistent with the low total number of varicosities in this layer. Seventy percent of the input from the LC to DG courses through the cingulum, with the fornix and ventral amygdaloid pathways each accounting for
15% (Loy et al., 1980). Developmentally, the NE input, as characterized by dopamine-b-hydroxylase immunocytochemistry, is very sparse in the DG at four days postnatal, and is heaviest at this time point in CA1. By 10 days postnatal, a band of fibers in the subgranular zone becomes detectable (Moudy et al., 1993). By 21 days, a mainly adult-like pattern is developed.

Adrenoceptor distribution Both a and b adrenoceptors occur in DG. In an early study using several methods to assess adrenoceptors, Crutcher and Davis (1980) reported b adrenoreceptors at uniform levels in the dentate (146 fmol/mg) and hippocampal gyri (178 fmol/ mg), while a1 receptors were more concentrated in the DG (269 fmol/mg). The uniformity of b adrenoceptor distribution led the authors to suggest these receptors might occur separately from LC terminals, while a1 adrenoreceptors showed better correspondence with LC innervation. Both receptors occurred most densely in the synaptosomal/ mitochondrial fractions. The a1 receptor population was not changed by 6-OHDA lesions (U’Prichard et al., 1980) suggesting a postsynaptic localization.

301

b Adrenoceptors Milner’s more recent work (Milner et al., 2000) using both EM and light microscope methodologies, suggests NE adrenoceptor distribution is best identified using EM. Her quantified EM receptor counts show laminar differences in b adrenoceptors that were not evident in light microscope studies. Approximately 500 b adrenoceptor-immunoreactive profiles per 6000 mm2 sample occur in the middle molecular layer, while only
250 profiles are seen in the inner and outer molecular layers, and in the subgranular and hilar zones;
125 b adrenoceptor-immunoreactive profiles/sample occur in the granule cell layer. The molecular layer distribution shows interesting parallels with b adrenoceptor physiological effects (see Physiology, below) and is consistent with Blackstad’s initial description of NE fiber innervation. Granule cell somata express b adrenoceptor immunoreactivity in the endoplasmic reticulum, consistent with granule cell production of b adrenoceptors. A few hilar interneurons, including those reactive for parvalbumin, which marks basket cells (Kosaka et al., 1987; Ribak et al., 1990, 1992), are also immunoreactive for b adrenoceptors. In the molecular layers, 40–50% of the b adrenoceptor reactive profiles are associated with dendrites, and 50–60% with astrocytes. In the granule cell layer, 30% are somatic, dendritic and astrocytic, respectively. In the subgranular region (an
55 mm zone below the granule cells), 65% are astrocytic, while 25% are dendritic. Below the subgranular zone, in the hilar region, 30% of the b adrenoceptor reactive sites are in axons or axon terminals and over 50% are in astrocytes. Some b adrenoceptors are present within parvalbumin-immunoreactive terminals. Reactive sites in axons are less common outside the hilar region, comprising only 2–10% of the b adrenoceptor-reactive sites in other layers (Milner et al., 2000). Dendritic receptors are associated with postsynaptic densities on both large (inner molecular layer) and small (middle and outer molecular layers) dendritic profiles. Immunoreactive sites also

occur in spines. Axonal b adrenoceptors occur in both axons and in axon terminals ending on spines (Milner et al., 2000). The b adrenoreceptors on astrocytes are usually next to terminals that make asymmetric synapses on dendrites and which are thought to be excitatory inputs. Many of the receptors on astrocytes are closely apposed to the terminals that form synapses on spines (Milner et al., 2000). The astrocytic adrenoceptor distribution, in close association with synapses, may support glycogen breakdown at active synaptic sites (see Physiology, below). Astrocytic b adrenoceptors also occur around blood vessels. NE axons, identified by tryrosine hydroxylase immunoreactivity, occur close (1–2 mm) to both astrocytic and dendritic b adrenoceptors, but direct contacts were not observed (Milner et al., 2000). The b adrenoceptors in DG are wellplaced to modify granule cell function, and some interneurons, either directly through postsynaptic receptors or indirectly, through effects on glial processes near synapses. The common postsynaptic location of b adrenoceptors on the dendrites of granule cells near, or with, asymmetric synaptic input and their predominance in the middle molecular layer, is consistent with a selective modulation of responses to glutamatergic input from the medial entorhinal cortex to the DG (see Physiology, below). Activation of presynaptic axonal b adrenoceptors and, possibly, astrocytic b adrenoreceptors, may play a role in the increase in glutamate release also reported with b-adrenoceptor activation (see Physiology, below). Molecular and binding studies suggest the majority of b adrenoceptors in the DG are b1 adrenoceptors (Minneman et al., 1979; Rainbow et al., 1984), but both b1 and b2 adrenoceptors occur in the molecular layer and hilus (Duncan et al., 1991; Booze et al., 1993) and granule cells produce mRNA for both b1 and b2 adrenoceptors with a stronger signal from the b2 mRNA probe (Nicholas et al., 1993b). The hippocampus has a high proportion of high-affinity b adrenoceptors, consistent with a strong functional role for these receptors (Arango et al., 1990).

302

a Adrenoceptors Three a1 receptor subtypes are recognized. Specific probes for each (a1a, a1b, a1d) reveal that granule cells have an intense mRNA signal for the a1d subtype (Pieribone et al., 1994; Day et al., 1997), but do not contain the a1b subtype mRNA. The a1a subtype is restricted to polymorph cells of the hilar region (Day et al., 1997). The a1 receptor is described as half as dense in hippocampus as in neocortex, despite the fact that the hippocampus has twice the NE innervation density of neocortex (Zilles et al., 1991). Zilles reports that a1 receptors are more concentrated in the inner 2/3rd of the molecular layer, but that overall a1 receptor density is similar in molecular, granule and hilar zones and does not correspond with NE innervation (both a1a and a1b receptors are represented at similar levels in contrast to the mRNA data). Zilles’ a1 pattern differs from earlier work using a different a1-ligand in which a1-receptor density closely followed NE fiber density (Jones et al., 1985). Since neither a2a nor b adrenoceptors are in clear synaptic association with NE terminals, a1 adrenoceptors are candidates, by default, for sites of synaptic specialization for tyrosine hydroxylase fibers. No a1 adrenoceptor subtype distribution has been described at the EM level in DG. Milner (Milner et al., 1998) characterized the a2a receptor at the light microscope and EM levels and describes its distribution as complementary to the b adrenoceptors in that the majority of receptors are presynaptic — in axons and axon terminals — rather than postsynaptic. Nonetheless, a2a receptors exist both pre- and postsynaptically, and while a2a receptors are implicated in the negative feedback regulation of NE release, most presynaptic profiles are in unmyelinated non-noradrenergic axons. Granule cells demonstrate immunoreactivity for a2a-like receptors in association with endoplasmic reticulum, suggesting granule cell synthesis of a2a receptors. The a2a-immunoreactive sites are evenly distributed in the middle and outer molecular layers with about 40% on axons, 30% on astrocytes and another 30% on dendrites or spines. There are about 1/3rd fewer immunoreactive a2a receptors in the inner molecular layer with only

10% on axons, 50% on dendrites and 40% on glia. Throughout the molecular layer, a2a receptors on dendrites are primarily on spines at asymmetric synapses. Astroctyic a2a receptors are also near asymmetric synapses. No direct synaptic contacts between the postsynaptic a2a receptors and tyrosine hyroxylase reactive axons are observed. The granule cell layer has the fewest a2a profiles (2/3rd less than the molecular layer) and 50% are somatic profiles. In the subgranular zone, there is a paucity of dendritic profiles, with mainly axonal (44%) and glial (46%) profiles. In the hilar region 60% are in axonal elements. The remaining a2a profiles are in glia. The axonal profiles throughout are predominantly in non-noradrenergic axons. A small subpopulation of hilar interneurons, with characteristics of the somatostatin interneuron subtype, contained a2a receptors (Milner et al., 1998). In situ studies examining mRNA for the three a2 receptor subtypes report more of a2c subtype in DG granule cells than a2a (Nicholas et al., 1993a; Scheinin et al., 1994) and none of a2b subtype. A mouse study did find evidence of granule cell a2b receptor expression, with the most signal in more mature granule cells (located at the border with the inner molecular layer) and none in the subgranular zone (Wang et al., 2002). Developmental studies report that the a2a mRNA signal is moderate at 1–3 days postnatal and then becomes weak after maturity (Winzer-Serhan et al., 1997a), while the a2c signal intensity is high from 11 to 14 days postnatal, and remains a moderately-intense signal through adulthood (Winzer-Serhan et al., 1997b). It would be of interest to know the localization of a2c receptors.

Physiology Glycogen metabolism The astrocytic metabolic enzyme, glycogen phosphorylase, which breaks down glycogen to glucose, is strongly activated by NE in DG, as demonstrated pharmacologically using an a2 adrenoceptor antagonist (Fara-On et al., 2005), or physiologically, using glutamate activation of the

303

LC (unpublished observations). In 1971, Phelps found high levels of glycogen accumulation in astrocytic processes near synapses in the DG in barbiturate-treated mice (Phelps, 1972). Earlier evidence that NE release depleted glycogen, while suppression of NE release increased glycogen, led Phelps to hypothesize that NE controlled glycogen breakdown in the DG at sites of synaptic activation through b adrenoceptors. This hypothesis is consistent with the common association of astrocytic b adrenoceptors and asymmetric synaptic contacts. At the light microscope level, a patchy, modular distribution of glycogen phosphorylase, which coexists with glycogen, is evident in the DG. Patches are more numerous during the dark phase of the daily cycle, when rodents are aroused, and predominate (
60%) in the middle molecular layer, with the remaining patches equally divided between inner and outer layers (Harley and Rusak, 1993). Glycogen phosphorylase activity is reduced during the theta EEG state, possibly in relation to a higher overall inhibition in the DG during theta rhythm (Uecker et al., 1997). NE activation of glycogen phosphorylase and glycogenolysis in vitro requires b adrenoceptor coupled cAMP activation (Edwards et al., 1974; Nahorski et al., 1975; Quach et al., 1978). NE also increases astrocyte metabolism through a adrenoceptor activation (Subbarao and Hertz, 1991).

Intracellular recording Granule cells Three studies have examined NE effects in vitro. Haas and Rose found that NE and the b adrenoceptor agonist, isoproterenol, produced a small depolarization of the membrane potential, an attenuated afterhyperpolarization, and a decrease in accommodation (Haas and Rose, 1987). The attenuated afterhyperpolarization was attributed to a cAMP-mediated block of Ca++-activated K+ currents. In a few granule cells, a b adrenoceptor initiated block of the transient outward K+ current (A-type current) was reported. Blocking the slow Ca++-activated afterhyperpolarization slows the

decay of excitatory postsynaptic currents (EPSCs) and increases temporal summation in other hippocampal principal cells (Lancaster et al., 2001). Gray and Johnston blocked K+ currents and showed that b adrenoceptor activation increased a voltage-dependent Ca++ current, which is activated when cells are depolarized to at least 15 mV, and is therefore likely to be L-type current (Gray and Johnston, 1987). They showed an increase in Ca++ channel open times and a dependence on cAMP elevation. No changes in the N-type Ca++ current were reported. The a2 agonist, clonidine, was without effect. NE also produced an increase in the amplitude and duration of Ca++-dependent action potentials, which contribute to back-propagating action potentials and spike-timing dependent plasticity (Dan and Poo, 2006). Thus, NE might be expected to enhance this form of plasticity. Intracellular calcium is also increased in granule cells by dose-dependent activation of a1 adrenoceptors (Kusaka et al., 2004). Lacaille and Schwartzkroin used focal application of a b adrenoceptor agonist in hippocampal slices to evaluate the effects of NE. They replicated the depolarization of the granule cell membrane potential (Lacaille and Schwartzkroin, 1988). They also showed that the depolarization was accompanied by an increased input resistance, which they attributed to a blockade of K+ channels that are open under normal conditions. Collectively, one would predict that these b adrenoceptor effects would enhance the excitability and responsivity of granule cells to afferent input. No other adrenoceptor subtype has been implicated in direct modulation of granule cell membrane characteristics.

Hilar interneurons Misgeld and Bijak sampled hilar interneurons in the subgranular zone (Bijak and Misgeld, 1995). Aspiny interneurons, which are GABAergic and inhibitory, have strong afterhyperpolarizations and little spike accommodation. EM studies show they receive NE input (Milner and Bacon, 1989a). Spiny hilar neurons, which are glutamatergic and

304

excitatory (mossy cells), have strong accommodation and little afterhyperpolarization. NE blocked both aspiny neuron afterhyperpolarizations and spiny neuron accommodation through b adrenoceptors. NE, or a b adrenoceptor agonist, increased the firing frequency of the excitatory spiny neurons to current injection, but did not alter the response of the aspiny inhibitory interneurons to current injection. Spontaneous firing of both interneuron types increased with NE or b adrenoceptor activation, as did the frequency of excitatory postsynaptic potentials (EPSPs) and inhibitory postsynaptic potentials (IPSPs). IPSPs increased in neighboring granule cells with increased activity in inhibitory interneurons. With glutamate release blocked, GABA-A receptor-mediated IPSPs still increased in granule cells, but NE no longer increased inhibitory interneuron firing. a adrenoceptor agonists suppressed inhibitory interneuron activity. Thus, NE activation of both b adrenoceptors and a adrenoceptors either did not change or else suppressed, inhibitory interneurons. The enhancement of spontaneous IPSPs, without increased inhibitory interneuron firing, suggests presynaptic b adrenoceptors enhance GABA release. Presynaptic b adrenoceptors also appear to enhance glutamate release. The outcome of natural NE release on inhibitory interneurons, and network inhibition, will reflect the balance of a and b adrenoceptor activation on all elements of the circuit. Extracellular unit recording (see Extracellular unit recording, below) shows that the effects of NE naturally released include decreases and increases in inhibitory interneuron firing, depending on the subtype of the interneuron.

Extracellular unit recording Segal and Bloom were the first to record changes in unit activity in DG in response to electrical stimulation of the LC (Segal and Bloom, 1976a). The cells they monitored had relatively high firing rates and therefore were likely to be interneurons (Jung and McNaughton, 1993). LC stimulation inhibited this cell type in anesthetized rats. In awake rats, a loud tone also inhibited these DG

cells, but when paired with sweet milk it evoked excitation; LC stimulation enhanced the cell excitation as well as inhibition (Segal and Bloom, 1976b). This suggests a circuit-dependent LC modulation of interneuron responses. The inhibition appeared to be a direct effect of NE release, since LC cells transplanted to a denervated hippocampus also inhibited spontaneously active units in the DG. A brief stimulus train to the LC transplant inhibited DG cell firing for
30 s (Bjorklund et al., 1979). This inhibition was antagonized by a b adrenoceptor antagonist. In two pharmacological studies, theta interneurons (putative inhibitory interneurons) in the hilus and near the granule cell layer were excited by NE and by a2 and b agonists, while granule cells were inhibited by NE and by a1 agonists, but excited by a2 and b agonists (Pang and Rose, 1987; Rose and Pang, 1989). The excitatory effects of a2 receptors appeared to be postsynaptic, since they occurred in the presence of high concentrations of magnesium in the extracellular buffer. A more recent study using glutamatergic activation of LC neurons demonstrates both inhibitory and excitatory effects on subgranular and hilar interneurons with natural NE release (Brown et al., 2005). Using NE release from LC activation (natural NE release) has the advantage of revealing in situ actions of NE rather than pharmacological actions, but non-NE effects may also be recruited. The nature of the LC-NE effect depended on the physiological identity of the interneurons. Feedforward interneurons, activated by perforant path input from the entorhinal cortex and firing with a lower threshold than granule cells, were consistently inhibited (
60 s) by brief LC activation (see Fig. 1). Feedback interneurons, recruited only after granule cells discharged, were either excited or inhibited, suggesting differences between subpopulations. Consistent with inhibition of feedforward interneurons was the observation of a strongly increased granule cell response after LC activation. These results reveal that natural release of NE produces selective modulation of DG circuitry. More work is needed to classify the feedback interneurons that respond differentially to NE release.

305

Fig. 1. Normalized firing rate of 16 feedforward interneurons in the dentate gyrus following ejection of glutamate in the locus coeruleus demonstrated the robust effects on this cell type. Glutamate was ejected at 0 min. Adapted with permission from The Society for Neuroscience, 2005.

EEG recording In the same study (Brown et al., 2005), LC activation enhanced DG EEG theta in the 4–8 Hz range while suppressing beta (12–20 Hz) and gamma (20–40 Hz) frequencies (see Fig. 2). This result appears consistent with concomitant inhibition of the feedforward interneuron population, which potentially mediates the more distant influences implicated in beta oscillations (Kopell et al., 2000), and excitation of a feedback subpopulation enhancing theta amplitude as reported for 5-HT modulation (Nitz and McNaughton, 1999). Gray originally proposed that LC activation increases hippocampal theta power specifically in a 7.5–7.8 Hz range in awake rats (Gray and Ball, 1970). He observed this increase when rats were responding to novelty or a disconfirmation of expectation. The low driving threshold for this frequency range was lost following NE depletion or blockade (Gray et al., 1975). We have observed that this theta frequency (7.5–7.8 Hz) increases with LC activation in awake rats (unpublished observations). Segal reported that electrical LC stimulation induces theta burst firing in medial septal neurons (Segal, 1976). Berridge and Foote later showed that tonic cholinergic LC activation produces a dramatic increase in hippocampal theta in anesthetized rats (Berridge and Foote, 1991), which depends on b adrenoceptors in medial septum (Berridge et al., 1996).

A pattern of enhanced theta and suppressed beta/gamma activity is consistent with the hypothesis that disengagement from established representations and enhancement of processes that promote incorporation of new information is an effect of LC activation (Brown et al., 2005). Consistent with this view, Bouret and Sara have proposed ‘network resetting’ as a primary LC function (Bouret and Sara, 2005).

Evoked potential recording NE potentiates the perforant path-evoked population spike both transiently, and in a sustained manner, depending on the degree of NE exposure in the DG. Transient potentiation of the perforant path-evoked population spike amplitude, which is b adrenoceptor-dependent, has been reported in numerous in vivo experiments after either exogenous NE application (Neuman and Harley, 1983; Winson and Dahl, 1985; Babstock and Harley, 1993; Harley et al., 1996; Chaulk and Harley, 1998) or activation of LC (Dahl and Winson, 1985; Harley and Milway, 1986; Harley et al., 1989; Washburn and Moises, 1989; Babstock and Harley, 1992; Frizzell and Harley, 1994; Klukowski and Harley, 1994). NE or b adrenoceptor agonists in vitro increase the perforant pathevoked EPSP slope, as well as the population spike, although enhanced E-S coupling is also observed (Lacaille and Harley, 1985). The duration

306

Fig. 2. Frequency-dependent increases and decreases in dentate gyrus EEG in the 4–40 Hz range following ejection of glutamate in the locus coeruleus. Asterisks indicate significant effects at po0.05. Upward arrow denotes time of ejection. Adapted with permission from The Society for Neuroscience, 2005.

of in vitro effects are transient when agonist concentrations are low or briefly applied (Lacaille and Harley, 1985; Stanton and Sarvey, 1985c). The discovery of sustained NE-induced increases in the perforant path-evoked potential or NE-induced long-term potentiation (NE-LTP) (Neuman and Harley, 1983) was the first direct evidence that LC-NE could provide a heterosynaptic signal to initiate long-lasting increases in DG responses to glutamate-mediated information. NE-LTP is consistent with Kety’s early hypothesis, based on pharmacological and behavioral evidence, that LC-NE strengthens brain responses to significant stimuli (Harley, 1987), and NE-LTP will be examined in detail in the next section.

NE-induced plasticity Spike potentiation in anesthetized rats While studies at the cellular and EEG level suggest transient LC-NE modulation of the DG network, studies of the perforant path-evoked potential, as noted, reveal a sustained component of LC-NE modulation of the DG response to entorhinal input. Iontophoresed NE was first reported in 1983 to produce a long-lasting potentiation of the perforant path-evoked DG population spike in anesthetized rats that endured for hours (Neuman and Harley, 1983). NE-induced LTP of the population spike has been repeatedly confirmed in vivo using

307

direct NE application (Winson and Dahl, 1985; Harley et al., 1996; Chaulk and Harley, 1998), LC activation by glutamate (Harley and Milway, 1986; Harley and Sara, 1992; Klukowski and Harley, 1994; Walling and Harley, 2004) and LC activation by orexin (Walling et al., 2004). Repeated LC electrical stimulation can also induce LTP of the perforant path-evoked potential (Harley et al., 1989). LC electrical stimulationinduced potentiation is controversial, however, with respect to dependence on b adrenoceptor activation. There is evidence both for (Washburn and Moises, 1989) and against (Harley et al., 1989) such dependence. DG evoked response potentiation associated with glutamate- or orexin-induced activation of LC neurons is consistently blocked by b adrenoceptor antagonists, either systemicallyadministered (Harley and Milway, 1986; Harley et al., 1989; Babstock and Harley, 1992; Walling and Harley, 2004; Walling et al., 2004) or locally-delivered to the DG (Harley and Evans, 1988). The requirement for DG b adrenoceptor activation is consistent with the intracellular evidence that b adrenoceptor activation increases the excitability of granule cells to exogenous input. NE-LTP does not depend, however, on continued exposure to NE, as demonstrated by the efficacy of brief iontophoretic applications (Neuman and Harley, 1983; Winson and Dahl, 1985) and by measurement of hippocampal NE during NE-LTP induced by exogenous NE application (Harley et al., 1996) or LC activation (Walling et al., 2004). (For further discussion, see LC firing and NE release patterns, below).

Synaptic and spike potentiation in vitro In vitro studies of NE and perforant path stimulation differ from in vivo studies because potentiation of both synaptic and population spike components of the perforant path-evoked potential occurred consistently in vitro but not in vivo (Lacaille and Harley, 1985; Stanton and Sarvey, 1987). In vitro NE-LTP depends on concomitant NMDA receptor activation (Burgard et al., 1989; Sarvey et al., 1989), as well as on b1 adrenoceptor activation and protein synthesis (Stanton and

Sarvey, 1985c), strongly suggesting an associative component to NE-LTP. However in vitro studies failed to find evidence that pairing electrical stimulation of perforant path input with NE bathapplication was critical for NE-LTP (Lacaille and Harley, 1985; Dahl and Sarvey, 1990) (but see Spike potentiation and pairing requirements, below). Frizzell and Harley found an NMDA receptorindependent potentiation of EPSP slope and population spike using LC-NE activation and ketamine anesthesia in vivo, but potentiation was typically short-lived (Frizzell and Harley, 1994). In vitro, Dahl reported that two applications of a low b adrenoceptor agonist dose (75 nM), spaced by 30 min, which, individually, elicit only weak transient potentiation, induce NE-LTP of the perforant path population spike only, and NMDA receptor activation was not required (Dahl and Li, 1994). This effect of a spaced NE agonist suggests a mechanism by which LC-NE might contribute to a stronger memory with spaced training.

Pathway selectivity The in vitro NE-LTP of EPSP slope occurred only for medial perforant path input, while the same application of NE with an a adrenoceptor antagonist, or of a b adrenoceptor agonist alone, produced a long-term depression (LTD) of the lateral perforant path EPSP slope (Dahl and Sarvey, 1989). It may be relevant for this selective effect that b adrenoceptors are densest in the terminal zone of the medial perforant path (Milner et al., 2000) and that enhancement of back-propagating calcium spikes by NE would be more likely to induce LTP in the adjacent medial perforant path target than the more distal lateral perforant path target (Dan and Poo, 2006). A possible explanation for the failure to induce NE-LTP of the field EPSP in vivo could be that stimulation in vivo, particularly the stronger stimulation that is required to evoke a DG population spike, would activate a mixture of medial and lateral perforant path fibers. NE-LTP in vivo occurs for short-latency population spikes, which reflect medial perforant path activation. In one attempt to examine the selectivity of the

308

NE effect in vivo, lateral perforant path input was activated by stimulating the lateral olfactory tract. The synaptic input produced by lateral olfactory stimulation was decreased by pairing with paragigantocellularis stimulation, a major glutamatergic afferent pathway to the LC (Babstock and Harley, 1993). Depression of the lateral perforant path input depended on b adrenoceptor activation. Since paragigantocellularis stimulation enhances the medial perforant path population spike through a b adrenoceptor-dependent mechanism (Babstock and Harley, 1992), the combined results are consistent with enhanced medial perforant path input and depressed lateral perforant path input with LC-NE activation in vivo. One functional interpretation of selectivity is suggested by a new perspective on the entorhinalhippocampal complex, which suggests two forms of spatial mapping occur in hippocampus. Global remapping, reflecting a completely new context, is associated with changes in medial perforant path ‘grid cell’ input; while rate remapping, which reflects firing changes in the same place cell map, is ascribed to alterations in nonspatial sensory input mediated by the lateral perforant path (McNaughton et al., 2006). If strong LC-NE activation potentiates medial, and depresses lateral, perforant path inputs, it would favor global remapping in DG over rate remapping. Global remapping would provide a new context for memory storage. It should be noted that high frequency-induced LTP of either the medial or lateral perforant path requires b adrenoceptor activation (Bramham et al., 1997). Thus, while b adrenoceptors may gate weaker inputs to favor the medial perforant path, when signals are strong, plasticity should be promoted by b adrenoceptor activation in both pathways similarly.

Spike potentiation and pairing requirements In vivo experiments provide evidence that LC-NE activation induces long-lasting increases in the medial perforant path-evoked population spike, suggesting that excitability in the postsynaptic

granule cells increases. Although mixed medial and lateral perforant path input is a possible explanation of the lack of EPSP potentiation observed in vivo, the ubiquity of studies that see no change in EPSP slope suggests an increase in E-S coupling, or granule cell excitability, is a prominent feature of LC-NE modulation. Recent evidence provides support for the hypothesis that an increase in intrinsic excitability, through such mechanisms as reduced K+ conductance in activated dendrites (Frick et al., 2004), is an important component of associative plasticity functioning similar to, and independent of, increases in synaptic drive per se (Daoudal and Debanne, 2003). Evidence for associative plasticity in LC-NE activation effects on DG has been lacking. In vitro studies have not found that NE-LTP requires concurrent pairing of bath-applied NE and perforant path stimulation (Lacaille and Harley, 1985; Dahl and Sarvey, 1990). However, we have recently found that pairing of LC glutamate activation and perforant path input is critical for NE-LTP in vivo (Reid and Harley, 2005). With LC activation 10 min after cessation of perforant path stimulation and 10 min prior to resumption of perforant path stimulation, no NE-LTP occurs. The same LC activation during concurrent perforant path stimulation produces reliable NE-LTP. Both EPSP slope and spike potentiation occurred in our in vivo study (Reid and Harley, 2005). The importance of timing in LC-NE activation for potentiation effects is consistent with evidence that conduction velocity in LC-NE axons varies across species to maintain a constant delay-to-signal arrival in forebrain structures (Aston-Jones et al., 1985). Two differences may be noted between the in vitro and in vivo pairing experiments. First, there was no delay between the offset of perforant path stimulation and the beginning of NE perfusion in the in vitro studies. Second, in vitro bath application of NE is associated with an extended elevation of cAMP levels in DG, permitting perforant path interaction with cAMP effects even at late time points (Stanton and Sarvey, 1985b). This is unlikely to be the case with LC activation (Siggins et al., 1973).

309

Delayed long-term synaptic potentiation in awake rats In contrast to the evidence that NE induces rapid potentiation of perforant path synaptic input from in vitro studies is new in vivo data from awake rats in which potentiation of perforant path synaptic input is first observed 24 h after LC activation (Walling and Harley, 2004). In the Walling and Harley study, LC burst activation initially had no effect on synaptic input, but induced the usual rapid and enduring potentiation of the perforant path-evoked population spike. Twenty-four hours later, both the perforant path field EPSP and the population spike were strongly potentiated (Fig. 3). The 24 h increase in field EPSP slope predicted the 24 h increase in population spike amplitude. This delayed NE-LTP of EPSP slope and population spike amplitude depended on b adrenoceptor activation at the time of LC activation, and on protein synthesis. Aplysia also demonstrates a 24 h delayed potentiation of synaptic input in response to heterosynaptic activation of a cAMP-dependent cascade (Brunelli et al., 1976; Schacher et al., 1988). Walling and Harley suggest mammalian delayed synaptic potentiation implies a special role for DG NE modulation in long-term memory. Such an association is likely to require relatively strong NE release. NE applied intracerebroventricularly and

measured by microdialysis in the DG is only associated with NE-LTP in the DG when the NE increase exceeds a critical threshold, estimated to be
400 nM, intrasynaptically (Harley et al., 1996).

LC firing and NE release patterns LC neurons fire tonically at rates from less than 1 to 5 Hz (Aston-Jones and Bloom, 1981b) as a function of level of arousal. They also fire phasically in bursts in which 2 or 3 spikes occur in rapid succession (10–15 Hz) followed by 200–500 ms pauses (Aston-Jones and Bloom, 1981a). Burst events are typically associated with responses to environmental stimuli (Aston-Jones and Bloom, 1981a). Harley and Sara have shown that glutamate activation of LC produces a strong burst followed by a pause, lasting minutes, in LC firing (Harley and Sara, 1992). Higher levels of NE release are associated with phasic rather than tonic firing patterns (Florin-Lechner et al., 1996). Glutamate administration to the LC in vivo increases NE release in hippocampus by
200%, measured by microdialysis, in the first 20 min after LC activation, subsequently, NE levels return to baseline (Walling et al., 2004). Orexin A is also a potent activator of LC neurons, and its administration to the LC led to a similar increase in NE, restricted to the first 20 min (Walling et al., 2004).

Fig. 3. Delayed potentiation of the perforant path-evoked field EPSP with immediate and delayed potentiation of the perforant pathevoked population spike in awake rats following infusion of glutamate in the locus coeruleus at the arrow. Adapted with permission from The Society for Neuroscience, 2005.

310

Both orexin A and glutamate activation of LC produce a b adrenoceptor-dependent increase in population spike amplitude that lasts for hours and is thus triggered, but not maintained by, hippocampal NE release. High frequency electrical stimulation of the perforant path also increases NE release in the DG (Bronzino et al., 2001), and may contribute to the level and duration of LTP induced (Bronzino et al., 1999). Voltammetry provides better temporal resolution of NE release. Electrical stimulation of the LC for 2 s at 50 Hz induces a 10 fold increase in oxidation signal in the DG (maximal extracellular signal
0.18 mM NE), appearing immediately after stimulation and returning to baseline within 10–15 s (Yavich et al., 2005). Repeated stimulation at 5 min intervals yields stable NE signals, while stimulation at 30 s intervals shows decreasing responses, suggesting a slow rate of vesicle refilling. Natural stimuli that activate LC (tail pinch or vibrissae stimulation) do not produce a visible NE signal without enhancement of NE release using the a2 receptor-antagonist, idazoxan. Glutamate activation of LC produces an average increase of
0.1 mM in hippocampal NE beginning 30 s after glutamate infusion, peaking at 1.5 min and returning to baseline by 7 min in urethane-anesthetized rats (Palamarchouk et al., 2000). NE patterns in awake rats with glutamate infusion were similar but peaked at 5 min and returned to baseline within 20 min (Palamarchouk et al., 2002).

NE release modulation by glutamate and vice-versa: local effects Both NMDA and AMPA receptors are present presynaptically on NE axon terminals in hippocampus, and when activated by glutamate or glutamate agonists, induce an increase in NE release (Pittaluga and Raiteri, 1992; Raiteri et al., 1992). NMDA-induced NE increases are greatest in DG microslices (Andres et al., 1993). Nictotine and somatostatin receptors are also present on NE terminals in hippocampus and induce NE release alone (nicotine) or in concert with NMDA receptor stimulation (nicotine and somtatostatin) in the

presence of normal magnesium levels and without membrane depolarization (Pittaluga et al., 2000, 2005; Risso et al., 2004). This provides a mechanism for local NE release in the absence of direct terminal activation. Finally, cognitive enhancers significantly facilitate NE release in response to glutamate, even in the presence of the endogenous NMDA receptor antagonist, kynurenate (Pittaluga et al., 1997). This facilitation is suggested to be an important mediator of cognitive enhancement (Desai et al., 1995; Pittaluga et al., 2001). Conversely, NE enhances glutamate release in DG, an effect which depends on presynaptic b adrenoceptors (Lynch and Bliss, 1986). NE can also enhance GABA release in hippocampus through a2 adrenoceptor activation (Pittaluga and Raiteri, 1987; Maura et al., 1988). Reversible EPSP slope potentiation of perforant path input to the DG has been related to b adrenoceptormediated phosphorylation of the presynaptic proteins synapsins I and II (Parfitt et al., 1991, 1992). The positive feedback between NE and glutamate release further suggests local increases in NE could occur independently of LC activation. The release of NE by glutamate and other neurotransmitters may explain extended limbic NE elevation when rats receive shock in a novel context (McIntyre et al., 2002). Suggestively, memory for the context correlates positively with NE levels at the time of acquisition. LTP of perforant path fibers also evokes longer lasting increases in NE release than LC activation alone (Bronzino et al., 1999, 2001).

LC-NE modulation of DG in relation to environmental events Novel objects investigated by a rat using a board with objects contained in holes (holeboard) elicit LC bursts (Vankov et al., 1995). Perforant pathevoked population spikes are potentiated for
1 min after a rat places its nose in a hole in response to a novel object (nose poke). Spike potentiation depends on b adrenoceptor activation (Kitchigina et al., 1997); as does the enhanced exploration of novelty (Sara et al., 1995). Prolonged spike enhancement occurs upon first exposure to

311

the novel holeboard environment, possibly because the novel environment elicits a larger LC response (Kitchigina et al., 1997). These effects appear to model the attentional (Berridge and Waterhouse, 2003), rather than the memory, effects of NE. Novel environment exposure can promote memory-like effects when a weak, high-frequency LTP stimulus is also presented (Straube et al., 2003). Placement in a novel environment converts earlyphase LTP to late-phase LTP in the DG. This conversion depends on b adrenoceptor activation, suggesting LC-NE activation by novelty induces an enhancement of the weak LTP stimulus. NE in the ventricles mediates a similar conversion of early to late-LTP in the DG (Almaguer-Melian et al., 2005). b adrenoceptor activation is critical for late DG LTP both in vitro (Stanton and Sarvey, 1985a) and in vivo (Straube and Frey, 2003). Only the use of strong LTP protocols (for example, 15 0.25 ms pulses at 200 Hz every 10 s for 200 s) with repeated 15 pulse protocols twice within 5 min can produce LTP in vivo that is independent of b adrenoceptor activation (Straube and Frey, 2003). In addition to facilitation by NE of enduring LTP in the entorhinal cortex-to-granule cell perforant pathway, there is LTP facilitation in the granule cell-to-CA3 pathway when subthreshold LTP stimuli are combined with b adrenoceptor activation (Hopkins and Johnston, 1984). Thus, weak patterns for LTP recruit enduring LTP in both the input and output components of DG circuitry when NE is elevated. Reward also promotes the conversion of early phase perforant path-LTP to late phase perforant path-LTP (Seidenbecher et al., 1997), and is known to activate LC neurons (Sara and Segal, 1991). Punishment has similar effects on activation of LC (Sara and Segal, 1991) and the conversion of early to late DG LTP. Conversions of early to late LTP by reward and punishment depend b adrenoceptor activation (Seidenbecher et al., 1997). These results suggest reward- or punishment-related LC activation ‘reinforces’ plasticity in the DG. LC-NE may also contribute to long-lasting circuit changes through the promotion of the survival of new neurons in DG (Rizk et al., 2006) and/or the promotion of neurogenesis (Kulkarni et al., 2002).

Restoration of recently-decayed LTP in the DG occurs following electrical stimulation of the LC (Ezrokhi et al., 1999). This phenomenon may relate to other evidence for a role of NE in memory retrieval (Sara and Devauges, 1989; Devauges and Sara, 1991; Murchison et al., 2004). NE’s promotion of theta EEG and the reported dependence of engram spread on b adrenoceptor activation could also contribute to retrieval processes (Flexner et al., 1985; Givens, 1996). Aston-Jones and Cohen’s proposal that phasic LC-NE optimizes decisiondriven behavior and decision-driven memory representations provides further theoretical support for an NE retrieval function (Aston-Jones and Cohen, 2005). Exploratory behavior is increased with DG NE infusions (Flicker and Geyer, 1982) and this effect is blocked by a b adrenoceptor antagonist. A tonic LC-NE driven modulation of diversive exploration (Usher et al., 1999; Mansour et al., 2003) is consistent with lesion evidence suggesting the DG has an important role in encoding spatial novelty (Lee et al., 2005; Jerman et al., 2006). Vasopressin in the DG facilitates memory for passive avoidance and increases NE turnover in DG (Kovacs et al., 1979b). Both consolidation and retrieval are enhanced, and these effects require intact LC-NE innervation (Kovacs et al., 1979a). Vasopressin in dorsal hippocampus also facilitates memory for spatial learning (Paban et al., 2003). NE dependence of this effect remains to be assessed. Corticotrophin releasing factor infused in the DG also enhances passive avoidance memory, while loss of local NE innervation, or blockade of local b adrenoceptors or NMDA receptors, prevents this memory enhancement (Lee et al., 1993). Lee et al. suggested that b adrenoceptors promote NMDA-mediated synaptic plasticity to support passive avoidance memory. A dependence of NEinduced plasticity on NMDA receptors was also described in some in vitro studies (Sarvey et al., 1989). As we come to better understand the behavioral role of DG, new probes of NE’s function in the DG should emerge. Aston-Jones and Cohen have proposed, based on data from primates, that phasic LC-NE activation facilitates decision-driven

312

responses or memories, while tonic LC-NE activity facilitates exploration (Aston-Jones and Cohen, 2005). The latter hypothesis is supported by increased exploration following NE infusion in DG (Flicker and Geyer, 1982), but the relation of phasic activation to decision-driven behavior and/or memory is unclear with respect to DG activity. Bouret and Sara’s more general description (Bouret and Sara, 2005) of the function of the LC-NE system as ‘network resetting’ is broadly consistent with NE-associated changes in EEG and with the promotion of long-term synaptic plasticity in DG input and output. The promotion of frequency-induced long-term potentiation (Dragoi et al., 2003) by NE reviewed above and the selective enhancement of the spatial map matrix input from the medial entorhinal cortex (Hafting et al., 2005) described earlier (see Pathway selectivity) lead to the prediction that, for the DG, strong activation of NE input would promote global remapping of spatial context. Weaker activation of NE input and an associated transient reduction in network inhibition (see Extracellular unit recording, above) might enhance input generalization, or pattern completion to assist retrieval, but strong activation would alter the DG map. Global remapping provides a new framework for the encoding of associative learning, reducing interference among associations and increasing memory capacity. Global remapping is assumed to underlie episodic memory. The prediction that strong LC-NE activation recruits global remapping in DG is consistent with the hypothesis that activation of the LC and the related sympathetic system is part of a response to strongly significant events that promotes long-term memories for significant events (Cahill et al., 1994). Global remapping with novelty exposure has been shown for CA3 cells (Leutgeb et al., 2006), which like DG (Chawla et al., 2005; Rolls and Kesner, 2006), are implicated in sparse encoding and orthogonal environmental representations. The occurrence of global remapping in DG concomitant with strong LC activation remains to be demonstrated, but is suggested by the data reviewed here. In summary, evoked potential and cellular data support various effects of NE on information processing in the DG. Transient effects of phasic

NE release include reduction of feedforward inhibition and increases in granule cell excitability, such that granule cell responses to entorhinal input increase. NE’s modulation of feedback interneurons enhances theta to facilitate throughput and associative change in synaptic strength depending on theta phase (Pavlides et al., 1988; Orr et al., 2001). NE enhances and depresses inputs in a pathway-dependent manner and recruits glial metabolic support for neuronal activation. Strong NE input engages processes that recruit long-term modifications of excitability and/or delayed increases in synaptic strength in the DG pathway. Together these effects support both attentional and memory roles for DG NE.

Acknowledgment The author wishes to express appreciation for the support of Canada’s NSERC Granting Council via Grant A9791.

References Almaguer-Melian, W., Rojas-Reyes, Y., Alvare, A., Rosillo, J.C., Frey, J.U. and Bergado, J.A. (2005) Long-term potentiation in the dentate gyrus in freely moving rats is reinforced by intraventricular application of norepinephrine, but not oxotremorine. Neurobiol. Learn. Mem., 83: 72–78. Andres, M.E., Bustos, G. and Gysling, K. (1993) Regulation of [3H]norepinephrine release by N-methyl-D-aspartate receptors in minislices from the dentate gyrus and the CA1-CA3 area of the rat hippocampus. Biochem. Pharmacol., 46: 1983–1987. Arango, V., Ernsberger, P., Reis, D.J. and Mann, J.J. (1990) Demonstration of high- and low-affinity beta-adrenergic receptors in slide-mounted sections of rat and human brain. Brain Res., 516: 113–121. Aston-Jones, G. and Bloom, F.E. (1981a) Norepinephrine-containing locus coeruleus neurons in behaving rats exhibit pronounced responses to non-noxious environmental stimuli. J. Neurosci., 1: 887–900. Aston-Jones, G. and Bloom, F.E. (1981b) Activity of norepinephrine-containing locus coeruleus neurons in behaving rats anticipates fluctuations in the sleep- waking cycle. J. Neurosci., 1: 876–886. Aston-Jones, G. and Cohen, J.D. (2005) An integrative theory of locus coeruleus-norepinephrine function: adaptive gain and optimal performance. Annu. Rev. Neurosci., 28: 403–450.

313 Aston-Jones, G., Foote, S.L. and Segal, M. (1985) Impulse conduction properties of noradrenergic locus coeruleus axons projecting to monkey cerebrocortex. Neuroscience, 15: 765–777. Babstock, D.M. and Harley, C.W. (1992) Paragigantocellularis stimulation induces beta-adrenergic hippocampal potentiation. Brain Res. Bull., 28: 709–714. Babstock, D.M. and Harley, C.W. (1993) Lateral olfactory tract input to dentate gyrus is depressed by prior noradrenergic activation using nucleus paragigantocellularis stimulation. Brain Res., 629: 149–154. Berridge, C.W., Bolen, S.J., Manley, M.S. and Foote, S.L. (1996) Modulation of forebrain electroencephalographic activity in halothane-anesthetized rat via actions of noradrenergic beta-receptors within the medial septal region. J. Neurosci., 16: 7010–7020. Berridge, C.W. and Foote, S.L. (1991) Effects of locus coeruleus activation on electroencephalographic activity in neocortex and hippocampus. J. Neurosci., 11: 3135–3145. Berridge, C.W. and Waterhouse, B.D. (2003) The locus coeruleus-noradrenergic system: modulation of behavioral state and state-dependent cognitive processes. Brain Res. Brain Res. Rev., 42: 33–84. Bijak, M. and Misgeld, U. (1995) Adrenergic modulation of hilar neuron activity and granule cell inhibition in the guinea-pig hippocampal slice. Neuroscience, 67: 541–550. Bjorklund, A., Segal, M. and Stenevi, U. (1979) Functional reinnervation of rat hippocampus by locus coeruleus implants. Brain Res., 170: 409–426. Blackstad, T.W., Fuxe, K. and Hokfelt, T. (1967) Noradrenaline nerve terminals in the hippocampal region of the rat and the guinea pig. Z. Zellforsch. Mikrosk. Anat., 78: 463–473. Booze, R.M., Crisostomo, E.A. and Davis, J.N. (1993) Betaadrenergic receptors in the hippocampal and retrohippocampal regions of rats and guinea pigs: autoradiographic and immunohistochemical studies. Synapse, 13: 206–214. Bouret, S. and Sara, S.J. (2005) Network reset: a simplified overarching theory of locus coeruleus noradrenaline function. Trends Neurosci., 28: 574–582. Bramham, C.R., Bacher-Svendsen, K. and Sarvey, J.M. (1997) LTP in the lateral perforant path is beta-adrenergic receptordependent. Neuroreport, 8: 719–724. Bronzino, J.D., Kehoe, P., Hendriks, R., Vita, L., Golas, B., Vivona, C. and Morgane, P.J. (1999) Hippocampal neurochemical and electrophysiological measures from freely moving rats. Exp. Neurol., 155: 150–155. Bronzino, J.D., Kehoe, P., Mallinson, K. and Fortin, D.A. (2001) Increased extracellular release of hippocampal NE is associated with tetanization of the medial perforant pathway in the freely moving adult male rat. Hippocampus, 11: 423–429. Brown, R.A.M., Walling, S.G., Milway, J.S. and Harley, C.W. (2005) Locus ceruleus activation suppresses feedforward interneurons and reduces beta-gamma electroencephalogram frequencies while it enhances theta frequencies in rat dentate gyrus. J. Neurosci., 25: 1985–1991.

Brunelli, M., Castellucci, V. and Kandel, E.R. (1976) Synaptic facilitation and behavioral sensitization in Aplysia: possible role of serotonin and cyclic AMP. Science, 194: 1178–1181. Burgard, E.C., Decker, G. and Sarvey, J.M. (1989) NMDA receptor antagonists block norepinephrine-induced long-lasting potentiation and long-term potentiation in rat dentate gyrus. Brain Res., 482: 351–355. Cahill, L., Prins, B., Weber, M. and McGaugh, J.L. (1994) Beta-adrenergic activation and memory for emotional events. Nature, 371: 702–704. Chaulk, P.C. and Harley, C.W. (1998) Intracerebroventricular norepinephrine potentiation of the perforant path-evoked potential in dentate gyrus of anesthetized and awake rats: a role for both alpha- and beta-adrenoceptor activation. Brain Res., 787: 59–70. Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P., Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton, B.L. and Barnes, C.A. (2005) Sparse, environmentally selective expression of Arc RNA in the upper blade of the rodent fascia dentata by brief spatial experience. Hippocampus, 15: 579–586. Crutcher, K.A. and Davis, J.N. (1980) Hippocampal alpha- and beta-adrenergic receptors: comparison of [3H]dihydroalprenolol and [3H]WB 4101 binding with noradrenergic innervation in the rat. Brain Res., 182: 107–117. Dahl, D. and Li, J. (1994) Induction of long-lasting potentiation by sequenced applications of isoproterenol. Neuroreport, 5: 657–660. Dahl, D. and Sarvey, J.M. (1989) Norepinephrine induces pathway-specific long-lasting potentiation and depression in the hippocampal dentate gyrus. Proc. Natl. Acad. Sci. U.S.A., 86: 4776–4780. Dahl, D. and Sarvey, J.M. (1990) Beta-adrenergic agonist-induced long-lasting synaptic modifications in hippocampal dentate gyrus require activation of NMDA receptors, but not electrical activation of afferents. Brain Res., 526: 347–350. Dahl, D. and Winson, J. (1985) Action of norepinephrine in the dentate gyrus. I. Stimulation of locus coeruleus. Exp. Brain Res., 59: 491–496. Dan, Y. and Poo, M.M. (2006) Spike timing-dependent plasticity: from synapse to perception. Physiol. Rev., 86: 1033–1048. Daoudal, G. and Debanne, D. (2003) Long-term plasticity of intrinsic excitability: learning rules and mechanisms. Learn. Mem., 10: 456–465. Day, H.E., Campeau, S., Watson Jr., S.J. and Akil, H. (1997) Distribution of alpha 1a-, alpha 1b- and alpha 1d-adrenergic receptor mRNA in the rat brain and spinal cord. J. Chem. Neuroanat., 13: 115–139. Desai, M.A., Valli, M.J., Monn, J.A. and Schoepp, D.D. (1995) 1-BCP, a memory-enhancing agent, selectively potentiates AMPA-induced [3H]norepinephrine release in rat hippocampal slices. Neuropharmacology, 34: 141–147. Devauges, V. and Sara, S.J. (1991) Memory retrieval enhancement by locus coeruleus stimulation: evidence for mediation by beta-receptors. Behav. Brain Res., 43: 93–97.

314 Dragoi, G., Harris, K.D. and Buzsaki, G. (2003) Place representation within hippocampal networks is modified by longterm potentiation. Neuron, 39: 843–853. Duncan, G.E., Little, K.Y., Koplas, P.A., Kirkman, J.A., Breese, G.R. and Stumpf, W.E. (1991) Beta-adrenergic receptor distribution in human and rat hippocampal formation: marked species differences. Brain Res., 561: 84–92. Edwards, C., Nahorski, S.R. and Rogers, K.J. (1974) In vivo changes of cerebral cyclic adenosine 30 ,50 -monophosphate induced by biogenic amines: association with phosphorylase activation. J. Neurochem., 22: 565–572. Ezrokhi, V.L., Zosimovskii, V.A., Korshunov, V.A. and Markevich, V.A. (1999) Restoration of decaying long-term potentiation in the hippocampal formation by stimulation of neuromodulatory nuclei in freely moving rats. Neuroscience, 88: 741–753. Fara-On, M., Evans, J.H. and Harley, C.W. (2005) Idazoxan activates rat forebrain glycogen phosphorylase in vivo: a histochemical study. Brain Res., 1059: 83–92. Flexner, J.B., Flexner, L.B., Church, A.C., Rainbow, T.C. and Brunswick, D.J. (1985) Blockade of beta 1- but not of beta 2adrenergic receptors replicates propranolol’s suppression of the cerebral spread of an engram in mice. Proc. Natl. Acad. Sci. U.S.A., 82: 7458–7461. Flicker, C. and Geyer, M.A. (1982) Behavior during hippocampal microinfusions. I. Norepinephrine and diversive exploration. Brain Res., 257: 79–103. Florin-Lechner, S.M., Druhan, J.P., Aston-Jones, G. and Valentino, R.J. (1996) Enhanced norepinephrine release in prefrontal cortex with burst stimulation of the locus coeruleus. Brain Res., 742: 89–97. Frick, A., Magee, J. and Johnston, D. (2004) LTP is accompanied by an enhanced local excitability of pyramidal neuron dendrites. Nat. Neurosci., 7: 126–135. Frizzell, L.M. and Harley, C.W. (1994) The N-methyl-D-aspartate channel blocker ketamine does not attenuate, but enhances, locus coeruleus-induced potentiation in rat dentate gyrus. Brain Res., 663: 173–178. Givens, B. (1996) Stimulus-evoked resetting of the dentate theta rhythm: relation to working memory. Neuroreport, 8: 159–163. Gray, J.A. and Ball, G.G. (1970) Frequency-specific relation between hippocampal theta rhythm, behavior, and amobarbital action. Science, 168: 1246–1248. Gray, J.A., McNaughton, N., James, D.T. and Kelly, P.H. (1975) Effect of minor tranquillisers on hippocampal theta rhythm mimicked by depletion of forebrain noradrenaline. Nature, 258: 424–425. Gray, R. and Johnston, D. (1987) Noradrenaline and beta-adrenoceptor agonists increase activity of voltage-dependent calcium channels in hippocampal neurons. Nature, 327: 620–622. Haas, H.L. and Rose, G.M. (1987) Noradrenaline blocks potassium conductance in rat dentate granule cells in vitro. Neurosci. Lett., 78: 171–174.

Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser, E.I. (2005) Microstructure of a spatial map in the entorhinal cortex. Nature, 436: 801–806. Haring, J.H. and Davis, J.N. (1985) Differential distribution of locus coeruleus projections to the hippocampal formation: anatomical and biochemical evidence. Brain Res., 325: 366–369. Harley, C., Milway, J.S. and Lacaille, J.C. (1989) Locus coeruleus potentiation of dentate gyrus responses: evidence for two systems. Brain Res. Bull., 22: 643–650. Harley, C. and Rusak, B. (1993) Daily variation in active glycogen phosphorylase patches in the molecular layer of rat dentate gyrus. Brain Res., 626: 310–317. Harley, C.W. (1987) A role for norepinephrine in arousal, emotion and learning: limbic modulation by norepinephrine and the Kety hypothesis. Prog. Neuropsychopharmacol. Biol. Psychiatry, 11: 419–458. Harley, C.W. and Evans, S. (1988) Locus coeruleus-induced enhancement of the perforant path-evoked potential: effects of intradentate beta blockers. In: Woody, C.D., Alkon, D.L. and McGaugh, J.L. (Eds.), Cellular Mechanisms of Conditioning and Behavioral Plasticity. Plenum Publishing Corporation, New York, New York, pp. 415–423. Harley, C.W., Lalies, M.D. and Nutt, D.J. (1996) Estimating the synaptic concentration of norepinephrine in dentate gyrus which produces beta-receptor mediated long-lasting potentiation in vivo using microdialysis and intracerebroventricular norepinephrine. Brain Res., 710: 293–298. Harley, C.W. and Milway, J.S. (1986) Glutamate ejection in the locus coeruleus enhances the perforant path-evoked population spike in the dentate gyrus. Exp. Brain Res., 63: 143–150. Harley, C.W. and Sara, S.J. (1992) Locus coeruleus bursts induced by glutamate trigger delayed perforant path spike amplitude potentiation in the dentate gyrus. Exp. Brain Res., 89: 581–587. Hopkins, W.F. and Johnston, D. (1984) Frequency-dependent noradrenergic modulation of long-term potentiation in the hippocampus. Science, 226: 350–352. Hortnagl, H., Berger, M.L., Sperk, G. and Pifl, C. (1991) Regional heterogeneity in the distribution of neurotransmitter markers in the rat hippocampus. Neuroscience, 45: 261–272. Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Disconnection analysis of CA3 and DG in mediating encoding but not retrieval in a spatial maze learning task. Learn. Mem., 13: 458–464. Jones, L.S., Gauger, L.L. and Davis, J.N. (1985) Anatomy of brain alpha 1-adrenergic receptors: in vitro autoradiography with [125I]-heat. J. Comp. Neurol., 231: 190–208. Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of unit activity in the hippocampal granular layer. Hippocampus, 3: 165–182. Kitchigina, V., Vankov, A., Harley, C. and Sara, S.J. (1997) Novelty-elicited, noradrenaline-dependent enhancement of excitability in the dentate gyrus. Eur. J. Neurosci., 9: 41–47.

315 Klukowski, G. and Harley, C.W. (1994) Locus coeruleus activation induces perforant path-evoked population spike potentiation in the dentate gyrus of awake rat. Exp. Brain Res., 102: 165–170. Koda, L.Y. and Bloom, F.E. (1977) A light and electron microscopic study of noradrenergic terminals in the rat dentate gyrus. Brain Res., 120: 327–335. Kopell, N., Ermentrout, G.B., Whittington, M.A. and Traub, R.D. (2000) Gamma rhythms and beta rhythms have different synchronization properties. Proc. Natl. Acad. Sci. U.S.A., 97: 1867–1872. Kosaka, T., Katsumaru, H., Hama, K., Wu, J.Y. and Heizmann, C.W. (1987) GABAergic neurons containing the Ca2+-binding protein parvalbumin in the rat hippocampus and dentate gyrus. Brain Res., 419: 119–130. Kovacs, G.L., Bohus, B. and Versteeg, D.H. (1979a) The effects of vasopressin on memory processes: the role of noradrenergic neurotransmission. Neuroscience, 4: 1529–1537. Kovacs, G.L., Bohus, B., Versteeg, D.H., de Kloet, E.R. and de, W.D. (1979b) Effect of oxytocin and vasopressin on memory consolidation: sites of action and catecholaminergic correlates after local microinjection into limbic-midbrain structures. Brain Res., 175: 303–314. Kulkarni, V.A., Jha, S. and Vaidya, V.A. (2002) Depletion of norepinephrine decreases the proliferation, but does not influence the survival and differentiation, of granule cell progenitors in the adult rat hippocampus. Eur. J. Neurosci., 16: 2008–2012. Kusaka, K., Morinobu, S., Kawano, K. and Yamawaki, S. (2004) Effect of neonatal isolation on the noradrenergic transduction system in the rat hippocampal slice. Synapse, 54: 223–232. Lacaille, J.C. and Harley, C.W. (1985) The action of norepinephrine in the dentate gyrus: beta-mediated facilitation of evoked potentials in vitro. Brain Res., 358: 210–220. Lacaille, J.C. and Schwartzkroin, P.A. (1988) Intracellular responses of rat hippocampal granule cells in vitro to discrete applications of norepinephrine. Neurosci. Lett., 89: 176–181. Lancaster, B., Hu, H., Ramakers, G.M. and Storm, J.F. (2001) Interaction between synaptic excitation and slow afterhyperpolarization current in rat hippocampal pyramidal cells. J. Physiol., 536: 809–823. Lapierre, Y., Beaudet, A., Demianczuk, N. and Descarries, L. (1973) Noradrenergic axon terminals in the cerebral cortex of rat. II. Quantitative data revealed by light and electron microscope radioautography of the frontal cortex. Brain Res., 63: 175–182. Lee, E.H., Lee, C.P., Wang, H.I. and Lin, W.R. (1993) Hippocampal CRF, NE, and NMDA system interactions in memory processing in the rat. Synapse, 14: 144–153. Lee, I., Hunsaker, M.R. and Kesner, R.P. (2005) The role of hippocampal subregions in detecting spatial novelty. Behav. Neurosci., 119: 145–153.

Leutgeb, S., Leutgeb, J.K., Moser, E.I. and Moser, M.B. (2006) Fast rate coding in hippocampal CA3 cell ensembles. Hippocampus, 16: 765–774. Loy, R., Koziell, D.A., Lindsey, J.D. and Moore, R.Y. (1980) Noradrenergic innervation of the adult rat hippocampal formation. J. Comp. Neurol., 189: 699–710. Lynch, M.A. and Bliss, T.V. (1986) Noradrenaline modulates the release of [14C]glutamate from dentate but not from CA1/CA3 slices of rat hippocampus. Neuropharmacology, 25: 493–498. Mansour, A.A., Babstock, D.M., Penney, J.H., Martin, G.M., McLean, J.H. and Harley, C.W. (2003) Novel objects in a holeboard probe the role of the locus coeruleus in curiosity: support for two modes of attention in the rat. Behav. Neurosci., 117: 621–631. Maura, G., Pittaluga, A., Ulivi, M. and Raiteri, M. (1988) Enhancement of endogenous GABA release from rat synaptosomal preparations is mediated by alpha 2-adrenoceptors pharmacologically different from alpha 2-autoreceptors. Eur. J. Pharmacol., 157: 23–29. McIntyre, C.K., Hatfield, T. and McGaugh, J.L. (2002) Amygdala norepinephrine levels after training predict inhibitory avoidance retention performance in rats. Eur. J. Neurosci., 16: 1223–1226. McNaughton, B.L., Battaglia, F.P., Jensen, O., Moser, E.I. and Moser, M.B. (2006) Path integration and the neural basis of the ‘cognitive map’. Nat. Rev. Neurosci., 7: 663–678. Milner, T.A. and Bacon, C.E. (1989a) GABAergic neurons in the rat hippocampal formation: ultrastructure and synaptic relationships with catecholaminergic terminals. J. Neurosci., 9: 3410–3427. Milner, T.A. and Bacon, C.E. (1989b) Ultrastructural localization of tyrosine hydroxylase-like immunoreactivity in the rat hippocampal formation. J. Comp. Neurol., 281: 479–495. Milner, T.A., Lee, A., Aicher, S.A. and Rosin, D.L. (1998) Hippocampal alpha2a-adrenergic receptors are located predominantly presynaptically but are also found postsynaptically and in selective astrocytes. J. Comp. Neurol., 395: 310–327. Milner, T.A., Shah, P. and Pierce, J.P. (2000) Beta-adrenergic receptors primarily are located on the dendrites of granule cells and interneurons but also are found on astrocytes and a few presynaptic profiles in the rat dentate gyrus. Synapse, 36: 178–193. Minneman, K.P., Hegstrand, L.R. and Molinoff, P.B. (1979) Simultaneous determination of beta-1 and beta-2-adrenergic receptors in tissues containing both receptor subtypes. Mol. Pharmacol., 16: 34–46. Moudy, A.M., Kunkel, D.D. and Schwartzkroin, P.A. (1993) Development of dopamine-beta-hydroxylase-positive fiber innervation of the rat hippocampus. Synapse, 15: 307–318. Murchison, C.F., Zhang, X.Y., Zhang, W.P., Ouyang, M., Lee, A. and Thomas, S.A. (2004) A distinct role for norepinephrine in memory retrieval. Cell, 117: 131–143.

316 Nahorski, S.R., Rogers, K.J. and Edwards, C. (1975) Cerebral glycogenolysis and stimulation of beta-adrenoreceptors and histamine H2 receptors. Brain Res., 92: 529–533. Neuman, R.S. and Harley, C.W. (1983) Long-lasting potentiation of the dentate gyrus population spike by norepinephrine. Brain Res., 273: 162–165. Nicholas, A.P., Pieribone, V. and Hokfelt, T. (1993a) Distributions of mRNAs for alpha-2 adrenergic receptor subtypes in rat brain: an in situ hybridization study. J. Comp. Neurol., 328: 575–594. Nicholas, A.P., Pieribone, V.A. and Hokfelt, T. (1993b) Cellular localization of messenger RNA for beta-1 and beta-2 adrenergic receptors in rat brain: an in situ hybridization study. Neuroscience, 56: 1023–1039. Nitz, D.A. and McNaughton, B.L. (1999) Hippocampal EEG and unit activity responses to modulation of serotonergic median raphe neurons in the freely behaving rat. Learn. Mem., 6: 153–167. Oleskevich, S., Descarries, L. and Lacaille, J.C. (1989) Quantified distribution of the noradrenaline innervation in the hippocampus of adult rat. J. Neurosci., 9: 3803–3815. Orr, G., Rao, G., Houston, F.P., McNaughton, B.L. and Barnes, C.A. (2001) Hippocampal synaptic plasticity is modulated by theta rhythm in the fascia dentata of adult and aged freely behaving rats. Hippocampus, 11: 647–654. Paban, V., Soumireu-Mourat, B. and escio-Lautier, B. (2003) Behavioral effects of arginine8-vasopressin in the Hebb-Williams maze. Behav. Brain Res., 141: 1–9. Palamarchouk, V.S., Swiergiel, A.H. and Dunn, A.J. (2002) Hippocampal noradrenergic responses to CRF injected into the locus coeruleus of unanesthetized rats. Brain Res., 950: 31–38. Palamarchouk, V.S., Zhang, J., Zhou, G., Swiergiel, A.H. and Dunn, A.J. (2000) Hippocampal norepinephrine-like voltammetric responses following infusion of corticotropin-releasing factor into the locus coeruleus. Brain Res. Bull., 51: 319–326. Pang, K. and Rose, G.M. (1987) Differential effects of norepinephrine on hippocampal complex-spike and theta-neurons. Brain Res., 425: 146–158. Parfitt, K.D., Doze, V.A., Madison, D.V. and Browning, M.D. (1992) Isoproterenol increases the phosphorylation of the synapsins and increases synaptic transmission in dentate gyrus, but not in area CA1, of the hippocampus. Hippocampus, 2: 59–64. Parfitt, K.D., Hoffer, B.J. and Browning, M.D. (1991) Norepinephrine and isoproterenol increase the phosphorylation of synapsin I and synapsin II in dentate slices of young but not aged Fisher 344 rats. Proc. Natl. Acad. Sci. U.S.A., 88: 2361–2365. Pavlides, C., Greenstein, Y.J., Grudman, M. and Winson, J. (1988) Long-term potentiation in the dentate gyrus is induced preferentially on the positive phase of theta-rhythm. Brain Res., 439: 383–387. Phelps, C.H. (1972) Barbiturate-induced glycogen accumulation in brain. An electron microscopic study. Brain Res., 39: 225–234.

Pieribone, V.A., Nicholas, A.P., Dagerlind, A. and Hokfelt, T. (1994) Distribution of alpha 1 adrenoceptors in rat brain revealed by in situ hybridization experiments utilizing subtype-specific probes. J. Neurosci., 14: 4252–4268. Pittaluga, A., Bonfanti, A. and Raiteri, M. (2000) Somatostatin potentiates NMDA receptor function via activation of InsP(3) receptors and PKC leading to removal of the Mg2+ block without depolarization. Br. J. Pharmacol., 130: 557–566. Pittaluga, A., Feligioni, M., Ghersi, C., Gemignani, A. and Raiteri, M. (2001) Potentiation of NMDA receptor function through somatostatin release: a possible mechanism for the cognition-enhancing activity of GABA(B) receptor antagonists. Neuropharmacology, 41: 301–310. Pittaluga, A., Feligioni, M., Longordo, F., Arvigo, M. and Raiteri, M. (2005) Somatostatin-induced activation and upregulation of N-methyl-D-aspartate receptor function: mediation through calmodulin-dependent protein kinase II, phospholipase C, protein kinase C, and tyrosine kinase in hippocampal noradrenergic nerve endings. J. Pharmacol. Exp. Ther., 313: 242–249. Pittaluga, A. and Raiteri, M. (1987) GABAergic nerve terminals in rat hippocampus possess alpha 2-adrenoceptors regulating GABA release. Neurosci. Lett., 76: 363–367. Pittaluga, A. and Raiteri, M. (1992) N-methyl-D-aspartic acid (NMDA) and non-NMDA receptors regulating hippocampal norepinephrine release. I. Location on axon terminals and pharmacological characterization. J. Pharmacol. Exp. Ther., 260: 232–237. Pittaluga, A., Vaccari, D. and Raiteri, M. (1997) The ‘‘kynurenate test,’’ a biochemical assay for putative cognition enhancers. J. Pharmacol. Exp. Ther., 283: 82–90. Quach, T.T., Rose, C. and Schwartz, J.C. (1978) [3H]Glycogen hydrolysis in brain slices: responses to neurotransmitters and modulation of noradrenaline receptors. J. Neurochem., 30: 1335–1341. Rainbow, T.C., Parsons, B. and Wolfe, B.B. (1984) Quantitative autoradiography of beta 1- and beta 2-adrenergic receptors in rat brain. Proc. Natl. Acad. Sci. U.S.A., 81: 1585–1589. Raiteri, M., Garrone, B. and Pittaluga, A. (1992) N-methyl-Daspartic acid (NMDA) and non-NMDA receptors regulating hippocampal norepinephrine release. II. Evidence for functional cooperation and for coexistence on the same axon terminal. J. Pharmacol. Exp. Ther., 260: 238–242. Reid, A.T. and Harley, C.W. (2005) Unpaired perforant path and locus ceruleus activations only transiently potentiate the evoked dentate gyrus population spike. Online Abstracts Viewer/Itinerary Planner and CD-ROM. Ref Type: Abstract. Ribak, C.E. (1992) Local circuitry of GABAergic basket cells in the dentate gyrus. Epilepsy Res. Suppl., 7: 29–47. Ribak, C.E., Nitsch, R. and Seress, L. (1990) Proportion of parvalbumin-positive basket cells in the GABAergic innervation of pyramidal and granule cells of the rat hippocampal formation. J. Comp. Neurol., 300: 449–461.

317 Risso, F., Grilli, M., Parodi, M., Bado, M., Raiteri, M. and Marchi, M. (2004) Nicotine exerts a permissive role on NMDA receptor function in hippocampal noradrenergic terminals. Neuropharmacology, 47: 65–71. Rizk, P., Salazar, J., Raisman-Vozari, R., Marien, M., Ruberg, M., Colpaert, F. and Debeir, T. (2006) The alpha2-adrenoceptor antagonist dexefaroxan enhances hippocampal neurogenesis by increasing the survival and differentiation of new granule cells. Neuropsychopharmacology, 31: 1146–1157. Rolls, E.T. and Kesner, R.P. (2006) A computational theory of hippocampal function, and empirical tests of the theory. Prog. Neurobiol., 79: 1–48. Rose, G.M. and Pang, K.C. (1989) Differential effect of norepinephrine upon granule cells and interneurons in the dentate gyrus. Brain Res., 488: 353–356. Sara, S.J. and Devauges, V. (1989) Idazoxan, an alpha-2 antagonist, facilitates memory retrieval in the rat. Behav. Neural Biol., 51: 401–411. Sara, S.J., Dyon-Laurent, C. and Herve, A. (1995) Novelty seeking behavior in the rat is dependent upon the integrity of the noradrenergic system. Brain Res. Cogn. Brain Res., 2: 181–187. Sara, S.J. and Segal, M. (1991) Plasticity of sensory responses of locus coeruleus neurons in the behaving rat: implications for cognition. Prog. Brain Res., 88: 571–585. Sarvey, J.M., Burgard, E.C. and Decker, G. (1989) Long-term potentiation: studies in the hippocampal slice. J. Neurosci. Methods, 28: 109–124. Schacher, S., Castellucci, V.F. and Kandel, E.R. (1988) cAMP evokes long-term facilitation in Aplysia sensory neurons that requires new protein synthesis. Science, 240: 1667–1669. Scheinin, M., Lomasney, J.W., Hayden-Hixson, D.M., Schambra, U.B., Caron, M.G., Lefkowitz, R.J. and Fremeau Jr., R.T. (1994) Distribution of alpha 2-adrenergic receptor subtype gene expression in rat brain. Brain Res. Mol. Brain Res., 21: 133–149. Segal, M. (1976) Brain stem afferents to the rat medial septum. J. Physiol., 261: 617–631. Segal, M. and Bloom, F.E. (1976a) The action of norepinephrine in the rat hippocampus. III. Hippocampal cellular responses to locus coeruleus stimulation in the awake rat. Brain Res., 107: 499–511. Segal, M. and Bloom, F.E. (1976b) The action of norepinephrine in the rat hippocampus. IV. The effects of locus coeruleus stimulation on evoked hippocampal unit activity. Brain Res., 107: 513–525. Seidenbecher, T., Reymann, K.G. and Balschun, D. (1997) A post-tetanic time window for the reinforcement of long-term potentiation by appetitive and aversive stimuli. Proc. Natl. Acad. Sci. U.S.A., 94: 1494–1499. Siggins, G.R., Battenberg, E.F., Hoffer, B.J., Bloom, F.E. and Steiner, A.L. (1973) Noradrenergic stimulation of cyclic adenosine monophosphate in rat Purkinje neurons: an immunocytochemical study. Science, 179: 585–588. Stanton, P.K. and Sarvey, J.M. (1985a) Depletion of norepinephrine, but not serotonin, reduces long-term

potentiation in the dentate gyrus of rat hippocampal slices. J. Neurosci., 5: 2169–2176. Stanton, P.K. and Sarvey, J.M. (1985b) The effect of highfrequency electrical stimulation and norepinephrine on cyclic AMP levels in normal versus norepinephrine-depleted rat hippocampal slices. Brain Res., 358: 343–348. Stanton, P.K. and Sarvey, J.M. (1985c) Blockade of norepinephrine-induced long-lasting potentiation in the hippocampal dentate gyrus by an inhibitor of protein synthesis. Brain Res., 361: 276–283. Stanton, P.K. and Sarvey, J.M. (1987) Norepinephrine regulates long-term potentiation of both the population spike and dendritic EPSP in hippocampal dentate gyrus. Brain Res. Bull., 18: 115–119. Straube, T. and Frey, J.U. (2003) Involvement of beta-adrenergic receptors in protein synthesis-dependent late long-term potentiation (LTP) in the dentate gyrus of freely moving rats: The critical role of the LTP induction strength. Neuroscience, 119: 473–479. Straube, T., Korz, V., Balschun, D. and Frey, J.U. (2003) Requirement of beta-adrenergic receptor activation and protein synthesis for LTP-reinforcement by novelty in rat dentate gyrus. J. Physiol., 552: 953–960. Subbarao, K.V. and Hertz, L. (1991) Stimulation of energy metabolism by alpha-adrenergic agonists in primary cultures of astrocytes. J. Neurosci. Res., 28: 399–405. Uecker, A., Barnes, C.A., McNaughton, B.L. and Reiman, E.M. (1997) Hippocampal glycogen metabolism, EEG, and behavior. Behav. Neurosci., 111: 283–291. U’Prichard, D.C., Reisine, T.D., Mason, S.T., Fibiger, H.C. and Yamamura, H.I. (1980) Modulation of rat brain alpha- and beta-adrenergic receptor populations by lesion of the dorsal noradrenergic bundle. Brain Res., 187: 143–154. Usher, M., Cohen, J.D., Servan-Schreiber, D., Rajkowski, J. and Aston-Jones, G. (1999) The role of locus coeruleus in the regulation of cognitive performance. Science, 283: 549–554. Vankov, A., Herve-Minvielle, A. and Sara, S.J. (1995) Response to novelty and its rapid habituation in locus coeruleus neurons of the freely exploring rat. Eur. J. Neurosci., 7: 1180–1187. Walling, S.G. and Harley, C.W. (2004) Locus ceruleus activation initiates delayed synaptic potentiation of perforant path input to the dentate gyrus in awake rats: a novel beta-adrenergic- and protein synthesis-dependent mammalian plasticity mechanism. J. Neurosci., 24: 598–604. Walling, S.G., Nutt, D.J., Lalies, M.D. and Harley, C.W. (2004) Orexin-A infusion in the locus ceruleus triggers norepinephrine (NE) release and NE-induced long-term potentiation in the dentate gyrus. J. Neurosci., 24: 7421–7426. Wang, G.S., Chang, N.C., Wu, S.C. and Chang, A.C. (2002) Regulated expression of alpha2B adrenoceptor during development. Dev. Dyn., 225: 142–152. Washburn, M. and Moises, H.C. (1989) Electrophysiological correlates of presynaptic alpha 2-receptor-mediated

318 inhibition of norepinephrine release at locus coeruleus synapses in dentate gyrus. J. Neurosci., 9: 2131–2140. Winson, J. and Dahl, D. (1985) Action of norepinephrine in the dentate gyrus. II. Iontophoretic studies. Exp. Brain Res., 59: 497–506. Winzer-Serhan, U.H., Raymon, H.K., Broide, R.S., Chen, Y. and Leslie, F.M. (1997a) Expression of alpha 2 adrenoceptors during rat brain development. I. Alpha 2A messenger RNA expression. Neuroscience, 76: 241–260. Winzer-Serhan, U.H., Raymon, H.K., Broide, R.S., Chen, Y. and Leslie, F.M. (1997b) Expression of alpha 2 adrenoceptors during

rat brain development. II. Alpha 2C messenger RNA expression and [3H]rauwolscine binding. Neuroscience, 76: 261–272. Yavich, L., Jakala, P. and Tanila, H. (2005) Noradrenaline overflow in mouse dentate gyrus following locus coeruleus and natural stimulation: real-time monitoring by in vivo voltammetry. J. Neurochem., 95: 641–650. Zilles, K., Gross, G., Schleicher, A., Schildgen, S., Bauer, A., Bahro, M., Schwendemann, G., Zech, K. and Kolassa, N. (1991) Regional and laminar distributions of alpha 1-adrenoceptors and their subtypes in human and rat hippocampus. Neuroscience, 40: 307–320.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 19

Endocannabinoids in the dentate gyrus Charles J. Frazier1,2, 1

Department of Pharmacodynamics, University of Florida, College of Pharmacy, Gainesville, FL 32610, USA 2 Department of Neuroscience, University of Florida, College of Medicine, Gainesville, FL 32610, USA

Abstract: Recent years have produced rapid and enormous growth in our understanding of endocannabinoid-mediated signaling in the CNS. While much of the recent progress has focused on other areas of the brain, a significant body of evidence has developed that indicates the presence of a robust system for endocannabinoid-mediated signaling in the dentate gyrus. This chapter will provide an overview of our current understanding of that system based on available anatomical and physiological data. Keywords: dentate gyrus; cannabinoids; CB1; retrograde transmission; short term plasticity; presynaptic modulation. depolarization-induced suppression of inhibition (DSI) (Ohno-Shosaku et al., 2001; Wilson and Nicoll, 2001). As originally described by Pitler and Alger in hippocampal pyramidal cells (1992), and by Marty in cerebellar Purkinje cells (Llano et al., 1991), DSI is a form of short term synaptic plasticity whereby depolarization of a single neuron results in transient inhibition of GABAergic afferents to that same neuron. A principal role for ECs in DSI is indicated by the fact DSI is blocked by CB1 antagonists, occluded by CB1 receptor agonists, and absent in CB1 / animals (Kreitzer and Regehr, 2001a; Ohno-Shosaku et al., 2001; Wilson and Nicoll, 2001; Wilson et al., 2001; Varma et al., 2001; Yoshida et al., 2002). The retrograde nature of the EC-mediated signaling was originally (and most simply) indicated by failure of depolarization-induced EC release (or bath application synthetic agonists) to inhibit responses to local application of exogenous GABA, despite effectively reducing evoked responses. This and other physiological data is strongly supported by an array of immunohistochemical studies that indicate prominent expression of CB1 receptors both

Introduction Both chronic and acute use of cannabis (marijuana) produces a wide array of physiological, cognitive, and analgesic effects in humans that have been widely recognized for centuries. The active ingredient in marijuana, delta-9-tetrahydrocannabinol (D9-THC), was first identified in 1964 (Gaoni and Mechoulam, 1964); the first primary cannabinoid receptor in the CNS (CB1) was identified and cloned in 1990 (Howlett et al., 1990; Matsuda et al., 1990), and the principal endogenous ligands, arachidonoylethanolamide (AEA) and 2-arachidonoylglycerol (2-AG), were identified soon thereafter (Devane et al., 1992; Sugiura et al., 1995). Efforts to understand the neurophysiological consequences of CB1 receptor activation received an enormous boost in 2001 with the discovery that endogenous cannabinoids (ECs) act as retrograde signaling molecules in a phenomenon known as Corresponding author. Tel.: +1 352 392 3447; Fax: +1 352 392 9187; E-mail: cjfraz@ufl.edu DOI: 10.1016/S0079-6123(07)63019-2

319

320

presynaptically and on GABAergic terminals (for general review see Freund et al., 2003, and below). Although it has only been a short time since these landmark studies, EC-dependent DSI governed by fundamentally similar mechanisms has now been identified in numerous other brain areas including the amygdala, substantia nigra, basal ganglia, neocortex, brainstem, and recently, in both granule cells and mossy cells of the dentate gyrus (Trettel and Levine, 2003; Yanovsky et al., 2003; Bodor et al., 2005; Isokawa and Alger, 2005; Mukhtarov et al., 2005; Zhu and Lovinger, 2005; Engler et al., 2006; Hofmann et al., 2006). Cumulatively, these findings highlight what is now the obviously broad significance of EC-mediated retrograde signaling in modulating inhibitory transmission in the CNS, but nevertheless, represent only a small percentage of recent advances. For example, it is now clear that cannabinoid receptors are expressed on many glutamatergic terminals as well. Accordingly, depolarization-induced EC-mediated inhibition of excitation (DSE) has now been described in the cerebellum (Kreitzer and Regehr, 2001b), ventral tegmental area (Melis et al., 2004a, b), hypothalamus (Di et al., 2005), amygdala (Domenici et al., 2006), and in the hippocampus (Ohno-Shosaku et al., 2002b), where the role of ECs in modulating Schaffer collateral inputs to CA1 pyramidal cells remains a particularly active area of investigation (Hajos and Freund, 2002; Hoffman et al., 2005; Katona et al., 2006; Takahashi and Castillo, 2006). Further, efforts to understand the role of postsynaptic metabotropic glutamate receptors in EC release have proven instrumental in identifying ECmediated long-term depression (LTD) in a number of areas. At present, both heterosynaptic and homosynaptic forms of EC-dependent LTD have been described, affecting both GABAergic and glutamatergic synapses respectively (Marsicano et al., 2002; Robbe et al., 2002; Chevaleyre and Castillo, 2003; Hoffman et al., 2003). While presynaptic expression still seems to be the norm in this form of EC-dependent synaptic plasticity, there has even been a recent example of EC-mediated, mGluR dependent LTD in the cerebellum that is expressed postsynaptically (Safo and Regehr, 2005). Given the extensive nature of EC-mediated effects on synaptic transmission, it should perhaps not come

as a surprise that fascinating metaplastic roles for ECs are also now coming to light. For example, Castillo and colleagues have recently and elegantly demonstrated that EC-mediated LTD can lower the threshold for traditional NMDA-dependent LTP in spatially confined parts of the CA1 pyramidal cell dendritic tree (Chevaleyre and Castillo, 2004). These types of findings are mentioned here in brief simply to underscore our growing appreciation for the enormous breadth of EC-dependent signaling in CNS physiology, and to highlight the extremely rapid nature of recent progress. In my view, this trend shows no sign of slowing. In fact, in this chapter, I will attempt to make the case that the dentate gyrus is an area of significant interest, and strong potential, for this field; albeit one that has received comparatively little attention to date. In making that argument I will highlight a significant body of existing evidence clearly indicating the presence of a robust system for EC-dependent signaling in the dentate gyrus, I will review the growing list of studies that reveal specific neurophysiological consequences of EC-dependent signaling at identified synapses in the dentate, and I will cover recent work suggesting several novel roles for EC signaling (from neuroprotection to neurogenesis) in which the dentate may be an area of particular interest. Potential future directions for some of this work will also be discussed. For a broader overview of specific roles for ECs in CNS physiology the interested reader is referred to any number of recent and outstanding reviews (Schlicker and Kathmann, 2001; Alger, 2002; Wilson and Nicoll, 2002; Freund et al., 2003; Chevaleyre et al., 2006; Marsicano and Lutz, 2006).

A brief overview of the endocannabinoid system There are currently two known cannabinoid receptors: the CB1 receptor, cloned in 1990 (Matsuda et al., 1990), and the CB2 receptor, identified 3 years later (Munro et al., 1993). Both receptors are classic metabotropic receptors coupled to Gi/o. Of these, the CB1 receptor appears to be intimately involved in mediating most central effects of ECs. It is among the most prominently expressed metabotropic receptors in the mammalian CNS, with

321

overall expression levels approaching that of ionotropic receptors for glutamate and GABA. Particularly high levels of CB1 expression are found in hippocampus (including the dentate gyrus), cortex, cerebellum, and basal ganglia (Herkenham et al., 1991a, c; Tsou et al., 1998a). This distribution is consistent with well-recognized effects of cannabinoids on movement and memory. The CB2 receptor by contrast, has been reported in the CNS only in striatum and cerebellum (Skaper et al., 1996; Van Sickle et al., 2005), but is far more abundant in the immune system where it is thought to be involved in regulation of both immune function and inflammatory responses (for review see Berdyshev, 2000; De et al., 2000). Since the CB2 receptor is believed to be absent in the dentate, it will not be reviewed further here. There are likely several endogenous ligands for the CB1 receptor, of which two have been particularly well characterized to date: arachidonoylethanolamide (anandamide, abbreviated AEA) and 2-AG. Both of these ligands are fatty acid derivatives, although they differ significantly in terms of their synthetic pathway. AEA is thought to be synthesized by phospholipase D induced hydrolysis of a lipid precursor, N-arachidonoyl phosphatidylethanolamine (N-arachidonoyl-PE) (Di Marzo et al., 1994, 1999; Cadas et al., 1997). One of the primary immediate precursors for synthesis of 2-AG, by contrast, appears to be diacylglycerol, which is produced from phosphatidylinositol by phospholipase C, and subsequently converted to 2-AG by diacylglycerol lipase (DAGL) (Stella et al., 1997; Piomelli, 2003). Because both AEA and 2-AG are membrane permeant, conventional vesicular release mechanisms are implausible. Thus, regulation of EC-mediated signaling is likely to depend principally on regulation of the synthetic pathways described above, on uptake from the extracellular space, and on subsequent breakdown. These processes also differ somewhat between AEA and 2-AG. Synthesis of AEA is increased significantly by rapid increases in intracellular calcium concentration (that reach low micromolar levels) such as occur with membrane depolarization. The calcium dependence of AEA production is tied to the fact that availability of Narachidonoyl-PE depends heavily on the activity of

a highly calcium-dependent enzyme, N-acyltransferase, which is capable of producing N-arachidonoyl-PE from phosphatidylethanolamine (Di Marzo et al., 1994; Cadas et al., 1996; Piomelli, 2003). While it is clear that increases in intracellular calcium can also increase production of 2-AG through a PLC and DAG dependent pathway (Stella et al., 1997), the overall story here is somewhat more complicated. For example, in some systems calcium dependent production of 2-AG is likely to be significantly enhanced if the calcium influx occurs coincident with synaptic activation of certain metabotropic receptors (notably metabotropic glutamate receptors and/or muscarinic acetylcholine receptors) capable of activating a calcium-dependent isoform of phospholipase C (Hashimotodani et al., 2005; Maejima et al., 2005). On the other hand, there also appear to be metabolic routes through which metabotropic receptors can promote largely calcium-independent production of 2-AG (Maejima et al., 2001; Varma et al., 2001; Ohno-Shosaku et al., 2002a, 2005; Fukudome et al., 2004). Once AEA has reached the extracellular space, uptake is mediated by a fast, selective, saturable, temperature-dependent, and thus apparently carrier-mediated process (Beltramo et al., 1997; Hillard et al., 1997; Piomelli et al., 1999). Subsequently, AEA is rapidly broken down to arachidonic acid and ethanolamine by fatty acid amide hydrolase (FAAH), an enzyme that has an expression pattern in the CNS that closely corresponds to that of the CB1 receptor (Hillard et al., 1995; Cravatt et al., 1996; Egertova et al., 1998; Piomelli et al., 1999; Deutsch et al., 2002). The corresponding story is again more complicated for 2-AG. While it is clear that the anandamide transporter can also facilitate uptake of 2-AG, and that FAAH can also break it down, it is too early to rule out the possibility of an additional transport mechanism for 2-AG, and it is clear that at least one other major route for breakdown, in this case mediated by monoglyceride lipase (MGL), is available (Dinh et al., 2002). Activation of metabotropic CB1 receptors located primarily on presynaptic terminals may be coupled to adenylyl cyclase, or directly to either Ca2+ or K+ conductances, and is typically (but perhaps not exclusively) observed to inhibit calcium dependent release of neurotransmitter. The

322

general consensus is that such activation of endogenous CB1 receptors typically occurs through retrograde transmission of endocannabinoids synthesized postsynaptically, and further that 2-AG and/or AEA are likely to be the retrograde messengers in most if not all forms of EC-mediated retrograde signaling identified to date. It is interesting to note that 2-AG is present in brain at concentrations approximately 200-fold greater than AEA, although this may simply reflect a more central role in lipid metabolism and not necessarily a similarly dominant role in retrograde signaling (Sugiura et al., 1995; Stella et al., 1997). For interesting and recent discussion of the relative roles of 2-AG and AEA in retrograde signaling see Chevaleyre et al. (2006). While this represents a brief and general summary of the EC system as currently understood in the brain (including the dentate) it is important to reiterate that this field is advancing rapidly. Significant advances in our conceptions of this system are certainly possible, and perhaps even probable. One area where our knowledge is currently developing at a rapid pace concerns CB1 receptor-mediated modulation of glutamatergic transmission (see below). Further, it is likely that additional sites of EC action will be identified, which may include currently uncloned CB-like receptors and/or conventional nonCB receptors not previously associated with EC action. Thus it seems likely that our understanding of the possible modulatory effects of ECs on synaptic transmission in the CNS will only broaden with time. A more extensive discussion of the molecular aspects of EC signaling is available in a number of other reviews (Di Marzo and Deutsch, 1998; Lutz, 2002; Piomelli, 2003; Sugiura et al., 2006).

Markers of the EC system in the dentate gyrus CB1 receptor expression in the dentate gyrus: initial findings indicate prominent role in GABAergic neurons Early efforts to map the distribution of putative CB receptors in brain were hindered by the fact that they relied largely on radiolabeled derivatives of D9THC that had generally poor affinity for the

receptors and a strong tendency to partition into biological membranes (Harris et al., 1978; Roth and Williams, 1979; Thomas et al., 1992). However, in 1991 two landmark studies capitalized on the favorable properties of a synthetic CB1 agonist, CP55,940, to provide one of the first complete surveys of putative cannabinoid receptor binding sites in the CNS (Herkenham et al., 1991a, b). These studies demonstrated that the molecular layer of the dentate gyrus was among the most intensely labeled areas in the rat brain. Comparatively low to moderate levels of 3H-CP55,940 binding were also reported in the hilus, with only sparse labeling over the granule cell layer. Interestingly, these studies also reported differential binding of 3H-CP55,940 along the septo-temporal axis of the hippocampus, with denser binding on the dorsal (septal) end. These findings, which were reported from rat brain, were largely consistent with later studies using the same radioligand in human (Glass et al., 1997). While the original studies with 3H-CP55,940 represented a significant advance in their day, it is noteworthy that they were completed at about the same time that the CB1 receptor was isolated and cloned (Matsuda et al., 1990). That feat enabled two additional approaches to the study of CB1 receptor expression and distribution. The first to be employed was in situ hybridization for CB1 mRNA (Mailleux and Vanderhaeghen, 1992; Mailleux et al., 1992; Matsuda et al., 1993; Marsicano and Lutz, 1999). In general, these studies noted high levels of CB1 mRNA in a subpopulation of cells in the subgranular region of the hilus, where 3H-CP55,940 was low to moderate, with comparatively less concentrated expression of CB1 mRNA in the molecular layer, where 3H-CP55,940 binding had been intense. This differential distribution between CB1 mRNA expression and 3HCP55,940 binding was the basis for one of the first arguments that functional CB1 expression may be relegated largely to processes distal from the soma. The second major advance enabled by cloning of the CB1 receptor was the immunohistochemical characterization of receptor distribution in the CNS. This technique offered several significant advantages over earlier efforts. First, because the initial antibodies employed were targeted to specific amino acid sequences on the N-terminal of

323

the cloned CB1 receptor, they offered significantly greater specificity than earlier radioligand binding studies. Second, the higher resolution of immunohistochemical techniques allowed the first detailed examination of CB1 expression at the subcellular level. Third, the eventual application of elegant dual-labeling immunohistochemical techniques allowed simultaneous analysis of CB1 expression and cell phenotype. As such, our current understanding of CB1 expression and distribution in the CNS depends heavily on these techniques, and results relevant to the dentate will be reviewed in some detail below. The first immunohistochemical studies of CB1 distribution in the CNS clearly demonstrated intense immunoreactivity in the molecular layer of the dentate gyrus (Pettit et al., 1998; Tsou et al., 1998a; Katona et al., 1999). Consistent with early binding studies, the intensity of this signal was reported to be highest in the inner third of the molecular layer, and on the septal end of the hippocampus (Fig. 1AA, Tsou et al., 1998a). However, based on the greater resolution of these techniques, it now became apparent that this area specifically contained many CB1-immunoreactive nerve fibers, but very few labeled cell bodies. By contrast, intensely immunoreactive neuronal cell bodies were identified in the hilus, just below the granule cell layer. These cells were reported to represent as much as 62% of all CB1-immunoreactive cell bodies in the dentate, and had characteristic pyramidal morphology, with a single primary dendrite that often extended without branching through the granule cell layer toward the molecular layer (Fig. 1AB C, Tsou et al., 1998a; Katona et al., 1999). The remaining CB1-positive cell bodies were located largely in the hilus and exhibited a variety of morphologies. The subsequent application of dual-labeling immunohistochemical techniques represented a major advance in this area, as it allowed the determination that the vast majority of CB1immunoreactive cell bodies were also positive for CCK. Specifically, between 80 and 95% of all CCK-positive cell bodies in the dentate were reported to also be CB1-positive, with a comparably high percentage of CB1-positive cells being CCKpositive. In stark contrast, virtually none of the parvalbumin-positive basket cells expressed CB1

(Katona et al., 1999; Tsou et al., 1999). Further examination of these results at the electron microscopic level indicated cellular expression of CB1 was largely or exclusively confined to cytoplasmic organelles, while surface expression of CB1 was very high on CCK-positive axon terminals that were often clustered around principal cell somas and proximal dendrites (Katona et al., 1999). Although these particular observations were originally made primarily in hippocampal cells in CA1 and CA3, it now seems reasonable to predict that surface expression of CB1 on GABAergic neurons in the dentate may also be restricted to nerve terminals. This hypothesis is consistent with the differential distribution of 3H-CP55,940 binding and in situ hybridization described earlier, and is also strongly supported by a study from the Freund laboratory the following year that specifically identified CB1-positive axon terminals in the hilus (Acsady et al., 2000). Although this study did not directly perform dual-labeling experiments for both CB1 and CCK, it did make the striking observation that both CB1-positive and (in separate experiments) CCK-positive terminals displayed very similar and yet highly unusual target selectivity. Specifically they found these terminals made abundant synaptic contacts on the soma and proximal dendrites of the GluR2/3 and calcitonin generelated peptide (CGRP)-positive hilar mossy cells, but largely avoided both parvalbumin and substance P (and thus also CCK)-immunoreactive interneurons in the hilus. Cumulatively, the data described so far clearly indicates prominent expression of CB1 receptors in the dentate, particularly on presynaptic terminals of CCK-positive interneurons, many of which have cell bodies located in the subgranular zone. The specific distribution of CB1-positive terminals further suggests a likely role for CB1 activation in regulating GABAergic transmission to both granule cells and hilar mossy cells. Nevertheless the apparent absence of CCK and CB1positive terminals making synaptic contact with other basket cells clearly distinguishes hilar circuitry from that in other hippocampal subfields, and may hold clues to understanding the unique physiology of this area. While the vast majority of the work described above was completed in either Sprague-Dawley or

324

Fig. 1. CB1 receptor expression in the dentate gyrus. (AA) An early study demonstrates intense immunoreactivity for CB1 receptors in the dentate gyrus, especially in the inner molecular layer (mo). (AB C) Although dentate granule cells appear to be CB1 negative, there are many intensely stained neurons found at the base of the granule cell layer, with characteristic pyramidal morphology. Other intensely stained neurons were found in the hilus (not shown). Note that labeled cell bodies were generally absent in the molecular layer. Scale bars: 200 mm, 50 mm, 20 mm in AA–C respectively. Adapted with permission from Tsou et al. (1998a). (B) A recent immunohistochemical analysis of CB1 receptor expression in wild type (WT), CaMK-CB1 / , GABA-CB1 / , and complete CB1 / mice provided clear evidence of CB1 expression on glutamatergic terminals (as well as GABAergic terminals) in the inner molecular layer of the dentate gyrus. Areas indicated by black boxes in the top row are shown in higher magnification in the bottom row. Scale bar in H is 100 mm. Adapted with permission from Monory et al. (2006).

Wistar rats, the basic features of CB1 expression in the dentate described so far appear to be largely conserved in gray mouse lemur, primate, and human brain (Ong and Mackie, 1999; Katona et al., 2000; Harkany et al., 2005). Further, although much of this work has relied on the N-terminal antibodies to the CB1 receptor that are identical or similar to the one originally developed by Tsou et al. (1998a), key features of this expression pattern have also been observed using antibodies that target intracellular sites on the C-terminal end of the CB1 receptor (Egertova and Elphick, 2000; Hajos et al., 2000). Nevertheless, at this juncture it is important

to emphasize that the data as summarized above do not represent the full picture with respect to likely sites of EC action in the dentate. Our current understanding of this topic with respect to glutamatergic systems is reviewed below.

CB1 receptor expression in the dentate gyrus: emphasis on glutamatergic neurons Compared to the results summarized above, the question of CB1 receptor expression by glutamatergic neurons has so far been a contentious

325

one. This is particularly true in areas CA1 and CA3 where apparent expression of low levels of CB1 mRNA in pyramidal neurons as indicated by in situ hybridization generally has not been easy to reconcile with the apparent lack of CB1 immunoreactivity on glutamatergic cell bodies or axon terminals as indicated by immunohistochemical techniques. This discrepancy has been made all the more acute by divergent reports from a number of outstanding laboratories regarding the physiological effects of ECs on glutamatergic transmission in the Schaffer collateral pathway (Hajos and Freund, 2002; Hoffman et al., 2005; Katona et al., 2006; Takahashi and Castillo, 2006). However, in general, the current details of this debate are outside the scope of this review. The dentate gyrus has to a large extent been spared from a similar debate in part because, unlike in area CA1 and CA3, in situ hybridization tends to agree with both immunohistochemical analysis and early radioligand binding studies in suggesting that dentate granule cells are CB1-negative. Nevertheless, it would be inaccurate to say there have been no divergent results in this area at all. For example, a unique N-terminal antibody targeting amino acids 83–91 of the CB1 receptor employed by Pettit et al. (1998) does label dentate granule cells, while the Tsou et al. (1998a) antibody that fails to label dentate granule cells in rats and humans has been reported to do so in non-human primates. While the precise reasons for these discrepancies in the dentate gyrus have still not been completely resolved, very recent work using novel C-terminal antibodies to CB1 has simultaneously reinforced the conclusion that dentate granule cells are CB1negative, while elegantly demonstrating a potentially major role for EC signaling via the other glutamatergic cell type in the dentate, the hilar mossy cells. The first of two significant reports in this area was from the laboratory of Tamas Freund (Katona et al., 2006). This study used a guinea pig anti-CB1 antibody originally characterized by Fukudome et al. (2004) that targets amino acids 443–473 on the C-terminal end of the protein. The pattern of CB1 immunoreactivity observed with this antibody had similar distribution to that previously reported, however the intensity of labeling

was notably greater than had been observed previously, particularly in the inner third of the molecular layer. Further, upon examination of the results at the electron microscopic level it became clear that, in contrast to results with earlier antibodies, as many as 80% of asymmetrical (i.e. excitatory) synapses in the inner molecular layer, and 30–50% of asymmetrical synapses in other layers were unambiguously positive for CB1. No differences in this pattern were observed between C57BL/6 and CD1 mice, and the authors noted that preliminary data suggests similar results will be obtained using this antibody in both rat and human hippocampus. A second extremely recent study has reproduced and extended these findings using yet another unique antibody, in this case a goat anti-CB1 antibody targeting the C-terminal 77 amino acids of the protein (Monory et al., 2006). While providing other significant insights to be reviewed later in this chapter, this study confirms the presence of CB1-positive glutamatergic terminals in the inner third of the molecular layer by elegantly demonstrating clear immunohistochemical labeling for CB1 that is resistant to selective CB1 knockouts targeted specifically at GABAergic neurons, but reduced by selective CB1 knockouts targeting glutamatergic cells (Fig. 1B, for more on this technology, also see Marsicano et al., 2003). The observation that apparently glutamate-specific CB1 labeling was concentrated in the inner third of the molecular layer, combined with the observation that dentate granule cells continued to lack immunoreactivity for CB1 in all models, suggested the hypothesis that glutamate-specific CB1 labeling could be related to presynaptic expression on the axon terminals of hilar mossy cells. Consistent with that hypothesis, these authors further use dual-labeling in situ hybridization techniques to show strong co-localization of mRNA for CB1 and for the vesicular glutamate transporter type 1 (VGlutT1) both in the soma of hilar mossy cells and in presynaptic terminals forming asymmetric synaptic contacts in the inner molecular layer. These findings do not conflict with earlier work using other antibodies that indicated robust expression of CB1 on CCK-positive axon terminals in this area. Instead, they indicate additional

326

expression of CB1 on glutamatergic terminals likely belonging to hilar mossy cells that was not previously detected. This is a striking conclusion by any measure, and I suspect it will have a significant impact on future work in this area. Nevertheless, the precise reasons that CB1 expression by glutamatergic terminals in the molecular layer was not detected earlier and continues to be somewhat unclear. One likely possibility is that the Nterminal antibodies generally employed in earlier work simply lack the sensitivity to detect lower levels of CB1 expression present on glutamatergic terminals. Another compatible possibility that has been raised, but to my knowledge not yet formally reported, is that there exists a ‘CB1 receptor-interacting protein’ that is exclusively expressed by excitatory cells and is capable of interfering with sites targeted by some of the earlier C-terminal antibodies (Niehaus et al., 2004).

CB1 receptor expression in the dentate gyrus: coexpression with other receptor subtypes While a concerted effort has been made to characterize CB1 expression with respect to glutamatergic vs. GABAergic neurons and presynaptic vs. somatic expression as described above, some otherwise isolated reports have specifically examined co-expression of CB1 with other receptor subtypes. Noteworthy examples in the dentate include a report by Cristino et al. (2006) who used dual-labeling immunohistochemical techniques to suggest co-expression of both CB1 and the vanilloid receptor TRPV1 in many non-pyramidal bipolar neurons of the dentate, particularly in the molecular layer. Further, Morales and Backman (2002) used double labeling in situ hybridization techniques to demonstrate a strong co-localization of mRNA for both CB1 receptors and 5-HT3A receptors in the subgranular interneurons of the dentate hilus. This result was generally consistent with Hermann et al. (2002) who also demonstrated co-localization of CB1 with 5-HT1B, and 5-HT3 receptors, and further noted the extent of co-expression was higher in neurons with high levels of CB1 mRNA. This study also was noteworthy in identifying co-expression of CB1 with D2 receptors throughout the dentate.

These findings suggest the possibility of significant cross talk between several transmitter systems in regulating the activity of CCK-positive interneurons in the dentate.

Other markers of the EC system in the dentate gyrus Although a few additional approaches have been employed, most efforts to demonstrate the presence of the EC system in the hippocampus and dentate gyrus without relying on detection of CB1 have focused on the two primary enzymes for degradation of AEA and 2-AG reviewed earlier: FAAH and MGL, respectively. At this point, immunohistochemical studies targeting various areas on the C-terminal end of FAAH have indicated that expression is predominantly postsynaptic and largely confined to hippocampal principal cells and dentate granule cells (Tsou et al., 1998b; Romero et al., 2002; Egertova et al., 2003; Gulyas et al., 2004). This distribution is consistent with that reported for mRNA for FAAH (Thomas et al., 1997), and indicates that FAAH expression is largely restricted to neurons known to receive CB1-positive afferent inputs. The diffuse labeling for FAAH outside the cell layers noted by Egertova et al. (2003) may be an artifact of tissue preparation, or as noted by the authors, could indicate an ability of principal cells to excrete FAAH into the extracellular space. Interestingly, Gulyas et al. (2004) has also noted that intracellular expression of FAAH was most often localized to the surface of organelles associated with calcium storage. In contrast to the predominantly postsynaptic localization of FAAH, MGL appears to be expressed primarily presynaptically in both the hippocampus and dentate gyrus. Specifically, immunoreactivity for MGL has been identified on the mossy fiber axon terminals of dentate granule cells, as well as on the terminals of CA3 pyramidal cells and some interneurons (Gulyas et al., 2004). In contrast, absence of MGL immunoreactivity in certain areas of CA3 and the molecular layer noted in the same study suggested that CA1 and perforant path inputs are likely to lack this enzyme.

327

Another noteworthy technique in this area has been to examine the expression of DAGL, a biosynthetic enzyme central to the production of 2AG. Two very recent studies that were published almost simultaneously have both demonstrated that DAGL, as detected by immunohistochemical analysis using several different antibodies, is highly expressed by principal neurons in the hippocampus and granule cells (and likely mossy cells) of the dentate gyrus, and has further indicated that such expression is almost completely restricted to the head and neck of dendritic spines (Katona et al., 2006; Yoshida et al., 2006). Katona et al. (2006) also showed specifically that presynaptic terminals opposite DAGL positive spines are CB1-positive, indicating that DAGL is perfectly positioned to promote retrograde transmission, presumably mediated by 2-AG at numerous sites throughout the hippocampus and dentate gyrus.

Physiological role for ECs in the dentate Early efforts to examine the neurophysiological effects of cannabinoids began well before isolation and cloning of the CB1 receptor. Both in vivo and in vitro approaches were employed; experimental measures ranged from auditory and sensory evoked potentials to extracellular field potentials and unit recordings. Using such techniques, a number of studies specifically demonstrated effects of cannabinoids on identified cells or circuits in the dentate gyrus (see Kujtan et al., 1983; Campbell et al., 1986a, b; Wilkison and Pontzer, 1987; Hampson et al., 1989). However, just as cloning of the CB1 receptor and development of powerful immunohistochemical techniques dramatically expanded our understanding of CB1 expression in the brain, major advances in electrophysiological techniques have now provided the technology necessary to dissect the effects of both endogenous and exogenous cannabinoids at the synaptic level. Admittedly, applications of such technology to the study of cannabinoid-dependent systems in the dentate gyrus are only recently beginning to appear in the literature. Nevertheless, the emphasis below will be on such recent findings, with a particular effort to note when and where physiological

findings to date are consistent with predictions based on the existing anatomical information reviewed above.

DSI in the dentate gyrus Anatomical studies reviewed earlier strongly suggest that CB1-positive GABAergic terminals of CCK-positive basket cells make numerous synaptic contacts with both granule cells and hilar mossy cells. These findings suggest such inputs may be modulated by CB1 agonists. Further, the presence of FAAH and/or DAGL in both granule cells and mossy cells suggests they may represent sources of ECs. Detailed physiological studies have now strongly reinforced these conclusions. The first evidence that GABAergic inputs to dentate granule cells were sensitive to exogenous cannabinoids was published in 2000 when Hajos et al. used whole-cell recording techniques to demonstrate that evoked inhibitory postsynaptic currents (eIPSCs) recorded from dentate granule cells were reduced to 3676% of baseline by bath application of WIN55,212-2 in wild type but not in CB1 / mice (Hajos et al., 2000). Several years later, Nakatsuka et al. (2003) similarly reported a WIN55,212-2-mediated, AM-251-sensitive reduction in both frequency and amplitude of spontaneous IPSCs (sIPSCs) recorded from dentate granule cells in the human dentate gyrus. These findings indicated that GABAergic inputs to dentate granule cells are indeed inhibited by CB1 receptor activation, however they did not directly address the question of whether granule cells were capable of activating these receptors by signaling via ECs. That question was answered in the affirmative with the first complete characterization of DSI in dentate granule cells in 2005 (Isokawa and Alger, 2005). This study specifically demonstrated that direct depolarization of a single dentate granule cell reduces the amplitude of both sIPSCs and IPSCs evoked by stimulation of the molecular layer (Fig. 2A). Endocannabinoids were implicated in this process because both CP55,942 and WIN55,212-2 mimicked and occluded DSI and because DSI was blocked by the CB1 antagonist

328

Fig. 2. DSI in dentate granule cells and hilar mossy cells. (AA) Dentate gyrus of the hippocampal slice (scale, 50 mm). (AB) Differential interference contrast image of dentate granule cells (scale, 10 mm). (AC) Single dentate granule cell filled, and visualized with fluorescence microscopy during whole cell recording (scale, 10 mm). (AD) DSI of eIPSCs (left, top) and sIPSCs (right, top) is observed in dentate granule cells. In both cases, magnitude of DSI is reduced by bath application of thapsigargin (TG, 2 mM, bottom). (AE) Thapsigargin also reduced the calcium transients produced by DSI-inducing depolarizing voltage steps. Adapted with permission from Isokawa and Alger (2005). (B) Depolarization induced release of endogenous cannabinoids from hilar mossy cells preferentially inhibits calcium-dependent exocytosis. DSI in of eIPSCs is apparent in an individual mossy cell. DSI of miniature IPSCs is absent in the same cell. Following bath application of KCl and CaCl2, DSI of mIPSCs becomes apparent. Summary data are presented in BD, where numbers on the bars are n values. Adapted with permission from Hofmann et al. (2006). (See Color Plate 19.2 in color plate section.)

SR141716A. These authors further demonstrated that the magnitude of DSI correlated with the extent of depolarization-induced increases in [Ca2+]i and that DSI could be blocked by chelating postsynaptic calcium with internal BAPTA. Finally, this study indicated a particular role for calcium release from ryanodine sensitive internal

calcium stores in the mechanism of DSI, as depletion of those stores reduced both [Ca2+]i and the magnitude of DSI. Interestingly, an earlier study had demonstrated that although DSI of sIPSCs was absent in dentate granule cells in control conditions, it could be detected 1 week or more following induction of

329

febrile seizures (Chen et al., 2003). Induction of febrile seizures also enhanced EC-mediated signaling in area CA1, apparently due to increases in CB1 expression that were restricted to CCK-positive GABAergic terminals. While the finding that febrile seizures upregulate EC-mediated signaling throughout the hippocampus and dentate gyrus remains quite intriguing, the apparent discrepancy with respect to the ability to detect DSI in dentate granule cells in control conditions remains unexplained. Also consistent with predictions based on anatomical data is a recent report from my laboratory which provides the first published characterization of DSI as observed in hilar mossy cells (Hofmann et al., 2006). Specifically, we reported DSI of eIPSCs in hilar mossy cells that depends on both postsynaptic calcium influx and presynaptic CB1 receptors. The magnitude of DSI in hilar mossy cells was directly dependent on depolarization duration, and enhanced by bath application of carbachol (CCh). Further, the presynaptic mechanism of inhibition was found to be largely or exclusively selective for calcium-dependent release events (Fig. 2B). We also used dual whole cell recording techniques to demonstrate that there are tight spatial constraints on diffusion of ECs released from hilar mossy cells. Finally, we were intrigued to observe two different forms of expression of DSI of sIPSCs that depended on whether the sIPSCs had high spectral power at theta frequencies. Cumulatively, these results suggested a prominent role for ECmediated signaling between hilar mossy cells and GABAergic afferents in normal hilar function, and raise interesting questions about the effects of both ECs and CCh on coordinated activity in GABAergic networks. An earlier report from another laboratory indicating DSI of sIPSCs could be detected in hilar mossy cells appeared in abstract form (Howard et al., 2003).

Evidence that ECs modulate glutamatergic transmission in the dentate gyrus The axon terminals of mossy fibers, perforant path inputs, mossy cells, and back-projecting CA3 pyramidal cells all form glutamatergic synapses

within the dentate gyrus. Nevertheless, I think it is fair to say that despite a number of studies clearly indicating antiepileptic (Consroe et al., 1975; Consroe and Wolkin, 1977; Wallace et al., 2001, 2002, 2003; Blair et al., 2006), and neuroprotective (Nagayama et al., 1999; Braida et al., 2000; Van der Stelt et al., 2001a, b, 2002; Khaspekov et al., 2004) effects of cannabinoids, there is a notable lack of physiological studies that directly test the hypothesis that ECs can modulate excitatory transmission at these synapses. For example, to my knowledge, there are currently no published reports directly testing effects of ECs on mossy fiber transmission. Preliminary data from my laboratory is consistent with the weight of existing anatomical evidence in suggesting that isolated mossy fiber inputs to CA3 pyramidal cells are indeed CB1-negative; however it must be noted that neither our preliminary work nor existing anatomical data can at present explicitly rule out the possibility of additional non-CB1 receptors for endocannabinoids at these terminals. There are somewhat more available data with respect to perforant path inputs to dentate granule cells. Although these inputs were initially reported to be insensitive to D9-THC using in vivo extracellular recording techniques (Wilkison and Pontzer, 1987), two later reports using different agonists and activation strategies have been consistent with the hypothesis that these terminals may express presynaptic CB1 receptors. One such study explicitly demonstrated that bath application of WIN55,212-2 produced a rightward shift in the input/output curve, and a reduction in paired pulse facilitation, of perforant path synaptic potentials recorded from the outer third of the molecular layer (Kirby et al., 1995). Further, another study examining field potential responses produced by stimulation of the medial perforant path strikingly found that reductions in fEPSPs produced by an acetylcholinesterase inhibitor (physostigmine) were blocked by the CB1 receptor antagonist AM-251, but not by methoctramine, an antagonist presumably selective for presynaptic (m2) muscarinic acetylcholine receptors. These results suggested that cholinergic facilitation of EC release can lead to tonic ECmediated inhibition of perforant path inputs to

330

dentate granule cells, and as such may be consistent with anatomical studies which indicate low levels of CB1 expression in the outer two thirds of the molecular layer. There is also very limited available data that directly addresses whether ECs play a role in modulating release of glutamate from mossy cell axon terminals. Likely sensitivity to inhibition by CB1 activation has been suggested, of course, by recent and striking immunohistochemical findings reviewed earlier, but it is also implicated by the demonstration that WIN55,212-2 sensitive eIPSCs can be produced in dentate granule cells by stimulation in the inner molecular layer, and perhaps most notably by the observation that CB1 receptor knockouts that are selective for glutamatergic neurons reduce the threshold for kainic acid-induced seizures (Marsicano et al., 2003; Monory et al., 2006). Nevertheless, it is likely that much remains to be learned from a detailed electrophysiological examination of EC-mediated signaling at this synapse. Finally, there is also a complete lack of information in the literature about the EC sensitivity of back projecting associational/commissural inputs to hilar neurons or dentate granule cells. As the functional consequences of CB1 receptor activation on Schaffer collateral inputs to CA1 pyramidal cells is currently a matter of much debate and investigation (see ‘Introduction’) it will not be reviewed further here.

A role for endocannabinoids in neurogenesis? In contrast to original conclusions, groundbreaking work over the last 10–15 years has clearly demonstrated that neurogenesis does occur in selective regions of the adult mammalian CNS, most notably in the olfactory bulb and dentate gyrus (for recent review, see Lledo et al., 2006). Indeed, the subgranular zone of the dentate gyrus, previously recognized for intense expression of CB1 receptors in a subset of inhibitory interneurons, has now also been recognized as a primary neuroproliferative zone. Further, Jiang et al. (2005) has used immunohistochemical techniques coupled with injections of 5-bromo-2-deoxyuridine (BrdU, a base analog of thymidine) to show that dividing cells in

the subgranular region of the adult rat dentate gyrus are immunoreactive for CB1. These striking findings, coupled with work completed in culture, have clearly suggested the hypothesis that EC-mediated signaling might modulate neurogenesis in vivo. Consistent with that hypothesis, Jiang et al. (2005) also demonstrated that chronic administration of a synthetic cannabinoid (HU210) increased neurogenesis in the adult dentate gyrus, coincident with anxiolytic and antidepressant like effects. At present, two other studies have made a compatible observation that BrdU-labeling of newborn cells in the adult dentate gyrus is dramatically reduced in CB1 knockout animals compared to wild type mice (Jin et al., 2004; Kim et al., 2006). However, the full story here is likely to be quite complicated because apparently paradoxical increases in neurogenesis produced by administration of CB1 antagonists, including SR141716A and AM-251, prior to BrdU-labeling have been noted by several laboratories (Rueda et al., 2002; Jin et al., 2004). Further, in at least one case, an AEA-mediated reduction in neurogenesis in the dentate gyrus has been reported (Rueda et al., 2002). One possible explanation is that apparent inhibitory effects of cannabinoids on neurogenesis, when they occur, may be CB1 receptor-independent, and perhaps mediated by direct activation of VR1. This is consistent with the observation by Jin et al. (2004) that SR141716A-mediated increases in neurogenesis were preserved in CB1 knockouts, but absent in VR1 knockouts, and also compatible with a much earlier report indicating that both AEA and SR141716A may interact with VR1 receptors (Zygmunt et al., 1999). In contrast, the most recent mechanism suggested for CB1-dependent increases in neurogenesis implicate down regulation of nitric oxide synthase (Di Marzo et al., 2002; Kim et al., 2006).

Non-CB receptor targets of ECs relevant to the dentate A final topic worthy of particular consideration here is the possible abundance of non-CB1 targets for EC-mediated signaling in the dentate gyrus. Although 2-AG has been postulated to be the

331

retrograde messenger in many forms of EC-mediated signaling, a role for AEA is far from ruled out. In fact, a role for AEA is implied by the presence of FAAH in hippocampal principal cells and dentate granule cells, and has occasionally been implicated in depolarization-induced as opposed to synaptically driven release of ECs (for review, see Chevaleyre et al., 2006). The extent to which EC-mediated signaling involves AEA is a particularly important question in large part because AEA and its metabolites, far more than 2-AG, have been associated with action at non-CB receptor sites (for review, see Di Marzo et al., 2002). One such site that may have an important role to play in the dentate gyrus is the VR1 receptor. Of course, VR1 receptor expression is typically associated with sensory fibers, and thus much of the data implicating an interaction of AEA and VR1 is based on apparent AEA-mediated increases in VR1-dependent release of CGRP. However, AEA and its analogs have also been shown to directly activate recombinant VR1 receptors in several types of assays and expression systems (Bisogno et al., 2001; Di Marzo et al., 2001; Ralevic et al., 2001). Further, recent immunohistochemical data has indicated a prominent expression of VR1 in the brain, with particularly high levels in several areas, including the dentate gyrus (Toth et al., 2005). Collectively these findings suggest an interaction between AEA and VR1, which may be relevant in understanding EC-mediated signaling in the dentate. This argument is further supported, albeit indirectly, by reports in other areas of the CNS where VR1 activation by AEA and/or capsaicin has been shown to directly modulate synaptic transmission (Marinelli et al., 2002, 2003). Other potential targets for AEA that could be significant in the dentate include a7-containing nicotinic acetylcholine receptors and 5-HT3 receptors. Interestingly, in both these cases, action of AEA has been shown to have inhibitory effects on recombinant receptors expressed in Xenopus oocytes (Oz et al., 2002, 2003), while additional lines of evidence implicating cannabinoids in serotonergic function have been previously reviewed (Morales, 2006). Further, a potential non-competitive interaction between AEA and muscarinic acetylcholine receptors has been suggested in both binding and

functional assays (Christopoulos and Wilson, 2001; Lanzafame et al., 2004).

Conclusions and future directions The data reviewed here indicate the presence of a robust system for EC-mediated signaling in the dentate gyrus that is clearly composed of both synthetic and degradative enzymes for endocannabinoids as well as cannabinoid receptors. A significant body of work reviewed here has largely coalesced to indicate that there is prominent expression of CB1 receptors on the axon terminals of CCK-positive basket cells in the dentate, and that these CB1-positive terminals likely form numerous synaptic contacts with both granule cells and hilar mossy cells. Consistent with that view, recent work has demonstrated robust modulation of GABAergic synapses to both granule cells and mossy cells that does indeed depend on retrograde activation of CB1 receptors by ECs (Isokawa and Alger, 2005; Hofmann et al., 2006). Until very recently, available data suggested that we would be unlikely to observe major effects of ECs at other types of synapses in the dentate. However, dramatic improvements in immunohistochemical techniques have now clearly challenged that conclusion by indicating robust expression of CB1 receptors on excitatory terminals in the inner third of the molecular layer that are likely to arise from hilar mossy cells, and that appear to be key players in determining threshold to kainic acid-induced seizures (Marsicano et al., 2003; Katona et al., 2006; Monory et al., 2006). These striking findings have strongly reinforced the notion of the dentate gyrus as a key player in EC-mediated signaling, and may highlight it as an attractive system for future research on EC-mediated modulation of glutamatergic transmission. That being said, it seems clear that we still have much to learn from a continued characterization of the EC system at the electron microscopic level, particularly when coupled with detailed physiological studies at the synaptic level. In fact, a number of issues beyond transmitter phenotype are likely to be of particular importance. These include developing a better understanding of

332

physiological stimuli that trigger EC-mediated signaling and more accurate measures of the scope of the effects. This will likely involve further examination of the role of postsynaptic metabotropic receptors in the release process and careful evaluation of potential heterosynaptic and metaplastic signaling mechanisms. At a larger level, it will also be important to understand broader effects of ECs on network activity, where both in vitro and in vivo studies focused on the dentate gyrus will likely be central to elucidating the neuroprotective and antiepileptic effects of cannabinoids. Finally, the potential role of ECs in neurogenesis is an extremely important topic that clearly has broad implications for neurobiology. Although the dentate gyrus has arguably been partially overlooked in the rapid pace of the last 5 years, it now seems clear that those motivated to understand EC-mediated signaling in the CNS have cause to look toward the dentate, while conversely, those who seek a more thorough understanding of the dentate may find significant value in further studies of cannabinoids.

References Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G. and Freund, T.F. (2000) Unusual target selectivity of perisomatic inhibitory cells in the hilar region of the rat hippocampus. J. Neurosci., 20: 6907–6919. Alger, B.E. (2002) Retrograde signaling in the regulation of synaptic transmission: focus on endocannabinoids. Prog. Neurobiol., 68: 247–286. Beltramo, M., Stella, N., Calignano, A., Lin, S.Y., Makriyannis, A. and Piomelli, D. (1997) Functional role of high-affinity anandamide transport, as revealed by selective inhibition. Science, 277: 1094–1097. Berdyshev, E.V. (2000) Cannabinoid receptors and the regulation of immune response. Chem. Phys. Lipids, 108: 169–190. Bisogno, T., Hanus, L., De, P.L., Tchilibon, S., Ponde, D.E., Brandi, I., Moriello, A.S., Davis, J.B., Mechoulam, R. and Di Marzo, V. (2001) Molecular targets for cannabidiol and its synthetic analogues: effect on vanilloid VR1 receptors and on the cellular uptake and enzymatic hydrolysis of anandamide. Br. J. Pharmacol., 134: 845–852. Blair, R.E., Deshpande, L.S., Sombati, S., Falenski, K.W., Martin, B.R. and Delorenzo, R.J. (2006) Activation of the cannabinoid type-1 receptor mediates the anticonvulsant properties of cannabinoids in the hippocampal neuronal culture models of acquired epilepsy and status epilepticus. J. Pharmacol. Exp. Ther., 317: 1072–1078.

Bodor, A.L., Katona, I., Nyiri, G., Mackie, K., Ledent, C., Hajos, N. and Freund, T.F. (2005) Endocannabinoid signaling in rat somatosensory cortex: laminar differences and involvement of specific interneuron types. J. Neurosci., 25: 6845–6856. Braida, D., Pozzi, M. and Sala, M. (2000) CP 55,940 protects against ischemia-induced electroencephalographic flattening and hyperlocomotion in Mongolian gerbils. Neurosci. Lett., 296: 69–72. Cadas, H., di, T.E. and Piomelli, D. (1997) Occurrence and biosynthesis of endogenous cannabinoid precursor, N-arachidonoyl phosphatidylethanolamine, in rat brain. J. Neurosci., 17: 1226–1242. Cadas, H., Gaillet, S., Beltramo, M., Venance, L. and Piomelli, D. (1996) Biosynthesis of an endogenous cannabinoid precursor in neurons and its control by calcium and cAMP. J. Neurosci., 16: 3934–3942. Campbell, K.A., Foster, T.C., Hampson, R.E. and Deadwyler, S.A. (1986a) Delta 9-tetrahydrocannabinol differentially affects sensory-evoked potentials in the rat dentate gyrus. J. Pharmacol. Exp. Ther., 239: 936–940. Campbell, K.A., Foster, T.C., Hampson, R.E. and Deadwyler, S.A. (1986b) Effects of delta 9-tetrahydrocannabinol on sensory-evoked discharges of granule cells in the dentate gyrus of behaving rats. J. Pharmacol. Exp. Ther., 239: 941–945. Chen, K., Ratzliff, A., Hilgenberg, L., Gulyas, A., Freund, T.F., Smith, M., Dinh, T.P., Piomelli, D., Mackie, K. and Soltesz, I. (2003) Long-term plasticity of endocannabinoid signaling induced by developmental febrile seizures. Neuron, 39: 599–611. Chevaleyre, V. and Castillo, P.E. (2003) Heterosynaptic LTD of hippocampal GABAergic synapses: a novel role of endocannabinoids in regulating excitability. Neuron, 38: 461–472. Chevaleyre, V. and Castillo, P.E. (2004) Endocannabinoid-mediated metaplasticity in the hippocampus. Neuron, 43: 871–881. Chevaleyre, V., Takahashi, K.A. and Castillo, P.E. (2006) Endocannabinoid-mediated synaptic plasticity in the CNS. Annu. Rev. Neurosci., 29: 37–76. Christopoulos, A. and Wilson, K. (2001) Interaction of anandamide with the M(1) and M(4) muscarinic acetylcholine receptors. Brain Res., 915: 70–78. Consroe, P. and Wolkin, A. (1977) Cannabidiol: antiepileptic drug comparisons and interactions in experimentally induced seizures in rats. J. Pharmacol. Exp. Ther., 201: 26–32. Consroe, P.F., Wood, G.C. and Buchsbaum, H. (1975) Anticonvulsant nature of marihuana smoking. JAMA, 234: 306–307. Cravatt, B.F., Giang, D.K., Mayfield, S.P., Boger, D.L., Lerner, R.A. and Gilula, N.B. (1996) Molecular characterization of an enzyme that degrades neuromodulatory fattyacid amides. Nature, 384: 83–87. Cristino, L., De, P.L., Pryce, G., Baker, D., Guglielmotti, V. and Di Marzo, V. (2006) Immunohistochemical localization of cannabinoid type 1 and vanilloid transient receptor potential vanilloid type 1 receptors in the mouse brain. Neuroscience, 139: 1405–1415.

333 De, P.L., Melck, D., Bisogno, T. and Di Marzo, V. (2000) Endocannabinoids and fatty acid amides in cancer, inflammation and related disorders. Chem. Phys. Lipids, 108: 191–209. Deutsch, D.G., Ueda, N. and Yamamoto, S. (2002) The fatty acid amide hydrolase (FAAH). Prostaglandins Leukot. Essent. Fatty Acids, 66: 201–210. Devane, W.A., Hanus, L., Breuer, A., Pertwee, R.G., Stevenson, L.A., Griffin, G., Gibson, D., Mandelbaum, A., Etinger, A. and Mechoulam, R. (1992) Isolation and structure of a brain constituent that binds to the cannabinoid receptor. Science, 258: 1946–1949. Di Marzo, V., Bisogno, T., De, P.L., Brandi, I., Jefferson, R.G., Winckler, R.L., Davis, J.B., Dasse, O., Mahadevan, A., Razdan, R.K. and Martin, B.R. (2001) Highly selective CB(1) cannabinoid receptor ligands and novel CB(1)/VR(1) vanilloid receptor ‘‘hybrid’’ ligands. Biochem. Biophys. Res. Commun., 281: 444–451. Di Marzo, V., Bisogno, T., De, P.L., Melck, D. and Martin, B.R. (1999) Cannabimimetic fatty acid derivatives: the anandamide family and other endocannabinoids. Curr. Med. Chem., 6: 721–744. Di Marzo, V., De, P.L., Fezza, F., Ligresti, A. and Bisogno, T. (2002) Anandamide receptors. Prostaglandins Leukot. Essent. Fatty Acids, 66: 377–391. Di Marzo, V. and Deutsch, D.G. (1998) Biochemistry of the endogenous ligands of cannabinoid receptors. Neurobiol. Dis., 5: 386–404. Di Marzo, V., Fontana, A., Cadas, H., Schinelli, S., Cimino, G., Schwartz, J.C. and Piomelli, D. (1994) Formation and inactivation of endogenous cannabinoid anandamide in central neurons. Nature, 372: 686–691. Di, S., Boudaba, C., Popescu, I.R., Weng, F.J., Harris, C., Marcheselli, V.L., Bazan, N.G. and Tasker, J.G. (2005) Activity-dependent release and actions of endocannabinoids in the rat hypothalamic supraoptic nucleus. J. Physiol., 569: 751–760. Dinh, T.P., Carpenter, D., Leslie, F.M., Freund, T.F., Katona, I., Sensi, S.L., Kathuria, S. and Piomelli, D. (2002) Brain monoglyceride lipase participating in endocannabinoid inactivation. Proc. Natl. Acad. Sci. U.S.A., 99: 10819–10824. Domenici, M.R., Azad, S.C., Marsicano, G., Schierloh, A., Wotjak, C.T., Dodt, H.U., Zieglgansberger, W., Lutz, B. and Rammes, G. (2006) Cannabinoid receptor type 1 located on presynaptic terminals of principal neurons in the forebrain controls glutamatergic synaptic transmission. J. Neurosci., 26: 5794–5799. Egertova, M., Cravatt, B.F. and Elphick, M.R. (2003) Comparative analysis of fatty acid amide hydrolase and cb(1) cannabinoid receptor expression in the mouse brain: evidence of a widespread role for fatty acid amide hydrolase in regulation of endocannabinoid signaling. Neuroscience, 119: 481–496. Egertova, M. and Elphick, M.R. (2000) Localisation of cannabinoid receptors in the rat brain using antibodies to the intracellular C-terminal tail of CB. J. Comp. Neurol., 422: 159–171.

Egertova, M., Giang, D.K., Cravatt, B.F. and Elphick, M.R. (1998) A new perspective on cannabinoid signalling: complementary localization of fatty acid amide hydrolase and the CB1 receptor in rat brain. Proc. Biol. Sci., 265: 2081–2085. Engler, B., Freiman, I., Urbanski, M. and Szabo, B. (2006) Effects of Exogenous and Endogenous Cannabinoids on GABAergic Neurotransmission between the Caudate-Putamen and the Globus Pallidus in the Mouse. J. Pharmacol. Exp. Ther., 316: 608–617. Freund, T.F., Katona, I. and Piomelli, D. (2003) Role of endogenous cannabinoids in synaptic signaling. Physiol. Rev., 83: 1017–1066. Fukudome, Y., Ohno-Shosaku, T., Matsui, M., Omori, Y., Fukaya, M., Tsubokawa, H., Taketo, M.M., Watanabe, M., Manabe, T. and Kano, M. (2004) Two distinct classes of muscarinic action on hippocampal inhibitory synapses: M2mediated direct suppression and M1/M3-mediated indirect suppression through endocannabinoid signalling. Eur. J. Neurosci., 19: 2682–2692. Gaoni, Y. and Mechoulam, R. (1964) Isolation, structure and partial synthesis of an active constituent of hashish. J. Am. Chem. Soc., 86: 1646–1647. Glass, M., Dragunow, M. and Faull, R.L. (1997) Cannabinoid receptors in the human brain: a detailed anatomical and quantitative autoradiographic study in the fetal, neonatal and adult human brain. Neuroscience, 77: 299–318. Gulyas, A.I., Cravatt, B.F., Bracey, M.H., Dinh, T.P., Piomelli, D., Boscia, F. and Freund, T.F. (2004) Segregation of two endocannabinoid-hydrolyzing enzymes into pre- and postsynaptic compartments in the rat hippocampus, cerebellum and amygdala. Eur. J. Neurosci., 20: 441–458. Hajos, N. and Freund, T.F. (2002) Pharmacological separation of cannabinoid sensitive receptors on hippocampal excitatory and inhibitory fibers. Neuropharmacology, 43: 503–510. Hajos, N., Katona, I., Naiem, S.S., Mackie, K., Ledent, C., Mody, I. and Freund, T.F. (2000) Cannabinoids inhibit hippocampal GABAergic transmission and network oscillations. Eur. J. Neurosci., 12: 3239–3249. Hampson, R.E., Foster, T.C. and Deadwyler, S.A. (1989) Effects of delta-9-tetrahydrocannabinol on sensory evoked hippocampal activity in the rat: principal components analysis and sequential dependency. J. Pharmacol. Exp. Ther., 251: 870–877. Harkany, T., Dobszay, M.B., Cayetanot, F., Hartig, W., Siegemund, T., Aujard, F. and Mackie, K. (2005) Redistribution of CB1 cannabinoid receptors during evolution of cholinergic basal forebrain territories and their cortical projection areas: a comparison between the gray mouse lemur (Microcebus murinus, primates) and rat. Neuroscience, 135: 595–609. Harris, L.S., Carchman, R.A. and Martin, B.R. (1978) Evidence for the existence of specific cannabinoid binding sites. Life Sci., 22: 1131–1137. Hashimotodani, Y., Ohno-Shosaku, T., Tsubokawa, H., Ogata, H., Emoto, K., Maejima, T., Araishi, K., Shin, H.S. and Kano, M. (2005) Phospholipase Cbeta serves as a coincidence detector through its Ca2+ dependency for triggering retrograde endocannabinoid signal. Neuron, 45: 257–268.

334 Herkenham, M., Lynn, A.B., de Costa, B.R. and Richfield, E.K. (1991a) Neuronal localization of cannabinoid receptors in the basal ganglia of the rat. Brain Res., 547: 267–274. Herkenham, M., Lynn, A.B., Johnson, M.R., Melvin, L.S., de Costa, B.R. and Rice, K.C. (1991b) Characterization and localization of cannabinoid receptors in rat brain: a quantitative in vitro autoradiographic study. J. Neurosci., 11: 563–583. Herkenham, M., Lynn, A.B., Johnson, M.R., Melvin, L.S., de Costa, B.R. and Rice, K.C. (1991c) Characterization and localization of cannabinoid receptors in rat brain: a quantitative in vitro autoradiographic study. J. Neurosci., 11: 563–583. Hermann, H., Marsicano, G. and Lutz, B. (2002) Coexpression of the cannabinoid receptor type 1 with dopamine and serotonin receptors in distinct neuronal subpopulations of the adult mouse forebrain. Neuroscience, 109: 451–460. Hillard, C.J., Edgemond, W.S., Jarrahian, A. and Campbell, W.B. (1997) Accumulation of N-arachidonoylethanolamine (anandamide) into cerebellar granule cells occurs via facilitated diffusion. J. Neurochem., 69: 631–638. Hillard, C.J., Wilkison, D.M., Edgemond, W.S. and Campbell, W.B. (1995) Characterization of the kinetics and distribution of N-arachidonylethanolamine (anandamide) hydrolysis by rat brain. Biochim. Biophys. Acta, 1257: 249–256. Hoffman, A.F., Macgill, A.M., Smith, D., Oz, M. and Lupica, C.R. (2005) Species and strain differences in the expression of a novel glutamate-modulating cannabinoid receptor in the rodent hippocampus. Eur. J. Neurosci., 22: 2387–2391. Hoffman, A.F., Oz, M., Caulder, T. and Lupica, C.R. (2003) Functional tolerance and blockade of long-term depression at synapses in the nucleus accumbens after chronic cannabinoid exposure. J. Neurosci., 23: 4815–4820. Hofmann, M.E., Nahir, B. and Frazier, C.J. (2006) Endocannabinoid mediated depolarization-induced suppression of inhibition in hilar mossy cells of the rat dentate gyrus. J. Neurophysiol., 96: 2501–2512. Howard, A.L., Ratzliff, A.D.H. and Soltesz, I. (2003) Exploring Endocannabinoid signaling in the dentate gyrus. Soc. Neurosci. Abstr., 808.11. Howlett, A.C., Bidaut-Russell, M., Devane, W.A., Melvin, L.S., Johnson, M.R. and Herkenham, M. (1990) The cannabinoid receptor: biochemical, anatomical and behavioral characterization. Trends Neurosci., 13: 420–423. Isokawa, M. and Alger, B.E. (2005) Retrograde endocannabinoid regulation of GABAergic inhibition in the rat dentate gyrus granule cell. J. Physiol., 567: 1001–1010. Jiang, W., Zhang, Y., Xiao, L., Van, C.J., Ji, S.P., Bai, G. and Zhang, X. (2005) Cannabinoids promote embryonic and adult hippocampus neurogenesis and produce anxiolytic- and antidepressant-like effects. J. Clin. Invest., 115: 3104–3116. Jin, K., Xie, L., Kim, S.H., Parmentier-Batteur, S., Sun, Y., Mao, X.O., Childs, J. and Greenberg, D.A. (2004) Defective adult neurogenesis in CB1 cannabinoid receptor knockout mice. Mol. Pharmacol., 66: 204–208. Katona, I., Sperlagh, B., Magloczky, Z., Santha, E., Kofalvi, A., Czirjak, S., Mackie, K., Vizi, E.S. and Freund, T.F.

(2000) GABAergic interneurons are the targets of cannabinoid actions in the human hippocampus. Neuroscience, 100: 797–804. Katona, I., Sperlagh, B., Sik, A., Kafalvi, A., Vizi, E.S., Mackie, K. and Freund, T.F. (1999) Presynaptically located CB1 cannabinoid receptors regulate GABA release from axon terminals of specific hippocampal interneurons. J. Neurosci., 19: 4544–4558. Katona, I., Urban, G.M., Wallace, M., Ledent, C., Jung, K.M., Piomelli, D., Mackie, K. and Freund, T.F. (2006) Molecular composition of the endocannabinoid system at glutamatergic synapses. J. Neurosci., 26: 5628–5637. Khaspekov, L.G., Brenz Verca, M.S., Frumkina, L.E., Hermann, H., Marsicano, G. and Lutz, B. (2004) Involvement of brain-derived neurotrophic factor in cannabinoid receptordependent protection against excitotoxicity. Eur. J. Neurosci., 19: 1691–1698. Kim, S.H., Won, S.J., Mao, X.O., Ledent, C., Jin, K. and Greenberg, D.A. (2006) Role for neuronal nitric oxide synthase in cannabinoid-induced neurogenesis. J. Pharmacol. Exp. Ther., 319: 150–154. Kirby, M.T., Hampson, R.E. and Deadwyler, S.A. (1995) Cannabinoids selectively decrease paired-pulse facilitation of perforant path synaptic potentials in the dentate gyrus in vitro. Brain Res., 688: 114–120. Kreitzer, A.C. and Regehr, W.G. (2001a) Cerebellar depolarization-induced suppression of inhibition is mediated by endogenous cannabinoids. J. Neurosci., 21: RC174. Kreitzer, A.C. and Regehr, W.G. (2001b) Retrograde inhibition of presynaptic calcium influx by endogenous cannabinoids at excitatory synapses onto Purkinje cells. Neuron, 29: 717–727. Kujtan, P.W., Carlen, P.L. and Kapur, B.M. (1983) Delta 9tetrahydrocannabinol and cannabidiol: dose-dependent effects on evoked potentials in the hippocampal slice. Can. J. Physiol. Pharmacol., 61: 420–426. Lanzafame, A.A., Guida, E. and Christopoulos, A. (2004) Effects of anandamide on the binding and signaling properties of M1 muscarinic acetylcholine receptors. Biochem. Pharmacol., 68: 2207–2219. Llano, I., Leresche, N. and Marty, A. (1991) Calcium entry increases the sensitivity of cerebellar Purkinje cells to applied GABA and decreases inhibitory synaptic currents. Neuron, 6: 565–574. Lledo, P.M., Alonso, M. and Grubb, M.S. (2006) Adult neurogenesis and functional plasticity in neuronal circuits. Nat. Rev. Neurosci., 7: 179–193. Lutz, B. (2002) Molecular biology of cannabinoid receptors. Prostaglandins Leukot. Essent. Fatty Acids, 66: 123–142. Maejima, T., Hashimoto, K., Yoshida, T., Aiba, A. and Kano, M. (2001) Presynaptic inhibition caused by retrograde signal from metabotropic glutamate to cannabinoid receptors. Neuron, 31: 463–475. Maejima, T., Oka, S., Hashimotodani, Y., Ohno-Shosaku, T., Aiba, A., Wu, D., Waku, K., Sugiura, T. and Kano, M. (2005) Synaptically driven endocannabinoid release requires Ca2+-assisted metabotropic glutamate receptor subtype 1 to

335 phospholipase C {beta}4 signaling cascade in the cerebellum. J. Neurosci., 25: 6826–6835. Mailleux, P., Parmentier, M. and Vanderhaeghen, J.J. (1992) Distribution of cannabinoid receptor messenger RNA in the human brain: an in situ hybridization histochemistry with oligonucleotides. Neurosci. Lett., 143: 200–204. Mailleux, P. and Vanderhaeghen, J.J. (1992) Distribution of neuronal cannabinoid receptor in the adult rat brain: a comparative receptor binding radioautography and in situ hybridization histochemistry. Neuroscience, 48: 655–668. Marinelli, S., Di Marzo, V., Berretta, N., Matias, I., Maccarrone, M., Bernardi, G. and Mercuri, N.B. (2003) Presynaptic facilitation of glutamatergic synapses to dopaminergic neurons of the rat substantia nigra by endogenous stimulation of vanilloid receptors. J. Neurosci., 23: 3136–3144. Marinelli, S., Vaughan, C.W., Christie, M.J. and Connor, M. (2002) Capsaicin activation of glutamatergic synaptic transmission in the rat locus coeruleus in vitro. J. Physiol., 543: 531–540. Marsicano, G., Goodenough, S., Monory, K., Hermann, H., Eder, M., Cannich, A., Azad, S.C., Cascio, M.G., Gutierrez, S.O., Van der Stelt, M., Lopez-Rodriguez, M.L., Casanova, E., Schutz, G., Zieglgansberger, W., Di Marzo, V., Behl, C. and Lutz, B. (2003) CB1 cannabinoid receptors and on-demand defense against excitotoxicity. Science, 302: 84–88. Marsicano, G. and Lutz, B. (1999) Expression of the cannabinoid receptor CB1 in distinct neuronal subpopulations in the adult mouse forebrain. Eur. J. Neurosci., 11: 4213–4225. Marsicano, G. and Lutz, B. (2006) Neuromodulatory functions of the endocannabinoid system. J. Endocrinol. Invest., 29: 27–46. Marsicano, G., Wotjak, C.T., Azad, S.C., Bisogno, T., Rammes, G., Cascio, M.G., Hermann, H., Tang, J., Hofmann, C., Zieglgansberger, W., Di Marzo, V. and Lutz, B. (2002) The endogenous cannabinoid system controls extinction of aversive memories. Nature, 418: 530–534. Matsuda, L.A., Bonner, T.I. and Lolait, S.J. (1993) Localization of cannabinoid receptor mRNA in rat brain. J. Comp. Neurol., 327: 535–550. Matsuda, L.A., Lolait, S.J., Brownstein, M.J., Young, A.C. and Bonner, T.I. (1990) Structure of a cannabinoid receptor and functional expression of the cloned cDNA. Nature, 346: 561–564. Melis, M., Perra, S., Muntoni, A.L., Pillolla, G., Lutz, B., Marsicano, G., Di Marzo, V., Gessa, G.L. and Pistis, M. (2004a) Prefrontal cortex stimulation induces 2-arachidonoyl-glycerol-mediated suppression of excitation in dopamine neurons. J. Neurosci., 24: 10707–10715. Melis, M., Pistis, M., Perra, S., Muntoni, A.L., Pillolla, G. and Gessa, G.L. (2004b) Endocannabinoids mediate presynaptic inhibition of glutamatergic transmission in rat ventral tegmental area dopamine neurons through activation of CB1 receptors. J. Neurosci., 24: 53–62. Monory, K., Massa, F., Egertova, M., Eder, M., Blaudzun, H., Westenbroek, R., Kelsch, W., Jacob, W., Marsch, R., Ekker, M., Long, J., Rubenstein, J.L., Goebbels, S., Nave, K.A., During, M., Klugmann, M., Wolfel, B., Dodt, H.U.,

Zieglgansberger, W., Wotjak, C.T., Mackie, K., Elphick, M.R., Marsicano, G. and Lutz, B. (2006) The endocannabinoid system controls key epileptogenic circuits in the hippocampus. Neuron, 51: 455–466. Morales, M. (2006) Cannabinoids and the central serotonergic system. In: Onaivi E.S., Sugiura T. and Di Marzo V. (Eds.), Endocannabinoids. CRC Press, Boca Raton, FL, pp. 249–260. Morales, M. and Backman, C. (2002) Coexistence of serotonin 3 (5-HT3) and CB1 cannabinoid receptors in interneurons of hippocampus and dentate gyrus. Hippocampus, 12: 756–764. Mukhtarov, M., Ragozzino, D. and Bregestovski, P. (2005) Dual Ca2+ modulation of glycinergic synaptic currents in rodent hypoglossal motoneurones. J. Physiol., 569: 817–831. Munro, S., Thomas, K.L. and bu-Shaar, M. (1993) Molecular characterization of a peripheral receptor for cannabinoids. Nature, 365: 61–65. Nagayama, T., Sinor, A.D., Simon, R.P., Chen, J., Graham, S.H., Jin, K. and Greenberg, D.A. (1999) Cannabinoids and neuroprotection in global and focal cerebral ischemia and in neuronal cultures. J. Neurosci., 19: 2987–2995. Nakatsuka, T., Chen, H.X., Roper, S.N. and Gu, J.G. (2003) Cannabinoid receptor-1 activation suppresses inhibitory synaptic activity in human dentate gyrus. Neuropharmacology, 45: 116–121. Niehaus, J.L., Wallis, K.T., Elphick, M.R. and Lewis, D.L., (2004) CRIP1B, a novel CB1 cannabinoid receptor interacting protein, interacts with CB1 at the distal C-terminal tail. Soc. Neurosci. Abstr., 623.19. Ohno-Shosaku, T., Hashimotodani, Y., Maejima, T. and Kano, M. (2005) Calcium signaling and synaptic modulation: regulation of endocannabinoid-mediated synaptic modulation by calcium. Cell Calcium, 38: 369–374. Ohno-Shosaku, T., Maejima, T. and Kano, M. (2001) Endogenous cannabinoids mediate retrograde signals from depolarized postsynaptic neurons to presynaptic terminals. Neuron, 29: 729–738. Ohno-Shosaku, T., Shosaku, J., Tsubokawa, H. and Kano, M. (2002a) Cooperative endocannabinoid production by neuronal depolarization and group I metabotropic glutamate receptor activation. Eur. J. Neurosci., 15: 953–961. Ohno-Shosaku, T., Tsubokawa, H., Mizushima, I., Yoneda, N., Zimmer, A. and Kano, M. (2002b) Presynaptic cannabinoid sensitivity is a major determinant of depolarizationinduced retrograde suppression at hippocampal synapses. J. Neurosci., 22: 3864–3872. Ong, W.Y. and Mackie, K. (1999) A light and electron microscopic study of the CB1 cannabinoid receptor in primate brain. Neuroscience, 92: 1177–1191. Oz, M., Ravindran, A., az-Ruiz, O., Zhang, L. and Morales, M. (2003) The endogenous cannabinoid anandamide inhibits alpha7 nicotinic acetylcholine receptor-mediated responses in Xenopus oocytes. J. Pharmacol. Exp. Ther., 306: 1003–1010. Oz, M., Zhang, L. and Morales, M. (2002) Endogenous cannabinoid, anandamide, acts as a noncompetitive inhibitor on 5-HT3 receptor-mediated responses in Xenopus oocytes. Synapse, 46: 150–156.

336 Pettit, D.A., Harrison, M.P., Olson, J.M., Spencer, R.F. and Cabral, G.A. (1998) Immunohistochemical localization of the neural cannabinoid receptor in rat brain. J. Neurosci. Res., 51: 391–402. Piomelli, D. (2003) The molecular logic of endocannabinoid signalling. Nat. Rev. Neurosci., 4: 873–884. Piomelli, D., Beltramo, M., Glasnapp, S., Lin, S.Y., Goutopoulos, A., Xie, X.Q. and Makriyannis, A. (1999) Structural determinants for recognition and translocation by the anandamide transporter. Proc. Natl. Acad. Sci. U.S.A., 96: 5802–5807. Pitler, T.A. and Alger, B.E. (1992) Postsynaptic spike firing reduces synaptic GABAA responses in hippocampal pyramidal cells. J. Neurosci., 12: 4122–4132. Ralevic, V., Kendall, D.A., Jerman, J.C., Middlemiss, D.N. and Smart, D. (2001) Cannabinoid activation of recombinant and endogenous vanilloid receptors. Eur. J. Pharmacol., 424: 211–219. Robbe, D., Kopf, M., Remaury, A., Bockaert, J. and Manzoni, O.J. (2002) Endogenous cannabinoids mediate long-term synaptic depression in the nucleus accumbens. Proc. Natl. Acad. Sci. U.S.A., 99: 8384–8388. Romero, J., Hillard, C.J., Calero, M. and Rabano, A. (2002) Fatty acid amide hydrolase localization in the human central nervous system: an immunohistochemical study. Brain Res. Mol. Brain Res., 100: 85–93. Roth, S.H. and Williams, P.J. (1979) The non-specific membrane binding properties of delta9-tetrahydrocannabinol and the effects of various solubilizers. J. Pharm. Pharmacol., 31: 224–230. Rueda, D., Navarro, B., Martinez-Serrano, A., Guzman, M. and Galve-Roperh, I. (2002) The endocannabinoid anandamide inhibits neuronal progenitor cell differentiation through attenuation of the Rap1/B-Raf/ERK pathway. J. Biol. Chem., 277: 46645–46650. Safo, P.K. and Regehr, W.G. (2005) Endocannabinoids control the induction of cerebellar LTD. Neuron, 48: 647–659. Schlicker, E. and Kathmann, M. (2001) Modulation of transmitter release via presynaptic cannabinoid receptors. Trends Pharmacol. Sci., 22: 565–572. Skaper, S.D., Buriani, A., Dal, T.R., Petrelli, L., Romanello, S., Facci, L. and Leon, A. (1996) The ALIAmide palmitoylethanolamide and cannabinoids, but not anandamide, are protective in a delayed postglutamate paradigm of excitotoxic death in cerebellar granule neurons. Proc. Natl. Acad. Sci. U.S.A., 93: 3984–3989. Stella, N., Schweitzer, P. and Piomelli, D. (1997) A second endogenous cannabinoid that modulates long-term potentiation. Nature, 388: 773–778. Sugiura, T., Kondo, S., Sukagawa, A., Nakane, S., Shinoda, A., Itoh, K., Yamashita, A. and Waku, K. (1995) 2-Arachidonoylglycerol: a possible endogenous cannabinoid receptor ligand in brain. Biochem. Biophys. Res. Commun., 215: 89–97. Sugiura, T., Oka, S., Ikeda, S. and Waku, K. (2006) Occurrence, biosynthesis, and metabolism of endocannabinoids. In: Onaivi E.S., Sugiura T. and Di Marzo V. (Eds.), Endocannabinoids. CRC Press, Boca Raton, FL, pp. 177–214.

Takahashi, K.A. and Castillo, P.E. (2006) The CB1 cannabinoid receptor mediates glutamatergic synaptic suppression in the hippocampus. Neuroscience, 139: 795–802. Thomas, B.F., Wei, X. and Martin, B.R. (1992) Characterization and autoradiographic localization of the cannabinoid binding site in rat brain using [3H]11-OH-delta 9-THCDMH. J. Pharmacol. Exp. Ther., 263: 1383–1390. Thomas, E.A., Cravatt, B.F., Danielson, P.E., Gilula, N.B. and Sutcliffe, J.G. (1997) Fatty acid amide hydrolase, the degradative enzyme for anandamide and oleamide, has selective distribution in neurons within the rat central nervous system. J. Neurosci. Res., 50: 1047–1052. Toth, A., Boczan, J., Kedei, N., Lizanecz, E., Bagi, Z., Papp, Z., Edes, I., Csiba, L. and Blumberg, P.M. (2005) Expression and distribution of vanilloid receptor 1 (TRPV1) in the adult rat brain. Brain Res. Mol. Brain Res., 135: 162–168. Trettel, J. and Levine, E.S. (2003) Endocannabinoids mediate rapid retrograde signaling at interneuron right-arrow pyramidal neuron synapses of the neocortex. J. Neurophysiol., 89: 2334–2338. Tsou, K., Brown, S., Sanudo-Pena, M.C., Mackie, K. and Walker, J.M. (1998a) Immunohistochemical distribution of cannabinoid CB1 receptors in the rat central nervous system. Neuroscience, 83: 393–411. Tsou, K., Mackie, K., Sanudo-Pena, M.C. and Walker, J.M. (1999) Cannabinoid CB1 receptors are localized primarily on cholecystokinin-containing GABAergic interneurons in the rat hippocampal formation. Neuroscience, 93: 969–975. Tsou, K., Nogueron, M.I., Muthian, S., Sanudo-Pena, M.C., Hillard, C.J., Deutsch, D.G. and Walker, J.M. (1998b) Fatty acid amide hydrolase is located preferentially in large neurons in the rat central nervous system as revealed by immunohistochemistry. Neurosci. Lett., 254: 137–140. Van der Stelt, M., Veldhuis, W.B., Bar, P.R., Veldink, G.A., Vliegenthart, J.F. and Nicolay, K. (2001a) Neuroprotection by delta9-tetrahydrocannabinol, the main active compound in marijuana, against ouabain-induced in vivo excitotoxicity. J. Neurosci., 21: 6475–6479. Van der Stelt, M., Veldhuis, W.B., van Haaften, G.W., Fezza, F., Bisogno, T., Bar, P.R., Veldink, G.A., Vliegenthart, J.F., Di, M.V. and Nicolay, K. (2001b) Exogenous anandamide protects rat brain against acute neuronal injury in vivo. J. Neurosci., 21: 8765–8771. Van der Stelt, M., Veldhuis, W.B., Maccarrone, M., Bar, P.R., Nicolay, K., Veldink, G.A., Di, M.V. and Vliegenthart, J.F. (2002) Acute neuronal injury, excitotoxicity, and the endocannabinoid system. Mol. Neurobiol., 26: 317–346. Van Sickle, M.D., Duncan, M., Kingsley, P.J., Mouihate, A., Urbani, P., Mackie, K., Stella, N., Makriyannis, A., Piomelli, D., Davison, J.S., Marnett, L.J., Di Marzo, V., Pittman, Q.J., Patel, K.D. and Sharkey, K.A. (2005) Identification and functional characterization of brainstem cannabinoid CB2 receptors. Science, 310: 329–332. Varma, N., Carlson, G.C., Ledent, C. and Alger, B.E. (2001) Metabotropic glutamate receptors drive the endocannabinoid system in hippocampus. J. Neurosci., 21: RC188.

337 Wallace, M.J., Blair, R.E., Falenski, K.W., Martin, B.R. and Delorenzo, R.J. (2003) The endogenous cannabinoid system regulates seizure frequency and duration in a model of temporal lobe epilepsy. J. Pharmacol. Exp. Ther., 307: 129–137. Wallace, M.J., Martin, B.R. and Delorenzo, R.J. (2002) Evidence for a physiological role of endocannabinoids in the modulation of seizure threshold and severity. Eur. J. Pharmacol., 452: 295–301. Wallace, M.J., Wiley, J.L., Martin, B.R. and Delorenzo, R.J. (2001) Assessment of the role of CB1 receptors in cannabinoid anticonvulsant effects. Eur. J. Pharmacol., 428: 51–57. Wilkison, D.M. and Pontzer, N.J. (1987) The actions of THC on the intact hippocampus: a comparison of dentate and CA1 responses. Brain Res. Bull., 19: 63–67. Wilson, R.I., Kunos, G. and Nicoll, R.A. (2001) Presynaptic specificity of endocannabinoid signaling in the hippocampus. Neuron, 31: 453–462. Wilson, R.I. and Nicoll, R.A. (2001) Endogenous cannabinoids mediate retrograde signalling at hippocampal synapses. Nature, 410: 588–592. Wilson, R.I. and Nicoll, R.A. (2002) Endocannabinoid signaling in the brain. Science, 296: 678–682.

Yanovsky, Y., Mades, S. and Misgeld, U. (2003) Retrograde signaling changes the venue of postsynaptic inhibition in rat substantia nigra. Neuroscience, 122: 317–328. Yoshida, T., Fukaya, M., Uchigashima, M., Miura, E., Kamiya, H., Kano, M. and Watanabe, M. (2006) Localization of diacylglycerol lipase-{alpha} around postsynaptic spine suggests close proximity between production site of an endocannabinoid, 2-arachidonoyl-glycerol, and presynaptic cannabinoid CB1 receptor. J. Neurosci., 26: 4740–4751. Yoshida, T., Hashimoto, K., Zimmer, A., Maejima, T., Araishi, K. and Kano, M. (2002) The cannabinoid CB1 receptor mediates retrograde signals for depolarization-induced suppression of inhibition in cerebellar Purkinje cells. J. Neurosci., 22: 1690–1697. Zhu, P.J. and Lovinger, D.M. (2005) Retrograde endocannabinoid signaling in a postsynaptic neuron/synaptic bouton preparation from basolateral amygdala. J. Neurosci., 25: 6199–6207. Zygmunt, P.M., Petersson, J., Andersson, D.A., Chuang, H., Sorgard, M., Di Marzo, V., Julius, D. and Hogestatt, E.D. (1999) Vanilloid receptors on sensory nerves mediate the vasodilator action of anandamide. Nature, 400: 452–457.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 20

Pro-inflammatory cytokines and their effects in the dentate gyrus Mark Pickering and John J. O’Connor UCD School of Biomolecular and Biomedical Science, Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin 4, Ireland

Abstract: The older notion of a central nervous system existing in essential isolation from the immune system has changed dramatically in recent years as the body of evidence relating to the interactions between these two systems has grown. Here we address the role of a particular subset of immune modulatory molecules, the pro-inflammatory cytokines, in regulating neuronal function and viability in the dentate gyrus of the hippocampus. These inflammatory mediators are known to be elevated in many neuropathological conditions, such as Alzheimer’s disease, Parkinson’s disease and ischaemic injury that follows stroke. Pro-inflammatory cytokines, such as tumour necrosis factor-a (TNF-a), interleukin 1-b (IL-1b) and interleukin 18 (IL-18), have been shown to regulate neurotoxicity; although, due to the complexity of the cytokine action in neurons and glia, the effect may be either facilitatory or protective, depending on the circumstances. As well as their role in neurotoxicity and neuroprotection, the pro-inflammatory cytokines have also been shown to be potent regulators of synaptic function. In particular, TNF-a, IL-1b and IL-18 have all been shown to inhibit long-term potentiation, a form of neuronal plasticity widely believed to underlie learning and memory, both in the early p38 mitogen activated protein kinase-dependant phase and the later protein synthesis-dependant phase. In this article we address the mechanisms underlying these cytokine effects in the dentate gyrus of the hippocampus. Keywords: TNF-a; IL-1b; IL-18; long-term potentiation; dentate gyrus; mGluR5 Chun, 2001). Chief amongst these are the cytokines — multifunctional proteins that play crucial roles in cellular communication and activation. Cytokines have been classified as pro-inflammatory or anti-inflammatory depending on the balance of their effects on the immune system (Mosmann et al., 1986). Cytokines may have an indirect modulatory effect on the nervous system via their effects on the hypothalamic-pituitaryadrenal axis (Besedovsky et al., 1991). However, here we will address the direct effect roles of the pro-inflammatory cytokines, in particular tumour necrosis factor-a (TNF-a), interleukin-1b (IL-1b)

Introduction The older notion of two ‘‘super-systems’’, nervous and immune, existing in relative isolation from each other due largely to the blood-brain barrier, has given way in recent years to a new consensus, as it became apparent that many immune molecules may be used by the nervous system in intercellular communication (Boulanger et al., 2001;

Corresponding author. Tel.: +353 1 716 6765; Fax: +353 1 716 7417; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63020-9

339

340

and interleukin-18 (IL-18) on the CNS in general, and the dentate gyrus in particular. Distribution of pro-inflammatory cytokines in the CNS Direct action of the pro-inflammatory cytokines TNF-a and IL-1b on the CNS has been known for some time (Plata-Salaman et al., 1988). Elevated CNS expression of various pro-inflammatory cytokines have been noted in many neuropathological situations, both chronic, such as Alzheimer’s disease (Cacquevel et al., 2004) and multiple sclerosis (Merrill, 1992), and acute, such as ischemia and stroke (Liu et al., 1994; Klein et al., 2000; Yu and Lau, 2000), and infection (Waage et al., 1989). Additionally, pro-inflammatory cytokines are constitutively expressed in unperturbed cells of the CNS (Benveniste and Benos, 1995; Yu and Lau, 2000). While pro-inflammatory cytokine binding sites are found throughout the hippocampus, in addition to other brain regions, there is some evidence that some of the functions of these cytokines are distinct in different hippocampal regions. For example, it has been demonstrated that local injection of IL-1 receptor antagonist (IL-1ra) into the dentate gyrus and CA3 regions of the hippocampus leads to a reduction in the febrile response of rats to peripheral injection of lipopolysaccharide (Cartmell et al., 1999), whereas IL-1ra injection has no effect in the CA1 region of the hippocampus. Here, we will address some of the many roles played by TNF-a and the pro-inflammatory IL1type cytokines in the dentate gyrus, highlighting differences between the dentate gyrus and the pyramidal cell regions of the hippocampus where appropriate. Action of TNF-a in the dentate gyrus The pro-inflammatory cytokine TNF-a is a 17-kDa peptide and forms multimers which are active in binding TNF receptors (TNFR) that are constitutively expressed on both neurons and glia in the central nervous system (Benveniste and Benos, 1995). TNF-a can be synthesized and released in the

brain by astrocytes, microglial and some neurons (Lieberman et al., 1989; Morganti-Kossman et al., 1997; Chung and Benveniste, 1990). Under various pathological conditions, such as trauma, ischemia and inflammatory diseases, the expression and release of TNF-a is rapidly increased, in some cases as early as 1 h after the brain insult and well before neuronal death (Liu et al., 1994; Wang et al., 1994; Allan and Rothwell, 2001) Two different TNFR (p55 and p75) have been identified (Beutler and Van Huffel, 1994a, b; Wajant and Scheurich, 2001) and shown to mediate differential cellular responses using distinct pathways (Kinouchi et al., 1991; Tartaglia et al., 1991). For instance, the signal transduction pathway used by the p55 TNFR results in the activation of the transcription factor nuclear factor kappaB (NF-kB) (Kolesnick and Golde, 1994; Goodman and Mattson, 1996; Mattson et al., 1997a, b) (see Fig. 1). TNF-a activates the NF-kB family of transcription factors, which are ubiquitously expressed and are pivotal in controlling diverse cellular processes, including immune responses, cell proliferation and differentiation (Israel, 2000; Silverman and Maniatis, 2001; Ghosh and Karin, 2002; Li and Verma, 2002). Lack of the p50 subunit of NF-kB increases the vulnerability of hippocampal neurons to excitotoxicity (Yu et al., 1999). Increasingly, it has become evident that NF-kB also plays important roles in the CNS (O’Neill and Kaltschmidt, 1997) and recently Sheridan et al. (2006) have shown that TNF-a can selectively upregulate NF-kB expression in hippocampal sub-neuronal populations. For example in the dentate gyrus and CA3 regions there were clear subpopulations with respect to activation of NF-kB. In the dentate gyrus there is a clear distinction between the outer 4/5th of the stratum granulosum, the cells of which expressed high nuclear NF-kB activity and the inner 1–2 cell layers that expressed much lower levels of the active form of this transcription factor. One of the most active areas of investigation regarding TNF-a is in relation to its role in neurotoxicity. Although TNF-a was originally named for its degeneration-inducing action in some types of tumor cells, our current understanding of TNF-a is such that it cannot be simply considered

341

Fig. 1. Putative TNF-a signalling pathway in the dentate gyrus. Downstream signalling by TNF-a involves recruitment by the TNF-a receptor of the death domain containing adaptor protein, TNF receptor-associated death domain (TRADD) and the TNF receptorassociated factor 2 (TRAF-2) which then recruits the receptor interacting protein (RIP) eventually leading to activation and translocation of NF-kB to the nucleus.

as an inducer or facilitator of cell death. On the one hand, TNF-a has been shown to induce cell death in septo-hippocampal cultures. Zhao et al. (2001) detected alpha-spectrin fragments in these septohippocampal cultures treated with TNF-a and found elevated 120 kDa fragments, indicative of caspase-3 activity, but not 145-kDa fragments, indicative of calpain activity. Unlike calpain, which is associated with both necrotic and apoptotic cell death, caspase-3 is exclusively characteristic of apoptosis-like cell death (Armstrong et al., 1996; Wang et al., 1996; Nath et al., 1998). On the other hand, a number of investigators have presented evidence that, under different conditions, TNF-a may play a protective role against

neuronal cell death. For example TNF-a treatment can protect against focal cerebral ischemia (Nawashiro et al., 1997). Also, while it has been shown that TNF-a mediates damage to myelin and oligodendrocytes (Selmaj and Raine, 1998), Garcia et al. (1992) found that it was not toxic to rat-cultured CNS neurons, including those from the spinal cord, ventral mesencephalon, cerebellum, septum and striatum, in vitro. TNF-a under in vitro conditions may protect neurons against metabolic, excitotoxic or oxidative insults by promoting maintenance of intracellular Ca2+ homeostasis, suppression of reactive oxygen species (Cheng et al., 1994), and by activation of transcription factor NF-kB (Barger et al., 1995).

342

Mice genetically deficient in TNF-receptor p55 (R1) or both R1 and p75 (R2), also show exacerbated neuronal damage compared to wild type controls following middle cerebral artery occlusion (Bruce et al., 1996; Gary et al., 1998). Some of the reasons for the manifest complexity of the relationship between TNF-a and neuronal cell death may be due to differences between the role of TNF-a in the early, acute post-injury phase, where it appears to be largely detrimental, and late post-injury phase, where it may be protective because it activates cellular repair mechanisms. It must also be remembered that there is some evidence that the different TNF receptors may have differing roles in TNF-a toxicity. It has been known for some time that the p55 TNF-R1 plays a major role in the anti-bacterial response and sensitivity to septic shock (Pfeffer et al., 1993), but the role of TNF-R2 was unclear until the membrane form of TNF receptor was recognized as the physiological activator of this TNF receptor (Grell et al., 1995). The more recent development of TNF-R1 and TNF-R2 knockout mice has allowed opportunities to tease apart the roles of the separate receptor pathways. While looking at cell death in retinal ischemia-reperfusion experiments in these knockout mice, Fontaine et al. (2002) demonstrated opposing actions of the two TNF receptors, with TNF-R1 aggravating neuronal damage and TNF-R2 promoting neuroprotection via an Akt/PKB signal pathway. The work of Fontaine et al. (2002) is in agreement with Kassiotis and Kollias (2001), who also showed a dual role for TNF-a in experimental autoimmune encephalomyelitis (EAE), a mouse model for multiple sclerosis. Indeed, this differential role for the two TNF receptors may be of key importance in understanding TNF-a systems. By ‘‘fine tuning’’ the balance between the two receptors, the exact role of TNF-a in cell damage can be controlled from species to species and tissue to tissue. This makes the TNF-a system very evolutionary versatile and advantageous. Another aspect of the complexity may relate to the role of TNF-a in the interactions between the neurons and glia. Increasingly, glial cells are no longer seen as passive supporters of neurons, but as active participants in information processing

(for review, see Haydon, 2001; Volterra and Steinhauser, 2004); neuronal–glial interactions are seen as important to many neural processes and phenomena. Of particular interest in the context of this review is the relationship between neurons, glial cells and TNF-a in glutamate excitotoxicity.

Role of TNF-a in excitotoxicity Excitotoxicity in general is linked to excessive glutamate activation of receptors, particularly the N-methyl-D-aspartate (NMDA) receptor. Cell death resulting from excessive levels of glutamate and overstimulation of glutamate receptors is known to be caused by impaired uptake of glutamate by glial cells (Choi, 1988). In vivo, it has been shown that mice lacking expression of the excitatory amino-acid transporter, EAAT2/GLT-1, develop epilepsy and increased susceptibility to acute injury as a result of excessive extracellular glutamate (Tanaka et al., 1997). The expression of this transporter has been shown recently to be both positively and negatively regulated by NF-kB (Sitcheran et al., 2005). This study showed that the increased binding of NF-kB to the EAAT2 promoter in H4 astroglioma cells was regulated by epidermal growth factor (EGF), but decreased expression was caused by TNF-a inducing the classical IkappaB degradation pathway to trigger NF-kB nuclear translocation and DNA binding to repress EAAT2 expression. In this situation, the presence of elevated TNF-a concentrations leads to elevated extracellular glutamate concentration, thereby increasing the risk of glutamate excitotoxicity. Indeed, Zou and Crews (2005) also showed that in rat organotypic hippocampal slice cultures, which possess a cytoarchitecture comparable to that in vivo, TNF-a increased glutamate neurotoxicity. They also demonstrated that the effect was mediated by NMDA and not AMPA receptors. In addition to this, TNF-a and glutamate have also been implicated together recently in b-amyloid-induced microglia-related cell death. Abundant activated microglia are prominent in the brains of patients with Alzheimer’s disease (Griffin et al., 1989), and are associated with beta

343

amyloid plaques (Griffin et al., 1995; Frautschy et al., 1998). It has been proposed that inefficient phagocytosis of peptide by microglia could lead to hyperactivation of cells and release of inflammatory mediators and neurotoxic factors, thereby contributing to neurodegenerative processes (Akiyama et al., 2000). It is widely believed that the microglia play a direct role in the neuronal death in Alzheimer’s disease. Floden et al. (2005) showed that applying media from b-amyloidstimulated microglial cultures to neurons led to neuronal cell death that was dependant on the synergistic co-activation of TNF-a and NMDA receptors; the NMDA receptor antagonists memantine and 2-amino-5-phosphopetanoic acid as well as soluble TNF-R1 applied to the neurons protected them from cell death. Interestingly, blockade of either the TNF-R1 or NMDA receptors alone was insufficient to induce neuronal cell death. There is, however, evidence that other glutamate receptors are involved in the relationship between neuronal cell death, glial cells and TNF-a. Taylor et al. (2005) examined the effect of metabotropic glutamate (mGlu) receptor stimulation on TNF-a release. They found that stimulating rat primarycultured microglial mGlus for 24 h induced microglial activation. This in turn induced caspase-3 activation in cerebellar granule neurons in culture, both those treated with microglial-conditioned media as well as in neuronal/microglial co-cultures. This microglial neurotoxicity was mediated by TNF-a released by the microglia via neuronal TNF-R1 and caspase-3 activation. Importantly, it was the specific group II mGluR agonist 2S,20 R,30 2-(20 ,30 -dicarboxy-cyclopropyl)glycine, which led to the release of TNF-a and consequent toxicity; N-acetyl-L-aspartyl-L-glutamate, a specific mGlu3 agonist, did not induce microglial activation or neurotoxicity. TNF-a was only neurotoxic in the presence of microglia or microglial-conditioned medium; possibly this was due to the presence of microglial-derived Fas ligand. However, TNF-a neurotoxicity was prevented when the neurons were exposed to conditioned medium from microglia stimulated by the specific group III agonist L2-amino-4-phosphono-butyric acid. Glial–glial interactions also influence the effects of TNF-a. Bezzi et al. (2001) showed that astrocyte

glutamate release induced by activation of the chemokine receptor CXCR4 is accompanied by release of TNF-a, and that the TNF-a release is dramatically enhanced by microglia. In light of this evidence of glial–glial and glial–neuronal interactions, it is possible to see how a cascading neuronal–glial interaction could occur, leading to cell death. Activation of TNF receptors on astrocytes leads to increased extracellular glutamate, which may lead to neurotoxicity in itself, while also activating mGlu2 on microglia, which releases more TNF-a, in addition to other pro-inflammatory cytokines. Also, it has been known for some time that the neurotoxic effects of TNF-a may be mediated by activation of glutamate AMPA receptor subtypes (Gelbard et al., 1993).

TNF-a and synaptic transmission The subject of TNF-a in glial–neuronal interactions also emerges when looking at synaptic transmission. Beattie et al. (2002) showed that glial TNF-a causes an increase in surface expression of neuronal AMPA receptors, which would increase synaptic efficacy, and that the removal of endogenous TNF-a induced a decrease in AMPA receptor expression at the cell surface. Further evidence for this relationship between TNF-a and AMPA receptor density came from Stellwagen and Malenka (2006), who found the same elevated AMPA receptor density in the presence of elevated TNF-a. This TNF-a-induced AMPA receptor exocytosis has recently been shown to be mediated by activation of TNF-R1 receptors through a PI3 kinase-dependent pathway. (Stellwagen et al., 2005). The newly expressed AMPA receptors were shown to have lower stoichiometric amounts of GluR2, making the receptors permeable to Ca2+ ions. Additionally, this study also showed a surprising concurrent endocytosis of inhibitory GABA receptors. TNF-a and synaptic plasticity Changes in neuronal excitability brought about by TNF-a have important implications for synaptic

344

plasticity (Carroll et al., 2001). Indeed, TNF-a is known to act as a regulator of synaptic plasticity in the dentate gyrus, in addition to playing a role in apoptotic events. As previously mentioned, elevated levels of TNF-a have been observed in several neuropathological states that are associated with learning and memory deficits, such as Alzheimer’s disease, leading to the search for a possible role in plasticity. To this end, much work has been done concerning the role of TNF-a in the hippocampus. TNF-a has been shown to regulate the development of the hippocampus; TNF-a alpha knockout mice demonstrate decreased arbourization of the apical dendrites of the CA1 and CA3 regions while at the same time causing accelerated development in the dentate gyrus (Golan et al., 2004), probably via activation of the TNF-R2 receptor, which, as mentioned earlier, does not lead to caspase-3 activation, but is known to transduce the trophic effect of TNF-a by leading to the activation of several transcription factors (Yang et al., 2002). Two forms of synaptic plasticity in the hippocampus, long-term potentiation (LTP) and longterm depression (LTD) involve glutamate receptor activation and increased intracellular Ca2+ levels, with LTP induction dependant on the activation of Ca2+/calmodulin kinase II, PKC and PKA, and LTD on activation of serine/threonine phosphatases 1, 2A and 2B (Mayford et al., 1995; Coussens and Teyler, 1996; Silva et al., 1997). LTP is a longlasting increase in synaptic efficacy, which is thought to be an important underlying mechanism of learning and memory formation (Bliss and Collindridge, 1993). TNF-a has in fact been shown to inhibit LTP in the CA1 region, as well as the dentate gyrus of the rat hippocampus, when pathophysiological levels of TNF-a are used (Tancredi et al., 1992; Cunningham et al., 1996; Butler et al., 2004; Cumiskey et al., 2004, 2007). There are, however, differences in the mechanisms by which TNF-a inhibits early phase (i.e., o120 min post-tetanic stimulation) and late phase (>120 min) LTP. Butler et al. (2002) investigated these differences in the dentate gyrus region of rat hippocampal slices. They found, using immunohistochemical techniques, that TNF-a activated the p38 MAP

kinase. Inhibition of p38 MAP kinase with SB 203580 blocked the impairment of the early phase of LTP by TNF-a, but the late phase was unaffected, suggesting that the inhibition of LTP by TNF-a is biphasic in character. If the early stage inhibition of LTP is mediated by a p38 MAP kinase pathway, it may be that the late phase inhibition of LTP by TNF-a is regulated by altered protein synthesis; not an unlikely possibility given the aforementioned regulation of NF-kB by TNFa. As well as a role for the p38 MAPK, (Cumiskey et al., 2004; Cumiskey and O’ Connor, 2006) also provided evidence for a role of metabotropic glutamate receptors in the TNF-a-mediated inhibition of LTP in the dentate gyrus (see Fig. 2). In addition to the effect of TNF-a on LTP, it has been shown that TNF receptor knockout mice demonstrate an impairment of LTD in the CA1 region of the hippocampus (Albensi and Mattson, 2000). The effect is mimicked by NF-kB decoy DNA, which implicates the TNF-a/NF-kB signalling pathway in LTD. Additionally, Wang et al. (2006) showed that metabotropic glutamate receptor mediated LTD in the dentate, induced with (RS)-3,5-dihydroxyphenylglycine (DHPG) could not be induced in TNF-R1 knockout mice. Interestingly, the findings relating to the effect of TNF-a on synaptic plasticity seem to have some behavioural correlates in vivo. TNF-a knockout mice showed increased performance in spatial memory and learning as measured in the Morris water maze task when compared to wild type animals (Golan et al., 2004). Conversely, Aloe et al. (1999) demonstrated a significant impairment in spatial learning in two lines of mice that over-expressed human recombinant TNF-a, also using the same water maze task. However, these results must be interpreted in the context of the study showing developmental differences mentioned previously; the ‘‘substrate’’ of learning and memory is not necessarily the same in the knockout and wild type mice. Recent work has also shown that TNF-a plays a crucial role in the development of synaptic scaling; a form of homeostatic synaptic plasticity that scales synaptic strengths to compensate for prolonged changes in activity. This form of plasticity is essential to stabilize activity across networks (Turrigiano and Nelson, 2004), as it regulates

345

Fig. 2. Schematic diagram of the role of the mGluR5 and NMDA receptor in mediating the TNF-a effects on LTP. TNF-a inhibits the early phase of LTP by activation of TNF-R1 (p55) and is dependent on p38 activation. Activation of the G-protein-linked mGluRs leads to inositol-1,4,5-trisphosphate (IP3) receptor-mediated Ca2+ release via phospholipase C (PLC). TNF-a may also alter Ca2+ concentrations. Combined activation of TNF-R and mGluR5 may lead to the impairment of LTP. TNF-a has negative effects on the NMDA receptor EPSPs.

synaptic ‘‘gain’’, thus preventing excessive quietening or activity across the network. Stellwagen and Malenka (2006) showed, using TNF receptor knockout mice, that signalling from the constitutive expression of TNF-a was not necessary for the induction of LTP or LTD, but was necessary for the development of synaptic scaling in the longer term. Additionally, they showed that the source of this TNF-a was not neuronal but was glial. This finding is not surprising when one considers that the glia are in a unique position to sense the overall level of local network activity (presumably using the index of glutamate overflow from synapses) and adjust the activity accordingly. While Stellwagen and Malenka (2006) did not directly implicate TNF-a in synaptic scaling in the dentate gyrus, the hippocampal slice component of their study instead focusing on the CA1 region in intact hippocampal slices, the evidence from dissociated hippocampal cultures (presumably containing a cohort of granule neurons) would seem to suggest that TNF-a may also play a role in the development of synaptic scaling in the dentate gyrus.

Action of IL-1b in the dentate gyrus IL-1b receptors As described for TNF-a, IL-1 receptors have also been shown to be present in many brain regions, with high levels in the hippocampus and hypothalamus (Ban et al., 1991). There are a number of receptors to which IL-1 can bind (O’Neill, 1997). The principal mediator of IL-1 signalling in the CNS is believed to be the type I IL-1 receptor (IL-1R1), which initiates intracellular events upon IL-1b binding (O’Neill, 1996), although the type II IL-1 receptor (IL-1R2) is also present, which although having a similar affinity for IL-1b to the type I receptor, has no ability to initiate signal transduction events, and so may serve as a ‘‘decoy’’ receptor, attenuating IL-1b-induced signalling. Other receptors are the IL-1 receptor accessory protein (IL-RAcp) that partakes in IL-1R1 signalling and the IL-1 receptor-related protein (IL-1Rcp), which is the receptor for IL-18, a member of the IL-1 family (Okamura et al., 1998), and will be discussed in detail later.

346

The presence of IL-1 receptors in brain was shown in 1990 by Takao et al. using 125I-labelled IL-1a (both IL-1a and IL-1b can activate IL-1 receptors) binding to crude membrane preparations of mouse hippocampus. Their autoradiographic localization studies revealed very low densities of [1251]-IL-1a binding sites throughout the brain, with highest densities present in the molecular and granular layers of the dentate gyrus of the hippocampus. This finding was later confirmed by further binding studies (Haour et al., 1990; Ban et al., 1991) and then by affinity cross-linking of IL-1a (Parnet et al., 1994). Further characterization of the distribution of IL-1 receptors in brain utilized antisense cRNA against the type I IL-1 receptor in mouse brain (Cunningham et al., 1992), and RT-PCR (Parnet et al., 1994). These studies demonstrate the presence of both type I and type II IL-1 receptors in mouse hippocampus, with particularly high expression in the cell-dense granule cell layer of the dentate gyrus. Much like TNF-R2 signalling, IL-1 signalling leads to a change in gene expression — up to 90 genes have been shown to be affected by IL-1b (O’Neill and Green, 1998). These changes include upregulation of mRNA transcription for a number of cytokines, growth factors, adhesion molecules and acute-phase proteins. Again, similar to TNFa, IL-1b regulates gene transcription is by activation of the transcription factor NF-kB. IL-1 binds to IL-1R1, which is complexed with the IL-1 Racp. This in turn leads to recruitment of the IL-1 receptor-associated kinase (IRAK; O’Neill and Green, 1998). IRAK then associates with TRAF6 (tumour necrosis factor (TNF)-receptor-associated factor), causing activation of IcB kinase (IKK). IKK activation leads to phosphorylation of IKB, which marks it for ubiquitination and subsequent degradation (O’Neill, 1997; O’Neill and Green, 1998). The p50-p65 heterodimer, now free of 1cB, can translocate to the nucleus, bind to its nuclear KB site and activate gene transcription, leading to expression of proteins such as IL-6, cycloxygenase2, inducible-nitric oxide synthase, IL-1b, TNF-a and interferons (O’Neill and Kaltschmidt, 1997; see Fig. 3). One family of protein kinases that have been very strongly implicated in IL-1 signalling are the

MAP kinases. The three MAP kinase cascades: the p42/44 MAP kinases (also known as extracellular regulated kinases — ERKs); the p38 MAP kinases (also known as reactivating kinase — RK) and jun kinases (JNK; also known as p54 or stressactivated protein kinases — SAPKs) have all been implicated in IL-1 signalling.

IL-1b and synaptic plasticity IL-1b has been found to exert a number of effects in the CNS, and has been implicated in a number of processes such as centrally-mediated fever, slow wave sleep patterns, appetite suppression and neurodegeneration (for reviews, see Rothwell and Hopkins, 1995; Rothwell, 1999). Oitzl et al. (1993) showed that IL-1b can disrupt the acquisition of spatial learning in the Morris water maze via a mechanism which seemed to be independent of IL-1b’s pyrogenic properties. More recently, it was shown that IL-1R knockout mice exhibited deficits in a number of learning paradigms, as well as in LTP, suggesting that there is a physiological role for IL-1b in learning and memory, and possibly the induction of synaptic plasticity (Avital et al., 2003); although, as with the TNFR knockout studies, the evidence is confounded by possible effects of cytokine receptor deficits on development. The first reported action of IL-1b on synaptic plasticity was inhibition of NMDA receptor-independent LTP in the mouse mossy-fibre-CA3 pathway in vitro (Katuski et al., 1990). Concentrations as low as 50 pg/ml were found to produce a significant inhibition of LTP, and this inhibitory effect was blocked by Lys-D-Pro-Thr, a tripeptide analogue of IL-1b. No effects on baseline transmission were observed. In the Schaeffer-collateral-CA1 pathway, Bellinger et al. (1993) reported an inhibition by IL-1b of the potentiation of both the field EPSP and the population spike induced by high-frequency stimulation. IL-1b treatment was also reported to cause a transient decrease in population spike amplitude and a persistent decrease in EPSP slope. Bellinger et al. (1993) reported preliminary results that suggested that treatment of slices with the IL-1ra led to the induction of a larger LTP than control conditions. A later study

347

Fig. 3. Putative IL-1b signalling pathway in the dentate gyrus. In IL-1 signalling, the signal-mediating IL-1 receptor type I (IL-1R1) forms a heterodimer with a second molecule the IL-1 receptor accessory protein (IL-1RAcP). Binding of IL-1b to this receptor complex leads to activation of the transcription factor NFB through different signalling molecules. DNA-binding motifs for NFB are found in the promotors and enhancers of many genes that are known to be activated upon inflammation.

of the MPP-dentate granule cell pathway showed an inhibition of the induction of LTP by treatment with IL-1b, and this effect was antagonized by coapplication of IL-1ra, and again, as in the CA3 region, IL-1b was found to have no effect on baseline transmission in the dentate gyrus (Cunningham et al., 1996). This IL-1b-induced inhibition of LTP was accompanied by a decrease in Ca2+ influx into slices compared with tetanized slices, suggesting that IL-1b inhibits LTP in the dentate gyrus by inhibiting the Ca2+ influx necessary for LTP induction. Schneider et al. (1998) showed that in the CA1 region, expression of IL-1b transcripts was significantly upregulated 1 h following tetanization both in vitro and in vivo, which may explain why Bellinger et al. (1993) found that treatment of slices solely with IL-1ra produces a larger potentiation of EPSPs than in untreated slices. Along with the increased gene expression of IL-1b during LTP,

Schneider et al. (1998) and Coogan et al. (1999) have shown that the maintenance phase of LTP can be blocked by administration of IL-1ra. This suggests a requirement for IL-1b in the sustained expression of LTP, i.e., the endogenous IL-1b production by tetanic stimulation is required for sustained expression of LTP, while, somewhat paradoxically, the presence of elevated IL-1b can inhibit the initial induction of LTP. It seems that it is the sequence and timing of IL-1b changes that is important in regulating its effect on LTP. In the dentate gyrus, the IL-1b-induced inhibition of LTP in slices was found to be concomitant with a decrease in Ca2+ influx following tetanic stimulation, either pre- or postsynaptically, as judged from synaptosomal preparations (Cunningham et al., 1996). Further investigations into the actions of IL-1b in the dentate gyrus showed that treatment of slices with IL-1b caused a large depression of the pharmacologically isolated NMDA

348

receptor EPSP (NMDA-EPSP; Coogan and O’Connor, 1997). These effects were thought to be specific for IL-1R1, because the depression was fully antagonized by IL-1ra. This might explain the decrease in Ca2+ influx, possibly through the NMDA receptor channel, and so may serve to inhibit LTP induction by preventing sufficient Ca2+ influx via the NMDA receptor to trigger the intracellular processes that underlie the expression of LTP (Bliss and Collindridge, 1993). The fact that IL-1b seems to have an effect on NMDA receptor-mediated EPSPs suggests a postsynaptic locus of action for IL-1b in the dentate gyrus. The inhibitory effects of IL-1b have not been restricted to tetanically induced LTP since it has been shown that IL-1b can inhibit a tetraethylammonium-induced synaptic potentiation in the rat dentate gyrus (Coogan and O’Connor, 1999). IL-1b has also been shown to inhibit the release of glutamate from synaptosomes prepared from hippocampi of young rats, via a mechanism that seemed to be coupled to phospholipase A2(Murray et al., 1997). However, IL-1b was not observed to have any effect on glutamate release from synaptosomes and hippocampi from aged animals (Murray et al., 1997; Lynch, 1998b) The inhibition of LTP by IL-1b in vivo has been proposed to be mediated by a decrease in glutamate release and also via an induced increase in the formation of reactive oxygen radicals, leading to an increase in lipid peroxidation and a decrease in arachidonic acid levels (Murray and Lynch, 1998). As arachidonic acid has been shown to act as a retrograde (postsynaptic-to-presynaptic) messenger during LTP induction (Medina and Izquierdo, 1995), this has been suggested as a mechanism of IL-1b-induced inhibition of LTP in the dentate gyrus. As discussed above, the mitogen-activated protein (MAP) kinases have been implicated in the induction of LTP. PD98059, a specific inhibitor of MAP kinase kinase (MEK) and thus, p42/44 MAP kinase activation (Pang et al., 1995), has been shown to block induction of LTP in the area CA1 as well as the dentate gyrus of the hippocampus (English and Sweatt, 1996; Coogan et al., 1999). When these effects were further investigated to try to elucidate the intracellular pathway through which IL-1b exerts its action on LTP, it was found that preincubation of slices with the p38 MAP kinase inhibitor SB203580 inhibited the effects of IL-1b on

both LTP induction and NMDA-EPSPs in the dentate gyrus (Coogan et al., 1999). SB203580 did cause a small but significant increase in baseline transmission. It is possible that there may be a high constitutive expression of p38 MAP kinase activity in the hippocampus, which may exert an inhibitory effect on synaptic transmission, and its inhibition may lead to the observed increase in synaptic transmission. When the effects of a lower dose of PD98059 (10 mM, a dose that does not inhibit LTP) were examined on the IL-1b inhibition of LTP, it was found that it did not antagonize both the inhibition of LTP or the inhibition of the isolated NMDA receptor-mediated EPSP (Coogan et al., 1999). In addition to p38 MAPK, a role for cjun n-terminal kinase in the inhibition of LTP by IL-1b and long-term depression has also been demonstrated (Curran et al., 2003). A role for MAP kinases in LTD has also been shown in the CA3–CA1 synapse (Bolshakov et al., 2000). IL-1b and Ca2+ channels While, as previously mentioned, IL-1 receptor activation leads to altered gene expression, since most of the inhibitory effects of IL-1b in the hippocampus have been reported following acute treatment of slices, it seems unlikely that gene transcription plays a role in the observed effects. Therefore it seems more likely that IL-1b mediates its effects by activation of cytoplasmic cascades with cytoplasmic/ membrane bound substrates. The mechanism by which IL-1b exerts its effects on hippocampal LTP remains controversial, although a number of possibilities have been raised. In dissociated guinea-pig CA1 pyramidal neurones it has been shown that IL1b rapidly induces a depression of inward, voltagedependent Ca2+ currents, (both peak and late current) in a dose-dependent fashion from 0.197 to 3.1 ng/ml. This could be reversed by pretreatment with IL-1ra (Plata-Salaman and Ffrench-Mullen, 1992), indicating a specific effect on IL-1R1. Further investigation of this effect showed it to be inhibited by a non-hydrolysable GTP analogue and pertussis toxin, and also by protein kinase C inhibitors H-7 and staurosporine (Plata-Salaman and Ffrench-Mullen, 1994), suggesting that the effect on

349

Ca2+ channels is mediated by G-proteins and protein kinase C, although H-7 and staurosporine can also inhibit other kinases, such as protein kinase A and Ca2+-calmodulin kinase (Hikada et al., 1991). IL-1ra was found to completely block (excess of 25) the PKC-dependent IL-1b induced decrease in Ca2+ currents. It is therefore likely that IL-1b interacts with the IL-1R1 membrane site. It must be noted, however, that low concentrations of IL-1b (3.5 ng/ ml) have been shown to decrease intracellular Ca2+ concentrations while higher levels (100 ng/ml) markedly increase it (Campbell and Lynch, 1998).

IL-18 IL-18 is a member of the IL-1 pro-inflammatory cytokine family, closely related to IL-1b (Bazan et al., 1996), and, like other cytokines, can be detected in many brain regions, including the hippocampus (Culhane et al., 1998). It is expressed mainly in glia (both astrocytes and microglia) and is secreted in response to LPS stimulation (Conti

et al., 1999). Again, like TNF-a and IL-1b, acute application of IL-18 has been shown to inhibit the induction of LTP in the dentate gyrus, without affecting baseline neurotransmission (Curran and O’Connor, 2001, 2003; Cumiskey et al., 2007). Another similarity between IL-1b and IL-18 is the observation that IL-18 also activates p38 MAP kinase (Thomassen et al., 1998), which, as mentioned above, is implicated in the IL-1b inhibition of LTP. Recent work has also shown interesting similarities between the mechanism of IL-18 inhibition of LTP and the mechanism by which TNF-a inhibits LTP. Cumiskey et al. (2007) showed that, like TNF-a, IL-18 inhibition of LTP involves mGluRs. The study suggested that IL-18 inhibits LTP in the dentate gyrus by an LTD-like mechanism. In these experiments, the application of the group II mGluR antagonist MTPG prevented the inhibition of LTP by IL-18. Group II and III mGluRs are negatively coupled to adenylyl cyclase (Tanabe et al., 1992), and their activation can induce synaptic depression (Manahan-Vaughan and Reymann, 1997), thus suggesting that IL-18 might be acting through an LTD-like mechanism.

Fig. 4. Schematic diagram of the role of the mGluRs and IL-18R in LTP in the dentate gyrus. IL-18 inhibits the early phase of LTP by activation of IL-18R. Activation of the G-protein-linked mGluR5 leads to inositol-1,4,5-trisphosphate (IP3) receptor-mediated Ca2+ release via phospholipase C (PLC). Activation of the group II mGluRs leads to inhibition of cAMP. p38 has also been implicated in IL-18 mediated inhibition of LTP.

350

An interesting aspect of the involvement of mGluRs in the effect of IL-18 on synaptic transmission and plasticity comes from an apparently contradictory study. Kanno et al. (2004) found that IL-18 facilitates basal synaptic transmission as a result of stimulating presynaptic glutamate release and enhancing AMPA receptor responses in the CA1, but had no significant effect on LTP in that region. The key difference would appear to be the region of the hippocampus under investigation. The expression of mGluRs is known to be very different in the CA1 and dentate regions (Fotuhi et al., 1994), with high levels of mGluR2, mGluR5 and mGluR7 and low levels of mGluR1, mGluR3 and mGluR4 in the dentate, while the CA1 exhibits low mGluR5 and high mGluR1 distribution. This would seem to suggest that the effect of IL-18 on synaptic transmission and plasticity is dependant in some way in the balance between mGluR1 and mGluR5 signalling. The exact nature of this relationship will remain unknown until further studies can shed light on the signalling paths from IL-18 to the mGluRs, which would also provide further insight into the complex nature of IL-18 signalling in the hippocampus (Fig. 4).

Concluding remarks The overriding theme among studies of pro-inflammatory cytokines in the dentate gyrus, and other brain regions, is that of complexity. A primary example is the potential for opposing actions. In addition, the action of these cytokines on neurons is often subtle and multi-faceted, with very different short- and long-term consequences, making their elucidation all the more challenging. A more comprehensive understanding cannot be achieved until the interactions between the cytokine receptors and other neuronal mechanisms are better understood. Acknowledgements We would like to thank the Health Education Authority, Ireland, PRTLI Cycle 3 for funding and Dr Derval Cumiskey for assistance on figure production.

References Akiyama, H., Barger, S., Barnum, S., Bradt, B., Bauer, J., Cole, G.M., Cooper, N.R., Eikelenboom, P., Emmerling, M., Fiebich, B.L., Finch, C.E., Frautschy, S., Griffin, W.S., Hampel, H., Hull, M., Landreth, G., Lue, L., Mrak, R., MacKenzie, I.R., McGeer, P.L., O’Banion, M.K., Pachter, J., Pasinetti, G., Plata-Salaman, C., Rogers, J., Rydel, R., Shen, Y., Streit, W., Strohmeyer, R., Tooyoma, I., Van Muiswinkel, F.L., Veerhuis, R., Walker, D., Webster, S., Wegrzyniak, B., Wenk, G. and Wyss-Coray, T. (2000) Inflammation and Alzheimer’s disease. Neurobiol. Aging, 21: 383–421. Albensi, B.C. and Mattson, M.P. (2000) Evidence for the involvement of TNF-a and NF-kappaB in hippocampal synaptic plasticity. Synapse, 35: 151–159. Allan, S.M. and Rothwell, N.J. (2001) Cytokines and acute neurodegeneration. Nat. Rev. Neurosci., 2: 734–744. Aloe, L., Properzi, F., Probert, L., Akassoglou, K., Kassiotis, G., Micera, A. and Fiore, L. (1999) Learning abilities, NGF and BDNF brain levels in two lines of TNFalpha transgenic mice, one characterized by neurological disorders, the other phenotypically normal. Brain Res., 840: 125–137. Armstrong, R.C., Aja, T., Xiang, J., Gaur, S., Krebs, J.F., Hoang, K., Bai, X., Korsmeyer, S.J., Karanewsky, D.S., Fritz, L.C. and Tomaselli, K.J. (1996) Fas-induced activation of the cell death-related protease CPP32 is inhibited by Bcl-2 and by ICE family protease inhibitors. J. Biol. Chem., 271: 16850–16855. Avital, A., Goshen, I., Kamsler, A., Segal, M., Iverfeldt, K., Richter-Levin, G. and Yirmiya, R. (2003) Impaired interleukin-1 signaling is associated with deficits in hippocampal memory processes and neural plasticity. Hippocampus, 13: 826–834. Ban, E., Milon, G., Prunhom, N., Fillion, G. and Haour, F. (1991) Receptors for interleukin-1 (a and b) in mouse brain: mapping and neuronal localization in hippocampus. Neuroscience, 43: 21–30. Barger, S.W., Horster, D., Furukawa, K., Goodman, Y., Krieglstein, J. and Mattson, M.P. (1995) Tumor necrosis factors alpha and beta protect neurons against amyloid betapeptide toxicity: evidence for involvement of a kappa B-binding factor and attenuation of peroxide and Ca2+ accumulation. Proc. Natl. Acad. Sci. U.S.A., 92: 9328–9332. Bazan, J.F., Timans, J.C. and Kastelein, R.A. (1996) A newly defined interleukin-1b. Nature, 379: 591. Beattie, E.C., Stellwagen, D., Morishita, W., Bresnahan, J.C., Ha, B.K., Von Zastrow, M., Beattie, M.S. and Malenka, R.C. (2002) Control of synaptic strength by glial TNFalpha. Science, 295: 2282–2285. Bellinger, F.P., Madamba, S. and Siggins, G.R. (1993) Interleukin-1b inhibits synaptic strength and long-term potentiation in the rat CAI hippocampus. Brain Res., 628: 227–234. Benveniste, E.N. and Benos, D.J. (1995) TNF-alpha- and IFNgamma-mediated signal transduction pathways: effects on glial cell gene expression and function. FASEB J., 9: 1577–1584.

351 Besedovsky, H.O., del Rey, A., Klusman, I., Furukawa, H., Monge Arditi, G. and Kabiersch, A. (1991) Cytokines as modulators of the hypothalamus-pituitary-adrenal axis. J. Steroid Biochem. Mol. Biol., 40: 613–618. Beutler, B. and Van Huffel, C. (1994a) An evolutionary and functional approach to the TNF receptor/ligand family. Ann. N.Y. Acad. Sci., 730: 118–133. Beutler, B. and Van Huffel, C. (1994b) Unraveling function in the TNF ligand and receptor families. Science, 264: 667–668. Bezzi, P., Domercq, M., Brambilla, L., Galli, R., Schols, D., De Clercq, E., Vescovi, A., Bagetta, G., Kollias, G., Meldolesi, J. and Volterra, A. (2001) CXCR4-activated astrocyte glutamate release via TNF: amplification by microglia triggers neurotoxicity. Nat. Neurosci., 4: 702–707. Bliss, T.V.P. and Collindridge, G.L. (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature, 361: 31–38. Bolshakov, V.Y., Carboni, L., Cobb, M.H., Siegelbaum, S.A. and Belardetti, F. (2000) Dual MAP kinase pathways mediate opposing forms of long-term plasticity at CA3–CA1 synapses. Nat. Neurosci., 3: 1107–1112. Boulanger, L.M., Huh, G.S. and Shatz, C.J. (2001) Neuronal plasticity and cellular immunity: shared molecular mechanisms. Curr. Opin. Neurobiol., 11: 568–578. Bruce, A.J., Boling, W., Kindy, M.S., Peschon, J., Kraemer, P.J., Carpenter, M.K., Holtsberg, F.W. and Mattson, M.P. (1996) Altered neuronal and microglial responses to excitotoxic and ischemic brain injury in mice lacking TNF receptors. Nat. Med., 2: 788–794. Butler, M., O’Connor, J.J. and Moynagh, P. (2004) Dissection of TNF-a inhibition of LTP reveals a p38 MAPK-dependent mechanism which maps to early but not late-phase LTP. Neuroscience, 124: 319–326. Butler, M.P., Moynagh, P.N. and O’Connor, J.J. (2002) Methods of detection of the transcription factor NF-kappa B in rat hippocampal slices. J. Neurosci. Methods, 119: 185–190. Cacquevel, M., Lebeurrier, N., Cheenne, S. and Vivien, D. (2004) Cytokines in neuroinflammation and Alzheimer’s disease. Curr. Drug Targets, 5(6): 529–534. Campbell, V. and Lynch, M.A. (1998) Biphasic modulation of [Ca2] by interleukin-1b in rat cortical synatosomes. Involvement of a PTX-sensitive G-protein and p42 MAP kinase. Neuroreport, 9: 9–12. Carroll, R.C., Beattie, E.C., von Zastrow, M. and Malenka, R.C. (2001) Role of AMPA receptor endocytosis in synaptic plasticity. Nat. Rev. Neurosci., 2(5): 315–324. Cartmell, T., Luheshi, G.N. and Rothwell, N.J. (1999) Brain sites of action of endogenous interleukin-1 in the febrile response to localized inflammation in the rat. J. Physiol., 518(2): 585–594. Cheng, B., Christakos, S. and Mattson, M.P. (1994) Tumor necrosis factors protect neurons against excitotoxic/ metabolic insults and promote maintenance of Ca2+ homeostasis. Neuron, 12: 139–153. Choi, D.W. (1988) Glutamate neurotoxicity and diseases of the nervous system. Neuron, 1: 623–634.

Chun, J. (2001) Selected comparison of immune and nervous system development. Adv. Immunol., 77: 297–322. Chung, I.Y. and Benveniste, E.N. (1990) Tumor necrosis factor-alpha production by astrocytes. Induction by lipopolysaccharide, IFN-gamma, and IL-1 beta. J. Immunol., 144: 2999–3007. Conti, B., Park, L.C., Calingasan, N.Y., Kim, Y., Kim, H., Bae, Y., Gibson, G.E. and Joh, T.H. (1999) Cultures of astrocytes and microglia express interleukin 18. Brain Res. Mol. Brain Res., 67: 46–52. Coogan, A. and O’Connor, J.J. (1997) Inhibition of NMDA receptor-mediated synaptic transmission in the rat dentate gyrus in vitro by IL-1b. Neuroreport, 8(9): 2107–2110. Coogan, A. and O’Connor, J.J. (1999) Interleukin-1b inhibits a tetraethylammonium-induced synaptic potentiation in the rat dentate gyrus in vitro. Eur. J. Pharmacol., 374(2): 197–206. Coogan, A., O’Neill, L.A.J. and O’Connor, J.J. (1999) The p38 MAP kinase inhibitor SB203580 antagonises the inhibitory effect of interleukin-1b on long-term potentiation in the rat dentate gyrus in vitro. Neuroscience, 93(1): 57–69. Coussens, C.M. and Teyler, T.J. (1996) Protein kinase and phosphatase activity regulate the form of synaptic plasticity expressed. Synapse, 24: 97–103. Culhane, A.C., Hall, M.D., Rothwell, N.J. and Luheshi, G.N. (1998) Cloning of rat brain interleukin-18 cDNA. Mol. Psychiatry, 3: 362–366. Cumiskey, D., Butler, M.P., Moynagh, P.N. and O’Connor, JJ. (2004) The inhibitory effect of tumour necrosis factor-a on long-term potentiation is attenuated by type 1 metabotropic glutamate receptor blockade. J. Physiol., 560P: C23. Cumiskey, D. and O’ Connor, J.J. (2006) A role for mGluRs and Ca2+ in the inhibitory effect of tumour necrosis factor-a on long-term potentiation. Proc. Br. Pharmacol. Soc., 3(4): 055P. Cumiskey, D., Pickering, M. and O’Connor, J.J. (2007) Interleukin-18 mediated inhibition of LTP in the rat dentate gyrus is attenuated in the presence of mGluR antagonists. Neurosci. Lett., 412: 206–210. Cunningham, A.J., Murray, C.A., O’Neill, L.A., Lynch, M.A. and O’Connor, J.J. (1996) Interleukin-1 beta (IL-1 beta) and tumour necrosis factor (TNF) inhibit long-term potentiation in the rat dentate gyrus in vitro. Neurosci. Lett., 203: 17–20. Cunningham, E.T., Wada, E., Carter, D.B., Tracey, D.E., Battery, J.F. and De Souza, E.B. (1992) In situ histochemical localization of type I interleukin-1 messenger RNA in the central nervous system, pituitary and adrenal gland of the mouse. J. Neurosci., 12: 1101–1114. Curran, B. and O’Connor, J.J. (2001) The novel pro-inflammatory cytokine interleukin-18 (IL-18) inhibits long-term potentiation in the rat hippocampus in vitro. Neuroscience, 108(1): 83–90. Curran, B. and O’Connor, J.J. (2003) The inhibition of longterm potentiation by pro-inflammatory cytokines is attenuated in the presence of nicotine. Neurosci. Lett., 344: 103–106.

352 Curran, B.P., Murray, H.J. and O’Connor, J.J. (2003) A role for c-jun n-terminal kinase in the inhibition of long-term potentiation by interleukin-1beta and long-term depression in the rat dentate gyrus in vitro. Neuroscience, 118: 347–357. English, J.D. and Sweatt, J.D. (1996) Activation of p42 MAPkinase in hippocampal LTP. J. Biol. Chem., 271: 24329–24332. Floden, A.M., Li, S. and Combs, C.K. (2005) Beta-amyloidstimulated microglia induce neuron death via synergistic stimulation of tumor necrosis factor alpha and NMDA receptors. J. Neurosci., 25: 2566–2575. Fontaine, V., Mohand-Said, S., Hanoteau, N., Fuchs, C., Pfizenmaier, K. and Eisel, U. (2002) Neurodegenerative and neuroprotective effects of tumor necrosis factor (TNF) in retinal ischemia: opposite roles of TNF receptor 1 and TNF receptor 2. J. Neurosci., 22: 1–7. Fotuhi, M., Standaert, D.G., Testa, C.M., Penney Jr., J.B. and Young, A.B. (1994) Differential expression of metabotropic glutamate receptors in the hippocampusand entorhinal cortex of the rat. Brain Res. Mol. Brain Res., 21(3–4): 283–292. Frautschy, SA., Yng, F., Irrizarry, M., Hyman, B., Saido, T.C., Hsiao, K. and Cole, G.M. (1998) Microglial response to amyloid plaques in APPsw transgenic mice. Am. J. Pathol., 152: 307–317. Garcia, J.E., Nonner, D., Ross, D. and Barrett, J.N. (1992) Neurotoxic components in normal serum. Exp. Neurol., 118: 309–316. Gary, D.S., Bruce-Keller, A.J., Kindy, M.S. and Mattson, M.P. (1998) Ischemic and excitotoxic brain injury is enhanced in mice lacking the p55 tumor necrosis factor receptor. J. Cereb. Blood Flow Metab., 18: 1283–1287. Gelbard, H.A., Dzenko, K.A., DiLoreto, D., del Cerro, C., del Cerro, M. and Epstein, L.G. (1993) Neurotoxic effects of tumor necrosis factor alpha in primary human neuronal cultures are mediated by activation of the glutamate AMPA receptor subtype: implications for AIDS neuropathogenesis. Dev. Neurosci., 15: 417–422. Ghosh, S. and Karin, M. (2002) Missing pieces in the NF-kappaB puzzle. Cell, 109(Suppl.): S81–S96. Golan, H., Levav1, T., Mendelsohn, A. and Huleihel, M. (2004) Involvement of tumor necrosis factor alpha in hippocampal development and function. Cereb. Cortex, 14: 97–105. Goodman, Y. and Mattson, M.P. (1996) Ceramide protects hippocampal neurons against excitotoxic and oxidative insults, and amyloid beta-peptide toxicity. J. Neurochem., 66: 869–872. Grell, M., Wajant, H., Lo¨hden, M., Maxeiner, B., Georgopoulos, S., Kollias, G., Lesslauer, W., Pfizenmaier, K. and Scheurich, P. (1995) The transmembrane form of tumor necrosis factor (TNF) is the prime activating ligand of the 80 kDa TNF receptor. Cell, 83: 793–802. Griffin, W.S.T., Sheng, J.G., Roberts, G.W. and Mrak, R.E. (1995) Interleukin-1 expression in different plaque types in Alzheimer’s disease, significance in plaque evolution. J. Neuropathol. Exp. Neurol., 54.

Griffin, W.S.T., Stanley, L.C., Ling, C., MacLeod, V., Perrot, L.J., White III, C.L., et al. (1989) Brain interleukin-1 and S100 immunoreactivity are elevated in Alzheimer disease and in Down syndrome. Proc. Natl. Acad. Sci. U.S.A., 86: 7611–7615. Haour, F.G., Ban, E.M., Milon, G.M., Baran, D. and Fillion, G.M. (1990) Brain interleukin-1 receptors: characterization and modulation after lipopolysaccharide injection. Prog. Neuroendrochronol., 3: 196–204. Haydon, P.G. (2001) Glia: listening and talking to the synapse. Nat. Rev. Neurosci., 2: 185–193. Hikada, H., Watanabe, M. and Kobayashi, R. (1991) Properties and use of H-series compounds as protein kinase inhibitors. Metab. Enzymol., 201: 328–336. Israel, A. (2000) The IKK complex: an integrator of all signals that activate NF-kappaB? Trends Cell Biol., 10: 129–133. Kanno, T., Nagata, T., Yamamoto, S., Okamura, H. and Nishizaki, T. (2004) Interleukin-265 18 stimulates synaptically released glutamate and enhances postsynaptic AMPA receptor responses in the CA1 region of mouse hippocampal slices. Brain Res., 1012: 190–193. Kassiotis, G. and Kollias, G. (2001) Uncoupling the proinflammatory from the immunosuppressive properties of tumor necrosis factor (TNF) at the p55 TNF receptor level: implications for pathogenesis and therapy of autoimmune demyelination. J. Exp. Med., 193: 427–434. Katuski, H., Nakai, S., Hirai, Y., Akaji, K.I., Kiso, Y. and Satoh, M. (1990) Interleukin-l, inhibits long-term potentiation in the CA3 region of mouse hippocampal slices. Eur. J. Pharmacol., 181: 323–326. Kinouchi, K., Brown, G., Pasternak, G. and Donner, D.B. (1991) Identification and characterization of receptors for tumor necrosis factor-alpha in the brain. Biochem. Biophys. Res. Commun., 181: 1532–1538. Klein, B.D., White, H.S. and Callahan, K.S. (2000) Cytokine and intracellular signaling regulation of tissue factor expression in astrocytes. Neurochem. Int., 36: 441–449. Kolesnick, R. and Golde, D.W. (1994) The sphingomyelin pathway in tumor necrosis factor and interleukin-1 signaling. Cell, 77: 325–328. Li, Q. and Verma, I.M. (2002) NF-kappaB regulation in the immune system. Nat. Rev. Immunol., 2: 725–734. Lieberman, A.P., Pitha, P.M., Shin, H.S. and Shin, M.L. (1989) Production of tumor necrosis factor and other cytokines by astrocytes stimulated with lipopolysaccharide or a neurotropic virus. Proc. Natl. Acad. Sci. U.S.A., 86: 6348–6352. Liu, T., Clark, R.K., McDonnell, P.C., Young, P.R., White, R.F., Barone, F.C. and Feuerstein, G.Z. (1994) Tumor necrosis factor-alpha expression in ischemic neurons. Stroke, 25: 1481–1488. Lynch, M.A. (1998) Age-related impairment in long-term potentiation in hippocampus: a role for the cytokine, interleukin-1 beta? Prog. Neurobiol., 56: 571–589. Manahan-Vaughan, D. and Reymann, K.G. (1997) Group 1 metabotropic glutamate receptors contribute to slow-onset potentiation in the rat CA1 region in vivo. Neuropharmacology, 36: 1533–1538.

353 Mattson, M.P., Barger, S.W., Furukawa, K., Bruce, A.J., Wyss-Coray, T., Mark, R.J. and Mucke, L. (1997a) Cellular signaling roles of TGF beta, TNF alpha and beta APP in brain injury responses and Alzheimer’s disease. Brain Res. Brain Res. Rev., 23: 47–61. Mattson, M.P., Goodman, Y., Luo, H., Fu, W. and Furukawa, K. (1997b) Activation of NF-kappaB protects hippocampal neurons against oxidative stress-induced apoptosis: evidence for induction of manganese superoxide dismutase and suppression of peroxynitrite production and protein tyrosine nitration. J. Neurosci. Res., 49: 681–697. Mayford, M., Wang, J., Kandel, E.R. and O’Dell, T.J. (1995) CaMKII regulates the frequency response function of hippocampal synapses for the production of both LTD and LTP. Cell, 81: 891–904. Medina, J.H. and Izquierdo, I. (1995) Retrograde messengers, long-term potentiation and memory. Behav. Brain Res., 58: 1–8. Merrill, J.E. (1992) Pro-inflammatory and antiinflammatory cytokines in multiple sclerosis and central nervous system acquired immunodeficiency syndrome. J. Immunother., 12: 167–170. Morganti-Kossman, M.C., Lenzlinger, P.M., Hans, V., Stahel, P., Csuka, E., Ammann, E., Stocker, R., Trentz, O. and Kossmann, T. (1997) Production of cytokines following brain injury: beneficial and deleterious for the damaged tissue. Mol. Psychiatry, 2: 133–136. Mosmann, T.R., Cherwinski, H., Bond, M.W., Giedlin, M.A. and Coffman, R.L. (1986) Two types of murine helper T cell clone. I. Definition according to profiles of lymphokine activities and secreted proteins. J. Immunol., 136: 2348–2357. Murray, C.A. and Lynch, M.A. (1998) Evidence that increased hippocampal expression of the cytokine interleukin-1b is a common trigger for age- and stress-induced impairments in long-term potentiation. J. Neurosci., 18: 2974–2981. Murray, C.A., McGahon, B., McBennet, S. and Lynch, M.A. (1997) Interleukin-l/1 inhibits glutamate release in hippocampus of young, but not aged, rats. Neurobiol. Ageing, 18: 343–348. Nath, R., Robert, A., McGinnis, K.M. and Wang, K.K.W. (1998) Evidence for activation of caspase-3-like protease in excitotoxins and hypoxia/hypoglycemia-injured cerebrocortical neurons. J. Neurochem., 71: 186–195. Nawashiro, H., Tasaki, K., Ruetzler, C.A. and Hallenbeck, J.M. (1997) TNF-alpha pretreatment induces protective effects against focal cerebral ischemia in mice. J. Cereb. Blood Flow Metab., 17: 483–490. O’Neill, L.A. and Kaltschmidt, C. (1997) NF-kappa B: a crucial transcription factor for glial and neuronal cell function. Trends Neurosci., 20: 252–258. O’Neill, L.A.J. (1996) Pharmacological targets in the immune response. Interleukin-l receptors and signal transduction. Biochem. Soc. Trans., 24: 207–211. O’Neill, L.A.J. (1997) Towards an understanding of the signal transduction pathway for interleukin-1. Biochim. Biophys. Acta, 1266: 31–44.

O’Neill, L.A.J. and Green, C. (1998) Signal transduction pathways activated by the IL-1 receptor family: ancient signalling machinery in mammals, insects and plants. J. Leukoc. Biol., 63: 650–657. Oitzl, M.S., van Oers, H., Schobitz, B. and de Kloet, E.R. (1993) Interleukin-1 beta, but not interleukin-6, impairs spatial navigation learning. Brain Res., 613: 160–163. Okamura, H., Tsutsui, H., Kashiwamura, S., Yoshimoto, T. and Nakanishi, K. (1998) Interleukin-18: a novel cytokine that augments both innate and acquired immunity. Adv. Immunol., 70: 281–312. Pang, L., Sawada, T., Deckers, S.J. and Saltiel, A.R. (1995) Inhibition of MAP kinase kinase blocks the differentiation of PC-12 cells induced by nerve growth factor. J. Biol. Chem., 270: 13585–13588. Parnet, P., Amindari, S., Wu, C., Brunke-Reese, D., Goujon, E., Weyhenmeyer, J.A., Dantzer, R. and Kelley, K.W. (1994) Expression of type I and type II interleukin-I receptors in mouse brain. Mol. Brain Res., 27: 63–70. Pfeffer, K., Matsuyama, T., Kunidg, T.M., Wakeham, A., Kishihara, K., Shahinian, A., Wiegmann, K., Ohashi, P.S., Kro¨nke, M. and Mak, T.W. (1993) Mice deficient for the 55 kd tumor necrosis factor receptor are resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell, 73: 457–467. Plata-Salaman, C.R. and Ffrench-Mullen, J.M.H. (1992) Interleukin-lb depresses calcium currents in CA 1 hippocampal neurons at pathophysiological concentrations. Brain Res. Bull., 29: 221–223. Plata-Salaman, C.R. and Ffrench-Mullen, J.M.H. (1994) Interleukin-lb inhibits Ca2, channel currents in hippocampal neurons through protein kinase C. Eur. J. Pharmacol., 266: 1–10. Plata-Salaman, C.R., Oomura, Y. and Kai, Y. (1988) Tumor necrosis factor and interleukin-1 beta: suppression of food intake by direct action in the central nervous system. Brain Res., 448: 106–114. Rothwell, N. (1999) Cytokines — killers in the brain. J. Physiol., 514: 3–17. Rothwell, N.J. and Hopkins, S.J. (1995) Cytokines and the nervous system II: actions and mechanisms of action. Trends Neurosci., 18: 130–136. Schneider, H., Prossi, F., Balschun, D., Wagner, A., Del Rey, A. and Besedovsky, H.O. (1998) A neuromodulatory role of interleukin-1b in the hippocampus. Proc. Natl. Acad. Sci. U.S.A., 95: 7778–7783. Selmaj, K.W. and Raine, C.S. (1998) Tumor necrosis factor mediates myelin and oligodendrocyte damage in vitro. Ann. Neurol., 23: 339–346. Sheridan, G., Pickering, M., Moynagh, P.N., O’Connor, J.J. and Murphy, K.J. (2006) Tumour necrosis factor-a selectively upregulates nuclear factor kb expression in hippocampal subneuronal populations. Proc. Br. Pharmacol. Soc., 3(4): 113P. Silva, A.J., Smith, A.M. and Giese, K.P. (1997) Gene targeting and the biology of learning and memory. Annu. Rev. Genet., 31: 527–546. Silverman, N. and Maniatis, T. (2001) NF-kappaB signaling pathways in mammalian and insect innate immunity. Genes Dev., 15: 2321–2342.

354 Sitcheran, R., Gupta, P., Fisher, P.B. and Baldwin, A.S. (2005) Positive and negative regulation of EAAT2 by NF-kappaB: a role for N-myc in TNFalpha-controlled repression. EMBO J., 24(3): 510–520. Stellwagen, D., Beattie, E.C., Seo, J.Y. and Malenka, R.C. (2005) Differential regulation of AMPA receptor and GABA receptor trafficking by tumor necrosis factor-alpha. J. Neurosci., 25: 3219–3228. Stellwagen, D. and Malenka, R.C. (2006) Synaptic scaling mediated by glial TNF-alpha. Nature, 440: 1054–1059. Tanabe, Y., Masu, M., Ishii, T., Shigemoto, R. and Nakanishi, S. (1992) A family of metabotropic glutamate receptors. Neuron, 8: 169–179. Tanaka, K., Watase, K., Manabe, T., Yamada, K., Watanabe, M., Takahashi, K., Iwama, H., Nishikawa, T., Ichihara, N., Kikuchi, T., Okuyama, S., Kawashima, N., Hori, S., Takimoto, M. and Wada, K. (1997) Epilepsy and exacerbation of brain injury in mice lacking the glutamate transporter GLT-1. Science, 276: 1699–1702. Tancredi, V., D’Arcangelo, G., Grassi, F., Tarroni, P., Palmieri, G., Santoni, A. and Eusebi, F. (1992) Tumor necrosis factor alters synaptic transmission in rat hippocampal slices. Neurosci. Lett., 146: 176–178. Tartaglia, L.A., Weber, R.F., Figari, I.S., Reynolds, C., Palladino Jr., M.A. and Goeddel, D.V. (1991) The two different receptors for tumor necrosis factor mediate distinct cellular responses. Proc. Natl. Acad. Sci. U.S.A., 88: 9292–9296. Taylor, D.L., Jones, F., Kubota, E.S. and Pocock, J.M. (2005) Stimulation of microglial metabotropic glutamate receptor mGlu2 triggers tumor necrosis factor alpha-induced neurotoxicity in concert with microglial-derived Fas ligand. J. Neurosci., 25(11): 2952–2964. Thomassen, E., Bird, T.A., Renshaw, B.R., Kennedy, M.K. and Sims, J.E. (1998) Binding of interleukin-18 to the interleukin-1 receptor homologous receptor IL-1Rrp1 leads to activation of signaling pathways similar to those used by interleukin-1. J. Interferon Cytokine Res., 18: 1077–1088. Turrigiano, G.G. and Nelson, S.B. (2004) Homeostatic plasticity in the developing nervous system. Nat. Rev. Neurosci., 5: 97–107. Volterra, A. and Steinhauser, C. (2004) Glial modulation of synaptic transmission in the hippocampus. Glia, 47(3): 249–257.

Waage, A., Halstensen, A., Shalaby, R., Brandtzaeg, P., Kierulf, P. and Espevik, T. (1989) Local production of tumor necrosis factor alpha, interleukin 1, and interleukin 6 in meningococcal meningitis. Relation to the inflammatory response. J. Exp. Med., 170: 1859–1867. Wajant, H. and Scheurich, P. (2001) Tumor necrosis factor receptor associated factor (TRAF) 2 and its role in TNF signaling. Int. J. Biochem. Cell Biol., 33: 19–32. Wang, K.K.W., Nath, R., Raser, K.J. and Hajimohammadreza, I. (1996) Maitotoxin induces calpain activation in SH-SY5Y neuroblastoma cells and cerebrocortical cultures. Arch. Biochem. Biophys., 331: 208–214. Wang, Q., Chang, L., Rowan, M.J. and Anwyl, R. (2006) Developmental dependence, the role of the kinases p38 MAPK and PKC, and the involvement of tumor necrosis factor-R1 in the induction of mGlu-5 LTD in the dentate gyrus. Neuroscience, 144: 110–118. Wang, X., Yue, T.L., Barone, F.C., White, R.F., Gagnon, R.C. and Feuerstein, G.Z. (1994) Concomitant cortical expression of TNF-alpha and IL-1 beta mRNAs follows early response gene expression in transient focal ischemia. Mol. Chem. Neuropathol., 23: 103–114. Yang, L., Lindholm, K., Konishi, Y., Li, R. and Shen, Y. (2002) Target depletion of distinct tumor necrosis factor receptor subtype reveals hippocampal neuron death and survival through different signal transduction pathways. J. Neurosci., 22: 3025–3032. Yu, A.C. and Lau, L.T. (2000) Expression of interleukin-1 alpha, tumor necrosis factor alpha and interleukin-6 genes in astrocytes under ischemic injury. Neurochem. Int., 36: 369–377. Yu, Z., Zhou, D., Bruce-Keller, A.J., Kindy, M.S. and Mattson, M.P. (1999) Lack of the p50 subunit of nuclear factor-kappaB increases the vulnerability of hippocampal neurons to excitotoxic injury. J. Neurosci., 19: 8856–8865. Zhao, X., Bausano, B., Pike, B.R., Newcomb-Fernandez, J.K., Wang, K.K., Shohami, E., Ringger, N.C., DeFord, S.M., Anderson, D.K. and Hayes, R.L. (2001) TNF-alpha stimulates caspase-3 activation and apoptotic cell death in primary septo-hippocampal cultures. J. Neurosci. Res., 64: 121–131. Zou, J.Y. and Crews, F.T. (2005) TNF alpha potentiates glutamate neurotoxicity by inhibiting glutamate uptake in organotypic brain slice cultures: neuroprotection by NF kappa B inhibition. Brain Res., 1034(1–2): 11–24.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 21

Role of corticosteroid hormones in the dentate gyrus Marian Joe¨ls Swammerdam Institute of Life Sciences, Center for NeuroScience (SILS-CNS), University of Amsterdam, Kruislaan 320, 1098 SM Amsterdam, The Netherlands

Abstract: Dentate granule cells are enriched with receptors for the stress hormone corticosterone, i.e., the high-affinity mineralocorticoid receptor (MR), which is already extensively occupied with low levels of the hormone, and the glucocorticoid receptor (GR), which is particularly activated after stress. More than any other cell type in the brain studied so far, dentate granule cells require hormone levels to be within the physiological range. In the absence of corticosteroids, proliferation and apoptotic cell death are dramatically enhanced. Dendritic morphology and synaptic transmission are compromised. Conversely, prolonged exposure of animals to a high level of corticosterone suppresses neurogenesis and presumably makes dentate granule cells more vulnerable to delayed cell death. These corticosteroid effects on dentate cell and network function are translated into behavioral consequences, in health and disease. Keywords: chronic stress; adrenalectomy; neurogenesis; apoptosis; electrophysiology

mineralocorticoid

glucocorticoid

receptor;

latter causes an array of behavioral changes including increased vigilance, alertness, focused attention, and selection of appropriate strategies, which together result in behavioral adaptation. Moreover, the hormones help to store relevant information in brain, for future use. The initial phases of behavioral adaptation primarily involve fast-acting catecholamines, i.e., mainly noradrenaline, neuropeptides such as corticotrophin releasing hormone and vasopressin (de Kloet et al., 2005), and possibly neurosteroids (Stell et al., 2003). Moreover, it has become clear in recent years that corticosterone may also play a role in these early phases, through non-genomic pathways (Di et al., 2003; Karst et al., 2005). The main role of corticosteroids, though, is to change brain function through slow and long-lasting actions. These long-lasting actions are mediated by corticosteroid hormone receptors, of which two main types have been recognized in the brain

Systems activated by stress When organisms are exposed to a stressful situation, sensory information about the situation is processed in the brain, and output regarding the event is forwarded to the hypothalamus (de Kloet et al., 1998, 2005; see Fig. 1). From there two hormonal systems are activated, aimed to face the challenge and eventually normalize the disturbed balance: the sympatho-adrenomedullar system, which leads to the release of adrenaline from the adrenal medulla; and the hypothalamo-pituitaryadrenal system, which results in elevated levels of the adrenocortical hormone (corticosterone in rodents, cortisol in humans). Adrenaline and corticosterone are not only active in peripheral organs but can also strongly affect brain function. The Corresponding author. Tel.: +31-20-5257626; Fax: +31-20-5257709; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63021-0

receptor;

355

356

Fig. 1. Exposure of a rat to stress may activate many brain regions (depending on the type of stressor), including the amygdala (Amy), hippocampus (Hipp) and prefrontal cortex (PFC). These areas project to the hypothalamus (HYP). Stimulation of cells in the hypothalamus leads to the activation of the fast-acting sympatho-adrenomedullar system (lower right) and the slower-acting hypothalamo-pituitary-adrenal axis (lower left). Both systems not only affect the function of peripheral organs but also feed back to the brain, via adrenaline and corticosterone respectively. Adrenaline can, via intermediate steps involving the nucleus tractus solitarius, give rise to central release of noradrenaline (NA) from the locus coeruleus (LC), which then exerts widespread influence on other areas such as the amygdala, prefrontal cortex and hippocampus. Corticosterone is distributed throughout the brain but acts only at those sites where receptors are enriched. Two receptor types are known in the brain, i.e., the mineralocorticoid (MR) and the glucocorticoid receptor (GR). Principal cells in the hippocampal CA1 area and the dentate gyrus (DG; see inset) are among the few cells in the brain that express MR (gray dots) as well as GR (black dots). CA3 neurons primarily express MR. In most other parts of the brain, GRs prevail. ANS ¼ autonomic nervous system; ACTH ¼ adrenocorticotropin hormone; CRH ¼ corticotropin releasing hormone. Adapted with permission from Elsevier (Joe¨ls et al., 2006).

(Reul and de Kloet, 1985). The first type, i.e., the mineralocorticoid receptor (MR), is identical to the protein found in peripheral organs like the kidney where it is involved in mineral balance. This receptor displays a very high affinity for the mineralocorticoid aldosterone as well as for corticosterone (Kd
0.5 nM). In kidney tissue, aldosterone preferentially binds to the MR, since

corticosterone is converted into a biologically inactive 11-keto isoform, by the enzyme 11-bhydroxysteroid dehydrogenase type 2 (Seckl and Walker, 2003). In most brain regions this enzyme is not present during adulthood, so corticosterone (which is present in a 100-fold excess over aldosterone) under normal conditions is the most likely occupant of MRs in brain. MRs show a restricted

357

distribution within the brain, with high abundance in some limbic areas, like the hippocampal subregions CA1, CA3 and dentate gyrus, the lateral septum and central amygdala, and in motor nuclei of the brain stem (de Kloet et al., 1998). The second receptor type in brain is the glucocorticoid receptor (GR), which has an approximately 10fold lower affinity for corticosterone than the MR. GRs have a relatively low affinity for aldosterone, but effectively bind synthetic steroids like dexamethasone. This receptor is much more ubiquitous than the MR and can be found in nearly all brain regions — both in neurons and glial cells — although some areas show a high enrichment, such as the paraventricular nucleus of the hypothalamus, the hippocampal CA1 region, and the dentate gyrus. Recent studies have shown that both the MR and GR gene give rise to several splice variants, and the proteins are known to have multiple isoforms due to alternative translation initiation sites as well as abundant posttranslational modifications caused by acetylation, phosphorylation, ubiquitination and sumoylation (Pascual-Le Tallec and Lombes, 2005; Zhou and Cidlowski, 2005). The functional relevance of these variants in brain, though, is presently not understood. As is evident from the above overview, some cells in brain express both MR and GR. This is for instance the case for granule cells in the dentate gyrus. The degree of activation of the two receptor types, however, depends on the available levels of the hormone. When the organism is at rest and circadian release of corticosteroids is at its nadir, MRs are already activated to a substantial degree by corticosterone, but occupation of GRs by corticosterone is below 20% (de Kloet et al., 1998). After exposure to stress or at the circadian peak, GRs will become increasingly occupied by corticosterone. Under physiological conditions, corticosteroid receptor occupancy in dentate granule cells will therefore range from a situation of predominant MR activation to a situation where both receptor types are fully occupied. Electrophysiological studies over the past decades have investigated both conditions. In addition, studies have examined what happens to dentate cells in the complete absence of corticosteroid

hormones, or when animals have been exposed to excess corticosteroid levels (e.g., due to chronic stress). Each of these conditions will be addressed in the following sections.

Dentate function in the absence of corticosteroid hormones Cell turnover and morphology The dentate gyrus is quite unique in that it requires corticosteroid receptor activation to prevent apoptosis and to preserve neuronal integrity. Nearly two decades ago it was reported for the first time that complete removal of corticosterone by adrenalectomy (ADX) can lead to substantial loss of dentate granule cells (Sloviter et al., 1989; Gould et al., 1990), due to a process showing all the hallmarks of apoptosis (Sloviter et al., 1993a). The degree of apoptosis varies greatly among individual animals, but rarely involves more than 25% of the neurons in the first week after ADX. Interestingly, about one out of five animals do not show any apoptosis upon ADX (Sloviter et al., 1993b; Stienstra et al., 1998). It is still not fully understood why these animals escape from apoptosis, but most likely these animals still have some residual corticosterone producing cells (despite the fact that circulating corticosterone levels are below the detection limit). Approximately 3 days are required for cells to fully undergo the apoptosis program upon removal of corticosterone (Hu et al., 1997). Substitution with a low dose of corticosterone, which preferentially occupies the MR, or with aldosterone can fully prevent apoptosis (Woolley et al., 1991; Stienstra et al., 1998), pointing to a crucial role of the MR in preventing apoptosis. Why some cells undergo apoptosis but neighboring cells that are seemingly comparable stay healthy is not easy to understand. Presumably, relatively ‘young’ granule cells are more resistant to apoptosis (Cameron and Gould, 1996). We tested the hypothesis that surviving cells differ in their gene expression profile triggered by ADX, resulting in resistance against apoptosis (Nair et al., 2004; Fig. 2). Multiple genes expressed in surviving individual granule cells were assessed by

358

Fig. 2. (A) Slices from animals that were adrenalectomized (ADX) 1, 2 or 3 days previously, or sham-operated rats, were stained in vitro with Hoechst 33258, a stain for nuclear chromatin. Apoptotic granule cells with condensed nuclear chromatin are brightly colored (arrows) and could not be recorded with whole cell patch clamp electrodes. Neighboring cells, which were not apoptotic, were recorded (cross), and RNA was subsequently isolated from the cell. This procedure allows the collection of RNA from cells that have not yet entered the apoptotic pathways. (B) After RNA collection, multiple gene expression in surviving individual granule cells was assessed by linear antisense RNA amplification and hybridization to slot blots containing various neuronal cDNAs. (C) Compared to cells from sham-operated control rats (average of all data from the 3 days, normalized to 10%, black squares), the ratio between Bcl-2 and Bax RNA expression was significantly enhanced in 2- and 3-day ADX cells (open squares), which at that time had still resisted entering the apoptotic pathway. A similar enhanced expression ratio was seen 3 days after ADX for calbindin relative to calcium-calmodulin kinase II (CamKII). (D) To confirm that Bcl-2 and calbindin conferred resistance to entering the apoptotic route, these two genes were overexpressed unilaterally in the dentate gyrus shortly before ADX; a control virus was injected into the contralateral hemisphere. A significant reduction in the number of apoptotic cells 3 days after ADX was seen in the ipsilateral versus contralateral (open bars) dentate gyrus of animals receiving the Bcl-2-expressing virus (gray bars), but not the calbindin-expressing virus (black bar). Adapted with permission from Blackwell (Nair et al., 2004).

linear antisense RNA amplification and hybridization to slot blots containing various neuronal cDNAs. Hierarchical clustering and principal component analysis was performed on two physiological variables and 14 mRNA ratios from cells collected 1, 2 or 3 days after ADX or sham operation. The results indicated that surviving 3-day ADX granule cells display lower membrane capacitance, lower relative N-methyl-D-aspartate (NMDA) R1 mRNA expression and higher

relative MR, alpha1A voltage-gated calcium channel, Bcl-2 and NMDA R2C mRNA expression; it is important to realize that overall dentate expression of these products may show changes in the opposite direction (e.g., Cardenas et al., 2002), most likely caused by the cells destined to die. Some 1- and 2-day ADX cells clustered with these 3-day survivors; therefore, one or more components of their mRNA expression profile may represent predictive markers for apoptosis resistance.

359

The functional relevance of two candidate genes was tested by in vivo local over-expression in the same model system; of these, Bcl-2 conferred partial protection when induced shortly before ADX. Collectively, the data indicate that cells with a low membrane capacitance and high input resistance — i.e., presumably ‘younger’ dentate granule cells (Liu et al., 1996) — can trigger a gene expression profile upon ADX which prevents initiation of the apoptotic program, whereas ‘older’ cells may no longer be able to do this. Probably not all degradation in the dentate following ADX is due to apoptosis. Interestingly, it was shown that loss of markers of mature neurons can occur, rather than cell death, and this can be reversed by a serotonin-1A receptor agonist (Huang et al., 1997). This could reflect that the latter restored antigenicity; more likely, though, it could indicate that ADX leads to neurite retraction, which is restored by a serotonin-1A receptor agonist. In agreement with the latter, 3-D reconstruction of surviving granule cells after ADX revealed that the dendritic length of higher order segments is significantly reduced in animals with undetectable corticosterone levels (Wossink et al., 2001). Yet, in ADX animals with residual circulating corticosterone no such reduced dendritic length was observed, indicating that minute amounts of corticosterone suffice to maintain dendritic integrity of granule cells. Growth factors, which are sensitive to the presence of corticosteroids in the dentate gyrus, may play a role in maintaining the dendritic structure (Chao and McEwen, 1994). ADX not only promotes apoptosis and neurodegeneration but also neurogenesis (Cameron and Gould, 1994). Already within a day of corticosterone removal the degree of cell proliferation in the dentate gyrus was enhanced (Cameron et al., 1995). This process may depend on intact entorhinal input and NMDA receptor-dependent transmission (Cameron et al., 1995). Data support that facilitation of neurogenesis after ADX not only requires MR activation but also GR activation (Wong and Herbert, 2005). The importance of corticosterone and the MR for neurogenesis and dentate granule cell number also follows from observations in mutant mice. Thus, in a pro-opiomelanocortin null mutant mouse (where

activation of the adrenal cortex by adrenocorticotropin is absent) granule cell density was decreased, due to diminished cell proliferation (Ostwald et al., 2006). Mice with a genetic disruption of the MR also exhibited decreased granule cell density and a significant reduction in neurogenesis (Gass et al., 2000).

Physiology If removal of corticosterone leads to loss of granule cells within 3 days, this would be expected to impair functional network properties in the dentate, particularly so if the loss involves mature neurons that are fully integrated into the network. The fact that proliferation is enhanced shortly after ADX is less relevant at early time points, because at that time the newborn cells are not yet incorporated into the network. Later on, of course, this burst in neurogenesis may have functional consequences too. Initial field potential recordings indeed corroborated the notion that apoptosis impairs synaptic transmission in the dentate gyrus. Thus, stimulation of the perforant path in vitro leads to a field excitatory postsynaptic potential (fEPSP) in the molecular layer of the dentate gyrus. Three days after ADX, both the slope and the amplitude of this fEPSP were reduced by approximately 30% (Stienstra et al., 1998). No reduction in fEPSP was observed in those animals that did not display apoptosis. Substitution of animals with a low dose of corticosterone, sufficient to activate the brain MRs and prevent apoptosis, fully normalized the fEPSPS properties. Yet, three observations indicate that the reduced field response is not a consequence of apoptotic cell loss. First, the reduced fEPSP slope and amplitude were already present 1 day after ADX, i.e., several days before the first apoptotic cells can be discerned (Stienstra et al., 1998). Second, administration of corticosterone in vitro to slices from 3-day ADX rats fully restored the amplitude and slope of the fEPSP, despite the presence of apoptotic cells in the slices (Stienstra and Joe¨ls, 2000). Finally, intracellular recording of dentate granule cells — which due to the method is confined to

360

non-apoptotic cells with an intact plasma membrane — also revealed a 40% reduction of the EPSP amplitude (Joe¨ls et al., 2001; Fig. 3A). This was observed both for the NMDA- and AMPA receptor-mediated components of the synaptic response. The latter suggests that postsynaptic glutamate receptors are unlikely to be directly changed after ADX, but rather that dendritic segments carrying both types of receptors, or presynaptic afferents, are lost. Most passive (resting membrane potential, input resistance) and active membrane properties (action potential height and width) were on average not significantly changed 3 days after ADX (Joe¨ls et al., 2001). As these properties depend on the age of the cells (Liu et al., 1996), these observations also indicate that the functional consequences of apoptotic cell death of mature dentate cells are

probably limited. Rather, in parallel with (and not as a consequence of) the change in cell turnover, an extensive synaptic reorganization seems to occur which largely alters synaptic properties of the surviving cells. In addition to these altered responses to lowfrequency stimulation of perforant path afferents, it has also been found that long-term potentiation (LTP) is reduced after ADX (Smriga et al., 1996; Krugers et al., 2007), although one report found no change (Pavlides et al., 1994). Again, the reduced LTP occurs independent of altered cell turnover, as LTP (but not cell turnover) can be fully normalized by adding corticosterone to slices from 3 day ADX rats (Krugers et al., 2007). Ionic conductances of dentate granule cells have also been studied after ADX. More specifically, the influx of calcium (Ca) through high

Fig. 3. (A) Typical excitatory postsynaptic potentials (EPSPs) induced in dentate granule cells by stimlation of the perforant path. The EPSP amplitude and area under the signal (voltage  time) was reduced after ADX (A2) compared to sham operated controls (A1) or ADX rats treated in vivo with a low dose of corticosterone sufficient to occupy the MR but not GR (A3). When the AMPA and NMDA receptor-mediated components of the EPSPs were pharmacologically isolated, it became apparent that both were reduced by 40–50% 3 days after ADX (A4). Adapted with permission from the American Physiological Society (Joe¨ls et al., 2001). (B) The distribution of the dentate Ca-current amplitudes in sham-operated control animals can be fitted well with two Gaussian curves (dark gray line), suggesting that under control conditions most cells have relatively small Ca-current amplitudes, but a subpopulation of dentate cells displays a large Ca-current amplitude. At short intervals after ADX (1–2 days, light gray line), both populations can still be discerned. The curve is shifted to the right, reflecting an ADX-dependent increase in Ca-current amplitude. Some days later (stippled line; 3–7 days after ADX, i.e., at a time point when some of the cells have died by apoptosis), the peak of the amplitude distribution is more or less at the same position, suggesting that in most neurons the Ca-current amplitude is not further enhanced. Importantly, though, the curve became unimodal: the subpopulation with large Ca-current amplitudes were no longer present, suggesting that those cells that had a large Ca-current amplitude to start with died by apoptosis. Adapted with permission from Blackwell (Karst and Joe¨ls, 2001).

361

voltage-activated channels was investigated (Karst and Joe¨ls, 2001). Shortly after ADX (i.e., 1–2 days later) Ca-current amplitude and density were increased, whereas ADX rats treated with a low dose of corticosterone (thus preferentially occupying the MR) exhibited small Ca currents. Interestingly, reduced Ca influx was observed at later times after ADX. Further investigation of the amplitude distribution (Fig. 3B) revealed that while the current amplitude distribution in general was shifted to the right after ADX, a particular subpopulation of cells with large Ca-current amplitudes had completely disappeared. It was reasoned that ADX enhances Ca influx of all dentate granule cells, similar to what had been observed earlier in CA1 neurons (Karst et al., 1994). In cells with a large Ca influx — i.e., presumably the relatively ‘old’ cells (Thibault and Landfield, 1996) — the ADXinduced increase in Ca influx may be just sufficient to make them vulnerable to cell death. Combined with a reduced potency to trigger genes that would confer resistance to apoptotic cell death, this enhanced Ca influx may cause the disappearance of these ‘older’ cells 3–7 days after ADX. In summary, low levels of corticosterone sufficient to activate MRs in dentate granule cells seem to be necessary to preserve neuronal integrity in this area. In the absence of corticosteroids and MR activation, apoptosis and neurogenesis are increased. Even prior to apoptosis, synaptic reorganization seems to occur, resulting in reduced synaptic efficacy and plasticity. At this early stage too, Ca influx is enhanced. Relatively ‘young’ granule cells may withstand these adverse conditions, as they probably have a small Ca influx to start with, and seem to be able to increase the expression of survival genes such as Bcl-2. Yet, ‘older’ cells are particularly vulnerable and may be destined to degenerate and/or enter the apoptotic pathway.

Physiological variations in corticosteroid levels Cell turnover In contrast to what is seen after ADX (or chronic stress, see below), neurogenesis and apoptosis in the dentate gyrus are only mildly affected by

exposure to a single stressor. It was found that proliferation is indeed suppressed shortly after stress exposure (Heine et al., 2004b). Apoptosis is increased, not only after stress (Heine et al., 2004b), but also after a single injection with a high dose of dexamethasone (Hassan et al., 1996), involving shifts in the expression of pro- and antiapoptotic genes (Almeida et al., 2000). The effects of acute stress, though, are largely normalized within 24 h (Heine et al., 2004b), indicating that the impact of a single stressor is probably limited.

Physiology In cells which carry both MRs and GRs, such as pyramidal neurons in the CA1 region or granule cells in the dentate gryus, physiological fluctuations in corticosterone levels will result in a receptor occupation ranging from predominant MR occupation when the organism is at rest and at the circadian nadir, to concomitant MR and GR occupation after stress or at the circadian peak. In CA1 neurons, these two conditions are associated with distinctly different physiological responses (Joe¨ls, 1997). Thus, compared to the situation where corticosteroid receptors are unoccupied, conditions of predominant MR occupation are associated with restricted Ca influx in the soma and primary dendrites, limited cell-firing frequency accommodation, stable glutamate/GABAergic transmission, small hyperpolarizing responses to serotonin (5-HT) and efficient LTP. All of these properties help to promote ongoing transfer of information through this area. Concomitant activation of GRs leads in a slow gene-mediated fashion to enhanced Ca influx, prominent cell-firing frequency accommodation, large hyperpolarizing responses to serotonin (5-HT), reduction of excitatory responses to noradrenaline, and reduction of LTP; there seems to be an activity (energy)dependent attenuation of glutamate and gamma amino butyric acid (GABA) mediated responses, although postsynaptic responses via AMPA receptors are enhanced within a time-window of 2–4 h after GR activation (Karst and Joe¨ls, 2005). All in all, GR activation generally may result in normalization of activity that was temporary

362

raised by transmitters that reach the CA1 region at the time of the stress exposure. As is also evident from the above discussion, steroid-dependency for many cell properties in the CA1 area follows a U-shaped curve (Joe¨ls, 2006). For instance, Ca currents are large in the absence of corticosterone, small with predominant MR activation, and large again when both receptor types are activated (Fig. 4). Although the consequences of physiological fluctuations in corticosteroid levels are far less investigated in the dentate gyrus than in the CA1 region, a U-shaped dose dependency may not exist for all properties in the dentate (Joe¨ls, 2006; see Fig. 4). In the dentate, as opposed to the CA1 region (Kerr et al., 1992, 1994), no significant difference in Ca-current amplitude was observed in slices with predominant MR versus slices with concomitant MR and GR activation (Van Gemert and Joe¨ls, 2006). It should be noted, though, that these data were obtained in animals that had been handled daily for 3 weeks, which in itself may affect cellular function; still, preliminary data in naı¨ ve animals confirm this observation (Van Gemert et al., unpublished observation). Similarly, AMPA receptor-mediated synaptic responses were comparable in slices (from handled rats) with predominant MR activation and slices where both receptor types were activated (Karst and Joe¨ls, 2003). This was also seen for the slope of the fEPSP (Stienstra and Joe¨ls, 2000). Hyperpolarizing responses to 5-HT do show some differences between conditions of low and high corticosterone concentrations (Karten et al., 2001), but these are not as distinct as in the CA1 area. A gene expression survey of single dentate cells from handled rats — involving, e.g., transcripts for the subunits of glutamate and GABAA receptors and of Ca channels — also revealed very little differences due to activation of GRs (Qin et al., 2004). Apparently, a rise in corticosterone level does often not lead to efficient activation of GR or GR-dependent pathways in the dentate gyrus. This could be due to local differences in GR variants, in posttranslational modification of the GR or in the presence of proteins that interact with GR. The steroid sensitivity of LTP in the dentate gyrus seems to be somewhat more complex to

interpret. Initial experiments reported that selective activation of MRs in ADX rats prolongs LTP (Pavlides et al., 1994), whereas very high doses of corticosterone reduce LTP in vivo (Pavlides et al., 1993) and show a strong tendency toward a decrease in vitro (Alfarez et al., 2003). Later studies showed that forced and uncontrollable swim stress prolongs LTP (Korz and Frey, 2003; Kavushansky et al., 2006), via a rapid MR-dependent mechanism (Korz and Frey, 2003) involving mitogenactivated protein kinases (Ahmed et al., 2006) and depending on intact input from the basolateral amygdala (Korz and Frey, 2005). This suggests that LTP in the dentate can be reduced by corticosterone (as in the CA1 area), but that stress in the intact brain causes a different effect, due to modulatory synaptic inputs and the effect of stress-released factors other than corticosteroid hormones (e.g., peptides or neurosteroids). In conclusion, activation of MRs in dentate granule cells is very important for the physiology of the area. In general it serves the same role as in the CA1 area, i.e., it promotes the transfer of signals through the dentate gyrus. To what extent additional GR activation plays a role in the dentate is at this moment less clear. The available data seems to indicate that the strong effects of GR activation (on top of MR activation) as seen in the CA1 area do not take place in the dentate gyrus.

Dentate function after chronic stress Cell turnover Cell proliferation and apoptosis have been extensively studied in the dentate in relation to chronic stress, a condition that is known to be a risk factor in the etiology of psychiatric diseases like major depression (McEwen, 2003; de Kloet et al., 2005). This research has recently been boosted by the observation that anti-depressants promote neurogenesis and that, in fact, this proliferative action is necessary for these compounds to exert their behavioral effects in an animal model (Santarelli et al., 2003; see further Chapter 38 in this book). In general, chronic stress — be it from repeated restraining, from a psychosocial source, or from a

363 100 90 80 70 60

response (% of maximum)

50 40

CA1 CA1

30 20

100 90 80 70 60 50 40

DG DG

30 20

MR MR GR GR stress Fig. 4. Dose-response relationships for cellular effects of corticosterone in the CA1 hippocampal area and the dentate gyrus (DG). All responses are expressed as a percentage of the maximal response. The concentration of corticosterone is a rough estimate of the local concentration, based on the solutions perfused on in vitro preparations or derived from the plasma concentration when fluctuations in hormone levels were measured in vivo. In the CA1 area, both the amplitude of depolarization-induced Ca currents (open squares) and the hyperpolarization caused by serotonin-1A receptor activation (filled circles) display a clear U-shaped dose dependency. The descending limb is linked to activation of MRs, while the ascending limb is associated with gradual GR activation, as occurs after stress; relative occupation of the two receptors with increasing corticosterone levels is depicted at the bottom. DG granule neurons show a clear effect on the field potential (filled squares) and single cell response (closed triangles) caused by activation of glutamatergic AMPA receptors; this effect is linked to MR activation. Although these cells also abundantly express GRs, high doses of corticosterone do not lead to additional changes, except when animals had been exposed to 3 weeks of unpredictable stress prior to recording (open triangles). Adapted with permission from Elsevier (Joe¨ls, 2006).

364

combination of physical and psychological stressors — has been reported to reduce cell proliferation in the dentate gyrus, at least in males (Czeh et al., 2001; Pham et al., 2003; Westenbroek et al., 2004; Heine et al., 2004b). This seems to be largely due to the elevation in corticosterone level associated with chronic stress, as repeated exogenous administration of corticosterone (which enhances corticosteroid levels but suppresses circulating levels of CRH and ACTH) also decreased the number of adult newborn cells, due to a reduction in cell survival (Wong and Herbert, 2004). Importantly, administration of a GR antagonist can fully normalize the reduction in cell proliferation and/or survival after repeated exogenous corticosterone administration (Mayer et al., 2006). It is presently still unclear, however, whether exogenous corticosterone and chronic stress indeed have the exact same effects on all aspects of neurogenesis, i.e., proliferation, migration, differentiation and survival of newborn cells. Whether or not chronic stress results in a reduction of the dentate volume may depend on the duration or severity of the stress protocol (Pham et al., 2003; Heine et al., 2004b). Surprisingly, apoptosis was reduced after chronic stress when analyzed over the entire dentate gyrus and hilus (Heine et al., 2004b). This may signify that chronic stress slows down the cell cycle, causing fewer neurons to be born, but also fewer to die. This is corroborated by the observation that chronic stress increases the level of p27Kip, an inhibitor of the cell cycle (Heine et al., 2004a). Both the effect on neurogenesis and on apoptosis observed after 3 weeks of unpredictable stress or restraint stress were largely reversed when animals were allowed to recover from stress for 3 weeks (Pham et al., 2003; Heine et al., 2004b). As progenitor cells do not express GRs (Garcia et al., 2004), corticosteroid-dependent effects on their proliferation requires other cells in the neighborhood which do contain GRs and can produce factors that indirectly affect the function of progenitor cells. Several potential intermediate factors have been investigated, but the emerging picture is somewhat confusing. It was found that, while acute stressors lead to a transient increase in brainderived neurotrophic factor (BDNF) mRNA levels

in the dentate gyrus (Marmigere et al., 2003), repeated restraint stress reduces BDNF mRNA levels (Smith et al., 1995). In contrast, NT-3 mRNA levels were increased. Stress did not affect the expression of neurotrophin-4, or tyrosine receptor kinases (trkB or C). More recently, though, it was found that chronic stress increases the levels of the catalytic but not of the truncated form of trkB (Nibuya et al., 1999). As many newborn cells are found in close association with the vasculature, mediators in this system were also investigated. A significant reduction in the level of vascular endothelial growth factor and its receptor Flk-1 was observed after 3 weeks of unpredictable stress (Heine et al., 2005). These levels were restored when the period of stress was followed by 3 weeks of rest. In addition to these growth factors, several cell adhesion factors have been studied. Effects of chronic over-exposure to corticosterone varied, however, because it was found to enhance the expression of neural cell adhesion molecule in the dentate gyrus (Pham et al., 2003), decrease it (Venero et al., 2002; Nacher et al., 2004) or leave it unaffected (Van Gemert et al., 2006). Transcripts for the adhesion molecule L1 were increased (Venero et al., 2002). Clearly, the crucial factors mediating the effect of chronic stress on progenitor cell function still need to be identified.

Physiology While temporary elevations in corticosterone level, leading to brief activation of GRs, seem to be relatively ineffective in the dentate gyrus of handled or naı¨ ve rats (as opposed to the CA1 area), such elevations do become very effective when occurring against a background of chronic stress. This was found to be true for GR-evoked changes in gene expression as well as functional endpoints. For instance, high doses of corticosterone in vitro do not change the AMPA receptor-mediated responses in handled animals, but a large increase in the amplitude of excitatory postsynaptic currents was observed with the same dose of corticosterone when slices were prepared from chronically-stressed rats (Karst and Joe¨ls, 2003). A significant enhancement was also seen with

365

respect to the amplitude of the sustained Ca current, when comparing corticosterone-treated cells from control with those of chronically-stressed rats (Van Gemert and Joe¨ls, 2006). This effect of chronic stress was also seen for the relative expression of the subunit, which forms the pore of the L-type (sustained) Ca channel (Qin et al., 2004). It has been found, both in vivo (Pavlides et al., 2002) and in vitro (Alfarez et al., 2003), that induction of LTP is very much impaired in the dentate gyrus of chronically stressed rats. Additional administration of a high dose of corticosterone is ineffective (Alfarez et al., 2003), as LTP cannot be reduced more than it is already in tissue exposed to low concentrations of the hormone. Taken together, it is evident that repeated exposure of the dentate gyrus to high levels of corticosterone reduces the number of newborn cells, through mechanisms that are presently only partly understood. Over a period of weeks this may not lead to a large reduction in the overall number of cells or volume of the adult dentate gyrus. Periods of stress relief may even partially normalize the changes in cell turnover. However, prolonged ongoing stress to which the organism cannot adapt may substantially change the composition of dentate gyrus granule cells. This could result in cognitive disturbances, and contribute to the precipitation of clinical symptoms. Interestingly, with respect to physiological properties, dentate granule cells seem to respond stronger to activation of GRs (and thus presumably to acute stressors) when the organism has experienced some weeks of unpredictable stress. The ensuing increased excitability in combination with enhanced Ca influx could make cells more vulnerable to excitotoxic damage.

Functional relevance of corticosteroid effects in the dentate gyrus Corticosteroid dose dependency in the dentate gyrus While many brain regions are affected by stressrelated hormones, the dentate gyrus seems to be exquisitely sensitive to concentrations of such

hormones that go beyond the borders of a restricted, physiological range. When corticosteroid hormone concentrations drop below a critical level, or when they are repeatedly elevated, neuronal integrity in the dentate gyrus is at stake. In the CA1 area of the hippocampus, conditions that lead to predominant activation of the MR have been proposed to facilitate ongoing activity as well as promote neuronal viability. In the dentate gyrus, however, activation of MRs is an absolute necessity to preserve neurites and to restrain excessive cell turnover. On the other end of the spectrum, repeated stress seems to slow the cycle of proliferating cells and put granule cells at risk, by increasing their response to glutamate and enhancing voltage-dependent Ca influx. Clearly, much of the unique sensitivity of the dentate gyrus to fluctuating corticosteroid levels is caused by the presence of progenitor cells in this area, which do not carry corticosteroid receptors themselves but nevertheless are very much influenced in their activity through neighboring cells that do express MR and GR. In CA1 pyramidal cells, which like dentate granule cells are among the few types of neurons that co-express MR and GR, physiological shifts in corticosteroid level (and hence in receptor occupation ratios) lead to distinct changes in functional properties. As discussed above, a U-shaped dose dependency has been observed for corticosteroid actions in this region. By contrast, dentate granule cells seem to react in a more linear fashion, with a relatively limited response to temporary activation of GRs. For instance, (1) absence of corticosterone leads to increased neurogenesis and (2) exposure to an acute stressor leads to a temporary reduction in neurogenesis; (3) repetitive stress exposure results in a more prominent and sustained suppression of neurogenesis (Fig. 5). Another example pertains to glutamate receptormediated transmission: (1) in the absence of corticosterone, responses to glutamate are small; (2) activation of MRs increases the responses; (3) while additional GR activation does not seem to further enhance the response in control rats, such activation does lead to a marked enhancement of AMPA receptor-mediated responses in animals with a history of chronic stress exposure.

366

A

C

handled 6000

# BrdUpositive cells

Mol. layer SGZ h hilus

GCL

*

Acute stress 24 hrs after acute stress

5000 4000 3000 2000 1000 0

# H3 Thymidine-labeled cells

9

*

D 3500

SHAM

8 7

ADX

6 5 4 3 2

# BrdUpositive cells

B

3000 2500

*

handled 24 hrs after chronic stress

2000 1500 1000

1

500

0

0

Fig. 5. (A) Immunohistochemical image of the dentate gyrus, showing proliferating cells mostly in the subgranular zone (SGZ) and far less in the granular cell layer (GCL), molecular layer and hilus. Proliferating cells were marked with Ki-67. (B) One week after ADX or sham operation, animals received an injection of 3H thymidine to label dividing cells. Three weeks later, dentate cells with a neuronal phenotype were counted and expressed in number per 106 mm2. The number of newborn neurons was significantly enhanced after ADX. Adapted with permission from Elsevier (Cameron and Gould, 1994). (C) Injection of BrdU just before exposure to acute stress revealed a significant reduction in the number of newborn cells in the dentate gyrus (excluding the hilus). When BrdU was injected one day after acute stress exposure, the difference in the number of newborn cells no longer was evident. Apparently, acute stress only results in a transient suppression of proliferation in the dentate gyrus. Adapted from Heine et al., 2004b. (D). If BrdU was injected one day after chronic stress exposure, the number of newborn cells was significantly reduced, indicating that chronic stress induces a much longer-lasting suppression of cell proliferation. Adapted from Heine et al., 2004b and unpublished observations.

Importantly, the altered glutamate responses do not depend on the changes in neurogenesis, but nevertheless seem to have the same steroid dependency. Why two cell types that apparently share the same steroid receptor expression pattern — i.e., CA1 pyramidal neurons and dentate granule cells — display such a different steroid dose dependency is not understood, but most likely local factors like the intracellular protein composition and network properties play an important role.

Functional relevance in health and disease The effects of corticosteroid hormones on physiological properties of hippocampal neurons will

have consequences for behavioral functions in which these neurons participate. Indeed, numerous studies have described how stress and/or selective manipulation of MR or GR occupancy alter hippocampal-dependent behavior. With respect to tests requiring spatial memory, it was found that MR occupation is important for reaction to novel information as well as determination of behavioral strategy (Oitzl and de Kloet, 1992; Oitzl et al., 1994). Additional GR activation within the learning context is required for consolidation of spatial information (Oitzl and de Kloet, 1992; Oitzl et al., 2001). Importantly, GR activation that is not related to the learning context may impair rather than improve memory consolidation (De Kloet et al., 1999). It has been proposed that convergence

367

in time and space (i.e., brain area) is necessary for the facilitating effects of stress on memory to take place (Joe¨ls et al., 2006). It should also be realized that due to stress, many transmitter and hormone systems are activated, so that not only corticosterone levels but also those of, e.g., CRH and noradrenaline will be elevated. This is of great relevance since dentate gyrus function, and in particular LTP, is known to be enhanced both by CRH (Wang et al., 2000) and noradrenaline (e.g., Bramham et al., 1997). Next to these short-lasting stressors, many studies have addressed the effects of chronic stress on memory function. It was found that chronic stress exposure impairs spatial memory in males (Luine et al., 1994; Wright and Conrad, 2005); yet, memory improvement was observed in females (Luine, 2002; McLaughlin et al., 2005). It is unclear, however, to what extent the spatial orientation studies reflect functions of the dentate gyrus. In this respect, it may be also of interest to consider effects of stress on anxiety-related behavior, which was shown to involve amygdala nuclei as well as hippocampal subregions (LeDoux, 2000; McGaugh, 2004). An extensive line of research has revealed that inhibitory avoidance memory is facilitated by stress (McGaugh and Roozendaal, 2002). Contextual fear conditioning is also promoted by additional stress exposure (Cordero et al., 2003; Donley et al., 2005). The facilitation of avoidance behavior critically depends on GR activation in the basolateral amygdala and requires interactions between noradrenaline and corticosterone to occur within a restricted time-frame (McGaugh and Roozendaal, 2002). Although intrahippocampal injections of a GR agonist could also improve avoidance memory, these effects only occurred when the basolateral amygdala was intact (Roozendaal et al., 1999). A clear interaction between basolateral amygdala and the hippocampus, particularly the dentate gyrus, was also demonstrated with electrophysiology. Thus, perforant path-induced LTP in the dentate was shown to be facilitated by amygdala stimulation applied within 1 h before tetanic stimulation of the perforant path (Akirav and Richter-Levin, 2002). With longer time delays, amygdala stimulation reduced perforant path LTP. It was shown that both effects involve noradrenaline as well as corticosterone.

More distinct functional changes may occur in the dentate gyrus when the corticosteroid levels are at extreme levels for a considerable period of time. As discussed above, this will occur when corticosteroid levels are extremely low, such as after ADX in experimental animals. In humans though, extreme adrenal insufficiency will rarely pass unnoticed for any length of time, so the consequences for brain function will be usually curtailed by replacement therapy. By contrast, prolonged periods of exposure to uncontrollable and/or unpredictable stressors are not uncommon. Animal studies indicate that renewed exposure to stress of an organism that already has a history of repeated stress experiences induces particularly strong changes in gene expression and cell physiology of dentate cells, such as enhanced response to excitatory input and increased voltage-dependent Ca influx. The consequences of these changes in physiology will be particularly prominent when they coincide with extensive depolarization of the dentate area. This could occur when the dentate is challenged by demanding behavioral events but also in association with pathological conditions such as ischemic insults (Krugers et al., 2000). This fits with the general theory that (prolonged) exposure to glucocorticoids may exacerbate damage inflicted to hippocampal cells by pathological conditions (Sapolsky, 1996; McEwen, 2004). A better understanding of the functional consequence of corticosteroid actions in the dentate, however, awaits studies that more specifically address this issue. It will be very informative to study the influences of variations in corticosteroid level using behavioral tests that more precisely reflect dentate gyrus function as well as pathological conditions that critically depend on the contribution of this area. References Ahmed, T., Frey, J.U. and Korz, V. (2006) Long-term effects of brief acute stress on cellular signaling and hippocampal LTP. J. Neurosci., 26: 3951–3958. Akirav, I. and Richter-Levin, G. (2002) Mechanisms of amygdala modulation of hippocampal plasticity. J. Neurosci., 22: 9912–9921. Alfarez, D.N., Joe¨ls, M. and Krugers, H.J. (2003) Chronic unpredictable stress impairs long-term potentiation in rat

368 hippocampal CA1 area and dentate gyrus in vitro. Eur. J. Neurosci., 17: 1928–1934. Almeida, O.F., Conde, G.L., Crochemore, C., Demeneix, B.A., Fischer, D., Hassan, A.H., Meyer, M., Holsboer, F. and Michaelidis, T.M. (2000) Subtle shifts in the ratio between pro- and antiapoptotic molecules after activation of corticosteroid receptors decide neuronal fate. FASEB J., 14: 779–790. Bramham, C.R., Bacher-Svendsen, K. and Sarvey, J.M. (1997) LTP in the lateral perforant path is beta-adrenergic receptordependent. Neuroreport, 8: 719–724. Cameron, H.A. and Gould, E. (1994) Adult neurogenesis is regulated by adrenal steroids in the dentate gyrus. Neuroscience, 61: 203–209. Cameron, H.A. and Gould, E. (1996) Distinct populations of cells in the adult dentate gyrus undergo mitosis or apoptosis in response to adrenalectomy. J. Comp. Neurol., 369: 56–63. Cameron, H.A., McEwen, B.S. and Gould, E. (1995) Regulation of adult neurogenesis by excitatory input and NMDA receptor activation in the dentate gyrus. J. Neurosci., 15: 4687–4692. Cardenas, S.P., Parra, C., Bravo, J., Morales, P., Lara, H.E., Herrera-Marschitz, M. and Fiedler, J.L. (2002) Corticosterone differentially regulates bax, bcl-2 and bcl-x mRNA levels in the rat hippocampus. Neurosci. Lett., 331: 9–12. Chao, H.M. and McEwen, B.S. (1994) Glucocorticoids and the expression of mRNAs for neurotrophins, their receptors and GAP-43 in the rat hippocampus. Brain Res. Mol. Brain Res., 26: 271–276. Cordero, M.I., Venero, C., Kruyt, N.D. and Sandi, C. (2003) Prior exposure to a single stress session facilitates subsequent contextual fear conditioning in rats. Evidence for a role of corticosterone. Horm. Behav., 44: 338–345. Czeh, B., Michaelis, T., Watanabe, T., Frahm, J., de Biurrun, G., van Kampen, M., Bartolomucci, A. and Fuchs, E. (2001) Stress-induced changes in cerebral metabolites, hippocampal volume, and cell proliferation are prevented by antidepressant treatment with tianeptine. Proc. Natl. Acad. Sci. U.S.A., 98: 12796–12801. De Kloet, E.R., Joe¨ls, M. and Holsboer, F. (2005) Stress and the brain: from adaptation to disease. Nat. Rev. Neurosci., 6: 463–475. De Kloet, E.R., Oitzl, M.S. and Joe¨ls, M. (1999) Stress and cognition: are corticosteroids good or bad guys? Trends Neurosci., 22: 422–426. De Kloet, E.R., Vreugdenhil, E., Oitzl, M.S. and Joe¨ls, M. (1998) Brain corticosteroid receptor balance in health and disease. Endocr. Rev., 19: 269–301. Di, S., Malcher-Lopes, R., Halmos, K.C. and Tasker, J.G. (2003) Nongenomic glucocorticoid inhibition via endocannabinoid release in the hypothalamus: a fast feedback mechanism. J. Neurosci., 23: 4850–4857. Donley, M.P., Schulkin, J. and Rosen, J.B. (2005) Glucocorticoid receptor antagonism in the basolateral amygdala and ventral hippocampus interferes with long-term memory of contextual fear. Behav. Brain Res., 164: 197–205.

Garcia, A., Steiner, B., Kronenberg, G., Bick-Sander, A. and Kempermann, G. (2004) Age-dependent expression of glucocorticoid — and mineralocorticoid receptors on neural precursor cell populations in the adult murine hippocampus. Aging Cell, 3: 363–371. Gass, P., Kretz, O., Wolfer, D.P., Berger, S., Tronche, F., Reichardt, H.M., Kellendonk, C., Lipp, H.P., Schmid, W. and Schu¨tz, G. (2000) Genetic disruption of mineralocorticoid receptor leads to impaired neurogenesis and granule cell degeneration in the hippocampus of adult mice. EMBO Rep., 1: 447–451. Gould, E., Woolley, C.S. and McEwen, B.S. (1990) Short-term glucocorticoid manipulations affect neuronal morphology and survival in the adult dentate gyrus. Neuroscience, 37: 367–375. Hassan, A.H., von Rosenstiel, P., Patchev, V.K., Holsboer, F. and Almeida, O.F. (1996) Exacerbation of apoptosis in the dentate gyrus of the aged rat by dexamethasone and the protective role of corticosterone. Exp. Neurol., 140: 43–52. Heine, V.M., Maslam, S., Joe¨ls, M. and Lucassen, P.J. (2004a) Increased P27KIP1 protein expression in the dentate gyrus of chronically stressed rats indicates G1 arrest involvement. Neuroscience, 129: 593–601. Heine, V.M., Maslam, S., Zareno, J., Joe¨ls, M. and Lucassen, P.J. (2004b) Suppressed proliferation and apoptotic changes in the rat dentate gyrus after acute and chronic stress are reversible. Eur. J. Neurosci., 19: 131–144. Heine, V.M., Zareno, J., Maslam, S., Joe¨ls, M. and Lucassen, P.J. (2005) Chronic stress in the adult dentate gyrus reduces cell proliferation near the vasculature and VEGF and Flk-1 protein expression. Eur. J. Neurosci., 21: 1304–1314. Hu, Z., Yuri, K., Ozawa, H., Lu, H. and Kawata, M. (1997) The in vivo time course for elimination of adrenalectomyinduced apoptotic profiles from the granule cell layer of the rat hippocampus. J. Neurosci., 17: 3981–3989. Huang, J., Strafaci, J.A. and Azmitia, E.C. (1997) 5-HT1A receptor agonist reverses adrenalectomy-induced loss of granule neuronal morphology in the rat dentate gyrus. Neurochem. Res., 22: 1329–1337. Joe¨ls, M. (1997) Steroid hormones and excitability in the mammalian brain. Front. Neuroendocrinol., 18: 2–48. Joe¨ls, M. (2006) Corticosteroid effects in the brain: U-shape it. Trends Pharmacol. Sci., 27: 244–250. Joe¨ls, M., Pu, Z., Wiegert, O., Oitzl, M.S. and Krugers, H.J. (2006) Learning under stress: how does it work? Trends Cogn. Sci., 10: 152–158. Joe¨ls, M., Stienstra, C. and Karten, Y. (2001) Effect of adrenalectomy on membrane properties and synaptic potentials in rat dentate granule cells. J. Neurophysiol., 85: 699–707. Karst, H., Berger, S., Turiault, M., Tronche, F., Schu¨tz, G. and Joe¨ls, M. (2005) Mineralocorticoid receptors are indispensable for nongenomic modulation of hippocampal glutamate transmission by corticosterone. Proc. Natl. Acad. Sci. U.S.A., 102: 19204–19207.

369 Karst, H. and Joe¨ls, M. (2001) Calcium currents in rat dentate granule cells are altered after adrenalectomy. Eur. J. Neurosci., 14: 503–512. Karst, H. and Joe¨ls, M. (2003) Effect of chronic stress on synaptic currents in rat hippocampal dentate gyrus neurons. J. Neurophysiol., 89: 625–633. Karst, H. and Joe¨ls, M. (2005) Corticosterone slowly enhances miniature excitatory postsynaptic current amplitude in mice CA1 hippocampal cells. J. Neurophysiol., 94: 3479–3486. Karst, H., Wadman, W.J. and Joe¨ls, M. (1994) Corticosteroid receptor-dependent modulation of calcium currents in rat hippocampal CA1 neurons. Brain Res., 649: 234–242. Karten, Y.J., Stienstra, C.M. and Joe¨ls, M. (2001) Corticosteroid effects on serotonin responses in granule cells of the rat dentate gyrus. J. Neuroendocrinol., 13: 233–238. Kavushansky, A., Vouimba, R.M., Cohen, H. and RichterLevin, G. (2006) Activity and plasticity in the CA1, the dentate gyrus, and the amygdala following controllable vs. uncontrollable water stress. Hippocampus, 16: 35–42. Kerr, D.S., Campbell, L.W., Thibault, O. and Landfield, P.W. (1992) Hippocampal glucocorticoid receptor activation enhances voltage-dependent Ca2+ conductances: relevance to brain aging. Proc. Natl. Acad. Sci. U.S.A., 89: 8527–8531. Korz, V. and Frey, J.U. (2003) Stress-related modulation of hippocampal long-term potentiation in rats: involvement of adrenal steroid receptors. J. Neurosci., 23: 7281–7287. Korz, V. and Frey, J.U. (2005) Bidirectional modulation of hippocampal long-term potentiation under stress and nostress conditions in basolateral amygdala-lesioned and intact rats. J. Neurosci., 25: 7393–7400. Krugers, H.J., van der Linden, S., van Olst, E., Alfarez, D.N., Maslam, S., Lucassen, P.J. and Joe¨ls, M. (2007) Dissociation between apoptosis, neurogenesis and synaptic potentiation in the dentate gyrus of adrenalectomized rats. Synapse, 61: 221–230. Krugers, H.J., Maslam, S., Korf, J. and Joe¨ls, M. (2000) The corticosterone synthesis inhibitor metyrapone prevents hypoxia/ischemia-induced loss of synaptic function in the rat hippocampus. Stroke, 31: 1162–1172. LeDoux, J.E. (2000) Emotion circuits in the brain. Annu. Rev. Neurosci., 23: 155–184. Liu, Y.B., Lio, P.A., Pasternak, J.F. and Trommer, B.L. (1996) Developmental changes in membrane properties and postsynaptic currents of granule cells in rat dentate gyrus. J. Neurophysiol., 76: 1074–1088. Luine, V. (2002) Sex differences in chronic stress effects on memory in rats. Stress, 5: 205–216. Luine, V., Villegas, M., Martinez, C. and McEwen, B.S. (1994) Repeated stress causes reversible impairments of spatial memory performance. Brain Res., 639: 167–170. Marmigere, F., Givalois, L., Rage, F., Arancibia, S. and TapiaArancibia, L. (2003) Rapid induction of BDNF expression in the hippocampus during immobilization stress challenge in adult rats. Hippocampus, 13: 646–655. Mayer, J.L., Klumpers, L., Maslam, S., de Kloet, E.R., Joe¨ls, M. and Lucassen, P.J. (2006) Brief treatment with the

glucocorticoid receptor antagonist mifepristone normalises the corticosterone-induced reduction of adult hippocampal neurogenesis. J. Neuroendocrinol., 18: 629–631. McEwen, B.S. (2003) Mood disorders and allostatic load. Biol. Psychiatry, 54: 200–207. McEwen, B.S. (2004) Protection and damage from acute and chronic stress: allostasis and allostatic overload and relevance to the pathophysiology of psychiatric disorders. Ann. N.Y. Acad. Sci., 1032: 1–7. McGaugh, J.L. (2004) The amygdala modulates the consolidation of memories of emotionally arousing experiences. Annu. Rev. Neurosci., 27: 1–28. McGaugh, J.L. and Roozendaal, B. (2002) Role of adrenal stress hormones in forming lasting memories in the brain. Curr. Opin. Neurobiol., 12: 205–210. McLaughlin, K.J., Baran, S.E., Wright, R.L. and Conrad, C.D. (2005) Chronic stress enhances spatial memory in ovariectomized female rats despite CA3 dendritic retraction: possible involvement of CA1 neurons. Neuroscience, 135: 1045–1054. Nacher, J., Gomez-Climent, M.A. and McEwen, B. (2004) Chronic non-invasive glucocorticoid administration decreases polysialylated neural cell adhesion molecule expression in the adult rat dentate gyrus. Neurosci. Lett., 370: 40–44. Nair, S.M., Karst, H., Dumas, T., Phillips, R., Sapolsky, R.M., Rumpff-van Essen, L., Maslam, S., Lucassen, P.J. and Joe¨ls, M. (2004) Gene expression profiles associated with survival of individual rat dentate cells after endogenous corticosteroid deprivation. Eur. J. Neurosci., 20: 3233–3243. Nibuya, M., Takahashi, M., Russell, D.S. and Duman, R.S. (1999) Repeated stress increases catalytic TrkB mRNA in rat hippocampus. Neurosci. Lett., 67: 81–84. Oitzl, M.S., Fluttert, M. and de Kloet, E.R. (1994) The effect of corticosterone on reactivity to spatial novelty is mediated by central mineralocorticosteroid receptors. Eur. J. Neurosci., 6: 1072–1079. Oitzl, M.S. and de Kloet, E.R. (1992) Selective corticosteroid antagonists modulate specific aspects of spatial orientation learning. Behav. Neurosci., 106: 62–71. Oitzl, M.S., Reichardt, H.M., Joe¨ls, M. and de Kloet, E.R. (2001) Point mutation in the mouse glucocorticoid receptor preventing DNA binding impairs spatial memory. Proc. Natl. Acad. Sci. U.S.A., 98: 12790–12795. Ostwald, D., Karpac, J. and Hochgeschwender, U. (2006) Effects on hippocampus of lifelong absence of glucocorticoids in the pro-opiomelanocortin null mutant mouse reveal complex relationship between glucocorticoids and hippocampal structure and function. J. Mol. Neurosci., 28: 291–302. Pascual-Le Tallec, L. and Lombes, M. (2005) The mineralocorticoid receptor: a journey exploring its diversity and specificity of action. Mol. Endocrinol., 19: 2211–2221. Pavlides, C., Kimura, A., Magarinos, A.M. and McEwen, B.S. (1994) Type I adrenal steroid receptors prolong hippocampal long-term potentiation. Neuroreport, 5: 2673–2677.

370 Pavlides, C., Nivon, L.G. and McEwen, B.S. (2002) Effects of chronic stress on hippocampal long-term potentiation. Hippocampus, 12: 245–257. Pavlides, C., Watanabe, Y. and McEwen, B.S. (1993) Effects of glucocorticoids on hippocampal long-term potentiation. Hippocampus, 3: 183–192. Pham, K., Nachler, J., Hof, P.R. and McEwen, B.S. (2003) Repeated restraint stress suppresses neurogenesis and induces biphasic PSA-NCAM expression in the adult rat dentate gyrus. Eur. J. Neurosci., 17: 879–886. Qin, Y., Karst, H. and Joe¨ls, M. (2004) Chronic unpredictable stress alters gene expression in rat single dentate granule cells. J. Neurochem., 89: 364–374. Reul, J.M. and de Kloet, E.R. (1985) Two receptor systems for corticosterone in rat brain: microdistribution and differential occupation. Endocrinology, 117: 2505–2511. Roozendaal, B., Nguyen, B.T., Power, A.E. and McGaugh, J.L. (1999) Basolateral amygdala noradrenergic influence enables enhancement of memory consolidation induced by hippocampal glucocorticoid receptor activation. Proc. Natl. Acad. Sci. U.S.A., 96: 11642–11647. Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F., Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O., Belzung, C. and Hen, R. (2003) Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science, 301: 805–809. Sapolsky, R.M. (1996) Stress, glucocorticoids, and damage to the nervous system: the current state of confusion. Stress, 1: 1–19. Seckl, J.R. and Walker, B.R. (2003) Minireview: 11betahydroxysteroid dehydrogenase type 1 — a tissue-specific amplifier of glucocorticoid action. Endocrinology, 142: 1371–1376. Sloviter, R.S., Dean, E. and Neubort, S. (1993a) Electron microscopic analysis of adrenalectomy-induced hippocampal granule cell degeneration in the rat: apoptosis in the adult central nervous system. J. Comp. Neurol., 330: 337–351. Sloviter, R.S., Sollas, A.L., Dean, E. and Neurbort, S. (1993b) Adrenalectomy-induced granule cell degeneration in the rat hippocampal dentate gyrus: characterization of an in vivo model of controlled neuronal death. J. Comp. Neurol., 330: 324–336. Sloviter, R.S., Valiquette, G., Abrams, G.M., Ronk, E.C., Sollas, A.L., Paul, L.A. and Neubort, S. (1989) Selective loss of hippocampal granule cells in the mature rat brain after adrenalectomy. Science, 243: 535–538. Smith, M.A., Makino, S., Kvetnansky, R. and Post, R.M. (1995) Stress and glucocorticoids affect the expression of brain-derived neurotrophic factor and neurotrophin-3 mRNAs in the hippocampus. J. Neurosci., 15: 1768–1777. Smriga, M., Saito, H. and Nishiyama, N. (1996) Hippocampal long- and short-term potentiation is modulated by adrenalectomy and corticosterone. Neuroendocrinology, 64: 35–41. Stell, B.M., Brickley, S.G., Tang, C.Y., Farrant, M. and Mody, I. (2003) Neuroactive steroids reduce neuronal excitability by selectively enhancing tonic inhibition mediated by delta

subunit-containing GABAA receptors. Proc. Natl. Acad. Sci. U.S.A., 100: 14439–14444. Stienstra, C.M., van der Graaf, F., Bosma, A., Karten, Y.J., Hesen, W. and Joe¨ls, M. (1998) Synaptic transmission in the rat dentate gyrus after adrenalectomy. Neuroscience, 85: 1061–1071. Stienstra, C.M. and Joe¨ls, M. (2000) Effect of corticosteroid treatment in vitro on adrenalectomy-induced impairment of synaptic transmission in the rat dentate gyrus. J. Neuroendocrinol., 12: 199–205. Thibault, O. and Landfield, P.W. (1996) Increase in single L-type calcium channels in hippocampal neurons during aging. Science, 272: 1017–1020. Van Gemert, N.G. and Joe¨ls, M. (2006) Effect of chronic stress and RU38486 treatment on voltage-dependent calcium currents in rat hippocampal dentate gyrus. J. Neuroendocrinol., 18: 732–741. Van Gemert, N.G., van Riel, E., Meijer, O.C., Fehr, S., Schachner, M. and Joe¨ls, M. (2006) No effect of prolonged corticosterone over-exposure on NCAM, SGK1, and RGS4 mRNA expression in rat hippocampus. Brain Res., 1093: 161–166. Venero, C., Tilling, T., Hermans-Borgmeyer, I., Schmidt, R., Schachner, M. and Sandi, C. (2002) Chronic stress induces opposite changes in the mRNA expression of the cell adhesion molecules NCAM and L1. Neuroscience, 115: 1211–1219. Wang, H.L., Tsai, L.Y. and Lee, E.H. (2000) Corticotropinreleasing factor produces a protein synthesis — dependent long-lasting potentiation in dentate gyrus neurons. J. Neurophysiol., 83: 343–349. Westenbroek, C., den Boer, J.A., Veenhuis, M. and ter Horst, G.J. (2004) Chronic stress and social housing differentially affect neurogenesis in male and female rats. Brain Res. Bull., 64: 303–308. Wong, E.Y. and Herbert, J. (2004) The corticoid environment: a determining factor for neural progenitors’ survival in the adult hippocampus. Eur. J. Neurosci., 20: 2491–2498. Wong, E.Y. and Herbert, J. (2005) Roles of mineralocorticoid and glucocorticoid receptors in the regulation of progenitor proliferation in the adult hippocampus. Eur. J. Neurosci., 22: 785–792. Woolley, C.S., Gould, E., Sakai, R.R., Spencer, R.L. and McEwen, B.S. (1991) Effects of aldosterone or RU28362 treatment on adrenalectomy-induced cell death in the dentate gyrus of the adult rat. Brain Res., 554: 312–315. Wossink, J., Karst, H., Mayboroda, O. and Joe¨ls, M. (2001) Morphological and functional properties of rat dentate granule cells after adrenalectomy. Neuroscience, 108: 263–272. Wright, R.L. and Conrad, C.D. (2005) Chronic stress leaves novelty-seeking behavior intact while impairing spatial recognition memory in the Y-maze. Stress, 8: 151–154. Zhou, J. and Cidlowski, J.A. (2005) The human glucocorticoid receptor: one gene, multiple proteins and diverse responses. Steroids, 70: 407–417.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 22

Neurotrophins in the dentate gyrus Devin K. Binder Department of Neurological Surgery, University of California, Irvine, CA 92868, USA

Abstract: Since the discovery of nerve growth factor (NGF) in the 1950s and brain-derived neurotrophic factor (BDNF) in the 1980s, a great deal of evidence has mounted for the roles of neurotrophins (NGF; BDNF; neurotrophin-3, NT-3; and neurotrophin-4/5, NT-4/5) in development, physiology, and pathology. BDNF in particular has important roles in neural development and cell survival, as well as appearing essential to molecular mechanisms of synaptic plasticity and larger scale structural rearrangements of axons and dendrites. Basic activity-related changes in the central nervous system (CNS) are thought to depend on BDNF modulation of synaptic transmission. Pathologic levels of BDNF-dependent synaptic plasticity may contribute to conditions such as epilepsy and chronic pain sensitization, whereas application of the trophic properties of BDNF may lead to novel therapeutic options in neurodegenerative diseases and perhaps even in neuropsychiatric disorders. In this chapter, I review neurotrophin structure, signal transduction mechanisms, localization and regulation within the nervous system, and various potential roles in disease. Modulation of neurotrophin action holds significant potential for novel therapies for a variety of neurological and psychiatric disorders. Keywords: brain-derived neurotrophic factor; neurotrophin-3; neurotrophin-4/5; nerve growth factor; epilepsy antiparallel b-strands and cysteine residues in a cystine knot motif. Initially produced as proneurotrophins, prohormone convertases such as furin cleave the proneurotrophins (molecular weight, MW
30 kDa) to the mature neurotrophin (MW
14 kDa) (Chao and Bothwell, 2002). Proneurotrophins have altered binding characteristics and distinct biologic activity in comparison with mature neurotrophins (Lee et al., 2001b). Mature neurotrophins are noncovalently linked homodimers with MW approximately 28 kDa. Dimerization appears essential for neurotrophin (NT) receptor activation. BDNF shares approximately 50% amino acid identity with NGF, NT-3, and NT-4/5. The BDNF gene (in humans mapped to chromosome 11p) has four 50 exons (exons I–IV) that are associated with distinct promoters, and one 30

Introduction to neurotrophins Neurotrophin structure Each neurotrophin (including nerve growth factor, NGF; brain-derived neurotrophic factor, BDNF; neurotrophin-3, NT-3; neurotrophin-4/5, NT-4/5; see below) consists of a noncovalently linked homodimer and contains (1) a signal peptide following the initiation codon; (2) a pro-region containing an N-linked glycosylation site and a proteolytic cleavage site for furin-like pro-protein convertases, followed by the mature sequence; (3) a distinctive three-dimensional structure containing two pairs of Corresponding author. Tel.: +1 (714) 456-6966; Fax: +1 (714) 456-8212; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63022-2

371

372

exon (exon V) that encodes the mature BDNF protein (Metsis et al., 1993; Timmusk et al., 1993b). Eight distinct mRNAs are transcribed, with transcripts containing exons I–III expressed predominantly in brain and exon IV found in lung and heart (Timmusk et al., 1993b).

Neurotrophin signal transduction Each NT binds one or more of the tropomyosinrelated kinase (trk) receptors, members of the family of receptor tyrosine kinases (RTKs) (Patapoutian and Reichardt, 2001). Trk proteins are transmembrane RTKs homologous to other RTKs such as the epidermal growth factor (EGF) receptor and insulin receptor family. Ligandinduced receptor dimerization results in kinase activation; subsequent receptor autophosphorylation on multiple tyrosine residues creates specific binding sites for intracellular target proteins, which bind to the activated receptor via SH2 domains (Barbacid, 1994; Patapoutian and Reichardt, 2001). These include PLCg1 (phospholipase C), p85 (the noncatalytic subunit of PI-3 kinase), and Shc (SH2containing sequence); activation of these target proteins can then lead to a variety of intracellular signaling cascades such as the Ras-MAP (mitogenactivated protein) kinase cascade and phosphorylation of cyclic AMP-response element binding protein (CREB) (Patapoutian and Reichardt, 2001; Segal, 2003). Binding specificity is conferred via the juxtamembrane Ig-like domain of the extracellular portion of the receptor in the following pattern (Urfer et al., 1995). TrkA binds NGF (with low-affinity binding by NT-3 in some systems); trkB binds BDNF and NT-4/5 with lower-affinity binding by NT-3; and trkC binds NT-3 (Barbacid, 1994). Trk receptors exist in both a full-length (trkB.FL) form as well as truncated (trkB.T1, trkB.T2) forms lacking the kinase domain (Eide et al., 1996; Fryer et al., 1997). Although most functions attributed to BDNF are associated with full-length trkB, several roles have been suggested for truncated receptors, including growth and development (Fryer et al., 1997; Yacoubian and Lo, 2000; Luikart et al., 2003) and negative modulation of trkB receptor expression

and function (Eide et al., 1996; Haapasalo et al., 2001, 2002). Expression of truncated trk receptors on astrocytes is upregulated following injury (Frisen et al., 1993) and may modulate neuronal vulnerability (Saarelainen et al., 2000a) and sequestration of BDNF in astrocytes (Biffo et al., 1995; Roback et al., 1995; Alderson et al., 2000). Recent studies have shown that BDNF activates glial calcium signaling by truncated trk receptors (Climent et al., 2000; Rose et al., 2003). In addition, all of the NTs bind to the p75 receptor, designated p75NTR. p75NTR, related to proteins of the tumor necrosis factor (TNFR) superfamily, has a glycosylated extracellular region involved in ligand binding, a transmembrane region, and a short cytoplasmic sequence lacking intrinsic catalytic activity (Chao and Hempstead, 1995; Dechant and Barde, 2002). NT binding to p75NTR is linked to several intracellular signal transduction pathways, including nuclear factorkB (NF-kB), Jun kinase, and sphingomyelin hydrolysis (Dechant and Barde, 2002). p75NTR signaling mediates biologic actions distinct from those of the trk receptors, notably the initiation of programed cell death (apoptosis) (CasacciaBonnefil et al., 1996; Frade et al., 1996; Roux et al., 1999; Dechant and Barde, 2002). It has also been suggested that p75 may serve to determine NT binding specificity (Esposito et al., 2001; Lee et al., 2001a; Zaccaro et al., 2001).

Nerve growth factor (NGF) NGF was discovered in the early 1950s by Rita Levi-Montalcini and Viktor Hamburger due to its trophic (survival and growth-promoting) effects on sensory and sympathetic neurons (Levi-Montalcini and Hamburger, 1951). In addition, NGF supports the survival and neurotransmitter synthesis of cholinergic neurons in the central nervous system (CNS). In the brain, it is synthesized primarily in cholinergic target tissues such as the cortex, hippocampal pyramidal layer, and striatum (Gall and Isackson, 1989; Rylett and Williams, 1994). The trkA NT receptor is expressed primarily on the axons of NGF-dependent cholinergic neurons (Sobreviela et al., 1994).

373 Table 1. Seizure regulation of NGF expression Reference

Methods

Results

NGF mRNA Gall and Isackson (1989) Gall and Lauterborn (1992)

Hilar electrolytic lesion Hilar electrolytic lesion

Increases in DG (4 h) and cortex (17 h) Increase in DG, max at 6 h (10  ) and 24 h (6  ) (biphasic) Increase max at 1 h (DG), 4 h (PC)

Ernfors et al. (1991) Bengzon et al. (1993)

Rapid kindling (ventral hippocampal stimulation) Traditional kindling (ventral hippocampal stimulation CA1–CA2)

Schmidt-Kastner et al. (1996) Mudo et al. (1996) Sato et al. (1996)

Pilocarpine Pilocarpine Traditional amygdala kindling

Morimoto et al. (1998)

Traditional amygdala kindling

NGF protein Bengzon et al. (1992)

Rapid hippocampal kindling (ventral hippocampal stimulation) NGF ELISA

NGF gene regulation NGF expression levels are regulated by activity. This has been most clearly demonstrated following the intense activity associated with seizures (Table 1). Hilar electrolytic lesion-induced (Gall and Isackson, 1989; Gall and Lauterborn, 1992; Lauterborn et al., 1994) or kindled seizures (Ernfors et al., 1991; Bengzon et al., 1993; Sato et al., 1996; Morimoto et al., 1998) induce a rapid and transient expression of NGF mRNA in dentate gyrus granule cells as well as piriform cortex. Similarly, pilocarpineinduced status epilepticus increases NGF mRNA expression in dentate gyrus, maximum at approximately 3 h (Mudo et al., 1996; Schmidt-Kastner et al., 1996). Basal levels of NGF protein in the hippocampus are very low (Narisawa-Saito and Nawa, 1996). Studies with NGF ELISAs indicate that NGF protein levels do increase following kindling, especially in dentate gyrus, piriform cortex, and parietal cortex (Bengzon et al., 1992). In contrast to the mRNA increases, NGF protein levels remain elevated for at least 7 days (Bengzon et al., 1992).

Increase — did not do time course but studied relationship between development of kindling and neurotrophin induction — found similar NGF induction (approximately 2  at 2 h time point) regardless of kindling stage Increase in DG max at 3–6 h Increase in DG max at 3 h (approximately 2.5  ) Increase in CA1, CA3, perirhinal cortex 1 h after stage 5 kindled seizure Increase in DG, max at 2 h (2  ) After 40 stimulations, NGF protein levels (measured by 2-site ELISA) increased to 150% in DG at 7 days, 260% in PC at 12 h, and 170% in parietal cortex at 24 h

Basal levels of trkA mRNA in the hippocampus are very low (Cellerino, 1996; Mudo et al., 1996), and do not appear to increase following kindling (Bengzon et al., 1993; Merlio et al., 1993) or pilocarpine status epilepticus (Mudo et al., 1996) (Table 2).

Brain-derived neurotrophic factor (BDNF) In 1982, BDNF, the second member of the ‘‘NT’’ family of neurotrophic factors, was shown to promote survival of a subpopulation of dorsal root ganglion neurons, and subsequently purified from pig brain (Barde et al., 1982). The amino acid sequence of BDNF was found to have a strong homology with NGF. Since then, other members of the NT family such as NT-3 (Maisonpierre et al., 1990b) and NT-4/5 (Hallbook et al., 1991; Ip et al., 1992) have been described, each with a distinct profile of trophic effects on subpopulations of neurons in the peripheral and CNS. BDNF mRNA has a widespread distribution in the CNS (Merlio et al., 1993; Conner et al., 1997),

374 Table 2. Seizure regulation of trk receptor expression Reference

Methods

Results

TrkA mRNA Cellerino (1996)

Basal hippocampal

Basal not detectable by 35S; can only detect with 33P Basal not detectable and pilocarpine status did not increase levels No change No change

Mudo et al. (1996)

Pilocarpine

Bengzon et al. (1993) Merlio et al. (1993)

Hippocampal kindling Rapid hippocampal kindling

TrkB mRNA Bengzon et al. (1993) Merlio et al. (1993) Elmer et al. (1996a, b) Humpel et al. (1993) Nibuya et al. (1995) Schmidt-Kastner et al. (1996) Mudo et al. (1996) Hughes et al. (1998) TrkC mRNA Bengzon et al. (1993) Merlio et al. (1993) Mudo et al. (1995)

Rapid hippocampal kindling (ventral hippocampal stimulation) Rapid hippocampal kindling (ventral hippocampal stimulation) Rapid hippocampal kindling (ventral hippocampal stimulation) PTZ ECS Pilocarpine Pilocarpine Hippocampal afterdischarge Rapid hippocampal kindling (ventral hippocampal stimulation) Rapid hippocampal kindling (ventral hippocampal stimulation) ICV KA or ICV bicuculline

including limbic forebrain, neocortex, and is abundant in all principal neurons of the hippocampus (Ernfors et al., 1990; Hofer et al., 1990; Wetmore et al., 1990). Like BDNF mRNA, constitutive BDNF protein expression is widespread (Conner et al., 1997; Yan et al., 1997b), localized on neuronal cell bodies, axons, and dendrites. The mossyfiber axons of hippocampal dentate granule cells display especially intense BDNF immunoreactivity (Conner et al., 1997). The principal receptor for BDNF, trkB, is a receptor tyrosine kinase, which is found in both catalytic and truncated forms in the adult forebrain (Fryer et al., 1996; Drake et al., 1999). TrkB mRNA and protein are found in hippocampus (Merlio et al., 1992; Altar et al., 1994; Yan et al., 1997a). Truncated trkB is also found in the ependymal cells lining the ventricular cavities, effectively limiting diffusion of intraventricularly administered BDNF (Yan et al., 1994; Anderson et al., 1995).

Increased in DG 2 h (2  ) after focal or generalized seizures Increased threefold in DG at 30 min after 40 stimulations Increased in DG max at 2 h Increased twofold in DG at 3 h Increased fivefold (DG) at 2 h Increase in DG, CA1–3 at 3–6 h Increase in DG, amygdala, PC at 3 h Increase in DG at 4 h Increased in DG 2 h (1.5  ) after focal or generalized seizures No change after 40 rapid K stimulations Increase max at 3 h (bicuculline) or 12 h (KA) confined to DG

Localization, transport and release Unlike the classical target-derived trophic factor model in which NTs — such as NGF — are retrogradely transported, there is now abundant evidence that BDNF is also anterogradely transported in brain. First, BDNF protein is localized to nerve terminals (Conner et al., 1997), and pathway transection or axonal transport inhibition abrogates this terminal expression (Altar et al., 1997; Conner et al., 1997; Altar and DiStefano, 1998). Second, higher resolution studies have shown that BDNF is associated with dense-core vesicles (Fawcett et al., 1997; Altar and DiStefano, 1998), which are the primary site for neuropeptide storage and release from nerve terminals. Third, further functional studies have supported the anterograde transport hypothesis (Fawcett et al., 1998, 2000). Fourth, pro-BDNF is shuttled from the trans-Golgi network into secretory granules,

375

where it is cleaved by prohormone convertase 1 (PC1) (Farhadi et al., 2000). In addition, emerging evidence suggests that both BDNF and trk receptors may undergo regulated intracellular transport. For example, seizures lead to redistribution of BDNF mRNA from hippocampal CA3 cell bodies to their apical dendrites (Bregola et al., 2000; Simonato et al., 2002). Trk signaling is now thought to include retrograde transport of intact NT-trk complexes to the neuronal cell body (Miller and Kaplan, 2001; Ginty and Segal, 2002). Recent evidence indicates that NTs are released acutely following neuronal depolarization (Griesbeck et al., 1999; Mowla et al., 1999; Hartmann et al., 2001; Egan et al., 2003; Goggi et al., 2003; Brigadski et al., 2005). In fact, direct activitydependent pre- to postsynaptic transneuronal transfer of BDNF has been demonstrated using fluorescently labeled BDNF (Kohara et al., 2001). The released form of BDNF is thought to be proBDNF (Mowla et al., 2001), raising the possibility of postsecretory proteolytic processing by membrane-associated or extracellular proteases in the modulation of BDNF action (Lee et al., 2001b).

BDNF gene regulation A multitude of stimuli have been described that alter BDNF gene expression in both physiologic and pathologic states (Lindholm et al., 1994). Physiologic stimuli are known to increase BDNF mRNA content. For example, light stimulation increases BDNF mRNA in visual cortex (Castre´n et al., 1992), osmotic stimulation increases BDNF mRNA in the hypothalamus (Castre´n et al., 1995; Dias et al., 2003), and whisker stimulation increases BDNF mRNA expression in somatosensory barrel cortex (Rocamora et al., 1996). Electrical stimuli that induce long-term potentiation (LTP) in the hippocampus, a cellular model of learning and memory, increase BDNF and NGF expression (Patterson et al., 1992; Castre´n et al., 1993; Bramham et al., 1996). Even physical exercise has been shown to increase NGF and BDNF expression in hippocampus (Neeper et al., 1995). Interestingly, BDNF levels vary across the estrous

cycle, which correlate with its effects on neural excitability (Scharfman et al., 2003). Distinct BDNF 50 exons are differentially regulated by stimuli such as neural activity. For example, exons I–III, but not exon IV, increase after kainic acid-induced seizures (Timmusk et al., 1993b) or other stimuli that increase activity (Lauterborn et al., 1996; Tao et al., 2002). Protein synthesis is required for the effects of activity on exon I and II, but not III and IV, raising the possibility that the latter act as immediate early genes (Lauterborn et al., 1996; Castre´n et al., 1998). The transcription factor CaRF (calcium response factor) activates transcription of exon III under the control of a calcium response element, CaRE1 (Tao et al., 2002). CREB, which can be stimulated by diverse stimuli ranging from activity to chronic antidepressant treatment (Nibuya et al., 1995, 1996; Shieh et al., 1998; Tao et al., 1998; Shieh and Ghosh, 1999), also modulates exon III transcription. Recent evidence also indicates that neural activity triggers calcium-dependent phosphorylation and release of MeCP2 (methyl-CpG binding protein 2) from BDNF promoter III to derepress transcription (Chen et al., 2003). Pathologic states are also associated with alteration in BDNF gene expression. For example, seizures dramatically upregulate BDNF mRNA (Table 3). A wide variety of seizure paradigms (kindling; kainic acid; pilocarpine; pentylenetetrazol, PTZ; electroconvulsive shock, ECS) rapidly and dramatically increase expression of BDNF mRNA in dentate gyrus as well as in other areas of the hippocampus and cortex (Ernfors et al., 1991; Isackson et al., 1991; Gall and Lauterborn, 1992; DugichDjordjevic et al., 1992a, b; Bengzon et al., 1993; Humpel et al., 1993; Nibuya et al., 1995; Mudo et al., 1996; Sato et al., 1996; Schmidt-Kastner et al., 1996). This is associated with a transient upregulation of BDNF protein (Nawa et al., 1995; Elmer et al., 1996b; Hughes et al., 1998). TrkB mRNA and protein in the dentate gyrus are also upregulated following various seizure protocols (Bengzon et al., 1993; Merlio et al., 1993; Elmer et al., 1996a) (Table 2). TrkB mRNA expression is increased in dentate granule cells 2–6 h after rapid electrical kindling, hippocampal after discharge, PTZ kindling, ECS, or pilocarpine

376 Table 3. Seizure regulation of BDNF expression Reference

Methods

Results Increase, onset o1.5 h, max at 6 h (12  ) Increase, onset 20 min, max at 4 h (12  ) Increase max at 30 min (DG/PC), 1 h (CA1)

Dugich-Djordjevic et al. (1992b)

Hilar electrolytic lesion Perforant path stimulation (1 AD) Rapid kindling (ventral hippocampal stimulation) KA

Dugich-Djordjevic et al. (1992a)

KA

Bengzon et al. (1993)

Traditional kindling (ventral hippocampal stimulation CA1–2)

Humpel et al. (1993)

Nibuya et al. (1995)

PTZ kindling (30 mg/kg i.p. followed by convulsive dose (50 mg/kg)) ECS

Schmidt-Kastner et al. (1996)

Pilocarpine

Mudo et al. (1996) Sato et al. (1996)

Pilocarpine Traditional amygdala kindling

BDNF mRNA Isackson et al. (1991) Gall and Lauterborn (1992) Ernfors et al. (1991)

BDNF protein Nawa et al. (1995)

Hilar electrolytic lesion

BDNF ELISA

Elmer et al. (1996a, b)

Rapid kindling (ventral hippocampal stimulation) BDNF ELISA

Hughes et al. (1998)

Hippocampal after discharge

status epilepticus (Bengzon et al., 1993; Humpel et al., 1993; Nibuya et al., 1995; Elmer et al., 1996a; Mudo et al., 1996; Schmidt-Kastner et al., 1996; Hughes et al., 1998). Subcellular studies have demonstrated targeting of BDNF and trkB mRNAs to dendrites in CA3 neurons following kindled seizures (Simonato et al., 2002).

Increase in all hippocampus (P21, P40), no change despite seizures (P8) Increase max at 30 min (10  in DG, 2–6  in CA1, CA3, CA4); later in cortex Increase — did not do time course but studied relationship between development of kindling and neurotrophin induction — found similar BDNF induction (approximately 9  at 2 h time point) regardless of kindling stage Increase max at 3 h (DG, PC, amygdala) after acute convulsive PTZ Increase 30-fold 2 h after ECS (DG), fivefold (PC) Increase max at 3–6 h (DG, other hippocampal, neocortex, PC, striatum, thalamus) Increase max at 3–6 h (DG, amygdala, PC) 4–5-fold increase in DG 1 h after stage 5 kindled seizure Highest levels of basal BDNF in hippocampus (followed by hypothalamus, neocortex, cerebellum, thalamus and striatum) Fourfold induction of BDNF protein levels in hippocampus, maximum at 24 h after HL, down by 1 week Basal levels of BDNF highest in DGCA3>CA1>PC After 1 AD: increase to 150% in DG at 6 h, 200% in CA3 at 12 h, 50% in PC at 6 h After 40 ADs: increase to 200% in DG at 6 and 24 h, 150% in CA3 at 6 h, 300% in PC at 2 and 6h After 7 ADs: increase to 200% in CA3 at 24 h Increase in DG at 4 h

Role(s) of BDNF during development In vitro and in vivo studies have demonstrated that BDNF has survival- and growth-promoting actions on a variety of CNS neurons, including dorsal root ganglion cells, dopaminergic and cholinergic neurons, retinal ganglion cells, and

377

hippocampal and cortical neurons (Johnson et al., 1986; Alderson et al., 1990; Hyman et al., 1991; Knusel et al., 1991; Acheson et al., 1995; Patel and McNamara, 1995; Huang and Reichardt, 2001). Certain peripheral sensory neurons, especially those in vestibular and nodose-petrosal ganglia, depend on the presence of BDNF because BDNF homozygous knockout (BDNF/) mice lack these neurons (Huang and Reichardt, 2001). Unlike NGF, sympathetic neurons are not affected, nor are motor neurons. BDNF/ mice fail to thrive, demonstrate lack of proper coordination of movement and balance, and ultimately die by 3 weeks of age. However, heterozygous BDNF knockout (BDNF+/) mice are viable, and exhibit a variety of phenotypes, including obesity (Lyons et al., 1999; Kernie et al., 2000), decreased seizure susceptibility (Kokaia et al., 1995), and impaired spatial learning (Linnarsson et al., 1997). Interestingly, conditional postnatal BDNF gene deletion (Rios et al., 2001) and reduction in trkB expression (Xu et al., 2003) also cause obesity. Physiologic regulation of BDNF gene expression may be very important in the development of the brain. For example, BDNF contributes to activity-dependent development of the visual cortex. Provision of excess BDNF (Cabelli et al., 1995) or blockade of BDNF signaling (Cabelli et al., 1997) leads to abnormal patterning of ocular dominance columns during a critical period of visual cortex development. This suggests a role for BDNF in axonal pathfinding during development. BDNF also has powerful effects on dendritic morphology (McAllister et al., 1997; Murphy et al., 1998; Horch and Katz, 2002; Tolwani et al., 2002).

Effects on synaptic transmission BDNF has an enormous range of physiologic actions at both developing and mature synapses, overall enhancing synaptic transmission by both pre- and postsynaptic mechanisms. The first studies of BDNF effects on synaptic transmission showed that BDNF increased the frequency of miniature excitatory postsynaptic currents (EPSCs) at Xenopus neuromuscular synapses (Lohof et al., 1993). Since then, numerous studies have examined the actions of

BDNF. Overall, BDNF appears to strengthen excitatory (glutamatergic) synapses and weaken inhibitory (GABAergic) synapses. Schuman and colleagues demonstrated that exposure of adult rat hippocampal slices to BDNF led to a long-lasting potentiation of synaptic strength at Schaffer collateral-CA1 synapses (Kang and Schuman, 1995). Subsequent studies have supported a role of BDNF in LTP (Korte et al., 1995; Korte et al., 1996; Patterson et al., 1996; Kang et al., 1997; Xu et al., 2000). For example, incubation of hippocampal or visual cortical slices with trkB inhibitors inhibits LTP (Figurov et al., 1996), and hippocampal slices from BDNF/ mice exhibit impaired LTP induction (Korte et al., 1995), which is restored by reintroduction of BDNF (Korte et al., 1996; Patterson et al., 1996). Whether BDNF-induced synaptic potentiation occurs primarily by a presynaptic action (e.g. through enhancement of glutamate release) or postsynaptically (e.g. via phosphorylation of neurotransmitter receptors) is intensely debated (Schinder and Poo, 2000). A number of studies have provided evidence for a presynaptic locus (Xu et al., 2000; Tyler et al., 2002; see also Kafitz et al., 1999), yet evidence for postsynaptic actions has also been obtained (Black, 1999; ThakkerVaria et al., 2001; reviewed in Poo, 2001). Both pre- and postsynaptic trkB receptors in the hippocampus may be important (Drake et al., 1999). A role for BDNF in GABAergic synapses was first raised by studies showing that BDNF influences GABAergic neuronal phenotype (Marty et al., 1996; McLean Bolton et al., 2000). Subsequently, BDNF was shown to decrease inhibitory (GABAergic) synaptic transmission (Tanaka et al., 1997; Frerking et al., 1998; Wardle and Poo, 2003). Recent evidence shows that BDNF can modulate the function of GABAA receptors via modulation of phosphorylation state (Jovanovic et al., 2004). Interestingly, BDNF may also regulate the efficacy of GABAergic synapses by direct downregulation of the neuronal K+-Cl cotransporter, which would impair neuronal Cl extrusion and weaken GABAergic inhibition (Rivera et al., 2002). Similarly, a recent paper found differential effects of BDNF on GABA-mediated currents in excitatory and inhibitory neuron subpopulations, selectively decreasing the efficacy of inhibitory neurotransmission

378

by downregulation of Cl transport (Wardle and Poo, 2003). Effect on neurogenesis An important feature of the dentate gyrus is the lifelong production of new granule cells from progenitor cells located in the subgranular zone (Gould and McEwen, 1993; Scharfman, 2004). BDNF has also been found to enhance neurogenesis. Intraventricular infusion of BDNF or adenoviral-induced BDNF activity increases the number of neurons in the adult olfactory bulb, striatum, septum, and thalamus (Zigova et al., 1998; Benraiss et al., 2001; Pencea et al., 2001), which can be potentiated by concurrent inhibition of glial differentiation of subependymal progenitor cells (Chmielnicki et al., 2004). Intrahippocampal infusion of BDNF into adult rats leads to increased dentate gyrus neurogenesis, accompanied by increased numbers of ectopic granule cells (Scharfman et al., 2005). BDNF+/ mice show decreased numbers of BrdU-labeled cells in the dentate gyrus (Lee et al., 2002). Studies of cultured progenitor cells have elucidated some of the signaling mechanisms, which appear to involve trkB activation, followed by activation of the MAP kinase and PI3kinase pathways (Barnabe-Heider and Miller, 2003) and downstream modification of basic helix-loop-helix transcription factors (Ito et al., 2003). Although some studies have concluded that the primary effect of BDNF is on proliferation (Katoh-Semba et al., 2002), other experiments suggest an important effect on survival (Lee et al., 2002). The effects of BDNF may depend on a previous history of ischemic damage (Larsson et al., 2002; Gustafsson et al., 2003). Effects on learning and memory Learning and memory depend on persistent selective modification of synapses between CNS neurons. Since BDNF appears to be involved in activitydependent synaptic plasticity, there is great interest in its role in learning and memory (Yamada and Nabeshima, 2003). The hippocampus, which is required for many forms of long-term memory in

humans and animals, appears to be an important site of BDNF action. Rapid and selective induction of BDNF expression in the hippocampus during contextual learning has been demonstrated (Hall et al., 2000), and function-blocking antibodies to BDNF (Alonso et al., 2002), BDNF knockout (Linnarsson et al., 1997), knockout of forebrain trkB signaling (Minichiello et al., 1999), or overexpression of truncated trkB (Saarelainen et al., 2000b) in mice impairs spatial learning. Another study demonstrated upregulation of BDNF in monkey parietal cortex associated with tool-use learning (Ishibashi et al., 2002). In humans, a valine to methionine polymorphism at the 50 pro-region of the human BDNF protein was found to be associated with poorer episodic memory; in vitro, neurons transfected with met-BDNF-green fluorescence protein (GFP) exhibited reduced depolarization-induced BDNF secretion (Egan et al., 2003). Neurotrophin-3 (NT-3) NT-3, first described in 1990 (Maisonpierre et al., 1990b), is similar to BDNF in several ways. Like BDNF, NT-3 mRNA and protein are widely distributed in the adult CNS (Maisonpierre et al., 1990a, b, Zhou and Rush, 1994; Katoh-Semba et al., 1996). While the preferred receptor for NT-3 is trkC, NT-3 can also bind to trkA and trkB (Barbacid, 1994; Ryden and Ibanez, 1996; Huang and Reichardt, 2003). Like BDNF, NT-3 is involved in synaptic transmission and neuronal excitability (Thoenen, 1995). Addition of NT-3 to hippocampal slices enhances synaptic strength at Schaffer collateral-CA1 synapses (Kang and Schuman, 1995). NT-3 enhances paired-pulse facilitation in the perforant path-dentate gyrus pathway (Kokaia et al., 1998; Asztely et al., 2000). Like BDNF, NT-3 reduces GABAergic inhibition (Kim et al., 1994). Also like BDNF, NT-3 enhances the survival and differentiation of neural progenitor cells (Barnabe-Heider and Miller, 2003). NT-3 gene regulation However, whereas NGF and BDNF levels increase after seizures, NT-3 levels are reduced in dentate

379 Table 4. Seizure regulation of NT-3 and NT-4 expression Reference NT-3 mRNA Gall and Lauterborn (1992) Bengzon et al. (1993)

Schmidt-Kastner and Olson (1995) Mudo et al. (1996) Kim et al. (1998) NT-4 mRNA Timmusk et al. (1993a, b)

Mudo et al. (1996)

Methods

Results

Hilar electrolytic lesion Traditional kindling (ventral hippocampal stimulation CA1-CA2)

Decrease, onset 12 h max 12 h (20%) in DG Decrease — did not do time course but studied relationship between development of kindling and neurotrophin induction — found similar NT3 decrease (
50%) regardless of kindling stage Decrease max at 3 h Decrease max at 12–24 h (40–50%) Decrease in DG

Pilocarpine Pilocarpine Amygdala kindling RNAse protection analysis from different tissues

Pilocarpine

gyrus granule neurons (Gall and Lauterborn, 1992; Bengzon et al., 1993; Schmidt-Kastner and Olson, 1995; Mudo et al., 1996; Kim et al., 1998) (Table 4). This suggests that the potential role of NT-3 in seizure progression is different. Whether trkC is elevated appears to depend on the model used; rapid kindling induces no change (Merlio et al., 1993) or a transient increase (Bengzon et al., 1993) in trkC mRNA levels, and ICV KA or ICV bicuculline transiently increase trkC mRNA in dentate granule cells (Mudo et al., 1995).

Neurotrophin-4/5 The fourth member of the NT family, NT-4/5, was discovered after NGF, BDNF, and NT-3 (Hallbook et al., 1991; Ip et al., 1992). Levels of NT-4/5 in the brain are very low at baseline (Timmusk et al., 1993a; Katoh-Semba et al., 2003) and are not increased by seizures (Timmusk et al., 1993a; Mudo et al., 1996) (Table 4). NT-4/5/ mice, unlike BDNF/ mice, are normal and long-lived with no obvious neurological deficits (Conover et al., 1995; Liu et al., 1995). The only loss of neurons in NT-4/5/ mice appears to be a reduction in the number of sensory neurons in the nodose-petrosal and geniculate ganglia (Conover et al., 1995; Liu et al., 1995). Provision of NT-4/

Could not detect NT-4 by in situ but did find very low levels in adult brain using RNAse protection No increase in NT-4 mRNA following either systemic KA or hippocampal stimulation Basal not detectable and pilocarpine status did not increase levels

5 protects hippocampal and cortical neurons against energy deprivation-induced injury (Cheng et al., 1994) and adrenalectomy-induced apoptosis of hippocampal granule cells (Qiao et al., 1996). Application of NT-4/5 enhances excitatory synaptic transmission in cultured hippocampal neurons (Lessmann et al., 1994). Roles of neurotrophins in epilepsy models The discovery that limbic seizures increase NGF mRNA levels (Gall and Isackson, 1989) led to the idea that seizure-induced expression of neurotrophic factors may contribute to the lasting structural and functional changes underlying epileptogenesis (Gall et al., 1991, 1997; Jankowsky and Patterson, 2001). NGF and epilepsy Is there a functional role for NGF gene upregulation in epileptogenesis? Indeed, intraventricular administration of NGF antibodies retards amygdala kindling (Funabashi et al., 1988) and blocks kindling-induced mossy-fiber sprouting (Van der Zee et al., 1995). Similarly, an NGF inhibitory peptide inhibits amygdala kindling and mossy-fiber sprouting (Rashid et al., 1995). Conversely, intraventricular NGF infusion was found to facilitate amygdala

380

and hippocampal kindling and increase mossy-fiber sprouting (Adams et al., 1997). Whether NGF exerts its effects on kindling and kindling-induced morphological changes via trkA or p75NTR has been investigated. Inhibition of Ras, a downstream effector of trkA, inhibits kindling and kindling-associated mossy-fiber sprouting (Li et al., 2003). Peptide inhibitors of NGF binding to trkA but not to p75NTR can inhibit kindling, whereas both trkA and p75NTR inhibition can inhibit mossy-fiber sprouting (Li et al., 2005). What is the locus of effects of NGF on hippocampal kindling? As described above, there is little evidence for trkA expression in hippocampus (Sobreviela et al., 1994; Cellerino, 1996). Similarly, there is little expression of p75 in hippocampus at baseline (Pioro and Cuello, 1990; Sobreviela et al., 1994). It is likely that NGF-dependent effects are due to modulation of the cholinergic system, as both trkA and p75NTR (Hofer et al., 1990) receptors are most strongly expressed in the basal forebrain cholinergic neurons which project to hippocampus (Sobreviela et al., 1994). Consistent with this hypothesis is that cholinergic agonists and antagonists produce effects on kindling and sprouting parallel to those of NGF (Adams et al., 2002).

BDNF and epilepsy Abundant in vitro and in vivo evidence implicates BDNF in the cascade of electrophysiologic and behavioral changes underlying the epileptic state (Binder et al., 2001). BDNF mRNA and protein are markedly upregulated in the hippocampus by seizure activity in animal models (Ernfors et al., 1991; Isackson et al., 1991; Lindvall et al., 1994; Nibuya et al., 1995). Infusion of trkB receptor body (a chimera of human IgG-Fc domain and the extracellular domain of the trkB receptor) (Binder et al., 1999b) or use of BDNF+/ (Kokaia et al., 1995) or truncated trkB-overexpressing (Lahteinen et al., 2002) mice inhibits epileptogenesis in animal models. Conversely, direct application of BDNF induces hyperexcitability in vitro (Scharfman, 1997; Scharfman et al., 1999), overexpression of

BDNF in transgenic mice leads to spontaneous seizures (Croll et al., 1999), and intrahippocampal infusion of BDNF is sufficient to induce seizure activity in vivo (Scharfman et al., 2002). A separate group of experiments has demonstrated that chronic BDNF infusion can inhibit kindling (Larmet et al., 1995; Osehobo et al., 1996; Reibel et al., 2000b). These inhibitory effects appear to be due to trkB receptor downregulation following chronic BDNF administration, and hence are still consistent with the ‘‘proepileptogenic BDNF’’ hypothesis. This interpretation is supported by the observation that chronic exposure to BDNF in vitro leads to downregulation of trkB mRNA and protein (Knusel et al., 1997). Similarly, continuous in vivo intrahippocampal BDNF infusion results in downregulation of trkB protein by as much as 80% (Frank et al., 1996). Thus, whereas chronic BDNF infusion inhibits kindling progression, acute microinjections of BDNF enhance epileptogenesis in the absence of effect on trkB expression (Xu et al., 2004). Furthermore, chronic infusions of BDNF may upregulate the inhibitory neuropeptide Y (NPY) (Reibel et al., 2000a). Whether BDNF has a significant effect on seizure-associated mossy-fiber sprouting is not clear. While mossy-fiber sprouting has been reported in BDNF+/ mice and following BDNF infusion (Kokaia et al., 1995; Scharfman et al., 2002), there is no effect on mossy-fiber sprouting in BDNFoverexpressing mice or following chronic infusion or bolus injection of BDNF in other studies (Qiao et al., 2001; Xu et al., 2004). However, BDNF overexpression does increase dendritic length and complexity in the hippocampus (Tolwani et al., 2002). The relative role of BDNF on effect synaptic changes vs. larger scale morphological changes during epileptogenesis remains to be clarified. The anatomic locus of action of NTs during epileptogenesis has been clarified with the study of trk receptor activation (see below).

NT-3 and epilepsy In comparison with BDNF, what is the evidence for a role of NT-3 in epileptogenesis? In NT-3+/

381

mice, which have
30% reduction in basal NT-3 mRNA levels, amygdala kindling was markedly retarded (Elmer et al., 1997). However, compensatory changes in BDNF and trkB mRNA levels in these mice made these data difficult to interpret (Elmer et al., 1997). Chronic intraventricular infusion of NT-3 retards the development of behavioral seizures (Xu et al., 2002), probably in part via downregulation of trk phosphorylation (Xu et al., 2002). What about the effects of NT-3 on kindlinginduced mossy-fiber sprouting in the dentate gyrus? Chronic infusion of NT-3 inhibits kindling-associated mossy-fiber sprouting (Xu et al., 2002). However, this effect is unclear as infusion of NT-3 in the absence of kindling actually enhances sprouting of mossy fibers in the inner molecular layer of the dentate gyrus and CA3 stratum oriens (Xu et al., 2002). NT-4 and epilepsy Unlike NGF, BDNF, and NT-3 levels, levels of NT-4/5 do not appear to be regulated by seizure activity (Timmusk et al., 1993a; Mudo et al., 1996). The amygdala kindling phenotype of NT-4/ 5/ mice was studied (He et al., 2006). No aspect of the development or persistence of amygdala kindling was different between NT-4/5/ and wild-type mice (He et al., 2006). Trk receptor activation following seizure activity The ability to monitor trk receptor activation following seizures using phospho-specific trk antibodies enabled identification of the anatomy, time course, and threshold characteristics of trk receptor activation in the hippocampus following seizure activity (Binder et al., 1999a). Kainate-induced status epilepticus or hippocampal electrographic seizures increase phospho-trk immunoreactivity selectively in the hippocampus, primarily confined to the dentate hilus and CA3 stratum lucidum. This seizureinduced phospho-trk immunoreactivity is marked but transient, maximal at 24–48 h but back to baseline by 1 week. The seizure duration threshold for increase in phospho-trk immunoreactivity appears to

correspond to the previously reported threshold for increase in BDNF gene expression. These observations are examined in greater detail in the next few sections. Anatomy of seizure-induced phospho-trk immunoreactivity Following seizure activity, phospho-trk immunoreactivity is selectively increased in dentate hilus and CA3 stratum lucidum of hippocampus (Binder et al., 1999a). This distribution precisely coincides with the ‘‘mossy fiber’’ pathway of dentate granule cell axon terminals. In addition, this anatomic pattern coincides with the distribution of both basal and seizure-induced BDNF protein. Basal BDNF protein is also localized in hilus and CA3 stratum lucidum (Conner et al., 1997), and seizures increase levels of BDNF protein in dentate gyrus and CA3 (Elmer et al., 1998) and BDNF immunoreactivity in hilus and CA3 stratum lucidum (Smith et al., 1997; Yan et al., 1997b; Rudge et al., 1998; Vezzani et al., 1999). This precise anatomic colocalization of increased phospho-trk immunoreactivity and increases in BDNF protein suggests that the phospho-trk immunoreactivity is caused by seizure-induced increases in BDNF. BDNF, but not NGF, is known to increase levels of NPY (Croll et al., 1994), and kindling and kainate-induced seizures increase NPY immunoreactivity in hilus and CA3 stratum lucidum (Marksteiner et al., 1990; Tønder et al., 1994), further implicating seizure-induced BDNF acting in the mossy-fiber pathway. While NGF mRNA content is upregulated by seizures, the anatomic distribution of increased NGF protein is not known. Thus, these anatomic considerations are most consistent with a role for BDNF. Time course of seizure-induced phospho-trk immunoreactivity The time course of known BDNF upregulation following seizures coincides temporally with increased phospho-trk immunoreactivity. Using hippocampal microdissection and a two-site ELISA for BDNF, Elmer et al. showed that after seven

382

ventral hippocampal electrographic seizures, the maximum increase in BDNF protein occurs at 12 h in dentate gyrus and 24 h in CA3 (Elmer et al., 1998). Similarly, maximum increases in BDNF protein following hilus lesion-induced (Nawa et al., 1995) or kainate-induced (Rudge et al., 1998) seizures occur at approximately 24 h in hippocampus. Importantly, BDNF protein levels in both of these studies returned to baseline after 1 week, similar to phospho-trk immunoreactivity. In contrast, Bengzon et al. found maximal NGF protein content (measured by two-site immunoassay) 7 days after a similar rapid kindling protocol (Bengzon et al., 1992) and did not see NGF protein increases at earlier time points. Similarly, Lowenstein et al. found maximal NGF-like neurotrophic activity of hippocampal extracts from animals 1 week after KA treatment (Lowenstein et al., 1993). Thus, the time-course data favor a role for BDNF rather than NGF in seizure-induced phospho-trk immunoreactivity.

Seizure duration threshold for increased phosphotrk immunoreactivity The seizure duration threshold for increase in phospho-trk immunoreactivity further supports a role for BDNF. Consistently, increased phosphotrk immunoreactivity was observed only in hippocampal kindled animals with ESD
70 s (Binder et al., unpublished data). In a similar ventral hippocampal stimulation protocol, Bengzon et al. observed increases in BDNF mRNA content in dentate granule cells in an all-or-none manner above an electrographic seizure duration of approximately 70 s (Bengzon et al., 1993). Like the increases in mRNA content, increases in phosphotrk immunoreactivity appeared to be ‘‘allor-none’’ as no differences were noted in intensity of immunoreactivity between kainate-treated and 7 hippocampal ES-treated animals despite marked differences in seizure duration (hours for kainate vs. seconds for 7 hippocampal ESs) (Binder et al., unpublished data). This strong similarity between thresholds as well as all-or-none characteristics suggests that such prior increases in BDNF

mRNA content may not only be necessary for any increase in phospho-trk immunoreactivity but also sufficient for maximal increase in phospho-trk immunoreactivity following seizures.

Evidence that the trk receptor activated by seizures is trkB Indirect evidence suggests that BDNF-induced trkB activation is responsible for the increased phospho-trk immunoreactivity following seizures. First, the mRNA content of NGF and BDNF is increased following seizures (Ernfors et al., 1991; Isackson et al., 1991; Lindvall et al., 1994; Nibuya et al., 1995) whereas dentate granule NT-3 mRNA content is decreased (Gall et al., 1991; Gall and Lauterborn, 1992; Bengzon et al., 1993; SchmidtKastner and Olson, 1995; Mudo et al., 1996). Second, protein levels of NGF and BDNF increase after seizure activity (Bengzon et al., 1992; Elmer et al., 1998). Third, the time-course data described above implicate BDNF rather than NGF. Fourth, mRNA levels of the other NT known to activate trkB, NT-4, are very low in adult brain (Timmusk et al., 1993a) and do not increase after seizures (Mudo et al., 1996). Fifth, unlike trkB and trkC, levels of expression of trkA in hippocampus are barely detectable (Barbacid, 1994; Cellerino, 1996), suggesting that trkA is unlikely to mediate seizure-induced increases in phospho-trk immunoreactivity. In order to more directly analyze the role of the trkB receptor in seizure-induced trk receptor activation, He et al. studied trk receptor phosphorylation in a mouse mutant with a single point mutation at the shc site (Y490 in humans, Y515 in mice) of the trkB receptor (He et al., 2002). Homozygous trkBshc/shc (Y515F) mice were generated by Minichiello et al. and interestingly display loss of NT-4-dependent neurons but have no major effects on BDNF responses (Minichiello et al., 1999). He et al. found that following amygdala kindling stimulation, phospho-trk immunoreactivity is increased in wild-type mice in a similar pattern (hilus and CA3 stratum lucidum) to that seen in the rat experiments (described above). The

383

trkBshc/shc homozygous mice displayed absence of seizure-induced phospho-trk immunoreactivity, and the heterozygotes displayed intermediate immunoreactivity (He et al., 2002). These experiments suggest that the trk receptor activated during kindling stimulation is indeed trkB. Interestingly, the Y515F point mutation had no effect on kindling development in the same study (He et al., 2002). This is remarkably consistent with the lack of effect of this mutation on synaptic LTP (Korte et al., 2000). More recently, this group has generated a distinct mouse with a point mutation at the PLC site. Unlike trkBshc/shc mice, trkBPLC/PLC mice exhibit impaired LTP (Minichiello et al., 2002). This direct comparison of distinct trkB tyrosine mutants implicates the PLC signaling pathway as opposed to the MAPK pathway in trkB activation-induced synaptic plasticity. Similarly, other studies have shown that specific stimuli may cause tyrosine-specific phosphorylation of the trkB receptor (i.e. at other tyrosines but not at the shc site). For example, Saarelainen et al., in studying the role of endogenous BDNF and trkB signaling in the mechanism of action of antidepressant drugs, found that acute and chronic antidepressant treatment caused trkB receptor phosphorylation and activation, but the pY674/5 site was selectively phosphorylated compared to the pY490 (shc) site (Saarelainen et al., 2003). The further development of phosphorylation statespecific antibodies to distinct tyrosines (pY674/ 5, pY785) may prove to be of use in dissecting tyrosine site-specific trkB signal transduction in vivo in a variety of paradigms. Furthermore, these results can be compared with antibodies that recognize activated intracellular signaling pathways (e.g. phosphoCREB) (Finkbeiner et al., 1997).

Cellular site of seizure-induced phospho-trk immunoreactivity What is the likely cellular site of seizure-induced phospho-trk immunoreactivity? The light microscopic distribution of phospho-trk immunoreactivity

after seizure (dentate hilus and CA3 stratum lucidum of hippocampus) corresponds to the mossy-fiber pathway of dentate granule cell axon terminals (Binder et al., 1999a). This suggests that the cellular site of phospho-trk immunoreactivity is either on mossy-fiber axons and/or targets. Localization on mossy-fiber axons represents a parsimonious explanation for both hilar and CA3 stratum lucidum immunoreactivity. In contrast, localization on targets requires immunoreactivity on both targets in hilus (hilar interneurons) and in CA3 stratum lucidum (pyramidal cell dendrites and/or stratum lucidum interneurons). Anatomic consideration of trkB-like immunoreactivity may lend insight into the likely cellular site of phospho-trk immunoreactivity. In some published experiments, an affinity-purified antibody directed against an extracellular trkB peptide sequence was used, which does not distinguish between full-length and truncated (Barbacid, 1994) trkB receptors. The earlier studies (using light microscopy) demonstrated that trkB-like immunoreactivity is preferentially distributed on cell bodies and dendrites of both cortical and hippocampal neurons (Fryer et al., 1996; Yan et al., 1997a). Pyramidal neurons in hippocampus in particular demonstrate marked trkB immunoreactivity on cell bodies and dendrites in comparison with axons (Fryer et al., 1996; Yan et al., 1997a). These studies utilized an antibody raised against the extracellular portion of trkB (trkB2336) common to both full-length and truncated forms. A more recent and comprehensive study of cellular and subcellular localization of trkB immunoreactivity was carried out by Drake et al. (1999). These investigators used acytoplasmic-domain antibody (trkB-in) to selectively label the full-length form of trkB and carried out both light and electron microscopic analysis. Their conclusion was that full-length trkB immunoreactivity exists in glutamatergic granule and pyramidal cells and was most intense in axons, axon terminals, and dendritic spines and to a lesser extent in somata and dendritic shafts. Occasionally, interneurons were also labeled. Thus, phospho-trkB immunoreactivity could represent pre- and/or postsynaptic activation of trkB receptors in the mossy-fiber pathway.

384

Potential models for induction of phospho-trk immunoreactivity by seizure activity Throughout the brain, BDNF immunoreactivity appears to be preferentially localized in cell bodies and axons compared to dendrites (Conner et al., 1997). In addition, unlike the classical target-derived trophic factor model in which NTs are retrogradely transported, abundant recent evidence suggests that CNS BDNF appears to be anterogradely transported (Von Bartheld et al., 1996a; Zhou and Rush, 1996; Altar et al., 1997; Conner et al., 1997; Fawcett et al., 1998; Tonra et al., 1998). This evidence, together with the anatomic distribution of BDNF immunoreactivity in hippocampus in a mossy fiber-like pattern, suggests that BDNF protein in hilus and CA3 stratum lucidum was synthesized in granule cell bodies and anterogradely transported to mossy-fiber terminals. Furthermore, following seizures there may be increased anterograde transport of BDNF. First, using hippocampal microdissections of dentate gyrus (which contained hilus) and CA3 (which contained stratum lucidum), Elmer et al. showed that maximal BDNF protein levels after seizures were at 12 h in dentate gyrus but 24 h in CA3 (Elmer et al., 1998). This suggests anterograde transport of seizure-induced BDNF protein. More recent evidence regarding the time course of BDNF immunoreactivity following seizures demonstrates that there is increased BDNF immunoreactivity in dentate granule cells at 4 h followed by subsequent increases in hilus and finally increases in CA3 stratum lucidum at about 24 h (Vezzani et al., 1999) (C. Gall, personal communication). Furthermore, this anterograde ‘‘movement’’ of BDNF immunoreactivity was abrogated by the axonal transport inhibitor colchicine (C. Gall, personal communication). These considerations lead to a model in which CA3 stratum lucidum phospho-trk immunoreactivity is a consequence of seizure-induced BDNF release from mossy-fiber axons activating trkB receptors on dendrites of CA3 pyramidal cells and hilar interneurons. Supporting a postsynaptic site for trk receptor activation is the evidence that fulllength trkB receptors are localized to the postsynaptic density (Wu et al., 1996). Alternatively,

dendritic BDNF mRNA targeting may underlie another potential cellular mechanism for BDNF translation, release, and trk receptor activation (Simonato et al., 2002). Determining the ultrastructural distribution of phospho-trk immunoreactivity would be necessary to distinguish these possibilities. Since the other primary target of mossy-fiber axons in CA3 is dendrites of stratum lucidum interneurons (Spruston et al., 1997), it is possible that phospho-trk immunoreactivity in stratum lucidum could reflect activation of trk receptors on interneurons as well as CA3 pyramidal cell dendrites. Indeed, quantitative analysis of mossy-fiber targets in CA3 suggests that the number of synaptic contacts onto GABAergic interneurons vastly outnumbers those onto CA3 dendrites (Acsady et al., 1998). Indeed, any interneuron with dendrites traversing stratum lucidum could be a target of mossyfiber axons. However, it is unclear whether functional trkB receptors exist on stratum lucidum interneurons, as in situ hybridization studies show trkB mRNA localization predominantly in granule and pyramidal cells of hippocampus (Bengzon et al., 1993) and only occasional interneurons were found to be trkB-immunoreactive in the EM study (Drake et al., 1999). Furthermore, recent evidence indicates that activated trk receptors may be endocytosed and retrogradely transported while still tyrosine phosphorylated (Grimes et al., 1996; Von Bartheld et al., 1996b; Bhattacharyya et al., 1997; Riccio et al., 1997; Senger and Campenot, 1997). Therefore, mossy fiber-like phospho-trk immunoreactivity could in part reflect not only distal synaptic sites of trk activation but also in-progress retrograde transport of activated trk from CA3 within the mossy fibers. Thus, the increase in phospho-trk immunoreactivity observed in the dentate hilus may represent activated trk from mossy-fiber terminals in hilus or CA3.

Role of BDNF in other pathologic conditions Pain BDNF also may play an important neuromodulatory role in pain transduction (Malcangio and

385

Lessmann, 2003). BDNF is synthesized by dorsal horn neurons and markedly upregulated in inflammatory injury to peripheral nerves (along with NGF) (Fukuoka et al., 2001). BDNF acutely sensitizes nociceptive afferents and elicits hyperalgesia which is abrogated by BDNF inhibitors (Kerr et al., 1999; Thompson et al., 1999; Pezet et al., 2002). Central pain sensitization is an activity-dependent increase in excitability of dorsal horn neurons leading to a clinically intractable condition termed ‘‘neuropathic pain’’ in which normally nonpainful somatosensory stimuli (touch and pressure) become exquisitely painful (allodynia). Electrophysiological and behavioral data demonstrate that inhibition of BDNF signal transduction inhibits central pain sensitization (Kerr et al., 1999; Pezet et al., 2002).

Neurodegenerative diseases The idea that degenerative diseases of the nervous system may result from insufficient supply of neurotrophic factors has generated great interest in BDNF as a potential therapeutic agent. Many reports have documented evidence of decreased expression of BDNF in neurological disease (Murer et al., 2001). Selective reduction of BDNF mRNA in the hippocampus has been reported in Alzheimer’s disease specimens (Phillips et al., 1991; Ferrer et al., 1999), although selective upregulation appears to occur in plaque-related glial cells in an animal model (Burbach et al., 2004). Decreased BDNF protein has been demonstrated in the substantia nigra in Parkinson’s disease (Howells et al., 2000). BDNF promotes survival of all major neuronal types affected in Alzheimer’s and Parkinson’s disease, such as hippocampal and neocortical neurons, cholinergic septal and basal forebrain neurons, and nigral dopaminergic neurons. Interestingly, recent work has implicated BDNF in Huntington’s disease as well. Huntingtin, the protein mutated in Huntington’s disease, upregulates BDNF transcription, and loss of huntingtinmediated BDNF transcription leads to loss of trophic support to striatal neurons which subsequently degenerate in the hallmark pathology of the disorder (Zuccato et al., 2001). A recent study has

demonstrated that huntingtin normally inhibits the neuron restrictive silencer element (NRSE) involved in tonic repression of transcription from BDNF promoter II (Zuccato et al., 2003). In all of these disorders, provision of BDNF or increasing endogenous BDNF production may conceivably be therapeutic if applied in the appropriate spatiotemporal context (Spires et al., 2004). Rett syndrome Rett syndrome is an X-linked postnatal neurodevelopmental disorder that strikes approximately 1 in 10,000 girls. It is characterized by regression of normal development after about the age of 1 year and eventually leads to several mental and physical impairment, including cognitive and movement deficits and breathing abnormalities. In 1999, Rett syndrome was linked to mutations in the MECP2 gene on the X chromosome (Amir et al., 1999). MeCP2, the protein product of the MECP2 gene, is a methyl-CpG binding protein, known to bind DNA regulatory regions to silence gene expression. In 2003, it was discovered that one of the genes normally turned off by MeCP2 is BDNF (Chen et al., 2003; Martinowich et al., 2003). Recently, a strain of mice missing the mouse version of MECP2 (Mecp2) has been found to have abnormally low levels of BDNF (Chang et al., 2006). These mice exhibit several features of human Rett syndrome. Increasing BDNF production in mice lacking Mecp2 restored mobility and extended life span (Chang et al., 2006). Neural activity triggers phosphorylation of MeCP2 that detaches it from the regulatory region of the BDNF gene and allows BDNF transcription (Zhou et al., 2006). Further study of MeCP2–BDNF interactions may lead to novel insights and treatment strategies for Rett syndrome. Interestingly, MECP2 abnormalities are starting to be found in other neurodevelopmental disorders such as autism, suggesting that BDNF dysregulation may also have a more widespread role in the pathophysiology of these conditions. Neuropsychiatric disease BDNF signaling may also be involved in affective behaviors (Altar, 1999). Environmental stresses

386

such as immobilization that induce depression also decrease BDNF mRNA (Smith et al., 1995). Conversely, physical exercise is associated with decreased depression and increased BDNF mRNA (Russo-Neustadt et al., 1999; Cotman and Berchtold, 2002). Existing treatments for depression are thought to act primarily by increasing endogenous monoaminergic (i.e. serotonergic and noradrenergic) synaptic transmission, and recent studies have shown that effective antidepressants increase BDNF mRNA (Dias et al., 2003) and protein (Chen et al., 2001; Altar et al., 2003). Exogenous delivery of BDNF promotes the function and sprouting of serotonergic neurons in adult rat brains (Mamounas et al., 1995), and BDNF-deficient mice are also deficient in serotonergic innervation (Lyons et al., 1999). Acute local BDNF infusion has antidepressant-like effects in rats (Shirayama et al., 2002). Thus, new pharmacologic strategies are focused on the potential antidepressant role of BDNF. It has also been hypothesized that BDNF may be involved in bipolar disorder (Tsai, 2004). Interestingly, lithium, a major drug for the treatment of bipolar disorder, increases BDNF and trkB activation in cerebral cortical neurons (Hashimoto et al., 2002). BDNF is an attractive candidate gene for susceptibility to bipolar disorder, and some (Neves-Pereira et al., 2002; Sklar et al., 2002) but not other (Hong et al., 2003; Nakata et al., 2003) studies suggest linkage between BDNF polymorphisms and disease susceptibility (Green and Craddock, 2003). How alterations in BDNF activity may relate to fluctuating bouts of mania and depression in bipolar disorder is still a matter of speculation.

Perspective Since the discovery of NGF in the 1950s and BDNF in the 1980s, a great deal of evidence has mounted for the roles of NGF, BDNF, NT-3, and NT-4/5 in development, physiology, and pathology. BDNF in particular has important roles in neural development and cell survival, as well as appearing essential to molecular mechanisms of synaptic plasticity and larger scale structural

rearrangements of axons and dendrites. Basic activity-related changes in the CNS are thought to depend on BDNF modulation of synaptic transmission. Pathologic levels of BDNF-dependent synaptic plasticity may contribute to conditions such as epilepsy and chronic pain sensitization, whereas application of the trophic properties of BDNF may lead to novel therapeutic options in neurodegenerative diseases and perhaps even in neuropsychiatric disorders. The role of BDNF in epilepsy provides a particularly good example of the pleiotropic effects of BDNF on excitability. The hippocampus and closely associated limbic structures are thought to be particularly important in the pro-epileptogenic effects of BDNF (Binder et al., 1999a), and increased BDNF expression in the hippocampus is found in specimens from patients with temporal lobe epilepsy (Mathern et al., 1997; Takahashi et al., 1999). It is hoped that understanding of the hyperexcitability associated with BDNF in epilepsy animal models may lead to novel anticonvulsant or antiepileptic therapies (Binder et al., 2001). Of course, simple up- or downregulation of NTs may lead to many nonspecific effects. For ultimate clinical application in specific conditions, it will be very helpful to elucidate the mechanisms of action of each of the effects of NT receptor activation. Therefore, much recent research has focused on downstream targets of the NT signaling pathways responsible for specific phenotypic effects. For example, BDNF activation of trkB down-regulates hippocampal KCC2, a K+-Cl cotransporter (Rivera et al., 2002); this suppresses chloridedependent fast GABAergic inhibition and may partially account for BDNF modulation of GABAergic synapses (Wardle and Poo, 2003). In addition, BDNF phosphorylates specific subunits of both, the NMDA receptor and the GABAA receptor, altering their function (Suen et al., 1997; Lin et al., 1998; Jovanovic et al., 2004). Long-term effects of BDNF must take into account the fact that it upregulates many other plasticity-related genes, such as NPY (Croll et al., 1994; Nawa et al., 1994). NPY, for example, may not only modulate excitability (Baraban et al., 1997; Reibel et al., 2000a) but also other phenomena such as neurogenesis (Howell et al., 2003).

387

New methods of modulation/upregulation of NTs may be required to achieve translational control of diseases of NT deficiency. ‘‘Ampakines’’ represent a new class of compounds that have been shown to upregulate BDNF over a long period of time (Lauterborn et al., 2003). Of course, proper dosing so as not to trigger downregulation of important NT signaling pathways will be critical to avoiding deleterious side effects of these potential new therapies. Nevertheless, these and similar compounds are under active clinical investigation as cognitive and memory enhancing drugs (Danysz, 1999; Johnson and Simmon, 2002). Acknowledgment I would like to thank Helen Scharfman for her detailed analysis and constructive criticism. References Acheson, A., Conover, J.C., Fandl, J.P., DeChiara, T.M., Russell, M., Thadani, A., Squinto, S.P., Yancopoulos, G.D. and Lindsay, R.M. (1995) A BDNF autocrine loop in adult sensory neurons prevents cell death. Nature, 374: 450–453. Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Adams, B., Sazgar, M., Osehobo, P., Van der Zee, C.E.E.M., Diamond, J., Fahnestock, M. and Racine, R.J. (1997) Nerve growth factor accelerates seizure development, enhances mossy fiber sprouting, and attenuates seizure-induced decreases in neuronal density in the kindling model of epilepsy. J. Neurosci., 17: 5288–5296. Adams, B., Vaccarella, L., Fahnestock, M. and Racine, R.J. (2002) The cholinergic system modulates kindling and kindling-induced mossy fiber sprouting. Synapse, 44: 132–138. Alderson, R.F., Alterman, A.L., Barde, Y.A. and Lindsay, R.M. (1990) Brain-derived neurotrophic factor increases survival and differentiated functions of rat septal cholinergic neurons in culture. Neuron, 5: 297–306. Alderson, R.F., Curtis, R., Alterman, A.L., Lindsay, R.M. and DiStefano, P.S. (2000) Truncated TrkB mediates the endocytosis and release of BDNF and neurotrophin-4/5 by rat astrocytes and schwann cells in vitro. Brain Res., 871: 210–222. Alonso, M., Vianna, M.R., Depino, A.M., Mello e Souza, T., Pereira, P., Szapiro, G., Viola, H., Pitossi, F., Izquierdo, I. and Medina, J.H. (2002) BDNF-triggered events in the rat hippocampus are required for both short- and long-term memory formation. Hippocampus, 12: 551–560.

Altar, C.A. (1999) Neurotrophins and depression. Trends Pharmacol. Sci., 20: 59–61. Altar, C.A., Cai, N., Bliven, T., Juhasz, M., Conner, J.M., Acheson, A.L., Lindsay, R.M. and Wiegand, S.J. (1997) Anterograde transport of brain-derived neurotrophic factor and its role in the brain. Nature, 389: 856–860. Altar, C.A. and DiStefano, P.S. (1998) Neurotrophin trafficking by anterograde transport. Trends Neurosci., 21: 433–437. Altar, C.A., Siuciak, J.A., Wright, P., Ip, N.Y., Lindsay, R.M. and Wiegand, S.J. (1994) In situ hybridization of trkB and trkC receptor mRNA in rat forebrain and association with high-affinity binding of [125I]BDNF, [125I]NT-4/5 and [125I]NT-3. Eur. J. Neurosci., 6: 1389–1405. Altar, C.A., Whitehead, R.E., Chen, R., Wortwein, G. and Madsen, T.M. (2003) Effects of electroconvulsive seizures and antidepressant drugs on brain-derived neurotrophic factor protein in rat brain. Biol. Psychiatry, 54: 703–709. Amir, R.E., Van den Veyver, I.B., Wan, M., Tran, C.Q., Francke, U. and Zoghbi, H.Y. (1999) Rett syndrome is caused by mutations in X-linked MECP2, encoding methylCpG-binding protein 2. Nat. Genet., 23: 185–188. Anderson, K.D., Alderson, R.F., Altar, C.A., DiStefano, P.S., Corcoran, T.L., Lindsay, R.M. and Wiegand, S.J. (1995) Differential distribution of exogenous BDNF, NGF, and NT-3 in the brain corresponds to the relative abundance and distribution of high-affinity and low-affinity neurotrophin receptors. J. Comp. Neurol., 357: 296–317. Asztely, F., Kokaia, M., Olofsdotter, K., Ortegren, U. and Lindvall, O. (2000) Afferent-specific modulation of shortterm synaptic plasticity by neurotrophins in dentate gyrus. Eur. J. Neurosci., 12: 662–669. Baraban, S.C., Hollopeter, G., Erickson, J.C., Schwartzkroin, P.A. and Palmiter, R.D. (1997) Knock-out mice reveal a critical antiepileptic role for neuropeptide Y. J. Neurosci., 17: 8927–8936. Barbacid, M. (1994) The trk family of neurotrophin receptors. J. Neurobiol., 25: 1386–1403. Barde, Y.A., Edgar, D. and Thoenen, H. (1982) Purification of a new neurotrophic factor from mammalian brain. EMBO J., 1: 549–553. Barnabe-Heider, F. and Miller, F.D. (2003) Endogenously produced neurotrophins regulate survival and differentiation of cortical progenitors via distinct signaling pathways. J. Neurosci., 23: 5149–5160. Bengzon, J., Kokaia, Z., Ernfors, P., Kokaia, M., Leanza, G., Nilsson, O.G., Persson, H. and Lindvall, O. (1993) Regulation of neurotrophin and trkA, trkB and trkC tyrosine kinase receptor messenger RNA expression in kindling. Neuroscience, 53: 433–446. Bengzon, J., Soderstrom, S., Kokaia, Z., Kokaia, M., Ernfors, P., Persson, H., Ebendal, T. and Lindvall, O. (1992) Widespread increase of nerve growth factor protein in the rat forebrain after kindling-induced seizures. Brain Res., 587: 338–342. Benraiss, A., Chmielnicki, E., Lerner, K., Roh, D. and Goldman, S.A. (2001) Adenoviral brain-derived neurotrophic factor induces both neostriatal and olfactory neuronal

388 recruitment from endogenous progenitor cells in the adult forebrain. J. Neurosci., 21: 6718–6731. Bhattacharyya, A., Watson, F.L., Bradlee, T.A., Pomeroy, S.L., Stiles, C.D. and Segal, R.A. (1997) Trk receptors function as rapid retrograde signal carriers in the adult nervous system. J. Neurosci., 17: 7007–7016. Biffo, S., Offenhauser, N., Carter, B.D. and Barde, Y.A. (1995) Selective binding and internalisation by truncated receptors restrict the availability of BDNF during development. Development, 121: 2461–2470. Binder, D.K., Croll, S.D., Gall, C.M. and Scharfman, H.E. (2001) BDNF and epilepsy: too much of a good thing? Trends Neurosci., 24: 47–53. Binder, D.K., Routbort, M.J. and McNamara, J.O. (1999a) Immunohistochemical evidence of seizure-induced activation of trk receptors in the mossy fiber pathway of adult rat hippocampus. J. Neurosci., 19: 4616–4626. Binder, D.K., Routbort, M.J., Ryan, T.E., Yancopoulos, G.D. and McNamara, J.O. (1999b) Selective inhibition of kindling development by intraventricular administration of TrkB receptor body. J. Neurosci., 19: 1424–1436. Black, I.B. (1999) Trophic regulation of synaptic plasticity. J. Neurobiol., 41: 108–118. Bramham, C.R., Southard, T., Sarvey, J.M., Herkenham, M. and Brady, L.S. (1996) Unilateral LTP triggers bilateral increases in hippocampal neurotrophin and trk receptor mRNA expression in behaving rats: evidence for interhemispheric communication. J. Comp. Neurol., 368: 371–382. Bregola, G., Frigati, L., Zucchini, S. and Simonato, M. (2000) Different patterns of induction of fibroblast growth factor-2 and brain-derived neurotrophic factor messenger RNAs during kindling epileptogenesis, and development of a herpes simplex vector for fibroblast growth factor-2 gene transfer in vivo. Epilepsia, 41(Suppl 6): S122–S126. Brigadski, T., Hartmann, M. and Lessmann, V. (2005) Differential vesicular targeting and time course of synaptic secretion of the mammalian neurotrophins. J. Neurosci., 25: 7601–7614. Burbach, G.J., Hellweg, R., Haas, C.A., Del Turco, D., Deicke, U., Abramowski, D., Jucker, M., Staufenbiel, M. and Deller, T. (2004) Induction of brain-derived neurotrophic factor in plaque-associated glial cells of aged APP23 transgenic mice. J. Neurosci., 24: 2421–2430. Cabelli, R.J., Hohn, A. and Shatz, C.J. (1995) Inhibition of ocular dominance column formation by infusion of NT-4/5 or BDNF. Science, 267: 1662–1666. Cabelli, R.J., Shelton, D.L., Segal, R.A. and Shatz, C.J. (1997) Blockade of endogenous ligands of trkB inhibits formation of ocular dominance columns. Neuron, 19: 63–76. Casaccia-Bonnefil, P., Carter, B.D., Dobrowsky, R.T. and Chao, M.V. (1996) Death of oligodendrocytes mediated by the interaction of nerve growth factor with its receptor p75. Nature, 383: 716–719. Castre´n, E., Berninger, B., Leingartner, A. and Lindholm, D. (1998) Regulation of brain-derived neurotrophic factor mRNA levels in hippocampus by neuronal activity. Prog. Brain Res., 117: 57–64.

Castre´n, E., Pitkanen, M., Sirvio, J., Parsadanian, A., Lindholm, D., Thoenen, H. and Riekkinen, P.J. (1993) The induction of LTP increases BDNF and NGF mRNA but decreases NT-3 mRNA in the dentate gyrus. Neuroreport, 4: 895–898. Castre´n, E., Thoenen, H. and Lindholm, D. (1995) Brain-derived neurotrophic factor messenger RNA is expressed in the septum, hypothalamus and in adrenergic brain stem nuclei of adult rat brain and is increased by osmotic stimulation in the paraventricular nucleus. Neuroscience, 64: 71–80. Castre´n, E., Zafra, F., Thoenen, H. and Lindholm, D. (1992) Light regulates expression of brain-derived neurotrophic factor mRNA in rat visual cortex. Proc. Natl. Acad. Sci. U.S.A., 89: 9444–9448. Cellerino, A. (1996) Expression of messenger RNA coding for the nerve growth factor receptor trkA in the hippocampus of the adult rat. Neuroscience, 70: 613–616. Chang, Q., Khare, G., Dani, V., Nelson, S. and Jaenisch, R. (2006) The disease progression of Mecp2 mutant mice is affected by the level of BDNF expression. Neuron, 49: 341–348. Chao, M.V. and Bothwell, M. (2002) Neurotrophins: to cleave or not to cleave. Neuron, 33: 9–12. Chao, M.V. and Hempstead, B.L. (1995) p75 and trk: a tworeceptor system. Trends Neurosci., 18: 321–326. Chen, B., Dowlatshahi, D., MacQueen, G.M., Wang, J.F. and Young, L.T. (2001) Increased hippocampal BDNF immunoreactivity in subjects treated with antidepressant medication. Biol. Psychiatry, 50: 260–265. Chen, W.G., Chang, Q., Lin, Y., Meissner, A., West, A.E., Griffith, E.C., Jaenisch, R. and Greenberg, M.E. (2003) Derepression of BDNF transcription involves calcium-dependent phosphorylation of MeCP2. Science, 302: 885–889. Cheng, B., Goodman, Y., Begley, J.G. and Mattson, M.P. (1994) Neurotrophin-4/5 protects hippocampal and cortical neurons against energy deprivation- and excitatory amino acid-induced injury. Brain Res., 650: 331–335. Chmielnicki, E., Benraiss, A., Economides, A.N. and Goldman, S.A. (2004) Adenovirally expressed noggin and brain-derived neurotrophic factor cooperate to induce new medium spiny neurons from resident progenitor cells in the adult striatal ventricular zone. J. Neurosci., 24: 2133–2142. Climent, E., Sancho-Tello, M., Minana, R., Barettino, D. and Guerri, C. (2000) Astrocytes in culture express the full-length Trk-B receptor and respond to brain derived neurotrophic factor by changing intracellular calcium levels: effect of ethanol exposure in rats. Neurosci. Lett., 288: 53–56. Conner, J.M., Lauterborn, J.C., Yan, Q., Gall, C.M. and Varon, S. (1997) Distribution of brain-derived neurotrophic factor (BDNF) protein and mRNA in the normal adult rat CNS–evidence for anterograde axonal transport. J. Neurosci., 17: 2295–2313. Conover, J.C., Erickson, J.T., Katz, D.M., Bianchi, L.M., Poueymirou, W.T., McClain, J., Pan, L., Helgren, M., Ip, N.Y., Boland, P., et al. (1995) Neuronal deficits, not involving motor neurons, in mice lacking BDNF and/or NT4. Nature, 375: 235–238.

389 Cotman, C.W. and Berchtold, N.C. (2002) Exercise: a behavioral intervention to enhance brain health and plasticity. Trends Neurosci., 25: 295–301. Croll, S.D., Suri, C., Compton, D.L., Simmons, M.V., Yancopoulos, G.D., Lindsay, R.M., Wiegand, S.J., Rudge, J.S. and Scharfman, H.E. (1999) Brain-derived neurotrophic factor transgenic mice exhibit passive avoidance deficits, increased seizure severity and in vitro hyperexcitability in the hippocampus and entorhinal cortex. Neuroscience, 93: 1491–1506. Croll, S.D., Wiegand, S.J., Anderson, K.D., Lindsay, R.M. and Nawa, H. (1994) Regulation of neuropeptides in adult rat forebrain by the neurotrophins BDNF and NGF. Eur. J. Neurosci., 6: 1343–1353. Danysz, W. (1999) CX-516 (Cortex Pharmaceuticals Inc). IDrugs, 2: 814–822. Dechant, G. and Barde, Y.A. (2002) The neurotrophin receptor p75(NTR): novel functions and implications for diseases of the nervous system. Nat. Neurosci., 5: 1131–1136. Dias, B.G., Banerjee, S.B., Duman, R.S. and Vaidya, V.A. (2003) Differential regulation of brain derived neurotrophic factor transcripts by antidepressant treatments in the adult rat brain. Neuropharmacology, 45: 553–563. Drake, C.T., Milner, T.A. and Patterson, S.L. (1999) Ultrastructural localization of full-length trkB immunoreactivity in rat hippocampus suggests multiple roles in modulating activity-dependent synaptic plasticity. J. Neurosci., 19: 8009–8026. Dugich-Djordjevic, M.M., Tocco, G., Lapchak, P.A., Pasinetti, G.M., Najm, I., Baudry, M. and Hefti, F. (1992a) Regionally specific and rapid increases in brain-derived neurotrophic factor messenger RNA in the adult rat brain following seizures induced by systemic administration of kainic acid. Neuroscience, 47: 303–315. Dugich-Djordjevic, M.M., Tocco, G., Willoughby, D.A., Najm, I., Pasinetti, G., Thompson, R.F., Baudry, M., Lapchak, P.A. and Hefti, F. (1992b) BDNF mRNA expression in the developing rat brain following kainic acid-induced seizure activity. Neuron, 8: 1127–1138. Egan, M.F., Kojima, M., Callicott, J.H., Goldberg, T.E., Kolachana, B.S., Bertolino, A., Zaitsev, E., Gold, B., Goldman, D., Dean, M., Lu, B. and Weinberger, D.R. (2003) The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function. Cell, 112: 257–269. Eide, F.F., Vining, E.R., Eide, B.L., Zang, K., Wang, X.Y. and Reichardt, L.F. (1996) Naturally occurring truncated trkB receptors have dominant inhibitory effects on brain-derived neurotrophic factor signaling. J. Neurosci., 16: 3123–3129. Elmer, E., Kokaia, M., Ernfors, P., Ferencz, I., Kokaia, Z. and Lindvall, O. (1997) Suppressed kindling epileptogenesis and perturbed BDNF and trkB gene regulation in NT-3 mutant mice. Exp. Neurol., 145: 93–103. Elmer, E., Kokaia, M., Kokaia, Z., Ferencz, I. and Lindvall, O. (1996a) Delayed kindling development after rapidly recurring seizures: relation to mossy fiber sprouting and neurotrophin, GAP-43 and dynorphin gene expression. Brain Res., 712: 19–34.

Elmer, E., Kokaia, Z., Kokaia, M., Carnahan, J., Nawa, H., Bengzon, J. and Lindvall, O. (1996b) Widespread increase of brain-derived neurotrophic factor protein in the rat forebrain after kindling-induced seizures. Soc. Neurosci. Abstr., 22: 2089. Elmer, E., Kokaia, Z., Kokaia, M., Carnahan, J., Nawa, H. and Lindvall, O. (1998) Dynamic changes of brain-derived neurotrophic factor protein levels in the rat forebrain after single and recurring kindling-induced seizures. Neuroscience, 83: 351–362. Ernfors, P., Bengzon, J., Kokaia, Z., Persson, H. and Lindvall, O. (1991) Increased levels of messenger RNAs for neurotrophic factors in the brain during kindling epileptogenesis. Neuron, 7: 165–176. Ernfors, P., Wetmore, C., Olson, L. and Persson, H. (1990) Identification of cells in rat brain and peripheral tissues expressing mRNA for members of the nerve growth factor family. Neuron, 5: 511–526. Esposito, D., Patel, P., Stephens, R.M., Perez, P., Chao, M.V., Kaplan, D.R. and Hempstead, B.L. (2001) The cytoplasmic and transmembrane domains of the p75 and Trk A receptors regulate high affinity binding to nerve growth factor. J. Biol. Chem., 276: 32687–32695. Farhadi, H.F., Mowla, S.J., Petrecca, K., Morris, S.J., Seidah, N.G. and Murphy, R.A. (2000) Neurotrophin-3 sorts to the constitutive secretory pathway of hippocampal neurons and is diverted to the regulated secretory pathway by coexpression with brain-derived neurotrophic factor. J. Neurosci., 20: 4059–4068. Fawcett, J.P., Alonso-Vanegas, M.A., Morris, S.J., Miller, F.D., Sadikot, A.F. and Murphy, R.A. (2000) Evidence that brain-derived neurotrophic factor from presynaptic nerve terminals regulates the phenotype of calbindin-containing neurons in the lateral septum. J. Neurosci., 20: 274–282. Fawcett, J.P., Aloyz, R., McLean, J.H., Pareek, S., Miller, F.D., McPherson, P.S. and Murphy, R.A. (1997) Detection of brain-derived neurotrophic factor in a vesicular fraction of brain synaptosomes. J. Biol. Chem., 272: 8837–8840. Fawcett, J.P., Bamji, S.X., Causing, C.G., Aloyz, R., Ase, A.R., Reader, T.A., McLean, J.H. and Miller, F.D. (1998) Functional evidence that BDNF is an anterograde neuronal trophic factor in the CNS. J. Neurosci., 18: 2808–2821. Ferrer, I., Marin, C., Rey, M.J., Ribalta, T., Goutan, E., Blanco, R., Tolosa, E. and Marti, E. (1999) BDNF and fulllength and truncated TrkB expression in Alzheimer disease. Implications in therapeutic strategies. J. Neuropathol. Exp. Neurol., 58: 729–739. Figurov, A., Pozzo-Miller, L.D., Olafsson, P., Wang, T. and Lu, B. (1996) Regulation of synaptic responses to high-frequency stimulation and LTP by neurotrophins in the hippocampus. Nature, 381: 706–709. Finkbeiner, S., Tavazoie, S.F., Maloratsky, A., Jacobs, K.M., Harris, K.M. and Greenberg, M.E. (1997) CREB: a major mediator of neuronal neurotrophin responses. Neuron, 19: 1031–1047. Frade, J.M., Rodriguez-Tebar, A. and Barde, Y.A. (1996) Induction of cell death by endogenous nerve growth factor through its p75 receptor. Nature, 383: 166–168.

390 Frank, L., Ventimiglia, R., Anderson, K., Lindsay, R.M. and Rudge, J.S. (1996) BDNF downregulates neurotrophin responsiveness, trkB protein and trkB mRNA levels in cultured rat hippocampal neurons. Eur. J. Neurosci., 8: 1220–1230. Frerking, M., Malenka, R.C. and Nicoll, R.A. (1998) Brainderived neurotrophic factor (BDNF) modulates inhibitory, but not excitatory, transmission in the CA1 region of the hippocampus. J Neurophysiol., 80: 3383–3386. Frisen, J., Verge, V.M., Fried, K., Risling, M., Persson, H., Trotter, J., Hokfelt, T. and Lindholm, D. (1993) Characterization of glial trkB receptors: differential response to injury in the central and peripheral nervous systems. Proc. Natl. Acad. Sci. U.S.A., 90: 4971–4975. Fryer, R.H., Kaplan, D.R., Feinstein, S.C., Radeke, M.J., Grayson, D.R. and Kromer, L.F. (1996) Developmental and mature expression of full-length and truncated trkB receptors in the rat forebrain. J. Comp. Neurol., 374: 21–40. Fryer, R.H., Kaplan, D.R. and Kromer, L.F. (1997) Truncated trkB receptors on nonneuronal cells inhibit BDNF-induced neurite outgrowth in vitro. Exp. Neurol., 148: 616–627. Fukuoka, T., Kondo, E., Dai, Y., Hashimoto, N. and Noguchi, K. (2001) Brain-derived neurotrophic factor increases in the uninjured dorsal root ganglion neurons in selective spinal nerve ligation model. J. Neurosci., 21: 4891–4900. Funabashi, T., Sasaki, H. and Kimura, F. (1988) Intraventricular injection of antiserum to nerve growth factor delays the development of amygdaloid kindling. Brain Res., 458: 132–136. Gall, C., Lauterborn, J., Bundman, M., Murray, K. and Isackson, P. (1991) Seizures and the regulation of neurotrophic factor and neuropeptide gene expression in brain. Epilepsy Res. Suppl., 4: 225–245. Gall, C.M. and Isackson, P.J. (1989) Limbic seizures increase neuronal production of messenger RNA for nerve growth factor. Science, 245: 758–761. Gall, C.M. and Lauterborn, J. (1992) In: Ribak C.E., Gall C.M. and Mody I. (Eds.), The Dentate Gyrus and its Role in Seizures. Elsevier, Amsterdam, pp. 171–185. Gall, C.M., Lauterborn, J.C., Guthrie, K.M. and Stinis, C.T. (1997) Seizures and the regulation of neurotrophic factor expression: associations with structural plasticity in epilepsy. Adv. Neurol., 72: 9–24. Ginty, D.D. and Segal, R.A. (2002) Retrograde neurotrophin signaling: Trk-ing along the axon. Curr. Opin. Neurobiol., 12: 268–274. Goggi, J., Pullar, I.A., Carney, S.L. and Bradford, H.F. (2003) The control of [125I]BDNF release from striatal rat brain slices. Brain Res., 967: 201–209. Gould, E. and McEwen, B.S. (1993) Neuronal birth and death. Curr. Opin. Neurobiol., 3: 676–682. Green, E. and Craddock, N. (2003) Brain-derived neurotrophic factor as a potential risk locus for bipolar disorder: evidence, limitations, and implications. Curr. Psychiatry Rep., 5: 469–476. Griesbeck, O., Canossa, M., Campana, G., Gartner, A., Hoener, M.C., Nawa, H., Kolbeck, R. and Thoenen, H. (1999) Are there differences between the secretion characteristics of

NGF and BDNF? Implications for the modulatory role of neurotrophins in activity-dependent neuronal plasticity. Microsc. Res. Tech., 45: 262–275. Grimes, M.L., Zhou, J., Beattie, E.C., Yuen, E.C., Hall, D.E., Valletta, J.S., Topp, K.S., LaVail, J.H., Bunnett, N.W. and Mobley, W.C. (1996) Endocytosis of activated trkA: evidence that nerve growth factor induces formation of singaling endosomes. J. Neurosci., 16: 7950–7964. Gustafsson, E., Lindvall, O. and Kokaia, Z. (2003) Intraventricular infusion of TrkB-Fc fusion protein promotes ischemia-induced neurogenesis in adult rat dentate gyrus. Stroke, 34: 2710–2715. Haapasalo, A., Koponen, E., Hoppe, E., Wong, G. and Castren, E. (2001) Truncated trkB.T1 is dominant negative inhibitor of trkB.TK+-mediated cell survival. Biochem. Biophys. Res. Commun., 280: 1352–1358. Haapasalo, A., Sipola, I., Larsson, K., Akerman, K.E., Stoilov, P., Stamm, S., Wong, G. and Castren, E. (2002) Regulation of TRKB surface expression by brain-derived neurotrophic factor and truncated TRKB isoforms. J. Biol. Chem., 277: 43160–43167. Hall, J., Thomas, K.L. and Everitt, B.J. (2000) Rapid and selective induction of BDNF expression in the hippocampus during contextual learning. Nat. Neurosci., 3: 533–535. Hallbook, F., Ibanez, C.F. and Persson, H. (1991) Evolutionary studies of the nerve growth factor family reveal a novel member abundantly expressed in Xenopus ovary. Neuron, 6: 845–858. Hartmann, M., Heumann, R. and Lessmann, V. (2001) Synaptic secretion of BDNF after high-frequency stimulation of glutamatergic synapses. EMBO J., 20: 5887–5897. Hashimoto, R., Takei, N., Shimazu, K., Christ, L., Lu, B. and Chuang, D.M. (2002) Lithium induces brain-derived neurotrophic factor and activates TrkB in rodent cortical neurons: an essential step for neuroprotection against glutamate excitotoxicity. Neuropharmacology, 43: 1173–1179. He, X.P., Butler, L., Liu, X. and McNamara, J.O. (2006) The tyrosine receptor kinase B ligand, neurotrophin-4, is not required for either epileptogenesis or tyrosine receptor kinase B activation in the kindling model. Neuroscience, 141: 515–520. He, X.P., Minichiello, L., Klein, R. and McNamara, J.O. (2002) Immunohistochemical evidence of seizure-induced activation of trkB receptors in the mossy fiber pathway of adult mouse hippocampus. J. Neurosci., 22: 7502–7508. Hofer, M., Pagliusi, S.R., Hohn, A., Leibrock, J. and Barde, Y.A. (1990) Regional distribution of brain-derived neurotrophic factor mRNA in the adult mouse brain. EMBO J., 9: 2459–2464. Hong, C.J., Huo, S.J., Yen, F.C., Tung, C.L., Pan, G.M. and Tsai, S.J. (2003) Association study of a brain-derived neurotrophic-factor genetic polymorphism and mood disorders, age of onset and suicidal behavior. Neuropsychobiology, 48: 186–189. Horch, H.W. and Katz, L.C. (2002) BDNF release from single cells elicits local dendritic growth in nearby neurons. Nat. Neurosci., 5: 1177–1184.

391 Howell, O.W., Scharfman, H.E., Herzog, H., Sundstrom, L.E., Beck-Sickinger, A. and Gray, W.P. (2003) Neuropeptide Y is neuroproliferative for post-natal hippocampal precursor cells. J. Neurochem., 86: 646–659. Howells, D.W., Porritt, M.J., Wong, J.Y., Batchelor, P.E., Kalnins, R., Hughes, A.J. and Donnan, G.A. (2000) Reduced BDNF mRNA expression in the Parkinson’s disease substantia nigra. Exp. Neurol., 166: 127–135. Huang, E.J. and Reichardt, L.F. (2001) Neurotrophins: roles in neuronal development and function. Annu. Rev. Neurosci., 24: 677–736. Huang, E.J. and Reichardt, L.F. (2003) Trk receptors: roles in neuronal signal transduction. Annu. Rev. Biochem., 72: 609–642. Hughes, P.E., Young, D., Preston, K.M., Yan, Q. and Dragunow, M. (1998) Differential regulation by MK801 of immediate-early genes, brain-derived neurotrophic factor and trk receptor mRNA induced by a kindling after-discharge. Mol. Brain Res., 53: 138–151. Humpel, C., Wetmore, C. and Olson, L. (1993) Regulation of brain-derived neurotrophic factor messenger RNA and protein at the cellular level in pentylenetetrazol-induced epileptic seizures. Neuroscience, 53: 909–918. Hyman, C., Hofer, M., Barde, Y.A., Juhasz, M., Yancopoulos, G.D., Squinto, S.P. and Lindsay, R.M. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature, 350: 230–232. Ip, N.Y., Ibanez, C.F., Nye, S.H., McClain, J., Jones, P.F., Gies, D.R., Belluscio, L., Le Beau, M.M., Espinosa III, R., Squinto, S.P., et al. (1992) Mammalian neurotrophin-4: structure, chromosomal localization, tissue distribution, and receptor specificity. Proc. Natl. Acad. Sci. U.S.A., 89: 3060–3064. Isackson, P.J., Huntsman, M.M., Murray, K.D. and Gall, C.M. (1991) BDNF mRNA expression is increased in adult rat forebrain after limbic seizures: temporal patterns of induction distinct from NGF. Neuron, 6: 937–948. Ishibashi, H., Hihara, S., Takahashi, M., Heike, T., Yokota, T. and Iriki, A. (2002) Tool-use learning induces BDNF expression in a selective portion of monkey anterior parietal cortex. Brain Res. Mol. Brain Res., 102: 110–112. Ito, H., Nakajima, A., Nomoto, H. and Furukawa, S. (2003) Neurotrophins facilitate neuronal differentiation of cultured neural stem cells via induction of mRNA expression of basic helix-loop-helix transcription factors Mash1 and Math1. J. Neurosci. Res., 71: 648–658. Jankowsky, J.L. and Patterson, P.H. (2001) The role of cytokines and growth factors in seizures and their sequelae. Prog. Neurobiol., 63: 125–149. Johnson, J.E., Barde, Y.A., Schwab, M. and Thoenen, H. (1986) Brain-derived neurotrophic factor supports the survival of cultured rat retinal ganglion cells. J. Neurosci., 6: 3031–3038. Johnson, S.A. and Simmon, V.F. (2002) Randomized, doubleblind, placebo-controlled international clinical trial of the Ampakine CX516 in elderly participants with mild cognitive impairment: a progress report. J. Mol. Neurosci., 19: 197–200.

Jovanovic, J.N., Thomas, P., Kittler, J.T., Smart, T.G. and Moss, S.J. (2004) Brain-derived neurotrophic factor modulates fast synaptic inhibition by regulating GABA(A) receptor phosphorylation, activity, and cell-surface stability. J. Neurosci., 24: 522–530. Kafitz, K.W., Rose, C.R., Thoenen, H. and Konnerth, A. (1999) Neurotrophin-evoked rapid excitation through TrkB receptors. Nature, 401: 918–921. Kang, H. and Schuman, E.M. (1995) Long-lasting neurotrophin-induced enhancement of synaptic transmission in the adult hippocampus. Science, 267: 1658–1662. Kang, H., Welcher, A.A., Shelton, D. and Schuman, E.M. (1997) Neurotrophins and time: different roles for trkB signaling in hippocampal long-term potentiation. Neuron, 19: 653–664. Katoh-Semba, R., Asano, T., Ueda, H., Morishita, R., Takeuchi, I.K., Inaguma, Y. and Kato, K. (2002) Riluzole enhances expression of brain-derived neurotrophic factor with consequent proliferation of granule precursor cells in the rat hippocampus. FASEB J., 16: 1328–1330. Katoh-Semba, R., Ichisaka, S., Hata, Y., Tsumoto, T., Eguchi, K., Miyazaki, N., Matsuda, M., Takeuchi, I.K. and Kato, K. (2003) NT-4 protein is localized in neuronal cells in the brain stem as well as the dorsal root ganglion of embryonic and adult rats. J. Neurochem., 86: 660–668. Katoh-Semba, R., Kaisho, Y., Shintani, A., Nagahama, M. and Kato, K. (1996) Tissue distribution and immunocytochemical localization of neurotrophin-3 in the brain and peripheral tissues of rats. J. Neurochem., 66: 330–337. Kernie, S.G., Liebl, D.J. and Parada, L.F. (2000) BDNF regulates eating behavior and locomotor activity in mice. EMBO J., 19: 1290–1300. Kerr, B.J., Bradbury, E.J., Bennett, D.L., Trivedi, P.M., Dassan, P., French, J., Shelton, D.B., McMahon, S.B. and Thompson, S.W. (1999) Brain-derived neurotrophic factor modulates nociceptive sensory inputs and NMDA-evoked responses in the rat spinal cord. J. Neurosci., 19: 5138–5148. Kim, H.G., Wang, T., Olafsson, P. and Lu, B. (1994) Neurotrophin 3 potentiates neuronal activity and inhibits gamma-aminobutyratergic synaptic transmission in cortical neurons. Proc. Natl. Acad. Sci. U.S.A., 91: 12341–12345. Kim, S.Y., Smith, M.A., Post, R.M. and Rosen, J.B. (1998) Attenuation of kindling-induced decreases in NT-3 mRNA by thyroid hormone depletion. Epilepsy Res., 29: 211–220. Knusel, B., Gao, H., Okazaki, T., Yoshida, T., Mori, N., Hefti, F. and Kaplan, D.R. (1997) Ligand-induced down-regulation of trk messenger RNA, protein and tyrosine phosphorylation in rat cortical neurons. Neuroscience, 78: 851–862. Knusel, B., Winslow, J.W., Rosenthal, A., Burton, L.E., Seid, D.P., Nikolics, K. and Hefti, F. (1991) Promotion of central cholinergic and dopaminergic neuron differentiation by brain-derived neurotrophic factor but not neurotrophin 3. Proc. Natl. Acad. Sci. U.S.A., 88: 961–965. Kohara, K., Kitamura, A., Morishima, M. and Tsumoto, T. (2001) Activity-dependent transfer of brain-derived neurotrophic factor to postsynaptic neurons. Science, 291: 2419–2423.

392 Kokaia, M., Asztely, F., Olofsdotter, K., Sindreu, C.B., Kullmann, D.M. and Lindvall, O. (1998) Endogenous neurotrophin-3 regulates short-term plasticity at lateral perforant path-granule cell synapses. J. Neurosci., 18: 8730–8739. Kokaia, M., Ernfors, P., Kokaia, Z., Elmer, E., Jaenisch, R. and Lindvall, O. (1995) Suppressed epileptogenesis in BDNF mutant mice. Exp. Neurol., 133: 215–224. Korte, M., Carroll, P., Wolf, E., Brem, G., Thoenen, H. and Bonhoeffer, T. (1995) Hippocampal long-term potentiation is impaired in mice lacking brain-derived neurotrophic factor. Proc. Natl. Acad. Sci. U.S.A., 92: 8856–8860. Korte, M., Griesbeck, O., Gravel, C., Carroll, P., Staiger, V., Thoenen, H. and Bonhoeffer, T. (1996) Virus-mediated gene transfer into hippocampal CA1 region restores long-term potentiation in brain-derived neurotrophic factor mutant mice. Proc. Natl. Acad. Sci. U.S.A., 93: 12547–12552. Korte, M., Minichiello, L., Klein, R. and Bonhoeffer, T. (2000) Shc-binding site in the TrkB receptor is not required for hippocampal long-term potentiation. Neuropharmacology, 39: 717–724. Lahteinen, S., Pitkanen, A., Saarelainen, T., Nissinen, J., Koponen, E. and Castren, E. (2002) Decreased BDNF signalling in transgenic mice reduces epileptogenesis. Eur. J. Neurosci., 15: 721–734. Larmet, Y., Reibel, S., Carnahan, J., Nawa, H., Marescaux, C. and Depaulis, A. (1995) Protective effects of brain-derived neurotrophic factor on the development of hippocampal kindling in the rat. Neuroreport, 6: 1937–1941. Larsson, E., Mandel, R.J., Klein, R.L., Muzyczka, N., Lindvall, O. and Kokaia, Z. (2002) Suppression of insult-induced neurogenesis in adult rat brain by brain-derived neurotrophic factor. Exp. Neurol., 177: 1–8. Lauterborn, J.C., Isackson, P.J. and Gall, C.M. (1994) Seizureinduced increases in NGF mRNA exhibit different time courses across forebrain regions and are biphasic in hippocampus. Exp. Neurol., 125: 22–40. Lauterborn, J.C., Rivera, S., Stinis, C.T., Hayes, V.Y., Isackson, P.J. and Gall, C.M. (1996) Differential effects of protein synthesis inhibition on the activity-dependent expression of BDNF transcripts: evidence for immediateearly gene responses from specific promoters. J. Neurosci., 16: 7428–7436. Lauterborn, J.C., Truong, G.S., Baudry, M., Bi, X., Lynch, G. and Gall, C.M. (2003) Chronic elevation of brain-derived neurotrophic factor by ampakines. J. Pharmacol. Exp. Ther., 307: 297–305. Lee, F.S., Kim, A.H., Khursigara, G. and Chao, M.V. (2001a) The uniqueness of being a neurotrophin receptor. Curr. Opin. Neurobiol., 11: 281–286. Lee, J., Duan, W. and Mattson, M.P. (2002) Evidence that brain-derived neurotrophic factor is required for basal neurogenesis and mediates, in part, the enhancement of neurogenesis by dietary restriction in the hippocampus of adult mice. J. Neurochem., 82: 1367–1375. Lee, R., Kermani, P., Teng, K.K. and Hempstead, B.L. (2001b) Regulation of cell survival by secreted proneurotrophins. Science, 294: 1945–1948.

Lessmann, V., Gottmann, K. and Heumann, R. (1994) BDNF and NT-4/5 enhance glutamatergic synaptic transmission in cultured hippocampal neurones. Neuroreport, 6: 21–25. Levi-Montalcini, R. and Hamburger, V. (1951) Selective growth-stimulating effects of mouse sarcoma on the sensory and sympathetic nervous system of the chick embryo. J. Exp. Zool., 116: 321–361. Li, S., Saragovi, H.U., Nedev, H., Zhao, C., Racine, R.J. and Fahnestock, M. (2005) Differential actions of nerve growth factor receptors TrkA and p75NTR in a rat model of epileptogenesis. Mol. Cell Neurosci., 29: 162–172. Li, S., Uri Saragovi, H., Racine, R.J. and Fahnestock, M. (2003) A ligand of the p65/p95 receptor suppresses perforant path kindling, kindling-induced mossy fiber sprouting, and hilar area changes in adult rats. Neuroscience, 119: 1147–1156. Lin, S.Y., Wu, K., Levine, E.S., Mount, H.T., Suen, P.C. and Black, I.B. (1998) BDNF acutely increases tyrosine phosphorylation of the NMDA receptor subunit 2B in cortical and hippocampal postsynaptic densities. Brain Res. Mol. Brain Res., 55: 20–27. Lindholm, D., Castren, E., Berzaghi, M., Blochl, A. and Thoenen, H. (1994) Activity-dependent and hormonal regulation of neurotrophin mRNA levels in the brain–implications for neuronal plasticity. J. Neurobiol., 25: 1362–1372. Lindvall, O., Kokaia, Z., Bengzon, J., Elmer, E. and Kokaia, M. (1994) Neurotrophins and brain insults. Trends Neurosci., 17: 490–496. Linnarsson, S., Bjorklund, A. and Ernfors, P. (1997) Learning deficit in BDNF mutant mice. Eur. J. Neurosci., 9: 2581–2587. Liu, X., Ernfors, P., Wu, H. and Jaenisch, R. (1995) Sensory but not motor neuron deficits in mice lacking NT4 and BDNF. Nature, 375: 238–241. Lohof, A.M., Ip, N.Y. and Poo, M.M. (1993) Potentiation of developing neuromuscular synapses by the neurotrophins NT-3 and BDNF. Nature, 363: 350–353. Lowenstein, D.H., Seren, M.S. and Longo, F.M. (1993) Prolonged increases in neurotrophic activity associated with kainate-induced hippocampal synaptic reorganization. Neuroscience, 56: 597–604. Luikart, B.W., Nef, S., Shipman, T. and Parada, L.F. (2003) In vivo role of truncated trkb receptors during sensory ganglion neurogenesis. Neuroscience, 117: 847–858. Lyons, W.E., Mamounas, L.A., Ricaurte, G.A., Coppola, V., Reid, S.W., Bora, S.H., Wihler, C., Koliatsos, V.E. and Tessarollo, L. (1999) Brain-derived neurotrophic factor-deficient mice develop aggressiveness and hyperphagia in conjunction with brain serotonergic abnormalities. Proc. Natl. Acad. Sci. U.S.A., 96: 15239–15244. Maisonpierre, P.C., Belluscio, L., Friedman, B., Alderson, R.F., Wiegand, S.J., Furth, M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990a) NT-3, BDNF, and NGF in the developing rat nervous system: parallel as well as reciprocal patterns of expression. Neuron, 5: 501–509. Maisonpierre, P.C., Belluscio, L., Squinto, S., Ip, N.Y., Furth, M.E., Lindsay, R.M. and Yancopoulos, G.D. (1990b)

393 Neurotrophin-3: a neurotrophic factor related to NGF and BDNF. Science, 247: 1446–1451. Malcangio, M. and Lessmann, V. (2003) A common thread for pain and memory synapses? Brain-derived neurotrophic factor and trkB receptors. Trends Pharmacol. Sci., 24: 116–121. Mamounas, L.A., Blue, M.E., Siuciak, J.A. and Altar, C.A. (1995) Brain-derived neurotrophic factor promotes the survival and sprouting of serotonergic axons in rat brain. J. Neurosci., 15: 7929–7939. Marksteiner, J., Ortler, M., Bellmann, R. and Sperk, G. (1990) Neuropeptide Y biosynthesis is markedly induced in mossy fibers during temporal lobe epilepsy of the rat. Neurosci. Lett., 112: 143–148. Martinowich, K., Hattori, D., Wu, H., Fouse, S., He, F., Hu, Y., Fan, G. and Sun, Y.E. (2003) DNA methylation-related chromatin remodeling in activity-dependent BDNF gene regulation. Science, 302: 890–893. Marty, S., Berninger, B., Carroll, P. and Thoenen, H. (1996) GABAergic stimulation regulates the phenotype of hippocampal interneurons through the regulation of brain-derived neurotrophic factor. Neuron, 16: 565–570. Mathern, G.W., Babb, T.L., Micevych, P.E., Blanco, C.E. and Pretorius, J.K. (1997) Granule cell mRNA levels for BDNF, NGF, and NT-3 correlate with neuron losses or supragranular mossy fiber sprouting in the chronically damaged and epileptic human hippocampus. Mol. Chem. Neuropathol., 30: 53–76. McAllister, A.K., Katz, L.C. and Lo, D.C. (1997) Opposing roles for endogenous BDNF and NT-3 in regulating cortical dendritic growth. Neuron, 18: 767–778. McLean Bolton, M., Pittman, A.J. and Lo, D.C. (2000) Brainderived neurotrophic factor differentially regulates excitatory and inhibitory synaptic transmission in hippocampal cultures. J. Neurosci., 20: 3221–3232. Merlio, J.P., Ernfors, P., Jaber, M. and Persson, H. (1992) Molecular cloning of rat trkC and distribution of cells expressing messenger RNAs for members of the trk family in the rat central nervous system. Neuroscience, 51: 513–532. Merlio, J.P., Ernfors, P., Kokaia, Z., Middlemas, D.S., Bengzon, J., Kokaia, M., Smith, M.L., Seisjo, B.K., Hunter, T. and Lindvall, O. (1993) Increased production of the trkB protein tyrosine kinase receptor after brain insults. Neuron, 10: 151–164. Metsis, M., Timmusk, T., Arenas, E. and Persson, H. (1993) Differential usage of multiple brain-derived neurotrophic factor promoters in the rat brain following neuronal activation. Proc. Natl. Acad. Sci. U.S.A., 90: 8802–8806. Miller, F.D. and Kaplan, D.R. (2001) On Trk for retrograde signaling. Neuron, 32: 767–770. Minichiello, L., Calella, A.M., Medina, D.L., Bonhoeffer, T., Klein, R. and Korte, M. (2002) Mechanism of TrkB-mediated hippocampal long-term potentiation. Neuron, 36: 121–137. Minichiello, L., Korte, M., Wolfer, D., Kuhn, R., Unsicker, K., Cestari, V., Rossi-Arnaud, C., Lipp, H.P., Bonhoeffer, T. and Klein, R. (1999) Essential role for TrkB receptors in hippocampus-mediated learning. Neuron, 24: 401–414.

Morimoto, K., Sato, K., Sato, S., Yamada, N. and Hayabara, T. (1998) Time-dependent changes in neurotrophic factor mRNA expression after kindling and long-term potentiation in rats. Brain Res. Bull., 45: 599–605. Mowla, S.J., Farhadi, H.F., Pareek, S., Atwal, J.K., Morris, S.J., Seidah, N.G. and Murphy, R.A. (2001) Biosynthesis and post-translational processing of the precursor to brain-derived neurotrophic factor. J. Biol. Chem., 276: 12660–12666. Mowla, S.J., Pareek, S., Farhadi, H.F., Petrecca, K., Fawcett, J.P., Seidah, N.G., Morris, S.J., Sossin, W.S. and Murphy, R.A. (1999) Differential sorting of nerve growth factor and brain-derived neurotrophic factor in hippocampal neurons. J. Neurosci., 19: 2069–2080. Mudo, G., Jiang, X.H., Timmusk, T., Bindoni, M. and Belluardo, N. (1996) Change in neurotrophins and their receptor mRNAs in the rat forebrain after status epilepticus induced by pilocarpine. Epilepsia, 37: 198–207. Mudo, G., Salin, T., Condorelli, D.F., Jiang, X.H., Dell’Albani, P., Timmusk, T., Metsis, M., Funakoshi, H. and Belluardo, N. (1995) Seizures increase trkC mRNA expression in the dentate gyrus of rat hippocampus. Role of glutamate receptor activation. J. Mol. Neurosci., 6: 11–22. Murer, M.G., Yan, Q. and Raisman-Vozari, R. (2001) Brainderived neurotrophic factor in the control human brain, and in Alzheimer’s disease and Parkinson’s disease. Prog. Neurobiol., 63: 71–124. Murphy, D.D., Cole, N.B. and Segal, M. (1998) Brain-derived neurotrophic factor mediates estradiol-induced dendritic spine formation in hippocampal neurons. Proc. Natl. Acad. Sci. U.S.A., 95: 11412–11417. Nakata, K., Ujike, H., Sakai, A., Uchida, N., Nomura, A., Imamura, T., Katsu, T., Tanaka, Y., Hamamura, T. and Kuroda, S. (2003) Association study of the brain-derived neurotrophic factor (BDNF) gene with bipolar disorder. Neurosci. Lett., 337: 17–20. Narisawa-Saito, M. and Nawa, H. (1996) Differential regulation of hippocampal neurotrophins during aging in rats. J. Neurochem., 67: 1124–1131. Nawa, H., Carnahan, J. and Gall, C. (1995) BDNF protein measured by a novel enzyme immunoassay in normal brain and after seizure: partial disagreement with mRNA levels. Eur. J. Neurosci., 7: 1527–1535. Nawa, H., Pelleymounter, M.A. and Carnahan, J. (1994) Intraventricular administration of BDNF increases neuropeptide expression in newborn rat brain. J. Neurosci., 14: 3751–3765. Neeper, S.A., Gomez-Pinilla, F., Choi, J. and Cotman, C. (1995) Exercise and brain neurotrophins. Nature, 373: 109. Neves-Pereira, M., Mundo, E., Muglia, P., King, N., Macciardi, F. and Kennedy, J.L. (2002) The brain-derived neurotrophic factor gene confers susceptibility to bipolar disorder: evidence from a family-based association study. Am. J. Hum. Genet., 71: 651–655. Nibuya, M., Morinobu, S. and Duman, R.S. (1995) Regulation of BDNF and trkB mRNA in rat brain by chronic electroconvulsive seizure and antidepressant drug treatments. J. Neurosci., 15: 7539–7547.

394 Nibuya, M., Nestler, E.J. and Duman, R.S. (1996) Chronic antidepressant administration increases the expression of cAMP response element binding protein (CREB) in rat hippocampus. J. Neurosci., 16: 2365–2372. Osehobo, P., Adams, B., Sazgar, M., Verdi, J., Racine, R. and Fahnestock, M. (1996) Effects of in vivo BDNF infusion on amygdala kindling, sprouting, and hilar area. Soc. Neurosci. Abstr., 22: 995. Patapoutian, A. and Reichardt, L.F. (2001) Trk receptors: mediators of neurotrophin action. Curr. Opin. Neurobiol., 11: 272–280. Patel, M.N. and McNamara, J.O. (1995) Selective enhancement of axonal branching of cultured dentate gyrus neurons by neurotrophic factors. Neuroscience, 69: 763–770. Patterson, S.L., Abel, T., Deuel, T.A., Martin, K.C., Rose, J.C. and Kandel, E.R. (1996) Recombinant BDNF rescues deficits in basal synaptic transmission and hippocampal LTP in BDNF knockout mice. Neuron, 16: 1137–1145. Patterson, S.L., Grover, L.M., Schwartzkroin, P.A. and Bothwell, M. (1992) Neurotrophin expression in rat hippocampal slices: a stimulus paradigm inducing LTP in CA1 evokes increases in BDNF and NT-3 mRNAs. Neuron, 9: 1081–1088. Pencea, V., Bingaman, K.D., Wiegand, S.J. and Luskin, M.B. (2001) Infusion of brain-derived neurotrophic factor into the lateral ventricle of the adult rat leads to new neurons in the parenchyma of the striatum, septum, thalamus, and hypothalamus. J. Neurosci., 21: 6706–6717. Pezet, S., Malcangio, M., Lever, I.J., Perkinton, M.S., Thompson, S.W., Williams, R.J. and McMahon, S.B. (2002) Noxious stimulation induces Trk receptor and downstream ERK phosphorylation in spinal dorsal horn. Mol. Cell Neurosci., 21: 684–695. Phillips, H.S., Hains, J.M., Armanini, M., Laramee, G.R., Johnson, S.A. and Winslow, J.W. (1991) BDNF mRNA is decreased in the hippocampus of individuals with Alzheimer’s disease. Neuron, 7: 695–702. Pioro, E.P. and Cuello, A.C. (1990) Distribution of nerve growth factor receptor-like immunoreactivity in the adult rat central nervous system. Effect of colchicine and correlation with the cholinergic system–I. Forebrain. Neuroscience, 34: 57–87. Poo, M.M. (2001) Neurotrophins as synaptic modulators. Nat. Rev. Neurosci., 2: 24–32. Qiao, X., Hughes, P.E., Venero, J.L., Dugich-Djordjevic, M.M., Nichols, N.R., Hefti, F. and Knusel, B. (1996) NT4/5 protects against adrenalectomy-induced apoptosis of rat hippocampal granule cells. Neuroreport, 7: 682–686. Qiao, X., Suri, C., Knusel, B. and Noebels, J.L. (2001) Absence of hippocampal mossy fiber sprouting in transgenic mice overexpressing brain-derived neurotrophic factor. J. Neurosci. Res., 64: 268–276. Rashid, K., Van der Zee, C.E., Ross, G.M., Chapman, C.A., Stanisz, J., Riopelle, R.J., Racine, R.J. and Fahnestock, M. (1995) A nerve growth factor peptide retards seizure development and inhibits neuronal sprouting in a rat model of epilepsy. Proc. Natl. Acad. Sci. U.S.A., 92: 9495–9499.

Reibel, S., Larmet, Y., Carnahan, J., Marescaux, C. and Depaulis, A. (2000a) Endogenous control of hippocampal epileptogenesis: a molecular cascade involving brain-derived neurotrophic factor and neuropeptide Y. Epilepsia, 41(Suppl 6): S127–S133. Reibel, S., Larmet, Y., Le, B.T., Carnahan, J., Marescaux, C. and Depaulis, A. (2000b) Brain-derived neurotrophic factor delays hippocampal kindling in the rat. Neuroscience, 100: 777–788. Riccio, A., Pierchala, B.A., Ciarallo, C.L. and Ginty, D.D. (1997) An NGF-trkA-mediated retrograde signal to transcription factor CREB in sympathetic neurons. Science, 277: 1097–1100. Rios, M., Fan, G., Fekete, C., Kelly, J., Bates, B., Kuehn, R., Lechan, R.M. and Jaenisch, R. (2001) Conditional deletion of brain-derived neurotrophic factor in the postnatal brain leads to obesity and hyperactivity. Mol. Endocrinol., 15: 1748–1757. Rivera, C., Li, H., Thomas-Crusells, J., Lahtinen, H., Viitanen, T., Nanobashvili, A., Kokaia, Z., Airaksinen, M.S., Voipio, J., Kaila, K. and Saarma, M. (2002) BDNF-induced TrkB activation down-regulates the K+-Cl- cotransporter KCC2 and impairs neuronal Cl- extrusion. J. Cell Biol., 159: 747–752. Roback, J.D., Marsh, H.N., Downen, M., Palfrey, H.C. and Wainer, B.H. (1995) BDNF-activated signal transduction in rat cortical glial cells. Eur. J. Neurosci., 7: 849–862. Rocamora, N., Welker, E., Pascual, M. and Soriano, E. (1996) Upregulation of BDNF mRNA expression in the barrel cortex of adult mice after sensory stimulation. J. Neurosci., 16: 4411–4419. Rose, C.R., Blum, R., Pichler, B., Lepier, A., Kafitz, K.W. and Konnerth, A. (2003) Truncated TrkB-T1 mediates neurotrophin-evoked calcium signalling in glia cells. Nature, 426: 74–78. Roux, P.P., Colicos, M.A., Barker, P.A. and Kennedy, T.E. (1999) p75 neurotrophin receptor expression is induced in apoptotic neurons after seizure. J. Neurosci., 19: 6887–6896. Rudge, J.S., Mather, P.E., Pasnikowski, E.M., Cai, N., Corcoran, T., Acheson, A., Anderson, K., Lindsay, R.M. and Wiegand, S.J. (1998) Endogenous BDNF protein is increased in adult rat hippocampus after a kainic acid induced excitotoxic insult but exogenous BDNF is not neuroprotective. Exp. Neurol., 149: 398–410. Russo-Neustadt, A., Beard, R.C. and Cotman, C.W. (1999) Exercise, antidepressant medications, and enhanced brain derived neurotrophic factor expression. Neuropsychopharmacology, 21: 679–682. Ryden, M. and Ibanez, C.F. (1996) Binding of neurotrophin-3 to p75LNGFR, TrkA, and TrkB mediated by a single functional epitope distinct from that recognized by trkC. J. Biol. Chem., 271: 5623–5627. Rylett, R.J. and Williams, L.R. (1994) Role of neurotrophins in cholinergic-neurone function in the adult and aged CNS. Trends Neurosci., 17: 486–490. Saarelainen, T., Hendolin, P., Lucas, G., Koponen, E., Sairanen, M., MacDonald, E., Agerman, K., Haapasalo, A.,

395 Nawa, H., Aloyz, R., Ernfors, P. and Castren, E. (2003) Activation of the TrkB neurotrophin receptor is induced by antidepressant drugs and is required for antidepressant-induced behavioral effects. J. Neurosci., 23: 349–357. Saarelainen, T., Lukkarinen, J.A., Koponen, S., Grohn, O.H., Jolkkonen, J., Koponen, E., Haapasalo, A., Alhonen, L., Wong, G., Koistinaho, J., Kauppinen, R.A. and Castren, E. (2000a) Transgenic mice overexpressing truncated trkB neurotrophin receptors in neurons show increased susceptibility to cortical injury after focal cerebral ischemia. Mol. Cell Neurosci., 16: 87–96. Saarelainen, T., Pussinen, R., Koponen, E., Alhonen, L., Wong, G., Sirvio, J. and Castren, E. (2000b) Transgenic mice overexpressing truncated trkB neurotrophin receptors in neurons have impaired long-term spatial memory but normal hippocampal LTP. Synapse, 38: 102–104. Sato, K., Kashihara, K., Morimoto, K. and Hayabara, T. (1996) Regional increases in brain-derived neurotrophic factor and nerve growth factor mRNAs during amygdaloid kindling, but not in acidic and basic fibroblast growth factor mRNAs. Epilepsia, 37: 6–14. Scharfman, H., Goodman, J., Macleod, A., Phani, S., Antonelli, C. and Croll, S. (2005) Increased neurogenesis and the ectopic granule cells after intrahippocampal BDNF infusion in adult rats. Exp. Neurol., 192: 348–356. Scharfman, H.E. (1997) Hyperexcitability in combined entorhinal/hippocampal slices of adult rat after exposure to brain-derived neurotrophic factor. J. Neurophysiol., 78: 1082–1095. Scharfman, H.E. (2004) Functional implications of seizure-induced neurogenesis. Adv. Exp. Med. Biol., 548: 192–212. Scharfman, H.E., Goodman, J.H. and Sollas, A.L. (1999) Actions of brain-derived neurotrophic factor in slices from rats with spontaneous seizures and mossy fiber sprouting in the dentate gyrus. J. Neurosci., 19: 5619–5631. Scharfman, H.E., Goodman, J.H., Sollas, A.L. and Croll, S.D. (2002) Spontaneous limbic seizures after intrahippocampal infusion of brain-derived neurotrophic factor. Exp. Neurol., 174: 201–214. Scharfman, H.E., Mercurio, T.C., Goodman, J.H., Wilson, M.A. and MacLusky, N.J. (2003) Hippocampal excitability increases during the estrous cycle in the rat: a potential role for brain-derived neurotrophic factor. J. Neurosci., 23: 11641–11652. Schinder, A.F. and Poo, M. (2000) The neurotrophin hypothesis for synaptic plasticity. Trends Neurosci., 23: 639–645. Schmidt-Kastner, R., Humpel, C., Wetmore, C. and Olson, L. (1996) Cellular hybridization for BDNF, trkB, and NGF mRNAs and BDNF-immunoreactivity in rat forebrain after pilocarpine-induced status epilepticus. Exp. Brain Res., 107: 331–347. Schmidt-Kastner, R. and Olson, L. (1995) Decrease of neurotrophin-3 mRNA in adult rat hippocampus after pilocarpine seizures. Exp. Neurol., 136: 199–204. Segal, R.A. (2003) Selectivity in neurotrophin signaling: theme and variations. Annu. Rev. Neurosci., 26: 299–330.

Senger, D.L. and Campenot, R.B. (1997) Rapid retrograde tyrosine phosphorylation of trkA and other proteins in rat sympathetic neurons in compartmented cultures. J. Cell Biol., 138: 411–421. Shieh, P.B. and Ghosh, A. (1999) Molecular mechanisms underlying activity-dependent regulation of BDNF expression. J. Neurobiol., 41: 127–134. Shieh, P.B., Hu, S.C., Bobb, K., Timmusk, T. and Ghosh, A. (1998) Identification of a signaling pathway involved in calcium regulation of BDNF expression. Neuron, 20: 727–740. Shirayama, Y., Chen, A.C., Nakagawa, S., Russell, D.S. and Duman, R.S. (2002) Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depression. J. Neurosci., 22: 3251–3261. Simonato, M., Bregola, G., Armellin, M., Del Piccolo, P., Rodi, D., Zucchini, S. and Tongiorgi, E. (2002) Dendritic targeting of mRNAs for plasticity genes in experimental models of temporal lobe epilepsy. Epilepsia, 43(Suppl 5): 153–158. Sklar, P., Gabriel, S.B., McInnis, M.G., Bennett, P., Lim, Y.M., Tsan, G., Schaffner, S., Kirov, G., Jones, I., Owen, M., Craddock, N., DePaulo, J.R. and Lander, E.S. (2002) Family-based association study of 76 candidate genes in bipolar disorder: BDNF is a potential risk locus. Brain-derived neutrophic factor. Mol. Psychiatry, 7: 579–593. Smith, M.A., Makino, S., Kvetnansky, R. and Post, R.M. (1995) Stress and glucocorticoids affect the expression of brain-derived neurotrophic factor and neurotrophin-3 mRNAs in the hippocampus. J. Neurosci., 15: 1768–1777. Smith, M.A., Zhang, L.-X., Lyons, W.E. and Mamounas, L.A. (1997) Anterograde transport of endogenous brain-derived neurotrophic factor in hippocampal mossy fibers. Neuroreport, 8: 1829–1834. Sobreviela, T., Clary, D.O., Reichardt, L.F., Brandabur, M.M., Kordower, J.H. and Mufson, E.J. (1994) TrkA-immunoreactive profiles in the central nervous system: colocalization with neurons containing p75 nerve growth factor receptor, choline acetyltransferase, and serotonin. J. Comp. Neurol., 350: 587–611. Spires, T.L., Grote, H.E., Varshney, N.K., Cordery, P.M., van Dellen, A., Blakemore, C. and Hannan, A.J. (2004) Environmental enrichment rescues protein deficits in a mouse model of Huntington’s disease, indicating a possible disease mechanism. J. Neurosci., 24: 2270–2276. Spruston, N., Lubke, J. and Frotscher, M. (1997) Interneurons in the stratum lucidum of the rat hippocampus: an anatomical and electrophysiological characterization. J. Comp. Neurol., 385: 427–440. Suen, P.-C., Wu, K., Levine, E.S., Mount, H.T.J., Xu, J.-L., Lin, S.-Y. and Black, I.B. (1997) Brain-derived neurotrophic factor rapidly enhances phosphorylation of the postsynaptic N-methyl-D-aspartate receptor subunit 1. Proc. Natl. Acad. Sci. U.S.A., 94: 8191–8195. Takahashi, M., Hayashi, S., Kakita, A., Wakabayashi, K., Fukuda, M., Kameyama, S., Tanaka, R., Takahashi, H. and Nawa, H. (1999) Patients with temporal lobe epilepsy show an increase in brain-derived neurotrophic factor protein and

396 its correlation with neuropeptide Y. Brain Res., 818: 579–582. Tanaka, T., Saito, H. and Matsuki, N. (1997) Inhibition of GABAa synaptic responses by brain-derived neurotrophic factor (BDNF) in rat hippocampus. J. Neurosci., 17: 2959–2966. Tao, X., Finkbeiner, S., Arnold, D.B., Shaywitz, A.J. and Greenberg, M.E. (1998) Ca2+ influx regulates BDNF transcription by a CREB family transcription factor-dependent mechanism. Neuron, 20: 709–726. Tao, X., West, A.E., Chen, W.G., Corfas, G. and Greenberg, M.E. (2002) A calcium-responsive transcription factor, CaRF, that regulates neuronal activity-dependent expression of BDNF. Neuron, 33: 383–395. Thakker-Varia, S., Alder, J., Crozier, R.A., Plummer, M.R. and Black, I.B. (2001) Rab3A is required for brain-derived neurotrophic factor-induced synaptic plasticity: transcriptional analysis at the population and single-cell levels. J. Neurosci., 21: 6782–6790. Thoenen, H. (1995) Neurotrophins and neuronal plasticity. Science, 270: 593–598. Thompson, S.W., Bennett, D.L., Kerr, B.J., Bradbury, E.J. and McMahon, S.B. (1999) Brain-derived neurotrophic factor is an endogenous modulator of nociceptive responses in the spinal cord. Proc. Natl. Acad. Sci. U.S.A., 96: 7714–7718. Timmusk, T., Belluardo, N., Metsis, M. and Persson, H. (1993a) Widespread and developmentally regulated expression of neurotrophin-4 mRNA in rat brain and peripheral tissues. Eur. J. Neurosci., 5: 605–613. Timmusk, T., Palm, K., Metsis, M., Reintam, T., Paalme, V., Saarma, M. and Persson, H. (1993b) Multiple promoters direct tissue-specific expression of the rat BDNF gene. Neuron, 10: 475–489. Tolwani, R.J., Buckmaster, P.S., Varma, S., Cosgaya, J.M., Wu, Y., Suri, C. and Shooter, E.M. (2002) BDNF overexpression increases dendrite complexity in hippocampal dentate gyrus. Neuroscience, 114: 795–805. Tønder, N., Kragh, J., Finsen, B., Bolwig, T.G. and Zimmer, J. (1994) Kindling induces transient changes in neuronal expression of somatostatin, neuropeptide Y, and calbindin in adult rat hippocampus and fascia dentata. Epilepsia, 35: 1299–1308. Tonra, J.R., Curtis, R., Wong, V., Cliffer, K.D., Park, J.S., Timmes, A., Nguyen, T., Lindsay, R.M., Acheson, A. and DiStefano, P.S. (1998) Axotomy upregulates the anterograde transport and expression of brain-derived neurotrophic factor by sensory neurons. J. Neurosci., 18: 4374–4383. Tsai, S.J. (2004) Is mania caused by overactivity of central brain-derived neurotrophic factor? Med. Hypotheses, 62: 19–22. Tyler, W.J., Perrett, S.P. and Pozzo-Miller, L.D. (2002) The role of neurotrophins in neurotransmitter release. Neuroscientist, 8: 524–531. Urfer, R., Tsoulfas, P., O’Connell, L., Shelton, D.L., Parada, L.F. and Presta, L.G. (1995) An immunoglobulin-like domain determines the specificity of neurotrophin receptors. EMBO J., 14: 2795–2805.

Van der Zee, C.E., Rashid, K., Le, K., Moore, K.A., Stanisz, J., Diamond, J., Racine, R.J. and Fahnestock, M. (1995) Intraventricular administration of antibodies to nerve growth factor retards kindling and blocks mossy fiber sprouting in adult rats. J. Neurosci., 15: 5316–5323. Vezzani, A., Ravizza, T., Moneta, D., Conti, M., Borroni, A., Rizzi, M., Samanin, R. and Maj, R. (1999) Brainderived neurotrophic factor immunoreactivity in the limbic system of rats after acute seizures and during spontaneo us convulsions: temporal evolution of changes as compared to neuropeptide Y. Neuroscience, 90: 1445–1461. Von Bartheld, C.S., Byers, M.R., Williams, R. and Bothwell, M. (1996a) Anterograde transport of neurotrophins and axodendritic transfer in the developing visual system. Nature, 379: 830–833. Von Bartheld, C.S., Williams, R., Lefcort, F., Clary, D.O., Reichardt, L.F. and Bothwell, M. (1996b) Retrograde transport of neurotrophins from the eye to the brain in chick embryos: roles of the p75NTR and trkB receptors. J. Neurosci., 16: 2995–3008. Wardle, R.A. and Poo, M.M. (2003) Brain-derived neurotrophic factor modulation of GABAergic synapses by postsynaptic regulation of chloride transport. J. Neurosci., 23: 8722–8732. Wetmore, C., Ernfors, P., Persson, H. and Olson, L. (1990) Localization of brain-derived neurotrophic factor mRNA to neurons in the brain by in situ hybridization. Exp. Neurol., 109: 141–152. Wu, K., Xu, J., Suen, P., Levine, E., Huang, Y., Mount, H.T.J., Lin, S. and Black, I.B. (1996) Functional trkB neurotrophin receptors are intrinsic components of the adult brain postsynaptic density. Mol. Brain Res., 43: 286–290. Xu, B., Gottschalk, W., Chow, A., Wilson, R.I., Schnell, E., Zang, K., Wang, D., Nicoll, R.A., Lu, B. and Reichardt, L.F. (2000) The role of brain-derived neurotrophic factor receptors in the mature hippocampus: modulation of long-term potentiation through a presynaptic mechanism involving TrkB. J. Neurosci., 20: 6888–6897. Xu, B., Goulding, E.H., Zang, K., Cepoi, D., Cone, R.D., Jones, K.R., Tecott, L.H. and Reichardt, L.F. (2003) Brainderived neurotrophic factor regulates energy balance downstream of melanocortin-4 receptor. Nat. Neurosci., 6: 736–742. Xu, B., Michalski, B., Racine, R.J. and Fahnestock, M. (2002) Continuous infusion of neurotrophin-3 triggers sprouting, decreases the levels of TrkA and TrkC, and inhibits epileptogenesis and activity-dependent axonal growth in adult rats. Neuroscience, 115: 1295–1308. Xu, B., Michalski, B., Racine, R.J. and Fahnestock, M. (2004) The effects of brain-derived neurotrophic factor (BDNF) administration on kindling induction, Trk expression and seizure-related morphological changes. Neuroscience, 126: 521–531. Yacoubian, T.A. and Lo, D.C. (2000) Truncated and fulllength TrkB receptors regulate distinct modes of dendritic growth. Nat. Neurosci., 3: 342–349.

397 Yamada, K. and Nabeshima, T. (2003) Brain-derived neurotrophic factor/TrkB signaling in memory processes. J. Pharmacol. Sci., 91: 267–270. Yan, Q., Matheson, C., Sun, J., Radeke, M.J., Feinstein, S.C. and Miller, J.A. (1994) Distribution of intracerebral ventricularly administered neurotrophins in rat brain and its correlation with trk receptor expression. Exp. Neurol., 127: 23–36. Yan, Q., Radeke, M.J., Matheson, C.R., Talvenheimo, J., Welcher, A.A. and Feinstein, S.C. (1997a) Immunocytochemical localization of trkB in the central nervous system of the adult rat. J. Comp. Neurol., 378: 135–157. Yan, Q., Rosenfeld, R.D., Matheson, C.R., Hawkins, N., Lopez, O.T., Bennett, L. and Welcher, A.A. (1997b) Expression of brain-derived neurotrophic factor protein in the adult rat central nervous system. Neuroscience, 78: 431–448. Zaccaro, M.C., Ivanisevic, L., Perez, P., Meakin, S.O. and Saragovi, H.U. (2001) p75 Co-receptors regulate liganddependent and ligand-independent Trk receptor activation, in part by altering Trk docking subdomains. J. Biol. Chem., 276: 31023–31029. Zhou, X.F. and Rush, R.A. (1994) Localization of neurotrophin-3-like immunoreactivity in the rat central nervous system. Brain Res., 643: 162–172.

Zhou, X.-F. and Rush, R.A. (1996) Endogenous brain-derived neurotrophic factor is anterogradely transported in primary sensory neurons. Neuroscience, 74: 945–951. Zhou, Z., Hong, E.J., Cohen, S., Zhao, W.N., Ho, H.Y., Schmidt, L., Chen, W.G., Lin, Y., Savner, E., Griffith, E.C., Hu, L., Steen, J.A., Weitz, C.J. and Greenberg, M.E. (2006) Brain-specific phosphorylation of MeCP2 regulates activitydependent Bdnf transcription, dendritic growth, and spine maturation. Neuron, 52: 255–269. Zigova, T., Pencea, V., Wiegand, S.J. and Luskin, M.B. (1998) Intraventricular administration of BDNF increases the number of newly generated neurons in the adult olfactory bulb. Mol. Cell Neurosci., 11: 234–245. Zuccato, C., Ciammola, A., Rigamonti, D., Leavitt, B.R., Goffredo, D., Conti, L., MacDonald, M.E., Friedlander, R.M., Silani, V., Hayden, M.R., Timmusk, T., Sipione, S. and Cattaneo, E. (2001) Loss of huntingtin-mediated BDNF gene transcription in Huntington’s disease. Science, 293: 493–498. Zuccato, C., Tartari, M., Crotti, A., Goffredo, D., Valenza, M., Conti, L., Cataudella, T., Leavitt, B.R., Hayden, M.R., Timmusk, T., Rigamonti, D. and Cattaneo, E. (2003) Huntingtin interacts with REST/NRSF to modulate the transcription of NRSE-controlled neuronal genes. Nat. Genet., 35: 76–83.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 23

Sex steroids and the dentate gyrus Tibor Hajszan1,2,, Teresa A. Milner3,4 and Csaba Leranth1,5 1

Department of Obstetrics, Gynecology, and Reproductive Sciences, Yale University School of Medicine, 333 Cedar Street, FMB 312, New Haven, CT 06520, USA 2 Department of Biophysics, Biological Research Center, Hungarian Academy of Sciences, Szeged, Hungary 3 Department of Neurology and Neuroscience, Division of Neurobiology, Weill-Cornell Medical College, 411 East 69th Street, New York, NY 10021, USA 4 Harold and Milliken Hatch Laboratory of Neuroendocrinology, The Rockefeller University, New York, NY, USA 5 Department of Neurobiology, Yale University School of Medicine, New Haven, CT, USA

Abstract: In the late 1980s, the finding that the dentate gyrus contains more granule cells in the male than in the female of certain mouse strains provided the first indication that the dentate gyrus is a significant target for the effects of sex steroids during development. Gonadal hormones also play a crucial role in shaping the function and morphology of the adult brain. Besides reproduction-related processes, sex steroids participate in higher brain operations such as cognition and mood, in which the hippocampus is a critical mediator. Being part of the hippocampal formation, the dentate gyrus is naturally involved in these mechanisms and as such, this structure is also a critical target for the activational effects of sex steroids. These activational effects are the results of three major types of steroid-mediated actions. Sex steroids modulate the function of dentate neurons under normal conditions. In addition, recent research suggests that hormone-induced cellular plasticity may play a larger role than previously thought, particularly in the dentate gyrus. Specifically, the regulation of dentate gyrus neurogenesis and synaptic remodeling by sex steroids received increasing attention lately. Finally, the dentate gyrus is influenced by gonadal hormones in the context of cellular injury, and the work in this area demonstrates that gonadal hormones have neuroprotective potential. The expression of estrogen, progestin, and androgen receptors in the dentate gyrus suggests that sex steroids, which could be of gonadal origin and/or synthesized locally in the dentate gyrus, may act directly on dentate cells. In addition, gonadal hormones could also influence the dentate gyrus indirectly, by subcortical hormone-sensitive structures such as the cholinergic septohippocampal system. Importantly, these three sex steroid-related themes, functional effects in the normal dentate gyrus, mechanisms involving neurogenesis and synaptic remodeling, as well as neuroprotection, have substantial implications for understanding normal cognitive function, with clinical importance for epilepsy, Alzheimer’s disease and mental disorders. Keywords: androgen; estrogen; progesterone; sex difference; electrophysiology; neurogenesis; synaptic remodeling; neuroprotection Introduction In 1989, Wimer and Wimer (1989) reported that in certain mouse strains, the dentate gyrus contains

Corresponding author. Tel.: +1 203 785 4748; Fax: +1 203 785 7684; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63023-4

399

400

more granule cells in the male than in the female. A few years later, the dentate granule cell layer was shown to be larger in males relative to females, both in adult and prepubescent rats, which is well correlated with performance in spatial memory tasks (Roof and Havens, 1992; Roof, 1993). Later, in the hilus of the dentate gyrus, the number of synapses formed by mossy fibers, the axons of dentate granule cells, was demonstrated to be higher in male rats than in females, consistent with the idea that more granule cells in males provide a more robust input to CA3 pyramidal neurons (Parducz and GarciaSegura, 1993). Using rigorous stereology techniques, however, two studies performed later have found no sexual dimorphism in the volume of the dentate granule cell layer in Sprague-Dawley (Isgor and Sengelaub, 1998) and Long-Evans rats (Jones and Watson, 2005). Paying particular attention to rodent strains in this respect is emphasized by another mouse study, demonstrating that the overall volume of the dentate granule cell layer in A/J mice is larger in males than in females, while there is no such sexual dimorphism in C57Bl/6J mice (Tabibnia et al., 1999). These key developments in our understanding of sex differences in the dentate gyrus were followed by other examples of sexual dimorphism in certain aspects of dentate gyrus function and morphology. For example, female rats produce more newly born cells than males in the dentate gyrus, but not in the subventricular zone (Tanapat et al., 1999). Expression of proteins regulated by estrogen, such as brain-derived neurotrophic factor (BDNF) (Sohrabji et al., 1995), also differs in males vs. females. Using immunocytochemistry, it appears that the mossy fibers contain the vast majority of BDNF protein (Conner et al., 1997). When a proestrous or estrous female rat with high estrogen levels was compared to a metestrous female, ovariectomized female, or male, which have low estrogen blood concentrations, BDNF protein expression in the mossy fibers was relatively low (Scharfman et al., 2003). These studies are consistent with BDNF expression in mossy fibercontaining micropunches of CA3 assayed by ELISA, demonstrating a higher level of BDNF in the female relative to the male rat (Franklin and Perrot-Sinal, 2006).

The finding that prepubescent female rats injected neonatally with testosterone develop a malelike dentate gyrus and perform better in the Morris water maze (Roof, 1993) suggests that the ‘‘gender’’ of the dentate gyrus and many other sexually dimorphic brain functions and structures could be readily manipulated via the hormonal milieu. The window during which these hormonal manipulations can affect dentate gyrus morphology seems to be restricted to the first few postnatal days (Roof, 1993), because no morphological responses to sex steroid treatment in the rat dentate gyrus were found before or after this period (Isgor and Sengelaub, 1998). The findings of Roof (1993) support the so-called ‘‘aromatization hypothesis,’’ i.e., the brain is masculinized by estrogen that is produced by aromatase from testosterone, which is readily available in the developing male but not in the female (Naftolin et al., 1975). Also consistent with this hypothesis, estrogen receptor (ER) mRNA levels in the dentate gyrus increase significantly between birth and postnatal day 4, and then decline by postnatal day 10, while adult male rat ER mRNA levels are similar to those found in newborn and postnatal day 10 animals (O’Keefe et al., 1995). On the other hand, it has been reported that the dentate granule cell layer is significantly larger in testicular feminization mutant (Tfm) male rats than in wild-type females (Jones and Watson, 2005). Because Tfm rats express a dysfunctional androgen receptor (AR), this finding implicates the AR in the development of the dentate gyrus, which contradicts the aromatization hypothesis. It should be noted, however, that Tfm rats retain a considerable portion of AR activity, whereas the dentate granule cell layer is not sexually dimorphic in Tfm mice with complete deletion of AR function (Tabibnia et al., 1999).

Sex steroids and dentate physiology The findings of sexual dimorphism strongly suggest that sex steroids, and their receptors, are important regulators of dentate gyrus organization, and emphasize their important role in development. The rest of this review, however, will focus on more recent and exciting research on the

401

activational effects of gonadal hormones in the adult brain. In adulthood, sex steroids participate not only in reproduction-related central actions, but also in higher brain functions such as cognition and mood regulation (Korol and Kolo, 2002; Seidman, 2003; Steiner et al., 2003; MacLusky et al., 2006). Because the hippocampal formation plays an essential role in declarative, spatial, and contextual memory, as well as in the regulation of mood and the hypothalamic-pituitary-adrenal axis, this limbic structure is a critical target of hormone action (McEwen and Alves, 1999; McEwen, 2003). As a part of the hippocampal formation, it is highly likely that the dentate gyrus also plays an important role in cognitive function and mood regulation. Although research has made progress in associating hippocampal subregions with specific functional roles in complex topics such as cognition and mood, proving that any specific aspect of these behaviors is dependent on the dentate gyrus has been more difficult, probably because complex hippocampal operations require uncompromised signal flow throughout the entire hippocampal formation, rather than only in select areas. Electrophysiological studies may provide the most specific insights into the ways sex steroids influence the adult dentate gyrus. Working with rat hippocampal slices in the presence or absence of 17b-estradiol, Kim et al. (2006) have reported that 17b-estradiol significantly potentiates the amplitude and slope of field excitatory postsynaptic potentials in dentate gyrus directly, as well as in CA3 following mossy fiber stimulation. Repetitive hilar stimuli frequently evoke multiple population spikes in CA3 at proestrus and estrus, but only rarely at other cycle stages, and never in slices of ovariectomized rats (Scharfman et al., 2003). This hyperexcitability in CA3 at proestrus was blocked by exposure to the high-affinity neurotrophin receptor antagonist K252a, or by an antagonist of the a7 nicotinic cholinergic receptor, whereas it was induced at metestrus by the addition of BDNF to hippocampal slices (Scharfman et al., 2003). These findings indicate that an estrogen-induced interaction of BDNF and a7 nicotinic receptors is important for estrous cycle-related changes in CA3 and dentate gyrus (Scharfman et al., 2003).

Considering androgen, intrahippocampal microinjection of neither dehydroepiandrosteronesulfate (DHEAS) nor trilostane, an inhibitor of the enzyme that metabolizes DHEAS, alters dentate field excitatory postsynaptic potential slopes or population spike amplitudes, but increases the amplitude of a late component of the postsynaptic potential. Both DHEAS and trilostane abolishes GABA-mediated paired-pulse inhibition. In addition, both DHEAS and trilostane markedly increases the spontaneous firing rate of dentate hilar interneurons and synchronizes their firing during hippocampal theta rhythm induced by tail pinch (Steffensen, 1995). Many studies indicate that sex steroid-induced electrophysiological changes may be due to modulation of dentate glutamate and GABA receptors. Indeed, ovariectomy in rats decreases [3H] glutamate binding to N-methyl-D-aspartate (NMDA) receptors in the dentate gyrus, while hormone replacement with estradiol, tamoxifen, or raloxifene prevents this decrease (Cyr et al., 2000, 2001). On the other hand, [3H] MK-801 binding shows that the density of noncompetitive NMDA antagonist sites is significantly increased in the dentate gyrus of ovariectomized compared to sham-operated rats (El-Bakri et al., 2004). 17b-estradiol returns [3H] MK-801 binding to the normal levels, while progesterone has no effect (Weiland, 1992; El-Bakri et al., 2004). In addition, estradiol treatment of ovariectomized rats significantly increases NMDA R1 subunit protein levels in granule cell somata, in comparison with nontreated animals, without concomitant changes in the corresponding mRNA hybridization signal (Gazzaley et al., 1996). Regarding other receptor types, ovarian steroids have no effect on the density of kainate or a-amino-3-hydroxy-5methylisoxazole-4-propionic acid (AMPA) receptors (Weiland, 1992; Cyr et al., 2000), while estradiol-benzoate increases [3H] muscimol binding in the dentate gyrus (Schumacher et al., 1989). In situ hybridization showed that progesterone suppresses mRNA levels of the a1 GABAA receptor subunit in the dentate gyrus of animals that were pretreated with estradiol (Weiland and Orchinik, 1995). Finally, there is a significant negative correlation between testosterone levels and the

402

mRNA level for the a1 GABAA receptor subunit (Orchinik et al., 1995). Sex steroids and dentate plasticity Although numerous findings support the view that molecular mechanisms, including receptor changes, play a critical role in alterations of neuronal activity and functional plasticity (see above), recent evidence suggests that sex steroids may also influence long-term potentiation (Lynch, 2004) and learning/memory via mediating aspects of structural plasticity, such as neurogenesis and synaptic remodeling (Kandel, 2001; Kasai et al., 2003). Sex steroids and neurogenesis Due to the recent confirmation that neurogenesis continues throughout the lifespan in the adult dentate gyrus (Altman and Das, 1965), the subgranular zone, where progenitors are primarily generated, has drawn considerable attention. Translational implications are one reason: dentate neurogenesis may be a critical factor in the pathophysiology of depression and in the mechanism of antidepressant action (Santarelli et al., 2003). The first indication that sex steroids may influence adult neurogenesis came from Tanapat et al. (1999), who demonstrated that female rats produce more newly born cells than males in the dentate gyrus (but not in the subventricular zone). They have also reported a fluctuation in cell proliferation during the estrous cycle: females produce more newly born cells during proestrus (when estrogen levels are highest) compared with estrus and diestrus. Ovariectomy diminishes, while acute treatment with estrogen rapidly increases, cell proliferation in ovariectomized rats, an effect that is reversed by the administration of progesterone (Tanapat et al., 1999, 2005; Falconer and Galea, 2003). Both the prolonged absence of ovarian hormones and chronic treatment decreases the potential of estrogen to stimulate cell proliferation (Tanapat et al., 2005), suggesting that both doseresponse and temporal characteristics of estrogen treatment may critically influence its neurogenic

efficacy. Indeed, Ormerod et al. (2003) have reported that relative to vehicle-treated rats, the number of new cells increases following a 4-h exposure but decreases following a 48-h exposure to estrogen in ovariectomized animals. This decrease at 48 h is abolished by adrenalectomy, suggesting a role of adrenal activity (Ormerod et al., 2003). In addition to being effective under normal conditions, estrogen is capable of influencing neurogenic potential when neurogenesis is examined in other contexts. For example, a strong reduction in cell proliferation occurs in the dentate gyrus and subventricular zone of mice sacrificed 20 days after streptozotocin administration, which induces a diabetic state. This reduction is completely relieved by 10 days of estradiol pellet implantation, which increases the circulating estrogen levels 30-fold (Saravia et al., 2006). In addition, there is a striking effect of aging on cell proliferation that appears to be influenced by estradiol. PerezMartin et al. (2005) have reported that treatment of 22-month-old ovariectomized animals for 10 weeks with a weekly subcutaneous injection of estradiol-valerianate, or with soy extract added to the drinking water, reverses the age-associated decline in dentate granule cell production (PerezMartin et al., 2005). Similarly, estrogen also normalizes the deficient granule cell proliferation in the dentate gyrus of aging mice (De Nicola et al., 2006). Several studies have provided insight into the mechanisms underlying the influence of estrogen on neurogenesis. Mazzucco et al. (2006) have shown that both diarylpropionitrile, an ERb agonist, and propyl-pyrazole triol, an ERa agonist, significantly enhances cell proliferation in the dentate gyrus of female rats. Other findings suggest that ERs are involved in the induction of adult neurogenesis by an interaction with insulinlike growth factor-1 (IGF-1), as estradiol and IGF-1 have a cooperative effect to promote neurogenesis (Perez-Martin et al., 2003). Administration of IGF-1 significantly increases dentate neurogenesis compared to rats treated with vehicle; and rats treated with both IGF-1 and estradiol show a higher level of cell proliferation than rats treated with IGF-1 or estradiol alone (Perez-Martin et al., 2003).

403

Serotonin has also been linked to the effect of estradiol on dentate neurogenesis. Administration of 5-hydroxytryptophan, a precursor to serotonin, restores cell proliferation that was decreased by ovariectomy, whereas estradiol is unable to reverse this change in ovariectomized rats treated with p-chlorophenylalanine, an inhibitor of serotonin synthesis (Banasr et al., 2001). These data implicate the central serotonergic system in the mediation of estrogen effects on dentate neurogenesis. Indeed, several studies indicate that estrogen influences the dentate serotonergic system (Bowman et al., 2002). In case of 5-HT1A receptors, estradiol treatment reduces 5-HT1A gene expression in the dentate gyrus (Birzniece et al., 2001), while ovariectomy increases 5-HT1A receptor stimulation, which is reversed by estradiol (Le Saux and Di Paolo, 2005). Besides the plethora of data with respect of estrogen and neurogenesis, there are almost no published work that address the neurogenic effect of androgen except an initial study in songbird suggesting that testosterone promotes neurogenesis (Louissaint et al., 2002). Another study has demonstrated that in a neuronal stem cell culture stimulated with epidermal growth factor, nandrolone, a synthetic androgen reduces cell proliferation (Brannvall et al., 2005). The decrease is abolished by flutamide, an AR antagonist. Nandrolone also reduces new cell production in the dentate gyrus, an effect observed in both female and male rats (Brannvall et al., 2005). For more details on gonadal hormone modulation of hippocampal neurogenesis in the adult, some excellent reviews are available (Gould et al., 2000; Galea et al., 2006). For more details on adult neurogenesis and its role in mood regulation, we refer the reader to other chapters within this volume.

Sex steroids and synaptic remodeling In 1992, the discovery that estradiol mediates fluctuation in hippocampal CA1 spine synapse density during the estrous cycle in the adult rat marked the beginning of a new era in the research of sex steroids (Woolley and McEwen, 1992). Subsequent extensive work has shown that sex steroids hold an

unparalleled synaptogenic power in the adult hippocampus. For example, hormone replacement induces changes on the order of 50–100% in the number of CA1 spine synapses of rats (MacLusky et al., 2006; Parducz et al., 2006). This synaptogenic efficacy seems to be rivaled only by antidepressant drugs (Hajszan et al., 2005). Moreover, estrogen-induced remodeling of hippocampal spine synapses is remarkably rapid, similar to the time course required for long-term potentiation induction (MacLusky et al., 2005). The temporal analogy suggests that formation of spine synapses may be involved in sex steroid-modulated cognitive functions. Indeed, a great deal of evidence has accumulated which suggests that rapid remodeling and stabilization of small spines (and their associated synaptic contacts) may represent a mechanism of memory formation and storage (Sorra and Harris, 2000; Kandel, 2001; Kasai et al., 2003). Unfortunately, several studies support the view that the dentate gyrus may miss this ‘‘synaptogenic party.’’ Woolley et al. (1990) have reported no significant changes in dendritic spine density across the estrous cycle in CA3 pyramidal cells or dentate granule cells of the rat. Using a similar Golgi-impregnation technique, Gould et al. (1990) have shown that ovariectomy or gonadal steroid replacement do not affect spine density of CA3 pyramidal cells or granule cells of the dentate gyrus. In a later study, Miranda et al. (1999) have demonstrated that there may be effects in the dentate gyrus, but they are likely to be dependent on age and the temporal pattern of estradiol replacement. In addition, Szymczak et al. (2006) suggest that ERb expression is negatively correlated with synapse formation. Another approach to the topic of synaptic remodeling is to address the expression of proteins that are associated with the pre- or postsynaptic apparatus. This approach has also failed to lead to a compelling body of evidence that hormonal fluctuations modulate dentate synaptic remodeling. For example, immunoreactivity for spinophilin, a marker of dendritic spines, is increased in the hilar region of the dentate gyrus, as well as in CA3, of ovariectomized rats treated with estrogen for 2 days (Brake et al., 2001). However, levels of

404

syntaxin and synaptophysin (presynaptic proteins associated with the transmitter release machinery), as well as spinophilin, are unaltered by hormone treatment in the dentate gyrus of rhesus monkeys (Choi et al., 2003). Although young female rhesus monkeys show a trend toward an estrogen-induced increase in immunoreactive spines in the dentate gyrus outer molecular layer, this effect appears to be statistically insignificant (Hao et al., 2003). Glia have also been associated with synaptic remodeling, as expansion of the dendritic tree and spine growth may occur at the expense of shrinking glial, primarily astroglial volume. As a result, presumably, abundance of astroglial processes and markers is negatively correlated with dendritic spine density. However, the ways sex steroids alter dentate gyrus glia do not appear to be consistent. Luquin et al. (1993) have shown that the surface density of astroglial cells is positively influenced by estrogen and progesterone. The surface density of astroglial cells was significantly increased over control values by 5 h after the injection of estrogen to ovariectomized rats, and as early as 1 h after the administration of progesterone; it reached maximal values by 24 h and returned to control levels by 48 h (Luquin et al., 1993). In contrast, levels of glial fibrillary acidic protein (GFAP) intron 1, a molecular marker of adult astrocytes, shows that GFAP transcription and mRNA are both decreased in the outer molecular layer of the dentate gyrus on the afternoon of proestrus, when plasma estradiol levels are highest (Stone et al., 1998b). In vitro, astrocytes show interesting bidirectional responses, such that estrogen treatment increases GFAP transcription in monotypic astrocytic cultures but decreases GFAP transcription in astrocytes cocultured with neurons (Stone et al., 1998b). In mice, Lei et al. (2003) have reported similar findings, i.e., longterm 17b-estradiol treatment in aged female mice significantly lowered the number of astrocytes in the dentate gyrus and CA1 compared with placebo. What may be in the background of such variable findings? Besides confounding variables such as the age of animals, strain, dose and temporal characteristics of hormone treatment, and/or the

dentate area examined, the chosen methodological approach alone may be critical. What measures most reliably the remodeling of synaptic connections, i.e., synaptogenesis or loss of synapses, is debatable. Above, we list several light microscopic approaches, such as Golgi-based estimation of dendritic spine density and histochemical detection of pre- and/or postsynaptic marker molecules, which are widely applied, mainly due to their relative methodological simplicity. However, it is impossible to decide at the light microscopic level what proportion of the measured molecular synaptic markers is actually associated with synapses, and what proportion represents extrasynaptic molecules that are processed and/or stored in different cellular compartments. Therefore, levels of synaptic marker molecules may and do change without alterations in the number of synapses (Li et al., 2004). Although Golgi-based estimation of dendritic spine density is a less controversial approach, it also has several limitations. The most important is that spines may or may not form synapses, and the Golgi method is incapable of differentiating between spines that have synapses and spines that do not. Thus, similar to synaptic marker molecules, measures of dendritic spines do not necessarily reflect the true number of spine synapses. The most important point in this debate is that the number of actual spine synapses is more relevant to the functional status of neurons than the number of dendritic spines or the levels of any molecular markers. Thus, when the true number of synapses is questioned, one should count synapses themselves using electron microscopic stereological techniques, because the above-discussed light microscopic markers are not reliable. Treating ovariectomized rats with 10 mg/day subcutaneous estradiol-bezoate for 2 days, a similar schedule that has been used in previous studies (Gould et al., 1990), estrogen increased the number of spine synapses in the CA1 stratum radiatum by 71.6% over oil-treated control values, by 50.1% in the CA3 stratum radiatum, and by 99.1% in the molecular layer of the dentate gyrus (Fig. 1). It is noteworthy that the number of granule cell spine synapses in the dentate gyrus doubles after estrogen administration.

405

Fig. 1. Effects of ovariectomy and estrogen replacement on the number of hippocampal spine synapses. Young adult female Sprague-Dawley rats (250 g) were ovariectomized and one week later, they received either 10 mg/rat/day estradiol-benzoate (EB, solid columns) or 200 ml/rat/day sesame oil vehicle (oil, open columns) subcutaneously for 2 days. Two days after the last injection, the animals were sacrificed by transcardial perfusion of fixative, and their brains were processed for electron microscopic stereological analysis. Spine synapses were counted in the CA1 and CA3 strata radiata, and in the molecular layer of the dentate gyrus (DG). Significantly different from the corresponding Oil group (t-test, po0.001 in CA1 and DG, po0.01 in CA3).

Neuroprotective effects Due to the vulnerability of some types of neurons in the dentate gyrus to insults, this area is a common subject of neurodegenerative/neuroprotective experiments. The work summarized below is focused around two topics for which neuroprotection is particularly germane: epilepsy and Alzheimer’s disease. Effects of estrogen on seizures vary, depending on the experimental approach, and many other factors. Estrogen may increase neuronal excitability and thus mediate proconvulsant effects that have been reported in the past, but reviews of the clinical and animal data show that estrogen may also have no effect or even anticonvulsant effects (Scharfman et al., 2003; Hajszan and MacLusky, 2006; Veliskova, 2006). The protective role of estrogen in seizure-induced damage is more straightforward. Severe seizures have been shown

to trigger excitotoxic cell death in the hilus of the dentate gyrus, and patients with temporal lobe epilepsy often exhibit neuronal loss in the hilus also (Margerison and Corsellis, 1966). Kainic acid is commonly used as a convulsant to elicit severe seizures (status epilepticus) and excitotoxic damage in the dentate gyrus of rats (Ben-Ari and Cossart, 2000). Estrogen administration is capable of preventing kainic acid-induced degeneration (Azcoitia et al., 1998; Veliskova et al., 2000). Estrogen may also mediate the protective actions of other steroids. For example, the neurosteroids pregnenolone and DHEA showed a dosedependent protective effect of hilar neurons against kainic acid. The administration of the aromatase inhibitor fadrozole, that blocks the conversion of these steroids into estrogen, prevented this effect (Veiga et al., 2003). Interestingly, 2-methoxyestradiol, an estradiol metabolite, induced significant neuronal loss in the hilus, detected 96 h after the treatment with this steroid. This finding suggests that endogenous metabolism of 17b-estradiol to 2-methoxyestradiol may counterbalance the neuroprotective effects of estrogen (Picazo et al., 2003). Regarding mechanisms for the neuroprotective effects of estrogen in studies of seizure-induced neuronal damage, Veliskova et al. (2000) as well as Haynes et al. (2003) suggest that intracellular ERs mediate the neuroprotective effect of estrogen, because tamoxifen pretreatment effectively abolished estrogen-induced neuroprotection. An interaction of ER and IGF-1 receptor signaling may also be important (Azcoitia et al., 1999a). Furthermore, GABAB receptors are likely to play a role, because there was a loss of GABAB receptor-mediated inhibition after kainic acid-induced status epilepticus in the rat dentate gyrus, and pretreatment with estrogen could prevent it (Velisek and Veliskova, 2002). Other steroids besides estrogen are also likely to reduce seizure-induced damage, and the progesterone metabolite 3a,5a-tetrahyroprogesterone (allopregnanolone) has been shown to be one example. Blocking progesterone’s metabolism to 3a,5a-tetrahyroprogesterone reduced progesterone’s protective effects in the dentate gyrus (Rhodes et al., 2004). In the kainate model,

406

3a,5a-tetrahyroprogesterone was able to protect the hilus from kainic acid (Ciriza et al., 2004). Another metabolite was also effective: 5a-hydroxyprogesterone (Ciriza et al., 2004). Other models of injury, which use adrenalectomy to examine neuronal loss in the dentate gyrus, focus on the granule cells, because adrenalectomy selectively kills granule cells (see chapter by M. Joels in this volume). In this model, estradiol treatment reduced pyknotic cell number compared to vehicle administration (Frye, 2001). Interestingly, a synthetic glucocorticoid, dexamethasone can also induce apoptosis in the dentate gyrus, and pretreatment with estrogen substantially attenuated the dexamethasone-induced neuronal damage (Haynes et al., 2003). Colchicine, a microtubule polymerization inhibitor, also selectively kills granule cells, an effect that is increased by ovariectomy and ameliorated by 17b-estradiol (Liu et al., 2001). Regarding other sex steroids, treatment of female or male rats with progesterone or its metabolites, 5a-dihydroprogesterone and 3a,5atetrahyroprogesterone similarly reduced the total number of adrenalectomy-induced pyknotic cells in the dentate gyrus compared with vehicle administration. In case of androgen, testosterone and its metabolites, 5a-dihydrotestosterone and 5aandrostane-3a,17b-diol significantly reduced the number of pyknotic cells in the dentate gyrus compared to vehicle-administered, adrenalectomized female rats (Frye and McCormick, 2000a, b). Estrogen also has been found to play a critical neuroprotective role in Azlheimer’s disease. Unilateral entorhinal cortex lesion (ECX) is frequently used as a model of Alzheimer’s disease-like deafferentation in the dentate gyrus. ECX elicits sprouting in the molecular layer, which is affected by gonadectomy and hormone replacement, but only in female rats: ovariectomy reduces fiber outgrowth and estrogen restores it (Stone et al., 2000). However, testosterone replacement had no effect on sprouting in castrated ECX males (Morse et al., 1986). Sprouting in hippocampal cultures of C57Bl/6J mice was increased by 75% after treatment with 17b-estradiol, which was blocked by an antagonist of nuclear receptors, tamoxifen (Teter et al., 1999). In intact female mice in vivo, lesions

of the lateral part of the entorhinal cortex increased axonal sprouting in the outer one-third of the molecular layer of the dentate gyrus. Ovariectomized mice receiving high and moderate estrogen supplementation displayed the same sprouting response. In ovariectomized nontreated mice, however, the sprouting response was significantly reduced (to nearly nothing) (Kadish and van Groen, 2002). Finally, Stone et al. (1998a) have shown that in wild-type ECX mice, ovariectomy decreases commissural/associational sprouting to the inner molecular layer of the dentate gyrus, which is reversed by estradiol replacement. In ECX apolipoprotein E-knockout mice, however, estradiol did not enhance sprouting, suggesting that sprouting may be stimulated by estrogen through its up-regulation of apolipoprotein E expression, leading to increased recycling of membrane lipids for use by sprouting neurons (Stone et al., 1998a). Estrogen and apolipoprotein E may therefore interact in their modulation of both Alzheimer’s disease risk and recovery from neuronal injury. Mediation of sex steroid effects Effects of sex steroids in the dentate gyrus depend on the location of their action, and the sites of steroid synthesis. Below we discuss the distribution of receptors, which is summarized in Fig. 2. Sex steroids are capable of acting directly on granule cells, as well as indirectly via nongranule cells within the dentate gyrus (interneurons, mossy cells). The influence of sex steroids could also be mediated by subcortical, hormone-sensitive structures, such as the septohippocampal cholinergic neurons. Sites of sex steroid synthesis are usually thought to be peripheral, but local synthesis is also possible, and is discussed further below. Distribution of ERa in the dentate gyrus Estrogen binding, as well as mRNA and immunoreactivity for ERa, have been detected in nuclei of scattered GABAergic interneurons, located predominantly in the subgranular region of the dentate gyrus (Loy et al., 1988; Shughrue et al., 1997;

407

Fig. 2. Subcellular localization of estrogen (ER), androgen (AR), and progestin (PR) receptors in the dentate gyrus. A subset of GABAergic interneurons contains nuclear ERa (dark pink). Granule cells, newly born cells (identified by DCX) and some GABAergic interneurons contain cytosolic and plasma membrane-associated ERb (blue). Dendritic spines, many originating from granule cells contain ERa, ERb, AR (dark green), and PR (purple). A few dendritic spines in the hilus, likely originating from mossy cells, contain ERa and ERb. ERa, ERb, AR, and PR are found in axons and axon terminals. Some ERa-containing terminals are cholinergic (acetylcholine, orange); some ERb-containing terminals resemble monoaminergic boutons. Lot of astrocytes (stars), mostly in the molecular layer, also contain ERa, ERb, AR, and PR. (See Color Plate 23.2 in color plate section.)

Weiland et al., 1997; Perlman et al., 2005). Nuclear ERa-immunoreactive interneurons co-express neuropeptide Y, calbindin-D28k and calretinin, but not cholecystokinin or parvalbumin (Nakamura and McEwen, 2005). In addition to nuclear receptors, ultrastructural studies have revealed ERa

at several extranuclear sites in the dentate gyrus (Milner et al., 2001). Specifically, ERa-immunoreactivity is affiliated with the cytoplasmic plasmalemma of select hilar interneurons and with endosomes of a few granule cell perikarya. Moreover, ERa-labeled profiles are dispersed

408

throughout the dentate gyrus. Approximately half of these labeled profiles are unmyelinated axons and axon terminals that contain numerous small, synaptic vesicles. ERa-labeled terminals form both symmetric and asymmetric synapses on dendritic shafts and spines, suggesting that ERa-positive axons arise from sources in addition to inhibitory interneurons. Dual labeling revealed that ERaimmunoreactivity is contained in axons and terminals labeled with vesicular acetylcholine transporter (Towart et al., 2003), suggesting that estrogen could rapidly and directly affect the local release and/or uptake of acetylcholine. About one-quarter of the ERa-immunoreactive profiles are dendritic spines, many originating from granule cells. In dendritic spines, ERa-immunoreactivity is often associated with the spine apparatus, suggesting that estrogen might act locally through ERa to influence protein synthesis during synaptic remodeling. The remaining one-quarter of ERalabeled profiles are from glial origin that resemble astrocytes and are often located near the spines of granule cells. Vesicular acetylcholine transportercontaining terminals often abut ERa-positive presynaptic and glial profiles and unlabeled terminals that contact ERa-immunoreactive spines (Towart et al., 2003), suggesting that acetylcholine release might play a critical role in estrogen-modulated structural plasticity. Collectively, these results imply that ERa may serve as both a genomic and nongenomic transducer of estrogen action in the dentate gyrus.

Distribution of ERb in the dentate gyrus The cellular and subcellular locations of ERbimmunoreactivity in the dentate gyrus are similar yet distinct from ERa. In monkey, dense ERb hybridization signal has been seen in the dentate gyrus, CA1, CA2, CA3, CA4, and the prosubiculum/subiculum areas of the hippocampus (Gundlah et al., 2000). In rodents, cells in or near the dentate granule cell layer transiently express high levels of estrogen binding and ERa protein in the nucleus during the first two postnatal weeks (O’Keefe et al., 1995; Solum and Handa, 2001). In adult rats and mice, ERb mRNA and protein has

been found in the perikarya of granule cells as well as cells in the dentate subgranular layer (Li et al., 1997; Shughrue et al., 1997; Mitra et al., 2003; Milner et al., 2005). Szymczak et al. (2006) have found that ERb mRNA and protein are displayed in high levels in the estrus and in low levels in the proestrus phase. Recently, robust mRNA expression for both the a and b subtypes of ERs has been found in proliferating and differentiating cells of neuronal phenotype in the subgranular zone of the dentate gyrus (Isgor and Watson, 2005). Furthermore, ERb-immunoreactive glia has been observed in the hilus of the dentate gyrus of male and female rats. ERb-immunoreactivity has been localized in glial processes and perikarya and, in some cases, in glial cell nuclei. Double immunocytochemical labeling of ERb and the specific astroglial marker, GFAP revealed that the ERb-immunoreactive glial cells are astrocytes (Azcoitia et al., 1999b). Ultrastructural analysis showed ERb-immunoreactivity at several extranuclear sites in the dentate gyrus (Milner et al., 2005). ERb-immunoreactivity is affiliated with cytoplasmic organelles, especially endomembranes and mitochondria, and with the membranes primarily of granule cell perikarya and proximal dendrites. Recent studies revealed that neuronal perikarya and dendrites labeled with doublecortin, a marker of newly generated cells, also contain extranuclear ERb-immunoreactivity in both the adult and neonatal dentate gyrus (Herrick et al., 2006). ERblabeled dendritic shafts and spines have mostly been found in the molecular layer. In dendritic processes, ERb-immunoreactivity is near the perisynaptic zone adjacent to synapses formed by unlabeled terminals. The ERb protein can also be found in preterminal axons and axon terminals, associated with clusters of small, synaptic vesicles. ERb-labeled axons are particularly dense in the hilus and outer molecular layer, forming both asymmetric and symmetric synapses with dendrites. Finally, ERb-immunoreactivity has been detected in glial profiles throughout the dentate gyrus, some of which appose doublecortin-labeled perikarya and dendrites (Herrick et al., 2006). These results suggest that ERb may serve primarily as a nongenomic transducer of estrogen actions in the dentate gyrus.

409

Distribution of progestin receptor (PR) in the dentate gyrus Cells containing PR mRNA have been detected in the dentate subgranular zone (Hagihara et al., 1992). By light microscopy, nuclear PR-immunoreactivity is undetectable in the dentate gyrus; however, ultrastructural analysis revealed that the PR protein is found at several extranuclear sites (Waters et al., 2005). In the molecular layer and hilus, PR-immunoreactivity is present in dendritic spines, closely associated with the postsynaptic density. The PR protein is expressed in axons and axon terminals that contain small synaptic vesicles. PR-positive terminals and en passant axonal boutons form synapses with dendritic spines. PRimmunoreactivity has also been found in glia, many resembling astrocytes and some forming presumed gap junctions with other astrocytic profiles. The considerable lack of nuclear PR labeling may indicate that progesterone uses nongenomic signaling mechanism in the dentate gyrus to directly affect dendritic spine morphology and synaptic plasticity.

Distribution of AR in the dentate gyrus Previous light microscopic studies have shown that AR mRNA, immunoreactivity, and binding are present in pyramidal cell nuclei but not granule cells (Commins and Yahr, 1985; Sar et al., 1990; Simerly et al., 1990; Kerr et al., 1995). However, AR-immunoreactivity is present in disperse, punctuate processes that are most dense in the pyramidal cell layer and diffusely distributed in the mossy fiber pathway (Tabori et al., 2005). Electron microscopic analysis revealed AR-immunoreactivity at several extranuclear sites in the dentate gyrus. AR labeling has been found in dendritic spines, many arising from granule cell dendrites. AR is affiliated with clusters of small, synaptic vesicles within preterminal axons and axon terminals, the majority of these being in the central hilus. AR-immunoreactive preterminal axons are most prominent in the CA3 stratum lucidum. ARlabeled terminals exclusively form asymmetric synapses. Throughout the dentate gyrus,

AR-immunoreactivity has also been detected in astrocytic profiles; many of them apposing terminals that synapse on unlabeled dendritic spines or forming gap junctions with other AR-positive or unlabeled astrocytes (Tabori et al., 2005). Together, these results suggest that ARs may serve as both a genomic and nongenomic transducer of androgen action in the dentate gyrus.

Role of local steroid synthesis Recent studies (Rune et al., 2006) suggest that locally synthesized steroids may contribute to hippocampal activational effects. Using slice cultures, Rune and colleagues have reported that the number of dentate proliferative cells decreases, whereas the number of apoptotic cells increases dose-dependently, in response to reduced estradiol release into the medium after treatment with letrozole, an aromatase enzyme inhibitor (Fester et al., 2006). This also holds true for cell cultures transfected with siRNA against steroidogenic acute regulatory protein (StAR). StAR transports cholesterol to the inner mitochondrial membrane, where it is converted by the cytochrome P-450 enzyme complex, and as such, it is the first step in the cascade of estrogen synthesis. Application of estradiol to the medium had no effect on proliferation and apoptosis, whereas the antiproliferative and pro-apoptotic effects of StAR knockdown and letrozole administration were restored by treatment of the cultures with estradiol (Fester et al., 2006). The data of Rune and colleagues are also supported by other studies. In situ hybridization revealed that StAR and aromatase are highly expressed in neuronal cells of the dentate gyrus. In addition, StAR- and aromatase-positive cells are strictly correlated with steroidogenic factor-1, a regulator of steroid biosynthesis, as shown by computer-assisted confocal microscopy in double labeling experiments (Wehrenberg et al., 2001). Similarly, Hojo et al. (2004) have reported in a rather comprehensive study that in the granule cells of the adult male rat dentate gyrus, significant colocalization is seen for both cytochromes P45017a (dehydroepiandrosterone synthase) and P450 aromatase by means of

410

immunohistochemical staining of slices. Only a weak immunoreaction of these P450s has been observed in astrocytes and oligodendrocytes (Hojo et al., 2004). More importantly, stimulation of hippocampal neurons with NMDA induced a significant net production of estradiol. The analysis of radioactive metabolites demonstrated the conversion from [3H] pregnenolone to [3H] estradiol through dehydroepiandrosterone and testosterone. This activity was abolished by the application of specific inhibitors of cytochrome P450s (Hojo et al., 2004). In summary, although these findings seem compelling, considering the fact that the majority of the above-mentioned data have been obtained from cultures, which may be quite different from conditions in vivo, one should exercise caution in interpreting such results.

Subcortical mediation of sex steroid effects In addition to direct effects of sex steroids on dentate cells, indirect actions may influence the dentate gyrus. We already mentioned above that the raphe serotonergic system mediates the dentate neurogenic effect of estrogen (Banasr et al., 2001). Consistent with this line of argument, raphe serotonergic neurons express ERa (Leranth et al., 1999). Another structure that appears to be critical is the septohippocampal cholinergic system. Septal cholinergic neurons express nuclear ERs (Shughrue et al., 2000) and hippocampal cholinergic axons and terminals contain extranuclear ERa (Towart et al., 2003). Moreover, estrogen affects septohippocampal cholinergic neurons both genomically and nongenomically (Gibbs and Aggarwal, 1998; Rudick et al., 2003). Androgen also can influence the expression of cholinergic markers in the dentate gyrus. Specifically, gonadectomy reduced the density of choline acetyltransferase immunoreactive fibers in the dentate gyrus, which was reversed by the addition of testosterone propionate (Nakamura et al., 2002). Although there is no direct data from the dentate gyrus, elsewhere in the hippocampus estrogen modulates the inhibition by specific GABAergic interneurons, which is

partially dependent on input from basal forebrain cholinergic neurons (Rudick et al., 2003). More evidence is available for the cholinergic role in dentate neurogenesis. Selective neurotoxic lesion of the forebrain cholinergic input with 192 IgG-saporin reduced dentate cell proliferation (Mohapel et al., 2005). Conversely, systemic administration of the cholinergic agonist physostigmine increased dentate neurogenesis. The neurogenic effect of acetylcholine appears to involve nicotinic receptors containing the b2 subunit (Harrist et al., 2004), as well as m2, m3, and m4 muscarinic receptors (Ma et al., 2000; Mohapel et al., 2005). Consistent with these findings, ovariectomy upregulated m4 receptors in the dentate gyrus, whereas estrogen treatment restored m4 binding to the level of the sham group (El-Bakri et al., 2002). However, other results suggest no septohippocampal involvement in the synaptogenic action of estrogen, because rats that received estrogen implants into the medial septum did not exhibit changes in astroglial process density in the dentate gyrus (Lam and Leranth, 2003).

Concluding remarks The studies discussed in this review clearly demonstrate that both the organization and functioning of the dentate gyrus is a significant target of sex steroids. The gonadal hormone modulation of physiological activity, neurogenesis, synaptic remodeling, and neurodegeneration/neuroprotection mechanisms are likely to be relevant to higher brain functions such as cognition and mood, in addition to their clinical implications in epilepsy, Alzheimer’s disease and mental disorders. Relative to hippocampal subfield CA1, however, the role of gonadal hormones in the dentate gyrus has not received substantial attention. In particular, our understanding of androgen effects on the dentate gyrus is extremely limited relative to estrogen. The available data suggest potent, complex, and potentially important effects, however, and therefore merit more attention in the future.

411

Abbreviations AR BDNF DHEAS ECX ER GFAP IGF-1 NMDA PR StAR Tfm

androgen receptor brain-derived neurotrophic factor dehydroepiandrosterone-sulfate entorhinal cortex lesion estrogen receptor glial fibrillary acidic protein insulin-like growth factor-1 N-methyl-D-aspartate progestin receptor steroidogenic acute regulatory protein testicular feminization mutant

Acknowledgements This work was supported by NIH grants MH074021 (T.H.); DA08259, NS07080 and HL18974 (T.A.M.); MH060858 and NS042644 (C.L.); as well as by a Hungarian National Office for Research and Technology grant RET-08/04.

References Altman, J. and Das, G.D. (1965) Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol., 124(3): 319–335. Azcoitia, I., Sierra, A. and Garcia-Segura, L.M. (1998) Estradiol prevents kainic acid-induced neuronal loss in the rat dentate gyrus. Neuroreport, 9(13): 3075–3079. Azcoitia, I., Sierra, A. and Garcia-Segura, L.M. (1999a) Neuroprotective effects of estradiol in the adult rat hippocampus: interaction with insulin-like growth factor-I signalling. J. Neurosci. Res., 58(6): 815–822. Azcoitia, I., Sierra, A. and Garcia-Segura, L.M. (1999b) Localization of estrogen receptor beta-immunoreactivity in astrocytes of the adult rat brain. Glia, 26(3): 260–267. Banasr, M., Hery, M., Brezun, J.M. and Daszuta, A. (2001) Serotonin mediates oestrogen stimulation of cell proliferation in the adult dentate gyrus. Eur. J. Neurosci., 14(9): 1417–1424. Ben-Ari, Y. and Cossart, R. (2000) Kainate, a double agent that generates seizures: two decades of progress. Trends Neurosci., 23(11): 580–587. Birzniece, V., Johansson, I.M., Wang, M.D., Seckl, J.R., Backstrom, T. and Olsson, T. (2001) Serotonin 5-HT(1A) receptor mRNA expression in dorsal hippocampus and raphe

nuclei after gonadal hormone manipulation in female rats. Neuroendocrinology, 74(2): 135–142. Bowman, R.E., Ferguson, D. and Luine, V.N. (2002) Effects of chronic restraint stress and estradiol on open field activity, spatial memory, and monoaminergic neurotransmitters in ovariectomized rats. Neuroscience, 113(2): 401–410. Brake, W.G., Alves, S.E., Dunlop, J.C., Lee, S.J., Bulloch, K., Allen, P.B., Greengard, P. and McEwen, B.S. (2001) Novel target sites for estrogen action in the dorsal hippocampus: an examination of synaptic proteins. Endocrinology, 142(3): 1284–1289. Brannvall, K., Bogdanovic, N., Korhonen, L. and Lindholm, D. (2005) 19-Nortestosterone influences neural stem cell proliferation and neurogenesis in the rat brain. Eur. J. Neurosci., 21(4): 871–878. Choi, J.M., Romeo, R.D., Brake, W.G., Bethea, C.L., Rosenwaks, Z. and McEwen, B.S. (2003) Estradiol increases pre- and post-synaptic proteins in the CA1 region of the hippocampus in female rhesus macaques (Macaca mulatta). Endocrinology, 144(11): 4734–4738. Ciriza, I., Azcoitia, I. and Garcia-Segura, L.M. (2004) Reduced progesterone metabolites protect rat hippocampal neurones from kainic acid excitotoxicity in vivo. J. Neuroendocrinol., 16(1): 58–63. Commins, D. and Yahr, P. (1985) Autoradiographic localization of estrogen and androgen receptors in the sexually dimorphic area and other regions of the gerbil brain. J. Comp. Neurol., 231(4): 473–489. Conner, J.M., Lauterborn, J.C., Yan, Q., Gall, C.M. and Varon, S. (1997) Distribution of brain-derived neurotrophic factor (BDNF) protein and mRNA in the normal adult rat CNS: evidence for anterograde axonal transport. J. Neurosci., 17(7): 2295–2313. Cyr, M., Ghribi, O. and Di Paolo, T. (2000) Regional and selective effects of oestradiol and progesterone on NMDA and AMPA receptors in the rat brain. J. Neuroendocrinol., 12(5): 445–452. Cyr, M., Thibault, C., Morissette, M., Landry, M. and Di Paolo, T. (2001) Estrogen-like activity of tamoxifen and raloxifene on NMDA receptor binding and expression of its subunits in rat brain. Neuropsychopharmacology, 25(2): 242–257. De Nicola, A.F., Saravia, F.E., Beauquis, J., Pietranera, L. and Ferrini, M.G. (2006) Estrogens and neuroendocrine hypothalamic-pituitary-adrenal axis function. Front. Horm. Res., 35: 157–168. El-Bakri, N.K., Adem, A., Suliman, I.A., Mulugeta, E., Karlsson, E., Lindgren, J.U., Winblad, B. and Islam, A. (2002) Estrogen and progesterone treatment: effects on muscarinic M(4) receptor subtype in the rat brain. Brain Res., 948(1–2): 131–137. El-Bakri, N.K., Islam, A., Zhu, S., Elhassan, A., Mohammed, A., Winblad, B. and Adem, A. (2004) Effects of estrogen and progesterone treatment on rat hippocampal NMDA receptors: relationship to Morris water maze performance. J. Cell. Mol. Med., 8(4): 537–544.

412 Falconer, E.M. and Galea, L.A. (2003) Sex differences in cell proliferation, cell death and defensive behavior following acute predator odor stress in adult rats. Brain Res., 975(1–2): 22–36. Fester, L., Ribeiro-Gouveia, V., Prange-Kiel, J., von Schassen, C., Bottner, M., Jarry, H. and Rune, G.M. (2006) Proliferation and apoptosis of hippocampal granule cells require local oestrogen synthesis. J. Neurochem., 97(4): 1136–1144. Franklin, T.B. and Perrot-Sinal, T.S. (2006) Sex and ovarian steroids modulate brain-derived neurotrophic factor (BDNF) protein levels in rat hippocampus under stressful and nonstressful conditions. Psychoneuroendocrinology, 31(1): 38–48. Frye, C.A. (2001) Estradiol tends to improve inhibitory avoidance performance in adrenalectomized male rats and reduces pyknotic cells in the dentate gyrus of adrenalectomized male and female rats. Brain Res., 889(1–2): 358–363. Frye, C.A. and McCormick, C.M. (2000a) Androgens are neuroprotective in the dentate gyrus of adrenalectomized female rats. Stress, 3(3): 185–194. Frye, C.A. and McCormick, C.M. (2000b) The neurosteroid, 3alpha-androstanediol, prevents inhibitory avoidance deficits and pyknotic cells in the granule layer of the dentate gyrus induced by adrenalectomy in rats. Brain Res., 855(1): 166–170. Galea, L.A., Spritzer, M.D., Barker, J.M. and Pawluski, J.L. (2006) Gonadal hormone modulation of hippocampal neurogenesis in the adult. Hippocampus, 16(3): 225–232. Gazzaley, A.H., Weiland, N.G., McEwen, B.S. and Morrison, J.H. (1996) Differential regulation of NMDAR1 mRNA and protein by estradiol in the rat hippocampus. J. Neurosci., 16(21): 6830–6838. Gibbs, R.B. and Aggarwal, P. (1998) Estrogen and basal forebrain cholinergic neurons: implications for brain aging and Alzheimer’s disease-related cognitive decline. Horm. Behav., 34(2): 98–111. Gould, E., Tanapat, P., Rydel, T. and Hastings, N. (2000) Regulation of hippocampal neurogenesis in adulthood. Biol. Psychiatry, 48(8): 715–720. Gould, E., Woolley, C.S., Frankfurt, M. and McEwen, B.S. (1990) Gonadal steroids regulate dendritic spine density in hippocampal pyramidal cells in adulthood. J. Neurosci., 10(4): 1286–1291. Gundlah, C., Kohama, S.G., Mirkes, S.J., Garyfallou, V.T., Urbanski, H.F. and Bethea, C.L. (2000) Distribution of estrogen receptor beta (ERbeta) mRNA in hypothalamus, midbrain and temporal lobe of spayed macaque: continued expression with hormone replacement. Brain Res. Mol. Brain Res., 76(2): 191–204. Hagihara, K., Hirata, S., Osada, T., Hirai, M. and Kato, J. (1992) Distribution of cells containing progesterone receptor mRNA in the female rat di- and telencephalon: an in situ hybridization study. Brain Res. Mol. Brain Res., 14(3): 239–249. Hajszan, T. and MacLusky, N.J. (2006) Neurologic links between epilepsy and depression in women: is hippocampal neuroplasticity the key? Neurology, 66(6 Suppl 3): S13–S22.

Hajszan, T., MacLusky, N.J. and Leranth, C. (2005) Shortterm treatment with the antidepressant fluoxetine triggers pyramidal dendritic spine synapse formation in rat hippocampus. Eur. J. Neurosci., 21(5): 1299–1303. Hao, J., Janssen, W.G., Tang, Y., Roberts, J.A., McKay, H., Lasley, B., Allen, P.B., Greengard, P., Rapp, P.R., Kordower, J.H., Hof, P.R. and Morrison, J.H. (2003) Estrogen increases the number of spinophilin-immunoreactive spines in the hippocampus of young and aged female rhesus monkeys. J. Comp. Neurol., 465(4): 540–550. Harrist, A., Beech, R.D., King, S.L., Zanardi, A., Cleary, M.A., Caldarone, B.J., Eisch, A., Zoli, M. and Picciotto, M.R. (2004) Alteration of hippocampal cell proliferation in mice lacking the beta 2 subunit of the neuronal nicotinic acetylcholine receptor. Synapse, 54(4): 200–206. Haynes, L.E., Lendon, C.L., Barber, D.J. and Mitchell, I.J. (2003) 17 Beta-oestradiol attenuates dexamethasone-induced lethal and sublethal neuronal damage in the striatum and hippocampus. Neuroscience, 120(3): 799–806. Herrick, S.P., Waters, E.M., Drake, C.T., McEwen, B.S. and Milner, T.A. (2006) Extranuclear estrogen receptor beta immunoreactivity is on doublecortin-containing cells in the adult and neonatal rat dentate gyrus. Brain Res., 1121: 46–58. Hojo, Y., Hattori, T.A., Enami, T., Furukawa, A., Suzuki, K., Ishii, H.T., Mukai, H., Morrison, J.H., Janssen, W.G., Kominami, S., Harada, N., Kimoto, T. and Kawato, S. (2004) Adult male rat hippocampus synthesizes estradiol from pregnenolone by cytochromes P45017alpha and P450 aromatase localized in neurons. Proc. Natl. Acad. Sci. U.S.A., 101(3): 865–870. Isgor, C. and Sengelaub, D.R. (1998) Prenatal gonadal steroids affect adult spatial behavior, CA1 and CA3 pyramidal cell morphology in rats. Horm. Behav., 34(2): 183–198. Isgor, C. and Watson, S.J. (2005) Estrogen receptor alpha and beta mRNA expressions by proliferating and differentiating cells in the adult rat dentate gyrus and subventricular zone. Neuroscience, 134(3): 847–856. Jones, B.A. and Watson, N.V. (2005) Spatial memory performance in androgen insensitive male rats. Physiol. Behav., 85(2): 135–141. Kadish, I. and van Groen, T. (2002) Low levels of estrogen significantly diminish axonal sprouting after entorhinal cortex lesions in the mouse. J. Neurosci., 22(10): 4095–4102. Kandel, E.R. (2001) The molecular biology of memory storage: a dialogue between genes and synapses. Science, 294(5544): 1030–1038. Kasai, H., Matsuzaki, M., Noguchi, J., Yasumatsu, N. and Nakahara, H. (2003) Structure-stability-function relationships of dendritic spines. Trends Neurosci., 26(7): 360–368. O’Keefe, J.A., Li, Y., Burgess, L.H. and Handa, R.J. (1995) Estrogen receptor mRNA alterations in the developing rat hippocampus. Brain Res. Mol. Brain Res., 30(1): 115–124. Kerr, J.E., Allore, R.J., Beck, S.G. and Handa, R.J. (1995) Distribution and hormonal regulation of androgen receptor (AR) and AR messenger ribonucleic acid in the rat hippocampus. Endocrinology, 136(8): 3213–3221.

413 Kim, M.T., Soussou, W., Gholmieh, G., Ahuja, A., Tanguay, A., Berger, T.W. and Brinton, R.D. (2006) 17beta-Estradiol potentiates field excitatory postsynaptic potentials within each subfield of the hippocampus with greatest potentiation of the associational/commissural afferents of CA3. Neuroscience, 141(1): 391–406. Korol, D.L. and Kolo, L.L. (2002) Estrogen-induced changes in place and response learning in young adult female rats. Behav. Neurosci., 116(3): 411–420. Lam, T.T. and Leranth, C. (2003) Gonadal hormones act extrinsic to the hippocampus to influence the density of hippocampal astroglial processes. Neuroscience, 116(2): 491–498. Le Saux, M. and Di Paolo, T. (2005) Changes in 5-HT1A receptor binding and G-protein activation in the rat brain after estrogen treatment: comparison with tamoxifen and raloxifene. J. Psychiatry Neurosci., 30(2): 110–117. Lei, D.L., Long, J.M., Hengemihle, J., O’Neill, J., Manaye, K.F., Ingram, D.K. and Mouton, P.R. (2003) Effects of estrogen and raloxifene on neuroglia number and morphology in the hippocampus of aged female mice. Neuroscience, 121(3): 659–666. Leranth, C., Shanabrough, M. and Horvath, T.L. (1999) Estrogen receptor-alpha in the raphe serotonergic and supramammillary area calretinin-containing neurons of the female rat. Exp. Brain Res., 128(3): 417–420. Li, C., Brake, W.G., Romeo, R.D., Dunlop, J.C., Gordon, M., Buzescu, R., Magarinos, A.M., Allen, P.B., Greengard, P., Luine, V. and McEwen, B.S. (2004) Estrogen alters hippocampal dendritic spine shape and enhances synaptic protein immunoreactivity and spatial memory in female mice. Proc. Natl. Acad. Sci. U.S.A., 101(7): 2185–2190. Li, X., Schwartz, P.E. and Rissman, E.F. (1997) Distribution of estrogen receptor-beta-like immunoreactivity in rat forebrain. Neuroendocrinology, 66(2): 63–67. Liu, Z., Gastard, M., Verina, T., Bora, S., Mouton, P.R. and Koliatsos, V.E. (2001) Estrogens modulate experimentally induced apoptosis of granule cells in the adult hippocampus. J. Comp. Neurol., 441(1): 1–8. Louissaint, A.J., Rao, S., Leventhal, C. and Goldman, S.A. (2002) Coordinated interaction of neurogenesis and angiogenesis in the adult songbird brain. Neuron, 34(6): 945–960. Loy, R., Gerlach, J.L. and McEwen, B.S. (1988) Autoradiographic localization of estradiol-binding neurons in the rat hippocampal formation and entorhinal cortex. Brain Res., 467(2): 245–251. Luquin, S., Naftolin, F. and Garcia-Segura, L.M. (1993) Natural fluctuation and gonadal hormone regulation of astrocyte immunoreactivity in dentate gyrus. J. Neurobiol., 24(7): 913–924. Lynch, M.A. (2004) Long-term potentiation and memory. Physiol. Rev., 84(1): 87–136. Ma, W., Maric, D., Li, B.S., Hu, Q., Andreadis, J.D., Grant, G.M., Liu, Q.Y., Shaffer, K.M., Chang, Y.H., Zhang, L., Pancrazio, J.J., Pant, H.C., Stenger, D.A. and Barker, J.L. (2000) Acetylcholine stimulates cortical precursor cell

proliferation in vitro via muscarinic receptor activation and MAP kinase phosphorylation. Eur. J. Neurosci., 12(4): 1227–1240. MacLusky, N.J., Hajszan, T., Prange-Kiel, J. and Leranth, C. (2006) Androgen modulation of hippocampal synaptic plasticity. Neuroscience, 138(3): 957–965. MacLusky, N.J., Luine, V.N., Hajszan, T. and Leranth, C. (2005) The 17alpha and 17beta isomers of estradiol both induce rapid spine synapse formation in the CA1 hippocampal subfield of ovariectomized female rats. Endocrinology, 146(1): 287–293. Margerison, J.H. and Corsellis, J.A. (1966) Epilepsy and the temporal lobes. A clinical, electroencephalographic and neuropathological study of the brain in epilepsy, with particular reference to the temporal lobes. Brain, 89(3): 499–530. Mazzucco, C.A., Lieblich, S.E., Bingham, B.I., Williamson, M.A., Viau, V. and Galea, L.A. (2006) Both estrogen receptor alpha and estrogen receptor beta agonists enhance cell proliferation in the dentate gyrus of adult female rats. Neuroscience, 141: 1793–1800. McEwen, B.S. (2003) Mood disorders and allostatic load. Biol. Psychiatry, 54(3): 200–207. McEwen, B.S. and Alves, S.E. (1999) Estrogen actions in the central nervous system. Endocr. Rev., 20(3): 279–307. Milner, T.A., Ayoola, K., Drake, C.T., Herrick, S.P., Tabori, N.E., McEwen, B.S., Warrier, S. and Alves, S.E. (2005) Ultrastructural localization of estrogen receptor beta immunoreactivity in the rat hippocampal formation. J. Comp. Neurol., 491(2): 81–95. Milner, T.A., McEwen, B.S., Hayashi, S., Li, C.J., Reagan, L.P. and Alves, S.E. (2001) Ultrastructural evidence that hippocampal alpha estrogen receptors are located at extranuclear sites. J. Comp. Neurol., 429(3): 355–371. Miranda, P., Williams, C.L. and Einstein, G. (1999) Granule cells in aging rats are sexually dimorphic in their response to estradiol. J. Neurosci., 19(9): 3316–3325. Mitra, S.W., Hoskin, E., Yudkovitz, J., Pear, L., Wilkinson, H.A., Hayashi, S., Pfaff, D.W., Ogawa, S., Rohrer, S.P., Schaeffer, J.M., McEwen, B.S. and Alves, S.E. (2003) Immunolocalization of estrogen receptor beta in the mouse brain: comparison with estrogen receptor alpha. Endocrinology, 144(5): 2055–2067. Mohapel, P., Leanza, G., Kokaia, M. and Lindvall, O. (2005) Forebrain acetylcholine regulates adult hippocampal neurogenesis and learning. Neurobiol. Aging, 26(6): 939–946. Morse, J.K., Scheff, S.W. and DeKosky, S.T. (1986) Gonadal steroids influence axon sprouting in the hippocampal dentate gyrus: a sexually dimorphic response. Exp. Neurol., 94(3): 649–658. Naftolin, F., Ryan, K.J., Davies, I.J., Reddy, V.V., Flores, F., Petro, Z., Kuhn, M., White, R.J., Takaoka, Y. and Wolin, L. (1975) The formation of estrogens by central neuroendocrine tissues. Recent Prog. Horm. Res., 31: 295–319. Nakamura, N., Fujita, H. and Kawata, M. (2002) Effects of gonadectomy on immunoreactivity for choline acetyltransferase in the cortex, hippocampus, and basal forebrain of adult male rats. Neuroscience, 109(3): 473–485.

414 Nakamura, N.H. and McEwen, B.S. (2005) Changes in interneuronal phenotypes regulated by estradiol in the adult rat hippocampus: a potential role for neuropeptide Y. Neuroscience, 136(1): 357–369. Orchinik, M., Weiland, N.G. and McEwen, B.S. (1995) Chronic exposure to stress levels of corticosterone alters GABAA receptor subunit mRNA levels in rat hippocampus. Brain Res. Mol. Brain Res., 34(1): 29–37. Ormerod, B.K., Lee, T.T. and Galea, L.A. (2003) Estradiol initially enhances but subsequently suppresses (via adrenal steroids) granule cell proliferation in the dentate gyrus of adult female rats. J. Neurobiol., 55(2): 247–260. Parducz, A. and Garcia-Segura, L.M. (1993) Sexual differences in the synaptic connectivity in the rat dentate gyrus. Neurosci. Lett., 161(1): 53–56. Parducz, A., Hajszan, T., Maclusky, N.J., Hoyk, Z., Csakvari, E., Kurunczi, A., Prange-Kiel, J. and Leranth, C. (2006) Synaptic remodeling induced by gonadal hormones: neuronal plasticity as a mediator of neuroendocrine and behavioral responses to steroids. Neuroscience, 138(3): 977–985. Perez-Martin, M., Azcoitia, I., Trejo, J.L., Sierra, A. and Garcia-Segura, L.M. (2003) An antagonist of estrogen receptors blocks the induction of adult neurogenesis by insulin-like growth factor-I in the dentate gyrus of adult female rat. Eur. J. Neurosci., 18(4): 923–930. Perez-Martin, M., Salazar, V., Castillo, C., Ariznavarreta, C., Azcoitia, I., Garcia-Segura, L.M. and Tresguerres, J.A. (2005) Estradiol and soy extract increase the production of new cells in the dentate gyrus of old rats. Exp. Gerontol., 40(5): 450–453. Perlman, W.R., Tomaskovic-Crook, E., Montague, D.M., Webster, M.J., Rubinow, D.R., Kleinman, J.E. and Weickert, C.S. (2005) Alteration in estrogen receptor alpha mRNA levels in frontal cortex and hippocampus of patients with major mental illness. Biol. Psychiatry, 58(10): 812–824. Picazo, O., Azcoitia, I. and Garcia-Segura, L.M. (2003) Neuroprotective and neurotoxic effects of estrogens. Brain Res., 990(1–2): 20–27. Rhodes, M.E., McCormick, C.M. and Frye, C.A. (2004) 3alpha,5alpha-THP mediates progestins’ effects to protect against adrenalectomy-induced cell death in the dentate gyrus of female and male rats. Pharmacol. Biochem. Behav., 78(3): 505–512. Roof, R.L. (1993) The dentate gyrus is sexually dimorphic in prepubescent rats: testosterone plays a significant role. Brain Res., 610(1): 148–151. Roof, R.L. and Havens, M.D. (1992) Testosterone improves maze performance and induces development of a male hippocampus in females. Brain Res., 572(1–2): 310–313. Rudick, C.N., Gibbs, R.B. and Woolley, C.S. (2003) A role for the basal forebrain cholinergic system in estrogen-induced disinhibition of hippocampal pyramidal cells. J. Neurosci., 23(11): 4479–4490. Rune, G.M., Lohse, C., Prange-Kiel, J., Fester, L. and Frotscher, M. (2006) Synaptic plasticity in the hippocampus: effects of estrogen from the gonads or hippocampus? Neurochem. Res., 31(2): 145–155.

Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F., Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O., Belzung, C. and Hen, R. (2003) Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science, 301(5634): 805–809. Sar, M., Lubahn, D.B., French, F.S. and Wilson, E.M. (1990) Immunohistochemical localization of the androgen receptor in rat and human tissues. Endocrinology, 127(6): 3180–3186. Saravia, F.E., Beauquis, J., Revsin, Y., Homo-Delarche, F., de Kloet, E.R. and De Nicola, A.F. (2006) Hippocampal neuropathology of diabetes mellitus is relieved by estrogen treatment. Cell. Mol. Neurobiol., 26: 943–957. Scharfman, H.E., Mercurio, T.C., Goodman, J.H., Wilson, M.A. and MacLusky, N.J. (2003) Hippocampal excitability increases during the estrous cycle in the rat: a potential role for brain-derived neurotrophic factor. J. Neurosci., 23(37): 11641–11652. Schumacher, M., Coirini, H. and McEwen, B.S. (1989) Regulation of high-affinity GABAA receptors in the dorsal hippocampus by estradiol and progesterone. Brain Res., 487(1): 178–183. Seidman, S.N. (2003) The aging male: androgens, erectile dysfunction, and depression. J. Clin. Psychiatry, 64(Suppl 10): 31–37. Shughrue, P.J., Lane, M.V. and Merchenthaler, I. (1997) Comparative distribution of estrogen receptor-alpha and -beta mRNA in the rat central nervous system. J. Comp. Neurol., 388(4): 507–525. Shughrue, P.J., Scrimo, P.J. and Merchenthaler, I. (2000) Estrogen binding and estrogen receptor characterization (ERalpha and ERbeta) in the cholinergic neurons of the rat basal forebrain. Neuroscience, 96(1): 41–49. Simerly, R.B., Chang, C., Muramatsu, M. and Swanson, L.W. (1990) Distribution of androgen and estrogen receptor mRNA-containing cells in the rat brain: an in situ hybridization study. J. Comp. Neurol., 294(1): 76–95. Sohrabji, F., Miranda, R.C. and Toran-Allerand, C.D. (1995) Identification of a putative estrogen response element in the gene encoding brain-derived neurotrophic factor. Proc. Natl. Acad. Sci. U.S.A., 92(24): 11110–11114. Solum, D.T. and Handa, R.J. (2001) Localization of estrogen receptor alpha (ER alpha) in pyramidal neurons of the developing rat hippocampus. Brain Res. Dev. Brain Res., 128(2): 165–175. Sorra, K.E. and Harris, K.M. (2000) Overview on the structure, composition, function, development, and plasticity of hippocampal dendritic spines. Hippocampus, 10(5): 501–511. Steffensen, S.C. (1995) Dehydroepiandrosterone sulfate suppresses hippocampal recurrent inhibition and synchronizes neuronal activity to theta rhythm. Hippocampus, 5(4): 320–328. Steiner, M., Dunn, E. and Born, L. (2003) Hormones and mood: from menarche to menopause and beyond. J. Affect. Disord., 74(1): 67–83. Stone, D.J., Rozovsky, I., Morgan, T.E., Anderson, C.P. and Finch, C.E. (1998a) Increased synaptic sprouting in response to estrogen via an apolipoprotein E-dependent mechanism: implications for Alzheimer’s disease. J. Neurosci., 18(9): 3180–3185.

415 Stone, D.J., Rozovsky, I., Morgan, T.E., Anderson, C.P., Lopez, L.M., Shick, J. and Finch, C.E. (2000) Effects of age on gene expression during estrogen-induced synaptic sprouting in the female rat. Exp. Neurol., 165(1): 46–57. Stone, D.J., Song, Y., Anderson, C.P., Krohn, K.K., Finch, C.E. and Rozovsky, I. (1998b) Bidirectional transcription regulation of glial fibrillary acidic protein by estradiol in vivo and in vitro. Endocrinology, 139(7): 3202–3209. Szymczak, S., Kalita, K., Jaworski, J., Mioduszewska, B., Savonenko, A., Markowska, A., Merchenthaler, I. and Kaczmarek, L. (2006) Increased estrogen receptor beta expression correlates with decreased spine formation in the rat hippocampus. Hippocampus, 16(5): 453–463. Tabibnia, G., Cooke, B.M. and Breedlove, S.M. (1999) Sex difference and laterality in the volume of mouse dentate gyrus granule cell layer. Brain Res., 827(1–2): 41–45. Tabori, N.E., Stewart, L.S., Znamensky, V., Romeo, R.D., Alves, S.E., McEwen, B.S. and Milner, T.A. (2005) Ultrastructural evidence that androgen receptors are located at extranuclear sites in the rat hippocampal formation. Neuroscience, 130(1): 151–163. Tanapat, P., Hastings, N.B. and Gould, E. (2005) Ovarian steroids influence cell proliferation in the dentate gyrus of the adult female rat in a dose- and time-dependent manner. J. Comp. Neurol., 481(3): 252–265. Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999) Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat. J. Neurosci., 19(14): 5792–5801. Teter, B., Harris-White, M.E., Frautschy, S.A. and Cole, G.M. (1999) Role of apolipoprotein E and estrogen in mossy fiber sprouting in hippocampal slice cultures. Neuroscience, 91(3): 1009–1016. Towart, L.A., Alves, S.E., Znamensky, V., Hayashi, S., McEwen, B.S. and Milner, T.A. (2003) Subcellular relationships between cholinergic terminals and estrogen receptor-alpha in the dorsal hippocampus. J. Comp. Neurol., 463(4): 390–401. Veiga, S., Garcia-Segura, L.M. and Azcoitia, I. (2003) Neuroprotection by the steroids pregnenolone and dehydroepiandrosterone is mediated by the enzyme aromatase. J. Neurobiol., 56(4): 398–406.

Velisek, L. and Veliskova, J. (2002) Estrogen treatment protects GABA(B) inhibition in the dentate gyrus of female rats after kainic acid-induced status epilepticus. Epilepsia, 43(Suppl 5): 146–151. Veliskova, J. (2006) The role of estrogens in seizures and epilepsy: the bad guys or the good guys? Neuroscience, 138(3): 837–844. Veliskova, J., Velisek, L., Galanopoulou, A.S. and Sperber, E.F. (2000) Neuroprotective effects of estrogens on hippocampal cells in adult female rats after status epilepticus. Epilepsia, 41(Suppl 6): S30–S35. Waters, E.M., Herrick, S.P., McEwen, B.S. and Milner, T.A. (2005) Subcellular localization of progestin receptors in the rat hippocampus. Society for Neuroscience Abstracts, pp. 403–412. Wehrenberg, U., Prange-Kiel, J. and Rune, G.M. (2001) Steroidogenic factor-1 expression in marmoset and rat hippocampus: co-localization with StAR and aromatase. J. Neurochem., 76(6): 1879–1886. Weiland, N.G. (1992) Estradiol selectively regulates agonist binding sites on the N-methyl-D-aspartate receptor complex in the CA1 region of the hippocampus. Endocrinology, 131(2): 662–668. Weiland, N.G. and Orchinik, M. (1995) Specific subunit mRNAs of the GABAA receptor are regulated by progesterone in subfields of the hippocampus. Brain Res. Mol. Brain Res., 32(2): 271–278. Weiland, N.G., Orikasa, C., Hayashi, S. and McEwen, B.S. (1997) Distribution and hormone regulation of estrogen receptor immunoreactive cells in the hippocampus of male and female rats. J. Comp. Neurol., 388(4): 603–612. Wimer, C.C. and Wimer, R.E. (1989) On the sources of strain and sex differences in granule cell number in the dentate area of house mice. Brain Res. Dev. Brain Res., 48(2): 167–176. Woolley, C.S., Gould, E., Frankfurt, M. and McEwen, B.S. (1990) Naturally occurring fluctuation in dendritic spine density on adult hippocampal pyramidal neurons. J. Neurosci., 10(12): 4035–4039. Woolley, C.S. and McEwen, B.S. (1992) Estradiol mediates fluctuation in hippocampal synapse density during the estrous cycle in the adult rat. J. Neurosci., 12(7): 2549–2554.

SECTION IV

Plasticity

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 24

Plastic processes in the dentate gyrus: a computational perspective Brian E. Derrick Department of Biology, The Cajal Neuroscience Research Institute, The University of Texas at San Antonio, 6900 N. Loop 1604 West, San Antonio, TX 78249-0662, USA

Abstract: The dentate gyrus has the capacity for numerous types of synaptic plasticity that use diverse mechanisms and are thought essential for the storage of information in the hippocampus. Here we review the various forms of synaptic plasticity that involve afferents and efferents of the dentate gyrus, and, from a computational perspective, relate how these plastic processes might contribute to sparse, orthogonal encoding, and the selective recall of information within the hippocampus. Keywords: dentate gyrus; LTP; LTD; metaplasticity; pattern separation hippocampal formation, has rekindled interest in the function of the dentate gyrus, as well as indicating a mechanism of plasticity besides LTP and long-term depression (LTD) that also might contribute to memory storage (Shors and Matzel, 1997). Anatomically, the dentate occupies an unusual place in the hippocampal formation. It is not part of the hippocampus (the dentate gyrus, the hippocampus proper — areas CA1–CA3 — together with the hilar region — are structures of the ‘‘hippocampal formation’’). In terms of evolution, it is a recent structure, attaining its present form with the emergence of mammals, the structure itself appears to have been almost an afterthought, tagged on to the hippocampus relatively late in evolution (Treves and Roudi, 2005). From the perspective of hippocampal function, here it has remained — simply the first relay of a three-part ‘‘trisynaptic circuit,’’ where it has been suggested to act as a ‘‘gate’’ to the hippocampus due primarily to its modulation of output with behavioral state (Winson and Abzug, 1978; Bramham and

Plastic processes of the dentate gyrus The dentate gyrus occupies an unusual place in both the history and anatomy of the hippocampal formation. Historically, it was the center of excitement, being the site where the phenomenon of longterm potentiation (LTP), an activity-dependent increase in synaptic efficacy, was first identified (Bliss and Gardner-Medwin, 1973; Bliss and Lomo, 1973). However, in terms of synaptic plasticity, the dentate gyrus has been relegated to ‘‘second place’’ behind the CA1 region, upon which the bulk of studies LTP have been focused for the past two decades. Recently, the dentate has again become a focus of much attention in that it is one of the few neural sites that display neurogenesis after development — dentate granule cells are continually generated well into adulthood (Altman and Das, 1965; Bayer, 1982). This rather remarkable process, which is unique to the dentate gyrus of the Corresponding author. Tel.: +1 210 458 4273; Fax: +1 210 458 5658; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63024-6

417

418

Srebro, 1989). However, we know now that information flow through the trisynaptic circuit may be only one (and not even the primary) pathway of input flow through the hippocampus (Mizumori et al., 1989; Yeckel and Berger, 1990; Do et al., 2002). Such findings offer new clues that have redefined the function and operation of both the dentate gyrus and the hippocampus. More surprising is that the dentate displays numerous forms of neural plasticity. In addition to the phenomenon of LTP, other plastic changes are observed at synapses to and from the dentate, such as LTD and metaplasticity. In addition, long-term changes in granule cell function, characterized by long-term alterations in neurotransmitter and neuromodulator synthesis (White et al., 1987; Hong et al., 1988), the formation of new synaptic contacts on granule cell targets (mossy fiber synaptogenesis, Ben-Ari and Represa, 1990; Cavazos et al., 1991), and the addition of new granule cells well into adulthood (adult granule cell neurogenesis, Altman and Das, 1965) are other plastic processes displayed by the dentate gyrus that, together, have the potential to contribute to accurate and efficient information storage within the hippocampal formation. Because many of these plastic changes involve the expression of numerous genes, and result in long-lasting morphological changes (such as synaptogenesis and neurogenesis), it seems odd that the one structure within the hippocampal formation that shows the most pronounced and sustained genetic, morphological, and functional plastic changes would be overshadowed by the CA1 region of the hippocampus. The purpose of this review is to introduce the myriad of plastic processes that are displayed by synapses of the dentate gyrus, and to describe their possible contribution to information storage on the basis of current information-theoretical views of information processing and memory storage that are thought operative in the hippocampal formation.

The dentate gyrus: A brief description of anatomy and architecture As is the case for most of the hippocampal formation, the dentate is essentially three-layered

archicortex (it is a surprise to many that the hippocampus is actually cortex, albeit the evolutionary older three-layered archicortex; Amaral and Witter, 1989). The principal cell of the dentate gyrus is the granule cell. In the rat, the dentate contains 1–2 million granule cells (Boss et al., 1985), a very large number of cells given that cortical input to the dentate (300,000–400,000 cells of the entorhinal cortex in the rat) is, by comparison, relatively small. The projection of a structure to one with a larger number of neurons, referred to as ‘‘fan out,’’ ‘‘input expansion,’’ or ‘‘input recoding’’ (Rolls and Treves, 1998), offers clues to the function of the dentate. It is an aspect of the dentate gyrus thought crucial for ‘‘pattern separation,’’ a process now thought to be one of the primary functions of the dentate, and essential for the formation of discrete ‘‘memories’’ within the hippocampus proper (Marr, 1971; Morris and McNaughton, 1987; Treves and Rolls, 1992, 1994). Classically, the dentate gyrus was first characterized as the initial stage of the three-stage ‘trisynaptic circuit’ of the hippocampal formation (Andersen et al., 1966). As such, the dentate gyrus represents the first stage of this circuit and the primary target of highly processed cortical information relayed from the entorhinal cortex via the main synaptic input to the dentate, the perforant pathway. The subsequent second stage of the socalled ‘trisynaptic circuit’ is the output from the dentate granule cells, the mossy fibers. The mossy fibers then project to the most proximal dendritic region of pyramidal cells within the CA3 region. The mossy fiber axons of the granule cells are relatively thin and unmyelinated, making them particularly slow as compared with the third and final stage of the trisynaptic circuit, the CA3 projection to CA1 mediated by CA3 ‘‘Schaffer collaterals.’’ Crucial to the more recent views of hippocampal operation is the finding that the primary input to CA3 pyramidal cells is not from the dentate gyrus. Rather, the primary synaptic input to CA3 pyramidal cells is recurrent input arising from other CA3 pyramidal cells (Amaral et al., 1993). This recurrent excitatory input (termed the commissural/associational CA3 pathway) is of particular interest computationally, as such recurrent networks are capable of performing autoassociative

419

Fig. 1. Schematic representation of the dentate–CA3 interface. Perforant path inputs to the CA3 region arrive directly via synapses on the distal CA3 dendrites, and indirectly by a disynaptic relay from the dentate gyrus via the sparse but powerful mossy fiber–CA3 projections. Recurrent connections constitute the greatest input to CA3 pyramidal cells, and are thought to serve as the substrate for autoassociative storage and recall. Adapted with permission from Amaral et al. (1990) and Rolls and Treves (1998).

function and thus represent a substrate for ‘‘content addressable’’ memories (or CAMs), a more general term for distributed memory devices (Kohonen, 1984). The view of the hippocampal formation as a ‘trisynaptic circuit’ is, however, an oversimplification of a much more elaborate and parallel input system within the hippocampus (Yeckel and Berger, 1998). Anatomical studies indicate that there are substantial direct perforant path projections to both the CA3 and CA1 regions of the hippocampus (Andersen et al., 1966; Amaral et al., 1990; Claiborne et al., 1993). The dentate and CA3 regions receive direct input from the same populations of neurons in layers II of the medial and lateral entorhinal cortex (Amaral and Witter, 1989). In addition, layer III neurons of the lateral and medial entorhinal cortex project directly to apical dendrites of CA1 pyramidal cells and these projections to the CA1 region constitute a separate ‘‘temporoammonic pathway.’’ These inputs, once ignored and deemed insignificant, were recognized as potentially important by neuroanatomists (Hjorth-Simonsen and Jeune, 1972; Steward 1976; Steward and Scoville, 1976; Buzsaki, 1988; Witter et al., 1988). Subsequent physiological studies in vivo (Yeckel and Berger, 1990; Berzhanskaya et al., 1998) demonstrated that the direct perforant path projections to the CA3 region are not only capable

of discharging pyramidal cells, but also display LTP (Do et al., 2002) as well as associative LTP (Martinez et al., 2002), a feature crucial to many of the models of CA3 computation (Treves and Rolls, 1994). A more realistic ‘‘schematic’’ view of dentate–CA3 hippocampal circuitry is shown in Fig. 1. An underappreciated feature of the cortical input to the hippocampal formation is that not only can the direct perforant path inputs activate CA3 pyramidal cells, but CA3 pyramidal cells are also the first cells to respond to perforant path input (Yeckel and Berger, 1990; Do et al., 2002). Extracellular measures of the summed output of the CA3 region and the dentate gyrus, as reflected in population spike latency, indicate that the CA3 region responds faster than the dentate gyrus, with spike discharges in vivo preceding granule cell spikes by 0.5–3 ms (Fig. 2). Thus, it is the CA3 region, rather than the dentate, that is the first population of principal cells to respond to cortical input relayed by the perforant pathway. As noted in the original study (Yeckel and Berger, 1990), this finding is in line with a number of studies indicating early evoked responses in the CA3 region that precede dentate responses (Segal, 1972). Because the CA3 region is the first region to respond with spike output to perforant path activity, the primacy of CA3 responses is likely relevant to the function of the rather unusual ‘‘backward’’ projections from CA3 pyramidal cells to hilar neurons

420

Fig. 2. Comparison of perforant path responses evoked in the dentate gyrus and CA3 region. Shown are field potentials recorded in the dentate gyrus (hatched lines) or the CA3 pyramidal layer (solid lines) following activation of the medial (top) or lateral perforant pathway (bottom traces). Responses were recorded in vivo from the same animals under pentobarbital anesthesia. Calibration: 0.5 mV, 5 ms (CA3); 1 mV, 5 ms (dentate).

(Scharfman, 1996). Because these interneurons can exert both excitatory and inhibitory actions on dentate granule cells, the projection from the CA3 region to these interneurons may serve to modulate dentate granule cell output in response to direct perforant path–CA3 input. The granule cells of the dentate gyrus dentate appear, by many accounts, unremarkable. They are small (6–10 mm) and have extensive, conical, spinous dendrites that extend almost 1 mm into the stratum moleculare (or the ‘‘molecular layer’’) of the dentate (Steward, 1976; Amaral and Witter, 1989). It is here in the molecular layer that the dentate gyrus receives its input from three primary sources: the lateral entorhinal cortex (LEC), which targets the outer one third of the dentate molecular layer; the medial entorhinal cortex (MEC), whose projections occupy the middle one third of the granule cell dendrites in the molecular layer; and the commissural/associational projections, originating from hilar glutamatergic ‘‘mossy’’ cells and terminating in the innermost one third of the molecular

layer (Amaral and Witter, 1989). The dentate associational projections reflect yet another recurrent excitatory pathway within the hippocampal formation, albeit one that is relayed disynaptically via mossy cells (Amaral, 1978; Amaral and Witter, 1995). The granule cells themselves display features that are unique as well: they have a high threshold for activation, and granule cell bursting occurs rarely in vivo (Jung and McNaughton, 1993), although granule cells can fire sustained bursts during theta rhythm (Munoz et al., 1990), a 4–12 Hz oscillation that is a prominent feature of hippocampal activity during exploration and, presumably, during learning (Buzsaki, 2002). Granule cells are notoriously silent, as evidenced by the difficulty in obtaining ‘‘place fields’’ — single unit firing that correlates with a specific place field in an environment (O’Keefe and Nadel, 1978) — and they often fire at very low rates (approximately 0.5 Hz, Jung and McNaughton, 1993). As discussed further below, these tight constraints on granule cell activity may be viewed as

421

another feature crucial for ‘‘pattern separation’’ and the formation of sparse outputs by the dentate gyrus (Skaggs and McNaughton, 1992). Another unusual aspect of the dentate is that the V-shaped blades of this structure appear not to be homogeneous, actually differing in a number of substantial ways. They appear to be under separate genetic control, as the lower (infrapyramidal) blade and the mossy fiber (suprapyramidal) layer differ among strains of mice (Schwegler et al., 1990). In addition, mossy fiber efferents arising from granule cells in the internal (suprapyramidal) layer ramify and extend through the stratum lucidum of the CA3 region, whereas the mossy fibers arising from the outer (infrapyramidal) blade of the dentate ramify extensively in the hilus and terminate primarily in the infrapyramidal layer of the CA3 region lying just below the CA3 pyramidal cell layer, or stratum pyramidale (Claiborne et al., 1986). The two blades may also have functional differences, because the infrapyramidal region develops greater mossy fiber sprouting following seizures or highfrequency mossy fiber stimulation (Cavazos et al., 1991; Escobar et al., 1997; Scharfman et al., 2002), as well as spatial learning (Ramirez-Amaya et al., 2001). In support of possible ‘‘divisions of labor’’ among the dentate blades, a greater degree of excitability is observed in the lower (infrapyramidal) blade, possibly due to differential innervation of granule cells by inhibitory interneurons (Scharfman et al., 2002). Moreover, dendrites on granule cells in the infrapyramidal layer show fewer spines (Claiborne et al., 1986). Interestingly, while both blades show the expression of Arc, an immediate-early gene associated with synaptic plasticity, following stimulation of the perforant path that induces LTP, Arc expression in behaving animals associated with learning is seen only in granule cells in the suprapyramidal (inner) blade lying between the hilus and the CA1 region (Chawla et al., 2005). However, the functional relevance of these differences among the blades of the dentate gyrus is unknown. The efferents of the granule cells, the mossy fibers, also display unusual features: these granule cell efferents are thin, unmyelinated fibers that, upon their exit from granule cells, form many bifurcations, projecting primarily to the proximal spines of CA3 pyramidal neurons. Although the

connections of granule cells to CA3 are sparse, with each granule cell making contact with only 8–10 CA3 pyramidal cells, they form unusual synapses on CA3 cells, with each mossy fiber bouton essentially engulfing a highly elaborated dendritic CA3 spine with multiple synaptic contacts, often referred to as ‘‘thorny excrescences’’ (Gonzales et al., 2001). The mossy fiber synapse itself is unusual in several respects, with the primary difference being that LTP observed at these synapses is of a type that does not require the activation of N-methyl-D-aspartate (NMDA) receptors (Harris and Cotman, 1986). In addition to the CA3 pyramidal cells, the mossy fibers also target numerous neurons within the hilar region. The array of neuronal subtypes within the hilar region of the dentate is impressive (Amaral, 1978); however the most studied neurons in the hilus are the so-called ‘‘mossy cells.’’ These cells are innervated by the mossy fibers and, by virtue of their excitatory glutamatergic projections, both ipsilaterally and contralaterally to the inner molecular layer, comprise the commissural/ associational system of the dentate gyrus. This recurrent excitatory input has been scrutinized recently (Kleschevnikov and Routtenberg, 2003), and studies have revealed that these projections show plasticity both at the mossy fiber–mossy cell synapse and at their contacts with granule cells in the inner molecular layer. Those familiar with the major target of the dentate granule cells, the CA3 region, may notice a similarity between the dentate and CA3 region in that both have extensive recurrent excitatory inputs that also project contralaterally. Thus, both the dentate and CA3 region have extensive, bilateral, and recurrent excitatory feedback. Thus these two structures can be thought of as a single, bilateral, highly re-entrant structure. The functional implications of this feature remain to be defined.

Information/theoretic views of the dentate gyrus Although the dentate gyrus defies any single account of function based on its operation (such as its early conceptualization as ‘‘gating’’ input to the CA3 region (Winson and Abzug, 1978), clues to its

422

functions may be provided by current informationtheoretic models of hippocampal information processing. To more fully appreciate the contribution of the dentate to hippocampus-based memory, it is necessary to consider information-theoretic models of information storage in the CA3 region of the hippocampus. The late David Marr’s original conceptualization of hippocampal function remains central to most information-theoretic theories of hippocampal operation today (Marr, 1971). As suggested by Marr, the CA3 region may serve as an ‘‘autoassociator’’ by virtue of its extensive recurrent (CA3–CA3) excitatory connections (Morris and McNaughton, 1987; Treves and Rolls, 1994; Rolls and Treves, 1998). Neural network models of autoassociators were pioneered by theoreticians such as Tuevo Kohonen (Kohonen et al., 1977; Kohonen, 1984) and later re-formalized by John Hopfield (Hopfield, 1982). In such models autoassociators are networks which by virtue of their extensive recurrent input are capable of unique computational operations. Representations in autoassociative systems involve the formation of ‘‘attractors’’ (or stable ensembles of active, recurrently connected cells as a result of plastic changes among recurrent synapses). Thus autoassociative architectures and their recurrent connections can allow for the formation of stable ‘‘attractor’’ states. This architecture also can enable the recall of an entire pattern of CA3 activity in response to the presentation of only a subset of the original pattern. This feature, termed ‘‘pattern completion,’’ is a fundamental computational feature of autoassociators (Kohonen et al., 1977). If one thinks of an ensemble of many recurrently connected CA3 neurons with strengthened synapses among them as a representation, one can see that the activation of a subset of this ensemble could, via recurrent activation of other CA3 neurons via strengthened synapses, recruit the entire original ensemble. Thus recurrent inputs provide the substrate of CA3 ‘‘autoassociation’’ and pattern completion from partial inputs, a property that may encode not only single stable ensembles of co-active CA3 pyramidal cells, but also sequences of stable states, provided that the output from the CA3 system can, via the CA3 recurrent inputs,

become associated with subsequent CA3 inputs over time (Levy, 1996). The view of autoassociative processes in the CA3 region, first formalized by Marr (1971), remains a feature of most models of hippocampal function (Morris and McNaughton, 1987; Treves and Rolls, 1994; Hasselmo et al., 1995; McClelland et al., 1995; Rolls and Treves, 1998) and is supported by both models of the CA3 region and behavioral studies (Nakazawa et al., 2002). As noted above, the recurrent CA3–CA3 connections constitute the majority of excitatory inputs to CA3 pyramidal cells, and these fibers also display Hebbian LTP (Harris and Cotman, 1986). It is thought that representations within the CA3 autoassociative system involve the formation of attractor states (simply relatively stable ensembles of active, recurrently connected CA3 cells) as a result of plastic changes at the recurrent CA3–CA3 synapses. These plastic changes are thought to occur by potentiation of active CA3–CA3 synapses whose activity is initiated by the direct perforant path–CA3 inputs, and ‘‘fixed’’ by strong postsynaptic activity initiated by the mossy fiber synapses from dentate granule cells. In this scheme, the mossy fiber synapses are thought to act as ‘‘detonator’’ synapses (Morris and McNaughton, 1987) that direct associative plasticity, and thus encoding, in this system. Due to the relative proximity of these synapses to the CA3 pyramidal cell body, activation of these ‘‘detonator’’ synapses can initiate CA3 cell firing and, via postsynaptic depolarization, induce LTP among other synapses to the CA3 pyramidal cell that were recently active (Levy and Steward, 1979, 1983). Thus, encoding of information in the CA3 region is thought to result from the activation of CA3 recurrent collaterals by perforant path input and the subsequent potentiation of recently active perforant path–CA3 and CA3–CA3 synapses by mossy fiber ‘‘detonator synapses.’’ Although autoassociators are capable of pattern completion given a partial input of the original pattern, they have their limitations. Crucial for efficient encoding and accurate recall in autoassociative systems is a sufficient separation of patterns that are encoded within the CA3 region, such that stored representations in the CA3 system have

423

minimal overlap with other, previously encoded ensembles (orthogonalization). This separation of patterns is necessary because accurate encoding and retrieval within autoassociative memory systems can occur only if correlations and redundancies among discrete attractor states (ensembles of active CA3 neurons) are reduced (see Rolls and Treves, 1998). Thus, encoding information in an autoassociator demands special considerations and requires the formation of orthogonal, or uncorrelated, ensembles of CA3 attractors. How this is done is the subject of much speculation, although a feature common to many models is that orthogonalization (sometimes referred to more generally as ‘‘pattern separation’’) requires sparse encoding, a process that is crucial for encoding in the CA3 region. This sparsification is accomplished primarily by transformation of perforant path inputs to the dentate into a sparse code that is then relayed to the CA3 region (Morris and McNaughton, 1987). The formation of sparse dentate outputs can increase the capacity of a recurrent autoassociative system, and by virtue of a fewer number of outputs (and plasticity among them) can also serve to reduce any similarities among ensembles formed within the CA3 region. Sparse encoding can serve as one process that minimizes similarity and interference among CA3 attractors, because representations using a minimal number of elements are less likely to overlap with other stored patterns (Rolls and Treves, 1998). In many models of hippocampal function, the task of ‘‘pattern separation’’ and orthogonalization of cortical inputs relayed to the CA3 region is ascribed to the dentate gyrus (Morris and McNaughton, 1987; Treves and Rolls, 1994; Rolls and Treves, 1998). The dentate gyrus can serve in the formation of sparse, orthogonal CA3 ensembles via several mechanisms. As discussed extensively by Rolls and Treves (1998), the dentate gyrus is thought to form sparse outputs to the CA3 region, a process that can be facilitated in a variety of ways by features of the dentate gyrus. Foremost among these features are the anatomical characteristics of the dentate gyrus–mossy fiber system. This system appears ideally suited to ensure the formation of sparse, uncorrelated input patterns relayed to the CA3 region. The rat dentate gyrus

consists of 1–2 million granule cells that receive input from only 300,000 cells in the entorhinal cortex (Boss et al., 1985). This ‘‘fan out’’ of cortical input to a structure with many more elements is one way to enforce pattern separation (referred to as ‘‘input expansion’’ or ‘‘expansion recoding;’’ Rolls and Treves, 1998). In addition, both the relative quiescence of granule cells in the dentate and the sparse connectivity of the mossy fiber projections (so that each granule cell forms mossy fiber contacts with only 10–15 CA3 pyramidal cells; Amaral et al., 1990) together make it unlikely that any given CA3 pyramidal cell can be activated by more than one granule cell. Thus the anatomically sparse dentate–CA3 connectivity by the mossy fibers is an important feature for the formation of sparse, orthogonal CA3 representations. Together, these aspects of the dentate are ideally suited for pattern separation, and the formation of sparse, orthogonal outputs to the CA3 region that allow the encoding of discrete, stable ensembles or ‘‘attractors’’ in the CA3 autoassociative system. While important, the anatomical features of the mossy fiber output are but one way the dentate may relay sparse transforms to the CA3 autoassociative system. As we will discuss later, when taken in a computational context, the plastic features of the dentate gyrus, including LTP, the various forms of LTD, and metaplastic processes, together, not only may ensure sparse, orthogonal outputs to the CA3 autoassociative system, but also may allow for the dentate to encode discrete attractors for similar, or even identical inputs that are encoded over time. In this view, the combination of plastic processes in the dentate that allow the formation of sparse, orthogonal CA3 inputs also may allow partial inputs to converge on distinct stable states encoded by very similar inputs over time — a feature characteristic of episodic memory, long thought of as the principal type of memory involving the human hippocampus (Tulving, 1983).

Synaptic plasticity in the dentate gyrus — LTP Foremost among the plastic processes studied in the dentate gyrus is LTP. The hypothesis that

424

Fig. 3. Perforant path–dentate LTP and heterosynaptic LTD. Plot of field EPSP slopes of medial (K) and lateral (J) perforant path–dentate responses following high-frequency stimulation of the medial perforant pathway. In addition to the potentiation observed at medial perforant path synapses, a sustained depression of inactive lateral perforant path synaptic responses (‘‘heterosynaptic LTD’’) is observed. Calibration: 0.5 mV, 5 ms. Villarreal and Derrick (2006).

memory involves changes in synaptic efficacy has remained a primary assumption for quite some time (Ramon y Cajal, 1937; Hebb, 1949; James, 1890). LTP is a rapidly induced and long-lasting change in synaptic strength that is observed following high-frequency afferent activity. LTP was first described formally in studies of the rabbit perforant path projections to the dentate gyrus using an in vivo preparation (Bliss and Lomo, 1973), and the first studies that addressed LTP longevity also were conducted using perforant path–dentate recordings (Barnes, 1979). However, since its discovery, and the subsequent discovery of similar plastic processes in other brain structures (from the amygdala to spinal cord), the bulk of studies of LTP, particularly in vitro, have focused on the Schaffer–CA1 synapse (Fig. 3). LTP has features that make it a compelling candidate mechanism that may contribute to information storage. LTP is induced rapidly after its induction, which, in the laboratory, requires a sufficient postsynaptic depolarization that usually is provided by high frequency stimulation of a sufficient number of afferents. Dentate LTP also is ‘‘input specific,’’ that is, LTP is confined to synapses of afferents activated with stimulation.

This synapse specificity is in contrast to the nonassociative phenomena, such as sensitization, wherein a general increase in responsivity of the postsynaptic cell is observed. Importantly, input specificity is a feature essential for any kind of memory that requires both high capacity and high fidelity, as this feature provides sparsity by increasing the number of modifiable inputs that can aid in encoding and retrieval in autoassociative systems. Another crucial feature of LTP making it an attractive potential mechanism for information storage is that it is remarkably persistent. Prior to the discovery of LTP, the only activity-dependent electrophysiological change of substantial duration was post-tetanic potentiation (PTP), a primarily presynaptic phenomenon lasting from seconds to minutes (Zucker and Regehr, 2002). By contrast, in the hippocampal formation, LTP can persist from hours to weeks or months, depending on the stimulation parameters. In an intact animal, LTP usually decays within several weeks (Barnes, 1979). A number of investigators have dismissed LTP as a long-term memory mechanism because its duration is far too short to account for memory

425

that lasts years (Shors and Matzel, 1997). While plasticity that lasts only days to weeks is too brief to store long-term memory, the limits of LTP longevity may be an important feature of LTP in the hippocampal formation. One reason is that LTP, at least within the hippocampus, need not be permanent. This is because, as is thought to be the case in humans (Scoville and Milner, 1957), the rat hippocampus also is thought to have a time-delimited role in memory, with persistent, ‘‘long term’’ memory being gradually consolidated in neocortical areas over time (Marr, 1971; Squire, 1992). Currently, the most popular view is that memories formed by the hippocampus are transferred to and consolidated in the neocortex, possibly during slow wave (NREM) sleep states (Wilson and McNaughton, 1993; Cartwright, 2004), a process that usually requires 2–3 weeks in the rat, as indicated by both lesion and imaging data (Bontempi et al., 1999). Thus, if the hippocampus does indeed serve as a temporary repository, or ‘‘buffer’’ for information, LTP may not need to persist for extended periods of time. In this scheme, any long-term retention of information within the hippocampus beyond the a few weeks might even be detrimental to hippocampal-based memory, resulting in interference with previously stored patterns of synaptic activity (Rosenzweig et al., 2002). How long, then, can LTP last? We reported that, in perforant path projections to the dentate gyrus, LTP decay is an active process mediated by NMDA receptors, and blocking these receptors can prevent decay (Villarreal et al., 2002). Therefore it appears that LTP doesn’t simply ‘‘fade;’’ instead, it is actively ‘‘erased’’ via processes that require NMDA receptor activation. A likely candidate mechanism for such ‘‘erasure’’ is LTD. As discussed more fully below, dentate LTD, like LTP, requires the activation of NMDA receptors (Desmond et al., 1991; Christie and Abraham, 1992b). Thus, NMDA receptor activation and a reversal of LTP by the induction of LTD (also referred to as ‘‘depotentiation,’’ both discussed below) may mediate LTP decay. However, NMDA receptors on dentate granule cells can be activated by low frequency perforant path synaptic activity and even miniature EPSPs (Dalby and

Mody, 2003). Consequently, background synaptic activity completely unrelated to learning may be sufficient to mediate LTP decay at perforant path–granule cell synapses, suggesting that synaptic activity unrelated to learning also may contribute to LTP decay. Regardless, because LTP decay requires activation of NMDA receptors, LTP decay still reflects an active process (Villarreal et al., 2002). Thus perforant path–dentate LTP, at least theoretically, is potentially permanent, and could last as long as the synapse itself, provided that it is not ‘‘erased’’ by subsequent synaptic activity and NMDA receptor activation. LTP also is ‘‘associative.’’ Specifically, if activity in a set of afferents to a common postsynaptic target induces LTP, then individual active synapses can be recruited to express LTP, provided that the synapse is coactive within a delimited window of time (Levy and Steward, 1983). The phenomenon of associativity is derived from the requirements for activation of the NMDA receptor (specifically, both presynaptic release of glutamate and postsynaptic depolarization, with the latter being essential for relieving the magnesium block of the NMDA channel). Associativity is a property derived directly from, and is essentially identical to, the property of ‘‘cooperativity’’ (McNaughton et al., 1978). Cooperativity indicates LTP induction has a threshold, and activation of a threshold number of afferents is essential to induce LTP. Because LTP is induced in a ‘‘cooperative’’ manner, it follows that associativity is an inherent property of ‘‘cooperativity.’’ Although it has been suggested that ‘‘cooperativity’’ and ‘‘associativity’’ are distinct, with cooperativity reflecting a threshold of afferent activity necessary to induce LTP and associativity involving distinct sets of afferents, it should be noted that the first demonstration of cooperativity used two distinct afferent systems to the dentate gyrus, the medial and lateral perforant pathways (McNaughton et al., 1978). However, because interactions among active afferent fibers could not be ruled out, the term cooperativity, rather than associativity, may have been chosen out of rigor, as the term associativity, as envisaged by Hebb, tacitly implies a postsynaptic integration of presynaptic activity, something that could not be ruled out with

426

‘‘cooperative’’ effects. It should be noted here that these studies (McNaughton et al., 1978; Levy and Steward, 1983) were conducted at a time when the NMDA receptor was unheard of. Subsequently, a convincing demonstration of Hebbian-like associativity was provided by studies showing that pairing of low intensity, low frequency afferent activity with postsynaptic depolarization is capable of inducing LTP (Gustafsson and Wigstrom, 1986; Gustafsson et al., 1987). It is often assumed that mechanisms underlying LTP can be generalized. However, what has emerged from recent studies is that the mechanisms underlying the induction, and possibly the maintenance and expression, of LTP not only differ at different synapses within the hippocampal formation, but may even involve a number of distinct mechanisms, even at a single synapse. These differences can arise from numerous variables, including what stimulation parameters are used, and even the animal’s behavioral state at the time of LTP induction (Davis et al., 2004; see below). Thus there may be a multitude of activity-dependent mechanisms that serve to increase synaptic strength, all of which have been labeled ‘‘LTP.’’ Yet these processes may have distinct mechanisms of induction (conditions necessary to initiate LTP) and/or maintenance (mechanisms initiated by LTP induction that underlie its expression for sustained periods of time). Such differences in both induction and maintenance mechanisms may be indicative of distinct ‘‘forms’’ of LTP. The heterogeneity and complexity of LTP are illustrated by the features of LTP induction in the primary input to the dentate — the perforant pathway. As noted above, the perforant pathway projections originate from layer II neurons within the medial and the lateral entorhinal cortices, and they target, respectively, the middle one third of the molecular layer for medial perforant path (MPP) afferents, and the outermost third of the molecular layer for the lateral perforant path (LPP) inputs (Steward, 1976; Amaral and Witter, 1989). Responses to the medial perforant path projection are, for a variety of reasons, the most robust, and many studies that employ stimulation of the perforant path often are likely to evoke responses

from medial perforant path afferents. This may be due, at least in part, to the fact that the medial and lateral projections are stratified in the angular bundle, so that only stimulation of fibers deep in the angular bundle evokes lateral perforant path responses (McNaughton and Barnes, 1977). Lateral perforant path–dentate responses differ from medial responses in a number of ways, including a slower rising phase of dentate field EPSPs (due to the more distal regions of the granule cell dendrite where lateral perforant path synapses terminate). In addition, medial perforant path responses show a pronounced frequency depression of synaptic responses, whereas lateral perforant path responses show robust frequency facilitation (McNaughton and Barnes, 1977). The reasons for these differences in afferent responses are not clear, although it likely involves differences among the afferents themselves, as depression and facilitation are also observed when paired pulses are delivered to medial and lateral perforant path projections to the CA3 region, (Do et al., 2002). It may come as no surprise that LTP in these two pathways differ in a number of important ways. The medial perforant pathway is a glutamatergic afferent system and displays LTP that depends on activation of the NMDA glutamate receptor. Thus LTP at medial perforant path synapses may reflect a form of LTP induction similar to the more widely-studied type at the Schaffer–CA1 synapse, as well as among neocortical regions (Kirkwood et al., 1993; Trepel and Racine, 1998), and appears to be the primary receptor involved in the induction of LTD as well (Desmond et al., 1991; Christie and Abraham, 1992b). NMDA receptors are activated by glutamate released by the presynaptic terminal. NMDA receptors are also voltage-dependent, such that current flow is prevented at hyperpolarized membrane potentials due to a cationic block of the channel by magnesium. However, the combination of presynaptic activity with postsynaptic depolarization sufficient to remove the cationic block of the NMDA receptor pore allows current flow (calcium influx via NMDA receptor channels), allowing for the induction of LTP or LTD. This property of NMDA receptors, the requirement for both presynaptic activity together with a

427

sufficient level of postsynaptic activity, is essentially a physical instantiation of the synaptic learning rule formalized by Donald Hebb in the 1940s (Hebb, 1949). It is now referred to as the ‘‘Hebb rule,’’ this simple local learning rule requiring correlated pre- and postsynaptic activity has significant computational power when employed in various neural network architectures. Networks with elements that use the Hebb rule display important computational features, and therefore it is thought to be a principal ‘‘coincidence detector’’ that mediates synaptic plasticity based on correlated pre- and postsynaptic activity. As with many synapses in the hippocampal formation, LTP at the medial perforant path–dentate synapse is blocked by antagonists selective for NMDA receptors. As with most known forms of LTP, the influx of calcium postsynaptically as a result of NMDA receptor activation is thought to be the crucial first step in LTP induction (Barrionuevo et al., 1980). The subsequent steps following calcium entry are not yet defined in detail, but as with LTP at other synapses, a concerted effort has been made to identify important molecular and biochemical events that eventually increase synaptic strength. Here a number of protein kinases have been implicated, including a-calcium/calmodulin kinase II (CaMKII; Thomas et al., 1994; Wu et al., 2006), protein kinase A (Wu et al., 2006), and the ERK/MAPK cascade (Maguire et al., 1999; Davis et al., 2000; Sweatt, 2001; Wu et al., 2006). The subsequent phosphorylation of the cAMP response element binding protein (CREB) is likely the first transcription factor activated by these kinases (Ying et al., 2002), and this is thought to initiate the expression of immediate-early genes (IEGs; Schulz et al., 1999; Davis et al., 2000). The later phase (>3 h) of LTP maintenance in the dentate requires the synthesis of proteins (Krug et al., 1984; Nguyen and Kandel, 1996) and transcription of new mRNAs (Frey et al., 1996), and appears to involve a later activation of protein kinase C (Colley et al., 1990; Thomas et al., 1994). Several IEGs are implicated in the expression of LTP. These genes are induced rapidly in a protein synthesis-independent manner, and generate a number of proteins that, together with other IEG proteins, can form transcription factors that

regulate the expression of other ‘‘late effector’’ genes. The major IEGs implicated in LTP include Arc (Steward et al., 1998), Homer (Kato et al., 1997), and Zif268 (Davis et al., 2000; Jones et al., 2001). Arc is a very interesting IEG in that the mRNA for this gene is transported into granule cells dendrites, where the synthesis of Arc proteins occurs (Steward et al., 1998; Guzowski et al., 2000). Similarly, the transport of CaMKII mRNAs to synaptic regions, and enhanced synthesis of CaMKII, also is observed (Havik et al., 2003). Finally, there is a protracted increase in the synthesis of NR2B subunits of the NMDA receptor (Thomas et al., 1996; Williams et al., 2003). Together, these proteins are suggested to participate in the formation of elaborate modifications to the postsynaptic density (PSD) of granule cell synapses, a process thought to involve NR2B NMDA receptor subunits and CaMKII via interactions with actin in PSD. Here, elaboration of the postsynaptic element and localization of CaMKII may mediate the insertion of new receptors and/or receptor subunits (Lisman and McIntyre, 2001), also referred to as ‘‘AMPA receptor trafficking’’ (see Malinow, 2003, for review). These genes also may mediate the morphological changes reported to occur at axospinous synapses in the molecular layer (Trommald et al., 1996), such as changes in postsynaptic structure (Desmond and Levy, 1983, 1986). However, few data exist for the role of AMPA trafficking in dentate granule cells, as most data supporting this process in LTP were obtained in studies of LTP in the Schaffer–CA1 pathway. Many of the genes thought to be involved in dentate LTP induction (genes that also code for proteins that can associate with the PSD and that are implicated in AMPA receptor trafficking) are also thought to be involved in LTP at the Schaffer–CA1 synapses. Thus NMDA receptor-dependent LTP as observed at medial perforant path–dentate and Schaffer–CA1 synapses may reflect a common, prototypical ‘‘form’’ of LTP induction and expression. Induction of LTP at the lateral perforant path synapse is quite different from the medial perforant path projections. As first demonstrated by Bramham et al. (1988), naloxone, an antagonist of the opioid receptor, blocks lateral perforant

428

path–dentate LTP in vivo. This was verified by studies in vitro, where naloxone blocked lateral perforant path–dentate LTP, and m opioid receptor-selective agonists facilitated LTP induction (Xie and Lewis, 1991). However, in these studies, naloxone had little or no effect on low frequency lateral perforant path–dentate responses. The lack of effect of this opioid receptor antagonist on responses evoked at low frequencies suggests that endogenous opioid peptides play a frequency-dependent role in LTP induction at lateral perforant path–dentate synapses (Bramham et al., 1988; Matthies et al., 2000; Krug et al., 2001). Although the requirement for opioid peptides in LTP induction may seem unusual, it may be less surprising when one considers that the lateral entorhinal cortex projections co-release opioid peptides with its principal transmitter, glutamate (Chavkin et al., 1985). These peptides, derived primarily from proenkephalin, with [Met]- and [Leu]-enkephalin being the primary opioid peptides produced by proenkephalin and released by the lateral perforant path, are stored in dense-core vesicles and are released in an activity-dependent manner (Neumaier and Chavkin, 1989; Wagner et al., 1990, 1991). The frequency dependence of opioid peptide release by the lateral perforant path is common to peptide co-transmitters; principally, their release often requires intense or repetitive activity (Hong et al., 1988). It is thought that such activity is essential for the release of neuropeptides that are stored in dense core vesicles, as they often are remote from both release sites and plasma membrane presynaptic terminals, and thus may require large amounts of calcium for their release. This aspect of peptide storage and release may also explain why opioid receptor blockade has little effect on synaptic responses evoked in opioidergic projections at low frequencies. Tacit in this interpretation is the idea that high-frequency lateral perforant path activity is necessary for LTP induction in order to achieve both postsynaptic depolarization and the frequency-dependent release of opioid neuropeptides. A similar dependence of lateral perforant path LTP on high-frequency afferent activity is also observed in vitro, as pairing low frequency presynaptic activity with postsynaptic depolarization,

while effective in inducing LTP in the medial perforant–dentate synapses, is not effective in inducing LTP in lateral perforant path–dentate synapses (Colino and Malenka, 1993). A similar lack of ‘‘single pulse associativity’’ also is observed at mossy fiber synapses on CA3 pyramidal cells, a synapse that also contains and releases opioid peptides (Chavkin et al., 1985) that are thought essential for Hebbian LTP induction at the mossy fiber synapse (Derrick et al., 1994a; Williams and Johnston, 1996). While this effect has been interpreted as reflecting a lack of associativity at mossy fiber synapses (Zalutsky and Nicoll, 1990), the lack of single pulse mossy fiber associativity may simply reflect this requirement for mossy fiber activity necessary for the frequency-dependent release of opioid peptides (Neumaier and Chavkin, 1989; Wagner et al., 1990, 1991; Derrick and Martinez, 1994a). Thus the requirement for high frequency activity for LTP induction at lateral perforant path–dentate synapses similarly may reflect a requirement for the frequency-dependent release of opioid peptides for LTP induction (Colino and Malenka, 1993). There are a number of other actions of both pro-enkephalin and pro-dynorphinderived opioid peptides released by dentate granule cells. Most notable is the rather surprising finding that dynorphins, which act at kappa receptors and exert primarily inhibitory effects, attenuate perforant path synaptic responses and LTP induction, most likely via dendritic release of dynorphin-derived peptides (Wagner et al., 1993), adding a different dimension to the roles of opioid peptides in hippocampal information processing. The actions of opioids within the hippocampus are varied and often depend on the receptors involved as well as the site of the receptors. However, the primary effect of many opioid peptides in the hippocampus is that they elicit excitation (Fry et al., 1979; Corrigall, 1983), which is observed with activation of both the m and d opioid receptor types (Piguet and North, 1983). The application of exogenous opioids can even elicit seizures, and repeated application can eventually elicit kindling (Cain and Corcoran, 1985). These excitatory effects are unusual for m and d opioid receptors, as these opioid receptor types are usually coupled to the superfamily of GTPase-binding proteins

429

that inhibit cellular responses (Piguet and North, 1983). Why would neuromodulators normally associated with inhibitory actions elicit excitation within the hippocampus? This paradox was clarified in subsequent studies indicating that one principal mechanism by which opioids mediate their excitatory effects is their attenuation of GABA release (Cohen et al., 1992). Thus the excitatory effects of proenkephalin-derived opioid peptides result from an attenuation of GABAergic inhibition, an effect thought to result from the m or d opioid receptor-mediated enhancement of potassium conductances in GABAergic interneurons, which inhibit the release of GABA (North et al., 1987). How could these disinhibitory effects be important in LTP induction? Agents that attenuate GABAergic inhibition generally facilitate LTP induction (Wigstrom and Gustaffson, 1985). This effect is thought to arise from the enhanced depolarization, and presumably, enhanced NMDA receptor activation that follows a reduction in GABAergic inhibition. In fact, the disinhibitory effect of opioid peptides likely underlies their requirement for the induction of lateral perforant path–dentate LTP, as the block of lateral perforant path LTP following naloxone can be prevented by co-application of a GABAA receptor antagonist (Bramham and Sarvey, 1996). This theory of the contribution of opioid peptides to LTP induction has been longstanding (Swearengen and Chavkin, 1987). Although its modulation by opioids has been clarified, the mechanisms underlying lateral perforant path LTP induction are not as simple. Studies by Bramham and colleagues have shown that, while the activation of both m and d-type opioid receptors is crucial for LTP induction (Bramham et al., 1991a; Bramham and Sarvey, 1996), the facilitation of LTP by opioid peptides is unlikely to arise via a facilitation of NMDA receptor activation. This is because LTP at the lateral perforant path–dentate synapse was found to be independent from NMDA receptors in vivo, as competitive NMDA receptor antagonists were ineffective in blocking its induction (Bramham et al., 1991b). Thus, the way that opioid receptor activation facilitates LTP induction at lateral perforant

path–dentate synapses, is not likely to be related to NMDA receptors. Perhaps the enhanced postsynaptic depolarization afforded by opioid peptides permits calcium influx, which is necessary for the induction of virtually all forms of LTP, from other sources. For example, voltage-dependent calcium channels can respond to the depolarization afforded by the disinhibitory effects of opioid peptides. In addition, d opioid receptors are suggested to activate phospholipase C and the formation of IP3, a second messenger that often regulates intracellular calcium release (Bramham, 1992). Thus direct actions of opioid peptides on principal cells may be necessary for LTP induction by enhancing intracellular Ca++ levels by enhancing release of Ca++ from intracellular stores. The suggestion of a distinct, NMDA receptorindependent, opioid receptor-dependent form of LTP in lateral perforant path afferents is also supported by our own studies in area CA3, where projections from the lateral perforant path terminates in the stratum lacunosum-moleculare. This direct perforant path projection to CA3 pyramidal cells is appreciable: the average CA3 pyramidal cell receives the same number of perforant path synapses as does the average granule cell (Amaral et al., 1990). Here, the story is quite similar to the dentate gyrus, in that while m, but not d, opioid receptor antagonists are effective in blocking LTP at lateral perforant path–CA3 synapses, NMDA receptor antagonists have no effect on LTP induction (Do et al., 2002; Kosub et al., 2005). This is in line with studies of opioid peptide receptor occupation, which show a high density of m opioid receptors in the stratum lacunosum-moleculare of the rat CA3 region (Neumaier and Chavkin, 1989). Conversely, NMDA receptor antagonists are effective in blocking LTP at medial perforant path–CA3 synapses (Do et al., 2002), whereas opioid receptor antagonists have little or no effect (Breindl et al., 1994; Do et al., 2002). In addition, stimulation of perforant path fibers often results in a depression of responses at neighboring, inactive synapses, a phenomenon termed ‘‘heterosynaptic LTD.’’ In our studies, we found that when LTP is induced in a lateral perforant pathway, the NMDA receptor antagonist CPP was ineffective in blocking lateral perforant path LTP induction,

430

CPP was effective in blocking the heterosynaptic LTD of medial perforant path–CA3 responses normally observed following lateral perforant path stimulation (Kosub et al., 2005). Thus, like their synapses in the dentate gyrus, synapses from the lateral perforant pathway to the CA3 region display LTP induction that is insensitive to NMDA receptor antagonists, but is sensitive to antagonists of the m opioid receptor. Although these results are consistent with the high density of m opioid receptors in the termination site of the lateral perforant path in area CA3, physiological data indicate that NMDA receptors are expressed in both stratum lacunosum-moleculare of area CA3 and the outer molecular layer of the dentate, although to a lesser degree than the regions where medial perforant path afferents terminate (see Monaghan and Cotman, 1985; Bramham, 1992). If LTP induction in both the dentate and the CA3 region is independent from NMDA receptors, what is the function of NMDA receptors located in their synaptic regions (Milner and Drake, 2001)? One possibility may be the induction of heterosynaptic LTD. Supporting this view, the NMDA receptors found in the molecular layer of the dentate may be extrasynaptic and found on dendrites and dendritic shafts (Monaghan and Cotman, 1985; Bramham, 1992). Because recent evidence implicates extrasynaptic NMDA receptors, possibly NMDA receptors containing the NR2B subunit, in LTD induction (Massey et al., 2004), it is possible that the principal process involving NMDA receptors at lateral perforant path targets (and likely localized to extrasynaptic sites) is the induction of various forms of perforant path LTD (see below). It is also important to note the independence of LTP at the lateral perforant path synapse from NMDA receptors is not unchallenged. Zhang and Levy (1992) showed that lateral perforant path–DG LTP could be blocked by systemic administration of ketamine, an uncompetitive NMDA receptor antagonist. A possible explanation for this effect may be the type of NMDA receptor involved in LTP induction. It is known that not all NMDA receptor heteromers (such as NMDA receptors that contain NR2B subunit) are sensitive to competitive NMDA antagonists like

CPP or AP-5 (Snyder et al., 2001). As uncompetitive antagonists generally are effective at most NMDA receptors, it remains possible that the NMDA receptors in the terminal fields of lateral perforant path projections may be a heteromeric variety (such as NR2B multimers, Massey et al., 2004) that are unaffected by typical NMDA receptor antagonists. In this case, the disinhibitory effects of opioid peptides might contribute to LTP induction via enhanced postsynaptic depolarization and a facilitation of the activation of NMDA receptor heteromers that may be insensitive to traditional competitive NMDA receptor antagonists. It remains unknown whether the LTP at medial and lateral perforant path terminals reflect distinct ‘‘forms’’ of LTP that display distinct mechanisms of both induction and maintenance, or if the LTP in the lateral and medial perforant pathways simply reflect differences in the mechanisms of inducing a common form of LTP. The issue is of interest, as there may be as many forms of LTP as there are forms of LTP induction. However, it is also quite possible that perforant path–dentate LTP may differ simply in its induction requirements, yet share common, downstream molecular processes of LTP maintenance. Nonetheless, if medial and lateral perforant path synapses display a similar form LTP that differs only in its induction mechanism, this also is of importance, as it suggests that distinct rules, or ‘‘constraints,’’ govern LTP induction among different synaptic populations of the hippocampal formation. Such difference may serve to restrict LTP in specific synaptic population to specific types of synaptic activity. For example, because opioid receptor activation is crucial for the induction of lateral perforant path LTP, and because opioid peptides display frequency-dependent release (Neumaier and Chavkin, 1989; Wagner et al., 1990, 1991; Derrick et al., 1994a, b), it would be expected that LTP at this synapse would only occur with high frequency presynaptic activity. Thus repetitive presynaptic activity, such as is observed with bursting (in this case, bursting in cells of the lateral entorhinal cortex giving rise to the lateral perforant pathway) may be absolutely required for LTP induction in these projections. Thus, the additional

431

requirement for opioid peptides, and their requirement for high frequency presynaptic activity or bursting for their release, may be viewed as a feature that imposes tighter ‘‘constraints’’ on the types of synaptic activity that are effective in inducing lateral perforant path LTP. Satisfaction of this additional constraint might only be provided by certain types of presynaptic activity (such as bursting), thereby restricting LTP induction in these synaptic populations to states, such as theta states, that are associated with bursting of granule cells (Munoz et al., 1990) or cells of the entorhinal cortex (Jeffery et al., 1995). While no data are available on whether or not medial and lateral perforant path synapses display distinct ‘‘forms’’ of LTP, it should be noted that different ‘‘forms’’ of LTP may occur even within a single synaptic population. Recent evidence suggests that different forms of LTP, reflected in differences in the longevity of LTP, and therefore LTP maintenance, can be regulated by a number of variables, such as the pattern of afferent activity used to induce LTP, or by the behavioral state of the animal during LTP induction. In the case of LTP induced with different types of stimulation, high-frequency stimulation that mimics naturallyoccurring theta rhythm is suggested to induce a ‘‘non-decremental’’ form of LTP at Schaffer–CA1 synapses (Staubli and Lynch, 1987), possibly by inducing a number of forms of LTP (Morgan and Teyler, 2001). In the case of different behavioral states, the longevity of LTP also can depend on behavioral state of the animal when LTP is induced. At perforant path–dentate synapses, LTP induced following a single stimulation session typically persists for 5–7 days (Barnes, 1979; Davis et al., 2004). However, the duration of LTP is extended to more than 2 weeks if perforant path LTP is induced while the animal is actively engaged in learning (such as during the initial exploration of a novel environment; Davis et al., 2004). Thus, the production of LTP with a sustained time course is highly suggestive of distinct mechanisms of LTP maintenance, and thus distinct ‘‘forms’’ of LTP. The mechanisms underlying these effects remain to be elucidated, although the latter effect of novelty on enhancing the longevity of perforant path–dentate LTP appears to require the presence

of specific monoamines often associated with arousal, such as dopamine, norepinephrine, and serotonin (Li et al., 2003; Straube et al., 2003; Sanberg et al., 2006). When LTP is induced either with theta bursts or by novel environments, LTP displays a distinct, extended time course. The implication of these findings is that distinct molecular mechanisms, or the modulation or modification of a common molecular mechanism, may underlie LTP maintenance when induced under different conditions. Thus, not only can different synapses display different ‘‘forms’’ of LTP, but a given synapse may also display distinct ‘‘forms’’ of LTP depending on the animal’s behavioral state during its induction. Thus, the present challenge is two-fold: to determine the necessary and sufficient mechanisms that may be shared by different ‘‘forms’’ of LTP, if they exist, and to determine the functional significance of the variations of LTP induction and maintenance. This approach has now become crucial, as differences in LTP forms or LTP induction mechanisms among the afferent populations and subregions in the hippocampal formation are likely to reveal important clues regarding the role of these distinct hippocampal afferents and hippocampal subregions in the processing and storage of information in the hippocampal formation.

Plasticity of mossy fibers It should also be mentioned that opioid peptides are also thought to play a role in LTP in efferents of the granule cells, the mossy fibers. In the first report by Harris and Cotman (1986) that LTP at the mossy fiber–CA3 synapse is insensitive to NMDA receptor antagonists, these investigators also suggested that the insensitivity of mossy fiber LTP to NMDA receptor antagonists may reflect a distinct ‘‘form’’ of LTP, although these investigators suggested that closer scrutiny of mossy fiber LTP maintenance might be important in order to verify this possibility. In addition, mossy fiber LTP induced in vivo displays an unusual, incremental, or slowly developing LTP that decay little over hours following induction, a feature not shared by commissural/ associational CA3 inputs that display NMDA

432

receptor-dependent LTP. These differences in both the mechanism of LTP induction and the incremental time course of LTP expression led us to suggest mossy fiber–CA3 LTP reflects a distinct form of LTP (Derrick and Martinez, 1989). Our subsequent studies (Derrick et al., 1991, 1992) showed that mossy fiber LTP induced in vivo requires the activation of opioid receptors, replicating an original report by Martin (1983). Subsequent studies revealed that the activation of the m opioid receptor is crucial for the induction of mossy fiber LTP (Derrick et al., 1992). Interestingly, subsequent studies revealed that induction and development of mossy fiber LTP is blocked by protein synthesis inhibitors (Barea-Rodriguez et al., 2000). As such inhibitors usually only alter the late phase (>3 h) of NMDA receptor-dependent LTP, this result offers further proof of distinct mechanisms of maintenance, and thus a distinct form of LTP. A 1991 report by Nicoll and colleagues came to a similar conclusion, but for different reasons. In this study, the induction of mossy fiber LTP was not only NMDA receptor-independent, but displayed induction mechanisms that were exclusively presynaptic (Zalutsky and Nicoll, 1990). However, the report by Zalutsky and Nicoll (1990) was in conflict with earlier reports indicating a dependence of mossy fiber LTP on both postsynaptic calcium (Williams and Johnston, 1989) and depolarization (Jaffe and Johnston, 1990). While at that time, our data favored a presynaptic locus for mossy fiber LTP induction (Derrick et al., 1992), our subsequent studies revealed an intensity-dependent threshold for mossy fiber LTP induction (Derrick and Martinez, 1994b). Moreover, associative mossy fiber LTP could be observed when weak mossy fiber synapses were co-active with intense activation of CA3 associational/commissural inputs. Importantly, this effect could be blocked by an NMDA receptor antagonist (Derrick et al., 1994b). This was a crucial finding, because NMDA antagonists, while effective in blocking LTP among commissural/ associational inputs, normally do not block mossy fiber LTP. That NMDA receptor antagonists blocked mossy fiber LTP induced associatively by coactivation of an NMDA receptor dependent pathway offers evidence that a common

postsynaptic factor that apparently is provided by NMDA receptor activation is crucial for the induction of mossy fiber LTP. What role might opioid peptides play in the induction of mossy fiber LTP? It seems likely that, as with lateral perforant path LTP, opioid receptor activation may facilitate LTP induction as a result of their disinhibitory effects. However, tacit in this view is that mossy fiber LTP induction involves postsynaptic depolarization. Although there is ample evidence that mossy fiber LTP depends on postsynaptic factors, including postsynaptic depolarization (Jaffe and Johnston, 1990; Derrick et al., 1994b), m opioid receptors may have entirely different actions at this synapse. Johnston and colleagues (Williams and Johnston, 1996) demonstrate that while naloxone blocks mossy fiber LTP in vitro, the addition of GABAA receptor antagonists does not obviate the block of mossy fiber LTP induction by naloxone. Thus, unlike the lateral perforant pathway, reducing GABAergic inhibition does not obviate the dependence of mossy fiber LTP on m opioid receptor activation. This suggests another role for m opioid receptors in mossy fiber LTP induction that is distinct from their actions on GABAergic inhibition. This same conclusion can be reached a priori from previous studies in vitro; while the mossy fiber–CA3 synapse normally does not display single-pulse associativity, single-pulse associativity can be observed at mossy fiber synapses when m opioid receptor agonists (which normally would be released with repetitive activity) are provided (Derrick et al., 1994a). However, single mossy fiber pulses delivered in conjunction with postsynaptic depolarization of CA3 pyramidal cells still fails to induce single-pulse associative mossy fiber LTP. This effect is difficult to reconcile from the perspective of opioid peptide-mediated disinhibition; if opioid receptor activation were necessary solely for disinhibitory effects (and enhanced postsynaptic depolarization), it would be expected that induced postsynaptic depolarization of CA3 pyramidal cells would obviate the need for any further depolarization afforded by the disinhibitory effects of opioid peptides. However, this appears not to be the case, leaving open the possibility that m opioid receptors may contribute

433

to mossy fiber LTP via other mechanisms. Among the possible effects of mu receptors are an enhancement of presynaptic activity mediated by reduced GABAB receptor activation (Mott and Lewis, 1992; Jin and Chavkin, 1999), or a m opioid receptor-mediated activation of second messenger cascades. For instance, while m opioid receptors typically decrease calcium influx and enhance potassium efflux in most neuronal populations, m opioid receptors can activate the MAP and ERK cascades in some preparations (Zimprich et al., 1995), and are even reported to facilitate transcription via the translocation of calmodulin to nuclear sites (Wang et al., 2000). The latter effect is of particular interest, as it appears that the slowly developing, ‘‘incremental’’ potentiation seen with mossy fiber LTP in vivo requires protein synthesis (Barea-Rodriguez et al., 2000). Nonetheless, the data suggesting exclusively presynaptic mossy fiber LTP is nonetheless compelling (Zalutsky and Nicoll, 1990; Weisskopf et al., 1993; Mellor and Nicoll, 2001; Castillo et al., 2002; Calixto et al., 2003). Computationally, if any synapse could operate effectively with a purely presynaptic form of LTP induction, particularly if its role were simply a ‘‘detonator’’ synapse, then the mossy fiber synapse would be it. It is already sparse anatomically, and the various features of dentate granule cells, in combination with competitive learning exemplified by LTP and LTD at perforant path–dentate synapses, together may provide sufficient constraints for the formation of sparse mossy fiber outputs to the CA3 region. These effects may obviate any need for coincident Hebbian mechanisms at mossy fiber synapses. Moreover, given the presumed ‘‘detonator’’ role of this synapse, mossy fiber synapses need not display ‘‘associativity;’’ rather, mossy fiber synaptic activity and its ‘‘detonator’’ properties are thought crucial for inducing associative LTP at other, coactive CA3 synapses. The view that mossy fiber activity plays a crucial role by virtue of its ‘‘detonator’’ properties and the associative induction of LTP at other, coactive CA3 synapses is supported by the temporal sequence of afferent activation of CA3 pyramids following perforant path activity. As noted above, the perforant path projections to the dentate gyrus and the CA3

region activate CA3 pyramidal cells 2–5 ms earlier than granule cells of the dentate (Fig. 2). Direct perforant path afferents are capable of discharging CA3 pyramids (which occurs in vivo with the same perforant path stimulation intensities that activate granule cells; see Do et al., 2002; Fig. 2). Thus the activation of CA3 region by perforant path would initiate recurrent CA3–CA3 activity rapidly, as the commissural/associational CA3–CA3 fibers have conduction velocities of 1.5 m/s in vivo (Do et al., 2002). By contrast, the parallel and simultaneous perforant path input to the granule cells elicits granule cell firing 1–3 ms later than the CA3 pyramidal cells. In addition, because the conduction velocity of the mossy fibers is quite slow, 0.4–0.8 m/s (Derrick et al., 1994a), the disynaptic activation of CA3 pyramidal cells by the mossy fibers would occur after both PP–CA3 and CA3–CA3 recurrent activity. This sequence of CA3 afferent activation (PP–CA3, then CA3–CA3, followed by mossy fiber–CA3 synapses) places mossy fiber–CA3 synaptic activity in the position of being the last CA3 synapse to be activated disynaptically by perforant path input. This sequence is crucial, because the induction of associative LTP at recently active synapses is asymmetric, and is observed only when presynaptic activity precedes postsynaptic depolarization (Levy and Steward, 1983; Gustafsson and Wigstrom, 1986; Gustafsson et al., 1987). This asymmetric feature, characterized more recently as asymmetric LTP mediated by spike timing-dependent plasticity (STDP), suggests that perforant path–CA3 and CA3–CA3 synapses activated prior to, or simultaneously with postsynaptic depolarization resulting from mossy fiber–CA3 ‘‘detonators’’ would induce associative LTP at the recently activated CA3 synapses. As mossy fiber ‘‘detonators’’ fire after PP–CA3 and CA3–CA3 activity, this delay in detonator activity would be optimal for the induction of associative LTP at recently activated PP–CA3 and CA3–CA3 synapses. A recent study suggests such an effect or mossy fiber activity of commissural/associational CA3 fibers, although it appears to follow rules atypical of those usually associated with STDP (Kobayashi and Poo, 2004). A primary assumption in this view of encoding in the CA3 region is that mossy fiber synaptic activity

434

can induce associative LTP at perforant path and CA3–CA3 synapses that are located in more distal dendritic regions of the CA3 pyramids. While this effect is plausible in the CA1 region, given the ability of action potentials to ‘‘backpropagate’’ to dendritic regions following CA1 pyramidal cell discharge (Magee and Johnston, 1997), the CA3 pyramids do not show a large apical dendrite that extends into the stratum radiatum, as do the CA1 pyramids. However, data obtained in vitro suggest that the mossy fibers are indeed able to induce LTP at more distal CA3–CA3 and PP–CA3 synapses (Chatterji et al., 1989; McMahon and Barrionuevo, 2002; Kobayashi and Poo, 2004). In addition, while there is evidence for the mossy fibers having ‘‘detonator’’ aspects and being able to initiate firing of CA3 pyramidal cell (Henze et al., 2002), this effect is frequency dependent. Thus repetitive activation of mossy fiber–CA3 synapses is necessary for mossy fiber synapses to exhibit their ‘‘detonator’’ effects. This requirement presents an additional problem because, as noted above, repetitive activation of the mossy fiber synapse would necessitate bursting in granule cells, and such bursting appears to be a rare event (Jung and McNaughton, 1993). However, granule cell bursting is observed in the intact animal during theta rhythm (Munoz et al., 1990). This observation suggests that granule cell bursting, essential for initiating the ‘‘detonator’’ features of the mossy fiber synapse, may be limited to periods during periods of theta rhythm. Thus while the mossy fiber synapse may display detonator properties, these properties may be limited to behavioral states that generate granule cell bursting, such as that observed during theta rhythm which occurs during exploratory behaviors. What might explain the differences in results from different laboratories regarding pre- and postsynaptic mechanisms of mossy fiber LTP? Barrionuevo and colleagues (Urban and Barrionuevo, 1996) suggested that mossy fiber LTP itself may express different forms of LTP. Perhaps, as observed with novelty-induced facilitation of LTP that occurs at medial perforant path–dentate synapses (Davis et al., 2004), this may indicate the induction of a different form of mossy fiber LTP that depends on conditions that co-occur with LTP induction. Thus, pre- or postsynaptic forms

of mossy fiber LTP may occur in distinct behavioral states, during particular patterns of afferent activity, which then may determine if pre- or postsynaptic forms of mossy fiber LTP are induced. For instance, it is possible that the postsynaptic, Hebbian form of mossy fiber LTP may be essential during encoding, whereas a strictly presynaptic, non-associative, and short-lived form of mossy fiber LTP may operate during recall. This occurrence of distinct forms of mossy fiber LTP may serve to prevent Hebbian associativity during recall states, an event that can have catastrophic effects on previously encoded patterns. This problem, often referred to as the ‘‘stability plasticity dilemma’’ (Grossberg, 1987) is a critical, but often ignored problem inherent in most neural network learning schemes (Hertz et al., 1991). Competitive learning and self-organization, as is suggested to occur in both perforant path–dentate (Hasselmo et al., 1995; Rolls and Treves, 1998) and mossy fiber–CA3 synapses (Derrick and Martinez, 1996), would, over time, result in continual changing of synaptic strength with each input, making any stable learning or categorization impossible (Hertz et al., 1991). Thus the expression of a distinct, presynaptic and non-associative form of mossy fiber LTP limited to periods of recall would limit the associative influence of mossy fiber synapses during recall. Thus a presynaptic, non-associative form of mossy fiber LTP could serve as a novel solution to the ‘‘stability–plasticity dilemma,’’ which is a problem inherent in all devices that require both encoding and retrieval (Hertz et al., 1991). Subsequent studies of the different types of mossy fiber LTP nonetheless suggest that postsynaptic factors are operative even with these apparently distinct forms of LTP induction (Yeckel et al., 1999). Furthermore, the participation of a number of factors thought to act transsynaptically at mossy fiber synapses (such as the transsynaptic activation of Eph receptors (Contractor et al., 2002; Armstrong et al., 2006) and the possible recruitment of extant, but ‘‘silent’’ mossy fiber synapses with LTP induction (Reid et al., 2004) also suggest a requirement for postsynaptic processes in many forms of mossy fiber plasticity, and with other studies, together support the view that

435

postsynaptic factors are involved in the induction of both mossy fiber LTP and mossy fiber LTD (Derrick et al., 1994b; Derrick and Martinez, 1996; Yeckel et al., 1999; Kapur et al., 2001; Contractor et al., 2002; Lei et al., 2003; Wang et al., 2004; Armstrong et al., 2006). Although the above characterization of LTP induction at the mossy fiber–CA3 synapse, as well as lateral perforant path inputs to both the dentate and CA3 regions, together suggest a mechanism of LTP induction involving opioid peptides, there is good evidence that a multiplicity of LTP and LTD forms may exist at the mossy fiber synapse (Lei et al., 2003). Perhaps distinct forms of LTP can be induced at mossy fiber synapses, and which form is induced may depend critically on the initial conditions during its induction. As we showed earlier in the dentate gyrus, LTP with an enhanced time course can be observed in vivo depending on the conditions during LTP induction (Davis et al., 2004; Sanberg et al., 2006). In this view, multiple forms of LTP are possible at a synapse, with the form of LTP that is expressed depending critically on the conditions during (and shortly after) LTP induction. Given that the mossy fiber–CA3 synapse displays the three known forms of LTP (homosynaptic, heterosynaptic and associative LTD; Derrick and Martinez, 1996), and given that a variety of cellular mechanisms (both pre- and postsynaptic) underlie what appears to be distinct forms of homosynaptic mossy fiber LTD (Manabe, 1997; Lei et al., 2003; Wang et al., 2004), the possibility that there are different forms of mossy fiber LTP remains the most parsimonious and compelling explanation for the disparate results seen with mossy fiber LTP. Further work hopefully will elaborate on the differences in the molecular machinery underlying and the maintenance and expression that are associated with the distinct induction mechanisms of mossy fiber LTP and LTD. The findings from such studies may reveal a common cascade among the different induction mechanisms, or reveal functionally distinct forms of plasticity at this synapse. The former result would make us reconsider the classification of LTP and its ‘‘forms’’ based simply upon differences in their induction mechanisms, and may also suggest that these distinct induction mechanisms may have functional significance, being

regulators of the kind of mossy fiber LTP that will be induced. Alternatively, the latter result would suggest distinct forms of LTD that also may have specialized functional roles in scaling, sparsification, depotentiation, or information storage. The evidence for both Hebbian and nonHebbian processes at the mossy fiber synapse gives rise to an important question: what would be the utility of having ‘‘detonator’’ synapses that are plastic? What function could mossy fibers have that would make Hebbian plasticity at a detonator synapse important? The dentate gyrus is thought crucial for the encoding of new information (McNaughton et al., 1989), and shows the greatest activity in humans with novelty (Zeineh et al., 2003). Yet most models of autoassociative recall do not indicate that plasticity among detonator synapses is essential (Rolls and Treves, 1998; Kali and Dayan, 2000). Moreover, behavioral studies indicate that the dentate gyrus is not essential for recollection of spatial memory or the activity of place fields (McNaughton et al., 1989; Mizumori et al., 1989), whereas others indicate the absence of mossy fiber LTP has little effect on spatial memory (Huang et al., 1995). Thus current evidence suggests that the dentate gyrus is crucial for the encoding, but not for the recall, of information. However, these behavioral finding must be interpreted carefully, as plasticity in dentate–CA3 inputs may contribute to aspects of hippocampal memory that may not be readily apparent when using simple behavioral paradigms that are used to assess only spatial memory. Given that the hippocampus is found to participate in more elaborate tasks (Agster et al., 2002;), including those that assess directly episodic-like memory (Fortin et al., 2002; Kart-Teke et al., 2006), it remains possible that the contribution of the mossy fibers to memory may become evident in more elaborate and demanding tasks. Given that the dentate gyrus itself displays place fields (Jung and McNaughton, 1993) and elaborate mechanisms of synaptic plasticity, as do the mossy fiber–CA3 outputs (Derrick and Martinez, 1996), it seems curious that elaborate plastic processes would even exist in the dentate and their mossy fiber efferents unless recall was an important aspect of mossy fiber function. The possible contribution

436

of the dentate gyrus to information storage may best be understood from a theoretical view and the presumed role of the direct perforant path inputs to CA3 to initiate recall (Treves and Rolls, 1992, 1994). However, such a view may be problematic, because activation of a subset of the direct perforant path–CA3 lines, which are thought to provide the partial input that initiates pattern completion in the CA3 region (Treves and Rolls, 1994), might fail to select accurately among the many potential CA3 attractors if they happen to share a critical subset of direct perforant path–CA3 inputs. Perhaps plasticity at perforant path–CA3, together with mossy fiber LTP may cooperate during recall, particularly recollection of more elaborate aspects of hippocampal memory not readily detectable by current behavioral methods that only assess spatial learning. In this view, perforant path–CA3 and dentate–CA3 inputs operate together during recall to select among specific attractors within the CA3 autoassociative system, with plasticity in both the direct perforant path–CA3 and mossy fiber–CA3 inputs acting as a simple pattern associator that is embedded within the CA3 autoassociative system. In this way, both mossy fiber and perforant path inputs could together select a smaller subset of CA3 pyramidal cells. Thus the additional input from modified mossy fiber–CA3 synapses thus might allow for a partial input arising from direct perforant path–CA3 synapses to select among potential CA3 attractors. Such a feature would allow for the selection of CA3 attractors that may share high degrees of similarity among their perforant path–CA3 inputs. In this view, plasticity at mossy fiber ‘‘detonators,’’ and subsequent mossy fiber activity during recall, may be crucial for the selective recall of specific patterns or sequences of events that may share similar features. Implicit in this view is that mossy fiber plasticity, and the dentate gyrus in general, may be crucial for both the formation and recall of discrete episodic-like memories.

LTD LTP is noteworthy in that its induction follows the rule of pre- and postsynaptic associativity as formalized by Hebb (1949). However, a mechanism

serving to increase synaptic strength cannot operate alone; otherwise, the strength of synapses could only increase, eventually reaching a point of saturation. Other mechanisms that permit either the reversal or the inverse of LTP are likely to be necessary. Such a phenomenon, termed LTD, is observed at the same synapses that display LTP. LTD was noted in early studies, although its possible role in information storage was only suggested (Barrionuevo et al., 1980). As it became apparent that any memory device that serves as a temporary repository for information must have some means to both increase and decrease synaptic strength, the mechanisms underlying LTD became a primary focus for many studies in the 1990s (Bear and Abraham, 1996). In contrast to LTP, individual ‘‘forms’’ of LTD in the dentate, as evidenced by their distinct mechanisms of induction, were noted early. Homosynaptic LTD, the term used to describe LTD that follows synaptic activity, is typically input specific and induced experimentally in many hippocampal synapses by repetitive low frequency (0.5–5 Hz) stimulation (Huang et al., 1992). At most dentate synapses, homosynaptic LTD, like LTP, is cooperative (Kerr and Abraham, 1995), associative (Christie and Abraham, 1992a), and also requires postsynaptic calcium, although the role of NMDA receptors is controversial; although some studies indicate that homosynaptic LTD is NMDA receptor-dependent (Thiels et al., 1996), other studies indicate that both homosynaptic LTD and depotentiation appear to involve metabotropic glutamate receptors (O’Mara et al., 1995a; ManahanVaughan, 1998; Kulla et al., 1999; Wu et al., 2004) and calcium influx from intracellular stores, rather than NMDA receptors (O’Mara et al., 1995b). For afferents to the dentate gyrus, studying homosynaptic LTD is problematic, as the induction requirements vary among laboratories, and it is notoriously difficult to elicit, both in vivo and in vitro in the slice preparation (Errington et al., 1995; Abraham, 1996; Abraham et al., 1996), although some laboratories have found that particular paradigms aid in its induction in vitro (Abraham, 1996) and in vivo (Manahan-Vaughan, 1998). Not only does this characteristic make LTD

437

difficult to study, but it also illustrates an important limitation of the in vitro preparation, as alterations in dentate function may be an inevitable consequence of slice preparation. The viability of studying plastic processes in dentate slices is not too surprising when one considers that preparation of hippocampal slices involves widespread deafferentation, activity, and such trauma can cause sustained plastic changes in dentate function. For example, preparation of hippocampal slices induce a number of IEGs such as c-fos (Dragunow et al., 1989). As IEGs often show a refractory period persisting for hours after their induction (Sheng and Greenberg, 1990), the dentate slice preparation may be of limited utility in dissecting the molecular events that mediate the variety of molecular events that underlie the various forms of plasticity in the dentate gyrus. LTD can be elicited at synapses adjacent to those that are active or potentiated. This form of LTD is referred to as ‘‘heterosynaptic’’ and is observed at synapses that were not activated by the stimulus used to induce potentiation. Heterosynaptic LTD is robust at the perforant path–dentate projections following the induction of LTP in one set of afferents (such as the medial perforant path), which then induces heterosynaptic LTD of responses at inactive synapses (in this case, either unstimulated medial perforant path synapses, or lateral perforant path synapses), and vice versa. Here, LTD induction appears to involve to both NMDA receptors (Desmond et al., 1991) and the L-type voltage-dependent calcium channels (VDCCs; Wickens and Abraham, 1991; Christie and Abraham, 1994; Camodeca et al., 1998), suggesting that the low levels of calcium necessary for heterosynaptic LTD may be provided by VDCCs activated in response to NMDA receptors, perhaps via distinct NMDA receptors at distinct, extrasynaptic locations (such as the NR2B subunits of the NMDA receptor, or similar receptors on dendrites or the spine base and neck; Massey et al., 2004). The third type of LTD, associative LTD, was once regarded as a less than replicable phenomenon (Bear and Malenka, 1994); however, studies showing homosynaptic LTD displays cooperativity (Kerr and Abraham, 1995) suggest

homosynaptic LTD in the dentate can be induced in an associative manner. However, associative LTD appears to involve cellular mechanisms of expression that are similar to heterosynaptic LTD, although their mechanisms of induction differ in that associative LTD in the dentate is reportedly NMDA receptor-independent (Christie et al., 1995; Abraham and Tate, 1997; Philpot et al., 1999). Depotentiation is related to LTD, but refers to a reversal of LTP (Levy and Steward, 1979; Barrionuevo et al., 1980; Fujii et al., 1991) and may be thought of as homosynaptic LTD of potentiated synapses. However, depotentiation is distinct from homosynaptic LTP in two important ways. First, depotentiation has a narrow window of induction; in the dentate, stimulation must occur minutes following LTP induction, otherwise LTP becomes resistant to depotentiation (Martin, 1998; Kulla et al., 1999). Second, depotentiation and homosynaptic LTD appear to involve distinct cellular mechanisms (Lee et al., 2000; Soderling and Derkach, 2000). The early (o3 h) (Otani et al., 1989) phase of LTP is suggested to involve phosphorylation of the GluR1 AMPA subunit at both CaMKII and PKA sites on this receptor subunit (Soderling and Derkach, 2000). Studies suggest that whereas depotentiation involves dephosphorylation at PKA serine/threonine residues, homosynaptic LTD at naive synapses may involve dephosphorylation of CaMKII serine/threonine residues (Lee et al., 2000). Just as the mechanisms underlying the induction of heterosynaptic LTD, homosynaptic LTD, and depotentiation appear distinct, so are the treatments that can affect their induction. For example, 5-HT4 agonists (Kulla and Manahan-Vaughan, 2002) and D1/D5 dopamine antagonists block depotentiation, but not LTP or heterosynaptic LTD (Otmakhova and Lisman, 1998; Kulla and Manahan-Vaughan, 2000). Conversely, drugs, such as nimodipine, an L-type calcium channel blocker, can block heterosynaptic LTD in the dentate gyrus (Wickens and Abraham, 1991) without altering LTP (Christie and Abraham, 1994). The number of studies of LTD in the dentate gyrus are limited, however, as LTD in the dentate is not easily induced in the slice preparation (Bear and Abraham,

438

1996), and may indicate the importance of extrinsic neuromodulators in the induction of LTD in the dentate gyrus (Pang et al., 1993). The diversity of the mechanisms that induce LTD and depotentiation suggest that different types of LTD may serve distinct functions (Kerr and Abraham, 1996). As noted earlier, mechanisms that serve to increase synaptic strength would be expected to be accompanied by other mechanisms that can reverse increases in synaptic strength; otherwise, interference would follow from an eventual saturation of synaptic plasticity. Thus LTD may play a role in reversing LTP (‘‘depotentiation’’). In line with this view, heterosynaptic LTD can reverse established dentate LTP (Doyere et al., 1997) a process likely to be essential in structures that have a transient role in information storage, such as the hippocampal formation (Bontempi et al., 1999). Normalization processes are thought to be a crucial aspect of neuronal homeostasis — maintaining neuronal excitability within a dynamic range of neuronal output, which is essential to maintain maximal variance in neuronal output. Thus LTP induced in a set of synapses on a neuron may result in a concomitant, equivalent net decrease in the strength of other, inactive, synapses on the same neuron. Other forms of LTD may play a role in sparse coding (Skaggs and McNaughton, 1992), ensuring that only the most active synapses increase in strength in response to a given input, whereas other less active synapses are depressed, preserving the sparse encoding essential for distributed memory systems that employ the Hebb rule (Skaggs and McNaughton, 1992). For example, both homosynaptic LTD and associative LTD may facilitate sparse encoding by depressing synapses that show ill-timed or only modest activity. In this view, these forms of LTD may filter out background synaptic ‘‘noise,’’ reducing and optimizing the number of modifiable elements that contribute to a given representation. Thus LTD may contribute to information storage by removing redundancies, and providing a sparse inverted code, or ‘‘transform,’’ of the same perforant path input that is relayed directly to distal dendrites of CA3 neurons (Morris and McNaughton, 1987). This feature may be crucial during encoding in order to

provide sparse, orthogonal inputs to the CA3 region (Rolls and Treves, 1998). Another interpretation is that LTD, operating in conjunction with LTP, together serves in information storage (Bear and Abraham, 1996; Bear, 1999; Braunewell and Manahan-Vaughan, 2001). In this way, LTP and LTD may act in concert, enhancing the dynamic range of plasticity at a given synapse, as bidirectional plasticity can enhance the capacity, fidelity, and flexibility of the networks (Dayan and Willshaw, 1991). Bidirectional plasticity observed at the perforant path–dentate synapse also is thought to be crucial in competitive learning schemes (Rolls and Treves, 1998). Thus, while competitive learning is often implicated in self-organization (Hertz et al., 1991), it also can be viewed simply as another way to remove redundancies and to sparsify dentate outputs to the CA3 region (Rolls and Treves, 1998). In this view, heterosynaptic LTD and homosynaptic LTD have the potential to contribute to normalization and sparse encoding (Morris, 1989).

Metaplasticity Metaplasticity refers to an alteration in the threshold or magnitude of LTP or LTD induction in a neuron as a result of prior neuronal or synaptic activity (see Abraham and Tate, 1997; Philpot et al., 1999). Metaplastic processes are thought to regulate subsequent changes in synaptic strength depending on the prior neuronal ‘history’ of synaptic activity. Thus, metaplastic effects are used to describe activity-dependent and sustained effects that can regulate the ability of a synapse, or synapses on a given neuron, to change synaptic strength in response to subsequent synaptic activity. As such, metaplasticity reflects a distinct mechanism of information storage (Philpot et al., 1999). Metaplastic effects follow predictions of several theories of information processing that account for activity-dependent self-organization during development. In particular, the Bienenstock, Cooper and Munroe (BCM) theory (Bienenstock et al., 1982; Bear et al., 1987; Shouval et al., 1997), originally used to account for visual cortex development,

439

employs processes that regulate synaptic strength within an entire neuron as a consequence of prior synaptic activity, and serves as a useful framework with which to predict metaplastic effects. Generally, the BCM theory predicts that if postsynaptic activity is insufficient for LTP induction, LTD will occur. This transition point where postsynaptic activity induces LTD or LTP, in the BCM algorithm, is denoted by ym (Fig. 4). Importantly, ym is not fixed; instead, it changes as a function of postsynaptic activity. For example, activity that is not sufficient for LTP induction not only induces LTD, but also ‘‘shifts’’ the threshold for LTP induction (indicated by ym) to the left, so that the threshold of postsynaptic activity necessary to induce LTP at synapses on that neuron is subsequently reduced. Conversely, if stimulation is sufficient to induce LTP, ym shifts to the right, increasing the threshold for inducing LTP, and favoring the induction of LTD with subsequent synaptic activity. An important aspect of changes in ym is that this process, as envisaged by Bienenstock et al. (1982), is ‘‘cell wide,’’ and any change in the threshold for synaptic plasticity that follows activity would affect all synapses of a neuron. Such a feature can serve a normalization, or scaling, function by maintaining a constant net excitatory input to a given neuron so that output remains within the neuron’s maximal dynamic range (Bienenstock et al., 1982; Barrionuevo and Brown, 1983). It is important to note that metaplastic effects are distinct from normalization or synaptic scaling, as metaplasticity refers specifically to the modification of thresholds necessary to induce LTP or LTD, whereas scaling reflects modifications in synaptic drive. When the rules of ‘‘cell wide’’ metaplasticity that alter LTP and LTD thresholds are employed in models of visual cortex, these rules can allow for self-organization, categorization, and a selective ‘tuning’ of specific cells within a network to specific inputs (Shouval et al., 1997). These predictions have been confirmed to some degree in the developing visual cortex, where rats reared in dark environments showed a greater propensity to display LTP than LTD. Thus, reduced input to the visual cortex favored subsequent induction of LTP

Fig. 4. The BCM rule and sliding modifications of LTP and LTP induction thresholds. (A) The induction of LTP or LTD depends on the level of postsynaptic activity. Postsynaptic depolarization (PSD) at or beyond the threshold for LTP induction (ym) results in the induction of LTP. (B) Prior neuronal activity sufficient for LTP induction shifts ym to the right, making subsequent postsynaptic activity more likely to induce LTD. (C) Prior activity insufficient to induce LTP shifts ym to the left, lowering the threshold for LTP induction. Note that a fixed level of postsynaptic activity that is initially the threshold for LTP induction (——) in (A) subsequently induces LTD (B) or LTP (C) depending on prior activity and the resultant shift in ym.

(shifting ym to the left). However, exposure of these rats to light for various periods of time resulted in a shift of LTD/LTP thresholds to values similar to controls, confirming activity-based alterations in LTP and LTD threshold in the visual cortex. This simple but powerful local learning rule, which appears operative in the visual cortex, can allow for emergent properties in neuronal networks, including self-organization and stimulus selectivity (Shouval et al., 1997). Within the hippocampus, metaplastic phenomena also are observed, as previous studies indicate

440

that low-frequency stimulation that is ineffective in inducing either LTP or LTD (‘priming’ stimulation) can subsequently lead to a cell-wide decrease (Holland and Wagner, 1998) in LTD threshold and enhance stimulation-induced reversal of LTP (or depotentiation). However, most hippocampal studies in vitro that utilize priming stimulation also report input-specific effects limited to stimulated synapses (Huang et al., 1992). Thus, the validity of the BCM theory is complicated by priming studies that report metaplastic effects that are not ‘‘cell wide,’’ but rather, are limited to activated synapses (‘‘input specific metaplasticity’’). Within the dentate gyrus, input specific effects are also observed with priming stimulation. Christie and Abraham (1992a) report that priming the lateral perforant path with low frequency (5 Hz) theta burst stimulation resulted in a facilitation of the induction of LTD. At first glance, these results seem to be in opposition to the BCM rule (as stimulation that does not evoke LTP should slide the LTD/LTP threshold to the left, favoring LTP). However, it should be noted that priming stimulation itself did not induce a change in synaptic strength. Thus these findings are consonant with the BCM model, as it would be expected that low frequency stimulation that was not effective in inducing either LTP or LTD may still result in a leftward shift in thresholds. Thus, both the threshold for LTD and LTP would decrease. However, in both studies of CA1 and the dentate, the threshold for LTP induction was increased (Fujii et al., 1991; Wexler and Stanton, 1993), suggesting that the threshold for LTP made a rightward, rather than a leftward shift. Subsequent studies by the same investigators found that modest lateral perforant pathway stimulation resulted in a facilitation of subsequent LTP induction (Christie et al., 1995). This still is problematic, as stronger stimulation would be expected to shift the BCM curve to the right, rather than the left, and increase the threshold for LTP induction, favoring LTD induction. While these findings suggest metaplastic effects governed by priming do not always follow the patter predicted by the BCM algorithm, or its cell wide functions, it seems likely that the use of priming

stimulation may underlie these discrepancies. By contrast, when stimulation is sufficient to induce LTP, subsequent perforant path–dentate potentiation with further perforant path stimulation is reduced (Frey et al., 1995). In addition, stimulation of lateral perforant path–dentate inputs that induces LTP in behaving animals reveals that subsequent stimulation at the same intensity induces LTP of a smaller magnitude for several weeks after initial LTP induction. Conversely, when LTD follow lateral perforant path stimulation, stimulation 1–3 weeks later at this same intensity results in a robust LTP (Villarreal and Derrick, 2000). Perhaps synaptic activation that does not induce plasticity, as observed with priming, may lead to a modification of metaplastic rules such that patterned synaptic activity that fails to induce LTP will always favor LTD (Christie and Abraham, 1992a), whereas low frequency activity that is random or not patterned might decrease the threshold for both LTD and LTP induction (Christie et al., 1995). Recent studies suggest that the dentate gyrus also can display ‘‘cell wide’’ metaplastic effects. This is observed following non-synaptic (antidromic) activation of dentate granule cells in vivo by hilar stimulation (Abraham et al., 2001). Perhaps the level of postsynaptic depolarization is the crucial factor in setting LTP and LTD thresholds, an effect that may not be effective when priming stimulation is used. Given that there is now evidence for both ‘‘input specific’’ and ‘‘cell-wide’’ metaplasticity in the dentate gyrus, how might the two forms of metaplasticity be important? Perhaps ‘‘cell wide’’ and ‘‘input specific’’ forms of metaplasticity serve functionally distinct roles at neuronal and synaptic levels, respectively. For example, input specific metaplasticity may operate during the initial stages of synaptic plasticity to ensure sparse encoding. In this scheme, synapses that have been recently potentiated may be refractory to further potentiation, while those neurons receiving postsynaptic input insufficient to induce LTP would be particularly sensitive to LTP induction with subsequent synaptic activity. Such an effect would ensure that recently potentiated synapses are ‘‘opted out’’ of potentiation by later inputs, further aiding in

441

dentate pattern separation. By contrast, cell-wide metaplastic effects that follow BCM rules may require synaptic plasticity (either LTD or LTP). Thus a BCM-like process that occurs with synaptic changes can allow for long-term metaplastic effects as predicted by the BCM theory, and allow for sustained stimulus selectivity, a feature that would be particularly advantageous in a competitive learning device that performs categorization functions. Interestingly, many priming effects that appear to follow BCM rules are observed primarily in the lateral perforant pathway and the mossy fiber pathway, pathways that contain and release proenkephalin-derived opioid peptides. However, no studies have addressed metaplasticity in medial perforant path afferents, or the contribution of opioid peptides to the induction or expression of metaplastic phenomena in opioidergic lateral perforant path afferents (Christie and Abraham, 1992a; Christie et al., 1995; Francesconi et al., 1997). The current evidence suggests it is a likely role for opioid peptides, given that the disinhibitory effects of opioid peptides can greatly influence the levels of postsynaptic activity, which, in turn, can influence NMDA receptor activation, as well as ym, and thus permit the differential induction of either LTP or LTD (Francesconi et al., 1997; Wagner et al., 2001). The mechanisms underlying metaplastic effects remain unknown, although alterations in glutamate receptors (Abraham and Tate, 1997; Castellani et al., 2001) and activation of group II (ManahanVaughan, 1998; Gisabella et al., 2003) and Group I (Wu et al., 2004) metabotropic glutamate receptors are suggested to mediate metaplastic effects, and the induction of LTD and depotentiation in general (Manahan-Vaughan, 1998). Recent theories also implicate alterations in NMDA receptors and their subunit composition as a likely means of regulating postsynaptic responses to synaptic activity (Castellani et al., 2001). Several studies suggest common mechanisms of metaplasticity among lateral and medial inputs to the dentate (Abraham et al., 2001) that may involve alterations in NMDA receptors. As NR2B NMDA receptors show a decreased calcium conductance once bound to calmodulin

(Rycroft and Gibb, 2002), and the duration of metaplastic effects in vivo (Villarreal and Derrick, 2000) parallels LTP longevity (Davis et al., 2004), alterations in the multimeric composition of NMDA receptors that involve the NR2B subunit of the NMDA receptor may reflect an important mechanism underlying metaplasticity. Subsequent down stream mechanisms are suggested to involve CaMKII (Zhang and Levy, 1992) and protein kinase C (Gisabella et al., 2003). Given that the primary function of the dentate is thought to be pattern separation and the generation of a sparse, inverted code that allows for orthogonal outputs to the CA3 region, input specific metaplasticity may play a role in the pattern separation processes of the dentate. For example, cells that were recently active and potentiated will display a rightward shift in ym, and thus an increase in LTP induction threshold, making these potentiated synapses refractory to any further induction of LTP, and even favor the induction of LTD with subsequent synaptic activity. By contrast, synapses that are activated to a degree that is insufficient to induce LTP would show a leftward shift in y, greatly reducing LTP induction threshold for subsequent LTP induction. Therefore, synapses that have been recently activated or modified may thus be ‘‘opted out’’ and resistant LTP induction or to any further increase in potentiation. Thus, analogous to competitive learning, input-specific forms of metaplasticity observed in perforant path–dentate synapses may also contribute to the ‘‘pattern separation’’ functions of the dentate gyrus, favoring LTP induction at synapses activated below LTP threshold, allowing for formation of sparse, non-overlapping patterns of granule cell activity, a function that is likely crucial for encoding information in the CA3 autoassociative system.

Conclusions The dentate gyrus displays plastic processes that may underlie learning and memory in the hippocampus. However, the same plastic processes that are thought to contribute to information storage also might serve to ensure the formation of sparse,

442

orthogonal transforms that are relayed to the CA3 region that can promote the encoding of discrete CA3 attractors, because these same plastic processes also can aid in the formation of sparse, orthogonal outputs to the CA3 region. First, both the anatomically sparse mossy fiber projections and the ‘‘input expansion’’ seen in the relay from the entorhinal cortex to the dentate gyrus are anatomical features that can facilitate dentate pattern separation. In addition, both competitive learning (as exemplified by LTP and LTD) and metaplasticity, often implicated in self-organization, also may contribute in ensuring sparse orthogonal dentate outputs relayed to the CA3 region, principally by removing redundancies from dentate outputs and by limiting synaptic plasticity at previously potentiated synapses, and favoring LTP only at synapses that have not been potentiated previously. Moreover, the requirement of opioid peptides for plasticity and their requirement for specific types of presynaptic activity for their release may constrain conditions during which plasticity can occur, further contributing to sparse encoding. Together, these mechanisms may insure that the dentate gyrus relays sparse, orthogonal transforms of the same perforant path input arriving distally on CA3 pyramids and aiding in the encoding of discrete attractors within the CA3 autoassociative matrix. Recent evidence suggests that the dentate gyrus, while essential for encoding, is not necessary for accurate recall (Mizumori et al., 1989; Lee and Kesner, 2004; Jerman et al., 2006). Why would there be synaptic plasticity in both afferents and efferents of granule cells, and elaborate mechanisms regulating it if the output is only crucial for encoding? One interpretation is that the ‘‘plastic’’ processes of LTP, LTD, and metaplasticity may operate primarily to enforce pattern separation by the dentate gyrus and sparse encoding in the CA3 system. Thus, the plastic processes in the dentate, rather than participating in memory and information storage per se, actually may serve primarily to preserve pattern separation functions of the dentate. These plastic processes could then provide mechanisms that allow for sustained pattern separation of dentate input patterns relayed to,

and directing storage in, the CA3 autoassociative system over time. In this view, the sustained plastic processes displayed by the dentate gyrus, rather than subserving ‘‘information storage’’ in a formal sense, maintains a ‘‘memory’’ for past inputs in order to maintain a continued generation of sparse, orthogonal patterns that are relayed to the CA3 region, where the actual information is both stored and recalled. Alternatively, bidirectional plasticity and metaplasticity in the dentate may play a more direct role in memory storage, rather than simply serving to maintain separation among patterns relayed and stored in the CA3 system. The dentate gyrus itself displays place fields (Jung and McNaughton, 1993) and elaborate mechanisms of long-lasting synaptic potentiation and depression, as does the mossy fiber–CA3 synapse (Derrick and Martinez, 1996). Thus it seems odd that such elaborate plastic processes would exist unless recall also was an important aspect of dentate–mossy fiber function. Perhaps the plasticity observed in both perforant path inputs and mossy fiber outputs of the dentate are crucial for both encoding and recall. In this case, sustained changes at both perforant path–dentate and dentate–CA3 synapses, in conjunction with activity initiated by direct perforant path–CA3 inputs, may allow for a more ‘‘fine tuned’’ selection of attractors in the CA3 region. This is because partial inputs arriving distally at perforant path–CA3 inputs and subsequent pattern completion have the possibility of converging on any number of possible CA3 attractors that share these same perforant path–CA3 inputs. Thus, sustained perforant path–dentate and mossy fiber–CA3 plasticity might contribute to recall, with the mossy fiber inputs acting as an additional ‘‘cue’’ that allows for the selective activation of CA3 attractors among many that may share common features. In this view, plasticity in the dentate gyrus and its CA3 outputs are critical for accurate recall among a number of potential CA3 attractor states. In this case, it would be predicted that damage to the dentate gyrus, or altering mechanisms of LTP at either perforant path–dentate of mossy fiber–CA3 synapses, would display behavioral deficits, but primarily in tasks that require

443

finer discriminations, such as disambiguation of encoded events that have common features (such as context). Thus the dentate gyrus, and plastic processes observed in the dentate CA3 relay, may be crucial in the formation of similar, but discrete events over time, a feature that may be crucial for accurate encoding and recollection of episodic memory.

References Abraham, W.C. (1996) Induction of heterosynaptic and homosynaptic LTD in hippocampal subregions in vivo. J. Physiol. Paris, 90: 305–306. Abraham, W.C., Mason-Parker, S.E., Bear, M.F., Webb, S. and Tate, W.P. (2001) Heterosynaptic metaplasticity in the hippocampus in vivo: a BCM-like modifiable threshold for LTP. Proc. Natl. Acad. Sci., 98(19): 10924–10929. Abraham, W.C., Mason-Parker, S.E. and Logan, B. (1996) Low-frequency stimulation does not readily cause long-term depression or depotentiation in the dentate gyrus of awake rats. Brain Res., 722(1–2): 217–221. Abraham, W.C. and Tate, W.P. (1997) Metaplasticity: a new vista across the field of synaptic plasticity. Prog. Brain Res., 52(4): 303–323. Agster, K.L., Fortin, N.J. and Eichenbaum, H. (2002) The hippocampus and disambiguation of overlapping sequences. J. Neurosci., 22(13): 5760–5768. Altman, J. and Das, G.D. (1965) Autoradiological and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol., 124: 319–335. Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182(4 Pt 2): 851–914. Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons, numbers and the hippocampal network. Prog. Brain Res., 83: 1–11. Amaral, D.G. and Witter, M.P. (1989) The three-di mensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31(3): 571–591. Amaral, D.G. and Witter, M.P. (1995) Hippocampal formation. In: Paxinos G. (Ed.), The Rat Nervous System (4th ed.). Academic Press, San Diego, CA, pp. 443–493. Andersen, P., Holmqvist, B. and Voorhoeve, P.E. (1966) Excitatory synapses on hippocampal apical dendrites activated by entorhinal stimulation. Acta Physiol. Scand., 66: 461–472. Armstrong, J.N., Saganich, M.J., Xu, N.J., Henkemeyer, M., Heinemann, S.F. and Contractor, A. (2006) B-ephrin reverse signaling is required for NMDA-independent long-term

potentiation of mossy fibers in the hippocampus. J. Neurosci., 26(13): 3474–3481. Barea-Rodriguez, E.J., Rivera, D.T., Jaffe, D.B. and Martinez Jr., J.L. (2000) Protein synthesis inhibition blocks the induction of mossy fiber long-term potentiation in vivo. J. Neurosci., 20(22): 8528–8532. Barnes, C.A. (1979) Memory deficits associated with senescence: a neurophysiological and behavioral study in the rat. J. Comp. Physiol. Psychol., 93(1): 74–104. Barrionuevo, G. and Brown, T.H. (1983) Associative long-term potentiation in hippocampal slices. Proc. Natl. Acad. Sci., 80: 7347–7351. Barrionuevo, G., Schottler, F. and Lynch, G. (1980) The effects of repetitive low frequency stimulation on control and ‘‘potentiated’’ synaptic responses in the hippocampus. Life Sci., 27(24): 2385–2391. Bayer, S.A. (1982) Changes in the total number of dentate granule cells in juvenile and adult rats: a correlated volumetric and thymidine study. Exp. Brain Res., 46: 315–323. Bear, M. and Abraham, W.C. (1996) Long-term depression in hippocampus. Annu. Rev. Neurosci., 19: 437–462. Bear, M.F. (1999) Homosynaptic long-term depression: a mechanism for memory? Proc. Natl. Acad. Sci. U.S.A., 96(17): 9457–9458. Bear, M.F., Cooper, L.N. and Ebner, F.F. (1987) A physiological basis for a theory of synapse modification. Science, 237(4810): 42–48. Bear, M.F. and Malenka, R.C. (1994) Synaptic plasticity: LTP and LTD. Curr. Opin. Neurobiol., 4(3): 389–399. Ben-Ari, Y. and Represa, A. (1990) Brief seizure episodes induce long-term potentiation and mossy fibre sprou ting in the hippocampus. Trends Neurosci., 13(8): 312–318. Berzhanskaya, J., Urban, N. and Barrionuevo, G. (1998) Electrophysiological and pharmacological characterization of the direct perforant path input to hippocampal area CA3. J. Neurophysiol., 79(4): 2111–2118. Bienenstock, E.L., Cooper, L.N. and Munro, P.W. (1982) Theory for the development of neuron selectivity: orientation specificity and binocular interaction in visual cortex. J. Neurosci., 2(1): 32–48. Bliss, T. and Gardner-Medwin, A. (1973) Long lasting potentiation of synaptic-transmission in the dentate area of the unanesthetized rabbit following stimulation of the perforant path. J. Physiol., 232(2): 357–374. Bliss, T.V.P. and Lomo, T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J. Physiol., 232: 331–356. Bontempi, B., Laurent-Demir, C., Destrade, C. and Jaffard, R. (1999) Time-dependent reorganization of brain circuitry underlying long-term memory storage. Nature, 400(6745): 671–675. Boss, B.D., Peterson, G.M. and Cowan, W.M. (1985) On the number of neurons in the dentate gyrus of the rat. Brain Res., 338: 144–150.

444 Bramham, C.R. (1992) Opioid receptor-dependent longterm potentiation: peptidergic regulation of synaptic plasticity in the hippocampus. Neurochem. Int., 20(4): 441–455. Bramham, C.R., Errington, M.L. and Bliss, T.V. (1988) Naloxone blocks the induction of long-term potentia tion in the lateral but not in the medial perforant pathway in the anesthetized rat. Brain Res., 449(1–2): 352–356. Bramham, C.R., Milgram, N.W. and Srebro, B. (1991a) Delta opioid receptor activation is required to induce LTP of synaptic transmission in the lateral perforant path in vivo. Brain Res., 567(1): 42–50. Bramham, C.R., Milgram, N.W. and Srebro, B. (1991b) Activation of AP5-sensitive NMDA receptors is not required to induce LTP of synaptic transmission in the lateral perforant path. Eur. J. Neurosci., 3: 1300–1308. Bramham, C.R. and Sarvey, J.M. (1996) Endogenous activation of mu and delta-1 opioid receptors is required for longterm potentiation induction in the lateral perforant path: dependence on GABAergic inhibition. J. Neurosci., 16: 8123–8131. Bramham, C.R. and Srebro, B. (1989) Synaptic plasticity in the hippocampus is modulated by behavioral state. Brain Res., 493(1): 74–86. Braunewell, K.H. and Manahan-Vaughan, D. (2001) Longterm depression: a cellular basis for learning? Rev. Neurosci., 12(2): 121–140. Breindl, A., Derrick, B.E., Rodriguez, S.B. and Martinez Jr., J.L. (1994) Opioid receptor-dependent long-term potentiation at the lateral perforant path-CA3 synapse in rat hippocampus. Brain Res. Bull., 33(1): 17–24. Buzsaki, G. (1988) Polysynaptic long-term potentiation: a physiological role of the perforant path-CA3/CA1 pyramidal cell synapse. Brain Res., 455: 192–195. Buzsaki, G. (2002) Theta oscillations in the hippocampus. Neuron, 33: 325–340. Cain, D.P. and Corcoran, M.E. (1985) Epileptiform effects of met-enkephalin, beta-endorphin and morphine: kindling of generalized seizures and potentiation of epileptiform effects by handling. Brain Res., 338(2): 327–336. Calixto, E., Thiels, E., Klann, E. and Barrionuevo, G. (2003) Early maintenance of hippocampal mossy fiber — long-term potentiation depends on protein and RNA synthesis and presynaptic granule cell integrity. J. Neurosci., 23(12): 4842–4849. Camodeca, N., Rowan, M.J. and Anwyl, R. (1998) Induction of LTD by increasing extracellular Ca2+ from a low level in the dentate gyrus in vitro. Neurosci. Lett., 255(1): 53–56. Cartwright, R.D. (2004) The role of sleep in changing our minds: a psychologist’s discussion of papers on memory reactivation and consolidation in sleep. Learn. Mem., 11(6): 660–663. Castellani, G.C., Quinlan, E.M., Cooper, L.N. and Shouval, H.Z. (2001) A biophysical model of bidirectional synaptic

plasticity: dependence on AMPA and NMDA receptors. Proc. Natl. Acad. Sci., 98(22): 12772–12777. Castillo, P.E., Schoch, S., Schmitz, F., Sudhof, T.C. and Malenka, RC. (2002) RIM1alpha is required for presynaptic long-term potentiation. Nature, 415(6869): 327–330. Cavazos, J.E., Golarai, G. and Sutula, T.P. (1991) Mossy fiber synaptic reorganization induced by kindling: time course of development, progression, and permanence. J. Neurosci., 11(9): 2795–2803. Chatterji, S., Stanton, P.K. and Sejnowski, T.J. (1989) Commissural synapses, but not mossy fiber synapses, in hippocampal field CA3 exhibit associative long-term potentiation. Brain Res., 495: 145–150. Chavkin, C., Shoemaker, W.J., McGinty, J.F., Bayon, A. and Bloom, F.E. (1985) Characterization of pro-dynorphin and proenkephalin neuropeptide systems in the rat hippocampus. J. Neurosci., 5: 808–816. Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P., Hoffman, K.L., Marriot, L.K., Worley, P.F., McNaughton, B.L. and Barnes, C.A. (2005) Sparse, environmentallyselective expression of Arc RNA in the upper blade of the rodent fascia dentate. Hippocampus, 15(5): 579–586. Christie, B.R. and Abraham, W.C. (1992a) Priming of associative LTD by theta frequency synaptic activity. Neuron, 8: 79–84. Christie, B.R. and Abraham, W.C. (1992b) NMDA-dependent heterosynaptic long-term depression in the dentate gyrus of anaesthetized rats. Synapse, 10(1): 1–6. Christie, B.R. and Abraham, W.C. (1994) L-type voltagesensitive calcium channel antagonists block heterosynaptic long-term depression in the dentate gyrus of anaesthetized rats. Neurosci. Lett., 167(1-2): 41–45. Christie, B.R., Stellwagen, D. and Abraham, W.C. (1995) Evidence for common expression mechanisms underlying heterosynaptic and associative long-term depression in the dentate gyrus. J. Neurophysiol., 74(3): 1244–1247. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus. J. Comp. Neurol., 246(4): 435–458. Claiborne, B.J., Xiang, Z. and Brown, T.H. (1993) Hippocampal circuitry complicates analysis of long-term poten tiation in mossy fiber synapses. Hippocampus, 3(2): 115–121. Cohen, G.A., Doze, V.A. and Madison, D.V. (1992) Opioid inhibition of GABA release from presynaptic terminals of rat hippocampal interneurons. Neuron, 9: 325–335. Colino, A. and Malenka, R.C. (1993) Mechanisms underlying the induction of long-term potentiation in rat medial and lateral perforant paths in vitro. J. Neurophysiol., 69: 1150–1159. Colley, P.A., Sheu, F.S. and Routtenberg, A. (1990) Inhibition of protein kinase C blocks two components of LTP persistence, leaving initial potentiation intact. J. Neurosci., 10(10): 3353–3360.

445 Contractor, A., Rogers, C., Maron, C., Henkemeyer, M., Swanson, G.T. and Heinemann, S.F. (2002) Trans-synaptic Eph receptor-ephrin signaling in hippocampal mossy fiber LTP. Science, 296(5574): 1864–1869. Corrigall, W.A. (1983) Opiates and the hippocampus: a review of the functional and morphological evidence. Pharmacol. Biochem. Behav., 18: 255–262. Dalby, N.O. and Mody, I. (2003) Activation of NMDA receptors in rat dentate gyrus granule cells by spontaneous and evoked transmitter release. J. Neurophysiol., 90(2): 786–797. Davis, C., Jones, F. and Derrick, B.E. (2004) Novelty enhances the induction and maintenance of LTP in the in the dentate gyrus. J. Neurosci., 24(29): 6497–6506. Davis, S., Vanhoutte, P., Pages, C., Caboche, J. and Laroche, S. (2000) The MAPK/ERK cascade targets both Elk-1 and cAMP response element-binding protein to control long-term potentiation-dependent gene expression in the dentate gyrus in vivo. J. Neurosci., 20(12): 4563–4572. Dayan, P. and Willshaw, D.J. (1991) Optimising synaptic learning rules in linear associative memories. Biol. Cybern., 65(4): 253–265. Derrick, B.E. and Martinez Jr., J.L. (1989) A unique opioid peptide-dependent form of long-term potentiation is found in the CA3 region of the rat hippocampus. Adv. Biosci., 75: 213–216. Derrick, B.E. and Martinez Jr., J.L. (1994a) Opioid receptor activation is one factor underlying the frequency dependence of mossy fiber LTP induction. J. Neurosci., 14(7): 4359–4367. Derrick, B.E. and Martinez Jr., J.L. (1994b) Frequencydependent associative long-term potentiation at the hippocampal mossy fiber-CA3 synapse. Proc. Natl. Acad. Sci. U.S.A., 91: 10290–10294. Derrick, B.E. and Martinez Jr., J.L. (1996) Associative, bidirectional modifications at the hippocampal mossy fibre-CA3 synapse. Nature, 381: 429–434. Derrick, B.E., Rodriguez, S.B., Lieberman, D.N. and Martinez Jr., J.L. (1992) Mu opioid receptors are associated with the induction of LTP at hippocampal mossy fiber synapses. J. Pharmacol. Exp. Ther., 263: 725–733. Desmond, N.L., Colbert, C.M., Zhang, D.X. and Levy, W.B. (1991) NMDA receptor antagonists block the induction of long-term depression in the hippocampal dentate gyrus of the anesthetized rat. Brain Res., 552: 93–98. Desmond, N.L. and Levy, W.B. (1983) Synaptic correlates of associative potentiation/depression: an ultrastructural study in the hippocampus. Brain Res., 265(1): 21–30. Desmond, N.L. and Levy, W.B. (1986) Changes in the postsynaptic density with long-term potentiation in the dentate gyrus. J. Comp. Neurol., 253(4): 476–482. Do, V.H., Martinez, C.O., Martinez Jr., J.L. and Derrick, B.E. (2002) Long-term potentiation in direct perforant path projections to hippocampal area CA3 in vivo. J. Neurophysiol., 87: 669–674. Doyere, V., Srebro, B. and Laroche, S. (1997) Heterosynaptic LTD and depotentiation in the medial perforant path of the

dentate gyrus in the freely moving rat. J. Neurophysiol., 77(2): 571–578. Dragunow, M., Abraham, W.C., Goulding, M., Mason, S.E., Robertson, H.A. and Faull, R.L. (1989) Long-term potentiation and the induction of c-fos mRNA and proteins in the dentate gyrus of unanesthetized rats. Neurosci. Lett., 101(3): 274–280. Errington, M.L., Bliss, T.V., Richter-Levin, G., Yenk, K., Doyere, V. and Laroche, S. (1995) Stimulation at 1–5 Hz does not produce long-term depression or depotentiation in the hippocampus of the adult rat in vivo. J. Neurophysiol., 74(4): 1793–1799. Escobar, M., Barea-Rodriguez, E.J., Derrick, B.E. and Martinez Jr., J.L. (1997) Mossy fiber synaptogenesis is induced by high frequency stimulation of mossy fibers. Brain Res., 751: 330–335. Fortin, N.J., Agster, K.L. and Eichenbaum, H.B. (2002) Critical role of the hippocampus in memory for sequences of events. Nat. Neurosci., 5(5): 458–462. Francesconi, W., Berton, F., Demuro, A., Madamba, S.G. and Siggins, G.R. (1997) Naloxone blocks long-term depression of excitatory transmission in rat CA1 hippocampus in vitro. Arch. Ital. Biol., 135(1): 37–48. Frey, U., Frey, S., Schollmeier, F. and Krug, M. (1996) Influence of actinomycin D, a RNA synthesis inhibitor, on long-term potentiation in rat hippocampal neurons in vivo and in vitro. J. Physiol., 490(Pt 3): 703–711. Frey, U., Schollmeier, K., Reymann, K.G. and Seidenbecher, T. (1995) Asymptotic hippocampal long-term potentiation in rats does not preclude additional potentiation at later phases. Neuroscience, 67(4): 799–807. Fry, J.P., Zieglgansberger, W. and Herz, A. (1979) Specific versus non-specific actions of opioids on hippocampal neurons in the rat brain. Brain Res., 163: 295–305. Fujii, S., Saito, K., Miyakawa, H., Ito, K. and Kato, H. (1991) Reversal of long-term potentiation (depotentiation) induced by tetanus stimulation of the input to CA1 neurons of guinea pig hippocampal slices. Brain Res., 555(1): 112–122. Gisabella, B., Rowan, M.J. and Anwyl, R. (2003) Mechanisms underlying the inhibition of long-term potentiation by preconditioning stimulation in the hippocampus in vitro. Neuroscience, 121(2): 297–305. Gonzales, R.B., DeLeon Galvan, C.J., Rangel, Y.M. and Claiborne, B.J. (2001) Distribution of thorny excrescences on CA3 pyramidal neurons in the rat hippocampus. J. Comp. Neurol., 430(3): 357–368. Grossberg, S. (1987) Competitive learning: from interactive activation to adaptive resonance. Cogn. Sci., 11: 23–63. Gustafsson, B. and Wigstrom, H. (1986) Hippocampal longlasting potentiation produced by pairing single volleys and brief conditioning tetani evoked in separate afferents. J. Neurosci., 6(6): 1575–1582. Gustafsson, B., Wigstrom, H., Abraham, W.C. and Huang, Y.Y. (1987) Long-term potentiation in the hippocampus using depolarizing current pulses as the conditioning stimulus

446 to single volley synaptic potentials. J. Neurosci., 7(3): 774–780. Guzowski, J.F., Lyford, G.L., Stevenson, G.D., Houston, F.P., McGaugh, J.L., Worley, P.F. and Barnes, C.A. (2000) Inhibition of activity-dependent arc protein expression in the rat hippocampus impairs the maintenance of long-term potentiation and the consolidation of long-term memory. J. Neurosci., 20(11): 3993–4001. Harris, E.W. and Cotman, C.W. (1986) Long-term potentiation of guinea pig mossy fiber responses is not blocked by N-methyl D-aspartate antagonists. Neurosci. Lett., 70: 132–137. Hasselmo, M.E., Schnell, E. and Barkai, E. (1995) Dynamics of learning and recall at excitatory recurrent synapses and cholinergic modulation in rat hippocampal region CA3. J. Neurosci., 15(7 Pt 2): 5249–5262. Havik, B., Rokke, H., Bardsen, K., Davanger, S. and Bramham, C.R. (2003) Bursts of high-frequency stimulation trigger rapid delivery of pre-existing alpha-CaMKII mRNA to synapses: a mechanism in dendritic protein synthesis during long-term potentiation in adult awake rats. Eur. J. Neurosci., 17(12): 2679–2689. Hebb, D.O. (1949) The Organization of Behavior. Wiley, New York. Henze, D.A., Wittner, L. and Buzsaki, G. (2002) Single granule cells reliably discharge targets in the hippocampal CA3 network in vivo. Nat. Neurosci., 5(8): 790–795. Hertz, J., Krogh, A. and Palmer, R.G. (1991) Introduction to the Theory of Neural Computation. Addison-Wesley Publishing, Redwood City, CA. Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. J. Comp. Neurol., 144: 215–232. Holland, L.L. and Wagner, J.J. (1998) Primed facilitation of homosynaptic long-term depression and depotentiation in rat hippocampus. J. Neurosci., 18(3): 887–894. Hong, J.S., McGinty, J.F., Grimes, L., Kanamatsu, T., Obie, J. and Mitchell, C.L. (1988) Seizure-induced alterations in the metabolism of hippocampal opioid peptides suggest opioid modulation of seizure-induced behaviors. In: McGinty, J.F., Friedman, D.P. (Eds.), Opioids in the Hippocampus. NIDA, Rockville, MD, pp. 48–66. Hopfield, J.J. (1982) Neural networks and physical systems with emergent collective computational abilities. Proc. Natl. Acad. Sci. U.S.A., 79(8): 2554–2558. Huang, Y.Y., Colino, A., Selig, D.K. and Malenka, R.C. (1992) The influence of prior synaptic activity on the induction of long-term potentiation. Science, 255(5045): 730–733. Huang, Y.Y., Kandel, E.R., Varshavsky, L., Brandon, E.P., Qi, M., Idzerda, R.L., McKnight, G.S. and Bourtchouladze, R. (1995) A genetic test of the effects of mutations in PKA on mossy fiber LTP and its relation to spatial and contextual learning. Cell, 83(7): 1211–1222.

Jaffe, D. and Johnston, D. (1990) Induction of long-term potentiation at hippocampal mossy fibers follow a Hebbian rule. J. Neurophysiol., 64: 948–960. James, W. (1890) Association. In: Psychology: A Briefer Course. Holt, New York, pp. 253–279. Jeffery, K.J., Donnett, J.G. and O’Keefe, J. (1995) Medial septal control of theta-correlated unit firing in the entorhinal cortex of awake rats. Neuroreport, 6(16): 2166–2170. Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Disconnection analysis of CA3 and DG in mediating encoding but not retrieval in a spatial maze learning task. Learn. Mem., 13(4): 458–464. Jin, W. and Chavkin, C. (1999) Mu opioids enhance mossy fiber synaptic transmission indirectly by reducing GABAB receptor activation. Brain Res., 821(2): 286–293. Jones, M.W., Errington, M.L., French, P.J., Fine, A., Bliss, T.V., Garel, S., Charnay, P., Bozon, B., Laroche, S. and Davis, S. (2001) A requirement for the immediate early gene Zif268 in the expression of late LTP and long-term memories. Nat. Neurosci., 4(3): 289–296. Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of unit activity in the hippocampal granular layer. Hippocampus, 3: 165–182. Kali, S. and Dayan, P. (2000) The involvement of recurrent connections in area CA3 in establishing the properties of place fields: a model. J. Neurosci., 20(19): 7463–7477. Kapur, A., Yeckel, M.F. and Johnston, D. (2001) Hippocampal mossy fiber activity evokes Ca2+ release in CA3 pyramidal neurons via a metabotropic glutamate receptor. J. Neurophysiol., 79(4): 2181–2190. Kart-Teke, E., De Souza Silva, M.A., Huston, J.P. and Dere, E. (2006) Wistar rats show episodic-like memory for unique experiences. Neurobiol. Learn. Mem., 85(2): 173–182. Kato, A., Ozawa, F., Saitoh, Y., Hirai, K. and Inokuchi, K. (1997) vesl, a gene encoding VASP/Ena family related protein, is upregulated during seizure, long-term potentiation and synaptogenesis. FEBS Lett., 412(1): 183–189. O’Keefe, J. and Nadel, L. (1978) Hippocampus as a Cognitive Map. Oxford University Press, New York. Kerr, D.S. and Abraham, W.C. (1995) Cooperative interactions among afferents govern the induction of homosynaptic longterm depression in the hippocampus. Proc. Natl. Acad. Sci., 92(25): 11637–11641. Kerr, D.S. and Abraham, W.C. (1996) LTD: many means to how many ends? Hippocampus, 6(1): 30–34. Kirkwood, A., Dudek, S.M., Gold, J.T., Aizenman, C.D. and Bear, M.F. (1993) Common forms of synaptic plasticity in the hippocampus and neocortex in vitro. Science, 260(5113): 1518–1521. Kleschevnikov, A.M. and Routtenberg, A. (2003) Long-term potentiation recruits a trisynaptic excitatory associative network within the mouse dentate gyrus. Eur. J. Neurosci., 18(6): 1717.

447 Kobayashi, K. and Poo, M.M. (2004) Spike train timingdependent associative modification of hippocampal CA3 recurrent synapses by mossy fibers. Neuron, 41(3): 445–454. Kohonen, T. (1984) Self organization and Associative Memory. Springer-Verlag, Berlin. Kohonen, T., Lehtio, P., Rovamo, J., Hyvarinen, J., Bry, K. and Vainio, L. (1977) A principle of neural associative memory. Neuroscience, 2: 1065–1076. Kosub, K., Do, V. and Derrick, B.E. (2005) NMDA antagonists block heterosynaptic LTD, but not LTP, following lateral perforant path stimulation. Neurosci. Lett., 374: 29–34. Krug, M., Brodemann, R. and Wagner. (2001) Simultaneous activation and opioid modulation of long-term potentiation in the dentate gyrus and the hippocampal CA3 region after stimulation of the perforant pathway in freely moving rats. Brain Res., 913(1): 68–77. Krug, M., Lossner, B. and Ott, T. (1984) Anisomycin blocks the late phase of long-term potentiation in the dentate gyrus of freely moving rats. Brain Res. Bull., 13(1): 39–42. Kulla, A. and Manahan-Vaughan, D. (2000) Depotentiation in the dentate gyrus of freely moving rats is modulated by D1/D5 dopamine receptors. Cereb. Cortex, 10(6): 614–620. Kulla, A. and Manahan-Vaughan, D. (2002) Modulation by serotonin 5-HT(4) receptors of long-term potentiation and depotentiation in the dentate gyrus of freely moving rats. Cereb. Cortex, 12(2): 150–162. Kulla, A., Reymann, K.G. and Manahan-Vaughan, D. (1999) Time-dependent induction of depotentiation in the dentate gyrus of freely moving rats: involvement of group 2 metabotropic glutamate receptors. Eur. J. Neurosci., 11(11): 3864–3872. Lee, H.K., Barbarosie, M., Kameyama, K., Bear, M.F. and Huganir, R.L. (2000) Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity. Nature, 405(6789): 955–959. Lee, I. and Kesner, R.P. (2004) Encoding versus retrieval of spatial memory: double dissociation between the dentate gyrus and the perforant path inputs into CA3 in the dorsal hippocampus. Hippocampus, 14(1): 66–76. Lei, S., Pelkey, K.A., Topolnik, L., Congar, P., Lacaille, J.C. and McBain, C.J. (2003) Depolarization-induced long-term depression at hippocampal mossy fiber-CA3 pyramidal neuron synapses. J. Neurosci., 23(30): 9786–9795. Levy, W.B. (1996) A sequence predicting CA3 is a flexible associator that learns and uses context to solve hippocampallike tasks. Hippocampus, 6(6): 579–590. Levy, W.B. and Steward, O. (1979) Synapses as associative memory elements in the hippocampal formation. Brain Res., 175: 233–245. Levy, W.B. and Steward, O. (1983) Temporal contiguity requirements for long-term associative potentiation/depression in the hippocampus. Neuroscience, 8(4): 791–797.

Li, S., Cullen, W.K., Anwyl, R. and Rowan, M.J. (2003) Dopamine-dependent facilitation of LTP induction in hippocampal CA1 by exposure to spatial novelty. Nat. Neurosci., 6(5): 526–531. Lisman, J.E. and McIntyre, C.C. (2001) Synaptic plasticity: a molecular memory switch. Curr. Biol., 11(19): R788– R791. Magee, J.C. and Johnston, D. (1997) A synaptically-controlled, associative signal for Hebbian plasticity in hippocampal neurons. Science, 275: 209–213. Maguire, C., Casey, M., Kelly, A., Mullany, P.M. and Lynch, M.A. (1999) Activation of tyrosine receptor kinase plays a role in expression of long-term potentiation in the rat dentate gyrus. Hippocampus, 9(5): 519–526. Malinow, R. (2003) AMPA receptor trafficking and long-term potentiation. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358(1432): 707–714. Manabe, T. (1997) Two forms of hippocampal long-term depression, the counterpart of long-term potentiation. Rev. Neurosci., 8(3–4): 179–193. Manahan-Vaughan, D. (1998) Priming of group 2 metabotropic glutamate receptors facilitates induction of long-term depression in the dentate gyrus of freely moving rats. Neuropharmacology, 37(12): 1459–1464. O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995a) Metabotropic glutamate receptor-induced homosynaptic long-term depression and depotentiation in the dentate gyrus of the rat hippocampus in vitro. Neuropharmacology, 34(8): 983–989. O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995b) Dantrolene inhibits long-term depression and depotentiation of synaptic transmission in the rat dentate gyrus. Neuroscience, 68(3): 621–624. Marr, D. (1971) Simple Memory: a theory for archicorticortex. Proc. R. Soc. Lond. B, 262: 23–81. Martin, M.R. (1983) Naloxone and long-term potentiation of hippocampal CA3 field potentials in vitro. Neuropeptides, 4(1): 45–50. Martin, S.J. (1998) Time-dependent reversal of dentate LTP by 5 Hz stimulation. Neuroreport, 9(17): 3775–3781. Martinez, C.O., Do, V.H. and Derrick, B.E. (2002) Associative LTP Among afferents to the CA3 region in vivo. Brain Res., 94: 86–94. Massey, P.V., Johnson, B.E., Moult, P.R., Auberson, Y.P., Brown, M.W., Molnar, E., Collingridge, G.L. and Bashir, Z.I. (2004) Differential roles of NR2A and NR2Bcontaining NMDA receptors in cortical long-term potentiation and long-term depression. J. Neurosci., 24: 7821–7828. Matthies, H., Schroeder, H., Becker, A., Loh, H., Hollt, V. and Krug, M. (2000) Lack of expression of long-term potentiation in the dentate gyrus but not in the CA1 region of the hippocampus of mu-opioid receptor-deficient mice. Neuropharmacology, 39(6): 952–960. McClelland, J.L., McNaughton, B.L. and O’Reilly, R.C. (1995) Why there are complementary learning systems in the

448 hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory. Psychol. Rev., 102: 419–457. McMahon, D.B. and Barrionuevo, G. (2002) Short- and longterm plasticity of the perforant path synapse in hippocampal area CA3. J. Neurophysiol., 88(1): 528–533. McNaughton, B.L. and Barnes, C.A. (1977) Physiological identification and analysis of dentate granule cell responses to stimulation of the medial and lateral perforant pathways in the rat. J. Comp. Neurol., 175: 439–454. McNaughton, B.L., Barnes, C.A., Meltzer, J. and Sutherland, R.J. (1989) Hippocampal granule cells are necessary for normal spatial learning but not for spatially-selective pyramidal cell discharge. Exp. Brain Res., 76: 485–496. McNaughton, B.L., Douglas, R.M. and Goddard, D.V. (1978) Synaptic enhancement in fascia dentata: cooperativity among coactive afferents. Brain Res., 157: 277–293. Mellor, J. and Nicoll, R.A. (2001) Hippocampal mossy fiber LTP is independent of postsynaptic calcium. Nat. Neurosci., 4(2): 125–126. Milner, T.A. and Drake, C.T. (2001) Ultrastructural evidence for presynaptic mu opioid receptor modulation of synaptic plasticity in NMDA-receptor-containing dendrites in the dentate gyrus. Brain Res. Bull., 54(2): 131–140. Mizumori, S.J., McNaughton, B.L., Barnes, C.A. and Fox, K.B. (1989) Preserved spatial coding in hippocampal CA1 pyramidal cells during reversible suppression of CA3c output: evidence for pattern completion in hippocampus. J. Neurosci., 9(11): 3915–3928. Monaghan, D.T. and Cotman, C.W. (1985) Distribution of N-methyl-D-aspartate sensitive l-[3H]glutamate-binding sites in rat brain. J. Neurosci., 5: 2909–2919. Morgan, S.L. and Teyler, T.J. (2001) Electrical stimuli patterned after the theta-rhythm induce multiple forms of LTP. J. Neurophysiol., 86(3): 1289–1296. Morris, R.G.M. (1989) Computational neuroscience: modeling the brain. In: Morris R.G.M. (Ed.), Parallel-Distributed Processing: Implication for Psychology and Neurobiology. Oxford University Press, NY, pp. 203–212. Morris, R.G.M. and McNaughton, B.L. (1987) Memory storage in a distributed model of hippocampal formation. Trends Neurosci., 10: 408–414. Mott, D.D. and Lewis, D.V. (1992) GABAB receptors mediate disinhibition and facilitate long-term potentiation in the dentate gyrus. Epilepsy Res. Suppl., 7: 119–134. Munoz, M.D., Nunez, A. and Garcia-Austt, E. (1990) In vivo intracellular analysis of rat dentate gyrus granule cells. Brain Res., 509: 91–98. Nakazawa, K., Quirk, M.C., Chitwood, R.A., Watanabe, M., Yeckel, M.F., Sun, L.D., Kato, A., Carr, C.A., Johnston, D., Wilson, M.A. and Tonegawa, S. (2002) Requirement for hippocampal CA3 NMDA receptors in associative memory recall. Science, 297(5579): 211–218. Neumaier, J.F. and Chavkin, C. (1989) Release of endogenous opioid peptides displaces [3H]diprenorphine binding in rat hippocampal slices. Brain Res., 493: 292–302.

Nguyen, P.V. and Kandel, E.R. (1996) A macromolecular synthesis-dependent late phase of long-term potentiation requiring cAMP in the medial perforant pathway of rat hippocampal slices. J. Neurosci., 16(10): 3189–3198. North, R.A., Williams, J.T., Surprenant, A. and Christie, M.J. (1987) Mu and delta receptors belong to a family of receptors that are coupled to potassium channels. Proc. Natl. Acad. Sci. U.S.A., 84(15): 5487–5491. Otani, S., Marshall, C.J., Tate, W.P., Goddard, G.V. and Abraham, W.C. (1989) Maintenance of long-term potentiation in rat dentate gyrus requires protein synthesis but not messenger RNA synthesis immediately post-tetanization. Neuroscience, 28(3): 519–526. Otmakhova, N.A. and Lisman, J.E. (1998) D1/D5 dopamine receptors inhibit depotentiation at CA1 synapses via cAMP-dependent mechanism. J. Neurosci., 18(4): 1270–1279. Pang, K., Williams, M.J. and Olton, D.S. (1993) Activation of the medial septal area attenuates LTP of the lateral perforant path and enhances heterosynaptic LTD of the medial perforant path in aged rats. Brain Res., 632(1–2): 150–160. Philpot, B., Bear, M.F. and Abraham, W.C. (1999) Metaplasticity: the plasticity of synaptic plasticity. In: Katz P.S. (Ed.), Beyond Neurotransmission: Neuromodulation and its Importance for Information processing. Oxford, London, pp. 161–197. Piguet, P. and North, R.A. (1983) Opioid actions at mu and delta receptors in the rat dentate gyrus in vitro. J. Pharmacol. Exp. Ther., 266(2): 1139–1149. Ramirez-Amaya, V., Balderas, I., Sandoval, J., Escobar, M.L. and Bermudez-Rattoni, F. (2001) Spatial long-term memory is related to mossy fiber synaptogenesis. J. Neurosci., 21(18): 7340–7348. Ramon y Cajal, S. (1937) Recuerdos de mi Vida. MIT Press, Cambridge. Reid, C.A., Dixon, D.B., Takahashi, M., Bliss, T.V. and Fine, A. (2004) Optical quantal analysis indicates that long-term potentiation at single hippocampal mossy fiber synapses is expressed through increased release probability, recruitment of new release sites, and activation of silent synapses. J. Neurosci., 24(14): 3618–3626. Rolls, E.T. and Treves, A. (1998) Neural Networks and Brain Function. Oxford University Press, New York. Rosenzweig, E.S., Barnes, C.A. and McNaughton, B.L. (2002) Making room for new memories. Nat. Neurosci., 5: 6–8. Rycroft, B.K. and Gibb, A.J. (2002) Direct effects of calmodulin on NMDA receptor single channel gating in rat hippocampal granule cells. J. Neurosci., 22: 8860– 8868. Sanberg, C.-D., Jones, F., Do, V., Dieguez, D. and Derrick, B.E. (2006) Antagonists of the 5-HT1a receptor block dentate LTP induction in novel, but not familiar environments. Learn. Mem., 13(1): 52–62.

449 Scharfman, H.E. (1996) Conditions required for polysynaptic excitation of dentate granule cells by area CA3 pyramidal cells in rat hippocampal slices. Neuroscience, 72(3): 655–658. Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and Goodman, J.H. (2002) Structural and functional asymmetry in the normal and epileptic rat dentate gyrus. J. Comp. Neurol., 454(4): 424–439. Schulz, S., Siemer, H., Krug, M. and Hollt, V. (1999) Direct evidence for biphasic cAMP responsive element-binding protein phosphorylation during long-term potentiation in the rat dentate gyrus in vivo. J. Neurosci., 19(13): 5683–5692. Schwegler, H., Crusio, W.E. and Brust, I. (1990) Hippocampal mossy fibers and radial-maze learning in the mouse: a correlation with spatial working memory but not with non-spatial reference memory. Neuroscience, 34(2): 293–298. Scoville, W.B. and Milner, B. (1957) Loss of recent memory by bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatry, 20: 11–21. Segal, M. (1972) Hippocampal unit responses to perforant path stimulation. Exp. Neurol., 35: 541–546. Sheng, M. and Greenberg, M.E. (1990) The regulation and function of c-fos and other immediate early genes in the nervous system. Neuron, 4(4): 477–485. Shors, T.J. and Matzel, L.D. (1997) Long-term potentiation: what’s learning got to do with it? Behav. Brain Sci., 20(4): 597–614. Shouval, H., Intrator, N. and Cooper, L.N. (1997) BCM network develops orientation selectivity and ocular dominance in natural scene environment. Vision Res., 37(23): 3339–3342. Skaggs, W.E. and McNaughton, B.L. (1992) Computational approaches to hippocampal function. Curr. Opin. Neurobiol., 2(2): 209–211. Snyder, J.S., Kee, N. and Wojtowicz, J.M. (2001) Effects of adult neurogenesis on synaptic plasticity in the rat dentate gyrus. J. Neurophysiol., 85(6): 2423–2431. Soderling, T.R. and Derkach, V.A. (2000) Postsynaptic protein phosphorylation and LTP. Trends Neurosci., 23(2): 75–80. Squire, L. (1992) Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychol. Rev., 99: 195–231. Staubli, U. and Lynch, G. (1987) Stable hippocampal long-term potentiation elicited by ‘theta’ pattern stimulation. Brain Res., 435(1–2): 227–234. Steward, O. (1976) Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat. J. Comp. Neurol., 167: 285–314. Steward, O. and Scoville, S.A. (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol., 169: 347–370. Steward, O., Wallace, C.S., Lyford, G.L. and Worley, P.F. (1998) Synaptic activation causes the mRNA

for the IEG Arc to localize selectively near activated postsynaptic sites on dendrites. Neuron, 21(4): 741– 751. Straube, T., Korz, V., Balschun, D. and Frey, J.U. (2003) Requirement of beta-adrenergic receptor activation and protein synthesis for LTP-reinforcement by novelty in rat dentate gyrus. J. Physiol., 552(Pt 3): 953–960. Swearengen, E. and Chavkin, C. (1987) NMDA receptor antagonist D-APV depresses excitatory activity produced by normorphine in rat hippocampal slices. Neurosci. Lett., 78: 80–84. Sweatt, J.D. (2001) The neuronal MAP kinase cascade: a biochemical signal integration system subserving synaptic plasticity and memory. J. Neurochem., 76(1): 1–10. Thiels, E., Xie, X., Yeckel, M.F., Barrionuevo, G. and Berger, T.W. (1996) NMDA receptor-dependent LTD in different subfields of hippocampus in vivo and in vitro. Hippocampus, 6(1): 43–51. Thomas, K.L., Davis, S., Hunt, S.P. and Laroche, S. (1996) Alterations in the expression of specific glutamate receptor subunits following hippocampal LTP in vivo. Learn. Mem., 3(2–3): 197–208. Thomas, K.L., Laroche, S., Errington, M.L., Bliss, T.V. and Hunt, S.P. (1994) Spatial and temporal changes in signal transduction pathways during LTP. Neuron, 13(3): 737–745. Trepel, C. and Racine, R.J. (1998) Long-term potentiation in the neocortex of the adult, freely moving rat. Cereb. Cortex, 8(8): 719–729. Treves, A. and Rolls, E.T. (1992) Computational constraints suggest the need for two distinct input systems to the hippocampal CA3 network. Hippocampus, 2(2): 189–199. Treves, A. and Rolls, E.T. (1994) Computational analysis of the role of the hippocampus in memory. Hippocampus, 4: 374–391. Treves, A. and Roudi, Y. (2005) On the evolution of the brain. In: Chow C.C., Gutkin B., Hansel D. and Meunier C. (Eds.), Methods and Models in Neurophysics. Elsevier, Holland, pp. 641–690. Trommald, M., Hulleberg, G. and Andersen, P. (1996) Longterm potentiation is associated with new excitatory spine synapses on rat dentate granule cells. Learn. Mem., 3(2–3): 218–228. Tulving, E. (1983) Elements of Episodic Memory. Clarendon Press, Oxford. Urban, N.N. and Barrionuevo, G. (1996) Induction of Hebbian and non-Hebbian mossy fiber long-term potentiation by distinct patterns of high-frequency stimulation. J. Neurosci., 16(13): 4293–4299. Villarreal, D., Do, V., Haddad, E. and Derrick, B.E. (2002) NMDA antagonists sustain LTP and spatial memory: evidence for active processes underlying LTP decay. Nat. Neurosci., 5: 48–52. Villarreal, D.M. and Derrick, B.E. (2000) Long-term metaplasticity: correspondence with LTP longevity

450 and the BCM Theory. Abst. Soc. for Neurosci., # 134.1 (online). Villarreal, D.M. and Derrick, B.E. (2006) Input-specific long-term potentiation reinduction at the rat perforant path to dentate gyrus. Soc. for Neurosci. Abst., # 731.2 (online). Wagner, J.J., Caudle, R.M., Neumaier, J.F. and Chavkin, C. (1990) Stimulation of endogenous opioid release displaces mu receptor binding in rat hippocampus. Neuroscience, 37: 45–53. Wagner, J.J., Etemad, L.R. and Thompson, A.M. (2001) Opioid-mediated facilitation of long-term depression in rat hippocampus. J. Pharmacol. Exp. Ther., 296(3): 776–781. Wagner, J.J., Evans, C.J. and Chavkin, C. (1991) Focal stimulation of the mossy fibers releases endogenous dynorphins that bind to k1-opioid receptors in guinea pig hippocampus. J. Neurochem., 57: 333–343. Wagner, J.J., Terman, G.W. and Chavkin, C. (1993) Endogenous dynorphins inhibit excitatory neurotransmission and block LTP induction in the hippocampus. Nature, 363(6428): 451–454. Wang, D., Tolbert, L.M., Carlson, K.W. and Sadee, W. (2000) Nuclear Ca2+/calmodulin translocation activated by mu-opioid (OP3) receptor. J. Neurochem., 74(4): 1418–1425. Wang, J., Yeckel, M.F., Johnston, D. and Zucker, R.S. (2004) Photolysis of postsynaptic caged Ca2+ can potentiate and depress mossy fiber synaptic responses in rat hippocampal CA3 pyramidal neurons. J. Neurophysiol., 91(4): 1596–1607. Weisskopf, M.G., Zalutsky, R.A. and Nicoll, R.A. (1993) The opioid peptide dynorphin mediates heterosynaptic depression of hippocampal mossy fibre synapses and modulates long-term potentiation. Nature, 362: 423–427. Wexler, E.M. and Stanton, P.K. (1993) Priming of homosynaptic long-term depression in hippocampus by previous synaptic activity. Neuroreport, 4(5): 591–594. White, J.D., Gall, C.M. and McKelvy, J.F. (1987) Enkephalin biosynthesis and enkephalin gene expression are increased in hippocampal mossy fibers following a unilateral lesion of the hilus. J. Neurosci., 7(3): 753–759. Wickens, J.R. and Abraham, W.C. (1991) The involvement of L-type calcium channels in heterosynaptic long-term depression in the hippocampus. Neurosci. Lett., 130(1): 128–132. Wigstrom, H. and Gustaffson, B. (1985) Facilitation of hippocampal long-lasting potentiation by GABA antagonists. Acta Physiol. Scand., 125: 159–172. Williams, J.M., Guevremont, D., Kennard, J.T., MasonParker, S.E., Tate, W.P. and Abraham, W.C. (2003) Long-term regulation of N-methyl-D-aspartate receptor subunits and associated synaptic proteins following

hippocampal synaptic plasticity. Neuroscience, 118(4): 1003–1013. Williams, S. and Johnston, D. (1989) Long-term potentiation of hippocampal mossy fiber synapses is blocked by postsynaptic injection of calcium chelators. Neuron, 3: 583–588. Williams, S.H. and Johnston, D. (1996) Actions of endogenous opioids on NMDA receptor-independent long-term potentiation in area CA3 of the hippocampus. J. Neurosci., 16: 3652–3660. Wilson, M.A. and McNaughton, B.L. (1993) Dynamics of the hippocampal ensemble code for space. Science, 261(5124): 1055–1058. Winson, J. and Abzug, C. (1978) Dependence upon behavior of neuronal transmission from perforant pathway through entorhinal cortex. Brain Res., 147(2): 422–427. Witter, M.P., Griffioen, A.W., Jorritsma-Byham, B. and Krijnen, J.L. (1988) Entorhinal projections to the hippocampal CA1 region in the rat: an underestimated pathway. Neurosci. Lett., 85(2): 193–198. Wu, J., Rowan, M.J. and Anwyl, R. (2004) Synaptically stimulated induction of group I metabotropic glutamate receptor-dependent long-term depression and depotentiation is inhibited by prior activation of metabotropic glutamate receptors and protein kinase C. Neuroscience, 123(2): 507–514. Wu, J., Rowan, M.J. and Anwyl, R. (2006) Long-term potentiation is mediated by multiple kinase cascades involving CaMKII or either PKA or p42/44 MAPK in the adult rat dentate gyrus in vitro. J. Neurophysiol., 95(6): 3519–3527. Xie, C.-W. and Lewis, D.V. (1991) Opioid mediated facilitation of long-term potentiation at the lateral perforant pathdentate granule cell synapse. J. Pharmacol. Exp. Ther., 256: 289–296. Yeckel, M.F. and Berger, T.W. (1990) Feedforward excitation of the hippocampus by afferents from the entorhinal cortex: redefinition of the role of the trisynaptic pathway. Proc. Natl. Acad. Sci. U.S.A., 87: 5832–5836. Yeckel, M.F. and Berger, T.W. (1998) Spatial distribution of potentiated synapses in hippocampus: dependence on cellular mechanisms and network properties. J. Neurosci., 18: 438–450. Yeckel, M.F., Kapur, A. and Johnston, D. (1999) Multiple forms of LTP in hippocampal CA3 neurons use a common postsynaptic mechanism. Nat. Neurosci., 2(7): 625–633. Ying, S.W., Futter, M., Rosenblum, K., Webber, M.J., Hunt, S.P., Bliss, T.V. and Bramham, C.R. (2002) Brain-derived neurotrophic factor induces long-term potentiation in intact adult hippocampus: requirement for ERK activation coupled to CREB and upregulation of Arc synthesis. J. Neurosci., 22(5): 1532–1540. Zalutsky, R.A. and Nicoll, R.A. (1990) Comparison of two forms of long-term potentiation in single hippocampal neurons. Science, 248: 1619–1624.

451 Zeineh, M.M., Engel, S.A., Thompson, P.M. and Bookheimer, S.Y. (2003) Dynamics of the hippocampus during encoding and retrieval of face-name pairs. Science, 299(5606): 577–580. Zhang, D.X. and Levy, W.B. (1992) Ketamine blocks the induction of LTP at the lateral entorhinal cortex-dentate gyrus synapses. Brain Res., 593(1): 124–127.

Zimprich, A., Simon, T. and Hollt, V. (1995) Transfected rat mu opioid receptors (rMOR1 and rMOR1B) stimulate phospholipase C and Ca2+ mobilization. Neuroreport, 7(1): 54–56. Zucker, R.S. and Regehr, W.G. (2002) Short-term synaptic plasticity. Annu. Rev. Physiol., 64: 355–405.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 25

Control of synaptic consolidation in the dentate gyrus: mechanisms, functions, and therapeutic implications Clive R. Bramham Department of Biomedicine and Bergen Mental Health Research Center, University of Bergen, Jonas Lies vei 91, N-5009 Bergen, Norway

Abstract: Synaptic consolidation refers to the development and stabilization of protein synthesis-dependent modifications of synaptic strength as observed during long-term potentiation (LTP) and long-term depression (LTD). Activity-dependent changes in synaptic strength are thought to underlie memory storage and other adaptive responses of the nervous systems of importance in mood stability, reward behavior, and pain control. This chapter focuses on the mechanisms and functions of synaptic consolidation in the dentate gyrus, a critical structure not only in hippocampal memory function, but also in regulation of stress responses and cognitive aspects of depression. Recent evidence suggests that synaptic consolidation at excitatory medial perforant path-granule cell synapses requires brain-derived neurotrophic factor (BDNF) signaling and induction of the immediate early gene activity-regulated cytoskeleton-associated protein (Arc). Arc mRNA is strongly induced and transported to dendritic processes following high-frequency stimulation (HFS) that induces LTP in the rat dentate gyrus in vivo. Sustained synthesis of Arc during a surprisingly protracted time-window is required for hyperphosphorylation of actin depolymerizing factor/ cofilin and local expansion of the actin cytoskeleton in vivo. Furthermore, this process of Arc-dependent synaptic consolidation is activated in response to brief infusion of BDNF. Microarray expression profiling has revealed a panel of BDNF-regulated genes that may cooperate with Arc during synaptic consolidation. In addition to regulating gene expression, BDNF signaling modulates the fine localization and biochemical activation of the translation machinery. By modulating the spatial and temporal translation of newly induced (Arc) and constitutively-expressed mRNA in dendrites, BDNF may effectively control the window of synaptic consolidation. Dysregulation of BDNF synthesis and Arc function, specifically within the dentate gyrus, is linked to behavioral symptoms and cognitive deficits in animal models of depression and Alzheimer’s disease. Therapeutics strategies targeting synaptic consolidation hold promise for the future. Keywords: long-term potentiation; brain-derived neurotrophic factor; BDNF; Neurotrophin; hippocampus; Arc; gene expression; Depression; memory

Corresponding author. Tel.: +47 55 58 60 32; Fax: +47 55 58 64 10; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63025-8

453

454

Control of synaptic consolidation in the dentate gyrus Synaptic consolidation Persistent activity-dependent changes in synaptic strength are believed to underlie a range of adaptive brain responses, including memory formation, mood stability, and drug addiction (Bliss and Collingridge, 1993; Nestler et al., 2002). However, our understanding of the mechanisms by which altered activity patterns trigger lasting changes in synaptic efficacy, exemplified by long-term potentiation (LTP) and long-term depression (LTD), is far from complete. A critical factor in high-frequency stimulation (HFS)-induced LTP is postsynaptic activation of N-methyl-D-aspartate receptors (NMDARs) and voltage-dependent calcium channels. The resulting activation of calcium-sensitive kinases and other calcium-sensitive effectors underlies a transient, early phase of LTP maintained by protein phosphorylation and membrane insertion of AMPA-type glutamate receptors. The generation of stable, late phase LTP requires at least one period of new gene expression and protein synthesis (Bliss and Collingridge, 1993; Frey et al., 1996; Nguyen and Kandel, 1996). Analogous to memory consolidation, the process of generating late LTP is referred to as synaptic consolidation. Late phase LTP is operationally defined on the basis of sensitivity of LTP maintenance to broad spectrum inhibitors of global gene expression and protein synthesis. Progress in unraveling the molecular basis of synaptic consolidation has been hampered by the lack of evidence that expression of a specific gene product or group of gene products mediates LTP. Alterations in gene expression in response to changes in neuronal activity are common, and, as pointed out by several authors (Sanes and Lichtman, 1999; Routtenberg and Rekart, 2005), it is likely that many regulated genes have no role or only a subsidiary role in generating stable LTP. BDNF as a trigger of synaptic consolidation Several lines of evidence suggest that BDNF, acting through its receptor tyrosine kinase (TrkB)

triggers synaptic consolidation at adult excitatory synpases. BDNF, unlike other members of the neurotrophin family (nerve growth factor, neurotrophin-3, and neurotrophin-4), is widely distributed across the adult brain and in hippocampal principal cells. BDNF signals through catalytic TrkB receptors present on pre- and postsynaptical elements of glutamatergic synapses (Drake et al., 1999). Postsynaptic TrkB receptors are found in the postsynaptic density (PSD) and TrkB co-immunoprecipitates with NMDAR complex proteins (Wu et al., 1996; Drake et al., 1999; Aoki et al., 2000; Husi et al., 2000). Postsynaptic release of BDNF from secretogranin II-immunoreactive vesicles occurs in response to HFS, and is sensitive to NMDAR blockade. BDNF may therefore be viewed as a co-neurotransmitter acting in tandem with glutamate at excitatory synapses. Two additional features important to mention in the context of synaptic consolidation are: (1) BDNF regulates protein synthesis through both transcriptional and post-transcriptional mechanisms, and (2) BDNF is capable of stimulating its own release, possibly allowing sustained, regenerative signaling at synapses (Santi et al., 2006). Genetic and pharmacological studies establishing the role of BDNF-TrkB in LTP have been reviewed in detail elsewhere (Bramham and Messaoudi, 2005). The contributions of BDNF signaling to LTP can be classified as permissive or instructive. Permissive mechanisms make synapses competent for LTP. For example, constitutive (non-evoked) presynaptic BDNF signaling increases the fraction of neurotransmitter vesicles docked in the active zone. This increase in the readily releasable pool of vesicles enables sustained glutamate release during trains of action potentials, and facilitates LTP induction (Figurov et al., 1996). Instructive BDNF signaling, in contrast, is initiated in response to HFS and is required for LTP development. Acute release of BDNF during HFS modulates the induction and early maintenance of LTP (Kossel et al., 2001). A series of studies based on focal application of BDNF suggests a mechanism involving activation of a tetrodotoxin (TTX)-insensitive voltage-dependent sodium channel leading to rapid depolarization and enhanced calcium influx into dendritic spines

455

during LTP induction (Blum and Konnerth, 2005). The mechanism of TrkB coupling to the TTX-insensitive channel is unclear at the moment. Formation of late LTP is coupled to a period of sustained BDNF release and TrkB receptor activation (Kang et al., 1997; Aicardi et al., 2004). Extracellular concentrations of BDNF are elevated for less than a minute following weak HFS, leading to decremental early LTP; whereas stimulus protocols that induce late LTP produce BDNF elevation lasting at least 5 min. Pharmacological inhibition of TrkB receptor activation during the first hour after HFS blocks synaptic consolidation in hippocampal region CA1 (Kang et al., 1997). Furthermore, bath application of BDNF induces a protein synthesis-dependent increase in synaptic efficacy termed BDNF-LTP (Kang and Schuman, 1996; Messaoudi et al., 1998). Induction of BDNF-LTP at medial perforant path (MPP)-granule cell synapses in the dentate gyrus is blocked by inhibitors of RNA synthesis and occluded by prior expression of late phase, but not early phase, of HFS-LTP (Messaoudi et al., 2002; Ying et al. 2002). BDNFLTP lasts at least 1 day in awake rats and, like HFS-LTP, is associated with increases in both synaptic efficacy and EPSP-spike coupling. BDNF-LTP induction is not affected by NMDA receptor blockade, consistent with the hypothesis that NMDA receptor activation stimulates BDNF release but is not involved in BDNF-LTP induction or expression. Like HFS-LTP, BDNFLTP induction requires extracellular signalregulated protein kinase (ERK) activation and is coupled to upregulation, dendritic transport, and translation of the immediate early gene activityregulated cytoskeleton-associated protein (Arc, aka Arg 3.1). These results suggest that endogenous BDNF activates synaptic consolidation and that exogenous BDNF mimics this effect.

Arc as mediator of synapse consolidation Persistent LTP is thought to occur when small spines are converted to large mushroom-shaped spines through a mechanism dependent on actin

polymerization (Matsuzaki et al., 2004). Insertion of glutamate receptors at postsynaptic membranes, increase in PSD diameter, and structural remodeling of spines are all connected to regulation of actin dynamics (Geinisman, 2000; Lisman and Zhabotinsky, 2001; Weeks et al., 2001; Zhou et al., 2001; Fukazawa et al., 2003; Harris et al., 2003; Matsuzaki et al., 2004; Okamoto et al., 2004; Zito et al., 2004; Oertner and Matus, 2005). Fukazawa et al. (2003) showed that LTP at MPP-granule cell synapses in anesthetized rats is associated with an increase in filamentous actin (F-actin) content at activated synapses and enhanced phosphorylation of cofilin, a major regulator of actin dynamics. Phosphorylation of cofilin on Ser3 inhibits activity and promotes actin polymerization. LIM domain kinase (LIMK), one of the major cofilin kinases in brain, modulates the morphology of dendritic spines through regulation of cofilin activity. Mice lacking the LIMK1 gene exhibit small, actin-deficient spines. Although the initial amplitude of LTP is actually larger, LIMK knockout mice fail to sustain late phase LTP (Meng et al., 2002). Taken together, this suggests a major role for cofilin regulation in actin-dependent synapse expansion underlying late LTP. How is gene expression coupled to actin dynamics and synapse expansion? LTP is associated with the induction of a number of immediate-early genes that encode a variety of transcription factors and other proteins, unassociated with transcription, and some of these have been causally linked to late phase LTP and long-term memory. Arc is unique because it encodes the only mRNA known to undergo transport to distal dendritic processes of granule cells within the first hour of LTP induction (Link et al., 1995; Lyford et al., 1995). Arc mRNA is enriched at stimulated synapses, and Arc protein is transiently elevated in dendritic spines following LTP induction (Steward and Worley, 2001; Moga et al., 2004; Rodriguez et al., 2005). Arc co-precipitates with crude F-actin and is found in the PSD of excitatory synapses (Lyford et al., 1995; Husi et al., 2000). Arc is dynamically expressed in principal neurons of many cortical and limbic structures during behavioral training, and this expression is necessary for long-term memory in a variety of memory tasks (Guzowski

456

et al., 1999, 2000; Plath et al., 2006; Vazdarjanova et al., 2006). Early work of Guzowski et al. using Arc antisense oligodeoxynucleotides (AS) suggested a role for Arc in LTP maintenance in the dentate gyrus. This issue has now been examined in more detail (Messaoudi et al., 2007). These authors show that Arc antisense application at 2 h (but not 4 h) after LTP induction leads to rapid and complete reversal of LTP. LTP reversal is coupled to rapid knockdown of newly synthesized Arc mRNA and protein, dephosphorylation of hyperphosphorylated cofilin, and loss of a discrete band of phalloidin staining in the dentate molecular layer, indicating disappearance of new F-actin at MPP synapses. Introduction of the F-actin stabilizer jasplakinolide during LTP maintenance blocks the reversal of LTP by Arc AS. Interestingly, LTP is

only transiently inhibited when Arc antisense is administered shortly before (5 min) HFS, suggesting that sustained synthesis of Arc during a time-window lasting at least 2 h is necessary to consolidate LTP. These results suggest a pivotal role for activity-induced Arc synthesis in control of synaptic consolidation and local expansion of the actin cytoskeleton. Furthermore, Arc AS infusion blocks BDNF-LTP and reverses maintenance of BDNF-LTP over a critical time-window. Thus, Arc-dependent consolidation is directly activated by local BDNF application. Figure 1 presents a working model of Arc function in LTP. Although the exact role of actin polymerization in synapse expansion is unknown, this process must involve the addition of newly synthesized or translocated proteins to the postsynaptic specialization. Rather than working alone, Arc is likely to

Fig. 1. BDNF as a trigger of synaptic consolidation. The mechanism of stable LTP formation at glutamatergic synapses is presented as a two-stage process: translation activation and Arc-dependent consolidation. In the translation activation stage, patterned highfrequency stimulation leads to sustained postsynaptic release of BDNF and activation of TrkB receptors pre- and postsynaptically. Postsynaptic TrkB leads to (1) rapid activation and translocation of the translation machinery in dendritic spines, and (2) Arc transcription in granule cell bodies. Translation activation is mediated by phosphorylation of the cap-binding protein eIF4E, and possibly by more mRNA-specific mechanisms such as relief of miRNA-mediated translation repression. Spines activated in this way may effectively capture and translate local mRNA pools. In this model, transcripts liberated from local RNA storage granules are translated first, followed by dendritic transport and sustained translation of newly synthesized Arc mRNA. During Arc-dependent consolidation, sustained translation of Arc is necessary for cofilin phosphorylation, local F-actin expansion, and formation of stable LTP. This is a working model; several points require further experimental validation (see text). Adapted from Soule et al. (2006) with permission from the Biochemical Society.

457

be part of a coordinate process of synapse growth involving rapid activity-induced regulation of many of genes. Wibrand et al. (2006) used cDNA microarray expression profiling to screen for genes that are co-upregulated with Arc mRNA 3 h after BDNF-LTP induction in the dentate gyrus of anesthetized rats. Of nine genes upregulated more than fourfold, five were selected for independent confirmation. Real-time PCR confirmed robust upregulation of all five genes: neuronal activity-regulated pentraxin (Narp), neuritin, ADP-ribosylation factor-like protein-4 (ARL4L), TGF-b-induced immediate early gene-1 (TIEG1), and calcium/calmodulin kinase-related peptide (CARP). In situ hybridization histochemistry further revealed upregulation of these genes in somata of postsynaptic granule cells following BDNF-LTP and HFS-LTP. Consistent with a process of synaptic growth and remodeling, Narp, neuritin, and TIEG1 are involved in synaptogenesis, axon guidance, and glutamate receptor clustering during development. Another immediate early gene critical for late phase LTP in the dentate gyrus and long-term memory is zif286, one of the four members of the early growth response (egr) family of zinc-finger transcription factors. Mice with constitutive deletion of zif268 (aka egr1) show impaired LTP maintenance one day after HFS as well as impaired long-term memory consolidation and reconsolidation (Jones et al., 2001; Bozon et al., 2003). Other studies examining the effects of antisense injection into the hippocampus concluded that synthesis of BDNF, but not zif268, is involved in initial memory consolidation (Lee et al., 2004). Conversely, zif268, but not BDNF, synthesis was implicated in reconsolidation. Immediate early gene responses are typically rapid and protein synthesis-independent. In addition to the classical protein synthesis-independent induction, Arc mRNA is induced through a delayed protein synthesis-dependent mechanism. Interestingly, recent work has identified Arc as a direct transcriptional target of zif268 and egr-3 (Li et al., 2005). Protein synthesis-dependent expression of Arc following exploration of a novel environment or seizures is at least partly dependent on egr-transcription factors including zif268.

This raises the intriguing possibility that zif268 stimulates the prolonged increase in Arc induction following HFS-LTP in the dentate gyrus. The fact that zif268 is not induced during BDNF-LTP suggests that Arc-dependent consolidation is independent of zif268 (Messaoudi et al., 2002; Ying et al., 2002; Wibrand et al., 2006). Endogenous BDNF may nonetheless contribute to zif268 expression during HFS-LTP. BDNF gene expression is enhanced following HFS-LTP induction in the CA1 region and dentate gyrus (Patterson et al., 1992; Castren et al., 1993; Bramham et al., 1996; Lee et al., 2005b). LTP at MPP-granule cell synapses of awake rats is associated with NMDA receptor-dependent increases in TrkB and BDNF mRNA expression (Bramham et al., 1996). TrkB mRNA is elevated unilaterally in dentate granule cells 2 h after LTP induction, whereas BDNF mRNA is elevated bilaterally at 6 and 24 h after LTP induction. The bilateral increase in BDNF expression is coupled to unilateral LTP induction, as commissural responses and contralateral MPP-granule cell responses were unaffected. Conceivably, such bilateral effects are due to altered network properties following LTP (Dragoi et al., 2003). Unilateral increases in BDNF gene expression occur in dentate granule cells 3 h after BDNF-LTP induction. Much of this response is due to regulation of the highly calcium-responsive exon III BDNF promoter (Tao et al., 1998; Wibrand et al., 2006). This suggests that BDNF signaling regulates BDNF transcription during long-term synaptic plasticity. However, the key question, whether new BDNF transcription actually contributes to synaptic strengthening, remains to be addressed. It is just as likely that BDNF transcription serves only to replenish depleted stores of protein following intensive BDNF secretion.

BDNF and translation control in dendrites Synaptic consolidation requires transcription, localization, and translation of mRNA. Each of these mechanisms is modulated by BDNF signaling (Klann and Dever, 2004; Richter and Sonenberg, 2005; Bramham and Wells, 2007).

458

The discovery of mRNA, ribosomes, and translation factors within dendrites and even within the spine itself suggested that synapses could be modified directly — and perhaps individually — through regulation of local protein synthesis. While dendritic protein synthesis is regulated by glutamate, acetylcholine, norepinephrine, dopamine, and probably other neurotransmitters, the BDNF-TrkB system stands out as a potentially autonomous modulator of translation at glutamate synapses. In the CA1 region of adult hippocampal slices, BDNF-LTP is induced in synaptic regions isolated from the pyramidal cell somata (Kang and Schuman, 1996). In primary hippocampal neuronal cultures, BDNF induces expression of a reporter construct in isolated dendrites (Aakalu et al., 2001). In synaptodendrosome (SD) preparations, subcellular fractions containing resealed terminal-spine contacts, BDNF rapidly stimulates translation of several mRNAs coupled to LTP or spine morphogenesis, including a-calcium/calmodulin-dependent protein kinase II (aCaMKII), Arc, LIMK1, and GluR1 (Yin et al., 2002; Schratt et al., 2004). Regulation of translation by BDNF has been shown in SDs prepared from forebrain and hippocampus of young (PN 15–22 day) rats (Yin et al., 2002; Schratt et al., 2004; Kanhema et al., 2006), as well as dentate gyrus from 2–4 month old rats (Kanhema et al., 2006).

Translation factor regulation Recent work has examined the biochemical mechanisms by which BDNF modulates local protein synthesis. Phosphorylation of the eukaryotic initiation factor 4E (eIF4E) is considered the rate-limiting step for translation of most mRNAs (those with a 7-methyl-guanosine ‘‘cap’’ at the 50 end). Phosphorylation of eIF4E on Ser209 is correlated with enhanced rates of translation, whereas hypophosphorylation is associated with decreased translation. eIF4E is phosphorylated by mitogenactivated integrating kinase 1 (MNK1), whose activity is regulated by ERK and p38 mitogen activated protein kinase (MAPK). The availability of

eIF4E is controlled by several binding proteins (BPs), most notably eIF4E-binding proteins (4E-BPs). Signaling through receptor-coupled phosphoinositide 3-kinase (PI3K) kinase and mammalian target of rapamycin (mTOR) leads to phosphorylation of 4E-BP and liberation of eIF4E. BDNF stimulates cap-dependent translation in dendrites through TrkB-coupled PI3KmTOR and Ras-ERK (Takei et al., 2001, 2002; Kelleher et al., 2004; Schratt et al., 2004; Takei et al., 2004; Kanhema et al., 2006). Genetic or pharmacological inhibition of mTOR or ERK impairs LTP maintenance and abolishes BDNF-induced LTP (Rosenblum et al., 2002; Tang et al., 2002; Ying et al., 2002; Cammalleri et al., 2003; Kelleher et al., 2004). BDNF-LTP in the dentate gyrus is coupled to transient ERKdependent phosphorylation of eIF4E and enhanced expression of eIF4E protein (Kanhema et al., 2006). Protein synthesis in synaptic plasticity is also controlled at the level of peptide chain elongation. Eukaryotic elongation factor 2 (eEF2) is a GTPbinding protein that mediates translocation of peptidyl-tRNAs from the A-site to the P-site on the ribosome. Phosphorylation of eEF2 on Thr56 inhibits eEF2-ribosome binding and arrests elongation (Browne and Proud, 2002). eEF2 phosphorylation observed during LTP could therefore explain decreased synthesis of some proteins in 2D gel studies. Paradoxically, however, some transcripts important for synaptic plasticity such as Arc and aCaMKII undergo enhanced translation under conditions of eEF2 phosphorylation and global suppression of protein synthesis (Scheetz et al., 2000; Chotiner et al., 2003; Kanhema et al., 2006). BDNF-LTP in vivo is coupled to ERKdependent phosphorylation of eEF2 and enhanced Arc expression in whole dentate gyrus homogenates (Kanhema et al., 2006). Consolidation of taste memory in the neocortex is also associated with ERK activation and transient eEF2 phosphorylation. (Belelovsky et al., 2005). In constrast, BDNF treatment of SDs does not alter eEF2 phosphorylation state, but leads to eIF4E phosphorylation and rapid synthesis of Arc, aCaMKII, LIMK, and other proteins (Yin et al., 2002; Schratt et al., 2004;

459

Kanhema et al., 2006). This suggests a compartment-specific regulation in which initiation is selectively enhanced by BDNF at synapses, while both initiation and elongation are modulated at synaptic sites. eEF2 is phosphorylated by eEF2 kinase, which itself is subject to tight regulation by calcium/ calmodulin, mTOR, ERK, and PKA through multiple phosphorylation sites. Recent studies have shown bidirectional effects of BDNF on eEF2 phosphorylation depending on the preparation studied (Inamura et al., 2005). At synapses, eEF2 is phosphorylated in response to NMDA (Scheetz et al., 2000), but not BDNF treatment (Kanhema et al., 2006). In primary cortical neurons, BDNF induces dephosphorylation of eEF2 and increases elongation rates (Inamura et al., 2005). Serotonin similarly reduces eEF2 phosphorylation in Aplysia synaptosomes. In this case, serotonin appears to offset an increase in eEF2 phosphorylation triggered by calcium influx (Carroll et al., 2004). Thus, the direction of eEF2 phosphorylation is context-dependent and controlled by multiple transmitters. While eEF2 appears to be important in transcript-specific and compartment-specific translation control in synaptic plasticity, further work is needed to resolve these mechanisms. In addition to cap-dependent mechanisms, mRNA translation may be initiated at internal ribosomal entry sites (IRES). A large percentage of total translation persists even when capdependent mechanisms are almost completely abolished. Evidence suggests that IRES-dependent translation is enhanced or maintained in the context of eIF4E hyperphosphorylation and saturation of cap-dependent mechanisms, and during conditions such as ischemia when eIF2a and eEF2 are phosphorylated (Dyer et al., 2003). A number of dendritic mRNA species, including Arc and aCAMKII, contain IRES sequences and are capable of IRES-dependent translation in cultured hippocampal neurons and non-neuronal cell lines (Pinkstaff et al., 2001). However, IRES-mediated translation has not been explored in the context of synaptic plasticity (Fig. 2).

miRNA Another area bound to be important for regulated protein synthesis in neuronal plasticity is microRNAs (miRNAs). miRNAs are small, noncoding RNAs that bind to the 30 UTR of target mRNAs and either block translation or induce transcript degradation (Schratt et al., 2006). As miRNAs generally bind numerous target mRNAs, and most mRNAs are regulated by more than one species of miRNA, the possibilities for fine regulation of translation are enormous. Schratt et al. showed that miR-134, a brainspecific miRNA found in the synaptodendritic compartment, negatively regulates dendritic spine morphogenesis in cultured hippocampal neurons by repressing translation of LIMK1 mRNA. Application of BDNF relieves miR134-mediated repression of LIMK1 translation and promotes spine morphogenesis. Interestingly, the RNA interference complex (RISC) proteins that mediate the effects of miRNA are themselves subject to regulation. Ashraf et al. recently demonstrated that RISC pathway proteins regulate dendritic mRNA transport and local protein synthesis during acquisition of long-term memory in flies (Ashraf et al., 2006). According to their model, training in olfactory/electric shock paradigm triggers proteosomal degradation of the SDE3 helicase Armitage, leading to relief of RISC-mediated suppression of several mRNAs, including aCaMKII, kinesin heavy chain, and staufen. Adding to this complexity, regulation of miRNA transcription may be coupled to specific physiological effects. Impey et al. showed that calcium/cyclic AMP responsive element binding protein (CREB)-dependent transcription of miR132 promotes neuronal morphogenesis by decreasing synthesis of the GTPase activating protein, p250GAP. Like many CREB-dependent effects in neurons, this mechanism was activated by BDNF (Vo et al., 2005). Furthermore, a recent study employing microarray expression profiling and PCR analysis has revealed specific upregulation and downregulation of miRNAs following LTP induction in the dentate gyrus in vivo (Wibrand et al., 2006).

460

Fig. 2. TrkB and translation control in dendritic spines. The cartoon depicts some of the major signaling pathways coupling TrkB to regulation of eIF4E and eEF2. TrkB activation of PI3K-mTOR and Ras-ERK promotes eIF4E phosphorylation and enhances translation initiation. Phosphorylation of eEF2 stalls ribosomes and arrests peptide chain elongation. BDNF-TrkB signaling has bidirectional effects on eEF2 phosphorylation. Adapted from Soule et al. (2006) with permission from the Biochemical Society.

Translocation Translocation and positioning of the translation machinery is another feature of activity-dependent regulation. RNA storage granules in dendrites discharge mRNA in response to strong depolarization (Krichevsky and Kosik, 2001; Kanai et al., 2004). During LTP, ribosomes move from dendritic shafts to spines (Ostroff et al., 2002), and pre-existing a-CaMKII mRNA moves to the synaptodendritic compartment (Havik et al., 2003). BDNF treatment of synaptoneurosomes

also induces a redistribution of eIF4E to an mRNA granule-rich cytoskeletal fraction (Smart et al., 2003).

Model of BDNF-controlled synaptic consolidation To summarize, current evidence supports a model for synaptic consolidation triggered by postsynaptic BDNF signaling and involving regulation of gene expression and local mRNA translation (Fig. 1). According to this model, bursts of

461

excitatory synaptic activity trigger sustained release of BDNF and activation of postsynaptic TrkB receptors. TrkB signaling rapidly activates translation in spines and induces transcription of Arc mRNA in cell bodies. Translational activation in spines consists of global (eIF4E), and probably more mRNA-specific (miRNA), mechanisms. Spines activated in this way may effectively capture and translate mRNAs liberated from local RNA storage granules, as well as newly induced Arc mRNA transport from the soma. Translation of Arc during a critical time-window is necessary for cofilin phosphorylation, local F-actin expansion, and formation of stable LTP. The model predicts that translation of Arc underlies actin polymerization-dependent enlargement of synapses.

Stages in synaptic consolidation: from Arc to PKCz? Arc is likely to interact with other core mechanisms of synaptic consolidation. The list of critical players includes N-cadherin, members of the integrin receptor family, matrix metalloproteinase-9, and the atypical protein kinase C isoform PKMz, (Bahr et al., 1997; Bozdagi et al., 2000; Ling et al., 2002; Chan et al., 2003; Nagy et al., 2006). Of these, PKMz warrants special note. PKMz is constitutively active because it consists of a PKMz catalytic domain, but lacks the autoinhibitory domain of most PKC isoforms. Evidence suggests that persistent PKMz activity is sufficient and necessary for a major component of late phase LTP. Thus, bath application of the myristoylated z-pseudosubstrate inhibititory peptide (ZIP) to hippocampal slices blocks potentiation produced by intracellular perfusion of PKMz and reverses the maintenance of LTP in the CA1 region (Ling et al., 2002; Serrano et al., 2005). A recent study further reported that maintenance of LTP in the dentate gyrus of awake rats is rapidly reversed by intrahippocampal injection of ZIP 24 h after LTP induction (Pastalkova et al., 2006). Moreover, injection of ZIP produces persistent loss of 1-day-old spatial information. LTP is associated with transcription and enhanced dendritic

translation of PKMz, which is thought to sustain LTP by increasing the number of active postsynaptic a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors (AMPAR) (Ling et al., 2006). Interestingly, the F-actin destabilizing agent latrunculin B blocks new synthesis of PKMz and attenuates LTP maintenance (Kelly et al., 2007). Taken together, this suggests a sequential mechanism of LTP maintenance in the dentate gyrus in which Arc-dependent consolidation couples to PKCz-dependent expression at the level of actin polymerization.

Extrinsic modulation of synaptic consolidation Modulatory transmitters such as norepinephrine, serotonin, dopamine, and acetylcholine modulate LTP induction and/or maintenance (Stanton and Sarvey, 1985a, b; Bramham and Srebro, 1989; Frey et al., 1991; Bramham et al., 1997; SwansonPark et al., 1999; Graves et al., 2001; Kulla and Manahan-Vaughan, 2002; Straube and Frey, 2003; Harley et al., 2005). These extrinsic neurotransmitter systems have diffuse, global patterns of innervation, and communicate through spatially dispersed, volume transmission. Neuronal firing of modulatory inputs is a function of the animal’s behavioral or attentional state, with changes in activity dictating the functional modes of target networks (i.e., local rhythmic activity, timing of synaptic events, frequency and duration of action potential firing). The classical modulatory transmitters may also set the biochemical tone in target neurons by modulating PKA and CREB activity. Acetylcholine, dopamine, and norepinephrine have all been found to modulate local protein synthesis in mammalian principal neurons (Feig and Lipton, 1993; Gelinas and Nguyen, 2005; Smith et al., 2005), and serotonergic regulation of local protein synthesis is well known from studies of Aplysia sensory neurons (Casadio et al., 1999). While these actions are spatially dispersed, they are likely to exert broad influence and act in concert with specific mechanisms triggered by glutamate and BDNF signaling. b-adrenergic receptor activation facilitates late LTP through a

462

mechanism involving local protein synthesis, but not transcription (Straube and Frey, 2003; Gelinas and Nguyen, 2005). Locus coeruleus activation induces a delayed protein synthesis-dependent LTP in the dentate gyrus (Walling and Harley, 2004). b-Adrenergic receptor activation is also necessary for protein synthesis underlying noveltyinduced facilitation of late LTP (Straube et al., 2003), and PKA activity is necessary for synaptic tagging and input specificity of late LTP in the CA1 region (Young et al., 2006). One intriguing possibility is that modulatory inputs are able to contract or expand the time-window of synaptic consolidation through effects on local mRNA translation.

Presynaptic mechanisms in synaptic consolidation Quantal analysis and biochemical studies support a contribution of enhanced glutamate release to LTP expression at MPP-granule cell synapses (Min et al., 1998; Errington et al., 2003). Evidence supporting a role for BDNF in enhanced glutamate transmitter release during LTP includes the following: (1) the maintenance phase of BDNFLTP, like HFS-LTP, is associated with a lasting increase in potassium-evoked glutamate release from synaptosomes (Gooney et al., 2004), (2) TrkB receptors are autophosphorylated in synaptosomes collected during the maintenance phase of both HFS- and BDNF-induced LTP (Gooney et al., 2002, 2004), (3) the Trk inhibitor, K252a, blocks the sustained enhancement in neurotransmitter release. Finally, LTP maintenance is associated with enhanced, depolarization-evoked release of BDNF from dentate gyrus tissue (Gooney and Lynch, 2001). The kinetics of these events suggest that BDNF acts rapidly on terminals to trigger a lasting increase in glutamate release. However, more delayed mechanisms involving retrograde nuclear signaling could contribute. The classic hypothesis of target-derived trophic support involves signaling from the nerve terminal to the nucleus (Ginty and Segal, 2002; Delcroix et al., 2003; Campenot and MacInnis, 2004). In sympathetic and sensory neurons, neurotrophin binding to presynaptic Trk

receptors activates retrograde signaling pathways in axons leading to activation of nuclear substrates, such as CREB, and modulation of gene expression. HFS of the perforant pathway leads to CREB phosphorylation in the entorhinal cortex and this effect is blocked by intracerebroventricular application of the Trk inhibitor K252a (Kelly et al., 2000; Gooney and Lynch, 2001). BDNFLTP, induced by local infusion into the dentate gyrus, similarly leads to retrograde activation of CREB in entorhinal cortex located some 4 mm away (Gooney et al., 2004). Such retrograde nuclear signaling could contribute to sustained expansion of presynaptic specializations and enhanced neurotransmitter release. BDNF treatment of cell cultures is associated with enhanced vesicle docking, expression of Rab3a, a small GTP-binding proteins involved in vesicle trafficking, and enhanced expression of SNARE proteins involved in vesicle exocytosis.

Lateral perforant path The lateral perforant input (LPP) projecting from layer II stellate neurons in the lateral entorhinal cortex to the outer-third of the dentate gyrus molecular layer has been less extensively studied in the context of LTP mechanisms. The LPP contains proenkephalin-derived peptides that are released in a frequency-dependent manner. Pharmacological studies show that m and d opioid receptor activation is necessary for LTP induction in the lateral perforant pathway, while antagonists for these receptors have no effect on LTP induction or maintenance in the MPP (Bramham, 1992). Several lines of evidence suggest that opioid peptides facilitate LTP induction by dampening GABAergic inhibition (Madison and Nicoll, 1988; Xie and Lewis, 1991; Hanse and Gustafsson, 1992; Bramham and Sarvey, 1996; Milner and Drake, 2001). Relief from GABAA receptor-mediated inhibition facilitates activation of NMDARs and voltage-dependent calcium channels. In transverse hippocampal slices, LPP LTP and spike-timingdependent LTP and LTD is NMDAR-dependent, whereas de-potentiation and de-depression require group 1 metabotropic glutamate receptor

463

activation (Xie and Lewis, 1991; Hanse and Gustafsson, 1992; Colino and Malenka, 1993; Bramham and Sarvey, 1996; Lin et al., 2006). In anesthetized rats, however, normal LPP LTP is obtained in the presence of NMDAR antagonists that abolish MPP LTP (Bramham et al., 1991). In vivo LTP induction in the LPP input to CA3 and is also opioid receptor-dependent and NMDAR-independent (Do et al., 2002; Kosub et al., 2005). The discrepancy between the hippocampal slice and in vivo preparations with regard to the NMDAR dependence of LPP LTP remains unresolved, although loss of fine inhibitory control in the slice preparation due to transection of the perpendicularly oriented interneurons during slice preparation may favor NMDA receptordependent mechanisms. The role of BDNF and other neurotrophic factors in the LPP is not well known. Studies of paired-pulse plasticity have revealed a remarkable pathway-specificity, such that NT-3 modulates only LPP responses and BDNF modulates only MPP responses (Kokaia et al., 1998; Asztely et al., 2000). Conditional NT-3 mutant mice have deficits in the LPP LTP, but not MPP, LTP induction (Shimazu et al., 2006). The molecular mechanisms of late phase LTP in the LPP have similarly received scant attention. However, the available evidence reveals certain similarities with MPP LTP mechanisms. For example: (1) Arc mRNA localizes selectively in the LPP termination zone following long (>15 min) sessions of repetitive HFS of the LPP (Steward et al., 1998), (2) LPP LTP induced by standard short bursts of HFS is associated with inputspecific accumulation of F-actin (Fukazawa et al., 2003), and (3) Like MPP LTP (Villarreal et al., 2002), stable LPP LTP in awake rats is degraded by NMDA receptor activation (Abraham et al., 2006). This NMDAR-dependent decay of LPP LTP is induced by high-frequency heterosynaptic stimulation of the MPP. Electron microscopic studies of synapse morphometry show that LTP of both the lateral and medial peforant path are associated with increases in the number of unperforated axospinous synapses. However, differences at the ultrastructural level probably exist as the number of perforated axospinous synapses is

elevated 24 h after LTP induction of the MPP, but not LPP (Mezey et al., 2004). Functions and clinical implications of synaptic consolidation in the dentate gyrus Perturbations in dentate gyrus synaptic plasticity are thought to contribute to a range of clinical conditions including memory loss, Alzheimer’s disease, and depression. There are many potential mechanisms involved, including those described above for synaptic consolidation, including BDNF, Arc, and other critical mediators. In addition, the role of newly generated granule cells and the synaptic plasticity of these new cells are likely to contribute. Alzheimer’s disease Modulation of BDNF synthesis would be expected to universally enhance the actions of BDNF. Given the diverse effects of BDNF, even within processes like LTP, it may be advantageous to target specific effector mechanisms such as Arcdependent synaptic consolidation. Recent animal studies connect Arc expression in dentate granule cells to Alzheimer’s disease. In transgenic mice expressing human amyloid precursor protein (hAPP) and hAPP-derived amyloid-b (Ab), Arc expression is reduced primarily in granule cells of the dentate gyrus (Palop et al., 2005). Other neuronal populations expressing hAPP, including CA1 pyramidal cells, show normal basal Arc expression and induction following exposure to a novel environment. In neuronal cell cultures, BDNF-mediated induction of Arc is suppressed even at sublethal levels of Ab (Wang et al., 2006). Downregulation of Arc expression in granule cells may therefore contribute to deficits in LTP and learning observed in rodent models of Alzheimer’s disease (Rowan et al., 2003; Gureviciene et al., 2004; Jacobsen et al., 2006). Although progress has been made in elucidating the computational role of the entorhinal cortex (which provides input to the dentate gyrus through the perforant path) and CA3 region (which receives output from granule cells), the function of

464

the dentate gyrus remains enigmatic. Recent work suggests that the dentate gyrus allows fine spatiotemporal separation of novel and complex cues, thereby disambiguating stimuli to allow sparse encoding of information (Kesner et al., 2004; Lee et al., 2005a). Arc mRNA levels in the dentate gyrus are elevated several hours following LTP induction or exploration of a novel environment, whereas Arc mRNA induction in CA1 region and other brain areas is only short-lived (minutes) (Ramirez-Amaya et al., 2005). It is tempting to speculate that the protracted timewindow of Arc-dependent consolidation is uniquely associated with the function of the dentate gyrus in disambiguating information for encoding in the entorhinal-hippocampal circuitry. Given the pivotal role of ongoing Arc synthesis, it’s possible that this window has a dynamic range that is regulated at the level of Arc mRNA translation. If so, it may prove possible to contract or expand the window of synaptic consolidation therapeutically.

Neurogenesis Among the unique features of the dentate gyrus is the lifelong production of new granule cells from progenitor cells located in the subgranular zone (Gould and Gross, 2002; Aimone et al., 2006). Although the function of neurogenesis is still unclear, it seems likely that the unique contributions of the dentate gyrus are tied to the availability of new granule cells. One hypothesis is that new granule cell populations are selectively engaged in new information storage, the unique associations of successive populations providing a kind of timestamp for episodic memory and recall (Aimone et al., 2006). Recent work indicates that HFS-LTP promotes the survival of newborn granule cells (Bruel-Jungerman et al., 2006), and NMDAR activation promotes integration of these cells within the dentate gyrus circuitry (Tashiro et al., 2006). BDNF-TrkB signaling is similarly implicated in survival of adult born granule cells (Sairanen et al., 2005; Scharfman et al., 2005). Castren has suggested that BDNF promotes the survival and maturation of new connections

through classic mechanism of target-derived support, with active synapses competing for limiting amounts of growth factor (Castren, 2005). Another possibility is that BDNF, as a direct consequence of NMDAR activation, drives synaptic consolidation of nascent contacts as it does at mature synapses.

Depression Alterations in BDNF signaling figure prominently in current theories of depression (Castren, 2004; Monteggia et al., 2004; Berton and Nestler, 2006; Kuipers and Bramham, 2006). Chronic stress leading to depression-like behavior in animals is associated with downregulation of BDNF transcription and TrkB signaling, while chronic antidepressant treatment increases BDNF synthesis reverses the effects of stress on BDNF expression and behavior. Furthermore, the dentate gyrus appears to be key structure in cognitive aspects of depression. First, chronic mild stress leading to anhedonia-like behavior in rats is associated with reductions in BDNF expression and CREB activity in the dentate gyrus, but not hippocampus proper (Gronli et al., 2006). Second, acute bilateral infusion of the BDNF into the dentate gyrus, using a protocol similar to that which induced BDNF-LTP, has antidepressant-like behavioral effects (Shirayama et al., 2002). Third, injection of virus expressing histone deacetylase, which modulates BDNF promoter activity, blocks the behavioral effects of antidepressants (Tsankova et al., 2006). It is noteworthy that several important events associated with BDNF-LTP are also induced by chronic antidepressant treatment. Chronic treatment with the selective serotonin reuptake inhibitor fluoxetine leads to enhanced TrkB signaling, CREB activation, and induction of several BDNF-LTP associated genes including BDNF, Arc, neuritin, TIEG1, and CARP (Alme et al., 2006). Dagestad et al. (2006) recently reported that chronic, but not acute, treatment with fluoxetine results in region-specific changes in translation factor activity. In the dentate gyrus, a pattern of dual eIF4E and eEF2 phosphorylation is seen

465

similar to that observed following BDNF-LTP induction (Dagestad et al., 2006). Taken together this suggests a direct connection between BDNFinduced synaptic strengthening in the dentate gyrus and antidepressant drug action. It will be important to determine whether BDNF-LTP actually develops during antidepressant treatment and whether specific inhibition of BDNF-LTP blocks the behavioral effects of local BDNF infusion.

Abbreviations AMPAR

Arc BDNF a-CaMKII CREB eEF2 eIF4E 4E-BP ERK HFS LIMK LTP Mnk1 mTOR NMDAR PI3K PSD SD

a-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid (AMPA) receptor activity-regulated cytoskeletonassociated protein brain-derived neurotrophic factor calcium/calmodulin-dependent protein kinase II calcium/cyclic AMP responsive element binding protein eukaryotic elongation factor 2 eukaryotic initiation factor 4E eIF4E-binding protein extracellular signal-regulated protein kinase high-frequency stimulation LIM domain kinase long-term potentiation mitogen-activated integrating kinase 1 mammalian target of rapamycin N-methyl-D-aspartate receptor phosphoinositide 3-kinase postsynaptic density synaptodendrosome

Acknowledgments Supported by the Norwegian Research Council and European Union grant 504231.

References Aakalu, G., Smith, W.B., Nguyen, N., Jiang, C. and Schuman, E.M. (2001) Dynamic visualization of local protein synthesis in hippocampal neurons. Neuron, 30(2): 489–502. Abraham, W.C., Mason-Parker, S.E., Irvine, G.I., Logan, B. and Gill, A.I. (2006) Induction and activity-dependent reversal of persistent LTP and LTD in lateral perforant path synapses in vivo. Neurobiol. Learn. Mem., 86(1): 82–90. Aicardi, G., Argilli, E., Cappello, S., Santi, S., Riccio, M., Thoenen, H. and Canossa, M. (2004) Induction of long-term potentiation and depression is reflected by corresponding changes in secretion of endogenous brain-derived neurotrophic factor. Proc. Natl. Acad. Sci. U.S.A., 101(44): 15788–15792. Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for adult neurogenesis in the encoding of time in new memories. Nat. Neurosci., 9(6): 723–727. Alme, M., Wibrand, K., Kuipers, S., Dagestad, G. and Bramham, C.R. (2006) Chronic fluoxetine treatment enhances expression of BDNF-regulated genes in prefrontal cortex. Soc. Neurosci., Abstr., 314.10. Aoki, C., Wu, K., Elste, A., Len, G., Lin, S., McAuliffe, G. and Black, I.B. (2000) Localization of brain-derived neurotrophic factor and TrkB receptors to postsynaptic densities of adult rat cerebral cortex. J. Neurosci. Res., 59(3): 454–463. Ashraf, S.I., McLoon, A.L., Sclarsic, S.M. and Kunes, S. (2006) Synaptic protein synthesis associated with memory is regulated by the RISC pathway in Drosophila. Cell, 124(1): 191–205. Asztely, F., Kokaia, M., Olofsdotter, K., Ortegren, U. and Lindvall, O. (2000) Afferent-specific modulation of shortterm synaptic plasticity by neurotrophins in dentate gyrus. Eur. J. Neurosci., 12(2): 662–669. Bahr, B.A., Staubli, U., Xiao, P., Chun, D., Ji, Z.X., Esteban, E.T. and Lynch, G. (1997) Arg-Gly-Asp-Ser-selective adhesion and the stabilization of long-term potentiation: pharmacological studies and the characterization of a candidate matrix receptor. J. Neurosci., 17(4): 1320–1329. Belelovsky, K., Elkobi, A., Kaphzan, H., Nairn, A.C. and Rosenblum, K. (2005) A molecular switch for translational control in taste memory consolidation. Eur. J. Neurosci., 22(10): 2560–2568. Berton, O. and Nestler, E.J. (2006) New approaches to antidepressant drug discovery: beyond monoamines. Nat. Rev. Neurosci., 7(2): 137–151. Bliss, T.V. and Collingridge, G.L. (1993) A synaptic model of memory: long-term potentiation in the hippocampus. Nature, 361(6407): 31–39. Blum, R. and Konnerth, A. (2005) Neurotrophin-mediated rapid signaling in the central nervous system: mechanisms and functions. Physiology (Bethesda), 20: 70–78. Bozdagi, O., Shan, W., Tanaka, H., Benson, D.L. and Huntley, G.W. (2000) Increasing numbers of synaptic puncta during late-phase LTP: N-cadherin is synthesized, recruited to synaptic sites, and required for potentiation. Neuron, 28(1): 245–259.

466 Bozon, B., Davis, S. and Laroche, S. (2003) A requirement for the immediate early gene zif268 in reconsolidation of recognition memory after retrieval. Neuron, 40(4): 695–701. Bramham, C.R. (1992) Opioid receptor dependent long-term potentiation: peptidergic regulation of synaptic plasticity in the hippocampus [see comments]. Neurochem. Int., 20(4): 441–455. Bramham, C.R., Bacher-Svendsen, K. and Sarvey, J.M. (1997) LTP in the lateral perforant path is beta-adrenergic receptordependent. Neuroreport, 8(3): 719–724. Bramham, C.R. and Messaoudi, E. (2005) BDNF function in adult synaptic plasticity: the synaptic consolidation hypothesis. Prog. Neurobiol., 76: 99–125. Bramham, C.R., Milgram, N.W. and Srebro, B. (1991) Activation of AP5-sensitive NMDA receptors is not required to induce LTP of synaptic transmission in the lateral perforant path. Eur. J. Neurosci., 3: 1300–1308. Bramham, C.R. and Sarvey, J.M. (1996) Endogenous activation of mu and delta-1 opioid receptors is required for longterm potentiation induction in the lateral perforant path: dependence on GABAergic inhibition. J. Neurosci., 16(24): 8123–8131. Bramham, C.R., Southard, T., Sarvey, J.M., Herkenham, M. and Brady, L.S. (1996) Unilateral LTP triggers bilateral increases in hippocampal neurotrophin and trk receptor mRNA expression in behaving rats: evidence for interhemispheric communication. J. Comp. Neurol., 368(3): 371–382. Bramham, C.R. and Srebro, B. (1989) Synaptic plasticity in the hippocampus is modulated by behavioral state. Brain Res., 493: 74–86. Bramham, C.R. and Wells, D.G. (2007) Dendritic mRNA: transport, translation, and function. Nat. Rev. Neurosci. (in press). Browne, G.J. and Proud, C.G. (2002) Regulation of peptidechain elongation in mammalian cells. Eur. J. Biochem., 269(22): 5360–5368. Bruel-Jungerman, E., Davis, S., Rampon, C. and Laroche, S. (2006) Long-term potentiation enhances neurogenesis in the adult dentate gyrus. J. Neurosci., 26(22): 5888–5893. Cammalleri, M., Lutjens, R., Berton, F., King, A.R., Simpson, C., Francesconi, W. and Sanna, P.P. (2003) Time-restricted role for dendritic activation of the mTOR-p70S6K pathway in the induction of late-phase long-term potentiation in the CA1. Proc. Natl. Acad. Sci. U.S.A., 100(24): 14368–14373. Campenot, R.B. and MacInnis, B.L. (2004) Retrograde transport of neurotrophins: fact and function. J. Neurobiol., 58(2): 217–229. Carroll, M., Warren, O., Fan, X. and Sossin, W.S. (2004) 5-HT stimulates eEF2 dephosphorylation in a rapamycin-sensitive manner in Aplysia neuritis. J. Neurochem., 90(6): 1464–1476. Casadio, A., Martin, K.C., Giustetto, M., Zhu, H., Chen, M., Bartsch, D., Bailey, C.H. and Kandel, E.R. (1999) A transient, neuron-wide form of CREB-mediated long-term facilitation can be stabilized at specific synapses by local protein synthesis. Cell, 99(2): 221–237. Castren, E. (2004) Neurotrophic effects of antidepressant drugs. Curr. Opin. Pharmacol., 4(1): 58–64.

Castren, E. (2005) Is mood chemistry? Nat. Rev. Neurosci., 6(3): 241–246. Castren, E., Pitkanen, M., Sirvio, J., Parsadanian, A., Lindholm, D., Thoenen, H. and Riekkinen, P.J. (1993) The induction of LTP increases BDNF and NGF mRNA but decreases NT-3 mRNA in the dentate gyrus. Neuroreport, 4(7): 895–898. Chan, C.S., Weeber, E.J., Kurup, S., Sweatt, J.D. and Davis, R.L. (2003) Integrin requirement for hippocampal synaptic plasticity and spatial memory. J. Neurosci., 23(18): 7107–7116. Chotiner, J.K., Khorasani, H., Nairn, A.C., O’Dell, T.J. and Watson, J.B. (2003) Adenylyl cyclase-dependent form of chemical long-term potentiation triggers translational regulation at the elongation step. Neuroscience, 116(3): 743–752. Colino, A. and Malenka, R.C. (1993) Mechanisms underlying induction of long-term potentiation in rat medial and lateral perforant paths in vitro. J. Neurophysiol., 69(4): 1150–1159. Dagestad, G., Kuipers, S.D., Messaoudi, E. and Bramham, C.R. (2006) Chronic fluoxetine induces region-specific changes in translation factor eIF4E and eEF2 activity in the rat brain. Eur. J. Neurosci., 23: 2814–2818. Delcroix, J.D., Valletta, J.S., Wu, C., Hunt, S.J., Kowal, A.S. and Mobley, W.C. (2003) NGF signaling in sensory neurons: evidence that early endosomes carry NGF retrograde signals. Neuron, 39(1): 69–84. Do, V.H., Martinez, C.O., Martinez Jr., J.L. and Derrick, B.E. (2002) Long-term potentiation in direct perforant path projections to the hippocampal CA3 region in vivo. J. Neurophysiol., 87(2): 669–678. Dragoi, G., Harris, K.D. and Buzsaki, G. (2003) Place representation within hippocampal networks is modified by longterm potentiation. Neuron, 39(5): 843–853. Drake, C.T., Milner, T.A. and Patterson, S.L. (1999) Ultrastructural localization of full-length trkB immunoreactivity in rat hippocampus suggests multiple roles in modulating activity-dependent synaptic plasticity. J. Neurosci., 19(18): 8009–8026. Dyer, J.R., Michel, S., Lee, W., Castellucci, V.F., Wayne, N.L. and Sossin, W.S. (2003) An activity-dependent switch to capindependent translation triggered by eIF4E dephosphorylation. Nat. Neurosci., 6(3): 219–220. Errington, M.L., Galley, P.T. and Bliss, T.V. (2003) Long-term potentiation in the dentate gyrus of the anaesthetized rat is accompanied by an increase in extracellular glutamate: real-time measurements using a novel dialysis electrode. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358(1432): 675–687. Feig, S. and Lipton, P. (1993) Pairing the cholinergic agonist carbachol with patterned Schaffer collateral stimulation initiates protein synthesis in hippocampal CA1 pyramidal cell dendrites via a muscarinic, NMDA-dependent mechanism. J. Neurosci., 13(3): 1010–1021. Figurov, A., Pozzo Miller, L.D., Olafsson, P., Wang, T. and Lu, B. (1996) Regulation of synaptic responses to high-frequency stimulation and LTP by neurotrophins in the hippocampus. Nature, 381(6584): 706–709.

467 Frey, U., Frey, S., Schollmeier, F. and Krug, M. (1996) Influence of actinomycin D, a RNA synthesis inhibitor, on longterm potentiation in rat hippocampal neurons in vivo and in vitro. J. Physiol. Lond., 490(Pt 3): 703–711. Frey, U., Matthies, H. and Reymann, K.G. (1991) The effect of dopaminergic D1 receptor blockade during tetanization on the expression of long-term potentiation in the rat CA1 region in vitro. Neurosci. Lett., 129(1): 111–114. Fukazawa, Y., Saitoh, Y., Ozawa, F., Ohta, Y., Mizuno, K. and Inokuchi, K. (2003) Hippocampal LTP is accompanied by enhanced F-actin content within the dendritic spine that is essential for late LTP maintenance in vivo. Neuron, 38(3): 447–460. Geinisman, Y. (2000) Structural synaptic modifications associated with hippocampal LTP and behavioral learning. Cereb. Cortex, 10(10): 952–962. Gelinas, J.N. and Nguyen, P.V. (2005) Beta-adrenergic receptor activation facilitates induction of a protein synthesis-dependent late phase of long-term potentiation. J. Neurosci., 25(13): 3294–3303. Ginty, D.D. and Segal, R.A. (2002) Retrograde neurotrophin signaling: Trk-ing along the axon. Curr. Opin. Neurobiol., 12(3): 268–274. Gooney, M. and Lynch, M.A. (2001) Long-term potentiation in the dentate gyrus of the rat hippocampus is accompanied by brain-derived neurotrophic factor-induced activation of TrkB. J. Neurochem., 77(5): 1198–1207. Gooney, M., Messaoudi, E., Maher, F.O., Bramham, C.R. and Lynch, M.A. (2004) BDNF-induced LTP in dentate gyrus is impaired with age: analysis of changes in cell signaling events. Neurobiol. Aging, 25(10): 1323–1331. Gooney, M., Shaw, K., Kelly, A., O’Mara, S.M. and Lynch, M.A. (2002) Long-term potentiation and spatial learning are associated with increased phosphorylation of TrkB and extracellular signal-regulated kinase (ERK) in the dentate gyrus: evidence for a role for brain-derived neurotrophic factor. Behav. Neurosci., 116(3): 455–463. Gould, E. and Gross, C.G. (2002) Neurogenesis in adult mammals: some progress and problems. J. Neurosci., 22(3): 619–623. Graves, L., Pack, A. and Abel, T. (2001) Sleep and memory: a molecular perspective. Trends Neurosci., 24(4): 237–243. Gronli, J., Bramham, C.R., Murison, R., Kanhema, T., Fiske, E., Bjorvatn, B., Sorensen, E., Ursin, R. and Portas, C.M. (2006) Chronic mild stress inhibits BDNF expression and CREB activation in the dentate gyrus but not in the hippocampus proper. Pharmacol. Biochem. Behav., 85(4): 842–849. Gureviciene, I., Ikonen, S., Gurevicius, K., Sarkaki, A., van Groen, T., Pussinen, R., Ylinen, A. and Tanila, H. (2004) Normal induction but accelerated decay of LTP in APP+PS1 transgenic mice. Neurobiol. Dis., 15(2): 188–195. Guzowski, J.F., Lyford, G.L., Stevenson, G.D., Houston, F.P., McGaugh, J.L., Worley, P.F. and Barnes, C.A. (2000) Inhibition of activity-dependent arc protein expression in the rat hippocampus impairs the maintenance of long-term

potentiation and the consolidation of long-term memory. J. Neurosci., 20(11): 3993–4001. Guzowski, J.F., Mcnaughton, B.L., Barnes, C.A. and Worley, P.F. (1999) Environment-specific expression of the immediate-early gene Arc in hippocampal neuronal ensembles. Nat. Neurosci., 2(12): 1120–1124. Hanse, E. and Gustafsson, B. (1992) Long-term potentiation and field EPSPs in the lateral and medial perforant paths in the dentate gyrus in vitro: a comparison. Eur. J. Neurosci., 4(11): 1191–1201. Harley, C.W., Walling, S.G. and Brown, R.A.M. (2005) The three faces of norepinephrine: plasticity at the perforant pathdentate gyrus synapse. In: Stanton P., Bramham C. and Scharfman H. (Eds.), Synaptic plasticity and transsynaptic signaling. Spring Science+Business Media, Berlin. Harris, K.M., Fiala, J.C. and Ostroff, L. (2003) Structural changes at dendritic spine synapses during long-term potentiation. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358(1432): 745–748. Havik, B., Rokke, H., Bardsen, K., Davanger, S. and Bramham, C.R. (2003) Bursts of high-frequency stimulation trigger rapid delivery of pre-existing alpha-CaMKII mRNA to synapses: a mechanism in dendritic protein synthesis during long-term potentiation in adult awake rats. Eur. J. Neurosci., 17(12): 2679–2689. Husi, H., Ward, M.A., Choudhary, J.S., Blackstock, W.P. and Grant, S.G. (2000) Proteomic analysis of NMDA receptoradhesion protein signaling complexes. Nat. Neurosci., 3(7): 661–669. Inamura, N., Nawa, H. and Takei, N. (2005) Enhancement of translation elongation in neurons by brain-derived neurotrophic factor: implications for mammalian target of rapamycin signaling. J. Neurochem., 95(5): 1438–1445. Jacobsen, J.S., Wu, C.C., Redwine, J.M., Comery, T.A., Arias, R., Bowlby, M., Martone, R., Morrison, J.H., Pangalos, M.N., Reinhart, P.H. and Bloom, F.E. (2006) Early-onset behavioral and synaptic deficits in a mouse model of Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A., 103(13): 5161–5166. Jones, M.W., Errington, M.L., French, P.J., Fine, A., Bliss, T.V., Garel, S., Charnay, P., Bozon, B., Laroche, S. and Davis, S. (2001) A requirement for the immediate early gene Zif268 in the expression of late LTP and long-term memories. Nat. Neurosci., 4(3): 289–296. Kanai, Y., Dohmae, N. and Hirokawa, N. (2004) Kinesin transports RNA: isolation and characterization of an RNAtransporting granule. Neuron, 43(4): 513–525. Kang, H. and Schuman, E.M. (1996) A requirement for local protein synthesis in neurotrophin-induced hippocampal synaptic plasticity. Science, 273(5280): 1402–1406. Kang, H., Welcher, A.A., Shelton, D. and Schuman, E.M. (1997) Neurotrophins and time: different roles for TrkB signaling in hippocampal long-term potentiation. Neuron, 19(3): 653–664. Kanhema, T., Dagestad, G., Panja, D., Tiron, A., Messaoudi, E., Havik, B., Ying, S.W., Nairn, A.C., Sonenberg, N. and Bramham, C.R. (2006) Dual regulation of translation

468 initiation and peptide chain elongation during BDNF-induced LTP in vivo: evidence for compartment-specific translation control. J. Neurochem., 19: 1328–1337. Kelleher III, R.J., Govindarajan, A., Jung, H.Y., Kang, H. and Tonegawa, S. (2004) Translational control by MAPK signaling in long-term synaptic plasticity and memory. Cell, 116(3): 467–479. Kelly, A., Mullany, P.M. and Lynch, M.A. (2000) Protein synthesis in entorhinal cortex and long-term potentiation in dentate gyrus. Hippocampus, 10(4): 431–437. Kelly, M.T., Yao, Y., Sondhi, R. and Sacktor, T.C. (2007) Actin polymerization regulates the synthesis of PKMzeta in LTP. Neuropharmacology, 52(1): 41–45. Kesner, R.P., Lee, I. and Gilbert, P. (2004) A behavioral assessment of hippocampal function based on a subregional analysis. Rev. Neurosci., 15(5): 333–351. Klann, E. and Dever, T.E. (2004) Biochemical mechanisms for translational regulation in synaptic plasticity. Nat. Rev. Neurosci., 5(12): 931–942. Kokaia, M., Asztely, F., Olofsdotter, K., Sindreu, C.B., Kullmann, D.M. and Lindvall, O. (1998) Endogenous neurotrophin-3 regulates short-term plasticity at lateral perforant path-granule cell synapses. J. Neurosci., 18(21): 8730–8739. Kossel, A.H., Cambridge, S.B., Wagner, U. and Bonhoeffer, T. (2001) A caged Ab reveals an immediate/instructive effect of BDNF during hippocampal synaptic potentiation. Proc. Natl. Acad. Sci. U.S.A., 98(25): 14702–14707. Kosub, K.A., Do, V.H. and Derrick, B.E. (2005) NMDA receptor antagonists block heterosynaptic long-term depression (LTD) but not long-term potentiation (LTP) in the CA3 region following lateral perforant path stimulation. Neurosci. Lett., 374(1): 29–34. Krichevsky, A.M. and Kosik, K.S. (2001) Neuronal RNA granules: a link between RNA localization and stimulationdependent translation. Neuron, 32(4): 683–696. Kuipers, S.D. and Bramham, C.R. (2006) Brain-derived neurotrophic factor mechanisms and function in adult synaptic plasticity: new insights and implications for therapy. Curr. Opin. Drug Discov. Devel., 9(5): 580–586. Kulla, A. and Manahan-Vaughan, D. (2002) Modulation by serotonin 5-HT(4) receptors of long-term potentiation and depotentiation in the dentate gyrus of freely moving rats. Cereb. Cortex, 12(2): 150–162. Lee, I., Hunsaker, M.R. and Kesner, R.P. (2005a) The role of hippocampal subregions in detecting spatial novelty. Behav. Neurosci., 119(1): 145–153. Lee, J.L., Everitt, B.J. and Thomas, K.L. (2004) Independent cellular processes for hippocampal memory consolidation and reconsolidation. Science., 304(5672): 839–843. Lee, P.R., Cohen, J.E., Becker, K.G. and Fields, R.D. (2005b) Gene expression in the conversion of early-phase to latephase long-term potentiation. Ann. N.Y. Acad. Sci., 1048: 259–271. Li, L., Carter, J., Gao, X., Whitehead, J. and Tourtellotte, W.G. (2005) The neuroplasticity-associated arc gene is a direct transcriptional target of early growth response (Egr) transcription factors. Mol. Cell Biol., 25(23): 10286–10300.

Lin, Y.W., Yang, H.W., Wang, H.J., Gong, C.L., Chiu, T.H. and Min, M.Y. (2006) Spike-timing-dependent plasticity at resting and conditioned lateral perforant path synapses on granule cells in the dentate gyrus: different roles of N-methylD-aspartate and group I metabotropic glutamate receptors. Eur. J. Neurosci., 23(9): 2362–2374. Ling, D.S., Benardo, L.S. and Sacktor, T.C. (2006) Protein kinase Mzeta enhances excitatory synaptic transmission by increasing the number of active postsynaptic AMPA receptors. Hippocampus, 16(5): 443–452. Ling, D.S., Benardo, L.S., Serrano, P.A., Blace, N., Kelly, M.T., Crary, J.F. and Sacktor, T.C. (2002) Protein kinase Mzeta is necessary and sufficient for LTP maintenance. Nat. Neurosci., 5(4): 295–296. Link, W., Konietzko, U., Kauselmann, G., Krug, M., Schwanke, B., Frey, U. and Kuhl, D. (1995) Somatodendritic expression of an immediate early gene is regulated by synaptic activity. Proc. Natl. Acad. Sci. U.S.A., 92(12): 5734–5738. Lisman, J.E. and Zhabotinsky, A.M. (2001) A model of synaptic memory: a CaMKII/PP1 switch that potentiates transmission by organizing an AMPA receptor anchoring assembly. Neuron, 31(2): 191–201. Lyford, G.L., Yamagata, K., Kaufmann, W.E., Barnes, C.A., Sanders, L.K., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., Lanahan, A.A. and Worley, P.F. (1995) Arc, a growth factor and activity-regulated gene, encodes a novel cytoskeletonassociated protein that is enriched in neuronal dendrites. Neuron, 14(2): 433–445. Madison, D.V. and Nicoll, R.A. (1988) Enkephalin hyperpolarizes interneurones in the rat hippocampus. J. Physiol., 398: 123–130. Matsuzaki, M., Honkura, N., Ellis-Davies, G.C. and Kasai, H. (2004) Structural basis of long-term potentiation in single dendritic spines. Nature, 429(6993): 761–766. Meng, Y., Zhang, Y., Tregoubov, V., Janus, C., Cruz, L., Jackson, M., Lu, W.Y., MacDonald, J.F., Wang, J.Y., Falls, D.L. and Jia, Z. (2002) Abnormal spine morphology and enhanced LTP in LIMK-1 knockout mice. Neuron, 35(1): 121–133. Messaoudi, E., Bardsen, K., Srebro, B. and Bramham, C.R. (1998) Acute intrahippocampal infusion of BDNF induces lasting potentiation of synaptic transmission in the rat dentate gyrus. J. Neurophysiol., 79(1): 496–499. Messaoudi, E., Kanhema, T., Soule, J., Tiron, A., Dagyte, G., da Silva, B. and Bramham, C.R. (2007) Sustained Arc synthesis controls LTP consolidation through regulation of local actin polymerization in the dentate gyrus in vivo. Submitted. Messaoudi, E., Ying, S.W., Kanhema, T., Croll, S.D. and Bramham, C.R. (2002) BDNF triggers transcription-dependent, late phase LTP in vivo. J. Neurosci., 22: 7453–7461. Mezey, S., Doyere, V., De Souza, I., Harrison, E., Cambon, K., Kendal, C.E., Davies, H., Laroche, S. and Stewart, M.G. (2004) Long-term synaptic morphometry changes after induction of long-term potentiation and long-term depression in the dentate gyrus of awake rats are not simply mirror phenomena. Eur. J. Neurosci., 19(8): 2310–2318.

469 Milner, T.A. and Drake, C.T. (2001) Ultrastructural evidence for presynaptic mu opioid receptor modulation of synaptic plasticity in NMDA-receptor-containing dendrites in the dentate gyrus. Brain Res. Bull., 54(2): 131–140. Min, M.Y., Asztely, F., Kokaia, M. and Kullmann, D.M. (1998) Long-term potentiation and dual-component quantal signaling in the dentate gyrus. Proc. Natl. Acad. Sci. U.S.A., 95(8): 4702–4707. Moga, D.E., Calhoun, M.E., Chowdhury, A., Worley, P., Morrison, J.H. and Shapiro, M.L. (2004) Activity-regulated cytoskeletal-associated protein is localized to recently activated excitatory synapses. Neuroscience, 125(1): 7–11. Monteggia, L.M., Barrot, M., Powell, C.M., Berton, O., Galanis, V., Gemelli, T., Meuth, S., Nagy, A., Greene, R.W. and Nestler, E.J. (2004) Essential role of brain-derived neurotrophic factor in adult hippocampal function. Proc. Natl. Acad. Sci. U.S.A., 101(29): 10827–10832. Nagy, V., Bozdagi, O., Matynia, A., Balcerzyk, M., Okulski, P., Dzwonek, J., Costa, R.M., Silva, A.J., Kaczmarek, L. and Huntley, G.W. (2006) Matrix metalloproteinase-9 is required for hippocampal late-phase long-term potentiation and memory. J. Neurosci., 26(7): 1923–1934. Nestler, E.J., Barrot, M., DiLeone, R.J., Eisch, A.J., Gold, S.J. and Monteggia, L.M. (2002) Neurobiology of depression. Neuron, 34(1): 13–25. Nguyen, P.V. and Kandel, E.R. (1996) A macromolecular synthesis-dependent late phase of long-term potentiation requiring cAMP in the medial perforant pathway of rat hippocampal slices. J. Neurosci., 16(10): 3189–3198. Oertner, T.G. and Matus, A. (2005) Calcium regulation of actin dynamics in dendritic spines. Cell Calcium, 37(5): 477–482. Okamoto, K., Nagai, T., Miyawaki, A. and Hayashi, Y. (2004) Rapid and persistent modulation of actin dynamics regulates postsynaptic reorganization underlying bidirectional plasticity. Nat. Neurosci., 7(10): 1104–1112. Ostroff, L.E., Fiala, J.C., Allwardt, B. and Harris, K.M. (2002) Polyribosomes redistribute from dendritic shafts into spines with enlarged synapses during LTP in developing rat hippocampal slices. Neuron, 35(3): 535–545. Palop, J.J., Chin, J., Bien-Ly, N., Massaro, C., Yeung, B.Z., Yu, G.Q. and Mucke, L. (2005) Vulnerability of dentate granule cells to disruption of arc expression in human amyloid precursor protein transgenic mice. J. Neurosci., 25(42): 9686–9693. Pastalkova, E., Serrano, P., Pinkhasova, D., Wallace, E., Fenton, A.A. and Sacktor, T.C. (2006) Storage of spatial information by the maintenance mechanism of LTP. Science, 313(5790): 1141–1144. Patterson, S.L., Grover, L.M., Schwartzkroin, P.A. and Bothwell, M. (1992) Neurotrophin expression in rat hippocampal slices: a stimulus paradigm inducing LTP in CA1 evokes increases in BDNF and NT-3 mRNAs. Neuron, 9(6): 1081–1088. Pinkstaff, J.K., Chappell, S.A., Mauro, V.P., Edelman, G.M. and Krushel, L.A. (2001) Internal initiation of translation of five dendritically localized neuronal mRNAs. Proc. Natl. Acad. Sci. U.S.A., 98(5): 2770–2775.

Plath, N., Ohana, O., Dammermann, B., Errington, M.L., Schmitz, D., Gross, C., Mao, X., Engelsberg, A., Mahlke, C., Welzl, H., Kobalz, U., Stawrakakis, A., Fernandez, E., Waltereit, R., Bick-Sander, A., Therstappen, E., Cooke, S.F., Blanquet, V., Wurst, W., Salmen, B., Bosl, M.R., Lipp, H.P., Grant, S.G., Bliss, T.V., Wolfer, D.P. and Kuhl, D. (2006) Arc/Arg3.1 is essential for the consolidation of synaptic plasticity and memories. Neuron, 52(3): 437–444. Ramirez-Amaya, V., Vazdarjanova, A., Mikhael, D., Rosi, S., Worley, P.F. and Barnes, C.A. (2005) Spatial explorationinduced Arc mRNA and protein expression: evidence for selective, network-specific reactivation. J. Neurosci., 25(7): 1761–1768. Richter, J.D. and Sonenberg, N. (2005) Regulation of cap-dependent translation by eIF4E inhibitory proteins. Nature, 433(7025): 477–480. Rodriguez, J.J., Davies, H.A., Silva, A.T., De Souza, I.E., Peddie, C.J., Colyer, F.M., Lancashire, C.L., Fine, A., Errington, M.L., Bliss, T.V. and Stewart, M.G. (2005) Longterm potentiation in the rat dentate gyrus is associated with enhanced Arc/Arg3.1 protein expression in spines, dendrites and glia. Eur. J. Neurosci., 21(9): 2384–2396. Rosenblum, K., Futter, M., Voss, K., Erent, M., Skehel, P.A., French, P., Obosi, L., Jones, M.W. and Bliss, T.V. (2002) The role of extracellular regulated kinases I/II in late-phase longterm potentiation. J. Neurosci., 22(13): 5432–5441. Routtenberg, A. and Rekart, J.L. (2005) Post-translational protein modification as the substrate for long-lasting memory. Trends Neurosci., 28(1): 12–19. Rowan, M.J., Klyubin, I., Cullen, W.K. and Anwyl, R. (2003) Synaptic plasticity in animal models of early Alzheimer’s disease. Philos. Trans. R. Soc. Lond. B Biol. Sci., 358(1432): 821–828. Sairanen, M., Lucas, G., Ernfors, P., Castren, M. and Castren, E. (2005) Brain-derived neurotrophic factor and antidepressant drugs have different but coordinated effects on neuronal turnover, proliferation, and survival in the adult dentate gyrus. J. Neurosci., 25(5): 1089–1094. Sanes, J.R. and Lichtman, J.W. (1999) Can molecules explain long-term potentiation? Nat. Neurosci., 2(7): 597–604. Santi, S., Cappello, S., Riccio, M., Bergami, M., Aicardi, G., Schenk, U., Matteoli, M. and Canossa, M. (2005) Hippocampal neurons recycle BDNF for activity-dependent secretion and LTP maintenance. EMBO J., 25(18): 4372–4380. Scharfman, H., Goodman, J., Macleod, A., Phani, S., Antonelli, C. and Croll, S. (2005) Increased neurogenesis and the ectopic granule cells after intrahippocampal BDNF infusion in adult rats. Exp. Neurol., 192(2): 348–356. Scheetz, A.J., Nairn, A.C. and Constantine-Paton, M. (2000) NMDA receptor-mediated control of protein synthesis at developing synapses. Nat. Neurosci., 3(3): 211–216. Schratt, G.M., Nigh, E.A., Chen, W.G., Hu, L. and Greenberg, M.E. (2004) BDNF regulates the translation of a select group of mRNAs by a mammalian target of rapamycin-phosphatidylinositol 3-kinase-dependent pathway during neuronal development. J. Neurosci., 24(33): 7366–7377.

470 Schratt, G.M., Tuebing, F., Nigh, E.A., Kane, C.G., Sabatini, M.E., Kiebler, M. and Greenberg, M.E. (2006) A brainspecific microRNA regulates dendritic spine development. Nature, 439(7074): 283–289. Serrano, P., Yao, Y. and Sacktor, T.C. (2005) Persistent phosphorylation by protein kinase Mzeta maintains late-phase long-term potentiation. J. Neurosci., 25(8): 1979–1984. Shirayama, Y., Chen, A.C., Nakagawa, S., Russell, D.S. and Duman, R.S. (2002) Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depression. J. Neurosci., 22(8): 3251–3261. Shimazu, K., Zhao, M., Sakata, K., Akbarian, S., Bates, B., Jaenisch, R. and Lu, B. (2006) NT-3 facilitates hippocampal plasticity and learning and memory by regulating neurogenesis. Learn. Mem., 13(3): 307–315. Smart, F.M., Edelman, G.M. and Vanderklish, P.W. (2003) BDNF induces translocation of initiation factor 4E to mRNA granules: evidence for a role of synaptic microfilaments and integrins. Proc. Natl. Acad. Sci. U.S.A., 100(24): 14403–14408. Smith, W.B., Starck, S.R., Roberts, R.W. and Schuman, E.M. (2005) Dopaminergic stimulation of local protein synthesis enhances surface expression of GluR1 and synaptic transmission in hippocampal neurons. Neuron, 45(5): 765–779. Soule, J., Messaoudi, E. and Bramham, C.R. (2006) Brainderived neurotrophic factor and control of synaptic consolidation in the adult brain. Biochem. Soc. Trans., 34(Pt 4): 600–604. Stanton, P.K. and Sarvey, J.M. (1985a) Blockade of norepinephrine-induced long-lasting potentiation in the hippocampal dentate gyrus by an inhibitor of protein synthesis. Brain Res., 361(1–2): 276–283. Stanton, P.K. and Sarvey, J.M. (1985b) Depletion of norepinephrine, but not serotonin, reduces long- term potentiation in the dentate gyrus of rat hippocampal slices. J. Neurosci., 5(8): 2169–2176. Steward, O., Wallace, C.S., Lyford, G.L. and Worley, P.F. (1998) Synaptic activation causes the mRNA for the IEG Arc to localize selectively near activated postsynaptic sites on dendrites. Neuron, 21(4): 741–751. Steward, O. and Worley, P.F. (2001) A cellular mechanism for targeting newly synthesized mRNAs to synaptic sites on dendrites. Proc. Natl. Acad. Sci. U.S.A., 98(13): 7062–7068. Straube, T. and Frey, J.U. (2003) Involvement of beta-adrenergic receptors in protein synthesis-dependent late longterm potentiation (LTP) in the dentate gyrus of freely moving rats: the critical role of the LTP induction strength. Neuroscience, 119(2): 473–479. Straube, T., Korz, V., Balschun, D. and Frey, J.U. (2003) Requirement of beta-adrenergic receptor activation and protein synthesis for LTP-reinforcement by novelty in rat dentate gyrus. J. Physiol., 552(Pt 3): 953–960. Swanson-Park, J.L., Coussens, C.M., Mason-Parker, S.E., Raymond, C.R., Hargreaves, E.L., Dragunow, M., Cohen,

A.S. and Abraham, W.C. (1999) A double dissociation within the hippocampus of dopamine D1/D5 receptor and beta-adrenergic receptor contributions to the persistence of long-term potentiation. Neuroscience, 92(2): 485–497. Takei, N., Inamura, N., Kawamura, M., Namba, H., Hara, K., Yonezawa, K. and Nawa, H. (2004) Brain-derived neurotrophic factor induces mammalian target of rapamycin-dependent local activation of translation machinery and protein synthesis in neuronal dendrites. J. Neurosci., 24(44): 9760–9769. Takei, N., Kawamura, M., Hara, K., Yonezawa, K. and Nawa, H. (2001) Brain-derived neurotrophic factor enhances neuronal translation by activating multiple initiation processes: comparison with the effects of insulin. J. Biol. Chem., 276(46): 42818–42825. Tang, S.J., Reis, G., Kang, H., Gingras, A.C., Sonenberg, N. and Schuman, E.M. (2002) A rapamycin-sensitive signaling pathway contributes to long-term synaptic plasticity in the hippocampus. Proc. Natl. Acad. Sci. U.S.A., 99(1): 467–472. Tao, X., Finkbeiner, S., Arnold, D.B., Shaywitz, A.J. and Greenberg, M.E. (1998) Ca2+ influx regulates BDNF transcription by a CREB family transcription factor-dependent mechanism. Neuron, 20(4): 709–726. Tashiro, A., Sandler, V.M., Toni, N., Zhao, C. and Gage, F.H. (2006) NMDA-receptor-mediated, cell-specific integration of new neurons in adult dentate gyrus. Nature, 442(7105): 929–933. Tsankova, N.M., Berton, O., Renthal, W., Kumar, A., Neve, R.L. and Nestler, E.J. (2006) Sustained hippocampal chromatin regulation in a mouse model of depression and antidepressant action. Nat. Neurosci., 9(4): 519–525. Vazdarjanova, A., Ramirez-Amaya, V., Insel, N., Plummer, T.K., Rosi, S., Chowdhury, S., Mikhael, D., Worley, P.F., Guzowski, J.F. and Barnes, C.A. (2006) Spatial exploration induces ARC, a plasticity-related immediate-early gene, only in calcium/calmodulin-dependent protein kinase II-positive principal excitatory and inhibitory neurons of the rat forebrain. J. Comp. Neurol., 498(3): 317–329. Villarreal, D.M., Do, V., Haddad, E. and Derrick, B.E. (2002) NMDA receptor antagonists sustain LTP and spatial memory: active processes mediate LTP decay. Nat. Neurosci., 5(1): 48–52. Vo, N., Klein, M.F., Varlamova, O., Keller, D.M., Yamamoto, T., Goodman, R.H. and Impey, S. (2005) A cAMP-response element binding protein-induced microRNA regulates neuronal morphogenesis. Proc. Natl. Acad. Sci. U.S.A., 102(45): 16426–16431. Walling, S.G. and Harley, C.W. (2004) Locus ceruleus activation initiates delayed synaptic potentiation of perforant path input to the dentate gyrus in awake rats: a novel beta-adrenergic- and protein synthesis-dependent mammalian plasticity mechanism. J. Neurosci., 24(3): 598–604. Wang, D.C., Chen, S.S., Lee, Y.C. and Chen, T.J. (2006) Amyloid-beta at sublethal level impairs BDNF-induced arc expression in cortical neurons. Neurosci. Lett., 398(1–2): 78–82. Weeks, A.C., Ivanco, T.L., LeBoutillier, J.C., Racine, R.J. and Petit, T.L. (2001) Sequential changes in the synaptic

471 structural profile following long-term potentiation in the rat dentate gyrus: III. Long-term maintenance phase. Synapse, 40(1): 74–84. Wibrand, K., Messaoudi, E., Havik, B., Steenslid, V., Løvlie, R., Steen, V.M. and Bramham, C.R. (2006) Identification of genes co-upregulated with Arc during BDNF-induced longterm potentiation in the adult rat dentate gyrus in vivo. Eur. J. Neurosci., 23: 1501–1511. Wu, K., Xu, J.L., Suen, P.C., Levine, E., Huang, Y.Y., Mount, H.T., Lin, S.Y. and Black, I.B. (1996) Functional trkB neurotrophin receptors are intrinsic components of the adult brain postsynaptic density. Brain Res. Mol. Brain Res., 43(1–2): 286–290. Xie, C.W. and Lewis, D.V. (1991) Opioid-mediated facilitation of long-term potentiation at the lateral perforant path-dentate granule cell synapse. J. Pharmacol. Exp. Ther., 256(1): 289–296. Yin, Y., Edelman, G.M. and Vanderklish, P.W. (2002) The brain-derived neurotrophic factor enhances synthesis of Arc

in synaptoneurosomes. Proc. Natl. Acad. Sci. U.S.A., 99(4): 2368–2373. Ying, S.W., Futter, M., Rosenblum, K., Webber, M.J., Hunt, S.P., Bliss, T.V. and Bramham, C.R. (2002) Brain-derived neurotrophic factor induces long-term potentiation in intact adult hippocampus: requirement for ERK activation coupled to CREB and upregulation of Arc synthesis. J. Neurosci., 22(5): 1532–1540. Young, J.Z., Isiegas, C., Abel, T. and Nguyen, P.V. (2006) Metaplasticity of the late-phase of long-term potentiation: a critical role for protein kinase A in synaptic tagging. Eur. J. Neurosci., 23(7): 1784–1794. Zhou, Q., Xiao, M. and Nicoll, R.A. (2001) Contribution of cytoskeleton to the internalization of AMPA receptors. Proc. Natl. Acad. Sci. U.S.A., 98(3): 1261–1266. Zito, K., Knott, G., Shepherd, G.M., Shenolikar, S. and Svoboda, K. (2004) Induction of spine growth and synapse formation by regulation of the spine actin cytoskeleton. Neuron, 44(2): 321–334.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 26

Comparison of cellular mechanisms of long-term depression of synaptic strength at perforant path–granule cell and Schaffer collateral–CA1 synapses Beatrice Po¨schel and Patric K. Stanton Department of Cell Biology and Anatomy, New York Medical College, Valhalla, NY 10595, USA

Abstract: This chapter compares the cellular mechanisms that have been implicated in the induction and expression of long-term depression (LTD) at Schaffer collateral–CA1 synapses to perforant path-dentate gyrus (DG) synapses. In general, Schaffer collateral LTD and long-term potentiation (LTP) both appear to be a complex combination of many alterations in synaptic transmission that occur at both presynaptic and postsynaptic sites, while at perforant path synapses, most evidence has focused on postsynaptic long-term alterations. Within the DG, the medial perforant path is far more studied than lateral perforant path synapses, where most evidence relates to the induction of heterosynaptic LTD at lateral perforant path synapses when LTP is induced in the medial perforant path. Of course, there remain many other classes of synapses in the DG where synaptic plasticity, including LTD, have been largely neglected. It is clear that a better understanding of the range of DG loci where long-lasting activity-dependent plasticity, both LTD and LTP, are expressed will be essential to improve our understanding of the cognitive roles of such DG plasticity. Keywords: CA1; dentate gyrus; hippocampus; long-term depression; perforant path; plasticity; Schaffer collateral Kobayashi et al., 1996), as well as many other synapses and regions. The cellular mechanisms for induction and expression of LTD can vary depending on the stimulation protocol, brain subregion, and stage of development (Domenici et al., 1998; Kemp et al., 2000; Malenka and Bear, 2004). The majority of studies elucidating the cellular mechanisms of hippocampal LTD have been performed in area CA1. Studies of LTD in area CA1 will be reviewed to develop a model of how LTD can be induced and expressed, to be used as a framework to compare LTD mechanisms in CA1 with those in the DG.

Hippocampal long-term depression (LTD) LTD is defined as a long lasting (hours to weeks), usually synapse specific, decrease of the strength of synaptic transmission. It appears to be the counterpart of long-term potentiation (LTP), a longlasting increase in synaptic strength. LTD can be induced at synapses in the hippocampal subregions CA1, CA3, and the dentate gyrus (DG) (Dudek and Bear, 1992; O’Mara et al., 1995a; Corresponding author. Tel.: +001-914-594-4883; Fax: +001-914-594-4653; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63026-X

473

474

LTD in area CA1 (Schaffer collateral– CA1 synapses) LTD can be divided into two phases: induction and expression. The induction phase involves cellular biochemical events which are triggered by particular patterns of synaptic activity within a timeframe of minutes, and which initiate longerlasting processes leading to the expression of LTD.

Induction LTD can be induced by different means, including both electrical stimulation and pharmacological manipulations. While it is often mistakenly assumed that these forms of LTD are similar mechanistically within and across brain regions, evidence is substantial that multiple forms of LTD with distinct induction and expression mechanisms coexist within single synapses, and that the profiles of expression of these forms can differ greatly between synapses. Most commonly, LTD is induced by a prolonged (10–15 min) train of low-frequency (1–2 Hz) stimulation (LFS; Dudek and Bear, 1992). This form of LTD is called homosynaptic because only the stimulated pathway exhibits a long-term change in synaptic strength. Heterosynaptic LTD involves interactions between separate synapses or groups of synapses that converge on the same neuron. In CA1, heterosynaptic LTD can be induced in a non-stimulated pathway by the delivery of LTPinducing stimuli to a parallel pathway that converges on the same neuron (Lynch et al., 1977). One computationally interesting form of heterosynaptic LTD is associative LTD (Stanton and Sejnowski, 1989). For associative forms of synaptic plasticity, the relative timing of activity in two (or more) input pathways is a key determinant of the direction of change in synaptic strength, i.e., whether LTP or LTD is induced. During induction of associative LTD, the activity of a test input is negatively correlated in time with activity of a conditioning train of bursts applied to a separate pathway, which leads to LTD in the test input. That is, if a series of brief high-frequency bursts of stimuli are applied to the conditioning input, and the test input is

stimulated between the bursts, then the test input expresses LTD. In a closely-related form of LTD called spike-timing LTD, repeated application of single presynaptic stimuli that closely follow single postsynaptic action potentials (APs) elicits LTD, while reversing the pairing (i.e., pre-AP) elicits LTP (Magee and Johnston, 1997; Markram et al., 1997). Induction of LTD by pharmacological agents allows for selective targeting of particular receptors that activate different induction pathways. Thus, particular signal cascades can be investigated with regard to their contribution to the induction of potentially separable forms of LTD that might nevertheless share similar expression mechanisms. In area CA1, LTD can be induced by agonist activation of a number of transmitter receptors (metabotropic glutamate receptors (mGluRs): Manahan-Vaughan and Reymann, 1995; N-methyl-D-aspartate receptors (NMDAR): Lee et al., 1998; Kamal et al., 1999) or activation/inactivation of intracellular messengers (i.e., simultaneous activation of protein kinase G (PKG) and inhibition of protein kinase A (PKA): Santschi et al., 1999; Bailey et al., 2003). The intracellular induction mechanisms of LTD vary depending on the brain region, developmental stage and triggering mechanism (Malenka and Bear, 2004). Different forms of LTD can coexist in one region, as in area CA1 (Kemp and Bashir, 1997; Manahan-Vaughan, 1997; Oliet et al., 1997; Nicoll et al., 1998), as well as other brain regions. The main forms of LTD at Schaffer collateral–CA1 synapses are NMDAR dependent and mGluRdependent LTD, which will be discussed below. After LTD-inducing synaptic stimulation, NMDAR- and mGluR-dependent LTD are induced simultaneously in most cases. They can be separated by selective pharmacological blockade of NMDARs or mGluRs, respectively, or selectively induced by agonist application.

NMDAR/mGluR-dependent LTD During initial investigations of LTD in CA1, LTD was elicited via a protocol of LFS trains (Dudek and Bear, 1992). It turned out that this stimulation induced a mixture of NMDAR- and mGluRdependent LTDs. Thus, studies of the intracellular

475

Fig. 1. Cascades involved with NMDAR/mGluR-dependent LTD at Schaffer collateral–CA1 synapses. AA: arachidonic acid, AC: adenylyl cyclase, Ca: Ca2+, cADPR: cyclic adenosine diphosphate ribose, CaM: calmodulin, CaMKII: calmodulin-dependent protein kinase II, cAMP: cyclic adenosine monophosphate, cGMP: cyclic guanosine monophosphate, CN: calcineurin, DAG: diacyl glycerol, HPETE: hydroperoxyeicosatetraenoic acid, I1: inhibitor 1, IP3: inositol-1,4,5-trisphosphate, IEG: immediate early genes, MAPK: mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, NO: nitric oxide, NOS: NO-synthetase, P: phosphate, PIP2: phosphatidylinositol-3,4-bisphosphate, PLA2: phospholipase A2, PKA: protein kinase A, PKC: protein kinase C, PKG: protein kinase G, PLC: phospholipase C, Post: postsynapse, PP1: protein phosphatase 1, Pre: presynapse, PS: protein synthesis, VGCC: voltage-gated Ca2+ channels.

mechanisms of stimulus-induced LTD did not distinguish between NMDAR- and mGluR-dependent pathways. Later it was found that selective activation of NMDARs and mGluRs can induce LTD independently, and that NMDAR- and mGluR-dependent LTD pathways could be studied independently. These studies will be reviewed below (Fig. 1). Induction. The first step in the induction of LTD is the activation of glutamate receptors and/or voltage-gated Ca2+ channels (VGCCs). NMDARs have been a primary focus of attention, because their requirement for membrane depolarization to relieve Mg2+-dependent channel block plus

glutamate to activate the receptor allows them to function as coincidence detectors of presynaptic transmitter release and postsynaptic depolarization (Cotman et al., 1988), and because they gate the influx of Ca2+, they are the key factors in establishing associative long-term changes (both LTP and LTD) of synaptic strength. It has been shown that application of NMDAR antagonists block the induction of LTD by prolonged LFS (1 Hz/15 min; Dudek and Bear, 1992; Mulkey and Malenka, 1992). Other studies have revealed that activation of mGluRs (Oliet et al., 1997; Manahan-Vaughan, 1997) and VGCCs (Christie et al., 1995; Cummings et al., 1996; Oliet et al., 1997) can also be necessary

476

for the induction of LTD in CA1, and the dependence on these receptors varies with the precise stimulus pattern and level of postsynaptic depolarization. Activation of NMDARs, mGluRs, or VGCCs each lead to elevations in intracellular [Ca2+]. Ca2+ is a key ion that triggers a variety of kinases and phosphatases, some mediating intracellular events, which lead to both short- and longterm changes in synaptic strength. During the induction of LTD, the rise in [Ca2+] is lower than that occurring during induction of LTP (Hansel et al., 1996; Yang et al., 1999), which is thought to lead to a preferred activation of phosphatases with lower Kd’s for activation than most Ca2+-dependent kinases (Lisman, 1989). Besides phosphatases (Mulkey et al., 1993; Thiels et al., 1998), LTD has been shown to also depend on multiple kinases. It has been reported that protein kinase C (PKC) (Stanton, 1995; Hrabetova and Sacktor, 1996), PKG (Reyes-Harde et al., 1999a, b), PKA (Brandon et al., 1995), Ca2+/calmodulin-dependent protein kinase II (CaMKII) (Stevens et al., 1994; Stanton and Gage, 1996), and mitogen-activated protein kinase (MAPK) (Thiels et al., 2002; Gallagher et al., 2004) can all play roles in the induction of LTD in area CA1. In some cases, these roles appear to be played presynaptically, and in others postsynaptically, reinforcing the notion of multiple forms of LTD. Expression. Phosphatases and kinases target several effector molecules that could cause changes in synaptic strength via effects on a wide range of protein mechanisms, such as receptor number and sensitivity, ion-channel open probability and conductance, cytoskeletal machinery, mRNA and protein synthesis (PS), and vesicle-release probability. (a) Phosphatase pathways Phosphatases whose activation has been suggested to be involved in the induction of LTD are protein phosphatase 2A (PP2A), calcineurin (CN; also, PP2B), and PP1. Inhibition of these phosphatases with the PP1/PP2A inhibitors okadaic acid and calyculin A (Mulkey et al., 1993) and CN inhibitors FK506 and cyclosporin A (Mulkey et al., 1994) have all been demonstrated to block the induction of LTD. Additionally, it has been shown that the induction of LTD in CA1 in vivo is

associated with an increase in PP1 activity lasting 35 min after LTD induction, as well as PP2A activity increases lasting 65 min (Thiels et al., 1998, 2000). CN is a phosphatase that is dependent on Ca2+ levels and is activated by Ca2+/CaM (Klee et al., 1979). Amongst its effects, CN has been demonstrated to dephosphorylate synapsin I (Hosaka et al., 1999), resulting in a gradual, activity-dependent reduction of neurotransmitter release, since vesicles are held in the reserve pool by dephoshorylated synapsin and are prevented from joining the readily releasa ble vesicle pool (RRP) (Hilfiker et al., 1999). CN also regulates the internalization of both alphaamino-3-hydroxy-5-methyl-4-isoxazoleproprionate (AMPA) receptors (AMPARs) (Beattie et al., 2000; Lin et al., 2000) and NMDARs (Shi et al., 2000). CN has been reported to contribute to downregulation of Ca2+ signaling by reducing both Ca2+ influx and Ca2+-induced Ca2+ release from intracellular stores via dephosphorylation and inactivation of VGCCs (L-, N-, P/Q-, or R-type; Armstrong, 1989) Additionally, it has been shown that NMDA-induced reduction in dendritic spine density in cultured hippocampal neurons is blocked by pre-treatment with CN inhibitors and that this is probably via preventing the CN-mediated promotion of actin depolymerization (Halpain et al., 1998). CN asserts its effects through disinhibition of PP1 (postsynaptic serine/threonine PP1), so that PP1 is only indirectly regulated by [Ca2+]. When activated, CN dephosphorylates and inactivates inhibitor-1 (Cohen, 1989; Mulkey et al., 1994), which is normally bound to PP1. Inactivation of inhibitor-1 releases and activates PP1. PP1 is also regulated by PKA, which phosphorylates and inactivates inhibitor-1 and thus inhibits PP1 (Cohen, 1989). PP1 regulates gene expression via the dephosphorylation and inactivation of the transcription factor cyclic adenosine monophosphate (cAMP) responsive element-binding protein (CREB; Hagiwara et al., 1992; Bito et al., 1996), which in its active form, initiates the transcription of genes that contain a cAMP response element (CRE) within their promoter sequence (Brindle and Montminy, 1992). Another substrate of PP1 (and/or PP2A) is AMPARs (Lee et al., 2000) which are dephosphorylated at the PKA site

477

(SER-845) (Lee et al., 1998, 2000), resulting in a decrease in their open probability (Banke et al., 2000). Like CN, PP1 has been implicated in activity-dependent endocytotic internalization of AMPARs, since CN and PP1 antagonists each inhibit NMDA- and insulin-induced endocytosis of AMPARs (Beattie et al., 2000; Ehlers, 2000; Lin et al., 2000). PP2A, as well as PP1, have both been reported to dephosphorylate the autophosphorylation site Ser-657/660 and the trans-phosphorylation site Ser-641 on the catalytic domain of PKC-a/PKCbII (Keranen et al., 1995; Thiels et al., 2000). PP1 can also dephosphorylate the autophosphorylation site Ser-641 on the catalytic domain of PKCbII (Keranen et al., 1995). However, the ability of phosphatases to dephosphorylate substrates in vitro or in vivo does not confirm that this dephosphorylation occurs during the induction of LTD. For instance, PP1 and PP2A have been shown to dephosphorylate Thr286 of CaMKII in vitro (Shields et al., 1985; Strack et al., 1997) but hippocampal LTD can still be induced in CaMKIIT286A mice (Krezel et al., 1999; Parsley et al., 2007), indicating that this dephosphorylation event is not required for at least some forms of LTD. Dephosphorylation events which have been linked to decreases in synaptic strength during LTD in area CA1 are: AMPAR dephosphorylation of SER-845 by CN and PP1/PP2A (Lee et al., 2000, 2003), AMPAR internalization (Carroll et al., 1999; Heynen et al., 2000), and PP1/PP2Amediated dephosphorylation of PKC at Ser-657/ 660 (Thiels et al., 2000). CN dephosphorylation of synapsin I (Hosaka et al., 1999) could, by reducing transmitter release (Hilfiker et al., 1999), contribute to presynaptic LTD. Additional substrates dephosphorylated by CN that might play important roles in the long-lasting remodeling of synaptic structure are microtubule-associated protein 2 (MAP2), protein tau and tubulin (Goto et al., 1985), though a requirement of these events for LTD has yet to be confirmed. (b) Kinase pathways Many kinases can be activated by Ca2+ and activated or inactivated by distinct mGluR and other G protein-coupled metabotropic pathways.

mGluRs are divided into three groups: I, II, and III. Group I mGluRs are positively coupled to adenylate cyclase and the cAMP-PKA pathway, and they also can activate PKC. mGluRs II and mGluRs III are negatively coupled to adenylate cyclase. NMDARs contribute to intracellular [Ca2+] elevation via Ca2+ influx through NMDAR channels, followed by Ca2+-induced Ca2+ release from ryanodine-sensitive stores (Alford et al., 1992). Group I mGluRs promote the release of cytosolic Ca2+ from inositol-1,4,5trisphosphate (IP3)-sensitive intracellular stores via positive coupling to IP3. It has been demonstrated that induction of LFS-LTD requires Ca2+ release from presynaptic, but not postsynaptic, ryanodine-sensitive stores, and Ca2+ release from postsynaptic IP3-sensitive stores (Reyes and Stanton, 1996). Intracellular Ca2+ can activate PKA via Ca2+/CaM-sensitive adenylyl cyclase (AC) (Eliot et al., 1989) which synthesizes cAMP, CaMKII via Ca2+/CaM (Colbran, 1992), PKG via the Ca2+/ CaM-NOS-NO-GC-cGMP pathway (Garbers, 1990; Bredt et al., 1991), and PKC by Ca2+ paired with diacylglycerol (DAG; Huang, 1989). During the induction of LTD, kinases are bidirectionally regulated, as are their effector pathways. While sometimes phosphorylated and activated kinases are necessary to induce LTD, in other cases kinases are dephosphorylated and inactivated to promote LTD. Ca2+/phospholipid-dependent protein kinase (PKC). Reports indicate that inhibition of PKC may be required for LTD, since a selective inhibitor of PKC, chelerythrine, induces a depression which shows mutual occlusion with LFS-LTD (Hrabetova and Sacktor, 1996) suggesting similar induction and/or expression mechanisms. Indeed, LTD is associated with a decrease in PKC activity (Hrabetova and Sacktor, 1996; Thiels et al., 2000). This decrease in PKC activity has been shown to be mediated by either a proteolysis-mediated decrease in PKCg and PKMz activity (Hrabetova and Sacktor, 1996) or by dephosphorylation of catalytically relevant autophosphorylation sites on the C terminus of PKC (Ser-657 on PKCa; Ser-660 on PKCbII) by PPs (Thiels et al., 2000). While the decrease in activity of PKCa, b, and g was only transient (35–65 min,

478

Thiels et al., 2000), downregulation of the autonomously active PKMz (Schwartz, 1993) persisted up to 120 min post-LFS (Hrabetova and Sacktor, 2001) making it a special candidate for the maintenance phase of LTD. It has been demonstrated that LTD is associated with a decrease in pre- as well as postsynaptic PKC substrate phosphorylation (Ramakers et al., 2000). Since it is known that PKC activators such as phorbol esters enhance presynaptic transmission (Chaki et al., 1994), it seems plausible that inactivation of constitutively active PKC might lead to depression of release. On the postsynaptic side, PKC is known to target the extracellular signal-regulated kinase (ERK) signaling cascade (Roberson et al., 1999) and could thus affect transcription processes during LTD (see below). On the other hand, there are also studies suggesting a requirement for activation of PKC during LTD. PKC is known to phosphorylate SER-880 of the AMPAR subunit GluR2 (McDonald et al., 2001), GluR2 SER-880 phosphorylation is increased after LTD induction (Kim et al., 2001), preventing the phosphorylation of SER-880 on GluR2 by point mutation inhibits the constitutive synaptic incorporation of GluR2 homomers, as well as LTD (Seidenman et al., 2003) and removal of GluR2 containing AMPARs from synapses is one potential expression mechanism for LTD (Seidenman et al., 2003). Thus, it is feasible that PKC activates a pathway that includes the phosphorylation of SER-880 on GluR2, followed by the endocytosis of GluR2 containing AMPARs to express LTD. This pathway could be implemented by the PKC isoforms (a, bI, bII, g) that have been reported to be transiently (o15 min) activated after induction of LTD (Hrabetova and Sacktor, 2001). Overall, multiple roles for both presynaptic and postsynaptic isoforms of PKC make the functions of these enzymes in induction and maintenance of LTD potentially quite complex. Ca2+/calmodulin-dependent protein kinase (CaMKII). It has been reported that mice that lack the gene for alpha CaMKII show reduced LTD (Stevens et al., 1994). Furthermore, induction of LFS-LTD is blocked by extracellular application of the CaMKII inhibitor KN-62 to hippocampal

slices, but not by postsynaptic intracellular infusion of KN-62 into single CA1 pyramidal neurons, indicating that the CaMKII that may be necessary for LTD induction is presynaptically localized (Stanton and Gage, 1996). cAMP-dependent protein kinase (PKA). There are contradictory studies concerning the role of PKA in the induction of LTD. Brandon et al. (1995) reported that the inhibition of PKA strongly inhibits LFS-induced LTD at mouse Schaffer collateral–CA1 synapses (Brandon et al., 1995), while studies in knockout mice have reported that LTD in field CA1 is dramatically reduced if either the PKA regulatory subunit RIb (Brandon et al., 1995) or the PKA catalytic subunit Cb1 (Qi et al., 1996) are missing. Interestingly Cb1- and RIb-containing PKA holoenzymes show a higher sensitivity to cAMP than Cb- and RIb-containing PKA holoenzymes (Cadd et al., 1990) and could be preferentially activated during LTD induction when Ca2+ influx, and perhaps, associated increases in [cAMP] may be less than those needed to elicit LTP. A number of studies explicitly contradict the notion that PKA activation is needed for LTD. Santschi et al. (1999) reported that multiple synthetic PKA inhibitors enhanced the magnitude of LTD at these same synapses in rat hippocampal slices, while Nicholls et al. (2006) found that LTD is enhanced in a transgenic mouse model where adenylate cyclase is constitutively inhibited by Gi2a and PKA stimulation prevented. Studies by Kameyama et al. (1998) similarly support a role for inhibition of PKA in promoting LTD; they found that inhibition of postsynaptic PKA induced a depression that occluded LFS-LTD and that LFS-LTD was inhibited by prior activation of postsynaptic PKA. Furthermore, LTD is blocked by the selective mGluRII antagonist EGLU, and by the A1 adenosine receptor blocker DPCPX, in juvenile rats (Santschi et al., 2006). Since mGluRIIs are negatively coupled to cAMP, it is a reasonable hypothesis that they mediate LTD by decreasing [cAMP] and PKA activity. The contradictory data suggesting a positive role for PKA activity in LTD might indicate either a developmental switch and/or dependence on genus, since Kameyama et al. (1998) and Santschi et al. (1999)

479

used juvenile rats (14–21 days), while Brandon et al. (1995) and Qi et al. (1996) studied adult mice (4–6 weeks). Distinct PKA isoforms may also play roles in different forms of synaptic plasticity, i.e., LTP or LTD (Brandon et al., 1997). cGMP-dependent protein kinase (PKG). LTD at Schaffer collateral–CA1 synapses can also be blocked by PKG inhibitors (Reyes-Harde et al., 1999a, b). It is thought that PKG is activated presynaptically during LTD induction via a cascade that includes the retrograde messenger nitric oxide (NO), activation of guanylyl cyclase, production of guanosine 30 ,50 -cyclic monophosphate (cGMP) and subsequent PKG activation. Although double knockout mice lacking both the neuronal and endothelial forms of NO synthase (nNOS and eNOS) did not show significantly reduced LTD (Son et al., 1996; 3 months) and LFS-LTD was not affected by the NOS inhibitor L-NG-nitroarginine (L-NOArg) (Cummings et al., 1994), several other studies have demonstrated a role for NO for LFSLTD. Izumi and Zorumski (1993) found a block of LFS-LTD by the NO inhibitor L-NG-monomethylargine (L-NMMA) as well as L-NOArg. Santschi et al. (1999) partially reduced the magnitude of LFS-LTD by application of the competitive NOS inhibitor L-nitroarginine (L-NA). Furthermore, Stanton et al. (2003) also reported a partial block of LFS-LTD by L-NA, as well as by the extracellular NO scavenger hemoglobin. Additional support for the NO-cGMP-PKG pathway comes from studies of Gage et al. (1997) and Reyes-Harde et al. (1999a, b); Gage et al. (1997) reported that presynaptic NO-sensitive guanylyl cyclase (NOGC) is necessary for the induction of LTD, since an extracellularly applied NOGC inhibitor prevented the induction of LTD, while selective intracellular blockade of postsynaptic NOGC did not. Furthermore, they also showed that the membrane-permeable cGMP analog 8-pCPTcGMP and the NO donor S-nitroso-Nacetylpenicillamine (SNAP) potentiated the induction of LTD by a weak submaximal LFS (1 Hz/400 stimuli). Reyes-Harde et al. (1999b) demonstrated that the potentiation of LTD by SNAP was mediated via the NOGC-cGMP-PKG cascade and required release of Ca2+ from Ca2+-mediated

ryanodine-sensitive stores, suggesting that this signal cascade and Ca2+ release from ryanodine-sensitive stores mediated a component of LFS-LTD. Since it is known that cGMP can enhance Ca2+release from ryanodine-sensitive stores via activation of cADP-ribose (Galione et al., 1993; Meszaros et al., 1993; Lee, 1997), it has been proposed that the PKG-ADP ribosyl cyclase-cADP ribose cascade links cGMP to Ca2+-release from ryanodine-sensitive stores during LTD (Reyes-Harde et al., 1999a). Reyes-Harde et al. (1999a) verified that cGMP does stimulate the synthesis of cADP ribose in hippocampal slices and found that antagonists of cADP ribose receptors in presynaptic terminals block the induction of a portion of LTD. Presynaptic vesicular release. LFS-LTD has been shown to be associated with a long-term reduction in neurotransmitter release (Stanton et al., 2001, 2003; Zhang et al., 2006). There is evidence that key opposing mediators of LTD on the presynaptic site are both PKA and PKG. In fact, it has been shown that a chemical LTD (cLTD) can be induced by simultaneously raising [cGMP] while inhibiting PKA (Santschi et al., 1999), and that this is associated with a decrease in the evoked release of the fluorescent marker of presynaptic activity FM1–43 from isolated rat hippocampal presynaptic terminals (Bailey et al., 2003). Furthermore, cLTD is occluded by LFS-LTD (Santschi et al., 1999) and suppresses LFS-evoked FM1–43 release (Stanton et al., 2001), demonstrating that it mimics the signal transduction pathways necessary for LFS-LTD, i.e., simultaneous increase of [cGMP] and inhibition of PKA, that lead to a presynaptic decrease in neurotransmitter release. It has also been shown that a selective PKG inhibitor completely prevents induction of LTD of presynaptic FM1–43 release from the RRP (Stanton et al., 2003). Further upstream, activation of NMDARs and group II mGluRs have both been indicated to play a role in the induction of presynaptic LFS-LTD. Inhibition of NMDARs or mGluRs (group I and II, but not group III) have been shown to block induction of LFS-LTD (Reyes-Harde et al., 1999b; Stanton et al., 2003; Santschi et al., 2006). The induction of LFS-LTD of FM1–43 release from the RRP by a 1–2 Hz

480

train is largely blocked by the NMDAR antagonist AP-5 (Stanton et al., 2003). The competitive NOS inhibitor L-NA completely prevented the LFS-induced RRP release and the extracellular NO scavenger hemoglobin partial blocked LFSinduced RRP release (Stanton et al., 2003) suggesting that NMDARs mediate presynaptic LTD via the retrograde messenger NO. Zhang et al. (2006) found that group II, but not group I mGluRs, contribute to the induction of presynaptic LTD by a different, paired-pulse stimulation protocol, indicating that, depending upon the particular stimulus protocol, either NMDARs or group I mGluRs can supply the necessary postsynaptic [Ca2+], while any presynaptic receptors, such as group II mGluRs, that inhibit adenylate cyclase can supply the necessary suppression of [cAMP] in the presynaptic terminal. Extracellular signal-regulated kinase (ERK). ERK/ MAPK cascade has also been implicated as a key signal transduction mechanism for coupling neuronal cell surface receptors to plasticity-induced transcription. It has been demonstrated that ERK activation is necessary for the induction of LTD, since an ERK inhibitor blocks induction of LTD (Thiels et al., 2002). Multiple second messengers, such as cAMP, PKA, [Ca2+], and DAG, can all control ERK signaling via the small G proteins Ras and Rap1 (Grewal et al., 1999) and thus link both NMDARs and mGluRs to this pathway. During induction of LTD, ERK stimulates specific transcription pathways that differ from LTP; in LTD, ERK activation does not result in the increased phosphorylation of the transcription factor CREB, but rather an increased phosphorylation of the transcription factor Elk-1 (Thiels et al., 2002). Phospholipase A2. It has been suggested that during LTD Ca2+ activates Ca2+-sensitive phospholipase A2 (PLA2) (Clark et al., 1995) besides phosphatases and kinases. The activation of PLA2 leads to arachidonic acid (AA) and lipoxygenase metabolite production. Findings which suggested involvement of the 12-1ipoxygenase pathway in LTD are: (1) A non-selective PLA2 inhibitor, bromophenacylbromide, significantly reduced the extent of Schaffer collateral–CA1 LTD in slices

from juvenile (P10–15; Fitzpatrick and Baudry, 1994) and adult rats (Normandin et al., 1996), but a more selective PLA2 inhibitor did not block induction of LTD in adults (Stanton, 1996), (2) inhibitors of lipoxygenase pathways of AA metabolism decreased the magnitude of LTD in juvenile (P20–25; Chabot et al., 1998) and adult rats (Normandin et al., 1996). The 12-1ipoxygenase pathway may have cellular effects compatible with LTD expression, such as inhibition of CaMKII (Piomelli et al., 1989) or reduction of transmitter release (Freeman et al., 1991) and changes in synaptic transmission through the alteration of the affinity states of AMPARs (Chabot et al., 1998). In detail, it was shown that 12-lipoxygenase products attenuate the potassiuminduced release of endogenous glutamate from hippocampal mossy fiber synaptosomes (Freeman et al., 1991) and low concentrations of PLA2 caused a decrease in AMPA binding due to a change in receptor affinity that was blocked by lipoxygenase inhibitors (Chabot et al., 1998). The PLA2-induced decrease in AMPA binding seems to be dependent on protein phosphatase dephosphorylation, as the PP inhibitor okadaic acid resulted in a significant reduction in the PLA2-induced decrease in AMPA (Chabot et al., 1998). As with many biochemical cascades, the involvement/requirement for PLA2 activation in LTD may prove to be stimulus protocol-specific. Maintenance. It is suggested that long-lasting expression of LTD of synaptic strength requires the regulated expression of proteins, some of them with low resting concentrations. It has been demonstrated that LTD, like LTP (Stanton and Sarvey, 1984), requires ongoing PS for stable expression. In slice cultures of 8–9 days old rats, the induction of LFS-LTD has been reported to be inhibited by the PS inhibitor anisomycin, as well as the transcriptional inhibitor actinomycin D (but see Huber et al., 2000; Kauderer and Kandel, 2000). While anisomycin blocks the late phase of LTD when applied prior to an LFS, it does not affect the maintenance of LTD when applied 30 min after LFS, suggesting that PS is only necessary during the induction and early phases of LTD to synthesize proteins necessary for maintenance of persistent depression.

481

LFS-LTD in adult rats (7–8 weeks) seems to be only dependent on PS, but not transcription, as its induction was blocked by anisomycin (in vivo: Manahan-Vaughan et al., 2000; in slices: Sajikumar and Frey, 2003) but not actinomycin D (ManahanVaughan et al., 2000). The blockade of LTD by anisomycin, applied before LFS, becomes effective at approximately 5 h after LFS, suggesting that, while existing protein levels can support alterations in synaptic strength lasting many hours, the maintenance of longer-lasting synaptic plasticity, like long-term memory, requires ongoing PS.

NMDAR-dependent LTD NMDAR-dependent LTD can be induced in area CA1 via application of NMDA (Lee et al., 1998; Kamal et al., 1999), or LFS-stimulation under pharmacological mGluR blockade (Oliet et al., 1997; Zhang et al., 2006) (Fig. 1). Brief application of NMDA (3 min, 20 mM) to hippocampal slices of young (P21–35: Lee et al., 1998; P14: Kamal et al., 1999) or adult rats (6 months; Kamal et al., 1999) induces NMDARdependent LTD blocked by the NMDAR antagonist 2-amino-5-phosphonopentanoic acid (AP5). NMDA-LTD and LFS-LTD have been shown to occlude one another (Lee et al., 1998; Kamal et al., 1999), suggesting shared induction and/or expression mechanisms. Indeed, in juvenile rats, NMDALTD, like LFS-LTD (Lee et al., 2000), is associated with dephosphorylation of the GluR1 subunit of AMPARs at the PKA-site (SER-845; Lee et al., 1998). Additionally, as for LFS-LTD (Kameyama et al., 1998), inhibition of PKA seems to play a role for NMDA-LTD in juvenile rats, since both NMDA-LTD and AMPAR dephosphorylation are inhibited by prior activation of postsynaptic PKA (Kameyama et al., 1998). Nevertheless, in juvenile rats, NMDA-LTD and the NMDA-induced dephosphorylation of AMPARs cannot be blocked by high concentrations of inhibitors of PP1 and PP2A (Kameyama et al., 1998), in marked contrast to earlier findings for LFS-LTD (Mulkey et al., 1993; Lee et al., 2000). Even though PP2B inhibitors did not affect NMDA-induced dephosphorylation of AMPARs, they partially inhibited NMDA-LTD

in juvenile rats (Kameyama et al., 1998), suggesting that PP2B dephosphorylates other substrates, but not AMPARs, during NMDA-LTD. Interestingly, the role of phosphatases for NMDA-LTD seems to be developmentally regulated. While inhibitors of PP1/PP2A do not affect NMDA-LTD in juvenile rats (P21–30: Kameyama et al., 1998), they have been reported to block NMDA-LTD in adult rats (6 months: Kamal et al., 1999). Furthermore, inhibition of PP2B completely blocks NMDA-LTD in adult rats (6 months: Kamal et al., 1999), but only partially blocks LTD in juvenile rats (P21–30: Kameyama et al., 1998) and is without effect on NMDA-LTD in even younger rats (P14; Kamal et al., 1999). Brief application of NMDA (20–50 mM for 1–2 min) with a concentration and application duration that is similar to those that induce NMDA-LTD (3 min, 20 mM), causes a rapid, nearly complete internalization of surface AMPARs in hippocampal cultures (2–3 weeks in vitro; Beattie et al., 2000; Ehlers, 2000), suggesting that NMDAR-LTD, like LFS-LTD, may be expressed by AMPAR endocytosis. Indeed, Nosyreva and Huber (2005) confirmed that NMDA-induced LTD (3 min, 20 mM NMDA) in CA1 of neonatal hippocampal slices was associated with a decrease in surface expression of GluR1 and GluR2/3 at 10 min and 60 min, respectively. Like LFS-LTD, the NMDA-induced AMPAR internalization was dependent on Ca2+ influx, being almost completely inhibited by removing extracellular Ca2+ (Beattie et al., 2000) and by the Ca2+ chelator BAPTA-AM (Ehlers, 2000). Like LFS-LTD, NMDA-induced AMPAR internalization was also dependent on CN, but not PP2A, since it was inhibited by the CN blocker FK-506, but not by low concentrations of okadaic acid (10 nM) that selectively inhibits PP2A (Beattie et al., 2000; Ehlers, 2000). There are contradictory findings concerning the role of PP1 in AMPAR endocytosis. While Beattie et al. (2000) could not inhibit NMDA-induced AMPAR internalization with the PP1/PP2A inhibitors calyculin A or okadaic acid (1 mM), Ehlers (2000) reported an almost complete inhibition of NMDA-induced AMPAR internalization by these inhibitors, as well as the PP1-selective inhibitor tautomycin. While the reason for these divergent

482

findings is unclear, there were slightly different induction protocols used; Beattie et al. (2000) prepared dissociated cell cultures from areas CA1/ CA3 from P0 rat pups, while Ehlers (2000) prepared hippocampal mixed cell cultures from 1 to 2 days old rats, with both maintained 2–3 weeks in vitro before commencing experiments. Beattie et al. (2000) induced AMPAR internalization by application of 50 mM NMDA for 1 min in the presence of both AMPAR and mGluR antagonists, while Ehlers (2000) applied 20 mM NMDA alone for 1–2 min, possibly suggesting an additional role for ambient activation of mGluRs in the activation of phosphatases that lead to AMPAR internalization. Interestingly, the findings of Ehlers (2000) seem to support the hypothesis that AMPAR dephosphorylation is a step toward AMPAR internalization. They observed that application of NMDA, which induces AMPAR endocytosis, additionally induces the rapid dephosphorylation of GluR1 at the SER-845 PKA site. Both events are temporally correlated, as NMDA-induced AMPAR internalization peaked shortly after maximal dephosphorylation of GluR1. Additionally, internalized GluR1 subunits showed no phosphorylation at the PKA site SER-845, but were still phosphorylated at the CaMKII site SER-831. Both AMPAR endocytosis and dephosphorylation of GluR1 at the SER-845 PKA site, were prevented by Ca2+-free medium, FK-506, high concentrations of okadaic acid (1 mM) or tautomycin, but not prevented by low concentrations of okadaic acid. Ehlers (2000) suggested that NMDAR activity results in Ca2+ influx and activation of a PP cascade (including PP2B and PP1) that triggers AMPAR dephosphorylation followed by endocytosis. There is evidence that NMDAR-dependent LTD is expressed not only postsynaptically as AMPAR dephosphorylation and endocytosis (Lee et al., 1998; Beattie et al., 2000; Ehlers, 2000; Zhang et al., 2006), but presynaptically as a longterm reduction in neurotransmitter release (Zhang et al., 2006). Zhang et al. (2006) showed, by evaluating changes in paired-pulse ratio (PPR), twophoton imaging of FM1–43 release and variancemean analyses, that NMDA-LTD is associated with reduced transmitter release probability in slices from juvenile rats (P14–19). Additionally,

they demonstrated that NOS activity is required for this presynaptic NMDAR-dependent LTD, as they could block it with the NOS inhibitor L-NA, suggesting NO as a retrograde transmitter that establishes presynaptic NMDAR-dependent LTD. Finally, it has been shown that T-type VGCCs and activation of PKC do not appear to play a role in NMDAR-dependent LTD (as opposed to mGluR-dependent LTD (Oliet et al., 1997), as the T-type VGCC channel blocker Ni2+ and the PKC inhibitory peptide, PKC19–36, did not affect the induction of NMDAR-dependent LTD (Oliet et al., 1997). mGluR-dependent LTD It is now clear that there exist multiple forms of mGluR-dependent LTD, depending on developmental stage and mGluR subtype involved. The mGluRs are a family of heterotrimeric guanosine triphosphate-binding protein (G protein)-coupled receptors (Sugiyama et al., 1987; Tanabe et al., 1992). They are divided into three groups, mGluR I, mGluR II, and mGluR III, which are made up of eight subtypes. This categorization is based on their amino acid sequence homology, agonist selectivity, and second messenger coupling (Conn and Pin, 1997). mGluR-dependent LTD in neonatal rats. In neonatal rats electrical stimulation (5 Hz/3 min) induces both NMDAR-independent and mGluR-dependent LTD (P3–7: Bolshakov and Siegelbaum, 1994). Metabotropic GluR-dependent LTD has been reported to require activation of group I mGluRs (but group II mGluR blockade was not tested; Oliet et al., 1997/P6–7; Nosyreva and Huber, 2005/ P8–15). mGluR activation is only necessary for induction of LTD in neonatal rats, since maintenance is not inhibited by the mGluRI inhibitor (+)-amethyl-4-carboxyphenylglycine (MCPG; Bolshakov and Siegelbaum, 1994) (Fig. 2). mGluR-dependent LTD, which is not affected by NMDAR blockade (Li et al., 2002), can be also induced by 1 Hz stimulation in neonatal rats (P6–8; Li et al., 2002; P4–10; Feinmark et al., 2003). mGluR-LTD induced by 1 Hz in neonatal rats has

483

Fig. 2. Cascades involved with mGluR-dependent LTD in neonatal rats at Schaffer collateral–CA1 synaspes. AA: arachidonic acid, Ca: Ca2+, HPETE: hydroperoxyeicosatetraenoic acid, IP3: inositol-1,4,5-trisphosphate, MAPK: mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, PIP2: phosphatidylinositol-3,4-bisphosphate, PLA2: phospholipase A2, PLC: phospholipase C, Post: postsynapse, VGCC: voltage-gated Ca2+ channels.

been claimed to be dependent on group II mGluRs only (as it was not affected by an mGluRI blocker; Li et al., 2002) but other groups detected a blockade of LTD after mGluR5 blockade (group II mGluR blockade was not tested; Feinmark et al., 2003). Induction of group I mGluR-LTD is prevented by the Ca2+ chelator EGTA and by the blocker of L-type VGCCs nitrendipine in slices from neonatal (P3–8) rats (Bolshakov and Siegelbaum, 1994), suggesting that a combination of Ca2+ influx via these channels and release from intracellular stores is needed for this form of LTD. Elevation of postsynaptic [Ca2+] and mGluR activation (with the specific mGluR agonist ACPD) is sufficient to induce LTD (Bolshakov and Siegelbaum, 1995), indicating a postsynaptic site of induction. Nevertheless, studies indicate that group I mGluR-LTD in neonatal rats is not expressed postsynaptically,

as postsynaptic mechanisms like PS or changes in the surface expression of AMPARs play no role in group I mGluR-LTD (Nosyreva and Huber, 2005). Instead, group I mGluR-LTD in neonatal rats appears to be expressed presynaptically as a long-term decrease in transmitter release, since the expression of LTD at this age is accompanied by changes in PPR (Nosyreva and Huber, 2005), coefficient of variation of EPSCs (Bolshakov and Siegelbaum, 1994), a decrease in the frequency but not amplitude, of miniature EPSCs (Bolshakov and Siegelbaum, 1994), and a decreased rate of release of the fluorescent marker FM1–43 (N-(3-triethylammoniumpropyl)-4-(4-(dibutylamino) styryl) pyridinium dibromide; Zakharenko et al., 2002). Since this form of LTD is induced postsynaptically but expressed presynaptically, a retrograde messenger is needed to transfer the signal from the

484

postsynaptic to the presynaptic site (Bolshakov and Siegelbaum, 1994). AA has been suggested to be the retrograde transmitter involved in neonatal mGluRLTD (P4–7; Bolshakov and Siegelbaum, 1995), because AA application can induce LTD (Bolshakov and Siegelbaum, 1995), an inhibitor of PLA2 that initiates the AA signaling cascade by liberating AA from membrane phospholipids, blocks LTD (Bolshakov and Siegelbaum, 1995) and metabolites of AA have been shown to mediate mGluR-LTD in neonatal rats (Feinmark et al., 2003). Feinmark et al. (2003) showed that 12(S)-hydroperoxyeicosa5Z,8Z,10E,14Z-tetraenoic acid (12-S-HPETE), a 12lipoxygenase metabolite of AA, is actively recruited during the induction of mGluR-LTD, since (1) a mouse in which the leukocyte-type 12-lipoxygenase (the neuronal isoform) was deleted through homologous recombination was deficient in mGluR-LTD, (2) pharmacological inhibition of 12-lipoxygenase also blocked induction of mGluR-LTD, and (3) direct application of 12-S-HPETE to hippocampal slices induced a LTD of synaptic transmission that mimicked and occluded mGluR-LTD induced by synaptic stimulation. Group I mGluR-LTD has also been reported to require the postsynaptic activation of p38 MAPK in neonatal rats, since introduction of the p38 MAPK inhibitor SB203580 into a postsynaptic CA1 pyramidal neuron greatly diminished the induction of this form of LTD, while the p42/44 inhibitor PD 98059 had no effect (P4–11; Bolshakov et al., 2000). Additionally, it has been demonstrated that dialyzing CA1 neurons with the active, dually phosphorylated p38 MAPK induces a progressive depression in the amplitude of the EPSC which occludes LTD in response to subsequent 5 Hz electrical stimulation. P38 MAPK provides a link to the AA pathway since P38 MAPK can activate cytosolic PLA2 (cPLA2) by phosphorylation (BorschHaubold et al., 1997). cPLA2 is also regulated by [Ca2+]; Ca2+ promotes binding of the enzyme to phospholipid substrates by causing the translocation of cPLA2 to nuclear and other cellular membranes through a Ca2+-dependent lipid-binding motif (Clark et al., 1995). mGluR-dependent LTD in juvenile rats. Group I mGluR-dependent LTD has been induced in

juvenile rats by LFS stimulation (5 Hz/3 min; P11-35; Oliet et al., 1997; Nicoll et al., 1998) and by paired-pulse LFS (PP-LFS; paired pulses at 50 ms interstimulus intervals, applied at a 1 Hz frequency for 15 min; P21–30; Huber et al., 2000) in the presence of the NMDAR antagonist AP5. Additionally, chemical activation of group I mGluRs with (S)-3,5-dihydroxyphenylglycine (DHPG) can induce an LTD in juvenile rats (Huber et al., 2001) which is independent of NMDAR activation (Huang and Hsu, 2006). LFS-induced group I mGluR-dependent and DHPG-LTD show mutual occlusion, which suggests similar induction/ expression mechanisms. Group I mGluR-dependent LTD in juvenile rats seems to be mediated by group I mGluRs, since it is blocked by the selective group 1 antagonist AIDA ([CRS]-1-aminoindan-1,5-dicarboxylic acid) (Oliet et al., 1997), but not by the selective group II mGluR antagonist EGLU (2S-aethylglutamic acid), or the selective group III mGluR antagonist MAP4 ((S)-2-amino-2-methyl4-phosphonobutanic acid) (Huang and Hsu, 2006). The responsible group I mGluR subtype appears to be mGluR5, since group I mGluR-LTD is not affected by the mGluR1 selective antagonist 4CPG ((S)-4-carboxyphenylglycine) (Oliet et al., 1997; Huang and Hsu, 2006), is blocked by the mGluR5 selective antagonist MPEP (2-methyl-6-(phenylethynyl)pyridine) (Huang et al., 2004; Huang and Hsu, 2006) and is absent in homozygote mGluR5 knockout mice (Huber et al., 2001). Established DHPG-LTD is completely reversed by application of the mGluR antagonists MCPG or LY341495 and LTD is reestablished once the antagonists are washed out, suggesting that the induction of DHPG-LTD involves either an alteration in functional state of the receptors or the switching on of a tonically active mGluR receptor (Huang and Hsu, 2006). In contrast to the requirement for persistent mGluRI activation, the Gq-mediated downstream pathways of group I mGluRs do not seem to be necessary for the maintenance of mGluRI-LTD; neither the phospholipase C (PLC) blocker U73122, the PKC inhibitor Bis-1 nor the depletion of intracellular Ca2+ pools by thapsigargin or cyclopiazonic acid have any effect on pre-established DHPG-LTD (Reyes and Stanton, 1996; Huang and Hsu, 2006). Instead, a downstream protein tyrosine phosphatase

485

(PTP) signaling cascade seems to be required for the expression of DHPG-LTD (see below; Huang and Hsu, 2006) (Fig. 3). The induction of group I mGluR-LTD requires postsynaptic activation of PKC (Oliet et al., 1997; Nicoll et al., 1998; Huang and Hsu, 2006), but is not affected by PP1/2A inhibitors (P11–35: Oliet et al., 1997; Nicoll et al., 1998; P21–28: Huang and Hsu, 2006), contrasting with NMDAR-LTD, which does appear to involve postsynaptic activation of PP2B (P21–30: Kameyama et al., 1998). Like NMDARLTD, induction of group I mGluR-LTD relies on increases in postsynaptic [Ca2+] (Oliet et al., 1997; Nicoll et al., 1998; but see Fitzjohn et al., 2001/ P12–18). In contrast to group I mGluR-LTD in neonatal rats (Bolshakov and Siegelbaum, 1994), induction

of group I mGluR-LTD in juvenile rats is prevented by blockers of T-type VGCCs, but not L-type VGCCs (Oliet et al., 1997; Nicoll et al., 1998), suggesting a developmental switch. Additionally, it has been shown that group I mGluRLTD in juvenile rats is prevented by inhibitors of mRNA translation but not transcription (Huber et al., 2000), indicating that new PS from existing mRNA species is a component of mGluR-LTD by this developmental stage. There are contradictory findings of the role of p38 MARK for the induction of group I mGluRLTD; one study reported that group I mGluRLTD was not affected by p38 MARK inhibitors, but was blocked by p42/44 MARK and ERK inhibitors instead (Gallagher et al., 2004). Another study found no effect of p42/44 MARK inhibition

Fig. 3. Cascades involved with mGluR-dependent LTD in juvenile rats at Schaffer collateral–CA1 synapses. Ca: Ca2+, DAG: diacyl glycerol, IP3: inositol-1,4,5-trisphosphate, MAPK: mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, P: phosphate, PIP2: phosphatidylinositol-3,4-bisphosphate, PKC: protein kinase C, PLC: phospholipase C, Post: postsynapse, PTP: protein tyrosine phosphatase, VGCC: voltage-gated Ca2+ channels.

486

on group I mGluR-LTD, but a potent effect of p38 MARK inhibition (Huang et al., 2004). Huang et al. (2004) further showed, with specific inhibitors, that during group I mGluR-LTD, p38 MARK is activated via a specific cascade that includes activation of mGluR5, release of G protein bg-subunits, which, in turn, promotes the exchange of GDP with GTP of the small GTPase Rap1, causing Rap1 to activate a sequential kinase cascades that includes MAPK kinase 3/6 and p38 MAPK. There is evidence for both pre- and postsynaptic expression of group I mGluR-LTD in juvenile rats. A presynaptic site of expression is suggested by studies showing that expression of LTD is accompanied by a change in PPR (Fitzjohn et al., 2001; Nosyreva and Huber, 2005), increase in EPSC failure rate (Fitzjohn et al., 2001), a change in the coefficient of variation of EPSCs (Fitzjohn et al., 2001), a decrease in the frequency but not the amplitude, of miniature EPSCs (Fitzjohn et al., 2001), and a decrease in the frequency, but not in amplitude, of Sr2+ asynchronously-evoked quantal events (which only occurs at the subset of synapses that are stimulated; Oliet et al., 1997). But these data are challenged by conflicting findings from Zhang et al. (2006) which showed a lack of presynaptic changes for DHPG-LTD and pharmacologically isolated group I mGluR-LTD, measured using changes in PPR, two-photon imaging of FM1–43 release, and variance-mean analysis. This study did, however, indicate that coactivation of group II mGluRs could recruit presynaptic changes in release. There is evidence that group I mGluR-LTD is expressed by postsynaptic NMDAR and AMPAR internalization, since group I mGluR activation induces a PS-dependent loss of postsynaptic AMPARs and NMDARs in cultured hippocampal neurons (Snyder et al., 2001), and group I mGluR activation induces an LTD in cultured slices which is dependent on AMPAR endocytosis (Xiao et al., 2001). In acute hippocampal slices from juvenile rats, group I mGluR activation produces a decrease in AMPAR surface expression (GluR1 and GluR2/3) which was completely blocked by the broad-range mGluR antagonist LY341495, while the late phase (60 min after LTD induction) was blocked by the translational inhibitor anisomyin

(Huang et al., 2004; Nosyreva and Huber, 2005; Huang and Hsu, 2006). During expression of group I mGluR-LTD, the decrease in AMPAR surface expression has been shown to be regulated by p38 MARK. p38 MARK accelerates a rapid loss of surface AMPARs by stimulating the activity of guanyl nucleotide dissociation inhibitor (GDI), extracting Rab5 from endosomal membranes and forming a GDI-Rab5 complex (Cavalli et al., 2001; Huang et al., 2004). This complex is required for the sequestration of the ligandreceptor-complex into clathrin-coated pits (McLauchlan et al., 1998). Huang and Hsu (2006) showed that a transient activation of p38 MARK is sufficient to stimulate AMPAR endocytosis during group I mGluR-dependent LTD, since maintenance of previously-induced DHPGLTD is not affected by the selective p38 MAPK inhibitor SB203580, and the increase in phosphorylation of p38 MAPK elicited by DHPG is transient. In contrast to p38 MAPK, postsynaptic PTPs appear to be necessary for the continuing expression of group I mGluR-dependent LTD, because DHPG-LTD is reversed by postsynaptic loading of pyramidal neurons with a PTP inhibitor (Huang and Hsu, 2006). Interestingly, it has been demonstrated that levels of tyrosine phosphorylation of GluR2 AMPAR subunits (phosphorylated under basal conditions (Ahmadian et al., 2004)), but not GluR1 subunits, are dramatically decreased during DHPG-LTD, suggesting that AMPARs are targets of PTPs during maintenance of group I mGluR-dependent LTD. Another chemically induced form of LTD at Schaffer collateral–CA1 synapses in hippocampal slices from juvenile rats (14–24 days old) involves the activation of group II mGluRs (Santschi et al., 2006; Zhang et al., 2006). While the selective mGluR II agonist DCGIV [(2S,20 R, 30 R)-2-(20 ,30 dicarboxycyclopropyl)glycine] alone only induces a short-term depression which fully reverses upon drug washout, a combination of DCGIV application and elevation of endogenous intracellular [cGMP] by application of the type V phosphodiesterase inhibitor zaprinast, is sufficient to induce LTD (Santschi et al., 2006). It is probable that mGluRs II promote LTD by inhibiting adenylate cyclase, because pairing application of zaprinast

487

with either activation of A1 adenosine receptors, which also inhibit adenylate cyclase, or with direct pharmacologic inhibition of PKA activity, are each sufficient to elicit LTD (Santschi et al., 2006). LTD elicited by any of these methods of pairing elevated [cGMP] with reduced [cAMP], is induced and expressed presynaptically: selective postsynaptic inhibition of PKA does not enhance LTD, while postsynaptic inhibition of PKG does not block any of these chemical forms of LTD (Santschi et al., 1999). Consistent with this conclusion, it has been shown that coapplication of zaprinast with a synthetic PKA inhibitor (H-89) significantly and persistently decreases evoked presynaptic vesicular release of the fluorescent marker FM1–43 from Schaffer collateral terminals in hippocampal slices (Stanton et al., 2001, 2003), and from isolated rat hippocampal presynaptic terminals (Bailey et al., 2003). mGluR-dependent LTD in adult rats. (a) DHPGinduced mGluRI-dependent LTD The group I mGluR agonist DHPG also induces LTD in slices from adult rats (8–10 weeks old), under conditions of increased excitability (in Mg2+free medium; Palmer et al., 1997). Interestingly, this form of LTD does not occlude stimulus-induced (2 Hz/10 min) LTD (as opposed to the studies of Huber and colleagues, see above). Based on studies in immature rats, it was expected that DHPG induces LTD via the activation of group I mGluR-coupled intracellular signal cascades like PKC and PKA activation, and release of Ca2+ from intracellular stores. Surprisingly, DHPG-induced LTD has been reported not to be affected by PKC or PKA inhibitors, or by depleting intracellular Ca2+ stores (Schnabel et al., 1999a, 2001). Instead, induction, but not expression, of DHPG-LTD is prevented by inhibitors of PTPs (Moult et al., 2002), but the molecular targets of tyrosine dephosphorylation remain to be identified. DHPG-LTD is enhanced by treatment with KN-62, an inhibitor of Ca2+/CaM-dependent protein kinases (CaMKs) (including CaMKII; Schnabel et al., 1999b), which indicates that some form of dephosphorylation of proteins which are phosphorylated by CaMKs may be important to DHPG-induced LTD. It could be that basal

CaMK activity may tend to inhibit DHPGinduced LTD. Another possibility is that DHPG simultaneously induces KN-62 insensitive LTD and a smaller KN-62-sensitive LTP; blockade of the latter leading to the apparent enhancement of the former. DHPG-LTD is also facilitated by inhibitors of PP1 and PP2A, but not affected by inhibition of PP2B (Schnabel et al., 2001), which indicates that some form of phosphorylation of targets of PP1 and PP2A is important to DHPG-induced LTD and that this phosphorylation is actively opposed by dephosphorylation by PP1 or PP2A. DHPG-induced LTD can be reversed by the non-selective mGluR antagonist MCPG (Palmer et al., 1997; Fitzjohn et al., 1998, 1999; Watabe et al., 2002), and by the group I/II mGluR antagonist LY341495 (Fitzjohn et al., 1998; Watabe et al., 2002), but LTD is reestablished following washout of either antagonist. One explanation for the transient effect of the mGluR antagonists could be that endogenous glutamate tonically activates mGluRIs. A second possibility is that mGlu receptor agonists and mGlu receptor antagonists switch the mode of activity of mGluRI (i.e., act as agonists and inverse agonists) with respect to downstream events, i.e., they activate opposing signaling pathways. MCPG-induced reversal of DHPG-induced LTD does not involve activation of CaMKII, PKA, or release of Ca2+ from intracellular stores (Schnabel et al., 1999a, b), but does involve the activation of PP1/PP2A (Schnabel et al., 2001). Thus, DHPG-induced LTD seems to depend on the net activity of some yet to be determined protein kinase(s) and PP1/PP2A. Activation of mGluRs by DHPG switches the receptor-effector cascade into a mode that promotes phosphorylation. Application of mGluRI antagonists can temporarily reverse the phosphorylation mode via transient activation of a cascade that involves dephosphorylation via PP1/PP2A. While DHPG-LTD is induced postsynaptically (group I mGluRs are only located postsynaptically; Conn and Pin, 1997), it has been suggested to be expressed presynaptically as inhibition of voltage-sensitive Ca2+ channels and K+ channels (Watabe et al., 2002; Tan et al., 2003) at Schaffer collateral–CA1 synapses. A presynaptic locus of

488

expression is supported by studies reporting that: (1) DHPG induces a persistent depression of NMDA, as well as AMPA, receptor-mediated components of EPSCs (Watabe et al., 2002), (2) the expression of LTD is accompanied by a change in PPR (Faas et al., 2002; Rouach and Nicoll, 2003; Tan et al., 2003); and (3) a lack of change in postsynaptic sensitivity to AMPA in CA1, measured by focally elicited AMPAR-currents evoked by rapid uncaging of glutamate, following DHPGLTD (Rammes et al., 2003). The depression induced by DHPG is dependent on postsynaptic, GTP-dependent signaling pathways, as it is blocked after replacement of GTP with guanosine-50 -O-(2-thiodiphosphate) (GDPgS; Watabe et al., 2002). (b) Stimulus-induced mGluR-dependent LTD Stimulus-induced (2 Hz, 900 pulses during GABAA-blockade) LTD in adult rats (7 weeks old) shows different mechanisms than DHPGinduced LTD (Otani and Connor, 1998). LTD depends on mGluR, but not NMDAR, activation in adult rats, as it is blocked by the mGluR antagonist MCPG, but not the NMDAR antagonist AP5. An increase in postsynaptic [Ca2+] is needed to induce this form of LTD, since postsynaptic infusion of the Ca2+ chelator BAPTA blocked induction of LTD. A transient (20–30 s) increase in intracellular [Ca2+] could be measured during the 2 Hz stimulation. This [Ca2+] increase was the result of voltage-gated influx, rather than intracellular Ca2+ release triggered by mGluR activation, since it was blocked by hyperpolarization of the neuronal soma to
110 mV, but not by MCPG. Injection of the PKC inhibitor peptide PKC19–36 into the postsynaptic neuron completely blocked LTD, verifying that PKC activation, probably via mGluRI, is necessary for the induction of this form of LTD. Summary: CA1 LTD. Current data on LTD at Schaffer collateral–CA1 synapses is becoming substantial and reveals a synapse where distinct preand postsynaptic cascades cause long-term alterations on both sides of the synapse. Because only the postsynaptic dendritic spine would seem to be privy to Hebbian information about both presynaptic glutamate release and global postsynaptic

voltage levels driven by many other synaptic contacts, it appears that long-term changes are driven by pairing of local spine [Ca2+] and postsynaptic depolarization within a certain range. Postsynaptically, increases in [Ca2+] that preferentially activate PPs and, perhaps independently, AMPAR internalization appear to elicit postsynaptic LTD. Presynaptically, the postsynaptic Ca2+-dependent activation of NOS generates NO which diffuses locally to proximal presynaptic terminals, where it must pair with activation of G protein-coupled receptors that inhibit adenylate cyclase to drive a cyclic GMP/PKG/cyclic ADPribose and Ca2+dependent presynaptic cascade that causes LTD of vesicular transmitter release from the rapidlyrecycling vesicle pool. NMDAR-dependent LTD appears to involve both presynaptic and postsynaptic alterations, while group I mGluR-dependent LTD is expressed purely postsynaptically, and the activation of group II mGluRs recruits presynaptic cascades that may or may not be shared with NMDAR-LTD. Thus, excitatory LTD in field CA1 is a complex of multiple alterations across the synapse. The extensive information now available about relatively rapid biochemical events during the induction of LTD is not, as of yet, matched by sufficient data about whether and how these changes lead to precise stimulation or suppression of particular genes, or of retraction or even complete elimination of synaptic contacts.

LTD in the dentate gyrus Induction LTD in the DG can also be induced by a variety of stimulus protocols. The most common induction protocol is LFS (1–2 Hz) of the medial perforant path (MPP) to induce homosynaptic LTD (in vitro: O’Mara et al., 1995a; in vivo: Abraham, 1996; Trommer et al., 1996; Po¨schel and Manahan-Vaughan, 2007). Heterosynaptic LTD has also been demonstrated at both medial and lateral perforant path (LPP) synapses by high-frequency stimulation (HFS) of the other synaptic input, (Abraham and Goddard, 1983; Doyere et al., 1997). Associative LTD can also be induced in the

489

lateral perforant path, where it seems to share expression mechanisms with heterosynaptic LTD (Christie et al., 1995). Christie et al. (1995) induced primed-associative LTD of the lateral path by alternating high-frequency conditioning bursts to the medial path and single shocks to the lateral path at 100 ms intervals, provided these were administered 10 min after a homosynaptic priming stimulation of the lateral path (5 Hz, 80 pulses). Lin et al. (2006) elicited LTD at lateral perforant path synapses by pairing antidromically-evoked postsynaptic APs (aAPs) with synaptic excitatory postsynaptic potentials (fEPSPs), providing that the fEPSP occurred rapidly after the aAP. Agonist activation of mGluRs can also induce multiple forms of LTD in the DG (O’Mara et al., 1995a; Camodeca et al., 1999; Huang et al., 1999a, b, c; Klausnitzer et al., 2004; Po¨schel and ManahanVaughan, 2005).

mGluR agonist-induced LTD in the medial perforant path Metabotropic GluR activation is reported to be necessary for the induction of LFS-LTD at medial perforant path-DG synapses (O’Mara et al., 1995a; Huang et al., 1997; Camodeca et al., 1999; Klausnitzer et al., 2004; Po¨schel et al., 2005). The mGluR family of heterotrimeric guanosine triphosphate-binding protein (G protein)-coupled receptors (Sugiyama et al., 1987; Tanabe et al., 1992) are divided into three groups: mGluR I, II, and III. Despite their different intracellular coupling, agonist activation of all three groups has surprisingly been shown to result in the induction of LTD (Huang et al., 1999a, b, c; Camodeca et al., 1999; Naie and Manahan-Vaughan, 2005; Po¨schel and ManahanVaughan, 2005) (Fig. 4) Group I mGluRs include the receptor subtypes mGluR1 and mGluR5 (and their splice variants). Group I mGluRs are coupled to a pathway that involves intracellular Ca2+ release from intracellular stores and PKC activation. They produce Gqmediated activation of PLC which hydrolyzes inositol-bisphosphate to IP3 and DAG. DAG, with Ca2+, coactivates protein kinase C, while IP3

causes the release of cytosolic Ca2+ from intracellular stores (Conn and Pin, 1997). The mGluRI agonist DHPG induces an mGluRIdependent LTD at medial perforant path synapses which does not require NMDAR activation (Camodeca et al., 1999). However, this LTD seems to occur only in juvenile/young adult animals (up to 5 weeks of age in rats; Camodeca et al., 1999; Wang et al., 2007). There exist conflicting results concerning the involvement of PKC activation in DHPG-LTD. While an early study reported a strong inhibition (but not complete block) of DHPG-LTD by the PKC inhibitor Bisindolylmaleimide I (Bis I, 1 mM, inhibits a, b, g, e at this concentration; Camodeca et al., 1999), a later study by the same group did not find an effect of Bis I on DHPG-LTD using identical protocols (Rush et al., 2002). Thus, it remains to be determined whether PKC activity is necessary for the induction of DHPG-LTD. Furthermore, it has been reported that DHPG-LTD is dependent on the activation of p38 MAPK (Rush et al., 2002), as well as non-receptor tyrosine kinases (Camodeca et al., 1999), since the p38 MAPK inhibitor SB203580 (4-(4-fluorophenyl)-2-(4-methylsulfonylphenyl)-5-(4pyridyl) imidazol) and non-receptor tyrosine kinase inhibitor lavendustin A each strongly inhibited the induction of DHPG-LTD. Additional support for a role for tyrosine kinases and p38 MAPK in DHPGLTD is the finding that activation of MAPK pathways (including p38 MAPK) by G protein-coupled receptors requires activation of the Src family of non-receptor tyrosine kinases (Luttrell et al., 1999). MAPKs phosphorylate transcription and translation factors, thereby regulating gene expression and PS (Thomas and Huganir, 2004; Kelleher et al., 2004). Consistent with this is a study showing that DHPGLTD is inhibited by PS blockade with either anisomycin or emetine (Naie and Manahan-Vaughan, 2005). It seems that DHPG-LTD is expressed only postsynaptically, as it is not accompanied by any change in paired-pulse depression (Camodeca et al., 1999). Some mGluRI subtypes (mGluR1a, mGluR5a and mGluR5b) stimulate cAMP formation by potentiating AC activation triggered by neurotransmitters that act on GS-coupled receptors (Conn and Pin, 1997; Moldrich et al., 2002). For instance, in cultured striatal neurons, the ability of group I

490

PKC? mGluRI

tyrosine kinase

p38 MAPK mRNAtranscription

mGluRII +

+

+

mRNAtranslation

?

PKC T-VGCC

2+

PKA

Ca

Post Fig. 4. Cascades involved with mGluR-induced LTD at medial perforant path-DG synapses. Ca: Ca2+, MAPK: mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, PKA: protein kinase A, PKC: protein kinase C, Post: postsynapse, VGCC: voltage-gated Ca2+ channels.

mGluRs to potentiate cAMP formation induced by D1 dopamine receptor agonists correlates with their ability to generate IP3, and this potentiation requires PKC activity (Paolillo et al., 1998). Nevertheless, mGluRI potentiation of the AC-cAMPPKA pathway does not seem to play a necessary role in inducing DHPG-LTD, because the PKA inhibitor H-89 does not affect/prevent induction of DHPG-LTD (Camodeca et al., 1999). Group II mGluRs consist of mGluR2 and mGluR3, which inhibit forskolin and receptorinduced increases in [cAMP]. A Gi-type of G protein is probably involved in this coupling (Conn and Pin, 1997). It has also been reported that activation of group II mGluRs potentiates increases in [cAMP] triggered by neurotransmitters that act on Gscoupled receptors in adult hippocampal slices (Conn and Pin, 1997; Moldrich et al., 2002), i.e., this

potentiation is dependent on the coactivation of group II mGluRs plus Gs-coupled receptors. The potentiation of cAMP accumulation seems to be mediated by synergistic interaction between group I and group II mGluRs, or group II mGluRs and b-adrenoceptors, which induces endogenous adenosine release (Moldrich et al., 2002). Adenosine subsequently activates both A2A receptors, which are positively coupled to cAMP, and A1 receptors coupled negatively to adenylate cyclase. Metabotropic GluRII-dependent LTD is induced in the DG by selective agonist activation of mGluRIIs with (2S,20 R,30 R)-2-(20 30 -dicarboxycyclopropyl) glycine (DCG-IV) or (1S,2S,5R,6S)-2aminobicyclo[3.1.0]hexane-2-6-dicarboxylic acid (LY354740): Huang et al., 1999a, b, c; (S)-4-carboxy-3-hydroxyphenylglycine (4C3HPG): Po¨schel and Manahan-Vaughan, 2005), and also by

491

activation of the mGluRII subtype mGluR3 (with N-acetylaspartylglutamate (NAAG): Huang et al., 1999b; Po¨schel and Manahan-Vaughan, 2005). Perfusion with mGluRII agonists in vitro induces a rapid depression of MPP-evoked EPSPs which is maintained in the presence of the agonist but, following washout, is partially reversed to a depression level which is maintained throughout the experiment (LTD was induced lasting at least 1 h following washout; Huang et al., 1999a, b). Intracerebroventricular injection of mGluRII agonists in vivo induces a rapid (mGluRII agonist) or slowly developing depression (mGluR3 agonist) of MPPevoked EPSPs which is maintained throughout the experiment (up to 25 h after LFS; Po¨schel and Manahan-Vaughan, 2005). The rapidly evoked reversible depression of the EPSP by mGluRII agonists in vitro was associated with a change in pairedpulse depression which was not maintained during LTD following washout of the agonist (Huang et al., 1999a, b), indicating that the reversible depression may be expressed via a presynaptic reduction in the probability of transmitter release, but that LTD is induced either by a presynaptic reduction in the number of active release sites, and/or postsynaptic changes. While Huang et al. (1999a) reported a partial block of mGluRII-induced LTD by the NMDAR antagonist AP5, Po¨schel and Manahan-Vaughan (2005) could not confirm a dependence of mGluR3-induced LTD on any NMDAR activation. This may mean that NMDAR blockade affects only mGluR2-dependent LTD. In vitro, mGluRII-induced LTD has been shown to depend on depolarization, since the cessation of test stimulation during mGluRII agonist application prevented the induction of LTD (Huang et al., 1999a). Furthermore, mGluRII-induced LTD, but not the reversible initial depression, in vitro requires the activation of VGCCs, and of PKA or PKC (Huang et al., 1999a, c). The downstream pathways of PKA and PKC have yet to be investigated, but PS does not seem to be required for the induction of mGluRII-induced LTD, since it is not blocked in vivo by the translation-inhibitor anisomycin or the transcription-inhibitor actinomycin D (Po¨schel and Manahan-Vaughan, 2005).

Group III mGluRs comprise the subtypes mGlu4, 6, 7, and 8 (and their splice variants) and are linked to the inhibition of Gs-activated ACs (Conn and Pin, 1997). The induction pathways of Group III mGluR-induced LTD have not been investigated, but it has been shown that Group III mGluR-induced LTD is expressed independently of PS (Naie and Manahan-Vaughan, 2005).

Electrically-induced LTD at medial PP-DG synapses Although an early study reported that the noncompetitive NMDAR open-channel blockers ketamine and phencyclidine were able to block the induction of medial perforant path LTD (Desmond et al., 1991), follow-up studies, using the specific NMDAR antagonist AP5, have not confirmed the dependence of LFS-LTD on NMDAR activation at medial PP-DG synapses (O’Mara et al., 1995a; in vitro: Trommer et al., 1996; Wang et al., 1997; in vivo: Po¨schel and Manahan-Vaughan, 2005, 2007). Nevertheless, postsynaptic intracellular [Ca2+] elevations are a necessary step in the induction of LFS-LTD at medial PP-DG synapses, and can be provided either via Ca2+ influx into dentate granule cells through low voltage-activated Ca2+ channels (T-type), or Ca2+ release from intracellular stores (O’Mara et al., 1995b; Wang et al., 1997), or both. The intracellular stores involved include both group I mGluR-activated IP3-receptor-dependent and ryanodinereceptor-sensitive postsynaptic stores (O’Mara et al., 1995b; Wang et al., 1997). Ryanodine-receptor-sensitive Ca2+ stores are a target of the NO-cGMP pathway via a PKG-dependent activation of a cyclase/hydrolase that generates cyclic ADPribose, a messenger that either enhances or causes release from RyR-sensitive stores which mediate Ca2+-dependent Ca2+ release (CDCR). The NO-cGMP pathway has been shown to play a role in the induction of LFS-LTD at PP-DG synapses, as inhibition of NOS, guanylyl cyclase and postsynaptic PKG all prevent the induction of LFS-LTD (Wu et al., 1997, 1998). Furthermore, cGMP-dependent LTD induced by the type V phosphodiesterase inhibitor zaprinast shows

492

mutual occlusion with LFS-LTD, and zaprinastinduced LTD is dependent on mGluR, but not NMDAR, activation (Wu et al., 1998). Since postsynaptic inhibition of PKG and postsynaptic blockade of ryanodine receptors were reportedly sufficient to strongly inhibit LFS-LTD in the DG, the NO-cGMP pathway seems to be induced postsynaptically in this region, with NO acting not as retrograde messenger as in CA1, but as an intercellular postsynaptic messenger in the DG (Fig. 5). LFS-LTD in the DG appears to be dependent on all subtypes of mGluRs at medial PP-DG synapses. It has been reported that LFS-LTD shows mutual occlusion with mGluRI-induced LTD (Camodeca

et al., 1999), and with mGluRII-induced LTD (Huang et al., 1999a), which suggests common induction and/or maintenance of intracellular pathways shared between a compound LFS-LTD and both mGluR I and II-dependent LTDs. Induction of LFS-LTD fully blocks the subsequent induction of either mGluRI-induced LTD or mGluRIIinduced LTD, whereas the induction of either mGluRI-induced LTD or mGluRII-induced LTD only partially inhibits the subsequent induction of LFS-LTD, suggesting that mGluRIs and mGluRIIs contribute with parallel/independent intracellular pathways to LFS-LTD. Furthermore, LFS-LTD is blocked by antagonists of mGluRIs ((AIDA): Camodeca et al., 1999), mGluRIIs (2S,1S0 ,2S0 -2-

Fig. 5. Cascades involved with LFS-induced LTD at medial perforant path-DG synapses. AC: adenylyl cyclase, Ca: Ca2+, cADPR: cyclic adenosine diphosphate ribose, CaM: calmodulin, cGMP: cyclic guanosine monophosphate, DAG: diacyl glycerol, IP3: inositol1,4,5-trisphosphate, MAPK: mitogen-activated protein kinase, mGluR: metabotropic glutamate receptor, NO: nitric oxide, NOS: NOsynthetase, PIP2: phosphatidylinositol-3,4-bisphosphate, PKA: protein kinase A, PKC: protein kinase C, PKG: protein kinase G, PLC: phospholipase C, Post: postsynapse, VGCC: voltage-gated Ca2+ channels.

493

methyl-2-(20 -carboxycyclopropyl)glycine (MCCG): Huang et al., 1997; NAAG: Po¨schel et al., 2005), and mGluRIIIs ((R.S)-r-cyclopropyl-4-phosphonophenylglycine (CPPG): Klausnitzer et al., 2004). Thus, all mGluR subgroups seem to contribute to the induction of LFS-LTD, despite their different intracellular signaling pathways. Kinases which are part of different mGluR intracellular pathways, such as PKA (Huang et al., 1999a, c), PKC (Wang et al., 1998; Huang et al., 1999a, c) and MAPK (Murray and O’Connor, 2003), have all been reported to be necessary for the induction of LFSLTD. Application of a PKC activator can also directly induce a depression which occludes subsequent induction of LFS-LTD (Wang et al., 1998) supporting a role for PKC in mediating some portion of LFS-LTD at MPP-DG synapses. The monomeric G protein Ras activates certain MAPK pathways (Thomas and Huganir, 2004) and the Ras inhibitor manumycin A has also been shown to significantly attenuate LFS-LTD (Murray and O’Connor, 2004). One stimulation protocol (step depolarization of a DG granule neuron from a holding potential of
70 mV to
50 mV for 1.1 s applied during HFS medial perforant path stimulus trains in the presence of an NMDAR blocker) can also induce LTD which depends only on mGluRI activation (Wu et al., 2001). This form of LTD required membrane depolarization, was mediated by Ca2+ influx via L-type Ca2+ channels plus Ca2+ release from ryanodine-receptor-sensitive intracellular Ca2+ stores, and required activation of both MAPK and PKC (Wu et al., 2001, 2004, Wang et al., 2007). Wu et al. (2001) detected decreases in potency and increases in failure rate after induction of mGluRI-dependent LTD. Although these changes can be attributed to either presynaptic or postsynaptic changes, Wu et al. (2001) argue in favor of a postsynaptic site of expression, stating that a decrease in potency could also be caused by a decrease in the probability of release if more than one synapse was being activated and the increase in failure rate, according to the silent synapse hypothesis of Liao et al. (1995), could also be caused by a postsynaptic change associated with a failure of detection of released transmitter as active synapses are converted into silent synapses (i.e.,

AMPARs are completely removed postsynaptically). This interpretation is supported by the findings of Camodeca et al. (1999) and Zhang et al. (2006), who found that DHPG-LTD in the DG was not accompanied by a change in paired-pulse depression (Camodeca et al., 1999) and group I mGluRs did not contribute to the presynaptic expression of LFS-LTD in field CA1 (Zhang et al., 2006).

Heterosynaptic LTD at lateral perforant path synapses Induction of heterosynaptic LTD in the lateral perforant path of the DG (by HFS stimulation of the medial PP) has been shown in vitro (Christie and Abraham, 1992b) and in vivo (Abraham et al., 2006) to depend on activation of both NMDARs and L-type VGCCs. Furthermore, immediate early gene expression has been associated with the persistence of heterosynaptic LTD at both medial and lateral perforant path synapses on dentate granule cells, i.e., the greatest immediate early gene expression occurred following a stimulus protocol which consistently gave the longest-lasting LTD (Abraham et al., 1994). Induction of associative LTD in the lateral perforant path, which is induced after priming of the lateral path, does not depend on NMDAR activation (Christie and Abraham, 1992a). Nevertheless, heterosynaptic LTD and prime-associative LTD of the lateral path seem to share a common expression mechanism, since they occlude one another (Christie et al., 1995). It remains to be determined what nonNMDARs are involved in either. The one study of spike-timing-dependent LTD at lateral perforant path synapses reported a dependence on NMDAR activation, and blockade by inhibitors of protein kinase C and PP2B (Lin et al., 2006). Morphological changes associated with LTD have been so far investigated in only two studies; Mezey et al. (2004), who reported that heterosynaptic LTD of the medial path of the DG results in an input-specific increase in axodendritic synapse density, whereas LTP at the lateral path was associated with an increase in perforated axospinous synapses in the potentiated area, and Zhou et al.

494

(2004), who reported an NMDAR, CN, and cofilindependent shrinkage of dendritic spines following induction of LTD at Schaffer collateral–CA1 synapses. Thus, the range of morphological alterations that may be associated with LTD, their functional effects on transmission, and the time course of such changes during the progression of LTD, are still unknown. Summary: DG LTD This chapter makes it clear that significantly less is known about LTD in the DG compared to field CA1, or even the hippocampus proper in general. The majority of this evidence points to postsynaptic alterations in MPP-DG LTD, and suggests that an NO/cGMP/PKG cascade produces long-term postsynaptic LTD here, in contrast to CA1. In this respect, MPP-LTD may have more in common with LTD in the striatum, than in field CA1. However, there are enough studies suggesting long-term presynaptic alterations in mGluR3dependent LTD (Po¨schel et al., 2005), and in the actions of brain-derived neurotrophic factor (BDNF) that promote/elicit LTP (Gooney et al., 2002), to leave the question of presynaptic longterm plasticity at MPP-DG synapses open to further study. Likewise, the relative lack of much knowledge about LPP-DG synapse LTD (or LTP) makes this a fertile area of future study. Most reports have examined heterosynaptic LTD of LPPDG synapses, which are likely to involve postsynaptic alterations that depend on the spread of postsynaptic depolarization and activation of voltage-dependent Ca2+ channels, and may differ substantially from homosynaptic LPP LTD. Even more so than in field CA1, the connections between biochemical cascades necessary for the induction of LTD, and eventual synaptic remodeling, are unexplored territory in the DG. Acknowledgments This work was supported by National Institutes of Health Grant R01-NS44421 (to P.K.S.) and National Institute of Drug Abuse Grant R01-HD45754 (to Dr. Peter Mundel).

References Abraham, W.C. (1996) Induction of heterosynaptic and homosynaptic LTD in hippocampal sub-regions in vivo. J. Physiol. Paris., 90: 305–306. Abraham, W.C., Christie, B.R., Logan, B., Lawlor, P. and Dragunow, M. (1994) Immediate early gene expression associated with the persistence of heterosynaptic long-term depression in the hippocampus. Proc. Natl. Acad. Sci. U.S.A., 91: 10049–10053. Abraham, W.C. and Goddard, G.V. (1983) Assymetric relationships between homosynaptic long-term potentiation and heterosynaptic long-term depression. Nature, 305: 717–719. Abraham, W.C., Maxon-Parker, S.E., Irvine, G.I., Logan, B. and Gill, A.I. (2006) Induction and activity-dependent reversal of persistent LTP and LTD in lateral perforant path synapses in vivo. Neurobiol. Learn. Mem., 86: 82–90. Ahmadian, G., Ju, W., Liu, L., Wyszynski, M., Lee, S.H., Dunah, A.W., Taghibiglou, C., Wang, Y., Lu, J., Wong, T.P., Sheng, M. and Wang, Y.T. (2004) Tyrosine phosphorylation of GluR2 is required for insulin-stimulated AMPA receptor endocytosis and LTD. EMBO J., 23: 1040–1050. Alford, S., Frenguelli, B.G. and Collingridge, G.L. (1992) Characterisation of Ca2+ signals induced in hippocampal CA1 neurones by the synaptic activation of NMDA receptors. J. Physiol. Lond., 469: 693–716. Armstrong, D.L. (1989) Calcium channels regulation by calcineurin, a Ca2+-activated phosphatase in mammalian brain. Trends Neurosci., 12: 117–122. Bailey, C.P., Trejos, J.A., Schanne, F.A. and Stanton, P.K. (2003) Pairing elevation of [cyclic GMP] with inhibition of PKA produces long-term depression of glutamate release from isolated rat hippocampal presynaptic terminals. Eur. J. Neurosci., 17: 903–908. Banke, T.G., Bowie, D., Lee, H., Huganir, R.L., Schousboe, A. and Traynelis, S.F. (2000) Control of GluR1 AMPA receptor function by cAMP-dependent protein kinase. J. Neurosci., 20: 89–102. Beattie, E.C., Carroll, R.C., Yu, X., Morishita, W., Yasuda, H., von Zastrow, M. and Malenka, R.C. (2000) Regulation of AMPA receptor endocytosis by a signaling mechanism shared with LTD. Nat. Neurosci., 3: 1291–1300. Bito, H., Deisseroth, K. and Tsien, R.W. (1996) CREB phosphorylation and dephosphorylation: a Ca2+- and stimulus duration-dependent switch for hippocampal gene expression. Cell, 87: 1203–1214. Bolshakov, V.Y., Carboni, L., Cobb, M.H., Siegelbaum, S.A. and Belardetti, F. (2000) Dual MAP kinase pathways mediate opposing forms of long-term plasticity at CA3-CA1 synapse. Nat. Neurosci., 3: 1107–1112. Bolshakov, V.Y. and Siegelbaum, S.A. (1994) Postsynaptic induction and presynaptic expression of hippocampal longterm depression. Science, 264: 1148–1152. Bolshakov, V.Y. and Siegelbaum, S.A. (1995) Hippocampal long-term depression: arachidonic acid as a potential retrograde messenger. Neuropharmacology, 34: 1581–1587.

495 Borsch-Haubold, A.G., Kramer, R.M. and Watson, S.P. (1997) Phosphorylation and activation of cytosolic phospholipase A2 by 38-kDa mitogen-activated protein kinase in collagenstimulated human platelets. Eur. J. Biochem., 245: 751–759. Brandon, E.P., Idzerda, R.L. and McKnight, G.S. (1997) PKA isoforms, neural pathways, and behaviour: making the connection. Curr. Opin. Neurobiol., 7: 397–403. Brandon, E.P., Zhuo, M., Huang, Y.Y., Qi, M., Gerhold, K.A., Burton, K.A., Kandel, E.R., McKnight, G.S. and Idzerda, R.L. (1995) Hippocampal long-term depression and depotentiation are defective in mice carrying a targeted disruption of the gene encoding the RI beta subunit of cAMP-dependent protein kinase. Proc. Natl. Acad. Sci. U.S.A., 92: 8851–8855. Bredt, D.S., Hwang, P.M., Glatt, C.E., Lowenstein, C., Reed, R.R. and Snyder, S.H. (1991) Cloned and expressed nitric oxide synthase structurally resembles cytochrome P-450 reductase. Nature, 351: 714–718. Brindle, P.K. and Montminy, M.R. (1992) The CREB family of transcription activators. Curr. Opin. Genet. Dev., 2: 199–204. Cadd, G.G., Uhler, M.D. and McKnight, G.S. (1990) Holoenzymes of cAMP-dependent protein kinase containing the neural form of type I regulatory subunit have an increased sensitivity to cyclic nucleotides. J. Biol. Chem., 265: 19502–19506. Camodeca, N., Breakwell, N.A., Rowan, M.J. and Anwyl, R. (1999) Induction of LTD by activation of group I mGluR in the dentate gyrus in vitro. Neuropharmacology, 38: 1597–1606. Carroll, R.C., Lissin, D.V., von Zastrow, M., Nicoll, R.A. and Malenka, R.C. (1999) Rapid redistribution of glutamate receptors contributes to long-term depression in hippocampal cultures. Nat. Neurosci., 2: 454–460. Cavalli, V., Vilbois, F., Corti, M., Marcote, M.J., Tamura, K., Karin, M., Arkinstall, S. and Gruenberg, J. (2001) The stressinduced MAP kinase p38 regulates endocytic trafficking via the GDI:Rab5 complex. Mol. Cell, 7: 421–432. Chabot, C., Gagne, J., Giguere, C., Bernard, J., Baudry, M. and Massicotte, G. (1998) Bidirectional modulation of AMPA receptor properties by exogenous phospholipase A2 in the hippocampus. Hippocampus, 8: 299–309. Chaki, S., Muramatsu, M. and Otomo, S. (1994) Involvement of protein kinase C activation in regulation of acetylcholine release from rat hippocampal slices by minaprine. Neurochem. Int., 24: 37–41. Christie, B.R. and Abraham, W.C. (1992a) Priming of associative long-term depression in the dentate gyrus by theta frequency synaptic activity. Neuron, 9: 79–84. Christie, B.R. and Abraham, W.C. (1992b) NMDA-dependent heterosynaptic long-term depression in the dentate gyrus of anaesthetized rats. Synapse, 10: 1–6. Christie, B.R., Stellwagen, D. and Abraham, W.C. (1995) Evidence for common expression mechanisms underlying heterosynaptic and associative long-term depression in the dentate gyrus. J. Neurophysiol., 74: 1244–1247. Clark, J.D., Schievella, A.R., Nalefski, E.A. and Lin, L.L. (1995) Cytosolic phospholipase A2. J. Lipid. Mediat. Cell Signal, 12: 83–117.

Cohen, P. (1989) The structure and regulation of protein phosphatases. Annu. Rev. Biochem., 58: 453–508. Colbran, R.J. (1992) Regulation and role of brain calcium/calmodulin-dependent protein kinase II. Neurochem. Int., 21: 469–497. Conn, P.J. and Pin, J.P. (1997) Pharmacology and functions of metabotropic glutamate receptors. Annu. Rev. Pharmacol. Toxicol., 37: 205–237. Cotman, C.W., Monaghan, D.T. and Ganong, A.H. (1988) Excitatory amino acid neurotransmission: NMDA receptors and Hebb-type synaptic plasticity. Ann. Rev. Neurosci., 11: 61–80. Cummings, J.A., Mulkey, R.M., Nicoll, R.A. and Malenka, R.C. (1996) Ca2+ signaling requirements for long-term depression in the hippocampus. Neuron, 16: 825–833. Cummings, J.A., Nicola, S.M. and Malenka, R.C. (1994) Induction in the rat hippocampus of long-term potentiation (LTP) and long-term depression (LTD) in the presence of a nitric oxide synthase inhibitor. Neurosci. Lett., 176: 110–114. Desmond, N.L., Colbert, C.M., Zhang, D.X. and Levy, W.B. (1991) NMDA receptor antagonists block the induction of long-term depression in the hippocampal dentate gyrus of the anesthetized rat. Brain Res., 552: 93–98. Domenici, M.R., Berretta, N. and Cherubini, E. (1998) Two distinct forms of long-term depression coexist at the mossy fiber-CA3 synapse in the hippocampus during development. Proc. Natl. Acad. Sci. U.S.A., 95: 8310–8315. Doyere, V., Srebro, B. and Laroche, S. (1997) Heterosynaptic LTD and depotentiation in the medial perforant path of the dentate gyrus in the freely moving rat. J. Neurophysiol., 77: 571–578. Dudek, S.M. and Bear, M.F. (1992) Homosynaptic long-term depression in area CA1 of hippocampus and effects of Nmethyl-D-aspartate receptor blockade. Proc. Natl. Acad. Sci. U.S.A., 89: 4363–4367. Ehlers, M.D. (2000) Reinsertion or degradation of AMPA receptors determined by activity-dependent endocytic sorting. Neuron, 28: 511–525. Eliot, L.S., Dudai, Y., Kandel, E.R. and Abrams, T.W. (1989) Ca2+/calmodulin sensitivity may be common to all forms of neural adenylate cyclase. Proc. Natl. Acad. Sci. U.S.A., 86: 9564–9568. Faas, G.C., Adwanikar, H., Gereau IV, R.W. and Saggau, P. (2002) Modulation of presynaptic calcium transients by metabotropic glutamate receptor activation: a differential role in acute depression of synaptic transmission and long-term depression. J. Neurosci., 22: 6885–6890. Feinmark, S.J., Begum, R., Tsvetkov, E., Goussakov, I., Funk, C.D., Siegelbaum, S.A. and Bolshakov, V.Y. (2003) 12-lipoxygenase metabolites of arachidonic acid mediate metabotropic glutamate receptor-dependent long-term depression at hippocampal CA3-CA1 synapses. J. Neurosci., 23: 11427–11435. Fitzjohn, S.M., Bortolotto, Z.A., Palmer, M.J., Doherty, A.J., Ornstein, P.L., Schoepp, D.D., Kingston, A.E., Lodge, D. and Collingridge, G.L. (1998) The potent mGlu receptor antagonist LY341495 identifies roles for both cloned and novel

496 mGlu receptors in hippocampal synaptic plasticity. Neuropharmacology, 37: 1445–1458. Fitzjohn, S.M., Kingston, A.E., Lodge, D. and Collingridge, G.L. (1999) DHPG-induced LTD in area CA1 of juvenile rat hippocampus; characterization and sensitivity to novel mGluR antagonists. Neuropharmacology, 38: 1577–1583. Fitzjohn, S.M., Palmer, M.J., May, J.E., Neeson, A., Morris, S.A. and Collingridge, G.L. (2001) A characterisation of long-term depression induced by metabotropic glutamate receptor activation in the rat hippocampus in vitro. J. Physiol., 537: 421–430. Fitzpatrick, J.S. and Baudry, M. (1994) Blockade of long-term depression in neonatal hippocampal slices by a phospholipase A2 inhibitor. Dev. Brain Res., 78: 81–86. Freeman, E.J., Damron, D.S., Terrian, D.M. and Dorman, R.V. (1991) 12-Lipoxygenase products attenuate the glutamate release and calcium accumulation evoked by depolarization of hippocampal mossy fiber nerve endings. J. Neurochem., 56: 1079–1082. Gage, A.T., Reyes, M. and Stanton, P.K. (1997) Nitric-oxideguanylyl-cyclase-dependent and independent components of multiple forms of long-term synaptic depression. Hippocampus, 7: 286–295. Galione, A., White, A., Willmott, N., Turner, M., Potter, B.V.L. and Watson, S.P. (1993) cGMP mobilizes intracellular Ca21 in sea urchin eggs by stimulating cyclic ADP-ribose synthesis. Nature, 365: 456–459. Gallagher, S.M., Daly, C.A., Bear, M.F. and Huber, K.M. (2004) Extracellular signal-regulated protein kinase activation is required for metabotropic glutamate receptor-dependent long-term depression in hippocampal area CA1. J. Neurosci., 24: 4859–4864. Garbers, D.L. (1990) The guanylyl cyclase receptor family. New Biol., 2: 499–504. Gooney, M., Shaw, K., Kelly, A., O’Mara, S.M. and Lynch, M.A. (2002) Long-term potentiation and spatial learning are associated with increased phosphorylation of TrkB and extracellular signal-regulated kinase (ERK) in the dentate gyrus: evidence for a role for brain-derived neurotrophic factor. Behav. Neurosci., 116: 455–463. Goto, S., Yamamoto, H., Fukunaga, K., Iwasa, T., Matsukado, Y. and Miyamoto, E. (1985) Dephosphorylation of microtubule-associated protein 2, tau factor, and tubulin by calcineurin. J. Neurochem., 45: 276–283. Grewal, S.S., York, R.D. and Stork, P.J. (1999) Extracellularsignal-regulated kinase signalling in neurons. Curr. Opin. Neurobiol., 9: 544–553. Hagiwara, M., Alberts, A., Brindle, P., Meinkoth, J., Feramisco, J., Deng, T., Karin, M., Shenolikar, S. and Montminy, M. (1992) Transcriptional attenuation following cAMP induction requires PP-1-mediated dephosphorylation of CREB. Cell, 70: 105–113. Halpain, S., Hipolito, A. and Saffer, L. (1998) Regulation of Factin stability in dendritic spines by glutamate receptors and calcineurin. J. Neurosci., 18: 9835–9844. Hansel, C., Artola, A. and Singer, W. (1996) Different threshold levels of postsynaptic [Ca2+]i have to be reached to

induce LTP and LTD in neocortical pyramidal cells. J. Physiol. Paris, 90: 317–319. Heynen, A.J., Quinlan, E.M., Bae, D.C. and Bear, M.F. (2000) Bidirectional, activity-dependent regulation of glutamate receptors in the adult hippocampus in vivo. Neuron, 28: 527–536. Hilfiker, S., Pieribone, V.A., Czernik, A.J., Kao, H.T., Augustine, G.J. and Greengard, P. (1999) Synapsins as regulators of neurotransmitter release. Philos. Trans. R. Soc. Lond. B Biol. Sci., 354: 269–279. Hosaka, M., Hammer, R.E. and Sudhof, T.C. (1999) A phospho-switch controls the dynamic association of synapsins with synaptic vesicles. Neuron, 24: 377–387. Hrabetova, S. and Sacktor, T.C. (1996) Bidirectional regulation of protein kinase M zeta in the maintenance of long-term potentiation and long-term depression. J. Neurosci., 16: 5324–5333. Hrabetova, S. and Sacktor, T.C. (2001) Transient translocation of conventional protein kinase C isoforms and persistent downregulation of atypical protein kinase Mzeta in long-term depression. Brain Res. Mol. Brain Res., 95: 146–152. Huang, C.C. and Hsu, K.S. (2006) Sustained activation of metabotropic glutamate receptor 5 and protein tyrosine phosphatases mediate the expression of (S)-3,5-dihydroxyphenylglycine-induced long-term depression in the hippocampal CA1 region. J. Neurochem., 96: 179–194. Huang, C.C., You, J.L., Wu, M.Y. and Hsu, K.S. (2004) Rap1induced p38 mitogen-activated protein kinase activation facilitates AMPA receptor trafficking via the GDI.Rab5 complex. Potential role in (S)-3,5-dihydroxyphenylglyceneinduced long term depression. Biol. Chem., 279: 12286–12292. Huang, K.P. (1989) The mechanism of protein kinase C activation. Trends Neurosci., 12: 425–432. Huang, L., Killbride, J., Rowan, M.J. and Anwyl, R. (1999a) Activation of mGluRII induces LTD via activation of protein kinase A and protein kinase C in the dentate gyrus of the hippocampus in vitro. Neuropharmacology, 38: 73–83. Huang, L.Q., Rowan, M.J. and Anwyl, R. (1997) mGluR II agonist inhibition of LTP induction, and mGluR II antagonist inhibition of LTD induction, in the dentate gyrus in vitro. Neuroreport, 8: 687–693. Huang, L., Rowan, M.J. and Anwyl, R. (1999b) Induction of long-lasting depression by (+)-alpha-methyl-4-carboxyphenylglycine and other group II mGlu receptor ligands in the dentate gyrus of the hippocampus in vitro. Eur. J. Pharmacol., 366: 151–158. Huang, L.Q., Rowan, M.J. and Anwyl, R. (1999c) Role of protein kinases A and C in the induction of mGluR dependent long-term depression in the medial perforant path of the rat dentate gyrus in vitro. Neurosci. Lett., 274: 71–74. Huber, K.M., Kayser, M.S. and Bear, M.F. (2000) Role for rapid dendritic protein synthesis in hippocampal mGluRdependent long-term depression. Science, 288: 1254–1257. Huber, K.M., Roder, J.C. and Bear, M.F. (2001) Chemical induction of mGluR5- and protein synthesis-dependent

497 long-term depression in hippocampal area CA1. J. Neurophysiol., 86: 321–325. Izumi, Y. and Zorumski, C.F. (1993) Nitric oxide and longterm synaptic depression in the rat hippocampus. Neuroreport, 4: 1131–1134. Kamal, A., Ramakers, G.M., Urban, I.J., De Graan, P.N. and Gispen, W.H. (1999) Chemical LTD in the CA1 field of the hippocampus from young and mature rats. Eur. J. Neurosci., 11: 3512–3516. Kameyama, K., Lee, H.K., Bear, M.F. and Huganir, R.L. (1998) Involvement of a postsynaptic protein kinase A substrate in the expression of homosynaptic long-term depression. Neuron, 21: 1163–1175. Kauderer, B.S. and Kandel, E.R. (2000) Capture of a protein synthesis-dependent component of long-term depression. Proc. Natl. Acad. Sci. U.S.A., 97: 13342–13347. Kelleher III, R.J., Govindarajan, A., Jung, H.Y., Kang, H. and Tonegawa, S. (2004) Translational control by MAPK signaling in long-term synaptic plasticity and memory. Cell, 116: 467–479. Kemp, N., McQueen, J., Faulkes, S. and Bashir, Z.I. (2000) Different forms of LTD in the CA1 region of the hippocampus: role of age and stimulus protocol. Eur. J. Neurosci., 12: 360–366. Keranen, L.M., Dutil, E.M. and Newton, A.C. (1995) Protein kinase C is regulated in vivo by three functionally distinct phosphorylations. Curr. Biol., 5: 1394–1403. Kim, C.H., Chung, H.J., Lee, H.K. and Huganir, R.L. (2001) Interaction of the AMPA receptor subunit GluR2/3 with PDZ domains regulates hippocampal long-term depression. Proc. Natl. Acad. Sci. U.S.A., 98: 11725–11730. Klausnitzer, J., Kulla, A. and Manahan-Vaughan, D. (2004) Role of the group III metabotropic glutamate receptor in LTP, depotentiation and LTD in dentate gyrus of freely moving rats. Neuropharmacology, 46: 160–170. Klee, C.B., Crouch, T.H. and Krinks, M.H. (1979) Calcineurin: a calcium and calmodulin-binding protein of the nervous system. Proc. Natl. Acad. Sci. U.S.A., 76: 6270–6273. Kobayashi, M., Manabe, T. and Takahashi, T. (1996) Presynaptic long-term depression at the hippocampal mossy fiberCA3 synapse. Science, 273: 648–650. Krezel, W., Giese, K.P., Silva, A.J. and Chapman, P.F. (1999) Long-term depression is unimpaired in hippocampus of adult alpha CaMKIIT286A mice. Soc. Neurosci. Abst., 25: 987. Lee, H.C. (1997) Mechanisms of calcium signaling by cyclic ADP-ribose and NAADP. Physiol. Rev., 77: 1133–1164. Lee, H.K., Barbarosie, M., Kameyama, K., Bear, M.F. and Huganir, R.L. (2000) Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity. Nature, 405: 955–959. Lee, H.K., Kameyama, K., Huganir, R.L. and Bear, M.F. (1998) NMDA induces long-term synaptic depression and dephosphorylation of the GluR1 subunit of AMPA receptors in hippocampus. Neuron, 21: 1151–1162. Lee, H.K., Takamiya, K., Han, J.S., Man, H., Kim, C.H., Rumbaugh, G., Yu, S., Ding, L., He, C. and Petralia, R.S. (2003) Phosphorylation of the AMPA receptor GluR1

subunit is required for synaptic plasticity and retention of spatial memory. Cell, 112: 631–643. Li, S.T., Kato, K., Tomizawa, K., Matsushita, M., Moriwaki, A., Matsui, H. and Mikoshiba, K. (2002) Calcineurin plays different roles in group II metabotropic glutamate receptorand NMDA receptor-dependent long-term depression. J. Neurosci., 22: 5034–5041. Liao, D., Hessler, N.A. and Malinow, R. (1995) Activation of postsynaptically silent synapses during pairing-induced LTP in CA1 region of hippocampal slice. Nature, 375(6530): 400–404. Lin, J.W., Ju, W., Foster, K., Lee, S.H., Ahmadian, G., Wyszynski, M., Wang, Y.T. and Sheng, M. (2000) Distinct molecular mechanisms and divergent endocytotic pathways of AMPA receptor internalization. Nat. Neurosci., 3: 1282–1290. Lin, Y.W., Yang, H.W., Wang, H.J., Gong, C.L., Chiu, T.H. and Min, M.Y. (2006) Spike-timing-dependent plasticity at resting and conditioned lateral perforant path synapses on granule cells in the dentate gyrus: different roles of N-methylD-aspartate and group I metabotropic glutamate receptors. Eur. J. Neurosci., 23: 2362–2374. Lisman, J. (1989) A mechanism for the Hebb and the anti-Hebb processes underlying learning and memory. Proc. Natl. Acad. Sci. U.S.A., 86: 9574–9578. Luttrell, L.M., Ferguson, S.S., Daaka, Y., Miller, W.E., Maudsley, S., Della Rocca, G.J., Kawakatsu, H., Owada, K., Luttrell, D.K., Caron, M.G. and Lefkowitz, R.J. (1999) Betaarrestin-dependent formation of beta2 adrenergic receptorSrc protein kinase complexes. Science, 283: 655–661. Lynch, G.S., Dunwiddie, T. and Gribkoff, V. (1977) Heterosynaptic depression: a postsynaptic correlate of long-term potentiation. Nature, 266: 737–739. Magee, J.C. and Johnston, D. (1997) A synaptically controlled, associative signal for Hebbian plasticity in hippocampal neurons. Science, 275: 209–213. Malenka, R.C. and Bear, M.F. (2004) LTP and LTD: an embarrassment of riches. Neuron, 44: 5–21. Manahan-Vaughan, D. (1997) Group 1 and 2 metabotropic glutamate receptors play differential roles in hippocampal long-term depression and long-term potentiation in freely moving rats. J. Neurosci., 17: 3303–3311. Manahan-Vaughan, D. and Reymann, K.G. (1995) 1S,3RACPD dose-dependently induces a slow-onset potentiation in the rat hippocampal CA1 region in vivo. Neuropharm., 34: 1103–1105. Manahan-Vaughan, D., Kulla, A. and Frey, J.U. (2000) Requirement of translation but not transcription for the maintenance of long-term depression in the CA1 region of freely moving rats. J. Neurosci., 20: 8572–8576. Markram, H., Lubke, J., Frotscher, M. and Sakmann, B. (1997) Regulation of synaptic efficacy by coincidence of postsynaptic APs and EPSPs. Science, 275: 213–215. McDonald, B.J., Chung, H.J. and Huganir, R.L. (2001) Identification of protein kinase C phosphorylation sites within the AMPA receptor GluR2 subunit. Neuropharmacology, 41: 672–679.

498 McLauchlan, H., Newell, J., Morrice, N., Osborne, A., West, M. and Smythe, E. (1998) A novel role for Rab5-GDI in ligand sequestration into clathrin-coated pits. Curr. Biol., 8: 34–45. Meszaros, L.G., Bak, J. and Chu, A. (1993) Cyclic ADP-ribose as an endogenous regulator of the non-skeletal type ryanodine receptor Ca2+ channel. Nature, 364: 76–79. Mezey, S., Doyere, V., De Souza, I., Harrison, E., Cambon, K., Kendal, C.E., Davies, H., Laroche, S. and Stewart, M.G. (2004) Long-term synaptic morphometry changes after induction of long-term potentiation and long-term depression in the dentate gyrus of awake rats are not simply mirror phenomena. Eur. J. Neurosci., 19: 2310–2318. Moult, P.R., Schnabel, R., Kilpatrick, I.C., Bashir, Z.I. and Collingridge, G.L. (2002) Tyrosine dephosphorylation underlies DHPG-induced LTD. Neuropharmacology, 43: 175–180. Mulkey, R.M., Endo, S., Shenolikar, S. and Malenka, R.C. (1994) Involvement of a calcineurin/inhibitor-1 phosphatase cascade in hippocampal long-term depression. Nature, 369: 486–488. Mulkey, R.M., Herron, C.E. and Malenka, R.C. (1993) An essential role for protein phosphatases in hippocampal longterm depression. Science, 261: 1051–1055. Mulkey, R.M. and Malenka, R.C. (1992) Mechanisms underlying induction of homosynaptic long-term depression in area CA1 of the hippocampus. Neuron, 9: 967–975. Murray, H.J. and O’Connor, J.J. (2003) A role for COX-2 and p38 mitogen activated protein kinase in long-term depression in the rat dentate gyrus in vitro. Neuropharmacology, 44: 374–380. Murray, H.J. and O’Connor, J.J. (2004) A role for monomeric G-proteins in synaptic plasticity in the rat dentate gyrus in vitro. Brain Res., 1000: 85–91. Naie, K. and Manahan-Vaughan, D. (2005) Investigations of the protein synthesis dependency of mGluR-induced longterm depression in the dentate gyrus of freely moving rats. Neuropharmacology, 49: 35–44. Nicholls, R.E., Zhang, X.L., Bailey, C.P., Conklin, B.R., Kandel, E.R. and Stanton, P.K. (2006) mGluR2 acts through inhibitory Galpha subunits to regulate transmission and long-term plasticity at hippocampal mossy fiber-CA3 synapses. Proc. Natl. Acad. Sci. U.S.A., 103: 6380–6385. Nicoll, R.A., Oliet, S.H. and Malenka, R.C. (1998) NMDA receptor-dependent and metabotropic glutamate receptordependent forms of long-term depression coexist in CA1 hippocampal pyramidal cells. Neurobiol. Learn. Mem., 70: 62–72. Normandin, M., Gagne, J., Bernard, J., Lie, R., Miceli, D., Baudry, M. and Massicotte, G. (1996) Involvement of the 12lipoxygenase pathway of arachidonic acid metabolism in homosynaptic long term depression of the rat hippocampus. Brain Res., 730: 40–46. Nosyreva, E.D. and Huber, K.M. (2005) Developmental switch in synaptic mechanisms of hippocampal metabotropic glutamate receptor-dependent long-term depression. J. Neurosci., 25: 2992–3001.

Oliet, S.H., Malenka, R.C. and Nicoll, R.A. (1997) Two distinct forms of long-term depression coexist in CA1 hippocampal pyramidal cells. Neuron, 18: 969–982. O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995a) Metabotropic glutamate receptor-induced homosynaptic long-term depression and depotentiation in the dentate gyrus of the rat hippocampus in vitro. Neuropharmacology, 34: 983–989. O’Mara, S.M., Rowan, M.J. and Anwyl, R. (1995b) Dantrolene inhibits long-term depression and depotentiation of synaptic transmission in the rat dentate gyrus. Neuroscience, 68: 621–624. Otani, S. and Connor, J.A. (1998) Requirement of rapid Ca2+ entry and synaptic activation of metabotropic glutamate receptors for the induction of long-term depression in adult rat hippocampus. J. Physiol., 511: 761–770. Palmer, M.J., Irving, A.J., Seabrook, G.R., Jane, D.E. and Collingridge, G.L. (1997) The group I mGlu receptor agonist DHPG induces a novel form of LTD in the CA1 region of the hippocampus. Neuropharmacology, 36: 1517–1532. Paolillo, M., Montecucco, A., Zanassi, P. and Schinelli, S. (1998) Potentiation of dopamine-induced cAMP formation by group I metabotropic glutamate receptors via protein kinase C in cultured striatal neurons. Eur. J. Neurosci., 10: 1937–1945. Parsley, S.L., Pilgram, S.M., Soto, F., Giese, K.P. and Edwards, F.A. (2007) Enriching the environment of alphaCaMKIIT286A mutant mice reveals that LTD occurs in memory processing but must be subsequently reversed by LTP. Learn Mem., 14: 75–83. Piomelli, D., Wang, J.K., Sihra, T.S., Nairn, A.C., Czernik, A.J. and Greengard, P. (1989) Inhibition of Ca2+/calmodulin-dependent protein kinase 11 by arachidonic acid and its metabolites. Proc. Natl. Acad. Sci. U.S.A., 86: 8550–8554. Po¨schel, B. and Manahan-Vaughan, D. (2005) Group II mGluR-induced long term depression in the dentate gyrus in vivo is NMDA receptor-independent and does not require protein synthesis. Neuropharmacology, 49: 1–12. Po¨schel, B. and Manahan-Vaughan, D. (2007) Persistent (424 h) long-term depression in the dentate gyrus of freely moving rats is not dependent on activation of NMDA receptors, L-type voltage-gated calcium channels or protein synthesis. Neuropharmacology, 52: 46–54. Po¨schel, B., Wroblewska, B., Heinemann, U. and ManahanVaughan, D. (2005) The metabotropic glutamate receptor mGluR3 is critically required for hippocampal long-term depression and modulates long-term potentiation in the dentate gyrus of freely moving rats. Cereb. Cortex, 15: 1414–1423. Qi, M., Zhuo, M., Skalhegg, B.S., Brandon, E.P., Kandel, E.R., McKnight, G.S. and Idzerda, R.L. (1996) Impaired hippocampal plasticity in mice lacking the Cbeta1 catalytic subunit of cAMP-dependent protein kinase. Proc. Natl. Acad. Sci. U.S.A., 93: 1571–1576. Ramakers, G.M., Heinen, K., Gispen, W.H. and de Graan, P.N. (2000) Long term depression in the CA1 field is associated with a transient decrease in pre- and postsynaptic PKC substrate phosphorylation. J. Biol. Chem., 275: 28682–28687.

499 Rammes, G., Palmer, M., Eder, M., Dodt, H.U., Zieglgansberger, W. and Collingridge, G.L. (2003) Activation of mGlu receptors induces LTD without affecting postsynaptic sensitivity of CA1 neurons in rat hippocampal slices. J. Physiol., 546: 455–460. Reyes, M. and Stanton, P.K. (1996) Induction of hippocampal long-term depression requires release of Ca2+ from separate presynaptic and postsynaptic intracellular stores. J. Neurosci., 16: 5951–5960. Reyes-Harde, M., Empson, R., Potter, B.V., Galione, A. and Stanton, P.K. (1999a) Evidence of a role for cyclic ADPribose in long-term synaptic depression in hippocampus. Proc. Natl. Acad. Sci. U.S.A., 96: 4061–4066. Reyes-Harde, M., Potter, B.V., Galione, A. and Stanton, P.K. (1999b) Induction of hippocampal LTD requires nitric-oxidestimulated PKG activity and Ca2+ release from cyclic ADPribose-sensitive stores. J. Neurophysiol., 82: 1569–1576. Roberson, E.D., English, J.D., Adams, J.P., Selcher, J.C., Kondratick, C. and Sweatt, J.D. (1999) The mitogen-activated protein kinase cascade couples PKA and PKC to cAMP response element binding protein phosphorylation in area CA1 of hippocampus. J. Neurosci., 19: 4337–4348. Rouach, N. and Nicoll, R.A. (2003) Endocannabinoids contribute to short-term but not long-term mGluR-induced depression in the hippocampus. Eur. J. Neurosci., 18: 1017–1020. Rush, A.M., Wu, J., Rowan, M.J. and Anwy, R. (2002) Group I metabotropic glutamate receptor (mGluR)-dependent longterm depression mediated via p38 mitogen-activated protein kinase is inhibited by previous high-frequency stimulation and activation of mGluRs and protein kinase C in the rat dentate gyrus in vitro. J. Neurosci., 22: 6121–6128. Sajikumar, S. and Frey, J.U. (2003) Anisomycin inhibits the late maintenance of long-term depression in rat hippocampal slices in vitro. Neurosci. Lett., 338: 147–150. Santschi, L., Reyes-Harde, M. and Stanton, P.K. (1999) Chemically induced, activity-independent LTD elicited by simultaneous activation of PKG and inhibition of PKA. J. Neurophysiol., 82: 1577–1589. Santschi, L.A., Zhang, X.L. and Stanton, P.K. (2006) Activation of receptors negatively coupled to adenylate cyclase is required for induction of long-term synaptic depression at Schaffer collateral-CA1 synapses. J. Neurobiol., 66: 205–219. Schnabel, R., Kilpatrick, I.C. and Collingridge, G.L. (1999a) An investigation into signal transduction mechanisms involved in DHPG-induced LTD in the CA1 region of the hippocampus. Neuropharmacology, 38: 1585–1596. Schnabel, R., Kilpatrick, I.C. and Collingridge, G.L. (2001) Protein phosphatase inhibitors facilitate DHPG-induced LTD in the CA1 region of the hippocampus. Br. J. Pharmacol., 132: 1095–1101. Schnabel, R., Palmer, M.J., Kilpatrick, I.C. and Collingridge, G.L. (1999b) A CaMKII inhibitor, KN-62, facilitates DHPG-induced LTD in the CA1 region of the hippocampus. Neuropharmacology, 38: 605–608. Schwartz, J.H. (1993) Cognitive kinases. Proc. Natl. Acad. Sci. U.S.A., 90: 8310–8313.

Seidenman, K.J., Steinberg, J.P., Huganir, R. and Malinow, R. (2003) Glutamate receptor subunit 2 Serine 880 phosphorylation modulates synaptic transmission and mediates plasticity in CA1 pyramidal cells. J. Neurosci., 23: 9220–9228. Shi, J., Townsend, M. and Constantine-Paton, M. (2000) Activity-dependent induction of tonic calcineurin activity mediates a rapid developmental downregulation of NMDA receptor currents. Neuron, 28: 103–114. Shields, S.M., Ingebritsen, T.S. and Kelly, P.T. (1985) Identification of protein phosphatase 1 in synaptic junctions: dephosphorylation of endogenous calmodulin-dependent kinase II and synapse-enriched phosphoproteins. J. Neurosci., 5: 3414–3422. Snyder, E.M., Philpot, B.D., Huber, K.M., Dong, X., Fallon, J.R. and Bear, M.F. (2001) Internalization of ionotropic glutamate receptors in response to mGluR activation. Nat. Neurosci., 4: 1079–1085. Son, H., Hawkins, R.D., Martin, K., Kiebler, M., Huang, P.L., Fishman, M.C. and Kandel, E.R. (1996) Long-term potentiation is reduced in mice that are doubly mutant in endothelial and neuronal nitric oxide synthase. Cell, 87: 1015–1023. Stanton, P.K. (1995) Transient protein kinase C activation primes long-term depression and suppresses long-term potentiation of synaptic transmission in hippocampus. Proc. Natl. Acad. Sci. U.S.A., 92: 1724–1728. Stanton, P.K. (1996) Phospholipase A2 activation is not required for long-term synaptic depression. Eur. J. Pharmacol., 273: R7–R9. Stanton, P.K. and Gage, A.T. (1996) Distinct synaptic loci of Ca2+/calmodulin-dependent protein kinase II necessary for long-term potentiation and depression. J. Neurophysiol., 76: 2097–2101. Stanton, P.K., Heinemann, U. and Muller, W. (2001) FM1-43 imaging reveals cGMP-dependent long-term depression of presynaptic transmitter release. J. Neurosci., 21: RC167. Stanton, P.K. and Sarvey, J.M. (1984) Blockade of long-term potentiation in rat hippocampal CA1 region by inhibitors of protein synthesis. J. Neurosci., 4: 3080–3088. Stanton, P.K. and Sejnowski, T.J. (1989) Associative long-term depression in the hippocampus induced by hebbian covariance. Nature, 339: 215–218. Stanton, P.K., Winterer, J., Bailey, C.P., Kyrozis, A., Raginov, I., Laube, G., Veh, R.W., Nguyen, C.Q. and Muller, W. (2003) Long-term depression of presynaptic release from the readily releasable vesicle pool induced by NMDA receptordependent retrograde nitric oxide. J. Neurosci., 23: 5936–5944. Stevens, C.F., Tonegawa, S. and Wang, Y. (1994) The role of calcium-calmodulin kinase II in three forms of synaptic plasticity. Curr. Biol., 4: 687–693. Strack, S., Barban, M.A., Wadzinski, B.E. and Colbran, R.J. (1997) Differential inactivation of postsynaptic density-associated and soluble Ca2+/calmodulin-dependent protein kinase II by protein phosphatases 1 and 2A. J. Neurochem., 68: 2119–2128.

500 Sugiyama, H., Ito, I. and Hirono, C. (1987) A new type of glutamate receptor linked to inositol phospholipid metabolism. Nature, 325: 531–533. Tan, Y., Hori, N. and Carpenter, D.O. (2003) The mechanism of presynaptic long-term depression mediated by group I metabotropic glutamate receptors. Cell Mol. Neurobiol., 23: 187–203. Tanabe, Y., Masu, M., Ishii, T., Shigemoto, R. and Nakanishi, S. (1992) A family of metabotropic glutamate receptors. Neuron, 8: 169–179. Thiels, E., Kanterewicz, B.I., Knapp, L.T., Barrionuevo, G. and Klann, E. (2000) Protein phosphatase-mediated regulation of protein kinase C during long-term depression in the adult hippocampus in vivo. J. Neurosci., 20: 7199–7207. Thiels, E., Kanterewicz, B.I., Norman, E.D., Trzaskos, J.M. and Klann, E. (2002) Long-term depression in the adult hippocampus in vivo involves activation of extracellular signalregulated kinase and phosphorylation of Elk-1. J. Neurosci., 22: 2054–2062. Thiels, E., Norman, E.D., Barrionuevo, C. and Klann, E. (1998) Transient and persistent increases in protein phosphatase activity during long-term depression in the adult hippocampus in vivo. Neuroscience, 86: 1023–1029. Thomas, G.M. and Huganir, R.L. (2004) MAPK cascade signalling and synaptic plasticity. Nat. Rev. Neurosci., 5: 173–183. Trommer, B.L., Liu, Y.B. and Pasternak, J.F. (1996) Longterm depression at the medial perforant path-granule cell synapse in developing rat dentate gyrus. Brain Res. Dev. Brain Res., 96: 97–108. Wang, Q., Chang, L., Rowan, M.J. and Anwyl, R. (2007) Developmental dependence, the role of the kinases p38 MAPK and PKC, and the involvement of tumor necrosis factor-R1 in the induction of mGlu-5 LTD in the dentate gyrus. Neuroscience, 144: 110–118. Wang, Y., Rowan, M.J. and Anwyl, R. (1997) Induction of LTD in the dentate gyrus in vitro is NMDA receptor independent, but dependent on Ca2+ influx via low-voltageactivated Ca2+ channels and release of Ca2+ from intracellular stores. J. Neurophysiol., 77: 812–825. Wang, Y., Wu, J., Rowan, M.J. and Anywl, R. (1998) Role of protein kinase C in the induction of homosynaptic long-term depression by brief low frequency stimulation in the dentate

gyrus of the rat hippocampus in vitro. J. Physiol., 513: 467–475. Watabe, A.M., Carlisle, H.J. and O’Dell, T.J. (2002) Postsynaptic induction and presynaptic expression of group 1 mGluR-dependent LTD in the hippocampal CA1 region. J. Neurophysiol., 87: 1395–1403. Wu, J., Rowan, M.J. and Anwyl, R. (2004) Synaptically stimulated induction of group I metabotropic glutamate receptordependent long-term depression and depotentiation is inhibited by prior activation of metabotropic glutamate receptors and protein kinase C. Neuroscience, 123: 507–514. Wu, J., Rush, A., Rowan, M.J. and Anwyl, R. (2001) NMDA receptor- and metabotropic glutamate receptor-dependent synaptic plasticity induced by high frequency stimulation in the rat dentate gyrus in vitro. J. Physiol., 533: 745–755. Wu, J., Wang, Y., Rowan, M.J. and Anwyl, R. (1997) Evidence for involvement of the neuronal isoform of nitric oxide synthase during induction of long-term potentiation and longterm depression in the rat dentate gyrus in vitro. Neuroscience, 78: 393–398. Wu, J., Wang, Y., Rowan, M.J. and Anwyl, R. (1998) Evidence for involvement of the cGMP-protein kinase G signaling system in the induction of long-term depression, but not long-term potentiation, in the dentate gyrus in vitro. J. Neurosci., 18: 3589–3596. Xiao, M.Y., Zhou, Q. and Nicoll, R.A. (2001) Metabotropic glutamate receptor activation causes a rapid redistribution of AMPA receptors. Neuropharmacology, 41: 664–671. Yang, S.N., Tang, Y.G. and Zucker, R.S. (1999) Selective induction of LTP and LTD by postsynaptic [Ca2]i elevation. J. Neurophysiol., 81: 781–787. Zakharenko, S.S., Zablow, L. and Siegelbaum, S.A. (2002) Altered presynaptic vesicle release and cycling during mGluR-dependent LTD. Neuron, 35: 1099–1110. Zhang, X.L., Zhou, Z.Y., Winterer, J., Muller, W. and Stanton, P.K. (2006) NMDA-dependent, but not group I metabotropic glutamate receptor-dependent, long-term depression at Schaffer collateral-CA1 synapses is associated with long-term reduction of release from the rapidly recycling presynaptic vesicle pool. J. Neurosci., 26: 10270–10280. Zhou, Q., Homma, K.J. and Poo, M.M. (2004) Shrinkage of dendritic spines associated with long-term depression of hippocampal synapses. Neuron, 44: 749–757.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 27

Structural reorganization of the dentate gyrus following entorhinal denervation: species differences between rat and mouse Thomas Deller, Domenico Del Turco, Angelika Rappert and Ingo Bechmann Institute of Clinical Neuroanatomy, J.W. Goethe-University, Theodor-Stern-Kai 7, D-60590 Frankfurt/Main, Germany

Abstract: Deafferentation of the dentate gyrus by unilateral entorhinal cortex lesion or unilateral perforant pathway transection is a classical model to study the response of the central nervous system (CNS) to denervation. This model has been extensively characterized in the rat to clarify mechanisms underlying denervation-induced gliosis, transneuronal degeneration of denervated neurons, and collateral sprouting of surviving axons. As a result, candidate molecules have been identified which could regulate these changes, but a causal link between these molecules and the postlesional changes has not yet been demonstrated. To this end, mutant mice are currently studied by many groups. A tacit assumption is that data from the rat can be generalized to the mouse, and fundamental species differences in hippocampal architecture and the fiber systems involved in sprouting are often ignored. In this review, we will (1) provide an overview of some of the basics and technical aspects of the entorhinal denervation model, (2) identify anatomical species differences between rats and mice and will point out their relevance for the axonal reorganization process, (3) describe glial and local inflammatory changes, (4) consider transneuronal changes of denervated dentate neurons and the potential role of reactive glia in this context, and (5) summarize the differences in the reorganization of the dentate gyrus between the two species. Finally, we will discuss the use of the entorhinal denervation model in mutant mice. Keywords: entorhinal cortex lesion; perforant pathway transection; sprouting; transneuronal degeneration; glia; inflammation; regeneration removal of the cellular debris, surviving nonlesioned axons sprout and reinnervate the denervated target cells, and denervated neurons remodel their dendritic arbor in response to denervation and reinnervation. In the past, much effort has been devoted to understanding these changes at the molecular level. The rationale for this is that the denervation of brain areas is a very common and detrimental side effect of all kinds of brain lesions (Steward, 1994b). It can occur following traumatic brain damage, ischemia, inflammation,

Introduction Injuries to the central nervous system (CNS) cause local damage at the lesion site as well as denervation of connected brain regions. In response to this denervation, complex cellular changes occur in the denervated brain area (see Steward, 1994b, for review): Glial cells react and participate in the Corresponding author. Tel.: +49-69-6301-6361; Fax: +49-69-6301-6425; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63027-1

501

502

and neurodegeneration and may contribute significantly to the severity of the clinical symptoms of these and other brain diseases. It is one of the motivations of the scientists working in the field today that understanding the reorganization processes in denervated brain areas and learning how to influence them in a beneficial fashion may lead to new treatment strategies for brain diseases. One of the classical model systems used to study the reorganization of denervated brain regions is the reorganization of the rat dentate gyrus following entorhinal denervation (Lynch and Cotman, 1975; Cotman and Nadler, 1978; Gall and Lynch, 1980; Gall et al., 1986; Steward, 1991; Deller and Frotscher, 1997; Frotscher et al., 1997; CollazosCastro and Nieto-Sampedro, 2001; Deller et al., 2001; Ramirez, 2001; Savaskan and Nitsch, 2001). This model is well established, and structural, functional, and molecular changes have been studied in detail over the years. A fairly large number of candidate molecules have been identified that could regulate the reorganization of the dentate gyrus, including growth-associated molecules, neurotrophic factors, cell adhesion molecules, extracellular matrix molecules, and a variety of cytokines (see Deller and Frotscher, 1997; Collazos-Castro and Nieto-Sampedro, 2001; Deller et al., 2001; Savaskan and Nitsch, 2001, for review). In the majority of studies, however, correlations between molecular changes and the reorganization process were reported, so it remains to be seen that these candidate molecules are causally linked to the postlesional reorganization process. An attractive experimental strategy to address these issues of causality is to use genetically engineered mice and to analyze the reorganization of the dentate gyrus following entorhinal denervation in these mutants (Fagan et al., 1998; Finsen et al., 1999; Jensen et al., 1999; Jensen et al., 2000b; Del Turco et al., 2001, 2003; White et al., 2001a, b; Kadish and van Groen, 2003; Rappert et al., 2004). Although the entorhinal denervation model has now frequently been used in mice, the reorganization of the denervated dentate gyrus has only been partially characterized in this species. In fact, in the absence of data from the mouse, results obtained after entorhinal denervation in mice have

often been interpreted solely on the basis of data obtained in rats. It is assumed that the reorganization of the mouse dentate gyrus is more or less similar to that of the rat dentate gyrus following entorhinal deafferentation. However, this assumption may not be true, given the important biological differences between these two species and even between different mouse strains. Therefore, it may be prudent to keep data from the two species separate. In this review, we will compare the reorganization of the dentate gyrus following entorhinal denervation in rats and mice. We will first give an overview over some of the fundamentals and technical aspects of the model system. Second, we will identify species differences between the non-lesioned dentate gyrus of rats and mice and will point out their relevance for the axonal reorganization process. Third, glial changes occurring in the dentate gyrus of both species after denervation will be described and considered as a local inflammatory response of the brain. Fourth, transneuronal changes of denervated dentate neurons will be considered, and the potential role of reactive glia in this context will be reviewed. Fifth, differences in the reorganization of the dentate gyrus between the two species will be summarized and, finally, the use of the entorhinal lesion model in mutant mice will be discussed.

Entorhinal denervation in rats and mice: technical aspects The entorhinal denervation model was established in the 1970s, when it was first shown that unilateral removal of entorhinal afferents to the rat dentate gyrus results in degenerative and regenerative changes in the so-called ‘‘entorhinal’’ zone of the molecular layer, i.e., the outer two-thirds of the molecular layer (Lynch et al., 1972; Cotman and Nadler, 1978). This model is considered particularly useful, because of the relatively simple and highly laminated cyto- and fiber architecture of the dentate gyrus, which makes it possible to distinguish between denervated and non-denervated layers and enables researchers to differentiate between the various cellular changes that occur within a denervated layer or in its vicinity (Cotman

503

and Nadler, 1978; Steward, 1991; Deller et al., 2001). A confusing element of the entorhinal denervation model in rats is the fact that the method to produce the lesion has changed over the years. In the beginning, entorhinal cortex lesions (ECL) were performed using electrolytic damage to the entorhinal cortex itself, thereby removing the cells of origin of the perforant pathway and denervating the dentate gyrus (Lynch et al., 1972; Steward et al., 1974; Zimmer and Hjorth-Simonsen, 1975; Rose et al., 1976; Gall et al., 1979a, b; Poirier et al., 1990; Parent et al., 1993; Guthrie et al., 1995, 1997; Jucker et al., 1996). Later, other methods were also used, for example, mechanical or electrolytic transection of the perforant pathway (Benowitz et al., 1990; Nitsch and Frotscher, 1992; Nitsch et al., 1992; Beck et al., 1993; Gwag et al., 1994; Deller et al., 1995a) or chemical lesions (Ueki et al., 1996). Often, these lesioning techniques are regarded as more or less equivalent, since all of them result in a robust denervation of the dentate gyrus. However, they may differ in their acute and chronic effects on the dentate gyrus: whereas electrolytic lesions strongly stimulate the perforant pathway and may have an epileptogenic effect, non-electrolytic lesions are less likely to do so (Dasheiff and McNamara, 1982; Campbell et al., 1984; Kelley and Steward, 1996a). Since a variety of molecules and genes are regulated by activity, gene expression — at least during the initial phase of the reorganization process — could be influenced by the lesioning technique (Kelley and Steward, 1996b, 1998; Steward et al., 1997). Moreover, as the extent of necrotic and apoptotic degeneration certainly varies between the different methods, it is likely that there are also fundamental differences with regard to the induced inflammatory and immune responses. In addition to such differences relating to the lesioning techniques, differences in rat strains were rarely considered. In summary, data from the rat ECL/perforant pathway transection model are plentiful, yet some caution is warranted when results from two labs using different lesioning techniques or different rat strains are compared. In mice, mechanical transection of the perforant pathway using wire and plate knifes has been the

preferred method to de-entorhinate the mouse dentate gyrus (Fig. 1; Finsen et al., 1999; Jensen et al., 1999, 2000b; Drojdahl et al., 2002, 2004; Ying et al., 2002; Del Turco et al., 2003; Kovac et al., 2004; Rappert et al., 2004; Nielsen et al., 2006; Pedersen et al., 2006). Chemical lesions of the entorhinal cortex itself were less frequently performed (Kadish and van Groen, 2002, 2003; Kadish et al., 2002). Thus, the lesioning techniques differ between rats and mice; in mice the mechanical perforant pathway transection is the most common, whereas electrolytic lesions are more typical in rats. Finally, the unilateral ECL/perforant pathway transection model should also be distinguished from several related models that have also been employed to study reactive changes in the denervated dentate gyrus in the past. Among these models are bilateral ECL (Lynch et al., 1976; Hardman et al., 1997; Faillace et al., 2006), a combination of ECL with fimbria-fornix-transection (Beck et al., 1993; Peterson, 1994), and fimbria-fornix-transection without ECL (Hjorth-Simonsen and Jeune, 1972; Lynch et al., 1974). All of these models remove additional or neurochemically different projection systems to the dentate gyrus, resulting in an altered reorganization process compared to unilateral ECL/perforant pathway transection.

The non-denervated dentate gyrus: neuroanatomical species differences The normal anatomy of the dentate gyrus is the basis for a meaningful interpretation of experimentally induced changes in this brain region. In the rat, the neuroanatomy of the dentate gyrus has been thoroughly investigated during the last 50 years. The cellular architecture of neuronal and non-neuronal cells, the connectivity, and the basic neurochemistry of the dentate gyrus were studied using anatomical techniques for cellular labeling, retro- and anterograde tracing, and immunohistochemistry at the light- and electron-microscopic level. These neuroanatomical facts have been reviewed in numerous papers and book chapters (e.g., Frotscher, 1988; Amaral and Witter, 1995) as well as in other chapters of this book. In contrast, the mouse dentate

504

Fig. 1. The mouse dentate gyrus after entorhinal denervation. (A, B) Fink-Heimer degeneration stain in a control mouse hippocampus (A) and 5 days after entorhinal denervation (B). Note darkly stained degeneration product in the denervated zone of the molecular layer (B, arrowheads). (C, D) Fluoro-Jade C degeneration stain of the contralateral (C) and ipsilateral (D) mouse dentate gyrus four days after denervation. Note increased reactivity throughout the denervated portion of the molecular layer (D). (E, F) Acetylcholinesterase (AChE) staining of the control mouse dentate gyrus (E) and 6 weeks after entorhinal denervation (F). A dense AChEpositive fiber band is present in the outer molecular layer (OML) and middle molecular layer (MML; arrowheads). EC, entorhinal cortex; CA1, CA3, Hippocampal subfields; GCL, granule cell layer; IML, inner molecular layer; ML; molecular layer; H, hilus. Scale bars: A, B: 300 mm; C–F: 100 mm.

505

gyrus has not been as thoroughly studied. Although data exist on its general cellular and fiber architecture (e.g., Stanfield and Cowan, 1979a, b; Stanfield et al., 1979; Soriano et al., 1994; Supe`r and Soriano, 1994; Deller et al., 1999a, b, 2002; Drakew et al., 2002; Gebhardt et al., 2002; Del Turco et al., 2004), data on the connectivity of identified cells are scarce. Only a handful of tracing studies were performed in mice, in which tracing was combined with electron microscopy to demonstrate synaptic connections of identified neurons (Blasco-Ibanez and Freund, 1997; Deller et al., 1999b; Del Turco et al., 2003; Otal et al., 2006). Similarly, the topographical organization of the afferent projection systems, the cells of origin of the various pathways, and the neurochemistry of the target cells have not been studied in detail. Strain differences have rarely been considered. Thus, neuroanatomical data on the non-denervated mouse dentate gyrus are often incomplete and lack the depth and precision of the rat data. Nevertheless, on the basis of the available neuroanatomical data several anatomical differences are evident, which are likely to result in significant differences in the reorganization of the dentate gyrus following entorhinal denervation. One relevant difference concerns the ipsilateral entorhinal projection to the dentate gyrus, i. e., the projection system that is lost as a consequence of ECL/perforant pathway transection. This projection has been studied in rats (e.g. Blackstad, 1958; Amaral and Witter, 1989; Deller, 1998) and mice (Deller et al., 1999a; van Groen et al., 2002, 2003) using anterograde tracing and almost identical patterns of termination were observed: In both species, the majority of entorhinodentate fibers terminates in a laminar fashion in the outer portion of the molecular layers of the dentate gyrus and even individual axons have a very similar morphology. However, the portion of the molecular layer covered by entorhinal fibers appears to be greater in the mouse. In this species, entorhinal fibers cover approximately four-fifth of the molecular layer (Deller et al., 1999a; van Groen et al., 2002, 2003) compared to twothirds in rats (Blackstad, 1958; Amaral and Witter, 1989; Deller, 1998). Thus, entorhinal denervation in mice will result in a more extensive denervation of granule cell dendrites than in rats. A second relevant difference concerns the commissural and

associational (C/A) projection to the inner molecular layer. Although this projection is formed by mossy cells in both species, these cells contain calretinin in mice but not in rats (Liu et al., 1996; Blasco-Ibanez and Freund, 1997; Fujise et al., 1998). In addition, the C/A projection, in mice covers only one-fifth of the molecular layer in contrast to onethird in rats, is more diffusely organized and the border of the C/A fiber plexus towards the entorhinal zone of the dentate gyrus is indistinct (Figs. 2a, b), suggesting that the border between the two layers is not as strict as in the rat (Del Turco et al., 2003). A third and striking anatomical difference relates to the crossed entorhinodentate projection. This projection system seems to be nonexistent in mice (van Groen et al., 2002, 2003), whereas it is readily detectable in rats (Steward et al., 1974; Steward, 1976a; Deller et al., 1996a). Thus, the entorhinodentate pathway and at least two other fiber systems are different in rats and mice, which could influence the pattern of axonal reorganization in the denervated dentate gyrus. Reorganization of axons and fiber systems in the denervated dentate gyrus The reorganization and collateral sprouting of non-lesioned axons in the denervated dentate gyrus is probably the most prominent and bestknown aspect of the entorhinal denervation model in rats. It was also the first to be described (Lynch et al., 1972). When discussing this aspect of the entorhinal denervation model, we should distinguish between changes at the level of overall number of synapses and terminals, and changes observed at the level of identified fiber systems.

Changes in synapse and bouton densities following entorhinal denervation in rat and mouse In rats, degenerative and regenerative changes of synapses and terminals have been studied in detail. Using electron microscopy, synapse counts following ECL show 80–90% of all synapses are lost in the denervated zone of the dentate gyrus, followed by reactive changes that replace up to 60–80% of all

506

507

lost synapses by 4 weeks postlesion (Matthews et al., 1976a, b; Steward and Vinsant, 1983). In a recent stereological study, these data were largely confirmed, although synapse recovery observed with the more sensitive stereological tools was somewhat less than reported before (Marrone et al., 2004b). To visualize total changes in presynaptic terminals, immunohistochemical markers for presynaptic boutons such as synaptophysin have also been employed in rats (Jucker et al., 1996; Miwa and Ueki, 1996; Stone et al., 1998; Kadish and van Groen, 2003). Although changes in synaptophysin-positive boutons following entorhinal denervation seem to correlate with the quantitative electron microscopic data, a systematic comparison of the two techniques has not yet been performed, and it remains to be seen whether or not changes in synaptophysinpositive puncta accurately reflect changes at the ultrastructural level. For the denervated mouse dentate gyrus, virtually no quantitative electron-microscopic data are available. Thus, the extent of both synapse loss and reactive synaptogenesis are unknown. Entorhinal denervation has, however, been shown qualitatively in mice using silver impregnation of degenerating fibers and terminals (Fig. 1; Steward, 1992; Del Turco et al., 2003). In addition, stereological estimates of the total volume of the dentate gyrus following entorhinal denervation indicate that the dentate gyrus shrinks to 60% of its original size, suggesting that considerable degenerative changes occur following entorhinal denervation (Phinney et al., 2004). Similarly, entorhinal reinnervation has to some extent been demonstrated qualitatively. Synaptophysin-immunostaining was used to demonstrate changes in presynaptic terminals and

revealed an increase in staining in the denervated dentate gyrus compared to the contralateral side by 4 weeks survival time (van Groen et al., 2002; Kadish and van Groen, 2002, 2003).

Sprouting of identified fiber systems following entorhinal denervation The recovery of synapse densities following entorhinal denervation in rats and mice raises the question which fibers contribute to this reinnervation and whether this process is, in fact, restorative in nature. As axons from the ipsilateral entorhinal cortex are completely removed from the dentate gyrus and these fibers cannot regenerate in the adult brain, reinnervation apparently occurs from non-lesioned axons. The new connections thus differ anatomically as well as functionally from the ones removed by entorhinal denervation. In rats, the sprouting fiber systems were identified using histochemical as well as retrograde and anterograde tracing methods. Four fiber systems, which are already present in the nonesioned dentate gyrus, are believed to contribute to the reinnervation process (see Cotman and Nadler, 1978; Steward, 1994b; Deller and Frotscher, 1997, for review): (i) glutamatergic C/A fibers to the inner molecular layer (mossy cell axons), (ii) GABAergic C/A fibers to the outer molecular layer (iii) glutamatergic cholinergic septohippocampal fibers, and (iv) crossed entorhinodentate fibers. In the following paragraphs, we will consider the contribution of each of these fiber systems to the reinnervation process and will point out potentially relevant species differences.

Fig. 2. Sprouting of commissural/associational fibers in the mouse dentate gyrus following entorhinal denervation. (A, B) Dentate gyrus of a control mouse shown at low (A) and high magnification (B). The commissural projections to the inner molecular layer (IML) and the middle and outer molecular layer (OML; arrow in B) were labeled using Phaseolus vulgaris-leucoagglutinin (PHAL)-tracing. (C, D) Dentate gyrus of a mouse 2 months after entorhinal denervation shown at low (C) and high (D) magnification. Compared with the control, the commissural fiber plexus has increased in width and PHAL-labeled commissural fibers are regularly observed in the MML (arrows). (E, F) Dentate gyrus of a control mouse stained for calretinin shown at low (E) and high (F) magnification. Calretininimmunostaining labels mossy cells in the hilus (H) and interneurons in the molecular layer. Calretinin-immunoreactive mossy cell axons are abundant in the IML. (G, H) Dentate gyrus of a mouse 2 months following entorhinal denervation shown at low (G) and high (H) magnification. Note that compared to the control, the density of calretinin-immunoreactive fibers has increased throughout the molecular layer and that the calretinin-positive fiber plexus in the IML has expanded in width (H). CA3, hippocampal subfield; GCL, granule cell layer; H, hilus; OML, outer molecular layer; IML, inner molecular layer; MML, middle molecular layer. Scale bars: A, C, E, G: 100 mm; B, D, F, H: 25 mm.

508

Reorganization of commissural/associational mossy cell axons after entorhinal denervation As mentioned above, the C/A projection to the inner molecular layer of the dentate gyrus in the rat arises from glutamate-immunoreactive mossy cells (Soriano and Frotscher, 1994). These neurons are located in the hilus and project both contralaterally (commissural fibers) and ipsilaterally (associational fibers). Their axons terminate in a characteristic laminar pattern in the inner onethird of the molecular layer, i.e., complementary to the entorhinodentate fiber projection to the outer two-thirds of the molecular layer (Amaral and Witter, 1995). The contribution of this projection to the reorganization of the rat dentate gyrus is age-dependent (Gall and Lynch, 1978, 1980, 1981; Sta¨ubli et al., 1984; Gall et al., 1986): In young rats, injured within the first two-postnatal weeks, extensive sprouting of C/A axon collaterals into the deafferented adjacent outer molecular layer can be observed and some of these fibers extend all the way to the hippocampal fissure. The degree of this reactive growth response diminishes with maturation, and is almost completely lost in the adult rat brain (Gall and Lynch, 1981; Deller et al., 1996c). In addition, a 30–40 mm outward expansion of the entire C/A fiber plexus was also observed (Cotman and Nadler, 1978; Gall and Lynch, 1981). Although the nature of this phenomenon was a matter of debate for many years (see Deller and Frotscher, 1997, for review), available data support the interpretation that the expansion of the C/A plexus is not the result of translaminar sprouting but rather caused by reorganization and shrinkage processes occurring within the non-denervated inner molecular layer itself (Hoff et al., 1981; Caceres and Steward, 1983; Frotscher et al., 1997; Marrone et al., 2004a, b; Deller et al., 2006). We conclude, therefore, that (1) C/A mossy cell axons do not leave their home territory to invade the denervated zone in adult rats, and (2) axonal remodeling occurs not only in the denervated zone but also within the adjacent non-denervated inner molecular layer. In adult C57BL/6 mice, the C/A projection to the inner molecular layer also arises from mossy

cells in the hilus (Blasco-Ibanez and Freund, 1997). In mice, these cells contain the calciumbinding protein calretinin (Figs. 2e, f), which facilitates the identification of their axons in the inner molecular layer (Liu et al., 1996; Blasco-Ibanez and Freund, 1997; Fujise et al., 1998; Deller et al., 1999a; Del Turco et al., 2003). The contribution of this projection to the reorganization of the mouse dentate gyrus following entorhinal denervation was first studied by Shi and Stanfield (1996), who used horseradish peroxidase tracing and observed an ingrowth of C/A fibers into the denervated molecular layer (Shi and Stanfield, 1996). We confirmed and extended these data using anterograde Phaseolus vulgaris-leucoagglutinin (PHAL)-tracing in combination with immunostaining for calretinin at the light as well as the electron-microscopic level (Del Turco et al., 2003). In this study we showed that, in contrast to the adult rat, many mossy cell axons leave the main C/A fiber plexus in the inner molecular layer in response to entorhinal denervation, invade the denervated adjacent zone (Figs. 2c, d, g, h), and form new synapses outside their normal terminal zone. This pattern is striking in its distinction from the layer-specific sprouting in adult rats and is rather reminiscent of the sprouting seen in immature rats (Gall and Lynch, 1981). In addition, an expansion of the entire C/A termination zone was also observed in mice (Del Turco et al., 2003; Phinney et al., 2004). Phinney et al. (2004) studied this expansion using three-dimensional stereological tools and compared the expansion of the inner molecular layer in single sections with changes of the total volume of the inner molecular layer following denervation. Although this study confirmed a significant increase in the width of the C/A termination zone using single sections, it did not detect an increase in its total volume. In fact, shrinkage of the dentate gyrus in all three dimensions, especially the compression of the dentate gyrus along the septotemporal axis, seemed to underlie the expansion. The authors concluded that the expansion of the inner molecular layer does not indicate an increase in the size of this lamina, and therefore should not be used as an indicator of C/A sprouting in C57BL/6 mice.

509

Reorganization of GABAergic commissural/ associational axons after entorhinal denervation The GABAergic C/A projection in rats that terminates in the outer molecular layer of the dentate gyrus (Amaral and Witter, 1989; Deller et al., 1995a, b, 1996b; Zappone and Sloviter, 2001) also exists in mice (Deller et al., 1999a; Gebhardt et al., 2002; Del Turco et al., 2004). In rats, the reaction of this projection was studied at the single-fiber level using anterograde tracing, electron microscopy, and GABA-postembedding (Deller et al., 1995a). This study demonstrated that GABAergic commissural axons sprout new collaterals within the denervated zone and form synapses there. We concluded from this observation that (1) GABAergic fibers can sprout and reinnervate the dentate gyrus in response to the loss of glutamatergic afferents, and (2) that C/A fiber sprouting in the adult rat is layer specific. Thus, both the C/A fibers to the outer molecular layer as well as C/A fiber to the inner molecular layer sprout within their home territories in response to entorhinal denervation in adult rats (Deller et al., 1995a, 1996c; Deller and Frotscher, 1997; Frotscher et al., 1997). In mice, a systematic analysis of the GABAergic C/A projection to the molecular layer following entorhinal denervation has not yet been reported. Reorganization of acetylcholinesterase (AChE): positive fibers after entorhinal denervation A third projection system contributing to the reorganization of the dentate gyrus following entorhinal denervation is the cholinergic septohippocampal projection. In non-lesioned rats, this projection arises from cholinergic neurons located in the diagonal band of Broca (Nyakas et al., 1987; Jakab and Leranth, 1995). In contrast to the C/A and entorhinal fibers, this projection system terminates in all layers of the rat and mouse dentate gyrus (Amaral and Witter, 1995). To visualize the cholinergic septohippocampal fibers in the rat dentate gyrus, acetylcholinesterase (AChE) histochemistry has been widely used, since it has been shown to be sufficiently specific in this brain area (Eckenstein and Sofroniew, 1983; Levey et al., 1983, 1984; Naumann et al., 1997).

Following entorhinal denervation in rats, changes in the density of AChE-fibers in the denervated zone were observed and interpreted as a sign of cholinergic sprouting (Lynch et al., 1972; Nadler et al., 1977a, b; Zimmer et al., 1986). These observations were later corroborated by anterograde tracing studies of septohippocampal fibers (Nadler and Evenson, 1982; Stanfield and Cowan, 1982; Nyakas et al., 1988). In spite of these data, however, doubts remained. It was pointed out that changes in AChE-staining could also be the result of denervation-induced changes in gene expression (Chen et al., 1983; McKeon et al., 1989) or tissue shrinkage (Storm-Mathisen, 1974). In addition, biochemical studies did not reveal any evidence for cholinergic sprouting (Aubert et al., 1994), and a non-stereological quantitative analysis of choline acetyltransferase (ChAT)-positive fibers did not confirm the AChE-data (Henderson et al., 1998). Thus, it was questioned whether AChE-histochemistry is an appropriate technique to visualize cholinergic sprouting in the denervated rat dentate gyrus (Aubert et al., 1994) and whether cholinergic sprouting occurs at all (Storm-Mathisen, 1974). The first of these issues could be resolved by using the immunotoxin 192 IgG-saporin, which selectively destroys cholinergic neurons in the basal forebrain (Wiley et al., 1991; Book et al., 1992; Heckers et al., 1994). Following entorhinal denervation, rats were treated with 192 IgG-saporin and the AChE-positive fiber band could be completely abolished (Naumann et al., 1997). Thus, AChEhistochemistry is appropriate to visualize septohippocampal fibers selectively in the denervated dentate gyrus. However, the second of these issues yet awaits clarification. Since it will be essential to rule out shrinkage as a cause, cholinergic sprouting should be reinvestigated in rats using assumption-free stereological methodology. In non-lesioned mice, the septohippocampal projection appears to be generally similar to the one observed in rats (Linke et al., 1994). Similarly, in denervated mice, an increase in the density of AChE-positive fibers is observed in the denervated outer molecular layer (Steward, 1992, 1994a; Shi and Stanfield, 1996; Kadish and van Groen, 2002; Phinney et al., 2004). Analogous to the situation in rats, this increase in AChE-fiber density (Figs. 1e, f)

510

was interpreted as a sign of cholinergic sprouting (Steward, 1992, 1994a; Shi and Stanfield, 1996; Kadish and van Groen, 2002). To verify this finding, and to account for tissue shrinkage that occurs in the denervated dentate gyrus, a three-dimensional stereological analysis of total AChE-fiber length was performed (Phinney et al., 2004). Although this study confirmed a significant increase in the density of AChE-fibers in single sections, an increase in the total length of AChE fibers could not be detected. Stereological measurements of the volume of the non-lesioned and denervated dentate gyrus revealed that shrinkage of the dentate gyrus in all three dimensions, rather than growth of fibers was responsible for the increase in AChE-fiber density. The authors concluded that the increase in AChE fiber density in the denervated outer molecular layer should not be used as an indicator of lesion-induced cholinergic sprouting in mice. In summary, there is at present no direct evidence for sprouting of septohippocampal fibers following entorhinal denervation in mice.

Reorganization of crossed entorhinodentate fibers after entorhinal denervation A fourth projection system, contributing to the reorganization of the denervated dentate gyrus in rats, is the crossed entorhinodentate (temporodentate) pathway. Its cells of origin, trajectory, and its termination pattern have been well characterized using retrograde and anterograde tracing techniques (Steward et al., 1974; Goldowitz et al., 1975; Steward, 1976b, 1980; Steward and Scoville, 1976; Steward and Vinsant, 1978; Amaral and Witter, 1995; Deller et al., 1996a). These studies revealed that the crossed entorhinodentate projection is weaker but anatomically homologous to the ipsilateral pathway. From a functional point of view this homology makes the crossed entorhinodentate pathway particularly interesting, because its sprouting could contribute to a restoration of function following entorhinal denervation. Accordingly, morphological, electrophysiological, and behavioral changes caused by the reorganization of crossed entorhinodentate fibers were studied in detail (Steward et al., 1976, 1988; Steward,

1976a; Cotman et al., 1977). Morphological analysis demonstrated robust sprouting of the crossed entorhinodentate projection: It increases its fiber and terminal density by sixfold and of its synapse density by approximately 100-fold. The reorganization of single fibers was also revealed using anterograde PHAL-tracing (Deller et al., 1996a). Whereas crossed entorhinodentate axons showed only few branches and formed almost exclusively en passant boutons in non-lesioned animals, crossed entorhinodentate fibers displayed high densities of small axonal extensions and tanglelike formations resembling axonal and synaptic clusters in denervated rats. In keeping with these morphological data, electrophysiological studies showed that the functional characteristics of the crossed entorhinodentate projection become similar to the ipsilateral projection following denervation (Steward et al., 1973; White et al., 1976; Harris et al., 1978; Reeves and Steward, 1986, 1988). Finally, behavioral experiments confirmed that sprouting of crossed entorhinodentate fibers can indeed ameliorate some, but not all, of the behavioral deficits observed after unilateral entorhinal denervation (Loesche and Steward, 1977; Scheff and Cotman, 1977; Ramirez and Stein, 1984; Schenk and Morris, 1985). In summary, these data indicate that (1) crossed entorhinodentate sprouting contributes significantly to the reinnervation of the rat dentate gyrus after entorhinal denervation, and (2) this sprouting of a homologous fiber system is probably functionally restorative. In mice, the crossed entorhinodentate pathway seems to be essentially non-existent (van Groen et al., 2002, 2003), and it is unlikely that the very few fibers observed in the contralateral dentate gyrus of mice could contribute significantly to the reinnervation of the dentate gyrus following denervation. However, it should also be pointed out that this has not yet been systematically analyzed and that it is unclear to what extent sprouting can be elicited from the few crossed entorhinodentate fibers observed in mice. In summary, we conclude that the entorhinodentate projection is almost absent in mice and that there is presently no evidence for crossed entorhinodentate fiber sprouting in this species.

511

Glial changes following entorhinal denervation In recent years, the role of non-neuronal cells in the reorganization of the dentate gyrus after entorhinal denervation has received considerable attention. The reaction of resident glia as well as the recruitment of blood cells into the zone of denervation have been studied, and a role of these cells in the reorganization process has been proposed; these non-neuronal cells are not only important for the removal of cellular debris in the zone of denervation, but they may also regulate axonal growth and transneuronal degeneration processes. Thus, non-neuronal cells may play a central role in the reorganization of the dentate gyrus following denervation (see below).

Microglia In rats, microglial cells react very rapidly to entorhinal denervation by proliferation, and by transformation from a ramified to an ‘‘activated’’ phenotype (Gall et al., 1979b; Fagan and Gage, 1990, 1994; Gehrmann et al., 1991; Jensen et al., 1994; Hailer et al., 1997). Using histological and immunohistochemical techniques, microglial reactions were visible as early as 24 h postlesion, were strongest at approximately 3 days postlesion, and returned to normal levels during the first 2 weeks after denervation. Electron microscopy and Mini Ruby-tracing demonstrated that reactive microglial cells phagocytose axonal debris and myelin sheaths (Gehrmann et al., 1991; Jensen et al., 1994; Bechmann and Nitsch, 1997). Interestingly, microglial activation was also observed, though to a much lesser extent, in adjacent non-denervated areas, suggesting that microglial cells are actively recruited to the denervated zone (Gall et al., 1979b; Jensen et al., 1994). Molecules that could regulate this highly specific recruitment are integrin adhesion molecules, such as leukocyte function antigen-1 (LFA-1), very late antigen-4 (VLA-4), and the ligand for LFA-1, intercellular adhesion molecule-1 (ICAM-1), which are all expressed by microglial cell in the zone of denervation (Hailer et al., 1997). In addition to the role of microglia as debris-removing phagocytes,

regulatory functions have also been proposed, and it was suggested that microglia could initiate a sequence of events, which could eventually induce or regulate axon sprouting (Gage et al., 1988; Fagan and Gage, 1990; Eyupoglu et al., 2003). In short, it was suggested that entorhinal terminal degeneration activates microglia. These activated cells phagocytose degenerating terminals and release interleukin-1, thereby activating the astroglia. Activated astrocytes, in turn, secrete neurotrophic factors, which induce sprouting of axons carrying specific receptors (Gage et al., 1988; Eyupoglu et al., 2003). In mice, the morphologic transformation of microglia is fairly similar to the changes observed in rats in that an accumulation of reactive microglia occurs within the denervated parts of the molecular layer (Fig. 3a) as early as 24 h after entorhinal denervation (Schauwecker and Steward, 1997; Jensen et al., 1999; Wirenfeldt et al., 2003; Rappert et al., 2004; Pedersen et al., 2006). At this time, the microglia have transformed into what is regarded as an ‘‘amoeboid’’ phenotype, and the relative depletion of the adjacent nondenervated inner molecular layer suggests that they migrate into the zone of axonal degeneration. In fact, we have demonstrated that the microglial accumulation is severely impaired in mice deficient for CXCR3, a chemokine receptor crucial for the attraction of microglia (Rappert et al., 2002, 2004). The other known factor contributing to the microglial accumulation is proliferation, which has been demonstrated in mice using modern stereological techniques (Wirenfeldt et al., 2003; Ladeby et al., 2005a, b). The molecular regulation of the microglial response following entorhinal denervation has been studied in greater detail in mice compared to rats. Besides CXCR3 (Rappert et al., 2004), the peripheral benzodiazepine binding site ligand PK11195 (Pedersen et al., 2006), and interferon-gamma (Jensen et al., 2000a), interleukin 1beta and the transforming growth factor beta1 (Eyupoglu et al., 2003) have been implicated. In addition to the role of microglia as phagocytes, however, regulatory functions were also proposed and, recently, experimental evidence for a link between microglia activation and dendritic atrophy of GABAergic parvalbumin-positive basket cells

512

Fig. 3. Microglial cells play a role in transneuronal degeneration of parvalbumin-positive dendrites. (A) Overview of the entorhinalhippocampal area of a control mouse after denervation of the middle molecular layer (MML) of the dentate gyrus. Mac-1 positive microglia accumulate in the zone of anterograde axonal degeneration. The insert shows the morphology of the activated microglial cells. This microglial response is less pronounced in CXCR3 / (KO) mice compared to wild-type (WT) controls. (B–E) If less microglial cells are recruited, denervated dendrites are preserved. The loss of parvalbumin-positive dendrites in the dentate gyrus is evident in WT mice (B, D) and clearly diminished in CXCR3 / animals (C, E). The boxes in (B) and (C) indicate the areas shown at higher magnification in (D) and (E), respectively. Scale bars: A: 300 mm; inset: 50 mm; B, C: 100 mm; D, E: 25 mm.

513

in the denervated dentate gyrus was shown (Rappert et al., 2004; see also below). Blood-derived cells Besides migration and proliferation of resident microglia, a third mechanism contributing to the accumulation of microglia/macrophages within the zone of denervation is the recruitment of blood-derived monocytic cells. At least in the mouse, these cells are able to invade the denervated hippocampus, where they transform into microglia-like elements within 72 h after entorhinal denervation (Bechmann et al., 2005; Ladeby et al., 2005b). As all antibodies used for microglia-detection also bind to monocytes/macrophages, the invasion of hematopoietic cells as a source of ‘‘new’’ microglia have been difficult to study. Fagan and Gage (1994) analyzed rats using i.v. injection of DiI-labeled low density lipoprotein (LDL) particles, but could not find any evidence for the infiltration of monocytes. In their preparations, T cells were also not visible (Fagan and Gage, 1994). Later, using the R73 antibody, we demonstrated the presence of CD4 a/b in the middle molecular layer after entorhinal denervation in the same rat strain (Bechmann et al., 2001a), but the question whether monocytes are also recruited has only been addressed in mice, where CCL-2 has been identified as a major chemoattractant molecule (Babcock et al., 2003). As described below, recent evidence links chemokine-mediated recruitment of microglia to transsynaptic dendritic changes in the dentate gyrus. Astroglia In rats, astroglial activation after entorhinal denervation was observed at slightly later time points than the activation of microglia and was characterized by proliferation, upregulation of glial fibrillary acidic protein (GFAP), and migration into the denervated zone of the dentate gyrus (Rose et al., 1976; Gall et al., 1979b; Gage et al., 1988; Fagan and Gage, 1990; Steward et al., 1993). Within the denervated zone, astrocytes were shown to

participate in the removal of degenerating terminals and spines (Bechmann and Nitsch, 1997; Deller et al., 1997). Increases in GFAP messenger RNA (mRNA) slightly preceded changes in GFAP and were observed as early as 12 h postlesion (Steward et al., 1990, 1993). Interestingly, GFAP mRNA changes depended on postlesional seizure activity, demonstrating that changes in GFAP expression in the early postlesional time period may be caused by changes in activity rather than denervation (Kelley and Steward, 1996b). In addition, it was shown that reactive astrocytes change the molecular composition of the denervated molecular layer (Gall et al., 1979b; Gage et al., 1988; Steward, 1991; Deller et al., 2000, 2001): Astrocyte-derived neurotrophic factors (Kawaja and Gage, 1991; Yoshida and Gage, 1991; Gomez-Pinilla et al., 1992; Guthrie et al., 1995, 1997; Lee et al., 1997), cell-adhesion molecules (Styren et al., 1994, 1995; Jucker et al., 1996), and a variety of extracellular matrix molecules (Deller et al., 2000) were studied and it was suggested that these astrocyte-derived molecules could regulate and pattern the sprouting response. In mice, less information is available. Astroglial changes were studied using GFAP-immunohistochemistry, revealing astroglial migration and an increased expression of GFAP-immunoreactivity throughout the denervated zone (Steward and Trimmer, 1997; Del Turco et al., 2003). Changes in GFAP mRNA were measured using northern blot (Steward and Trimmer, 1997; Steward et al., 1997) and at the single-cell level using laser microdissection in combination with quantitative realtime-polymerase chain reaction (qRT-PCR) (Burbach et al., 2004a, b). These postlesional changes in GFAP mRNA depend, as in the rat, on activity and protein synthesis (Steward, 1994a; Steward et al., 1997). In contrast to rats, however, studies analyzing changes in the expression of astrocyte-derived molecules following entorhinal denervation are rare (Poulsen et al., 2000; Mayer et al., 2005; Schafer et al., 2005). In summary, morphological evidence supports a phagocytic role of astrocytes in the denervated dentate gyrus of the adult rat and correlative data implicate astrocytes in the patterning of the sprouting response. In mice, the cellular response

514

of astrocytes appears to be similar to the one observed in rats. However, information about molecular changes induced by reactive astrocytes is scarce and the role of astrocytes in this species remains to be elucidated.

2006), suggesting that sprouting axons could be myelinated by newly formed oligodendroglial cells. In summary, oligodendrocytes react to entorhinal denervation in rats as well as mice. Their surface molecules may be involved in the regulation of sprouting, and newly formed oligodendrocytes may contribute to the myelination of sprouting axons.

Oligodendroglia and NG2-positive cells It is now well established that oligodendrocytes and some of their surface molecules are potent inhibitors of axonal growth (Kapfhammer and Schwab, 1992; Schwab et al., 1993). Since these inhibitory molecules could influence sprouting, they were studied following entorhinal denervation. In addition, the entorhinal denervation model was used to study the myelination of newly formed (sprouting) axons in the normal adult CNS. Changes in putative oligodendroglial cell precursors, NG2-cells (Dawson et al., 2000, 2003) were also analyzed. In rats, the first evidence for a reaction of oligodendroglial cells to entorhinal denervation was provided by Gall et al. (1979b), who observed signs of oligodendroglial cell proliferation. Later, changes in myelination and Nogo-expression were analyzed (Meier et al., 2003, 2004). These studies revealed correlations between the axonal sprouting response and the time course of demyelination, suggesting that oligodendroglial-derived inhibitory molecules could regulate sprouting. Consistent with the results of Gall et al. (1979a, b), we recently showed that oligodendroglial precursor cells, NG2-positive cells, also react to entorhinal denervation in rats (Dehn et al., 2006). In mice, an analysis of the dynamics of oligodendrocytes following entorhinal denervation was performed by Drojdahl and colleagues (2004). In addition, changes in oligodendroglial-derived molecules, such as changes in the expression of myelin basic protein (Jensen et al., 2000b; Drojdahl et al., 2004) and myelin-associated glycoprotein (Mingorance et al., 2005), were reported. Similar to the observations in rats, proliferation of NG2-positive oligodendroglial precursors and the formation of new oligodendrocytes was demonstrated following entorhinal denervation in mice (Nielsen et al.,

Dendritic changes following entorhinal denervation Denervated neurons in the CNS remodel their dendritic field. This reorganization process involves spines and dendrites, and may, if the denervation is extensive, lead to the death of the denervated target cell (Steward, 1994b; Kovac et al., 2004). In the rat, dentate gyrus granule cells (Nafstad, 1967; Steward and Scoville, 1976; Amaral and Witter, 1995) and parvalbuminpositive GABAergic interneurons (Kosaka et al., 1987; Zipp et al., 1989; Nitsch et al., 1990) are the targets of entorhinal afferents. These fibers impinge on their distal dendrites and form between 80 and 90% of all synapses on their distal dendritic field (Matthews et al., 1976a; Steward and Vinsant, 1983). In non-lesioned C57BL/6 mice, the anatomical situation appears to be similar (Deller, 1998; Gebhardt et al., 2002; Del Turco et al., 2004). Thus, ECL/perforant pathway transection will denervate both cell types in both species. Transneuronal changes in granule cells and parvalbumin-positive neurons In rats, transneuronal changes in granule cell morphology after entorhinal denervation have been studied using Golgi-impregnated (Parnavelas et al., 1974; Caceres and Steward, 1983) and intracellularly labeled (Diekmann et al., 1996) granule cells. These studies agree that the dendritic arbor of granule cells is remodeled following denervation, and that granule cells lose spines and dendrites during the first 2 weeks after entorhinal denervation. By 30 days postlesion, there was a recovery of the dendritic arbor detected by the Golgi method (Caceres and Steward, 1983), whereas permanent transneuronal changes of granule cell dendrites were reported

515

using the intracellularly labeled granule cells (Diekmann et al., 1996). Dendritic remodeling of parvalbumin-positive GABAergic neurons was also studied following entorhinal denervation (Nitsch and Frotscher, 1991, 1992, 1993; Nitsch et al., 1992). Similar to granule cells, the distal dendrites of parvalbuminpositive neurons demonstrated degenerative changes (Nitsch and Frotscher, 1991, 1992, 1993). These degenerative changes persisted, with only a slight recovery by 2 months postlesion (Nitsch and Frotscher, 1991). In mice, morphological changes of granule cells have not yet been studied at the level of single cells, and at present only indirect — and controversial — data are available. Following ibotenic acid injections into the entorhinal cortex, Kadish and van Groen (2003) did not observe atrophy of the denervated molecular layer of the mouse dentate gyrus, and suggested that granule cell dendritic atrophy is limited in this species. In contrast, others (Phinney et al., 2004) who used mechanical transection and measured the width and the volume of the molecular layer using stereological methods demonstrated a>40% reduction in volume 45 days postlesion, suggesting substantial granule cell atrophy. A detailed temporal analysis on the postlesional morphological changes of single granule cells is needed to clarify this issue, and its potential differences from the rat. In contrast to granule cells, morphological changes of parvalbumin-positive neurons have been studied in more detail in mouse (Rappert et al., 2004). A temporal analysis demonstrated changes in parvalbumin-positive dendrites that were quite similar to rats. Parvalbumin-positive dendrites were rapidly lost in the denervated molecular layer of C57BL/6 mice and no recovery was observed up to 31 days postlesion (Fig. 3). However, a detailed morphological analysis of identified parvalbumin-positive cells was not performed.

Transneuronal changes: insights into cellular mechanisms What are the potential mechanisms involved in the transneuronal dendritic degeneration of granule

cells and parvalbumin-positive neurons following entorhinal denervation? Although several hypotheses have been proposed, including excitotoxicity, elevation of intracellular calcium levels, and lack of growth factors (Steward, 1991; Nitsch and Frotscher, 1992; Sattler et al., 1998; Arundine and Tymianski, 2003; Kovac et al., 2004), the role of these mechanisms and molecules has not yet been clarified. As far as granule cells are concerned, no studies to date have explained the extensive transneuronal degeneration of the denervated granule cell arbor. In the case of the degeneration of parvalbumin-positive neurons, Nitsch and Frotscher (1992) showed that the application of the glutamate receptor antagonist MK801 prior to entorhinal denervation could prevent the atrophy of parvalbumin-immunoreactive dendrites (Nitsch and Frotscher, 1992). They suggested that an NMDA-receptor mediated influx of calcium could cause the degeneration of parvalbumin-positive dendrites. In mice, degeneration of parvalbuminpositive dendrites was recently investigated (Rappert et al., 2004) and microglia and chemokines were implicated in the regulation of transneuronal degeneration. Since this study proposed a novel mechanism for denervation-induced transneuronal degeneration a more detailed review of this study and the potential role of chemokines appears to be warranted.

Transneuronal degeneration of parvalbumin-positive dendrites: a role for the chemokine receptor CXCR3 Chemokines are a diverse family of pro-inflammatory cytokines. These small homologous peptides (8–14 kDa) (Rollins, 1997; Baggiolini, 1998) play an important role in recruiting inflammatory cells into tissues in response to infection and inflammation (Karpus and Ransohoff, 1998; Mackay, 2001). Additionally, these molecules are involved in the pathogenesis of many important neuroinflammatory diseases ranging from multiple sclerosis and stroke to HIV encephalopathy (Luster, 1998; Asensio and Campbell, 1999). For many chemokine receptors in the adult CNS, the corresponding ligand can only be detected under pathological conditions, for example,

516

in the vicinity of an inflammatory lesion (Asensio and Campbell, 1999; Bacon and Harrison, 2000) suggesting that chemokine-receptors and their ligands are primarily involved in the regulation of pathophysiological processes. In line with this interpretation, several studies have shown that chemokines recruit leukocytes into inflamed brain tissue (Fife et al., 2000; Izikson et al., 2000; Siebert et al., 2000; Huang et al., 2001; Babcock and Owens, 2003). Messenger RNA for CCL5, CCL2, CXCL10, CCL3, CCL4 and CXCL2 were specifically induced in the lesion-reactive hippocampus and at the site of axonal transection by 24 h after axotomy, while other chemokines like XCL1, CCL11 and CCL1 were not detected (Babcock et al., 2003). Interestingly, CCL2 was upregulated in the denervated hippocampus already at 3 h after axotomy, while an increase in CCL5 was first detected after 24–48 h. The different time course of induction of these two chemokines suggests that chemokines could have different functions following ECL. In fact, as pointed out above, analyses of chemokine receptor-deficient mice indicated that the CCR2 ligand CCL2 (MCP-1) is critical for leukocyte infiltration, whereas CCR5 ligands such as CCL5 (RANTES) are not (Babcock et al., 2003). What could be the role of the chemokines and their receptors in the reorganization of the dentate gyrus following ECL? Since chemokine receptors are expressed in neurons and glia, and chemokines are mostly produced by glial cells, it has been suggested that they are involved in glia-to-glia, gliato-neuron, and neuron-to-glia interactions (Hesselgesser and Horuk, 1999; Ambrosini and Aloisi, 2004; Adler et al., 2005). In the context of ECL the chemokine receptor CXCR3 and its ligands CXCL9, CXL10, CXCL11 and CCL21 are especially interesting. For example, CCL21 is specifically expressed by damaged neurons (Biber et al., 2001; de Jong et al., 2005) and not by any glial cell type (Biber et al., 2001; Alt et al., 2002; Columba-Cabezas et al., 2003). Since CCL21 induces intracellular calcium signals and chemotaxis of cultured microglia (Biber et al., 2001; Rappert et al., 2002), a potential role of CCL21 in neuronmicroglia communication has been proposed — but has to be proven in vivo. As a second ligand,

CXCL10 mRNA has been found in the lesioned hippocampus of the mouse (Babcock et al., 2003) and the protein could be localized on neurons by our group (Rappert et al., 2004). In mice deficient for CXCR3, the accumulation of microglia in zones of denervation was dramatically reduced, demonstrating a crucial role for CXCR3 in microglial recruitment and suggesting CXCL10 as the signaling partner (Rappert et al., 2004). The importance of CXCR3 is underscored by the preservation of parvalbumin-positive dendrites after entorhinal denervation in CXCR3-deficient mice (Rappert et al., 2004). The absence of the dendritic loss in the CXCR3 knockout animals is not simply delayed, since a difference in the dendritic length compared to wild-type animals was still evident 31 days after entorhinal denervation. Although a reduction of dendrites developed, there were significantly more dendrites in the KO in comparison to the wild-type animal (Figs 3b–e). This is consistent with previous observations that the elimination of microglia in rat entorhinohippocampal slice cultures diminishes lesion-induced dendritic loss (Eyupoglu et al., 2003). Although there is no mechanistic explanation for this phenomenon, these data argue against the concept that ‘‘retraction’’ is dependent only on the loss of a specific input, and demonstrates a causal link between microglial response and the structural outcome of postlesional changes. Whether the microglia actively eliminate denervated dendritic segments or provide signals inducing their retraction remains to be seen.

Summary In the present review, we have summarized and compared the available data on the structural reorganization of the dentate gyrus following entorhinal denervation in rats and mice. We conclude on the basis of this comparison, that in mice neither the normal anatomy of the dentate gyrus and nor many aspects concerning its structural reorganization following denervation have yet been described in depth. The available data, however, already point to important species differences in both aspects, anatomy and reorganization. In the

517

following paragraphs, we will discuss these species differences in the context of sprouting and inflammation and will make suggestions for the use of the entorhinal denervation model in mutant mice.

How can the species difference in the sprouting response be explained? Our comparison of sprouting following entorhinal denervation in rats and mice revealed major species differences. Some of these differences may readily be explained on the basis of differences in normal anatomy. For example, if crossed entorhinodentate fibers do not exist in mice, sprouting of this fiber projection is not to be expected. Other species differences, for example the response of the C/A fibers to the inner molecular layer are more difficult to explain. This projection system exists in rats as well as mice and only subtle anatomical differences are observed between the two species (Del Turco et al., 2003). Nevertheless, in adult rats these C/A fibers stay in their home territory whereas they sprout across the laminar boundaries in mice. In the absence of major anatomical differences, an explanation for this phenomenon should be sought on the molecular level. In recent years, we have studied molecules which could regulate the layer-specific sprouting of C/A fibers to the inner molecular layer in rats (Frotscher et al., 1997; Deller et al., 2000, 2001). Whereas glia-derived neurotrophic factors and cytokines are probably involved in the induction of new axonal collaterals (Gomez-Pinilla et al., 1992; Guthrie et al., 1995, 1997; Fagan et al., 1997; Lee et al., 1997; Woods et al., 1998, 1999), other molecules may be better suited for the guidance of newly formed fibers. A group of molecules which could regulate a layer-specific growth response by forming a growth barrier between the non-denervated inner and the denervated middle- and outer molecular layer are extracellular matrix molecules synthesized by reactive glia (Rose et al., 1976; Gage et al., 1988). These molecules are thought to form boundaries for growing axons during development (Faissner and Steindler, 1995; Ho¨ke and Silver, 1996; Pearlman and Sheppard, 1996; Margolis and Margolis, 1997; Bovolenta and

Fernaud-Espinosa, 2000), and it could be demonstrated that several growth inhibitory extracellular matrix molecules such as tenascin-C (Deller et al., 1997), DSD-1-proteoglycan (Deller et al., 1997), neurocan (Haas et al., 1999), brevican (Thon et al., 2000), and NG2 (Dehn et al., 2006) are upregulated within the denervated zone of the molecular layer after entorhinal denervation. Because these molecules form a sharp border towards the adjacent non-denervated zones, we suggested that these molecules delineate boundaries of axonal growth following entorhinal lesion (Deller et al., 2000, 2001). This hypothesis, although certainly attractive, still awaits experimental verification. In mice, the situation is different and C/A fibers to the inner molecular layer sprout across laminar boundaries. One explanation could be that in this species, the expression pattern, the time course of expression, or the order of assembly of growth inhibitory extracellular matrix molecules could be different. So far, only few of the extracellular matrix molecules mentioned above have been investigated: Brevican (Mayer et al., 2005) and NG2 (Nielsen et al., 2006) were analyzed and denervation-induced changes were reported which were similar to those observed for these molecules in rats. However, subtle differences with regard to the time course of expression and the cellular distribution of the two molecules were also observed. Whether or not these differences — or differences in the expression of other extracellular matrix molecules — could result in an absent or less effective growth barrier between the non-denervated inner molecular layer and the denervated zone in mice remains to be seen. Another explanation could be that mossy cells in mice and rats differ in their ability to grow, i.e., in their intrinsic growth competence (Benowitz and Routtenberg, 1997; Caroni, 1997, 1998). It is well established that the growth competence of neurons depends on the expression of growth-associated proteins, such as growth-associated protein 43 or cortical cytoskeleton-associated protein 23 (Aigner et al., 1995; Benowitz and Routtenberg, 1997; Caroni et al., 1997; Frey et al., 2000; Laux et al., 2000). These molecules are strongly expressed during brain development and are downregulated postnatally (Caroni, 1997). Intriguingly, the ability

518

of rat mossy cells to sprout across laminar boundaries (Lynch et al., 1973; Gall and Lynch, 1980, 1981; Gall et al., 1986) and the downregulation of growth-associated proteins (Caroni, 1997) occurs during the same developmental time window, suggesting that the ability of mossy cell axons to sprout into the denervated dentate gyrus indeed depends on their intrinsic growth competence. In contrast, in mice, in which translaminar mossy cell sprouting is found in the adult, mossy cell growth competence is relatively strong. Recent observations in transgenic mice overexpressing growth-associated proteins indicate that the intrinsic growth-competence of mossy cells affects the strength and extent of the translaminar commissural sprouting response in mice following entorhinal denervation (Del Turco et al., 2001). Finally, competitive interactions of sprouting C/A mossy cell axons with other sprouting fiber systems could also play a role (Cotman and Nadler, 1978; Nadler et al., 1980; Steward, 1994b). In rats, crossed entorhinodentate fibers sprout within the denervated molecular layer (Steward et al., 1974; Deller et al., 1996a). Since recent data indicate that C/A fibers are repelled by entorhinal fibers (Borrell et al., 1999; Deller et al., 1999a), sprouting of this pathway would block translaminar ingrowth of C/A fibers into the denervated zone in rats. In contrast, C/A fibers could grow into the denervated mouse fascia dentata because virtually all entorhinal fibers are removed from the mouse dentate gyrus after unilateral entorhinal denervation. If this hypothesis were true, then the absence of the crossed entorhinodentate pathway in mice would suffice to explain the major species differences between rats and mice with regard to the axonal reorganization of the dentate gyrus following denervation.

Inflammatory changes following denervation and the immune response to entorhinal denervation The morphologic transformation of astrocytes and microglia and their recruitment into the zones of axonal degeneration are similar in mice and rats. As described above, much has been speculated with regard to the contribution of glia to postlesional

reorganization, but as yet an impact on the sprouting response has not been proven. However, in both rats and mice, interfering with microglial recruitment impacts on the loss of denervated dendrites (Eyupoglu et al., 2003; Rappert et al., 2004). Phagocytosis of distal segments by activated microglia is an attractive explanation for this observation, which is currently tested in our laboratory. The immune system of rodents is capable of mounting strong immune responses to myelinassociated epitopes, but entorhinal denervation does not induce evident autoimmune demyelination. This could be attributed to a state of immune ignorance, as Fagan and Gage (1994) did not find infiltration of haematogenous cells into zones of axonal degeneration after entorhinal denervation in the rat. In the same strain, we detected infiltrating CD4 T cells interacting with MHC-II positive microglia. These myelin-phagocytosing cells exhibited the phenotype of immature antigen-presenting cells which is likely to cause T cell anergy (Bechmann et al., 2001a). Moreover, astrocytes expressed the death-ligand CD95L (FasL) which causes apoptosis of activated T cells (Bechmann et al., 1999, 2000, 2002). Thus, immune tolerance seems to be actively maintained and, in fact, we have shown that autoimmunity to myelin is diminished following entorhinal denervation in mice. Strikingly, axonal antigens are found in cervical lymph nodes after entorhinal denervation and their appearance is accompanied by T cell apoptosis (Mutlu et al., 2007). Monocytic cells are known to pass the vascular wall to reside in perivascular spaces in rats and mice under normal conditions (Bechmann et al., 2001b, c), but they cross the glia limitans and transform into microglia after entorhinal denervation in mice (Bechmann et al., 2005). CCL2 has been identified as the key chemokine driving this recruitment (Babcock et al., 2003), while it is unclear which signals adopt the invading macrophages into ramified microglia. One of the key questions is whether such ‘‘microglia’’ transform into dendritic cells capable of leaving the neuropil to present antigens in lymphoid organs. In fact, after entorhinal denervation in mice, T cells specifically target one cervical lymph node via the cribroid plate and the nasal mucosa (Goldmann et al., 2006).

519

It is also noteworthy that permeability of the blood-brain barrier for IgG has been demonstrated after entorhinal denervation in the rat a decade ago (Jensen et al., 1997), but the role of antibodies in this context and their impact on postlesional reorganization is completely unclear. It is now clear that entorhinal denervation in mice and rats involves a systemic immune response, which impacts on the structural outcome at least at the levels of phagocytosis of growth-inhibiting debris. T cells have also been shown to secrete neurotrophins (Kerschensteiner et al., 1999), but entorhinal denervation-studies in mice bearing myelin-specific T cell receptors have not yet been performed. It is evident from countless immunological studies that there are significant species and strain differences in regard to the strength and consequence of immune responses to brain antigens (Gold et al., 2006). In regard to the neurological outcome, they can be protective or detrimental. The same type of lesion can evoke overall protective or detrimental immune responses solely depending on strain subjected to the damage (Kipnis et al., 2001). Thus, beyond species differences, strain differences must also be expected when evaluating the role of immune cells in postlesional plasticity. The intrinsic immune suppression of the brain may have evolved to cope with situations bearing high evolutionary pressure such as infection, where it is better for the individual to tolerate certain infectious agents than to eliminate all infected neurons. Under conditions such as axonal injuries, which certainly provided low pressure, it may be beneficial for the neurological outcome to therapeutically foster immune responses (Bechmann, 2005). This can be tested in mouse models of entorhinal denervation.

Conclusions for the use of the entorhinal denervation model in rats and mice In summary, ECL/perforant pathway transections have been used to denervate the dentate gyrus of rats and mice in order to study the reorganization of the brain following injury. Under the assumption that the reorganization of the dentate gyrus is similar in the two related rodent species, data obtained in one,

usually the rat, were used to interpret data obtained in the other, usually the mouse. Since this assumption was never proven, we have compiled and compared the data available on structural changes following entorhinal denervation in both species in this review. This comparative analysis revealed major differences between the entorhinal model system in rats and mice. We conclude, that a non-critical transfer of data from one species to the other is not warranted. Given the difficulty in generalizing data from the rat to mouse, all observations made in rats need first to be verified in the context of the mouse model before mutants should be analyzed. In addition, strain differences should also be considered and it will probably be necessary to reinvestigate certain aspects specifically in the background strain of a given mutant. Only then, will it be possible to gain meaningful and fundamental biological insights from the entorhinal denervation model in mice. Although certainly demanding, such an approach may yield additional scientific benefits, since the investigation of species and strain differences in hippocampal connectivity and plasticity may open up new strategies for a comparative, yet functionally oriented analysis. This way, a methodologically more careful approach would be richly awarded and of considerable value to the scientific community.

Acknowledgments This review is dedicated to Michael Frotscher, M.D., who is one of the pioneers of hippocampal research and the entorhinal denervation model, and who contributed to many of the studies reviewed here. This study was supported by the Deutsche Forschungsgemeinschaft (DE 551/8-1, SFB 507 B16), German Israeli Foundation (T.D.), and Gisela Stadelmann-Stiftung (D.D.T.)

References Adler, M.W., Geller, E.B., Chen, X. and Rogers, T.J. (2005) Viewing chemokines as a third major system of communication in the brain. AAPS J., 7: E865–E870.

520 Aigner, L., Arber, S., Kapfhammer, J.P., Laux, T., Schneider, C., Botteri, F., Brenner, H.R. and Caroni, P. (1995) Overexpression of the neural growth-associated protein GAP-43 induces nerve sprouting in the adult nervous system of transgenic mice. Cell, 83: 269–278. Alt, C., Laschinger, M. and Engelhardt, B. (2002) Functional expression of the lymphoid chemokines CCL19 (ELC) and CCL 21 (SLC) at the blood-brain barrier suggests their involvement in G-protein-dependent lymphocyte recruitment into the central nervous system during experimental autoimmune encephalomyelitis. Eur. J. Immunol., 32: 2133–2144. Amaral, D.G. and Witter, M.P. (1989) The three dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31: 571–591. Amaral, D.G. and Witter, M.P. (1995) The hippocampal formation. In: Paxinos G. (Ed.), The Rat Nervous System (2nd ed.). Academic Press, San Diego, CA, pp. 443–494. Ambrosini, E. and Aloisi, F. (2004) Chemokines and glial cells: a complex network in the central nervous system. Neurochem. Res., 29: 1017–1038. Arundine, M. and Tymianski, M. (2003) Molecular mechanisms of calcium-dependent neurodegeneration in excitotoxicity. Cell Calcium, 34: 325–337. Asensio, V.C. and Campbell, I.L. (1999) Chemokines in the CNS: plurifunctional mediators in diverse states. Trends Neurosci., 22: 504–512. Aubert, I., Poirier, J., Gauthier, S. and Quirion, R. (1994) Multiple cholinergic markers are unexpectedly not altered in the rat dentate gyrus following entorhinal cortex lesions. Neuroscience, 14: 2476–2484. Babcock, A. and Owens, T. (2003) Chemokines in experimental autoimmune encephalomyelitis and multiple sclerosis. Adv. Exp. Med. Biol., 520: 120–132. Babcock, A.A., Kuziel, W.A., Rivest, S. and Owens, T. (2003) Chemokine expression by glial cells directs leukocytes to sites of axonal injury in the CNS. J. Neurosci., 23: 7922–7930. Bacon, K.B. and Harrison, J.K. (2000) Chemokines and their receptors in neurobiology: perspectives in physiology and homeostasis. J. Neuroimmunol., 104: 92–97. Baggiolini, M. (1998) Chemokines and leukocyte traffic. Nature, 392: 565–568. Bechmann, I. (2005) Failed central nervous system regeneration: a downside of immune privilege? Neuromolecular Med., 7: 217–228. Bechmann, I., Goldmann, J., Kovac, A.D., Kwidzinski, E., Simburger, E., Naftolin, F., Dirnagl, U., Nitsch, R. and Priller, J. (2005) Circulating monocytic cells infiltrate layers of anterograde axonal degeneration where they transform into microglia. FASEB J., 19: 647–649. Bechmann, I., Kwidzinski, E., Kovac, A.D., Simburger, E., Horvath, T., Gimsa, U., Dirnagl, U., Priller, J. and Nitsch, R. (2001c) Turnover of rat brain perivascular cells. Exp. Neurol., 168: 242–249. Bechmann, I., Lossau, S., Steiner, B., Mor, G., Gimsa, U. and Nitsch, R. (2000) Reactive astrocytes upregulate Fas (CD95) and Fas ligand (CD95L) expression but do not undergo

programmed cell death during the course of anterograde degeneration. Glia, 32: 25–41. Bechmann, I., Mor, G., Nilsen, J., Eliza, M., Nitsch, R. and Naftolin, F. (1999) FasL (CD95L, Apo1L) is expressed in the normal rat and human brain: evidence for the existence of an immunological brain barrier. Glia, 27: 62–74. Bechmann, I. and Nitsch, R. (1997) Astrocytes and microglial cells incorporate degenerating fibers following entorhinal lesion: a light, confocal, and electron microscopical study using a phagocytosis-dependent labeling technique. Glia, 20: 145–154. Bechmann, I., Peter, S., Beyer, M., Gimsa, U. and Nitsch, R. (2001a) Presence of B7-2 (CD86) and lack of B7-1 (CD80) on myelin phagocytosing MHC-II-positive rat microglia is associated with nondestructive immunity in vivo. FASEB J., 15: 1086–1088. Bechmann, I., Priller, J., Kovac, A., Bontert, M., Wehner, T., Klett, F.F., Bohsung, J., Stuschke, M., Dirnagl, U. and Nitsch, R. (2001b) Immune surveillance of mouse brain perivascular spaces by blood-borne macrophages. Eur. J. Neurosci., 14: 1651–1658. Bechmann, I., Steiner, B., Gimsa, U., Mor, G., Wolf, S., Beyer, M., Nitsch, R. and Zipp, F. (2002) Astrocyte-induced T cell elimination is CD95 ligand dependent. J. Neuroimmunol., 132: 60–65. Beck, K.D., Lamballe, F., Klein, R., Barbacid, M., Schauwecker, P.E., McNeill, T.H., Finch, C.E., Hefti, F. and Day, J.R. (1993) Induction of noncatalytic TrkB neurotrophin receptors during axonal sprouting in the adult hippocampus. J. Neurosci., 13: 4001–4014. Benowitz, L.I., Rodriguez, W.R. and Neve, R.L. (1990) The pattern of GAP-43 immunostaining changes in the rat hippocampal formation during reactive synaptogenesis. Mol. Brain Res., 8: 17–23. Benowitz, L.I. and Routtenberg, A. (1997) GAP-43: an intrinsic determinant of neuronal development and plasticity. Trends Neurosci., 20: 84–91. Biber, K., Sauter, A., Brouwer, N., Copray, S.C. and Boddeke, H.W. (2001) Ischemia-induced neuronal expression of the microglia attracting chemokine secondary lymphoid-tissue chemokine (SLC). Glia, 34: 121–133. Blackstad, T.W. (1958) On the termination of some afferents to the hippocampus and fascia dentata: an experimental study in the rat. Acta Anat. (Basel), 35: 202–214. Blasco-Ibanez, J.M. and Freund, T.F. (1997) Distribution, ultrastructure and connectivity of calretinin-immunoreactive mossy cells of the mouse dentate gyrus. Hippocampus, 7: 307–320. Book, A.A., Wiley, R.G. and Schweitzer, J.B. (1992) Specificity of 192 IgG-saporin for NGF receptor- positive cholinergic basal forebrain neurons in the rat. Brain Res., 590: 350–355. Borrell, V., Ruiz, M., del Rio, J.A. and Soriano, E. (1999) Development of commissural connections in the hippocampus of reeler mice: evidence of an inhibitory influence of Cajal-Retzius cells. Exp. Neurol., 156: 268–282. Bovolenta, P. and Fernaud-Espinosa, I. (2000) Nervous system proteoglycans as modulators of neurite outgrowth. Prog. Neurobiol., 61: 113–132.

521 Burbach, G.J., Dehn, D., Del Turco, D., Staufenbiel, M. and Deller, T. (2004a) Laser microdissection reveals regional and cellular differences in GFAP mRNA upregulation following brain injury, axonal denervation, and amyloid plaque deposition. Glia, 48: 76–84. Burbach, G.J., Dehn, D., Nagel, B., Del Turco, D. and Deller, T. (2004b) Laser microdissection of immunolabeled astrocytes allows quantification of astrocytic gene expression. J. Neurosci. Methods, 138: 141–148. Caceres, A. and Steward, O. (1983) Dendritic reorganization in the denervated dentate gyrus of the rat following entorhinal cortical lesions: a Golgi and electron microscopic analysis. J. Comp. Neurol., 136: 287–295. Campbell, K.A., Bank, B. and Milgram, N.W. (1984) Epileptogenic effects of electrolytic lesions in the hippocampus: role of iron deposition. Exp. Neurol., 86: 506–514. Caroni, P. (1997) Intrinsic neuronal determinants that promote axonal sprouting and elongation. Bio Essays, 19: 767–775. Caroni, P. (1998) Driving the growth cone. Science, 281: 1465–1466. Caroni, P., Aigner, L. and Schneider, C. (1997) Intrinsic neuronal determinants locally regulate extrasynaptic and synaptic growth at the adult neuromuscular junction. J. Cell Biol., 136: 679–692. Chen, L.L., Van Hoesen, G.E., Barnes, C.L. and West, J.R. (1983) Enhanced acetylcholinesterase staining in the hippocampal perforant pathway zone after combined lesions of the septum and entorhinal cortex. Brain Res., 272: 354–359. Collazos-Castro, J.E. and Nieto-Sampedro, M. (2001) Developmental and reactive growth of dentate gyrus afferents: Cellular and molecular interactions. Restor. Neurol. Neurosci., 19: 169–187. Columba-Cabezas, S., Serafini, B., Ambrosini, E. and Aloisi, F. (2003) Lymphoid chemokines CCL19 and CCL21 are expressed in the central nervous system during experimental autoimmune encephalomyelitis: implications for the maintenance of chronic neuroinflammation. Brain Pathol., 13: 38–51. Cotman, C.W., Gentry, C. and Steward, O. (1977) Synaptic replacement in the dentate gyrus after unilateral entorhinal lesion: electron microscopic analysis of the extent of replacement of synapses by the remaining entorhinal cortex. J. Neurocytol., 6: 455–464. Cotman, C.W. and Nadler, J.V. (1978) Reactive synaptogenesis in the hippocampus. In: Cotman C.W. (Ed.), Neuronal Plasticity. Raven Press, New York, pp. 227–271. Dasheiff, R.M. and McNamara, J.O. (1982) Electrolytic entorhinal cortex lesions cause seizures. Brain Res., 231: 444–450. Dawson, M.R., Levine, J.M. and Reynolds, R. (2000) NG2expressing cells in the central nervous system: are they oligodendroglial progenitors? J. Neurosci. Res., 61: 471–479. Dawson, M.R., Polito, A., Levine, J.M. and Reynolds, R. (2003) NG2-expressing glial progenitor cells: an abundant and widespread population of cycling cells in the adult rat CNS. Mol. Cell Neurosci., 24: 476–488.

Dehn, D., Burbach, G.J., Schafer, R. and Deller, T. (2006) NG2 upregulation in the denervated rat fascia dentata following unilateral entorhinal cortex lesion. Glia, 53: 491–500. Del Turco, D., Gebhardt, C., Burbach, G.J., Pleasure, S.J., Lowenstein, D.H. and Deller, T. (2004) Laminar organization of the mouse dentate gyrus: insights from BETA2/Neuro D mutant mice. J. Comp. Neurol., 477: 81–95. Del Turco, D., Gebhardt, C., Woods, A.G., Kapfhammer, J.P., Naumann, T., Frotscher, M., Caroni, P. and Deller, T. (2001) Overexpression of the growth associated protein CAP-23 results in translaminar commissural sprouting in the hippocampus after entorhinal cortex lesion in adult mice. Soc. Neurosci. Abstr., 27: 698.12. Del Turco, D., Woods, A.G., Gebhardt, C., Phinney, A.L., Jucker, M., Frotscher, M. and Deller, T. (2003) Comparison of commissural sprouting in the mouse and rat fascia dentata after entorhinal cortex lesion. Hippocampus, 13: 685–699. Deller, T. (1998) The anatomical organization of the rat fascia dentate — new aspects of laminar organization as revealed by anterograde tracing with Phaseolus vulgaris-leucoagglutinin. Anat. Embryol., 197: 89–103. Deller, T., Bas, O.C., Vlachos, A., Merten, T., Del Turco, D., Dehn, D., Mundel, P. and Frotscher, M. (2006) Plasticity of synaptopodin and the spine apparatus organelle in the rat fascia dentata following entorhinal cortex lesion. J. Comp. Neurol., 499: 471–484. Deller, T., Drakew, A. and Frotscher, M. (1999a) Different primary target cells are important for fiber lamination in the fascia dentata: a lesson from reeler mutant mice. Exp. Neurol., 156: 239–253. Deller, T., Drakew, A., Heimrich, B., Fo¨rster, E., Tielsch, A. and Frotscher, M. (1999b) The hippocampus of the reeler mutant mouse: Fiber segregation in area CA1 depends on the position of the postsynaptic target cells. Exp. Neurol., 156: 254–267. Deller, T. and Frotscher, M. (1997) Lesion-induced plasticity of central neurons: sprouting of single fibers in the rat hippocampus after unilateral entorhinal lesion. Prog. Neurobiol., 53: 687–727. Deller, T., Frotscher, M. and Nitsch, R. (1995a) Morphological evidence for the sprouting of inhibitory commissural fibers in response to the lesion of the excitatory entorhinal input to the rat dentate gyrus. J. Neurosci., 15: 6868–6878. Deller, T., Frotscher, M. and Nitsch, R. (1996a) Sprouting of crossed entorhinodentate fibers after a unilateral entorhinal lesion: anterograde tracing of fiber reorganization with Phaseolus vulgaris-leucoagglutinin (PHAL). J. Comp. Neurol., 365: 42–55. Deller, T., Haas, C.A., Deissenrieder, K., Del Turco, D., Coulin, C., Gebhardt, C., Drakew, A., Schwarz, K., Mundel, P. and Frotscher, M. (2002) Laminar distribution of synaptopodin in normal and reeler mouse brain depends on the position of spine-bearing neurons. J. Comp. Neurol., 453: 33–44. Deller, T., Haas, C.A. and Frotscher, M. (2000) Reorganization of the rat fascia dentata after a unilateral entorhinal cortex lesion: Role of the extracellular matrix. Ann. N.Y. Acad. Sci., 911: 207–220.

522 Deller, T., Haas, C.A. and Frotscher, M. (2001) Sprouting in the hippocampus after entorhinal cortex lesion is layer-specific but not translaminar: which molecules may be involved? Restor. Neurol. Neurosci., 19: 159–167. Deller, T., Haas, C.A., Naumann, T., Joester, A., Faissner, A. and Frotscher, M. (1997) Upregulation of astrocyte-derived tenascin-c correlates with neurite outgrowth in the rat dentate gyrus after unilateral entorhinal cortex lesion. Neuroscience, 81: 829–846. Deller, T., Nitsch, R. and Frotscher, M. (1995b) Phaseolus vulgaris-leucoagglutinin (PHAL) tracing of commissural fibers to the rat dentate gyrus: Evidence for a previously unknown commissural projection to the outer molecular layer. J. Comp. Neurol., 352: 55–68. Deller, T., Nitsch, R. and Frotscher, M. (1996b) Heterogeneity of the commissural projection to the rat dentate gyrus: a Phaseolus vulgaris-leucoagglutinin tracing study. Neuroscience, 75: 111–121. Deller, T., Nitsch, R. and Frotscher, M. (1996c) Layer-specific sprouting of commissural fibers to the rat fascia dentata after unilateral entorhinal cortex lesion: a Phaseolus vulgaris-leucoagglutinin tracing study. Neuroscience, 71: 651–660. Diekmann, S., Ohm, T.G. and Nitsch, R. (1996) Long-lasting transneuronal changes in rat dentate granule cell dendrites after entorhinal cortex lesion-a combined intracellular injection and electron microscopy study. Brain Pathol., 6: 205–214. Drakew, A., Deller, T., Heimrich, B., Gebhardt, C., Del Turco, D., Tielsch, A., Forster, E., Herz, J. and Frotscher, M. (2002) Dentate granule cells in reeler mutants and VLDLR and ApoER2 knockout mice. Exp. Neurol., 176: 12–24. Drojdahl, N., Fenger, C., Nielsen, H.H., Owens, T. and Finsen, B. (2004) Dynamics of oligodendrocyte responses to anterograde axonal (Wallerian) and terminal degeneration in normal and TNF-transgenic mice. J. Neurosci. Res., 75: 203–217. Drojdahl, N., Hegelund, I.V., Poulsen, F.R., Wree, A. and Finsen, B. (2002) Perforant path lesioning induces sprouting of CA3-associated fibre systems in mouse hippocampal formation. Exp. Brain Res., 144: 79–87. Eckenstein, F. and Sofroniew, M.V. (1983) Identification of central cholinergic neurons containing both choline acetytransferase and acetylcholinesterase and of central neurons containing only acetylcholinesterase. J. Neurosci., 3: 2286–2291. Eyupoglu, I.Y., Bechmann, I. and Nitsch, R. (2003) Modification of microglia function protects from lesion-induced neuronal alterations and promotes sprouting in the hippocampus. FASEB J., 17: 1110–1111. Fagan, A.M. and Gage, F.H. (1990) Cholinergic sprouting in the hippocampus: a proposed role for il-1. Exp. Neurol., 110: 105–120. Fagan, A.M. and Gage, F.H. (1994) Mechanismus of sprouting in the adult central nervous system: cellular responses in areas of terminal degeneration and reinnervation in the rat hippocampus. Neuroscience, 58: 705–725. Fagan, A.M., Murphy, B.A., Patel, S.N., Kilbridge, J.F., Mobley, W.C., Bu, G. and Holtzman, D.M. (1998)

Evidence for normal aging of the septo-hippocampal cholinergic system in apoE ( / ) mice but impaired clearance of axonal degeneration products following injury. Exp. Neurol., 151: 314–325. Fagan, A.M., Suhr, S.T., Lucidi-Phillipi, C.A., Peterson, D.A., Holtzman, D.M. and Gage, F.H. (1997) Endogenous FGF-2 is important for cholinergic sprouting in the denervated hippocampus. J. Neurosci., 17: 2499–2511. Faillace, M.P., Zwiller, J., Di Scala, G. and Bernabeu, R. (2006) Odor increases [3H]phorbol dibutyrate binding to protein kinase C in olfactory structures of rat brain. Effect of entorhinal cortex lesion. Brain Res., 1068: 16–22. Faissner, A. and Steindler, D. (1995) Boundaries and inhibitory molecules in developing neural tissues. Glia, 13: 233–254. Fife, B.T., Huffnagle, G.B., Kuziel, W.A. and Karpus, W.J. (2000) CC chemokine receptor 2 is critical for induction of experimental autoimmune encephalomyelitis. J. Exp. Med., 192: 899–905. Finsen, B., Jensen, M.B., Lomholt, N.D., Hegelund, I.V., Poulsen, F.R. and Owens, T. (1999) Axotomy-induced glial reactions in normal and cytokine transgenic mice. Adv. Exp. Med. Biol., 468: 157–171. Frey, D., Laux, T., Xu, L., Schneider, C. and Caroni, P. (2000) Shared and unique roles of CAP23 and GAP43 in actin regulation, neurite outgrowth, and anatomical plasticity. J. Cell Biol., 149: 1443–1454. Frotscher, M. (1988) Neuronal elements in the hippocampus and their synaptic connections. Adv. Anat. Embryol., 111: 2–17. Frotscher, M., Heimrich, B. and Deller, T. (1997) Sprouting in the hippocampus is layer-specific. Trends Neurosci., 20: 218–223. Fujise, N., Liu, Y., Hori, N. and Kosaka, T. (1998) Distribution of calretinin immunoreactivity in the mouse dentate gyrus: II. Mossy cells, with special reference to their dorsoventral difference in calretinin immunoreactivity. Neuroscience, 82: 181–200. Gage, F.H., Olejniczak, P. and Armstrong, D.M. (1988) Astrocytes are important for sprouting in the septohippocampal circuit. Exp. Neurol., 102: 2–13. Gall, C., Ivy, G. and Lynch, G. (1986) Neuroanatomical plasticity. Its role in organizing and reorganizing the central nervous system. Hum. Growth, 2: 411–436. Gall, C. and Lynch, G. (1978) Rapid axon sprouting in the neonatal rat hippocampus. Brain Res., 153: 357–362. Gall, C. and Lynch, G. (1981) Fiber architecture of the dentate gyrus following ablation of the entorhinal cortex in rats of different ages: evidence for two forms of axon sprouting in the immature brain. Neuroscience, 6: 903–910. Gall, C. and Lynch, G.S. (1980) The regulation of fiber growth and synaptogenesis in the developing hippocampus. Curr. Top. Dev. Biol., 1: 159–180. Gall, C., McWilliams, R. and Lynch, G. (1979a) The effect of collateral sprouting on the density of innervation of normal target sites: implications for theories on the regulation of the size of the developing synaptic domains. Brain Res., 175: 37–47.

523 Gall, C., Rose, G. and Lynch, G. (1979b) Proliferative and migratory activity of glial cells in the partially deafferented hippocampus. J. Comp. Neurol., 183: 539–550. Gebhardt, C., Del Turco, D., Drakew, A., Tielsch, A., Herz, J., Frotscher, M. and Deller, T. (2002) Abnormal positioning of granule cells alters afferent fiber distribution in the mouse fascia dentata: morphologic evidence from reeler, apolipoprotein E receptor 2-, and very low density lipoprotein receptor knockout mice. J. Comp. Neurol., 445: 278–292. Gehrmann, J., Schoen, S.W. and Kreutzberg, G.W. (1991) Lesion of the rat entorhinal cortex leads to a rapid microglial reaction in the dentate gyrus. Acta Neuropathol., 82: 442–455. Gold, R., Linington, C. and Lassmann, H. (2006) Understanding pathogenesis and therapy of multiple sclerosis via animal models: 70 years of merits and culprits in experimental autoimmune encephalomyelitis research. Brain, 129: 1953–1971. Goldmann, J., Kwidzinski, E., Brandt, C., Mahlo, J., Richter, D. and Bechmann, I. (2006) T cells traffic from brain to cervical lymph nodes via the cribroid plate and the nasal mucosa. J. Leukoc. Biol., 80: 797–801. Goldowitz, D., White, W.F., Steward, O., Cotman, C.W. and Lynch, G. (1975) Anatomical evidence for a projection from the entorhinal cortex to the contralateral dentate gyrus of the rat. Exp. Neurol., 47: 433–441. Gomez-Pinilla, F., Lee, J.W. and Cotman, C.W. (1992) Basic FGF in adult rat brain: cellular distribution and response to entorhinal lesion and fimbria-fornix transection. J. Neurosci., 12: 345–355. van Groen, T., Kadish, I. and Wyss, J.M. (2002) Species differences in the projections from the entorhinal cortex to the hippocampus. Brain Res. Bull., 57: 553–556. van Groen, T., Miettinen, P. and Kadish, I. (2003) The entorhinal cortex of the mouse: organization of the projection to the hippocampal formation. Hippocampus, 13: 133–149. Guthrie, K.M., Nguyen, T. and Gall, C.M. (1995) Insulin-like growth factor-1 mRNA is increased in deafferented hippocampus: spatiotemporal correspondence of a trophic event with axon sprouting. J. Comp. Neurol., 352: 147–160. Guthrie, K.M., Woods, A.G., Nguyen, T. and Gall, C.M. (1997) Astroglial ciliary neurotrophic factor mRNA expression is increased in fields of axonal sprouting in deafferented hippocampus. J. Comp. Neurol., 386: 137–148. Gwag, B.J., Sessler, F.M., Kimmerer, K. and Springer, J.E. (1994) Neurotrophic factor mRNA expression in the dentate gyrus is increased following angular bundle transection. Brain Res., 647: 23–29. Haas, C.A., Rauch, U., Thon, N., Merten, T. and Deller, T. (1999) Entorhinal cortex lesion in adult rats induce the expression of the neuronal chondroitin sulfate proteoglycan neurocan in reactive astrocytes. J. Neurosci., 19: 9953–9963. Hailer, N.P., Bechmann, I., Heizmann, S. and Nitsch, R. (1997) Adhesion molecule expression on phagocytic microglial cells following anterograde degeneration of perforant path axons. Hippocampus, 7: 341–349.

Hardman, R., Evans, D.J., Fellows, L., Hayes, B., Rupniak, H.T., Barnes, J.C. and Higgins, G.A. (1997) Evidence for recovery of spatial learning following entorhinal cortex lesions in mice. Brain Res., 758: 187–200. Harris, E.W., Lasher, S.S. and Steward, O. (1978) Habitutation like decrements in transmission along the normal and lesioninduced temporodentate pathways in the rat. Brain Res., 151: 623–631. Heckers, S., Ohtake, T., Wiley, R.G., Lappi, D.A., Geula, C. and Mesulam, M.M. (1994) Complete and selective cholinergic denervation of rat neocortex and hippocampus but not amygdala by an immunotoxin against the p75 NGF receptor. J. Neurosci., 14: 1271–1289. Henderson, Z., Harrison, P.S., Jagger, E. and Beeby, J.H. (1998) Density of choline acetyltransferase-immunreactive terminals in the rat dentate gyrus after entorhinal cortex lesions: a quantitative light microscope study. Exp. Neurol., 152: 50–63. Hesselgesser, J. and Horuk, R. (1999) Chemokine and chemokine receptor expression in the central nervous system. J. Neurovirol., 5: 13–26. Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. J. Comp. Neurol., 144: 215–232. Hoff, S.F., Scheff, S.W., Kwan, A.Y. and Cotman, C.W. (1981) A new type of lesion-induced synaptogenesis: I. Synaptic turnover in non-denervated zones of the dentate gyrus in young adult rats. Brain Res., 222: 1–13. Ho¨ke, A. and Silver, J. (1996) Proteoglycans and other repulsive molecules in glial boundaries during development and regeneration of the nervous system. Prog. Brain Res., 108: 149–163. Huang, D.R., Wang, J., Kivisakk, P., Rollins, B.J. and Ransohoff, R.M. (2001) Absence of monocyte chemoattractant protein 1 in mice leads to decreased local macrophage recruitment and antigen-specific T helper cell type 1 immune response in experimental autoimmune encephalomyelitis. J. Exp. Med., 193: 713–726. Izikson, L., Klein, R.S., Charo, I.F., Weiner, H.L. and Luster, A.D. (2000) Resistance to experimental autoimmune encephalomyelitis in mice lacking the CC chemokine receptor (CCR)2. J. Exp. Med., 192: 1075–1080. Jakab, R.L. and Leranth, C. (1995) Septum. In: Paxinos G. (Ed.), The Rat Nervous System (2nd ed.). Academic Press, San Diego, CA, pp. 405–442. Jensen, M.B., Finsen, B. and Zimmer, J. (1997) Morphological and immunophenotypic microglial changes in the denervated fascia dentata of adult rats: correlation with blood-brain barrier damage and astroglial reactions. Exp. Neurol., 143: 103–116. Jensen, M.B., Gonza`lez, B., Castellano, B. and Zimmer, J. (1994) Microglial and astroglial reactions to anterograde axonal degeneration: a histochemical and immuncytochemical study of the adult rat fascia dentata after entorhinal perforant path lesions. Exp. Brain Res., 98: 245–260. Jensen, M.B., Hegelund, I.V., Lomholt, N.D., Finsen, B. and Owens, T. (2000a) IFNgamma enhances microglial reactions

524 to hippocampal axonal degeneration. J. Neurosci., 20: 3612–3621. Jensen, M.B., Hegelund, I.V., Poulsen, F.R., Owens, T., Zimmer, J. and Finsen, B. (1999) Microglial reactivity correlates to the density and the myelination of the anterogradely degenerating axons and terminals following perforant path denervation of the mouse fascia dentata. Neuroscience, 93: 507–518. Jensen, M.B., Poulsen, F.R. and Finsen, B. (2000b) Axonal sprouting regulates myelin basic protein gene expression in denervated mouse hippocampus. Int. J. Dev. Biol., 18: 221–235. de Jong, E.K., Dijkstra, I.M., Hensens, M., Brouwer, N., van Amerongen, M., Liem, R.S., Boddeke, H.W. and Biber, K. (2005) Vesicle-mediated transport and release of CCL21 in endangered neurons: a possible explanation for microglia activation remote from a primary lesion. J. Neurosci., 25: 7548–7557. Jucker, M., D‘Amato, F., Mondadori, C., Mohajeri, H., Magyar, J., Bartsch, U. and Schachner, M. (1996) Expression of the neural adhesion molecule L1 in the deafferented dentate gyrus. Neuroscience, 75: 703–715. Kadish, I. and van Groen, T. (2002) Low levels of estrogen significantly diminish axonal sprouting after entorhinal cortex lesions in the mouse. J. Neurosci., 22: 4095–4102. Kadish, I. and van Groen, T. (2003) Differences in lesioninduced hippocampal plasticity between mice and rats. Neuroscience, 11: 499–509. Kadish, I., Pradier, L. and van Groen, T. (2002) Transgenic mice expressing the human presenilin 1 gene demonstrate enhanced hippocampal reorganization following entorhinal cortex lesion. Brain Res. Bull., 57: 587–594. Kapfhammer, J.P. and Schwab, M.E. (1992) Modulators of neuronal migration and neurite growth. Curr. Opin. Cell Biol., 4: 863–868. Karpus, W.J. and Ransohoff, R.M. (1998) Chemokine regulation of experimental autoimmune encephalomyelitis: temporal and spatial expression patterns govern disease pathogenesis. J. Immunol., 161: 2667–2671. Kawaja, M.D. and Gage, F.H. (1991) Reactive astrocytes are substrates for the growth of adult CNS axons in the presence of elevated levels of nerve growth factor. Neuron, 7: 1019–1030. Kelley, M.S. and Steward, O. (1996a) The process of reinnervation in the dentate gyrus of adult rats: physiological events at the time of the lesion and during the early postlesion period. Exp. Neurol., 139: 73–82. Kelley, M.S. and Steward, O. (1996b) The role of postlesion seizures and spreading depression in the upregulation of glial fibrillary acidic protein mRNA after entorhinal cortex lesions. Exp. Neurol., 139: 83–94. Kelley, M.S. and Steward, O. (1998) Injury-induced physiological events that may modulate gene expression in neurons and glia. Rev. Neurosci., 8: 147–177. Kerschensteiner, M., Gallmeier, E., Behrens, L., Leal, V.V., Misgeld, T., Klinkert, W.E., Kolbeck, R., Hoppe, E., Oropeza-Wekerle, R.L., Bartke, I., Stadelmann, C., Lassmann,

H., Wekerle, H. and Hohlfeld, R. (1999) 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: 865–870. Kipnis, J., Yoles, E., Schori, H., Hauben, E., Shaked, I. and Schwartz, M. (2001) Neuronal survival after CNS insult is determined by a genetically encoded autoimmune response. J. Neurosci., 21: 4564–4571. Kosaka, T., Katsumaru, H., Hama, K., Wu, J.Y. and Heizmann, C.W. (1987) GABAergic neurons containing CA2+ binding protein parvalbumin in the rat hippocampus and dentate gyrus. Brain Res., 419: 119–130. Kovac, A.D., Kwidzinski, E., Heimrich, B., Bittigau, P., Deller, T., Nitsch, R. and Bechmann, I. (2004) Entorhinal cortex lesion in the mouse induces transsynaptic death of perforant path target neurons. Brain Pathol., 14: 249–257. Ladeby, R., Wirenfeldt, M., Dalmau, I., Gregersen, R., Garcia-Ovejero, D., Babcock, A., Owens, T. and Finsen, B. (2005a) Proliferating resident microglia express the stem cell antigen CD34 in response to acute neural injury. Glia, 50: 121–131. Ladeby, R., Wirenfeldt, M., Garcia-Ovejero, D., Fenger, C., Dissing-Olesen, L., Dalmau, I. and Finsen, B. (2005b) Microglial cell population dynamics in the injured adult central nervous system. Brain Res. Brain Res. Rev., 48: 196–206. Laux, T., Fukami, K., Thelen, M., Golub, T., Frey, D. and Caroni, P. (2000) GAP43, MARCKS, and CAP23 modulate PI(4,5)P(2) at plasmalemmal rafts, and regulate cell cortex actin dynamics through a common mechanism. J. Cell Biol., 149: 1455–1472. Lee, M.Y., Deller, T., Kirsch, M., Frotscher, M. and Hofmann, H.D. (1997) Differential regulation of CNTF and CNTF receptor alpha expression in astrocytes and neurons of the fascia dentata following entorhinal cortex lesion. J. Neurosci., 17: 1137–1146. Levey, A.I., Wainer, B.H., Mufson, E.J. and Mesulam, M.M. (1983) Co-localization of acetylcholinesterase and choline acetyltransferase in the rat cerebrum. Neuroscience, 9: 9–22. Levey, A.I., Wainer, B.H., Rye, D.B., Mufson, E.J. and Mesulam, M.M. (1984) Choline acetyltransferase-immunoreactive neurons intrinsic to rodent cortex and distinction from acetylcholinesterase-positive neurons. Neuroscience, 13: 341–353. Linke, R., Schwegler, H. and Boldyreva, M. (1994) Cholinergic and GABAergic septo-hippocampal projection neurons in mice: a retrograde tracing study combined with double immunocytochemistry for choline acetyltransferase and parvalbumin. Brain Res., 653: 73–80. Liu, Y., Fujise, N. and Kosaka, T. (1996) Distribution of calretinin immunoreactivity in the mouse dentate gyrus. I. General description. Exp. Brain Res., 108: 389–403. Loesche, J. and Steward, O. (1977) Behavioral correlates of denervation and reinnervation of the hippocampal formation of the rat: recovery of alternation performance following unilateral entorhinal cortex lesions. Brain Res. Bull., 2: 31–39.

525 Luster, A.D. (1998) Chemokines–chemotactic cytokines that mediate inflammation. N. Engl. J. Med., 338: 436–445. Lynch, G. and Cotman, C. (1975) The hippocampus as a model for studying anatomical plasticity in the adult brain. In: Isaacson R.L. and Pribram K.H. (Eds.), The Hippocampus, Vol. 1. Plenum Press, New York, pp. 123–154. Lynch, G., Gall, C., Rose, G. and Cotman, C.W. (1976) Changes in the distribution of the dentate gyrus associational system following unilateral or bilateral entorhinal lesion in the adult rat. Brain Res., 110: 57–71. Lynch, G., Matthews, D.A., Mosko, S., Parks, T. and Cotman, C.W. (1972) Induced acetylcholinesterase-rich layer in rat dentate gyrus following entorhinal lesions. Brain Res., 42: 311–318. Lynch, G., Stanfield, B. and Cotman, C.W. (1973) Developmental differences in postlesion axonal growth in the hippocampus. Brain Res., 59: 155–168. Lynch, G.S., Stanfield, B., Parks, T. and Cotman, C.W. (1974) Evidence for selective post-lesion axonal growth in the dentate gyrus of the rat. Brain Res., 69: 1–11. Mackay, C.R. (2001) Chemokines: immunology’s high impact factors. Nat. Immunol., 2: 95–101. Margolis, R.U. and Margolis, R.K. (1997) Chondroitin sulfate proteoglycans as mediators of axon growth and pathfinding. Cell Tissue Res., 290: 343–348. Marrone, D.F., LeBoutillier, J.C. and Petit, T.L. (2004a) Changes in synaptic ultrastructure during reactive synaptogenesis in the rat dentate gyrus. Brain Res., 1005: 124–136. Marrone, D.F., LeBoutillier, J.C. and Petit, T.L. (2004b) Comparative analyses of synaptic densities during reactive synaptogenesis in the rat dentate gyrus. Brain Res., 996: 19–30. Matthews, D.A., Cotman, C.W. and Lynch, G. (1976a) An electron microscopic study of lesion-induced synaptogenesis in the dentate gyrus of the adult rat. I. Magnitude and time course of degeneration. Brain Res., 115: 1–21. Matthews, D.A., Cotman, C.W. and Lynch, G. (1976b) An electron microscopic study of lesion-induced synaptogenesis in the dentate gyrus of the adult rat. II. Reappearance of morphologically normal synaptic contacts. Brain Res., 115: 23–41. Mayer, J., Hamel, M.G. and Gottschall, P.E. (2005) Evidence for proteolytic cleavage of brevican by the ADAMTSs in the dentate gyrus after excitotoxic lesion of the mouse entorhinal cortex. BMC Neurosci., 6: 52. McKeon, R.J., Vietje, B.P. and Wells, J. (1989) Increase in acetylcholinesterase in the molecular layer of the denate gyrus in the absence of septal inputs following selective granule cell lesions. Brain Res., 503: 317–321. Meier, S., Brauer, A.U., Heimrich, B., Nitsch, R. and Savaskan, N.E. (2004) Myelination in the hippocampus during development and following lesion. Cell Mol. Life Sci., 61: 1082–1094. Meier, S., Brauer, A.U., Heimrich, B., Schwab, M.E., Nitsch, R. and Savaskan, N.E. (2003) Molecular analysis of Nogo expression in the hippocampus during development and following lesion and seizure. FASEB J., 17: 1153–1155.

Mingorance, A., Fontana, X., Soriano, E. and del Rio, J.A. (2005) Overexpression of myelin-associated glycoprotein after axotomy of the perforant pathway. Mol. Cell Neurosci., 29: 471–483. Miwa, C. and Ueki, A. (1996) Effects of entorhinal cortex lesion on learning behavior and on hippocampus in the rat. Psychiatry Clin. Neurosci., 50: 223–230. Mutlu, L., Brandt, C., Kwidzinski, E., Sawitzki, B., Gimsa, U., Mahlo, J., Aktas, O., Nitsch, R., van Zwam, M., Laman, J. and Bechmann, I. (2007) Tolerogenic effect of fiber tract injury: Peripheral apoptosis of T cells and reduction of EAE severity following entorhinal cortex lesion. Exp. Brain Res., 178: 542–553. Nadler, J.V., Cotman, C., Paoletti, C. and Lynch, G.S. (1977a) Histochemical evidence of altered development of cholinergic fibers in the rat dentate gyrus following lesions. II. Effects of partial entorhinal and simultaneous multiple lesions. J. Comp. Neurol., 171: 589–604. Nadler, J.V., Cotman, C.W. and Lynch, G.S. (1977b) Histochemical evidence of altered development of cholinergic fibers in the rat dentate gyrus following lesions. I. Time course after complete unilateral entorhinal lesion at various ages. J. Comp. Neurol., 171: 561–587. Nadler, J.V. and Evenson, D.A. (1982) Autoradiographic evidence that septohippocampal fibers reinnervate fascia dentata denervated by entorhinal lesion during development. Anat. Embryol., 165: 113–123. Nadler, J.V., Perry, B.W. and Cotman, C.W. (1980) Interaction with CA4-derived fibers accounts for distribution of septohippocampal fibers in rat fascia denata after entorhinal lesion. Exp. Neurol., 68: 185–194. Nafstad, P.H.J. (1967) An electron microscope study on the termination of the perforant path fibres in the hippocampus and fascia denata. Z. Zellforsch. Mikrosk. Anat., 76: 532–542. Naumann, T., Deller, T., Bender, R. and Frotscher, M. (1997) 192 IgG-saporin-induced loss of cholinergic neurons in the septum abolishes cholinergic sprouting after unilateral entorhinal lesion in the rat. Eur. J. Neurosci., 9: 1304–1313. Nielsen, H.H., Ladeby, R., Drojdahl, N., Peterson, A.C. and Finsen, B. (2006) Axonal degeneration stimulates the formation of NG2+ cells and oligodendrocytes in the mouse. Glia, 54: 105–115. Nitsch, R., Bader, S. and Frotscher, M. (1992) Reorganization of input synapses of parvalbumin-containing neurons in the rat fascia dentata following entorhinal lesion. Neurosci. Lett., 135: 33–36. Nitsch, R. and Frotscher, M. (1991) Maintenance of peripheral dendrites of GABAergic neurons requires specific input. Brain Res., 554: 304–307. Nitsch, R. and Frotscher, M. (1992) Reduction of posttraumatic transneuronal ‘‘early gene’’ activation and dendritic atrophy by the N-methyl-D-aspartate receptor antagonist MK-801. Proc. Natl. Acad. Sci. U.S.A., 89: 5197–5200. Nitsch, R. and Frotscher, M. (1993) Transneuronal changes in dendrites of GABAergic parvalbumin-containing neurons of

526 the rat fascia dentata following entorhinal lesion. Hippocampus, 3: 481–490. Nitsch, R., Soriano, E. and Frotscher, M. (1990) The parvalbumin-containing nonpyramidal neurons in the rat hippocampus. Anat. Embryol., 181: 413–425. Nyakas, C., Luiten, P.G.M., Balkan, B. and Spencer, D.G.J. (1988) Changes in septohippocampal projections after lateral entorhinal or combined entorhinal-raphe` lesions as studied by anterograde tracing methods. Brain Res. Bull., 21: 285–293. Nyakas, C., Luiten, P.G.M., Spencer, D.G. and Traber, J. (1987) Detailed projection patterns of septal and diagonal band efferents to the hippocampus in the rat with emphasis on innervation of CA1 and dentate gyrus. Brain Res. Bull., 18: 533–545. Otal, R., Burgaya, F., Frisen, J., Soriano, E. and Martinez, A. (2006) Ephrin-A5 modulates the topographic mapping and connectivity of commissural axons in murine hippocampus. Neuroscience, 141: 109–121. Parent, A., Dea, D., Quirion, R. and Poirier, J. (1993) [3H]phorbol ester binding sites and neuronal plasticity in the hippocampus following entorhinal cortex lesions. Brain Res., 607: 23–32. Parnavelas, J.G., Lynch, G., Brecha, N., Cotman, C.W. and Globus, A. (1974) Spine loss and regrowth in hippocampus following deafferentation. Nature, 248: 71–73. Pearlman, A.L. and Sheppard, A.M. (1996) Extracellular matrix in early cortical development. Prog. Brain Res., 108: 117–134. Pedersen, M.D., Minuzzi, L., Wirenfeldt, M., Meldgaard, M., Slidsborg, C., Cumming, P. and Finsen, B. (2006) Up-regulation of PK11195 binding in areas of axonal degeneration coincides with early microglial activation in mouse brain. Eur. J. Neurosci., 24: 991–1000. Peterson, G.M. (1994) Sprouting of central noradrenergic fibers in the dentate gyrus following combined lesions of its entorhinal and septal afferents. Hippocampus, 4: 635–648. Phinney, A.L., Calhoun, M.E., Woods, A.G., Deller, T. and Jucker, M. (2004) Stereological analysis of the reorganization of the dentate gyrus following entorhinal cortex lesion in mice. Eur. J. Neurosci., 19: 1731–1740. Poirier, J., May, P.C., Osterburg, H.H., Geddes, J., Cotman, C.W. and Finch, C.E. (1990) Selective alterations of RNA in rat hippocampus after entorhinal cortex lesioning. Proc. Natl. Acad. Sci. U.S.A., 87: 303–307. Poulsen, F.R., Lagord, C., Courty, J., Pedersen, E.B., Barritault, D. and Finsen, B. (2000) Increased synthesis of heparin affin regulatory peptide in the perforant path lesioned mouse hippocampal formation. Exp. Brain Res., 135: 319–330. Ramirez, J.J. (2001) The role of axonal sprouting in functional reorganization after CNS injury: lessons from the hippocampal formation. Restor. Neurol. Neurosci., 19: 237–262. Ramirez, J.J. and Stein, D.G. (1984) Sparing and recovery of spatial alternation performance after entorhinal cortex lesions in rats. Behav. Brain Res., 13: 53–61. Rappert, A., Bechmann, I., Pivneva, T., Mahlo, J., Biber, K., Nolte, C., Kovac, A.D., Gerard, C., Boddeke, H.W., Nitsch,

R. and Kettenmann, H. (2004) CXCR3-dependent microglial recruitment is essential for dendrite loss after brain lesion. J. Neurosci., 24: 8500–8509. Rappert, A., Biber, K., Nolte, C., Lipp, M., Schubel, A., Lu, B., Gerard, N.P., Gerard, C., Boddeke, H.W. and Kettenmann, H. (2002) Secondary lymphoid tissue chemokine (CCL21) activates CXCR3 to trigger a Cl current and chemotaxis in murine microglia. J. Immunol., 168: 3221–3226. Reeves, T.M. and Steward, O. (1986) Emergence of the capacity for LTP during reinnervation of the dentate gyrus: evidence that abnormally shaped spines can mediate LTP. Exp. Brain Res., 65: 167–175. Reeves, T.M. and Steward, O. (1988) Changes in the firing properties of neurons in the dentate gyrus with denervation and reinnervation: implications for behavioral recovery. Exp. Neurol., 102: 37–49. Rollins, B.J. (1997) Chemokines. Blood, 90: 909–928. Rose, G., Lynch, G. and Cotman, C.W. (1976) Hypertrophy and redistribution of astrocytes in the deafferented dentate gyrus. Brain Res. Bull., 1: 87–92. Sattler, R., Charlton, M.P., Hafner, M. and Tymianski, M. (1998) Distinct influx pathways, not calcium load, determine neuronal vulnerability to calcium neurotoxicity. J. Neurochem., 71: 2349–2364. Savaskan, N.E. and Nitsch, R. (2001) Molecules involved in reactive sprouting in the hippocampus. Rev. Neurosci., 12: 195–215. Schafer, M., Brauer, A.U., Savaskan, N.E., Rathjen, F.G. and Brummendorf, T. (2005) Neurotractin/kilon promotes neurite outgrowth and is expressed on reactive astrocytes after entorhinal cortex lesion. Mol. Cell Neurosci., 29: 580–590. Schauwecker, P.E. and Steward, O. (1997) Genetic influences on cellular reactions to brain injury: activation of microglia in denervated neuropil in mice carrying a mutation (WLDs) that causes delayed wallerian degeneration. J. Comp. Neurol., 380: 82–94. Scheff, S.W. and Cotman, C.W. (1977) Recovery of spontaneous alternation following lesions of the entorhinal cortex in adult rats: possible correlation to axon sprouting. Behav. Bio., 21: 286–293. Schenk, F. and Morris, R.G.M. (1985) Dissociation between components of spatial memory in rats after recovery from the effects of retrohippocampal lesions. Exp. Brain Res., 58: 11–28. Schwab, M.E., Kapfhammer, J.P. and Bandtlow, C.E. (1993) Inhibitors of neurite growth. Annu. Rev. Neurosci., 16: 565–595. Shi, B. and Stanfield, B.B. (1996) Differential sprouting responses in axonal fiber systems in the dentate gyrus following lesions of the perforant path in WLDs mutant mice. Brain Res., 740: 89–101. Siebert, H., Sachse, A., Kuziel, W.A., Maeda, N. and Bruck, W. (2000) The chemokine receptor CCR2 is involved in macrophage recruitment to the injured peripheral nervous system. J. Neuroimmunol., 110: 177–185. Soriano, E. and Frotscher, M. (1994) Mossy cells of the rat fascia dentata are glutamate-immunoreactive. Hippocampus, 4: 65–70.

527 Soriano, E., del Rio, J.A., Martinez, A. and Super, H. (1994) Organization of the embryonic and early postnatal murine hippocampus. I. Immunocytochemical characterization of neuronal populations in the subplate and marginal zone. J. Comp. Neurol., 342: 571–595. Stanfield, B.B., Caviness, V.S.J. and Cowan, W.M. (1979) The organization of certain afferents to the hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol., 185: 461–483. Stanfield, B.B. and Cowan, W.M. (1979a) The development of the hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol., 185: 423–460. Stanfield, B.B. and Cowan, W.M. (1979b) The morphology of the hippocampus and dentate gyrus in normal and reeler mice. J. Comp. Neurol., 185: 393–422. Stanfield, B.B. and Cowan, W.M. (1982) The sprouting of septal afferents to the dentate gyrus after lesions of the entorhinal cortex in adult rats. Brain Res., 232: 162–170. Sta¨ubli, U., Gall, C. and Lynch, G. (1984) The distribution of the commissural-associational afferents of the dentate gyrus after perforant path lesions in one-day-old rats. Brain Res., 292: 156–159. Steward, O. (1976a) Reinnervation of the dentate gyrus by homologous afferents following entorhinal cortical lesion in adult rats. Science, 194: 426–428. Steward, O. (1976b) Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat. J. Comp. Neurol., 167: 285–314. Steward, O. (1980) Trajectory of contralateral entorhinal axons which reinnervate the fascia dentata of the rat following ipsilateral entorhinal lesions. Brain Res., 183: 277–289. Steward, O. (1991) Synapse replacement on cortical neurons following denervation. Cereb. Cortex, 9: 81–132. Steward, O. (1992) Signals that induce sprouting in the central nervous system: sprouting is delayed in a strain of mouse exhibiting delayed axonal degeneration. Exp. Neurol., 118: 340–351. Steward, O. (1994a) Cholinergic sprouting is blocked by repeated induction of electroconvulsive seizures, a manipulation that induces a persistent reactive state in astrocytes. Exp. Neurol., 129: 103–111. Steward, O. (1994b) Reorganization of neuronal circuitry following central nervous system trauma: naturally occurring processes and opportunities for therapeutic intervention. In: Salzman S.K. and Faden A.I. (Eds.), The Neurobiology of Central Nervous System Trauma. Oxford University Press, New York, pp. 266–287. Steward, O., Cotman, C.W. and Lynch, G. (1974) Growth of a new fiber projection in the brain of adult rats: reinnervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral entorhinal lesion. Exp. Brain Res., 20: 45–66. Steward, O., Cotman, C.W. and Lynch, G.S. (1973) Re-establishment of electrophysiologically functional entorhinal cortical input to the dentate gyrus deafferented by ipsilateral entorhinal lesions: innervation by the contralateral cortex. Exp. Brain Res., 18: 396–414.

Steward, O., Cotman, C.W. and Lynch, G.S. (1976) A quantitative autoradiographic and electrophysiological study of the reinnervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral lesions. Brain Res., 114: 181–200. Steward, O., Kelley, M.S. and Schauwecker, P.E. (1997) Signals that regulate astroglial gene expression: induction of GFAP mRNA following seizures or injury is blocked by protein synthesis inhibitors. Exp. Neurol., 148: 100–109. Steward, O., Kelley, M.S. and Torre, E.R. (1993) The process of reinnervation in the dentate gyrus of adult rats: temporal relationship between changes in the level of glial fibrillary acidic protein (GFAP ) and GFAP mRNA in reactive astrocytes. Exp. Neurol., 124: 167–183. Steward, O. and Scoville, S.A. (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol., 169: 347–370. Steward, O., Torre, E.R., Phillips, L. and Trimmer, P.A. (1990) The process of reinnervation in the dentate gyrus of adult rats: time course of increases in mRNA for glial fibrillary acidic protein. J. Neurosci., 10: 2373–2384. Steward, O. and Trimmer, P.A. (1997) Genetic influences on cellular reactions to CNS injury: the reactive response of astrocytes in denervated neuropil regions in mice carrying a mutation (Wld(S)) that causes delayed Wallerian degeneration. J. Comp. Neurol., 380: 70–81. Steward, O. and Vinsant, S.L. (1978) Identification of the cells of origin of a central pathway which sprouts following lesions in mature rats. Brain Res., 147: 223–243. Steward, O. and Vinsant, S.L. (1983) The process of reinnervation in the dentate gyrus of the adult rat: a quantitative electron microscopic analysis of terminal proliferation and reactive synaptogenesis. J. Comp. Neurol., 214: 370–386. Steward, O., Vinsant, S.L. and Davis, L. (1988) The process of reinnervation in the dentate gyrus of adlut rats: an ultrastructural study of changes in presynaptic terminals as a result of sprouting. J. Comp. Neurol., 267: 203–210. Stone, D.J., Rozovsky, I., Morgan, T.E., Anderson, C.P. and Finch, C.E. (1998) Increased synaptic sprouting in response to estrogen via an apolipoprotein E-dependent mechanism: implications for Alzheimer’s disease. J. Neurosci., 18: 3180–3185. Storm-Mathisen, J. (1974) Choline acetyltransferase and acetylcholinesterase in fascia dentata following lesion of the entorhinal afferents. Brain Res., 80: 181–197. Styren, S.D., Lagenaur, C.F., Miller, P.D. and DeKosky, S.T. (1994) Rapid expression and transport of embryonic N-CAM in dentate gyrus following entorhinal cortex lesion: ultrastructural analysis. J. Comp. Neurol., 349: 486–492. Styren, S.D., Miller, P.D., Lagenaur, C.F. and DeKosky, S.T. (1995) Alternate strategies in lesion-induced reactive synaptogenesis: differential expression of L1 in two populations of sprouting axons. Exp. Neurol., 131: 165–173. Supe`r, H. and Soriano, E. (1994) The organization of the embryonic and early postnatal murine hippocampus. II. Development of entorhinal, commissural, and septal connections studied with the lipophilic tracer dil. J. Comp. Neurol., 344: 101–120.

528 Thon, N., Haas, C.A., Rauch, U., Merten, T., Fa¨ssler, R., Frotscher, M. and Deller, T. (2000) The chondroitin sulphate proteoglycan brevican is upregulated by astrocytes after entorhinal cortex lesions in adult rats. Eur. J. Neurosci., 12: 2547–2558. Ueki, A., Miwa, C., Oohara, K. and Miyoshi, K. (1996) Histological evidence for cholinergic alteration in the hippocampus following entorhinal cortex lesion. J. Neurol. Sci., 142: 7–11. White, F., Nicoll, J.A. and Horsburgh, K. (2001a) Alterations in ApoE and ApoJ in relation to degeneration and regeneration in a mouse model of entorhinal cortex lesion. Exp. Neurol., 169: 307–318. White, F., Nicoll, J.A., Roses, A.D. and Horsburgh, K. (2001b) Impaired neuronal plasticity in transgenic mice expressing human apolipoprotein E4 compared to E3 in a model of entorhinal cortex lesion. Neurobiol. Dis., 8: 611–625. White, W.F., Goldowitz, D., Lynch, G. and Cotman, C.W. (1976) Electrophysiological analysis of the projection from the contralateral entorhinal cortex to the dentate gyrus in normal rats. Brain Res., 114: 201–209. Wiley, R.G., Oeltmann, T.N. and Lappi, D.A. (1991) Immunolesioning: selective destruction of neurons using immunotoxin to rat NGF receptor. Brain Res., 562: 149–153. Wirenfeldt, M., Dalmau, I. and Finsen, B. (2003) Estimation of absolute microglial cell numbers in mouse fascia dentata using unbiased and efficient stereological cell counting principles. Glia, 44: 129–139. Woods, A.G., Guthrie, K.M., Kurlawalla, M.A. and Gall, C.M. (1998) Deafferentation-induced increases in hippocampal insulin-like growth factor-1 messenger RNA expression

are severely attenuated in middle aged and aged rats. Neuroscience, 83: 663–668. Woods, A.G., Poulsen, F.R. and Gall, C.M. (1999) Dexamethasone specifically suppresses microglial trophic responses to hippocampal deafferentation. Neuroscience, 91: 1277–1289. Ying, G.X., Huang, C., Jiang, Z.H., Liu, X., Jing, N.H. and Zhou, C.F. (2002) Up-regulation of cystatin C expression in the murine hippocampus following perforant path transections. Neuroscience, 112: 289–298. Yoshida, K. and Gage, F.H. (1991) Fibroblast growth factors stimulate nerve growth factor synthesis and secretion by astrocytes. Brain Res., 538: 118–126. Zappone, C.A. and Sloviter, R.S. (2001) Commissurally projecting inhibitory interneurons of the rat hippocampal dentate gyrus: a colocalization study of neuronal markers and the retrograde tracer Fluoro-Gold. J. Comp. Neurol., 441: 324–344. Zimmer, J. and Hjorth-Simonsen, A. (1975) Crossed pathways from the entorhinal area to the fascia dentata. II. Provokable in rats. J. Comp. Neurol., 161: 71–102. Zimmer, J., Laurberg, S. and Sunde, N. (1986) Non-cholinergic afferents determine the distribution of the cholinergic septohippocampal projection: a study of the AChE staining pattern in the rat fascia dentata and hippocampus after lesions, x-irradiation, and intracerebral grafting. Exp. Brain Res., 64: 158–168. Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989) Entorhinal fibers form synaptic contacts on parvalbuminimmunoreactive neurons in the rat fascia dentata. Brain Res., 495: 161–166.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 28

Adult neurogenesis in the intact and epileptic dentate gyrus Jack M. Parent Department of Neurology, University of Michigan Medical Center, 109 Zina Pitcher Place, 5021 BSRB, Ann Arbor, MI 48109-2200, USA

Abstract: Neurogenesis persists throughout life in the adult mammalian dentate gyrus. Adult-born dentate granule cells integrate into existing hippocampal circuitry and may provide network plasticity necessary for certain forms of hippocampus-dependent learning and memory. Neural stem cells and neurogenesis in the adult dentate gyrus are regulated by a variety of environmental, physiological, and molecular factors. These include aging, stress, exercise, neurovascular components of the stem cell niche, growth factors, neurotransmitters, and hormones. Seizure activity also influences dentate granule cell neurogenesis. Production of adult-born neurons increases in rodent models of temporal lobe epilepsy, and both newborn and pre-existing granule neurons contribute to aberrant axonal reorganization in the epileptic hippocampus. Prolonged seizures also disrupt the migration of dentate granule cell progenitors and lead to hilar-ectopic granule cells. The ectopic granule neurons appear to integrate abnormally and contribute to network hyperexcitability. Similar findings of granule cell layer dispersion and ectopic granule neurons in human TLE suggest that aberrant neurogenesis contributes to epileptogenesis or learning and memory disturbances in this epilepsy syndrome. Keywords: neurogenesis; stem cell; epileptogenesis; temporal lobe epilepsy; migration; dentate granule cell (DGCs). Persistent DGC neurogenesis occurs in all mammalian species examined to date, including human and nonhuman primates (Eriksson et al., 1998; Gould et al., 1998; Kornack and Rakic, 1999). Although the majority of DGCs in the rat are produced near the end of the first postnatal week, new DGCs continue to be generated at a lower rate throughout adulthood and into senescence (Schlessinger et al., 1975; Kuhn et al., 1996) (Fig. 1). The human dentate gyrus, for example, generates neurons well into the seventh decade of life (Eriksson et al., 1998; Kempermann et al., 2004). Estimates suggest that approximately 9000 DGCs are added daily to the dentate gyrus of young adult rats, and 6% of the granule cell

Dentate granule cell neurogenesis persists throughout life New neurons are generated throughout life in the mammalian dentate gyrus (Altman and Das, 1965; Kaplan and Hinds, 1977; Cameron et al., 1993; Kuhn et al., 1996). This process, termed neurogenesis, begins with neural stem or progenitor cells that give rise to transit amplifying precursors to expand the numbers of progeny. In the dentate gyrus, these progeny form dentate granule cells

Corresponding author. Tel.: +1 734 763 3776; Fax: +1 734 763 7686; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63028-3

529

530

Fig. 1. Neonatal and adult dentate granule cell neurogenesis. (A, B) Bromodeoxyuridine (BrdU) immunostaining of coronal sections through the dentate gyrus of 10 day (A) and 10 week (B) old rats shows proliferating granule cell progenitors concentrated at the inner part of the granule cell layer at the hilar (h) border. Dentate granule cell neurogenesis peaks at the end of the 1st postnatal week (A) but persists into adulthood (B). BrdU was injected 6 days earlier for both. (C) Retroviral nuclear localization signal-b-galactosidase reporter labeling of proliferating cell clusters in the dentate subgranular zone (arrows) of an adult rat. Inset shows the right cluster at higher magnification. Retrovirus was injected 3 days earlier. (D) Doublecortin immunolabeling of immature neurons (arrows) in the adult rat, with cell bodies close to the SGZ but processes throughout the granule cell and molecular layer gcl, granule cell layers.

layer is thought to consist of recently born neurons (Cameron and McKay, 2001). Dentate gyrus neural stem cells reside in the subgranular zone (SGZ) at the border of the hilus and DGC layer. The stem cells are a subtype of radial glia-like astrocyte that expresses glial fibrillary acidic protein (GFAP) and nestin (Seri et al., 2001; Filippov et al., 2003). They proliferate and give rise to clusters of transit amplifying precursors that divide further to generate neuroblasts (Fig. 1). The SGZ precursors and neuroblasts are characterized by a sequence of marker expression as they mature (reviewed in Kempermann et al., 2004). The progenitors express doublecortin (DCX) and polysialylated neural cell adhesion molecule (PSA-NCAM) during cell division and as early postmitotic neurons (Fig. 1D). With further differentiation, postmitotic neurons continue to expresss DCX and PSA-NCAM, but also begin to express calretinin and Prox1; subsequently, they differentiate further into mature granule neurons that are immunoreactive for Prox1 and calbindin.

Integration and function of adult-born DGCs Increasing attention has focused on the potential for adult-generated DGCs to integrate into existing neural circuitry. Combined retrograde tracer and mitotic labeling studies using tritiated thymidine or its analogue, bromodeoxyuridine (BrdU), in adult rodent indicate that mossy fibers of newborn DGCs project to appropriate targets in hippocampal area CA3 (Stanfield and Trice, 1988; Hastings and Gould, 1999; Markakis and Gage, 1999). Acute hippocampal slice recordings from rodents in which adult-generated DGCs were labeled with retroviral reporters or by transgenic manipulations show that the cells integrate into hippocampal circuitry and acquire some electrophysiological characteristics of mature DGCs (Song et al., 2002b; van Praag et al., 2002; Wadiche et al., 2005; Ge et al., 2006). The newborn neurons appear to receive gamma-aminobutyric acid (GABA)ergic synaptic inputs at 1 week after birth and glutamatergic inputs within 2 weeks.

531

Adult-generated DGCs initially show depolarizing responses to GABA (Tozuka et al., 2005; Wadiche et al., 2005; Ge et al., 2006), similar to immature neurons during early development. Unlike those in the neonate, however, adult-generated DGCs take much longer (perhaps 3–6 months) to obtain a full dendritic arborization pattern (van Praag et al., 2002). Compared to neighboring mature DGCs, immature DGCs in the adult have a higher input resistance and a subthreshold calcium conductance that may enhance their excitability. In vitro electrophysiological studies also show a decreased threshold for long-term potentiation (LTP) or long-term depression, as well as higher-amplitude LTP, in newborn DGCs compared to their mature counterparts (Wang et al., 2000; Schmidt-Hieber et al., 2004). The precise function of DGCs generated in adulthood is unknown. Most of the experimental evidence to date, however, supports a role for DGC neurogenesis in learning and memory function (see Doetsch and Hen, 2005 for review). Stimulation of adult DGC neurogenesis with behavioral interventions such as exercise or environmental enrichment is associated with better performance on certain hippocampal learning tasks (Bruel-Jungerman et al., 2005; van Praag et al., 2005; Meshi et al., 2006), although the link between increased neurogeneis and enrichmentenhanced performance has recently been called into question (Meshi et al., 2006). The degree of adult neurogenesis in some inbred mouse strains also correlates positively with learning ability (Kempermann and Gage, 2002). More direct evidence that neurogenesis supports hippocampal-dependent learning and memory derives from experiments in which the depletion of adult-generated neurons disrupts certain forms of learning (Shors et al., 2001, 2002; Winocur et al., 2006). These data are supported by the findings described above that suggest LTP is more easily elicited and of higher amplitude in immature, adult-born DGCs (Wang et al., 2000; SchmidtHieber et al., 2004). Recent animal studies also provide evidence that DGC neurogenesis in the adult is necessary for the positive effects of antidepressant medication (Santarelli et al., 2003). Despite these accumulating data, knowledge of the

entire spectrum of biological functions involving adult DGC neurogenesis awaits more specific means of manipulating DGC progenitors in the adult.

Regulation of adult DGC neurogenesis A host of physiological states and molecular factors appear to regulate adult hippocampal neurogenesis (see Table 1). Neuronal production declines markedly with age (Seki and Arai, 1993; Kuhn et al., 1996), probably due to reduced DGC precursor proliferation and altered differentiation or delayed maturation of their progeny (Kempermann et al., 1998; Rao et al., 2005). The mechanisms likely reflect active suppression of neurogenesis with age because removal of corticosteroids by adrenalectomy or the infusion of specific growth factors significantly restores DGC production in aged rats (Cameron and McKay, 1999; Lichtenwalner et al., 2001; Jin et al., 2003). In addition to age, acute or chronic stress decreases DGC neurogenesis in young adult rodents and primates (Mirescu and Gould, 2006). Adult neurogenesis is stimulated by environmental enrichment, which increases the survival and neuronal differentiation of DGCs generated in early adulthood through senescence (Kempermann et al., 1997, 1998, 2002; Nilsson et al., 1999). Exercise also enhances adult neurogenesis, but this effect relates to increased precursor proliferation rather than survival (van Praag et al., 1999). Growth factors may play a significant role in mediating these effects, as blocking brain uptake of insulin-like growth factor-1 (IGF-1) or vascular endothelial growth factor (VEGF) prevents the increase in DGC production induced by enrichment or exercise (Trejo et al., 2001; Fabel et al., 2003; Cao et al., 2004). Administration of these growth factors also increases adult neurogenesis in the rodent dentate gyrus (Aberg et al., 2000; Jin et al., 2002; Cao et al., 2004; Schanzer et al., 2004), but it is unclear whether this effect results from increased progenitor cell proliferation (Jin et al., 2002) or survival (Schanzer et al., 2004). Other mediators of adult hippocampal neurogenesis include hormones and neurotransmitters.

532 Table 1. Modulation of adult hippocampal neurogenesis Variable

Physiologic state Aging

Stress Exercise

Change in neurogenesis

Mechanism

Change in neurogenesis

Decreased

Proliferation/ neuronal differentiation Proliferation Proliferation/ ?survival/ ?differentiation Survival

Decreased

m

m

Decreased Increased

m m

m?

Decreased Increased

Environmental Increased enrichment Growth factors bFGF Increased (early postnatal) BDNF Increased VEGF

Increased

IGF-1

Increased

Proliferation Differentiation/ survival Proliferation/ survival Proliferation/ survival/ ?differentiation

Neurotransmitters/neuromodulators Glutamate Decreased Proliferation GABA Increased Differentiation/ ?proliferation Serotonin Increased Proliferation Nitrous oxide Decreased Decreased proliferation (promotes differentiation) Transcription factors Wnt Increased Proliferation/ ?differentiation Sonic Increased Proliferation hedgehog Sox2 Increased ?Proliferation/ ?differentiation

Proliferation

Neuronal differentiation

m m

Increased

m

Increased

k

m

Decreased Increased

m m?

m

Increased Decreased

m k

m

Increased

m

m?

Increased

m

Adrenalectomy and corticosteroid replacement studies have shown that circulating stress hormones suppress the rate of DGC production in adulthood (reviewed in Mirescu and Gould, 2006). In contrast, cell proliferation in the rat dentate gyrus increases during proestrus, and the effect is abolished by ovariectomy (Tanapat et al., 1999). Glutamatergic inputs also appear to influence dentate gyrus neurogenesis in the adult rodent. Activation of N-methyl-D-aspartate (NMDA) receptors attenuates cell proliferation, and suppressing glutamatergic neurotransmission with NMDA

m? m

Increased

Increased (postnatal) Increased

Survival

m?

m m

m?

receptor antagonists or lesion of DGC afferents exerts the opposite effect (Cameron et al., 1995). Serotonin (5-hydroxytryptamine) also influences adult neurogenesis. Chronic treatment with antidepressant drugs that act as serotonin reuptake inhibitors increases cell proliferation in the rodent dentate gyrus (Malberg et al., 2000), and blockade of serotonin synthesis or administration of a serotonin neurotoxin decreases new cell production (Brezun and Daszuta, 1999). In terms of other neuromodulators, the number of newly generated DGCs is negatively regulated by nitric oxide,

533

which reduces DGC precursor proliferation but promotes neuronal differentiation (Packer et al., 2003). Neuronal addition thus continues in adulthood under the influence of a balance between multiple positive and negative regulators of neurogenic activity. The local microenvironment in regions of ongoing neurogenesis in the adult mammalian brain provides what is referred to as a ‘‘stem cell niche.’’ This niche governs neuronal production and includes progenitors, astrocytes, microvasculature, and microglia (Wurmser et al., 2004). Astrocytes appear to influence all phases of adult neurogenesis, including neural precursor cell proliferation, migration, and differentiation (Lim and AlvarezBuylla, 1999; Song et al., 2002a). Studies of neurogenesis in the adult rodent dentate gyrus and subventricular zone (SVZ) implicate radial glialike astrocytes as the neural stem cells from which neuroblasts derive (Doetsch et al., 1999; Seri et al., 2001; Garcia et al., 2004; Merkle et al., 2004). A vascular presence in the stem cell niche likely modulates neurogenesis as precursors in the dentate gyrus and SVZ reside in close proximity to the microvasculature (Palmer et al., 2000; Shen et al., 2004), and their proliferative activity appears to be tightly linked with angiogenesis. In the adult, suppression of hippocampal neurogenesis by irradiation may reflect disruption of local angiogenesis (Monje et al., 2002), and infusion of VEGF into the adult brain increases the production both of endothelial cells and DGCs (Jin et al., 2002). In contrast to astrocytes and endothelial cells, recent studies suggest that reactive microglia disrupt neurogenesis in the adult dentate gyrus. Microglial reaction and inflammation induced by brain irradiation, seizures, or lipopolysaccharide treatment decreases the survival of adult-generated DGCs, and suppressing microglial activation with minocycline or indomethacin treatment partially restores neurogenesis (Ekdahl et al., 2003; Monje et al., 2003).

Seizure-induced hippocampal neurogenesis The persistence of neural stem cells in the adult mammalian forebrain raises the possibility for

endogenous repair after brain injury or neurodegeneration. Various brain insults in the adult, including mechanical lesions, hypoglycemia, fluid percussion injury, and stroke, stimulate progenitor cell proliferation in the hippocampal dentate gyrus (Gould and Tanapat, 1997; Liu et al., 1998; Jin et al., 2001; Suh et al., 2005). As in the intact brain, many newly generated cells die, but the vast majority of those that survive differentiate into neurons, and typically less than 20% of adult-born cells become glia or remain undifferentiated. Similar findings are seen after seizure-induced injury in temporal lobe epilepsy (TLE) models, which were the first brain injury models used to demonstrate altered hippocampal neurogenesis (Parent et al., 1997). Studies of adult rodent models of limbic epileptogenesis or acute seizures indicate that prolonged seizures potently stimulate DGC neurogenesis (Bengzon et al., 1997; Gray and Sundstrom, 1998; Parent et al., 1997, 1998; Scott et al., 1998). In the adult rodent kainate and pilocarpine models of temporal lobe epilepsy, chemoconvulsant-induced status epilepticus (SE) increases dentate gyrus cell proliferation approximately 5–10-fold after a latent period of at least several days (Parent et al., 1997; Gray and Sundstrom, 1998) (Fig. 2A and B). Approximately 80–90% of the newly generated cells differentiate into DGCs. DGC neurogenesis is also enhanced by kindlinginduced epileptogenesis. Electrical kindling of the amygdala (Parent et al., 1998; Scott et al., 1998), hippocampus (Bengzon et al., 1997) or perforant path (Nakagawa et al., 2000) acutely increases dentate gyrus cell proliferation and neurogenesis. Similar neurogenic effects occur after acute seizures induced by intermittent perforant path or hippocampal stimulation in adult rats (Bengzon et al., 1997; Parent et al., 1997). Even single, discrete seizure-like afterdischarges induced by hippocampal stimulation lead to increased numbers of newly differentiated DGCs (Bengzon et al., 1997). Although more severe seizures enhance neurogenesis to a greater extent, the survival of the newborn DGCs may decrease with increased seizure severity (Mohapel et al., 2004). This effect may relate to the degree of microglial activation as

534

Fig. 2. Increased cell proliferation and altered dentate gyrus neuroblast migration after SE. (A, B) BrdU incorporation (arrows) in adult rat dentate gyrus 7 days after saline (A) or pilocarpine (B) treatment followed by a 2 day post-BrdU survival. Subgranular zone BrdU labeling increases markedly after SE (B) compared to the control (A). (C, D) Doublecortin (DCX) immunoreactivity in the dentate gyrus of a control (Con; C) and an adult rat 14 days after pilocarpine-induced SE (14d; D). Note the increased hilar DCX expression and chains of DCX-positive cells extending into the hilus (arrows in D) after SE. E, PSA-NCAM+ neuroblast chains (red, arrows) alongside GFAP+ hilar astrocytes (green). Dashed lines in C–E denote the granule cell layer (gcl) — hilar (h) border. (See Color Plate 28.2 in color plate section.)

it is reversed in part by minocycline treatment (Ekdahl et al., 2003).

Seizure-induced hippocampal neurogenesis: functional significance In the pilocarpine model of TLE, aberrant mossy fiber sprouting occurs in parallel with the generation of increased numbers of DGCs (Parent et al., 1997). Evidence suggests that developing axons from newborn DGCs contribute to status epilepticus (SE)-induced aberrant mossy fiber reorganization in both hippocampal area CA3 and the dentate inner molecular layer (Parent et al., 1997). Eliminating DGC progenitors by brain irradiation before and after pilocarpine treatment, however, does not suppress mossy fiber sprouting (Parent et al., 1999). Thus, seizure-induced injury appears to induce both pre-existing and adult-born DGCs

to project axons aberrantly in the epileptogenic hippocampal formation. An increasingly recognized abnormality of DGC neurogenesis in experimental TLE is the ectopic location of adult-born granule neurons. The DGC layer in human TLE is often abnormal due to dispersion and the presence of ectopic granule-like neurons in the hilus and inner molecular layer (Houser, 1990; Parent et al., 2006a). An early report of seizure-induced neurogenesis described newly differentiating neurons with granule cell morphology in the dentate hilus and inner molecular layer after pilocarpine-induced SE in adult rats (Parent et al., 1997). Hilar-ectopic DGCs are found in several different experimental TLE models and may persist for many months (Parent et al., 1997, 2006a; Scharfman et al., 2000; Dashtipour et al., 2001). The cells resemble the granule-like neurons identified in surgical specimens from humans with temporal lobe epilepsy, both in terms of their morphology and marker

535

Fig. 3. Schematic showing the effects of prolonged seizures on caudal subventricular zone (SVZ) and dentate gyrus progenitors. In the intact adult rat (top), progenitors located in the infracallosal SVZ (purple) give rise to white matter oligodendrocytes (orange), while those in the dentate gyrus (green) give rise to DGCs in the granule cell layer (gcl). After seizure-induced hilar and pyramidal cell layer (pcl) injury (bottom; jagged arrows), progenitors migrate aberrantly from the caudal SVZ to form glia (a; purple) in the hippocampus proper or from the dentate subgranular zone to form hilar-ectopic DGCs (b). (See Color Plate 28.3 in color plate section.)

expression (Scharfman et al., 2000; Dashtipour et al., 2001; Parent et al., 2006a). Studies of neurogenesis in the pilocarpine TLE model indicate that the hilar-ectopic DGCs migrate aberrantly from the dentate subgranular proliferative zone to the hilus (Parent et al., 2006a) (Fig. 3). Using intracellular recordings in hippocampal slices from epileptic adult rats, Scharfman and colleagues (Scharfman et al., 2000) have shown that the hilar-ectopic granule-like cells are hyperexcitable and fire in abnormal bursts synchronously with CA3 pyramidal cells. Expression of the activation-induced immediate early gene c-fos by hilar-ectopic DGCs during spontaneous limbic seizures also suggests that they participate in epileptic circuitry (Scharfman et al., 2002). In addition, many putatively newborn DGCs located in the hilus or hilar aspect of the DGC layer after

seizures exhibit a much higher percentage of persistent basal dendrites than is normally found (Scharfman et al., 2000; Dashtipour et al., 2001). Work by Dashtipour and colleagues suggests that the basal dendrites of hilar-ectopic DGCs receive increased excitatory input (Dashtipour et al., 2001), suggesting that this structural plasticity may be a mechanism for seizure generation or propagation. Further evidence supporting the epileptogenic nature of hilar-ectopic DGCs comes from the work of Jung and colleagues (Jung et al., 2004). Their group used the antimitotic agent AraC to inhibit DGC neurogenesis after pilocarpine treatment, and found that these rats developed fewer and shorter spontaneous recurrent seizures than controls. Taken together, these data suggest that hilar-ectopic DGCs integrate abnormally, are hyperexcitable, and thus may contribute

536

Brain insult in childhood (e.g., prolonged febrile seizure)

Altered developmental cues

Aberrant migration of DGC progenitors

RECURRENT SEIZURES

IMPAIRED LEARNING/ MEMORY

Ectopically integrated DGCs

Fig. 4. Aberrant neurogenesis hypothesis of epileptogenesis and memory disturbance in mesial temporal lobe epilepsy. Note the potential maladaptive positive feedback loop of recurrent seizures on aberrant neurogenesis.

to seizure generation or propagation. The presence of hilar-ectopic DGCs in hippocampi surgically resected to treat intractable epilepsy (Houser, 1990; Parent et al., 2006a) raises the possibility that abnormalities in dentate gyrus persistent neurogenesis contribute to epileptogenesis or cognitive dysfunction in human TLE (Fig. 4).

Seizure-induced neurogenesis: mechanisms The mechanisms by which seizure activity stimulates neurogenesis or gliogenesis are unknown. Experiments in which proliferating cells were labeled with BrdU prior to seizure induction have shown that epileptic activity stimulates dentate gyrus and caudal SVZ precursors that are actively proliferating prior to injury (Parent et al., 1999, 2006b). Huttman and colleagues used a reporter mouse in which green fluorescent protein selectively labeled nestin-expressing stem-like cells to show that kainate-induced seizures specifically increased the proliferation of the radial glia-like progenitors in dentate gyrus (Huttmann et al., 2003). Seizures may act to increase neurogenesis indirectly through activation of astrocytes, as astrocytes stimulate hippocampal neurogenesis via wnt signaling and perhaps other mechanisms (Song et al., 2002a; Lie et al., 2005). Alternatively, the SE-induced death of some mature DGCs may increase cell turnover in the dentate gyrus via other mechanisms. Cell death is associated with subsequent cell birth in a number of postnatal neurogenic systems, including the dentate gyrus

and olfactory bulb of adult rodents (Gould and McEwen, 1993; Biebl et al., 2000). Seizures also increase the expression of molecules with the potential to increase neurogenesis or gliogenesis such as growth factors (Humpel et al., 1993) and neurotrophins (Isackson et al., 1991). Specific neurotransmitters or neuromodulatory systems that normally influence DGC neurogenesis (see Table 1) also may be altered by seizure activity. In terms of potential mechanisms underlying hilar- or molecular-layer ectopic DGC formation, molecular factors that influence neuronal migration during development are prime candidates. The expression of one of these developmental factors, reelin, persists in the hippocampal dentate gyrus of adult humans and rodents, and has been implicated in DGC layer dispersion in human TLE (Haas et al., 2002). The expression of reelin decreases markedly in the dentate gyrus after pilocarpine-induced SE (JMP, personal communication), suggesting that loss of this migration guidance factor may be responsible for the aberrant migration of DGC progenitors during epileptogenesis. Another potential mechanism is loss of GABA caused by SE-induced depletion of dentate interneurons, as this neurotransmitter influences DGC differentiation (see Table 1) and GABA decreases neuroblast migration in the other adult neurogenic region, the SVZ-olfactory bulb pathway (Liu et al., 2005). In an interesting study, Scharfman and colleagues showed that hippocampal brain-derived neurotrophic factor (BDNF) infusion in adult rat increased DGC neurogenesis and led to the appearance of ectopic DGCs

537

(Scharfman et al., 2005). This result raises the possibility that BDNF is involved in seizure-induced, aberrant DGC neurogenesis as well. The multiple steps of DGC neurogenesis are influenced by a delicate balance of factors affecting DGC progenitors in the adult. Improved understanding of the biological functions of adult-born DGCs, their regulation, and how they are influenced by brain pathology may provide important insights into the pathophysiology of TLE, its associated cognitive impairments and other brain disorders. Manipulation of adult neurogenesis for therapeutic purposes therefore is likely to involve suppression of aberrant integration in certain instances rather than simply stimulation of neurogenesis for neuronal replacement.

Abbreviations BDNF BrdU DCX DGC GABA GFAP IGF-1 LTP mTLE NMDA PSA-NCAM SE SGZ SVZ TLE VEGF

brain-derived neurotrophic factor bromodeoxyuridine doublecortin dentate granule cell gamma-aminobutyric acid glial fibrillary acidic protein insulin-like growth factor-1 long-term potentiation medial temporal lobe epilepsy N-methyl-D-aspartate polysialylated neural cell adhesion molecule status epilepticus subgranular zone subventricular zone temporal lobe epilepsy vascular endothelial growth factor

References Aberg, M.A., Aberg, N.D., Hedbacker, H., Oscarsson, J. and Eriksson, P.S. (2000) Peripheral infusion of IGF-I selectively induces neurogenesis in the adult rat hippocampus. J. Neurosci., 20: 2896–2903.

Altman, J. and Das, G.D. (1965) Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol., 124: 319–335. Bengzon, J., Kokaia, Z., Elmer, E., Nanobashvili, A., Kokaia, M. and Lindvall, O. (1997) Apoptosis and proliferation of dentate gyrus neurons after single and intermittent limbic seizures. Proc. Natl. Acad. Sci. U.S.A., 94: 10432–10437. Biebl, M., Cooper, C.M., Winkler, J. and Kuhn, H.G. (2000) Analysis of neurogenesis and programmed cell death reveals a self-renewing capacity in the adult rat brain. Neurosci. Lett., 291: 17–20. Brezun, J.M. and Daszuta, A. (1999) Depletion in serotonin decreases neurogenesis in the dentate gyrus and the subventricular zone of adult rats. Neuroscience, 89: 999–1002. Bruel-Jungerman, E., Laroche, S. and Rampon, C. (2005) New neurons in the dentate gyrus are involved in the expression of enhanced long-term memory following environmental enrichment. Eur. J. Neurosci., 21: 513–521. Cameron, H.A., McEwen, B.S. and Gould, E. (1995) Regulation of adult neurogenesis by excitatory input and NMDA receptor activation in the dentate gyrus. J. Neurosci., 15: 4687–4692. Cameron, H.A. and McKay, R.D. (1999) Restoring production of hippocampal neurons in old age. Nat. Neurosci., 2: 894–897. Cameron, H.A. and McKay, R.D. (2001) Adult neurogenesis produces a large pool of new granule cells in the dentate gyrus. J. Comp. Neurol., 435: 406–417. Cameron, H.A., Woolley, C.S., McEwen, B.S. and Gould, E. (1993) Differentiation of newly born neurons and glia in the dentate gyrus of the adult rat. Neuroscience, 56: 337–344. Cao, L., Jiao, X., Zuzga, D.S., Liu, Y., Fong, D.M., Young, D. and During, M.J. (2004) VEGF links hippocampal activity with neurogenesis, learning and memory. Nat. Genet., 36: 827–835. Dashtipour, K., Tran, P.H., Okazaki, M.M., Nadler, J.V. and Ribak, C.E. (2001) Ultrastructural features and synaptic connections of hilar ectopic granule cells in the rat dentate gyrus are different from those of granule cells in the granule cell layer. Brain Res., 890: 261–271. Doetsch, F., Caille, I., Lim, D.A., Garcia-Verdugo, J.M. and Alvarez-Buylla, A. (1999) Subventricular zone astrocytes are neural stem cells in the adult mammalian brain. Cell, 97: 703–716. Doetsch, F. and Hen, R. (2005) Young and excitable: the function of new neurons in the adult mammalian brain. Curr. Opin. Neurobiol., 15: 121–128. Ekdahl, C.T., Claasen, J.H., Bonde, S., Kokaia, Z. and Lindvall, O. (2003) Inflammation is detrimental for neurogenesis in adult brain. Proc. Natl. Acad. Sci. U.S.A., 100: 13632–13637. Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M., Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Neurogenesis in the adult human hippocampus. Nat. Med., 4: 1313–1317. Fabel, K., Fabel, K., Tam, B., Kaufer, D., Baiker, A., Simmons, N., Kuo, C.J. and Palmer, T.D. (2003) VEGF is

538 necessary for exercise-induced adult hippocampal neurogenesis. Eur. J. Neurosci., 18: 2803–2812. Filippov, V., Kronenberg, G., Pivneva, T., Reuter, K., Steiner, B., Wang, L.P., Yamaguchi, M., Kettenmann, H. and Kempermann, G. (2003) Subpopulation of nestin-expressing progenitor cells in the adult murine hippocampus shows electrophysiological and morphological characteristics of astrocytes. Mol. Cell Neurosci., 23: 373–382. Garcia, A.D., Doan, N.B., Imura, T., Bush, T.G. and Sofroniew, M.V. (2004) GFAP-expressing progenitors are the principal source of constitutive neurogenesis in adult mouse forebrain. Nat. Neurosci., 7: 1233–1241. Ge, S., Goh, E.L., Sailor, K.A., Kitabatake, Y., Ming, G.L. and Song, H. (2006) GABA regulates synaptic integration of newly generated neurons in the adult brain. Nature, 439: 589–593. Gould, E. and McEwen, B.S. (1993) Neuronal birth and death. Curr. Opin. Neurobiol., 3: 676–682. Gould, E. and Tanapat, P. (1997) Lesion-induced proliferation of neuronal progenitors in the dentate gyrus of the adult rat. Neuroscience, 80: 427–436. Gould, E., Tanapat, P., McEwen, B.S., Flugge, G. and Fuchs, E. (1998) Proliferation of granule cell precursors in the dentate gyrus of adult monkeys is diminished by stress. Proc. Natl. Acad. Sci. U.S.A., 95: 3168–3171. Gray, W.P. and Sundstrom, L.E. (1998) Kainic acid increases the proliferation of granule cell progenitors in the dentate gyrus of the adult rat. Brain Res., 790: 52–59. Haas, C.A., Dudeck, O., Kirsch, M., Huszka, C., Kann, G., Pollak, S., Zentner, J. and Frotscher, M. (2002) Role for reelin in the development of granule cell dispersion in temporal lobe epilepsy. J. Neurosci., 22: 5797–5802. Hastings, N.B. and Gould, E. (1999) Rapid extension of axons into the CA3 region by adult-generated granule cells. J. Comp. Neurol., 413: 146–154. Houser, C.R. (1990) Granule cell dispersion in the dentate gyrus of humans with temporal lobe epilepsy. Brain Res., 535: 195–204. Humpel, C., Lippoldt, A., Chadi, G., Ganten, D., Olson, L. and Fuxe, K. (1993) Fast and widespread increase of basic fibroblast growth factor messenger RNA and protein in the forebrain after kainate-induced seizures. Neuroscience, 57: 913–922. Huttmann, K., Sadgrove, M., Wallraff, A., Hinterkeuser, S., Kirchhoff, F., Steinhauser, C. and Gray, W.P. (2003) Seizures preferentially stimulate proliferation of radial glia-like astrocytes in the adult dentate gyrus: functional and immunocytochemical analysis. Eur. J. Neurosci., 18: 2769–2778. Isackson, P.J., Huntsman, M.M., Murray, K.D. and Gall, C.M. (1991) BDNF mRNA expression is increased in adult rat forebrain after limbic seizures: temporal patterns of induction distinct from NGF. Neuron, 6: 937–948. Jin, K., Minami, M., Lan, J.Q., Mao, X.O., Batteur, S., Simon, R.P. and Greenberg, D.A. (2001) Neurogenesis in dentate subgranular zone and rostral subventricular zone after focal cerebral ischemia in the rat. Proc. Natl. Acad. Sci. U.S.A., 98: 4710–4715.

Jin, K., Sun, Y., Xie, L., Batteur, S., Mao, X.O., Smelick, C., Logvinova, A. and Greenberg, D.A. (2003) Neurogenesis and aging: FGF-2 and HB-EGF restore neurogenesis in hippocampus and subventricular zone of aged mice. Aging Cell, 2: 175–183. Jin, K., Zhu, Y., Sun, Y., Mao, X.O., Xie, L. and Greenberg, D.A. (2002) Vascular endothelial growth factor (VEGF) stimulates neurogenesis in vitro and in vivo. Proc. Natl. Acad. Sci. U.S.A., 99: 11946–11950. Jung, K.H., Chu, K., Kim, M., Jeong, S.W., Song, Y.M., Lee, S.T., Kim, J.Y., Lee, S.K. and Roh, J.K. (2004) Continuous cytosine-b-D-arabinofuranoside infusion reduces ectopic granule cells in adult rat hippocampus with attenuation of spontaneous recurrent seizures following pilocarpine-induced status epilepticus. Eur. J. Neurosci., 19: 3219–3226. Kaplan, M.S. and Hinds, J.W. (1977) Neurogenesis in the adult rat: electron microscopic analysis of light radioautographs. Science, 197: 1092–1094. Kempermann, G. and Gage, F.H. (2002) Genetic determinants of adult hippocampal neurogenesis correlate with acquisition, but not probe trial performance, in the water maze task. Eur. J. Neurosci., 16: 129–136. Kempermann, G., Gast, D. and Gage, F.H. (2002) Neuroplasticity in old age: sustained fivefold induction of hippocampal neurogenesis by long-term environmental enrichment. Ann. Neurol., 52: 135–143. Kempermann, G., Jessberger, S., Steiner, B. and Kronenberg, G. (2004) Milestones of neuronal development in the adult hippocampus. Trends Neurosci., 27: 447–452. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) More hippocampal neurons in adult mice living in an enriched environment. Nature, 386: 493–495. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Experience-induced neurogenesis in the senescent dentate gyrus. J. Neurosci., 18: 3206–3212. Kornack, D.R. and Rakic, P. (1999) Continuation of neurogenesis in the hippocampus of the adult macaque monkey. Proc. Natl. Acad. Sci. U.S.A., 96: 5768–5773. Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Neurogenesis in the dentate gyrus of the adult rat: age-related decrease of neuronal progenitor proliferation. J. Neurosci., 16: 2027–2033. Lichtenwalner, R.J., Forbes, M.E., Bennett, S.A., Lynch, C.D., Sonntag, W.E. and Riddle, D.R. (2001) Intracerebroventricular infusion of insulin-like growth factor-I ameliorates the age-related decline in hippocampal neurogenesis. Neuroscience, 107: 603–613. Lie, D.C., Colamarino, S.A., Song, H.J., Desire, L., Mira, H., Consiglio, A., Lein, E.S., Jessberger, S., Lansford, H., Dearie, A.R. and Gage, F.H. (2005) Wnt signalling regulates adult hippocampal neurogenesis. Nature, 437: 1370–1375. Lim, D.A. and Alvarez-Buylla, A. (1999) Interaction between astrocytes and adult subventricular zone precursors stimulates neurogenesis. Proc. Natl. Acad. Sci. U.S.A., 96: 7526–7531.

539 Liu, J., Solway, K., Messing, R.O. and Sharp, F.R. (1998) Increased neurogenesis in the dentate gyrus after transient global ischemia in gerbils. J. Neurosci., 18: 7768–7778. Liu, X., Wang, Q., Haydar, T.F. and Bordey, A. (2005) Nonsynaptic GABA signaling in postnatal subventricular zone controls proliferation of GFAP-expressing progenitors. Nat. Neurosci., 8: 1179–1187. Malberg, J.E., Eisch, A.J., Nestler, E.J. and Duman, R.S. (2000) Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. J. Neurosci., 20: 9104–9110. Markakis, E.A. and Gage, F.H. (1999) Adult-generated neurons in the dentate gyrus send axonal projections to field CA3 and are surrounded by synaptic vesicles. J. Comp. Neurol., 406: 449–460. Merkle, F.T., Tramontin, A.D., Garcia-Verdugo, J.M. and Alvarez-Buylla, A. (2004) Radial glia give rise to adult neural stem cells in the subventricular zone. Proc. Natl. Acad. Sci. U.S.A., 101: 17528–17532. Meshi, D., Drew, M.R., Saxe, M., Ansorge, M.S., David, D., Santarelli, L., Malapani, C., Moore, H. and Hen, R. (2006) Hippocampal neurogenesis is not required for behavioral effects of environmental enrichment. Nat. Neurosci., 9: 729–731. Mirescu, C. and Gould, E. (2006) Stress and adult neurogenesis. Hippocampus, 16: 233–238. Mohapel, P., Ekdahl, C.T. and Lindvall, O. (2004) Status epilepticus severity influences the long-term outcome of neurogenesis in the adult dentate gyrus. Neurobiol. Dis., 15: 196–205. Monje, M.L., Mizumatsu, S., Fike, J.R. and Palmer, T.D. (2002) Irradiation induces neural precursor-cell dysfunction. Nat. Med., 8: 955–962. Monje, M.L., Toda, H. and Palmer, T.D. (2003) Inflammatory blockade restores adult hippocampal neurogenesis. Science, 302: 1760–1765. Nakagawa, E., Aimi, Y., Yasuhara, O., Tooyama, I., Shimada, M., McGeer, P.L. and Kimura, H. (2000) Enhancement of progenitor cell division in the dentate gyrus triggered by initial limbic seizures in rat models of epilepsy. Epilepsia, 41: 10–18. Nilsson, M., Perfilieva, E., Johansson, U., Orwar, O. and Eriksson, P.S. (1999) Enriched environment increases neurogenesis in the adult rat dentate gyrus and improves spatial memory. J. Neurobiol., 39: 569–578. Packer, M.A., Stasiv, Y., Benraiss, A., Chmielnicki, E., Grinberg, A., Westphal, H., Goldman, S.A. and Enikolopov, G. (2003) Nitric oxide negatively regulates mammalian adult neurogenesis. Proc. Natl. Acad. Sci. U.S.A., 100: 9566–9571. Palmer, T.D., Willhoite, A.R. and Gage, F.H. (2000) Vascular niche for adult hippocampal neurogenesis. J. Comp. Neurol., 425: 479–494. Parent, J.M., von dem Bussche, N. and Lowenstein, D.H. (2006b) Prolonged seizures recruit caudal subventricular zone glial progenitors into the injured hippocampus. Hippocampus, 16: 321–328. Parent, J.M., Elliott, R.C., Pleasure, S.J., Barbaro, N.M. and Lowenstein, D.H. (2006a) Aberrant seizure-induced

neurogenesis in experimental temporal lobe epilepsy. Ann. Neurol., 59: 81–91. Parent, J.M., Janumpalli, S., McNamara, J.O. and Lowenstein, D.H. (1998) Increased dentate granule cell neurogenesis following amygdala kindling in the adult rat. Neurosci. Lett., 247: 9–12. Parent, J.M., Tada, E., Fike, J.R. and Lowenstein, D.H. (1999) Inhibition of dentate granule cell neurogenesis with brain irradiation does not prevent seizure-induced mossy fiber synaptic reorganization in the rat. J. Neurosci., 19: 4508–4519. Parent, J.M., Yu, T.W., Leibowitz, R.T., Geschwind, D.H., Sloviter, R.S. and Lowenstein, D.H. (1997) Dentate granule cell neurogenesis is increased by seizures and contributes to aberrant network reorganization in the adult rat hippocampus. J. Neurosci., 17: 3727–3738. van Praag, H., Kempermann, G. and Gage, F.H. (1999) Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nat. Neurosci., 2: 266–270. van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer, T.D. and Gage, F.H. (2002) Functional neurogenesis in the adult hippocampus. Nature, 415: 1030–1034. van Praag, H., Shubert, T., Zhao, C. and Gage, F.H. (2005) Exercise enhances learning and hippocampal neurogenesis in aged mice. J. Neurosci., 25: 8680–8685. Rao, M.S., Hattiangady, B., Abdel-Rahman, A., Stanley, D.P. and Shetty, A.K. (2005) Newly born cells in the ageing dentate gyrus display normal migration, survival and neuronal fate choice but endure retarded early maturation. Eur. J. Neurosci., 21: 464–476. Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F., Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O., Belzung, C. and Hen, R. (2003) Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science, 301: 805–809. Schanzer, A., Wachs, F.P., Wilhelm, D., Acker, T., CooperKuhn, C., Beck, H., Winkler, J., Aigner, L., Plate, K.H. and Kuhn, H.G. (2004) Direct stimulation of adult neural stem cells in vitro and neurogenesis in vivo by vascular endothelial growth factor. Brain Pathol., 14: 237–248. Scharfman, H., Goodman, J., Macleod, A., Phani, S., Antonelli, C. and Croll, S. (2005) Increased neurogenesis and the ectopic granule cells after intrahippocampal BDNF infusion in adult rats. Exp. Neurol., 192: 348–356. Scharfman, H.E., Goodman, J.H. and Sollas, A.L. (2000) Granule-like neurons at the hilar/CA3 border after status epilepticus and their synchrony with area CA3 pyramidal cells: functional implications of seizure-induced neurogenesis. J. Neurosci., 20: 6144–6158. Scharfman, H.E., Sollas, A.L. and Goodman, J.H. (2002) Spontaneous recurrent seizures after pilocarpine-induced status epilepticus activate calbindin-immunoreactive hilar cells of the rat dentate gyrus. Neuroscience, 111: 71–81. Schlessinger, A.R., Cowan, W.M. and Gottlieb, D.I. (1975) An autoradiographic study of the time of origin and the pattern of granule cell migration in the dentate gyrus of the rat. J. Comp. Neurol., 159: 149–175.

540 Schmidt-Hieber, C., Jonas, P. and Bischofberger, J. (2004) Enhanced synaptic plasticity in newly generated granule cells of the adult hippocampus. Nature, 429: 184–187. Scott, B.W., Wang, S., Burnham, W.M., De Boni, U. and Wojtowicz, J.M. (1998) Kindling-induced neurogenesis in the dentate gyrus of the rat. Neurosci. Lett., 248: 73–76. Seki, T. and Arai, Y. (1993) Highly polysialylated neural cell adhesion molecule (NCAM-H) is expressed by newly generated granule cells in the dentate gyrus of the adult rat. J. Neurosci., 13: 2351–2358. Seri, B., Garcia-Verdugo, J.M., McEwen, B.S. and AlvarezBuylla, A. (2001) Astrocytes give rise to new neurons in the adult mammalian hippocampus. J. Neurosci., 21: 7153–7160. Shen, Q., Goderie, S.K., Jin, L., Karanth, N., Sun, Y., Abramova, N., Vincent, P., Pumiglia, K. and Temple, S. (2004) Endothelial cells stimulate self-renewal and expand neurogenesis of neural stem cells. Science, 304: 1338–1340. Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and Gould, E. (2001) Neurogenesis in the adult is involved in the formation of trace memories. Nature, 410: 372–376. Shors, T.J., Townsend, D.A., Zhao, M., Kozorovitskiy, Y. and Gould, E. (2002) Neurogenesis may relate to some but not all types of hippocampal-dependent learning. Hippocampus, 12: 578–584. Song, H., Stevens, C.F. and Gage, F.H. (2002a) Astroglia induce neurogenesis from adult neural stem cells. Nature, 417: 39–44. Song, H.J., Stevens, C.F. and Gage, F.H. (2002b) Neural stem cells from adult hippocampus develop essential properties of functional CNS neurons. Nat. Neurosci., 5: 438–445.

Stanfield, B.B. and Trice, J.E. (1988) Evidence that granule cells generated in the dentate gyrus of adult rats extend axonal projections. Exp. Brain Res., 72: 399–406. Suh, S.W., Fan, Y., Hong, S.M., Liu, Z., Matsumori, Y., Weinstein, P.R., Swanson, R.A. and Liu, J. (2005) Hypoglycemia induces transient neurogenesis and subsequent progenitor cell loss in the rat hippocampus. Diabetes, 54: 500–509. Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999) Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat. J. Neurosci., 19: 5792–5801. Tozuka, Y., Fukuda, S., Namba, T., Seki, T. and Hisatsune, T. (2005) GABAergic excitation promotes neuronal differentiation in adult hippocampal progenitor cells. Neuron, 47: 803–815. Trejo, J.L., Carro, E. and Torres-Aleman, I. (2001) Circulating insulin-like growth factor I mediates exercise-induced increases in the number of new neurons in the adult hippocampus. J. Neurosci., 21: 1628–1634. Wadiche, L.O., Bromberg, D.A., Bensen, A.L. and Westbrook, G.L. (2005) GABAergic signaling to newborn neurons in dentate gyrus. J. Neurophysiol., 94: 4528–4532. Wang, S., Scott, B.W. and Wojtowicz, J.M. (2000) Heterogenous properties of dentate granule neurons in the adult rat. J. Neurobiol., 42: 248–257. Winocur, G., Wojtowicz, J.M., Sekeres, M., Snyder, J.S. and Wang, S. (2006) Inhibition of neurogenesis interferes with hippocampus-dependent memory function. Hippocampus, 16: 296–304. Wurmser, A.E., Palmer, T.D. and Gage, F.H. (2004) Neuroscience: cellular interactions in the stem cell niche. Science, 304: 1253–1255.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 29

Unmasking recurrent excitation generated by mossy fiber sprouting in the epileptic dentate gyrus: an emergent property of a complex system Thomas P. Sutula1, and F. Edward Dudek2 1

Department of Neurology H6/570 CSC, University of Wisconsin, 600 Highland Avenue, Madison, WI 53792, USA 2 Department of Physiology, University of Utah, Salt Lake City, UT, USA

Abstract: Seizure-induced sprouting of the mossy fiber pathway in the dentate gyrus has been observed nearly universally in experimental models of limbic epilepsy and in the epileptic human hippocampus. The observation of progressive mossy fiber sprouting induced by kindling demonstrated that even a few repeated seizures are sufficient to alter synaptic connectivity and circuit organization. As it is now recognized that seizures induce synaptic reorganization in hippocampal and cortical pathways, the implications of seizure-induced synaptic reorganization for circuit properties and function have been subjects of intense interest. Detailed anatomical characterization of the sprouted mossy fiber pathway has revealed that the overwhelming majority of sprouted synapses in the inner molecular layer of the dentate gyrus form recurrent excitatory connections, and are thus likely to contribute to recurrent excitation and potentially to enhanced susceptibility to seizures. Nevertheless, difficulties in detecting functional abnormalities in circuits reorganized by mossy fiber sprouting and the fact that some sprouted axons appear to form synapses with inhibitory interneurons have been cited as evidence that sprouting may not contribute to seizure susceptibility, but could form recurrent inhibitory circuits and be a compensatory response to prevent seizures. Quantitative analysis of the synaptic connections of the sprouted mossy fiber pathway, assessment of the functional features of sprouted circuitry using reliable physiological measures, and the perspective of complex systems analysis of neural circuits strongly support the view that the functional effects of the recurrent excitatory circuits formed by mossy fiber sprouting after seizures or injury emerge only conditionally and intermittently, as observed with spontaneous seizures in human epilepsy. The recognition that mossy fiber sprouting is induced after hippocampal injury and seizures and contributes conditionally to emergence of recurrent excitation has provided a conceptual framework for understanding how injury and seizure-induced circuit reorganization may contribute to paroxysmal network synchronization, epileptogenesis, and the consequences of repeated seizures, and thus has had a major influence on understanding of fundamental aspects of the epilepsies. Keywords: hippocampus; seizures; epilepsy; synaptic reorganization; granule cells; plasticity

Corresponding author. Tel.: +1 608 263 5448; Fax: +1 608 263 9405; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63029-5

541

542

Introduction Axon sprouting and growth are important cellular processes in the coordinated sequence of cell birth, differentiation, migration, and neurite outgrowth that culminate in the formation of synapses, circuits, and networks in the developing nervous system. The earliest events in formation of neural circuits are genetically determined, but as development proceeds, activity-dependent sprouting and axon growth play an important role in organizing and refining neural connections (Sur and Rubenstein, 2005). Electrical activity in immature circuits promotes synapse formation and organizes synaptic connectivity into networks supporting adult function, behavior, and adaptation. The observation, nearly 20 years ago, that repeated brief seizures evoked by kindling induced sprouting of the mossy fiber pathway in the dentate gyrus of adult rats demonstrated that the activity-dependent processes of sprouting and axon growth required for circuit formation in the developing nervous system also remodel and reorganize adult pathways after brief seizures (Sutula et al., 1988).

The association of axon sprouting with injury in neural circuits Sprouting in the adult nervous system was first recognized in response to injury and damage. In addition to the mossy fiber pathway, the septohippocampal, associational, and crossed entorhinal pathways within the dentate gyrus and hippocampal formation undergo sprouting and rearrangement of synaptic connectivity in response to a variety of lesions and injuries (Steward et al., 1973, 1974, 1976; Steward, 1976; Steward and Messenheimer, 1978; Nadler et al., 1980a, 1981; Staubli et al., 1984; Steward, 1992). The first suggestion of a link between seizures, sprouting, and circuit reorganization was the observation that sprouted axons of the crossed entorhinal pathway to the dentate gyrus gained access to hippocampal networks modified by kindling (Messenheimer et al., 1979). Evidence of relationships between injury, sprouting, and seizures was provided by studies of kainic acid-induced status epilepticus.

Status epilepticus induced by chemoconvulsants or electrical stimulation induces macroscopic damage in the hilus of the dentate gyrus, subfields of the hippocampus, and a variety of extrahippocampal regions, and is accompanied by reactive sprouting of the mossy fiber axons in the dentate gyrus (Ben-Ari et al., 1979, 1980; Nadler et al., 1980b; Ben-Ari, 1985). Mossy fiber sprouting in these models is a consequence of both direct neurotoxic damage and excitotoxicity induced by status epilepticus (Ben-Ari et al., 1980). The massive, macroscopic damage associated with status epilepticus in these models precluded definitely distinguishing whether axon sprouting and reorganization was a consequence of direct injury, lesion-induced deafferentation, seizures, or a combination of these processes.

Progressive mossy fiber sprouting induced by repeated brief seizures Axon sprouting was definitively identified as a consequence of the recurring brief seizures that define epilepsy by the observation in kindled rats that mossy fiber axons expanded their terminal field to the supragranular region of the dentate gyrus in the absence of overt damage or injury (Sutula et al., 1988). Mossy fiber sprouting, as demonstrated by Timm histochemistry, develops after only a few brief seizures, progresses with repeated seizures, one is permanent (Sutula et al., 1988; Represa et al., 1989a; Cavazos et al., 1991, 1992). Further evidence that mossy fiber sprouting is routinely and reliably induced by seizures was provided by studies indicating that sprouting is induced by seizures propagating into the hippocampus from remote regions (Golarai et al., 1992). Mossy fiber sprouting has been observed in genetic mouse models of epilepsy (Stanfield, 1989; Qiao and Noebels, 1993), and after seizures evoked by maximal electroshock, flurothyl, pentylenetetrazol, intrahippocampal tetanus toxin, and by hyperthermia in immature rats (Golarai et al., 1992; Holmes et al., 1998, 1999; Anderson et al., 1999; Gombos et al., 1999; Vaidya et al., 1999; Bender et al., 2003). Mossy fiber sprouting has been observed in primates after complex partial seizures

543

induced by alumina gel injection (Ribak et al., 1998). Sprouting has been observed in CA3, CA1, and neocortex (Represa and Ben-Ari, 1992; Salin et al., 1995; Esclapez et al., 1999; Lehmann et al., 2001; Smith and Dudek, 2001; Cavazos et al., 2004; Cross and Cavazos, 2007), and in the human epileptic temporal lobe (Represa et al., 1989b; Sutula et al., 1989; Houser et al., 1990; El Bahh et al., 1999). Axonal remodeling may be a continuing process in chronic poorly controlled epilepsy as expression of growth associated proteins B-50 and polysialyated neural cell adhesion molecule (PSANCAM) are observed in resected human epileptic hippocampus (Mikkonen et al., 1998). Because mossy fiber sprouting is progressively induced by repeated brief seizures, poorly controlled seizures in humans may induce progressive sprouting and synaptic reorganization.

Seizure-induced mossy fiber sprouting is accompanied by progressive neuronal loss As sprouting is induced by status epilepticus with extensive injury and by direct damage to hippocampal pathways, and also by kindling which is not accompanied by obvious overt damage, there has been uncertainty about whether seizures can induce sprouting in the absence of neuronal damage. Mossy fiber sprouting has been induced by repetitive stimulation trains that evoke long-term potentiation (Adams et al., 1997), and in association with spatial learning in behavioral tasks such as the Morris water maze, which would not be expected to produce neuronal loss (RamirezAmaya et al., 1999; Holahan et al., 2006). In the case of mossy fiber sprouting induced by seizures, however, it is difficult to conclusively eliminate the possibility that mild neuronal loss may be occurring even with the most sensitive immunohistochemical and stereological techniques for detection of apoptosis and neuronal loss. Studies from multiple laboratories using techniques including unbiased estimates of neuronal numbers, TUNEL staining for apoptotic neurons, and silver impregnation for degenerating cells have demonstrated that single and repeated brief seizures induce neuronal death which contributes to reorganization of

neural circuits (Cavazos and Sutula, 1990; Cavazos et al., 1991, 1994; Bengzon et al., 1997; Pretel et al., 1997; Zhang et al., 1998; Dalby et al., 1998; Haas et al., 2001; Kotloski et al., 2002). A single kindled seizure doubles the rate of apoptosis in the hilar neurons of the dentate gyrus (Bengzon, 1997), and kindling induces apoptosis in CA1, subiculum, and neocortex (Pretel, 1997). Cumulative neuronal loss has been detected after repeated evoked secondary generalized seizures evoked by kindling using stereological counting methods in the CA1, CA3 subfields of the hippocampus and the hilus of the dentate gyrus resembling classical human hippocampal sclerosis (Cavazos and Sutula, 1990; Cavazos et al., 1994; Kotloski et al., 2002). As neuronal loss is progressively induced by kindling and occurs in topographically specific patterns in hippocampal and limbic areas as a function of the site of seizure initiation, mossy fiber sprouting after seizures in the adult nervous system is closely associated with seizure-induced neuronal loss, implying that even a relatively small number of repeated seizures which induce sprouting are potentially accompanied not only by neuronal loss, but by a variety of cellular alterations.

Mossy fiber sprouting is associated with a variety of seizure-induced cellular alterations Repeated seizures not only induce mossy fiber sprouting and neuronal loss, but also a diverse range of cellular and molecular alterations that may potentially modify functional properties of neural circuits. The range of alterations induced in neurons and circuits by seizures spans molecular to network levels. Neurons are born and added to neural circuits in the dentate gyrus after status epilepticus and only a few partial seizures induced by kindling stimulation (Parent et al., 1998), integrate into local circuits, and become activated during recurring seizures (Scharfman et al., 2002). Seizures alter dendritic branching and promote spine loss (Jiang M et al., 1998), and induce growth of basal dendrites by granule cells of the dentate gyrus (Buckmaster and Dudek, 1999; Ribak et al., 2000). Repeated brief seizures

544

evoked by kindling upregulate GFAP (glial fibrillary acidic protein) mRNA and protein levels in a time-dependent manner (Hansen et al., 1990; Bonthius et al., 1995), cause glial cell hypertrophy and proliferation (Khurgel and Ivy, 1996; Stringer, 1996). Kindled seizures modify vesicle release proteins (Matveeva et al., 2006), alter extracellular matrix proteins (Niquet et al., 1995), and modify motor cortex maps (van Rooyen et al., 2006). The association of mossy fiber sprouting with a diverse variety of activity-dependent seizure-induced alterations in hippocampal pathways in both experimental and human epilepsy has implications for assessment of the functional effects of sprouting and these other associated alterations, and poses significant experimental challenges for interpretation of how and to what degree any one of these seizure-induced alterations contribute to functional properties of reorganized neural circuits. Assessing the functional effects of mossy fiber sprouting The functional effects of sprouting are unavoidably confounded by numerous neuronal and glial alterations. Despite this potential complexity, the pattern of alterations of the terminal field of the reorganized sprouted pathway and the synaptic targets of sprouted axons are important determinants of the functional effects of mossy fiber sprouting. Detailed anatomic characterization of the terminal field of the sprouted mossy pathway and quantitative morphological analysis of the synaptic targets of sprouted mossy fiber terminals have generated clear and unambiguous predictions about the possible functional effects of the sprouted pathway in hippocampal circuits. Functional implications of seizure-induced alterations in the terminal field of the mossy fiber pathway Reorganization of terminal fields by sprouting would be expected to have potentially significant functional consequences in neural networks. While mossy fiber sprouting into the supragranular region of the dentate gyrus has received a great

deal of attention because it is easily detected by a variety of methods and is a prominent histological alteration, seizures also induce axon growth in the hilus, promote infrapyramidal to suprapryamidal (interblade) connections not observed in normal rats, and expand the terminal field of the mossy fiber pathway in the inner molecular layer of the dentate gyrus along the septotemporal axis of the hippocampus over distances as long as 700 mm (Sutula et al., 1998; Buckmaster and Dudek, 1999) (Fig. 1). As an example of the potential functional effects of sprouting, reorganization along the septotemporal axis may have specific effects on hippocampal dependent spatial memory, as rightand left-specific place cells are organized in lamellar patterns along this axis (Hampson et al., 1999). Afferent projections to the dentate gyrus are topographically organized in the rat along the septotemporal axis (Dolorfo and Amaral, 1998), and as sprouting in the inner molecular layer varies along this axis as a function of site of seizure induction, might disrupt the normal topographic organization of afferent inputs undergoing processing in hippocampal circuitry, which would be expected to result in functional abnormalities including paradigm specific behavioral and cognitive dysfunction.

Functional implications of seizure-induced alterations in synaptic targets of mossy fibers Formation of synaptic contacts by sprouted axons on principal cells or inhibitory interneurons determines whether seizure-induced reorganization results in new excitatory or inhibitory neural circuits. There is considerable information about the targets of mossy fiber axons in both normal and reorganized circuitry after seizures. The number of synapses formed on principal cells or interneurons determines whether the overall impact of the new circuits is likely to be net excitation or inhibition. While both types of synaptic contacts have been observed in both normal and epileptic hippocampus, the new synapses formed by sprouted axons in the inner molecular layer are predominantly on principal cells and form recurrent excitatory circuits (Franck et al., 1995;

545

Fig. 1. Anatomical features and alterations in terminal fields of sprouted mossy fibers in the reorganized dentate gyrus. (A) Sprouting by a biocytin-filled mossy fiber axon of a granule cell from a rat that received kainic acid in the supragranular region and inner molecular layer and growth of the axon plexus in the hilus with increased numbers of synaptic varicosities. In addition, growth of basal dendrites is well documented, but not shown. (B) Sprouted axons not only extend into the supragranular and inner molecular layer, but cross the hilus from the infrapyramidal blade and thereby increase connectivity between blades of the dentate gyrus. (C) Sprouted mossy fiber axons establish connectivity in the supragranular layer along the septotemporal axis as demonstrated by a biocytin filled granule cell in a saggital section extending from the septal to temporal pole of the hippocampus and dentate gyrus. (D) The sprouted axons of the granule cell in (C) shown at higher magnification establish connectivity in the supragranular layer extending nearly 700 mm along the septotemporal axis. Scale bars (in mm): A, B, C ¼ 50, D ¼ 500. Adapted with permission from Sutula et al., 1998.

Okazaki et al., 1995; Wenzel et al., 2000; Buckmaster et al., 2002; Cavazos et al., 2003) (Fig. 2). Analysis of the synaptic targets of sprouted mossy fibers at the ultrastructural EM level has been facilitated by the ability to identify mossy fiber terminals based on size, morphological features, immunoreactivity to dynophin, and labeling by Timm histochemistry (Zhang and Houser, 1999; Buckmaster et al., 2002; Cavazos et al., 2003). The typical mossy fiber in the normal dentate gyrus forms about 8–10 giant terminals that have asymmetric (excitatory) synaptic contacts with neurons in the hilus and with pyramidal neurons in CA3 (Amaral, 1979; Amaral and Dent, 1981; Claiborne

et al., 1986; Lim et al., 1997; Zhang and Houser, 1999; Gonzales et al., 2001). Although the ‘‘giant mossy fiber terminal’’ is a characteristic distinguishing anatomical feature of mossy fiber axons in the normal dentate gyrus and hippocampus, the more numerous terminals of mossy fibers of the normal dentate gyrus are small and form asymmetric excitatory contacts with mossy cells and with inhibitory interneurons in the hilus (Acsady et al., 1998). About 98% of the small terminals in the hilus are on inhibitory interneurons (Acsady et al., 1998). These observations imply that while the net contribution of the mossy fiber pathway to CA3 is excitatory, collaterals in the hilus may

546

Fig. 2. Examples of synaptic targets of sprouted mossy fibers in the reorganized dentate gyrus. Presynaptic terminals from sprouted mossy fibers form axo-spinous, axo-dendritic, and axo-somatic synapses, and occasionally form synapses with interneurons. (A) A presynaptic terminal of a sprouted mossy fiber axon labeled with silver grains from pre-embedding Timm methods in the inner molecular layer of the dentate gyrus of a kindled rat that experienced 50 class V seizures forms a synapse with a spine head (s) connected to a dendrite (d) in the inner molecular layer. The same terminal also makes contact with another dendrite in cross section. (B) A presynaptic terminal labeled by dynorphin-A immunoreactivity from a sprouted mossy fiber axon forms an asymmetrical synapse with a dendritic spine in the inner molecular layer of a rat that experienced 50 class V kindled seizures. The spine appears perforated and contains a spine apparatus. (C) A dendrite (d) in the inner molecular layer of the dentate gyrus is extensively contacted by synaptic terminals from sprouted mossy fiber axons labeled by preembedding Timm technique in a kindled rat that experienced 50 class V seizures. (D) A dentate basket cell with prominent nuclear infolding receives numerous abnormally large presynaptic terminals (shown in inset) exhibiting the morphological characteristics of mossy fiber boutons (cytoplasm with tightly packed vesicles, mitochondria, and dense core vesicles), which form asymmetrical synapses with the dendrite. (E) A sprouted mossy fiber presynaptic terminal (arrow) labeled by dynorphin-A is apposed to the plasma membrane of the cell bodies of granule cells (gc) in the granule cell layer of the dentate gyrus of a rat that experienced 50 class V kindled seizures. (F) The sprouted synaptic terminal in (E) shown at higher magnification forms an asymmetric postsynaptic density on the granule cell. Scale bars (in mm): A, B, C, E, F ¼ 1; D ¼ 5, inset ¼ 1.25. Adapted with permission from Cavazos et al. (2003, panels A, B, C, E, F), and Sloviter et al. (2006; panel D).

predominantly contribute to local recurrent inhibitory circuits. In addition, mossy fibers in the normal dentate gyrus can form dense innervation patterns along interneurons extending dendrites into the molecular layer (Ribak and Peterson, 1991). These regionally specific anatomical

features of the normal mossy fiber pathway indicate that functional contributions of asymmetric excitatory terminals of mossy fibers to net excitation and inhibition may vary systematically in different regions of hippocampal circuitry, even from the hilus to CA3.

547

It should not be surprising that similar local circuit complexity is encountered in the reorganized dentate gyrus. Except for the occasional interneuron dendrite that receives dense innervation by mossy fibers in the normal dentate gyrus (Ribak and Peterson, 1991), the terminal field of mossy fibers does not project to the inner molecular layer and this region is therefore devoid of recurrent circuits. Seizures alter the terminal field and the synaptic targets of mossy fibers in the inner molecular layer by sprouting and growth of new mossy fiber axons which predominantly contact excitatory principal cells and to a much lesser extent form synapses with interneurons (Frotscher and Zimmer, 1983; Franck et al., 1995; Okazaki et al., 1995; Kotti et al., 1997; Wenzel et al., 2000; Pierce and Milner, 2001; Buckmaster et al., 2002; Cavazos et al., 2003; Sloviter et al., 2006) (Figs. 2 and 3). Quantitative analyses of the relative frequency of synapses have provided unequivocal evidence that the overwhelming number of new synapses are on granule cells and form excitatory circuits. In the most comprehensive study of synaptic targets of sprouted mossy fibers, 96% of sprouted terminals contacted granule cell dendrites in the inner molecular layer and thus formed recurrent excitatory circuits (Buckmaster et al., 2002) (Fig. 3). In another study, the number of asymmetric excitatory axo-spinous synapses characteristic of granule cell synapses exceeded asymmetric axo-dendritic shaft synapses putatively on inhibitory dendrites by nearly 2 to 1 in the reorganized inner molecular layer of kainic acidtreated rats (Cavazos et al., 2003). Perforated axospinous synapses on granule cells are observed in kindled rats (Geinisman et al., 1992; Cavazos et al., 2003), as well as axo-somatic synapses which are not typically observed in normal rats (Cavazos et al., 2003). While some authors have emphasized the very occasional to rare synapses formed by sprouted mossy fibers on interneurons (Kotti et al., 1997; Sloviter et al., 2006), the overwhelming anatomical evidence from the most comprehensive and quantitative analyses clearly indicates that the predominant circuit formed by seizureinduced sprouting in the inner molecular layer is a recurrent excitatory circuit (Coulter, 2002; Sutula, 2002; Dudek and Shao, 2004).

The quantitative anatomical studies indicating that the overwhelming number of new synapses formed by sprouted mossy fibers in the inner molecular layer form new recurrent excitatory circuits are not incompatible with the fact that many mossy fiber terminals in the hilus of normal rats form recurrent inhibitory circuits (Acsady et al., 1998), and that these circuits and potentially new mossy fiber connections to inhibitory interneurons formed in the hilus by sprouted axons may contribute to inhibition in the reorganized dentate gyrus. The important point, however, is that new recurrent excitatory circuits are formed in the inner molecular layer, a region that normally lacks any recurrent excitatory connections among granule cells. It is noteworthy that some studies have reported expression of the 67 kDa isoform of glutamic acid decarboxylase (GAD67, the synthetic enzyme for GABA) in mossy fiber terminals after status epilepticus (Sloviter et al., 1996), suggesting that sprouted asymmetric MF synapses may contribute to inhibition in some circumstances (Gutierrez and Heinemann, 2001), although the overall physiological significance of this alteration remains unclear. So while mossy fibers in the normal and reorganized dentate gyrus form regionally specific and heterogeneous synaptic connections with both CA3 pyramidal neurons and interneurons in the hilus, seizures importantly induce a new and predominantly excitatory circuit among granule cells that does not exist in the normal dentate gyrus.

The paradox of epileptic circuits: episodic dysfunction and synchronization in association with permanent structural reorganization Despite the abundant evidence for mossy fiber sprouting and a variety of chronic structural and molecular alterations in the epileptic hippocampus, it has been surprisingly difficult to detect evidence for functional abnormalities in epileptic circuitry in normal physiological conditions both in vitro and in vivo, including in the resected human epileptic hippocampus (Cronin and Dudek, 1988; Cronin et al., 1992; Golarai and Sutula, 1996; Wuarin and Dudek, 1996, 2001;

548

Fig. 3. Sprouted mossy fibers synapse with GABA-negative and GABA-positive targets identified by EM immunohistochemistry. The predominant circuit formed by sprouted mossy fibers in the supragranular and inner molecular layer is a recurrent excitatory circuit. (A) A biocytin-labeled axon (black) forms synaptic contacts (arrowheads) with two GABA-negative spines. Nearby GABA-positive structures are labeled with 10 nm-diameter colloidal gold particles, and a GABA-positive axon terminal forms a symmetric synapse (arrow) with a granule cell body. (B) A biocytin-labeled axon (black) in the molecular layer forms a synaptic contact (arrowhead) with a GABA-positive dendritic shaft. (C) Sprouted mossy fibers synapse preferentially with GABA-negative dendritic spines. Example of a reconstructed sprouted mossy fiber with synaptic contacts indicated by markers. Squares indicate that the postsynaptic target was a dendritic spine; circles indicate a dendritic shaft. Open markers indicate that the postsynaptic target was GABA-negative; filled markers indicate GABA-positive. Borders between strata (h, hilus; gcl, granule cell layer; ml, molecular layer) are indicated by lines. The vast majority of synapses (93–96% in the granule cell layer and inner molecular layer respectively) form synapses with GABA-negative structures demonstrating that sprouted mossy fibers in the supragranular and inner molecular predominantly form recurrent excitatory circuits. Adapted with permission from Buckmaster et al., 2002.

549

Buckmaster and Dudek, 1997a, b, 1999; Lynch and Sutula, 2000; Sutula, 2002). Membrane properties and evoked responses in resected human epileptic hippocampal slices in standard physiological conditions appear surprisingly normal (Kim et al., 1993; Dudek et al., 1995). Initial analyses of evoked field potentials in the dentate gyrus of hippocampal slices from kainic acid-treated rats and resected hippocampal slices from patients with temporal lobe epilepsy demonstrated an association between the duration and complexity of antidromically evoked extracellular field potentials and the extent of mossy fiber sprouting examined by the Timm method (Tauck and Nadler, 1985; Masukawa et al., 1992), but other studies revealed only minimal evidence for spontaneous epileptic burst discharges in hippocampal slices from kainic acid-treated epileptic rats (Cronin and Dudek, 1988; Cronin et al., 1992; Nissinen et al., 2001). Attempts to correlate the extent of mossy fiber sprouting with spontaneous seizures in chronic in vivo models and human epileptic hippocampus also failed to reveal a direct relationship between sprouting and seizure frequency, and were interpreted as evidence that mossy fiber sprouting does not contribute to abnormal excitation and epilepsy (Longo and Mello, 1998, 1999; Pitkanen et al., 2000; Nissinen et al., 2001). Studies indicating that prevention of mossy fiber sprouting by co-administration of the protein synthesis inhibitor cycloheximide did not prevent development of seizures after pilocarpine (Longo and Mello, 1998, 1999), which were subsequently not replicated (Williams et al., 2002; Toyoda and Buckmaster, 2005), also were cited as evidence that sprouting is not related to epileptogenesis. Other studies noting that seizures could be induced in the absence of sprouting similarly were interpreted as evidence that sprouting is not necessary for seizure induction by kindling (Armitage et al., 1998; Mohapel et al., 2000), which should not be surprising given that previous studies demonstrated that kindled seizures can be evoked from limbic pathways when granule cells in the dentate gyrus have been selectively lesioned by colchicine (Dasheiff and McNamara, 1982; Sutula et al., 1986). Skepticism about the potential importance of mossy fiber sprouting thus developed as multiple studies did

not detect simple direct relationships between sprouting and outcome variables such as spontaneous seizure frequency, and led some to conclude that sprouting and other cellular alterations in the reorganized epileptic hippocampus have negligible functional importance for epileptogenesis (Benardo, 2002). Contrary to this viewpoint, synaptic transmission in the terminal field of the sprouted mossy fiber pathway has been directly demonstrated in vivo by current source density analysis in kindled rats (Golarai and Sutula, 1996) (Fig. 4A–E). An inward current (sink) corresponding spatially to the terminal field of the sprouted mossy fiber terminals in the inner molecular layer of the dentate gyrus develops in kindled rats at a latency consistent with disynaptic transmission in response to perforant path stimulation. The difficulty of detecting functional alterations in chronically reorganized epileptic circuitry despite multiple neuronal and circuit alterations and functional synaptic transmission in the terminal field of the sprouted mossy fiber pathway might appear to be a paradox. This paradox should not be entirely surprising, however, given the clinical fact that patients with epilepsy manifest seizures only sporadically and briefly, so continuous evidence for network synchronization and underlying imbalance of excitation and inhibition should not be expected. Although chronic susceptibility to recurring network synchronization and seizures is the defining feature of epilepsy, seizures are an infrequent emergent event even in intractable patients, and may not be associated with continuously detectable functional abnormalities (Litt et al., 2001; Litt and Echauz, 2002; Worrell et al., 2004). The paradox of infrequent expression of spontaneous functional abnormality despite the presence of permanent structural alterations has implications for design and interpretation of experiments seeking to identify functional abnormalities associated with epileptic circuits reorganized by sprouting and a variety of cellular alterations. The paradox has produced confusion that is in part based on lack of understanding about emergent functional properties in complex systems such as neural circuits.

Fig. 4. Synaptic transmission in the terminal field of the sprouted mossy fiber pathway and unmasking of functional abnormalities in the reorganized dentate gyrus with mossy fiber sprouting. (A) Surface plot of current source density (CSD) as function of depth and time in the dentate gyrus of a rat that experienced three afterdischarges but prior to development of mossy fiber sprouting. Inward currents (sinks) are upward and outward currents (sources) are downward. (B) Corresponding CSD contour plot demonstrates that the inward current sink (solid lines) develops in the middle and outer molecular layer at short latency, and is followed by a lower amplitude sink developing in the inner molecular layer at 16 ms. (C) Surface plot of CSD in the dentate gyrus of a rat with mossy fiber sprouting after 105 class V seizures demonstrates a prominent inward current sink developing in the inner molecular layer at a latency of 9–10 ms (arrow) corresponding to the terminal field of the sprouted mossy fiber pathway. (D) Corresponding CSD contour plot also demonstrates the prominent inward current sink developing in the inner molecular layer at a latency of 9–10 ms (arrow). (E) Schematic summary of the spatial and temporal features of inward current sinks corresponding to the terminal field of sprouted mossy fiber pathway compared to normal control rats without sprouting. Black areas in the schematic dendrites are inward currents of highest amplitude and gray areas are inward currents of lower amplitude. The initial inward current at 3 ms in both the normal and sprouted dentate gyrus is in the distal dendrites (black area), corresponding to monosynaptic transmission in the perforant path. At 9–12 ms in the normal DG, inward current are not yet developed in the proximal dendrites of the inner molecular layer. In the reorganized dentate gyrus with sprouting, a large inward current develops in the proximal dendrites at 9–12 ms, which colocalizes with the laminar distribution of sprouted mossy fiber terminals. This current is consistent with disynaptic transmission in the sprouted pathway in response to perforant path stimulation. (F) Spontaneous epileptic bursts are unmasked in the dentate gyrus of hippocampal slices from rats with mossy fiber sprouting in 10–30 mM bicuculline. The spontaneous bursts consist of prolonged negative shifts in the extracellular field potential and synchronous action potentials. Slices without sprouting had small intracellular depolarizations followed by hyperpolarizations and positive-going extracellular fields; no action potentials or extracellular population spikes were observed. In slices with sprouting, large spontaneous depolarizations that evoked intracellular action potentials could occur synchronously with slow extracellular negative shifts and population spikes. Arrows show expansion of traces. (G) Unmasking of functional abnormalities in the reorganized dentate gyrus with mossy fiber sprouting by altering the extracellular ionic environment. In standard bathing solution (3.5 mM K+), antidromic stimulation evoked epileptiform activity only in the presence of robust recurrent mossy fiber growth. Increasing Timm scores indicate increased density of mossy fiber sprouting. For each value of the Timm score, raising [K+]o increased the percentage of slices that responded with epileptiform activity. In slices with a Timm score of 1, epileptiform activity appeared only when [K+]o was increased. Adapted with permission from Golarai and Sutula (1996; panels A–E), Cronin et al. (1992; panel F), and Hardison et al. (2002; panel G).

551

Episodic paroxysmal abnormalities in epileptic circuitry: the phenomena of unmasking and emergent functional properties in reorganized circuitry Evidence for functional abnormalities in brain slices from experimental models of epilepsy examined by in vitro physiological methods generally emerges only when the reorganized circuitry is perturbed by alterations in the extracellular environment such as elevated [K+]o, or reduced [Ca2+]o, or by disinhibition by GABAA antagonists (Cronin et al., 1992; Wuarin and Dudek, 1996; Patrylo and Dudek, 1998; Molnar and Nadler, 1999; Patrylo et al., 1999; Hardison et al., 2000; Lynch and Sutula, 2000; Wuarin and Dudek 2001; Sutula, 2002) (Fig. 4F and G). Surgically resected human epileptic hippocampus similarly appears normal unless the extracellular environment is altered directly or in response to tetanic stimulation (Masukawa et al., 1989, 1992; Wuarin et al., 1990; Kim et al., 1993; Dudek et al., 1995). These observations suggest that the reorganized circuits are a substrate for dysfunction, but functional abnormality becomes unmasked and emerges only in the context of some other perturbation or transient alterations. These observations in hippocampal slices and the fact that spontaneous seizures occur sporadically and unpredictably in both experimental and human epilepsy indicate that seizures and underlying epileptic functional abnormalities are emergent properties of epileptic neural circuitry altered by primary pathologies or by seizure-induced reorganization.

Unmasking emergent functional abnormalities in recurrent sprouted circuits by modifying GABAergic neurotransmission Functional abnormalities in reorganized epileptic circuits including spontaneous synchronous network discharge become unmasked by reducing GABAergic inhibition. The fact that epileptic network synchronization can emerge even in normal neural circuits when inhibition is reduced indicates the narrow dynamic balance of excitatory and

inhibitory processes, and complicates efforts to directly identify the contributions of new recurrent excitatory circuits in the epileptic dentate gyrus. How and when functional effects emerge at the network and behavioral levels as a consequence of the abnormal excitatory disynaptic activation of granule cell dendrites associated with sprouted terminal field in the inner molecular layer is a critical question. The phenomena of unmasking and dynamic emergent properties must be considered in the design and interpretation of experiments seeking to identify the functional consequences of sprouting and other circuit alterations associated with epilepsy. Emergent functional effects of new recurrent excitatory circuits formed by mossy fiber sprouting will be influenced by the state of inhibition in the dentate gyrus, which must remain sufficiently robust even in intractable epilepsy to prevent runaway excitation and continuous synchronization. Dynamic alterations in inhibition are therefore a key factor in assessing functional effects of sprouting in reorganized epileptic circuity of the dentate gyrus.

The state of inhibition in the reorganized dentate gyrus: a critical variable for emergence of recurrent excitation The state of GABAergic inhibition thus plays an important and indeed critical role in regulating emergence of network synchronization and other functional abnormalities in both the normal and reorganized dentate gyrus. GABAergic neurotransmission is not a unitary or simple process but rather represents a diverse range of potentially complex and dynamic cellular and physiological mechanisms which undergo activity-dependent modification (McCarren and Alger, 1985; Brooks-Kayal et al., 1998; Coulter, 2001; Cossart et al., 2005). In some circumstances, such as in early postnatal development and during acute conditions when intracellular Cl
increases, GABAergic neurotransmission can contribute to excitation as well as cellular inhibition (Staley et al., 1995; Khalilov et al., 1999; Leinekugel et al., 1999; Dzhala and Staley, 2003). Assessing the functional effects of sprouting therefore requires analysis of

552

alterations in both inhibitory and excitatory synaptic transmission and their dynamic interactions in circuits that have undergone reorganization as a result of seizures and primary injury or pathology. Dynamic alterations in excitatory and inhibitory synaptic transmission and seizure-induced formation of new circuits vary significantly in different regions of the hippocampus and independently as a function of time after seizures. The time course of alterations in synaptic transmission evolve over periods of as long as months after the last episode of network synchronization in circuits undergoing remodeling and reorganization, so experiments seeking to understand the emergence of functional abnormalities associated with sprouting also need to evaluate how inhibition and alterations in connectivity systematically vary across these prolonged periods in order to understand functional effects of sprouting and phenomena such as latent periods after initial injuries. The normal dentate gyrus has powerful systems of inhibition as demonstrated by the fact that granule cells are resistant to repetitive discharge in response to synaptic activation, infrequently generate spontaneous activity, and typically do not generate repetitive discharges even when GABAA inhibition is blocked (Fricke and Prince, 1984; Lynch and Sutula, 2000; Lynch et al., 2000). The state of inhibition in the dentate gyrus is predictably modified by episodic network synchronization, specifically as a function of both the timing and duration of seizures. A variety of physiological and pharmacological methods both in vivo and in vitro have demonstrated that repeated brief seizures evoked by kindling increase inhibition in the dentate gyrus (Tuff et al., 1983a, b; de Jonge and Racine, 1987; Stringer and Lothman, 1989; Otis et al., 1994; Buhl et al., 1996; Nusser et al., 1998; Sayin et al., 2003), while sustained seizures as during status epilepticus initially have opposite effects and produce rundown or acute reduction of inhibition (Kapur et al., 1989c; Kapur and Macdonald, 1997), which is later followed by increases in inhibition in granule cells (Cohen et al., 2003). After a few repeated brief seizures evoked during the early stages of kindling, paired pulse inhibition in the dentate gyrus is acutely increased at interpulse intervals of 15–40 ms, which are associated

with the GABA-dependent Cl
mediated conductance change and provide an indirect measure of GABA-dependent inhibition of evoked granule cell discharge under appropriately controlled stimulation and timing parameters (Tuff et al., 1983a, b; de Jonge and Racine, 1987; Stringer and Lothman, 1989; Sayin et al., 2003). The increase in inhibition implied by the indirect measure of paired pulse inhibition has been directly confirmed by demonstration of increased frequency and amplitude of IPSCs in granule cells (Otis et al., 1994; Buhl et al., 1996; Nusser et al., 1998). It is also clear that inhibition after status epilepticus is increased in the dentate gyrus after an initial period of depression (Hellier et al., 1999; Gorter et al., 2002; Harvey and Sloviter, 2005; Sloviter et al., 2006). Although inhibition in the dentate gyrus is initially increased by kindling, inhibition is eventually lost after 100 evoked Class V seizures in association with emergence of spontaneous seizures, as directly confirmed by reduction in amplitude and alterations in kinetics of the monosynaptic IPSC measured by single electrode voltage clamp techniques (Sayin et al., 2003). The loss of inhibition and emergence of spontaneous seizures is associated with reduction of CCK and GAT-1 subpopulations of interneurons, which provide axo-somatic and axo-axonic inhibitory terminals on granule cells. These observations are consistent with the view that recurrent excitatory circuits which are progressively formed by sprouting in kindled rats gradually increase capacity to generate recurrent excitation in the dentate gyrus but are insufficient at early stages of seizureinduced reorganization to generate spontaneous seizures unless inhibition is reduced. With seizureinduced loss of interneurons and critical axosomatic and axo-axonic inhibitory terminals at advanced stages of kindling, recurrent excitation generated by sprouted circuits among granule cells, while still for the most part checked by inhibition, periodically becomes sufficient to drive spontaneous emergent network synchronization. These studies in rats experiencing brief seizures evoked by kindling and status epilepticus indicate that the state of inhibition in the dentate gyrus varies both acutely and chronically in response to

553

ongoing episodes of network synchronization as a function of the duration, frequency, and number of seizures. Efforts to assess the potential functional properties of recurrent circuits formed by sprouting must be considered with recognition of these dynamic alterations in inhibition as a consequence of timing and duration of recent seizures, and pursued with a variety of electrophysiological techniques. In addition to the effects of timing and duration of seizures, it is also clear from in vivo and in vitro studies that the effects of seizures on network inhibition vary significantly as function of location in the hippocampus, and that results in other regions of hippocampal and neocortical circuitry do not necessarily apply to the dentate gyrus. The effects of seizures in the dentate gyrus specifically contrast with CA1, where in vivo methods have demonstrated reduction in inhibition after brief evoked seizures and status epilepticus (Kapur et al., 1989a–c; Michelson et al., 1989; Gorter et al., 2002). In further contrast to the dentate gyrus, GABAA receptor dependent currents in granule cells are increased after status epilepticus induced by pilocarpine but are reduced in pyramidal neurons of CA1 (Gibbs et al., 1997), and dendritic GABAergic inhibition in CA1 as measured by analysis of frequency and amplitude of IPSCs is reduced while somatic inhibition is increased (Cossart et al., 2001). Audiogenic kindling reduces GABAA dependent IPCSs in the inferior colliculus (Evans et al., 2006), and increases amplitude and duration of mIPSCs in piriform cortex (Gavrilovici et al., 2006). These contrasting observations in the dentate gyrus, CA1, and other regions clearly demonstrate that assessment of inhibitory and excitatory processes in both normal and reorganized circuitry cannot be simply characterized by a single technique or in a single model or in a single region of hippocampal circuitry, and that multiple techniques are required to systematically characterize the effects of seizures on these synaptic and circuit properties. This is an important perspective for efforts to assess the effects of sprouting or any other putative cellular causes of seizures and chronic epilepsy.

Direct evidence for recurrent excitation unmasked by disinhibition in the reorganized dentate gyrus The functional abnormalities that become unmasked in the reorganized dentate gyrus by reducing inhibition or altering the extracellular ionic environment include direct evidence for development of recurrent excitation in association with mossy fiber sprouting. Evidence for recurrent excitation in association with mossy fiber sprouting has been demonstrated by focal stimulation by glutamate application using microdrop or flash photolysis techniques (Wuarin and Dudek, 1996, 2001; Molnar and Nadler, 1999; Lynch and Sutula, 2000). Focal application of glutamate microdrops to dendrites and cell bodies of granule cells remote from the recorded granule cell in hippocampal slices from normal rats evokes no responses when inhibition is blocked, but microdrop application in disinhibited slices with sprouting evokes EPSPs at long and variable latency (Wuarin and Dudek, 1996; Lynch and Sutula, 2000). In kindled rats with mossy fiber sprouting, trains of EPSPs and population discharges can be evoked by glutamate microstimulation remote from the recording site at one week after induction of kindled seizures, when sprouting is first detectable by Timm histochemistry, but are not evoked in hippocampal slices from kindled rats examined at 24 h after a single afterdischarge prior to the development of sprouting (Lynch and Sutula, 2000). Stimulation by flash photolysis of caged glutamate at sites remote from the recording site in granule cells also evokes EPSCs in granule cells (Molnar and Nadler, 1999; Wuarin and Dudek, 2001) which increase in correlation with the development of sprouting (Fig. 5A and B). The long and variable latency of responses evoked by focal glutamate stimulation in these studies suggested that recurrent excitation was generated by multisynaptic rather than monosynaptic circuits, but monosynaptic EPSPs can be evoked at short latency (2.670.36 ms) between blades of the dentate gyrus under conditions in which recurrent inhibitory circuits are blocked by bicuculline, polysynaptic activity is suppressed by 10 mM Ca2+ in the bathing medium, and perforant path activation is prevented by knife cuts (Lynch and

554

Fig. 5. Formation of recurrent excitatory circuits in the reorganized dentate gyrus with mossy fiber sprouting. (A) Evidence for development of excitatory connectivity in the reorganized dentate gyrus with mossy fiber sprouting. Photostimulation of caged glutamate in the granule cell layer in conditions of reduced inhibition (bicuculline, 30 mM) and elevated [K+]o (6 mM) which evokes no responses in the normal dentate gyrus evokes epileptiform bursts of action potentials in a granule cell from a rat 33 weeks after kainate treatment. The patch pipette in the diagram indicates the position of the recorded granule cell in the outer blade. The numbers show the locations of the photostimulations in the diagram and the corresponding evoked bursts of action potentials. The granule cell was recorded in the whole cell current-clamp configuration at resting membrane potential. Arrowheads show stimulation artifact produced by the flash. (B) Excitatory connectivity as revealed by flash photolysis increases with time after treatment with kainic acid in association with increasing mossy fiber sprouting. Plot of the percentage of granule cells from saline- and kainate-injected animals responding to photostimulation with an increase in EPSCs as a function of time after treatment. The number of cells tested is indicated for each age group. (C) Monosynaptic excitatory connections between blades of the dentate gyrus in a hippocampal slice from a rat treated with kainic acid. The transected transverse hippocampal slice contained only the infrapyramidal and suprapyramidal blades of the dentate gyrus, the intervening hilus, a small sector of CA3c, and CA1. The transection removed the crest of the dentate gyrus, which severed perforant path connections between the blades. In bathing solution containing 10 mM bicuculline to suppress inhibitory postsynaptic potentials and 10 mM [Ca2+]o to suppress polysynaptic activity, stimulation of the molecular layer of the infrapyramidal blade (site indicated by arrows) with a 100 ms constant-voltage pulse of 7 V evoked an EPSP in a granule cell in the suprapyramidal blade (location indicated by the arrowhead). The latency of the EPSP was 2.7 ms. Focal electrical microstimulation evoked EPSPs in 5 of 18 suprapyramidal granule cells in kainic acid-treated rats with an average latency of 2.5970.36 ms. In contrast, stimulation of the infrapyramidal blade in hippocampal slices from normal rats failed to evoke EPSPs in 15 suprapyramidal granule cells. Calibration bars: 2 mV, 8 ms. (D) Monosynaptic connections directly demonstrated between granule cells in slices from rats with mossy fiber sprouting. Recordings from a pair of simultaneously recorded granule cells are shown in (5D1): presynaptic neuron (top), postsynaptic neuron (bottom). Intracellular current (a 150 ms rectangular current pulse; start and end of the pulse are marked by the dots) triggered an action potential (AP) in the presynaptic cell. Immediately thereafter, a small depolarization occurred in the postsynaptic cell. An arrow marks the capacitative artifact of the presynaptic cell’s AP. Calibration: presynaptic cell, 20 mV, 30 ms; postsynaptic cell, 4 mV, 30 ms. (5D2): recordings from the same pair of neurons with higher gain. Several postsynaptic responses are overlapped to show the variability in the response to the presynaptic AP. Calibration: presynaptic cell, 20 mV, 4 ms; postsynaptic cell, 3 mV, 4 ms. (5D3): in a different pair of granule cells, tonic intracellular current was used to depolarize both the putative presynaptic (top) and postsynaptic (bottom) cells. A spontaneous AP in the presynaptic cell triggered an AP in the second cell. Membrane potentials: top,
55 mV; bottom,
54 mV. Calibration: 15 mV, 25 ms. Adapted with permission from Wuarin and Dudek (2001; panel A), Hardison et al. (2000; panel B), Lynch et al. (2000; panel C), and Scharfman et al. (2003; panel D).

555

Sutula, 2000) (Fig. 5C). These studies from multiple laboratories are evidence that seizures induce recurrent excitatory connectivity in the dentate gyrus which emerges only under conditions of reduced inhibition or alterations in the extracellular ionic environment. This physiological evidence for formation of recurrent excitatory connections has been confirmed by dual recordings from pairs of granule cells in the dentate gyrus reorganized by mossy fiber sprouting (Fig. 5D1–3). Examination of 903 granule cell pairs in hippocampal slices from epileptic pilocarpine-treated rats revealed monosynaptic connections in 1/150 granule cell pairs which were not observed in any of 285 pairs from normal animals (Scharfman et al., 2003). While the number of pairs was small, the results directly confirm the presence of recurrent excitatory circuits in epileptic dentate gyrus when inhibition was intact, and leave open the possibility that reduction in inhibition might reveal additional evidence for recurrent excitatory circuits masked by powerful inhibition present in both the normal and reorganized dentate gyrus.

Dynamic and evolving plasticity in reorganized excitatory and inhibitory circuits of the epileptic dentate gyrus: implications for assessment of the functional effects of sprouting The molecular and cellular processes underlying network inhibition and excitation may be altered by primary pathologies that cause epilepsy. In addition, network inhibition and excitation are systematically altered as a consequence of repeated network synchronization and seizures, i.e. by seizure-induced plasticity or kindling. Inhibition and excitation undergo independent acute and longterm alterations that extend for as long as months after seizures. For example, in addition to the rewiring of synaptic connections by mossy fiber sprouting, enhancement of kainate-receptor and NMDA-receptor dependent glutamatergic transmission also results in alterations in excitatory synaptic properties that may contribute to network synchronization (Sayin et al., 1999; Behr et al., 2001; Epsztein et al., 2005). If the primary

pathological process and the subsequent processes of seizure-induced plasticity following an initial episode of network synchronization are sustained or permanent, recurring seizures that define epilepsy can emerge. The strength and dynamic balance of excitatory and inhibitory transmission, which undergo activity-dependent alterations after seizures in reorganizing circuitry with sprouting, will evolve and vary as a function of time and duration of recent seizures. These potentially complex interactions and the temporal relationships of acute and chronic seizure-induced plasticity support the view that recurrent excitation and network synchronization be considered as emergent circuit properties in different experimental models. As the neurobiological phenomena of plasticity associated with repeated network synchronization and seizures in neural circuits, the gradually evolving time course of circuit alteration induced by kindling has been informative for assessment of how mossy fiber sprouting contributes to network dysfunction and epileptogenesis. Unlike models of status epilepticus which produce massive initial damage followed by mossy fiber sprouting, kindling induces gradually progressive sprouting accompanied by incremental but cumulative neuronal loss. These features of gradually evolving cumulative circuit alterations after brief seizures evoked by kindling have enabled physiological analysis at different time points after a range of induced seizures, and thereby provide an experimental opportunity to examine both acute and chronic effects of repeated network synchronization in normal and reorganized circuits with sprouting. Exploitation of these features of kindling has enabled recognition of the distinct contributions of increased NMDA dependent synaptic currents, fading of inhibition in association with seizure-induced loss of interneuron subpopulations, and identification of recurrent excitatory circuits at both early and advanced stages of circuit reorganization (Golarai and Sutula, 1996; Sayin et al., 1999, 2003; Lynch et al., 2000; Lynch and Sutula, 2000). Similar experimental designs have been employed for analysis of structural and functional alterations in models of status epilepticus (Houser and Esclapez, 1996; Buckmaster and Dudek, 1997a, b,

556

1999; Cossart et al., 2001; Austin and Buckmaster, 2004; Peng et al., 2004; Peng and Houser, 2005; Ang et al., 2006). Attempts to define the functional effects of sprouting in both kindling that gradually evolves into spontaneous seizures and after status epilepticus that more quickly induces spontaneous seizures after a latent period have provided some straightforward lessons about the pitfalls of simple approaches in complex systems. For example, efforts to characterize the functional effects of sprouting anticipating simple associations with the presence or absence of recurrent excitatory or inhibitory circuits formed by sprouted mossy fibers which ignore independently evolving activity-dependent plasticity in excitatory and inhibitory synaptic transmission are subject to significant misinterpretation (Harvey and Sloviter, 2005; Sloviter et al., 2006). At both early and advanced stages of reorganization with minimal or extensive sprouting, recurrent excitation and seizures are emergent events in a complex system. How recurrent excitatory circuits formed by mossy fiber sprouting contribute to recurrent excitation needs to be considered from the perspective of complex systems.

Epilepsy as a ‘‘complex systems’’ disorder There is increasing awareness that biological systems often behave as ‘‘complex systems’’, and that neural circuitry of the brain and dentate gyrus fulfill criteria of a ‘‘complex system’’ (Koch and Laurent, 1999). As the circuits formed by sprouting are but one component of a complex system of molecular and cellular pathways in the reorganized dentate gyrus, when and how these new circuits are activated and contribute to network function can be anticipated to be governed by principles of ‘‘complex systems’’. The phenomena of complex systems are well-known in engineering design, but have received attention in biology only relatively recently (Gallagher and Appenzeller, 1999). Complex systems are difficult to fully explain through reductionistic understanding of their component parts, and typically demonstrate functional properties that are context dependent and cannot be readily predicted from isolated analysis of

component parts. The functional properties of complex systems typically include emergent events which are context dependent. The heterogeneous etiologies of epilepsy and the diverse variety of underlying molecular and cellular mechanisms are consistent with the view that epilepsy is a ‘‘complex systems’’ disorder which cannot be fully explained by simple understanding of one or perhaps even a few processes or components. Multiple molecular and physiological mechanisms work alone or together to promote epileptogenesis (Dudek et al., 2002; Sutula, 2002; Dudek and Shao, 2004), as demonstrated by the rapidly growing list of transgenic mice demonstrating epilepsy. An example is the interaction between cellular mechanisms studied in an isolated neuron, and the profound functional alterations that may emerge when neurons with subtle abnormalities are components of a neural network. Alterations that appear subtle at one level may produce cumulative and profound alterations manifesting as emergent properties at another level, such as recurrent excitation and behavioral seizures. Emergent properties in complex systems may not be associated with continuously detectable abnormalities (Gallagher and Appenzeller, 1999; Sole and Goodwin, 2000). Application of reductionistic approaches to ‘‘complex systems’’ often result in causal ‘‘gaps’’ between one level of understanding and the next (Sole and Goodwin, 2000). Phenomena at one level cannot be viewed in isolation, but need to be considered together with other processes at the same and at other levels. These features apply to the interactions of inhibition and excitation generated at both synaptic and network levels in the circuitry of the normal and reorganized dentate gyrus and hippocampus. Observations that spontaneous seizures are infrequent even in intractable epilepsy and functional abnormalities that become unmasked by disinhibition or extracellular ionic alterations in hippocampal slices from epileptic rats and human epileptic temporal lobe fulfill the defining criteria for emergent properties of a complex system. Understanding of constituent parts of a complex system, for example the functional properties generated by the sprouted mossy fiber pathway,

557

requires not only precise knowledge of the parts, but also the context in which the part operates. These principles are highly relevant to the interplay of activity-dependent alterations in inhibition, unmasking of recurrent excitation, generation of episodic network synchronization, and recurring behavioral seizures which are the defining features of epilepsy and are prototypical emergent events in the complex system of neural circuitry in the dentate gyrus.

Recurrent excitation in the dentate gyrus reorganized by sprouting: an emergent property of a complex system While the heterogeneity and complexity of epilepsy at the systems biology level is obvious, studies of potential underlying mechanisms for seizures and chronic epilepsies frequently employ experimental designs which anticipate linear relationships between a given alteration such as an ion channel mutation with altered biophysical properties and emergent properties such as network synchronization and behavioral seizures. At the circuit level, numerous past studies of specific molecular or cellular alterations associated with epilepsy have been pursued as possible mechanistic ‘‘causes’’ of the disorder, and then dismissed when it is recognized that sporadically expressed emergent phenomena, such as network synchronization or behavioral seizures, are not universally or linearly related to the specific defect (Dudek, 2002). This interpretive flaw is common in epilepsy research, and has been contributed to skepticism about the function and importance of mossy fiber sprouting (Armitage et al., 1998; Longo and Mello, 1998, 1999; Timofeeva and Peterson, 1999; Mohapel et al., 2000; Nissinen et al., 2001; Romcy-Pereira and Garcia-Cairasco, 2003; Sloviter et al., 2006). Experiments seeking to define the functional effects of mossy sprouting in the complex system of the reorganized dentate gyrus must include design and interpretive perspectives recognizing that sprouting is but one alteration among the many molecular and cellular alterations in epileptic neural circuitry. Characterization of its effects will require more than correlational assessment of

time dependent dynamic changes in neural circuits, but in addition manipulation of multiple variables to isolate context dependent processes. This level of complexity demands detailed quantitative characterization of the structural features of sprouting and reorganized circuitry, physiological assessment using multiple techniques, and experimental designs which recognize evolving and time dependent alterations in neurons and circuits subject to activity dependent modification across intervals spanning milliseconds to months or more. Analysis of relationships between structural alterations such as mossy fiber sprouting in the inner molecular layer of the dentate gyrus and context dependent emergent events such as recurrent excitation will be flawed if analysis is pursued without appreciation of regional variations and critical quantitative features such as numbers of excitatory and inhibitory synapses and targets. Attempting to define functional effects of sprouting based on simple association, correlation, or with anticipation of linear relationships will invariably overlook important context dependent emergent properties and will be subject to misinterpretation and flawed perspectives. The preceding sections have reviewed some of the anatomical and physiological considerations that need to be addressed in order to characterize the functional effects of sprouting in the reorganized dentate gyrus. Recurrent excitatory circuits between granule cells are the predominant recurrent connection formed by sprouted mossy fiber axons in the inner molecular layer of the reorganized dentate gyrus. The available quantitative characterizations of the synaptic connections of the sprouted mossy fiber pathway, assessment of the functional features of sprouted circuitry using multiple physiological measures and experimental designs, and the perspective of complex systems analysis of neural circuits strongly support the conclusion that mossy fiber sprouting induced in the inner molecular layer by seizures or injury forms predominantly recurrent excitatory circuits in this region whose functional effects emerge only conditionally and intermittently. The episodic emergence of recurrent excitation in this region of circuitry that is commonly involved in human focal and limbic epilepsy is a potentially important

558

functional property of circuitry reorganized by primary pathologies and recurring seizures, but should not be anticipated to be a necessary condition or absolute requirement for network synchronization underlying the heterogeneous variety of epileptic syndromes.

Anderson, A.E., Hrachovy, R.A., Antalffy, B.A., Armstrong, D.L. and Swann, J.W. (1999) A chronic focal epilepsy with mossy fiber sprouting follows recurrent seizures induced by intrahippocampal tetanus toxin injection in infant rats. Neuroscience, 92: 73–82. Ang, C.W., Carlson, G.C. and Coulter, D.A. (2006) Massive and specific dysregulation of direct cortical input to the hippocampus in temporal lobe epilepsy. J. Neurosci., 26:

Recurrent excitation as a functional effect of mossy fiber sprouting: experimental challenges and therapeutic opportunities The dentate gyrus and hippocampus are prominently involved in the most common form of intractable human epilepsy. The recognition of mossy fiber sprouting in experimental models and in human epileptic temporal lobe and appreciation of its capacity to episodically generate recurrent excitation has been a milestone for epilepsy research and a major influence on understanding of fundamental aspects of the epilepsies. The perspective that mossy fiber sprouting contributes conditionally to emergence of recurrent excitation provides a conceptual framework for understanding how injury and seizure-induced circuit reorganization may contribute to paroxysmal network synchronization, epileptogenesis, and the consequences of repeated seizures. The accomplishments of nearly two decades of investigation on mossy fiber sprouting set the stage for efforts to modify seizure-induced sprouting and processes of plasticity in the reorganizing dentate gyrus as a major translational and therapeutic opportunity for epilepsy research. References Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Adams, B., Lee, M., Fahnestock, M. and Racine, R.J. (1997) Long-term potentiation trains induce mossy fiber sprouting. Brain Res., 775: 193–197. Amaral, D.G. (1979) Synaptic extensions from the mossy fibers of the fascia dentata. Anat. Embryol., 155: 241–251. Amaral, D.G. and Dent, J.A. (1981) Development of the mossy fibers of the dentate gyrus: I. A light and electron microscopic study of the mossy fibers and their expansions. J. Comp. Neurol., 195: 51–86.

11850–11856. Armitage, L.L., Mohapel, P., Jenkins, E.M., Hannesson, D.K. and Corcoran, M.E. (1998) Dissociation between mossy fiber sprouting and rapid kindling with low-frequency stimulation of the amygdala. Brain Res., 781: 37–44. Austin, J.E. and Buckmaster, P.S. (2004) Recurrent excitation of granule cells with basal dendrites and low interneuron density and inhibitory postsynaptic current frequency in the dentate gyrus of macaque monkeys. J. Comp. Neurol., 476: 205–218. Behr, J., Heinemann, U. and Mody, I. (2001) Kindling induces transient NMDA receptor-mediated facilitation of high-frequency input in the rat dentate gyrus. J. Neurophysiol., 85: 2195–2202. Benardo, L.S. (2002) Evidence against a pathogenic role for mossy fiber sprouting. Epilepsy Curr., 2: 96–97. Ben-Ari, Y. (1985) Limbic seizure and brain damage produced by kainic acid: mechanisms and relevance to human temporal lobe epilepsy. Neuroscience, 14(2): 375–403. Ben-Ari, Y., Tremblay, E., Ottersen, O.P. and Meldrum, B.S. (1980) The role of epileptic activity in hippocampal and ‘‘remote’’ cerebral lesions induced by kainic acid. Brain Res., 191: 79–97. Ben-Ari, Y., Tremblay, E., Ottersen, O.P. and Naquet, R. (1979) Evidence suggesting secondary epileptogenic lesions after kainic acid: pre-treatment with diazepam reduces distant but not local damage. Brain Res., 165: 362–365. Bender, R.A., Dube, C., Gonzalez-Vega, R., Mina, E.W. and Baram, T.Z. (2003) Mossy fiber plasticity and enhanced hippocampal excitability, without hippocampal cell loss or altered neurogenesis, in an animal model of prolonged febrile seizures. Hippocampus, 13: 399–412. Bengzon, J., Kokaia, Z., Elmer, E., Nanobashvili, A., Kokaia, M. and Lindvall, O. (1997) Apoptosis and proliferation of dentate gyrus neurons after single and intermittent limbic seizures. Proc. Natl. Acad. Sci. U.S.A., 94: 10432–10437. Bonthius, D.J., Lothman, E.W. and Steward, O. (1995) The role of extracellular ionic changes in upregulating the mRNA for glial fibrillary acidic protein following spreading depression. Brain Res., 674: 314–328. Brooks-Kayal, A.R., Shumate, M.D., Jin, H., Rikhter, T.Y. and Coulter, D.A. (1998) Selective changes in single cell GABA(A) receptor subunit expression and function in temporal lobe epilepsy. Nat. Med., 4: 1166–1172. Buckmaster, P.S. and Dudek, F.E. (1997a) Network properties of the dentate gyrus in epileptic rats with hilar neuron loss and granule cell axon reorganization. J. Neurophysiol., 77: 2685–2696.

559 Buckmaster, P.S. and Dudek, F.E. (1997b) Neuron loss, granule cell axon reorganization, and functional changes in the dentate gyrus of epileptic kainate-treated rats. J. Comp. Neurol., 385: 385–404. Buckmaster, P.S. and Dudek, F.E. (1999) In vivo intracellular analysis of granule cell axon reorganization in epileptic rats. J. Neurophysiol., 81: 712–721. Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002) Axon sprouting in a model of temporal lobe epilepsy creates a predominantly excitatory feedback circuit. J. Neurosci., 22: 6650–6658. Buhl, E.H., Otis, T.S. and Mody, I. (1996) Zinc-induced collapse of augmented inhibition by GABA in a temporal lobe epilepsy model. Science, 271: 369–373. Cavazos, J.E., Das, I. and Sutula, T.P. (1994) Neuronal loss induced in limbic pathways by kindling: evidence for induction of hippocampal sclerosis by repeated brief seizures. J. Neurosci., 14: 3106–3121. Cavazos, J.E., Golarai, G. and Sutula, T.P. (1991) Mossy fiber synaptic reorganization induced by kindling: time course of development, progression, and permanence. J. Neurosci., 11: 2795–2803. Cavazos, J.E., Golarai, G. and Sutula, T.P. (1992) Septotemporal variation of the supragranular projection of the mossy fiber pathway in the dentate gyrus of normal and kindled rats. Hippocampus, 2: 363–372. Cavazos, J.E., Jones, S.M. and Cross, D.J. (2004) Sprouting and synaptic reorganization in the subiculum and CA1 region of the hippocampus in acute and chronic models of partialonset epilepsy. Neuroscience, 126: 677–688. Cavazos, J.E. and Sutula, T.P. (1990) Progressive neuronal loss induced by kindling: a possible mechanism for mossy fiber synaptic reorganization and hippocampal sclerosis. Brain Res., 527: 1–6. Cavazos, J.E., Zhang, P., Qazi, R. and Sutula, T.P. (2003) Ultrastructural features of sprouted mossy fiber synapses in kindled and kainic acid-treated rats. J. Comp. Neurol., 458: 272–292. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus. J. Comp. Neurol., 246: 435–458. Cohen, A.S., Lin, D.D., Quirk, G.L. and Coulter, D.A. (2003) Dentate granule cell GABA(A) receptors in epileptic hippocampus: enhanced synaptic efficacy and altered pharmacology. Eur. J. Neurosci., 17: 1607–1616. Cossart, R., Bernard, C. and Ben-Ari, Y. (2005) Multiple facets of GABAergic neurons and synapses: multiple fates of GABA signalling in epilepsies. Trends Neurosci., 28: 108–115. Cossart, R., Dinocourt, C., Hirsch, J.C., Merchan-Perez, A., De Felipe, J., Ben-Ari, Y., Esclapez, M. and Bernard, C. (2001) Dendritic but not somatic GABAergic inhibition is decreased in experimental epilepsy. Nat. Neurosci., 4: 52–62. Coulter, D.A. (2001) Epilepsy-associated plasticity in gammaaminobutyric acid receptor expression, function, and inhibitory synaptic properties. Int. Rev. Neurobiol., 45: 237–252.

Coulter, D.A. (2002) Sprouted mossy fibers form primarily excitatory connections. Epilepsy Curr., 2: 194–195. Cronin, J. and Dudek, F.E. (1988) Chronic seizures and collateral sprouting of dentate mossy fibers after kainic acid treatment in rats. Brain Res., 474: 181–184. Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E. (1992) Electrophysiology of dentate granule cells after kainateinduced synaptic reorganization of the mossy fibers. Brain Res., 573: 305–310. Cross, D.J. and Cavazos, J.E. (2007) Synaptic reorganization in subiculum and CA3 after early-life status epilepticus in the kainic acid rat model. Epilepsy Res., 73(2): 156–165. Dalby, N.O., West, M. and Finsen, B. (1998) Hilar somatostatin-mRNA containing neurons are preserved after perforant path kindling in the rat. Neurosci. Lett., 255: 45–48. Dasheiff, R.M. and McNamara, J.O. (1982) Intradentate colchicine retards the development of amygdala kindling. Ann. Neurol., 11: 347–352. Dolorfo, C.L. and Amaral, D.G. (1998) Entorhinal cortex of the rat: topographic organization of the cells of origin of the perforant path projection to the dentate gyrus. J. Comp. Neurol., 398: 25–48. Dudek, F.E. and Shao, L.R. (2004) Mossy fiber sprouting and recurrent excitation: direct electrophysiologic evidence and potential implications. Epilepsy Curr., 4: 184–187. Dudek, F.E., Staley, K.J. and Sutula, T.P. (2002) The search for animal models of epileptogenesis and pharmacoresistance: are there biologic barriers to simple validation strategies? Epilepsia, 43: 1275–1277. Dudek, F.E., Wuarin, J.P., Tasker, J.G., Kim, Y.I. and Peacock, W.J. (1995) Neurophysiology of neocortical slices resected from children undergoing surgical treatment for epilepsy. J. Neurosci. Methods, 59: 49–58. Dzhala, V.I. and Staley, K.J. (2003) Excitatory actions of endogenously released GABA contribute to initiation of ictal epileptiform activity in the developing hippocampus. J. Neurosci., 23: 1840–1846. El Bahh, B., Lespinet, V., Lurton, D., Coussemacq, M., Le Gal La Salle, G. and Rougier, A. (1999) Correlations between granule cell dispersion, mossy fiber sprouting, and hippocampal cell loss in temporal lobe epilepsy. Epilepsia, 40: 1393–1401. Epsztein, J., Represa, A., Jorquera, I., Ben-Ari, Y. and Crepel, V. (2005) Recurrent mossy fibers establish aberrant kainate receptor-operated synapses on granule cells from epileptic rats. J. Neurosci., 25: 8229–8239. Esclapez, M., Hirsch, J.C., Ben-Ari, Y. and Bernard, C. (1999) Newly formed excitatory pathways provide a substrate for hyperexcitability in experimental temporal lobe epilepsy. J. Comp. Neurol., 408: 449–460. Evans, M.S., Cady, C.J., Disney, K.E., Yang, L. and Laguardia, J.J. (2006) Three brief epileptic seizures reduce inhibitory synaptic currents, GABA(A) currents, and GABA(A)-receptor subunits. Epilepsia, 47: 1655–1664. Franck, J.E., Pokorny, J., Kunkel, D.D. and Schwartzkroin, P.A. (1995) Physiologic and morphologic characteristics of

560 granule cell circuitry in human epileptic hippocampus. Epilepsia, 36: 543–558. Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of dentate gyrus granule cells. J. Neurophysiol., 51: 195–209. Frotscher, M. and Zimmer, J. (1983) Lesion-induced mossy fibers to the molecular layer of the rat fascia dentata: identification of postsynaptic granule cells by the Golgi-EM technique. J. Comp. Neurol., 215: 299–311. Gallagher, R. and Appenzeller, T. (1999) Beyond reductionism. Science, 284: 79. Gavrilovici, C., D’Alfonso, S., Dann, M. and Poulter, M.O. (2006) Kindling-induced alterations in GABA(A) receptormediated inhibition and neurosteroid activity in the rat piriform cortex. Eur. J. Neurosci., 24: 1373–1384. Geinisman, Y., Morrell, F. and deToledo-Morrell, L. (1992) Increase in the number of axospinous synapses with segmented postsynaptic densities following hippocampal kindling. Brain Res., 569: 341–347. Gibbs III, J.W., Shumate, M.D. and Coulter, D.A. (1997) Differential epilepsy-associated alterations in postsynaptic GABA(A) receptor function in dentate granule and CA1 neurons. J. Neurophysiol., 77: 1924–1938. Golarai, G., Cavazos, J.E. and Sutula, T.P. (1992) Activation of the dentate gyrus by pentylenetetrazol evoked seizures induces mossy fiber synaptic reorganization. Brain Res., 593: 257–264. Golarai, G. and Sutula, T.P. (1996) Functional alterations in the dentate gyrus after induction of long-term potentiation, kindling, and mossy fiber sprouting. J. Neurophysiol., 75: 343–353. Gombos, Z., Spiller, A., Cottrell, G.A., Racine, R.J. and McIntyre Burnham, W. (1999) Mossy fiber sprouting induced by repeated electroconvulsive shock seizures. Brain Res., 844: 28–33. Gonzales, R.B., DeLeon Galvan, C.J., Rangel, Y.M. and Claiborne, B.J. (2001) Distribution of thorny excrescences on CA3 pyramidal neurons in the rat hippocampus. J. Comp. Neurol., 430: 357–368. Gorter, J.A., van Vliet, E.A., Aronica, E. and Lopes da Silva, F.H. (2002) Long-lasting increased excitability differs in dentate gyrus vs. CA1 in freely moving chronic epileptic rats after electrically induced status epilepticus. Hippocampus, 12: 311–324. Gutierrez, R. and Heinemann, U. (2001) Kindling induces transient fast inhibition in the dentate gyrus — CA3 projection. Eur. J. Neurosci., 13: 1371–1379. Haas, K.Z., Sperber, E.F., Opanashuk, L.A., Stanton, P.K. and Moshe, S.L. (2001) Resistance of immature hippocampus to morphologic and physiologic alterations following status epilepticus or kindling. Hippocampus, 11: 615–625. Hampson, R.E., Simeral, J.D. and Deadwyler, S.A. (1999) Distribution of spatial and nonspatial information in dorsal hippocampus. Nature, 402: 610–614. Hansen, A., Jorgensen, O.S., Bolwig, T.G. and Barry, D.I. (1990) Hippocampal kindling alters the concentration of glial fibrillary acidic protein and other marker proteins in rat brain. Brain Res., 531: 307–311.

Hardison, J.L., Okazaki, M.M. and Nadler, J.V. (2000) Modest increase in extracellular potassium unmasks effect of recurrent mossy fiber growth. J. Neurophysiol., 84: 2380–2389. Harvey, B.D. and Sloviter, R.S. (2005) Hippocampal granule cell activity and c-Fos expression during spontaneous seizures in awake, chronically epileptic, pilocarpine-treated rats: implications for hippocampal epileptogenesis. J. Comp. Neurol., 488: 442–463. Hellier, J.L., Patrylo, P.R., Dou, P., Nett, M., Rose, G.M. and Dudek, F.E. (1999) Assessment of inhibition and epileptiform activity in the septal dentate gyrus of freely behaving rats during the first week after kainate treatment. J. Neurosci., 19: 10053–10064. Holahan, M.R., Rekart, J.L., Sandoval, J. and Routtenberg, A. (2006) Spatial learning induces presynaptic structural remodeling in the hippocampal mossy fiber system of two rat strains. Hippocampus, 16: 560–570. Holmes, G.L., Gairsa, J.L., Chevassus-Au-Louis, N. and BenAri, Y. (1998) Consequences of neonatal seizures in the rat: morphological and behavioral effects. Ann. Neurol., 44: 845–857. Holmes, G.L., Sarkisian, M., Ben-Ari, Y. and Chevassus-AuLouis, N. (1999) Mossy fiber sprouting after recurrent seizures during early development in rats. J. Comp. Neurol., 404: 537–553. Houser, C.R. and Esclapez, M. (1996) Vulnerability and plasticity of the GABA system in the pilocarpine model of spontaneous recurrent seizures. Epilepsy Res., 26: 207–218. Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O., Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered patterns of dynorphin immunoreactivity suggest mossy fiber reorganization in human hippocampal epilepsy. J. Neurosci., 10: 267–282. Jiang, M., Lee, C.L., Smith, K.L. and Swann, J.W. (1998) Spine loss and other persistent alterations of hippocampal pyramidal cell dendrites in a model of early-onset epilepsy. J. Neurosci., 18: 8356–8368. de Jonge, M. and Racine, R.J. (1987) The development and decay of kindling-induced increases in paired-pulse depression in the dentate gyrus. Brain Res., 412: 318–328. Kapur, J., Bennett Jr., J.P., Wooten, G.F. and Lothman, E.W. (1989a) Evidence for a chronic loss of inhibition in the hippocampus after kindling: biochemical studies. Epilepsy Res., 4: 100–108. Kapur, J. and Macdonald, R.L. (1997) Rapid seizure-induced reduction of benzodiazepine and Zn2+ sensitivity of hippocampal dentate granule cell GABA(A) receptors. J. Neurosci., 17: 7532–7540. Kapur, J., Michelson, H.B., Buterbaugh, G.G. and Lothman, E.W. (1989b) Evidence for a chronic loss of inhibition in the hippocampus after kindling: electrophysiological studies. Epilepsy Res., 4: 90–99. Kapur, J., Stringer, J.L. and Lothman, E.W. (1989c) Evidence that repetitive seizures in the hippocampus cause a lasting reduction of GABAergic inhibition. J. Neurophysiol., 61: 417–426.

561 Khalilov, I., Dzhala, V., Ben-Ari, Y. and Khazipov, R. (1999) Dual role of GABA in the neonatal rat hippocampus. Dev. Neurosci., 21: 310–319. Khurgel, M. and Ivy, G.O. (1996) Astrocytes in kindling: relevance to epileptogenesis. Epilepsy Res., 26: 163–175. Kim, Y.I., Peacock, W.J. and Dudek, F.E. (1993) Properties and synaptic mechanisms of bicuculline-induced epileptiform bursts in neocortical slices from children with intractable epilepsy. J. Neurophysiol., 70: 1759–1766. Koch, C. and Laurent, G. (1999) Complexity and the nervous system. Science, 284: 96–98. Kotloski, R., Lynch, M., Lauersdorf, S. and Sutula, T. (2002) Repeated brief seizures induce progressive hippocampal neuron loss and memory deficits. Prog. Brain Res., 135: 95–110. Kotti, T., Riekkinen Sr., P.J. and Miettinen, R. (1997) Characterization of target cells for aberrant mossy fiber collaterals in the dentate gyrus of epileptic rat. Exp. Neurol., 146: 323–330. Lehmann, T.N., Gabriel, S., Eilers, A., Njunting, M., Kovacs, R., Schulze, K., Lanksch, W.R. and Heinemann, U. (2001) Fluorescent tracer in pilocarpine-treated rats shows widespread aberrant hippocampal neuronal connectivity. Eur. J. Neurosci., 14: 83–95. Leinekugel, X., Khalilov, I., McLean, H., Caillard, O., Gaiarsa, J.L., Ben-Ari, Y. and Khazipov, R. (1999) GABA is the principal fast-acting excitatory transmitter in the neonatal brain. Adv. Neurol., 79: 189–201. Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997) Connections of the hippocampal formation in humans: I. The mossy fiber pathway. J. Comp. Neurol., 385: 325–351. Litt, B. and Echauz, J. (2002) Prediction of epileptic seizures. Lancet Neurol., 1: 22–30. Litt, B., Esteller, R., Echauz, J., D’Alessandro, M., Shor, R., Henry, T., Pennell, P., Epstein, C., Bakay, R., Dichter, M. and Vachtsevanos, G. (2001) Epileptic seizures may begin hours in advance of clinical onset: a report of five patients. Neuron, 30: 51–64. Longo, B.M. and Mello, L.E. (1998) Supragranular mossy fiber sprouting is not necessary for spontaneous seizures in the intrahippocampal kainate model of epilepsy in the rat. Epilepsy Res., 32: 172–182. Longo, B.M. and Mello, L.E. (1999) Effect of long-term spontaneous recurrent seizures or reinduction of status epilepticus on the development of supragranular mossy fiber sprouting. Epilepsy Res., 36: 233–241. Lynch, M., Sayin, U., Golarai, G. and Sutula, T. (2000) NMDA receptor-dependent plasticity of granule cell spiking in the dentate gyrus of normal and epileptic rats. J. Neurophysiol., 84: 2868–2879. Lynch, M. and Sutula, T. (2000) Recurrent excitatory connectivity in the dentate gyrus of kindled and kainic acid-treated rats. J. Neurophysiol., 83: 693–704. Masukawa, L.M., Higashima, M., Kim, J.H. and Spencer, D.D. (1989) Epileptiform discharges evoked in hippocampal brain slices from epileptic patients. Brain Res., 493: 168–174. Masukawa, L.M., Uruno, K., Sperling, M., O’Connor, M.J. and Burdette, L.J. (1992) The functional relationship between

antidromically evoked field responses of the dentate gyrus and mossy fiber reorganization in temporal lobe epileptic patients. Brain Res., 579: 119–127. Matveeva, E.A., Vanaman, T.C., Whiteheart, S.W. and Slevin, J.T. (2006) Asymmetric accumulation of hippocampal 7S SNARE complexes occurs regardless of kindling paradigm. Epilepsy Res., 73(3): 266–274. McCarren, M. and Alger, B.E. (1985) Use-dependent depression of IPSPs in rat hippocampal pyramidal cells in vitro. J. Neurophysiol., 53: 557–571. Messenheimer, J.A., Harris, E.W. and Steward, O. (1979) Sprouting fibers gain access to circuitry transsynaptically altered by kindling. Exp. Neurol., 64: 469–481. Michelson, H.B., Kapur, J. and Lothman, E.W. (1989) Reduction of paired pulse inhibition in the CA1 region of the hippocampus by pilocarpine in naive and in amygdala-kindled rats. Exp. Neurol., 104: 264–271. Mikkonen, M., Soininen, H., Kalvianen, R., Tapiola, T., Ylinen, A., Vapalahti, M., Paljarvi, L. and Pitkanen, A. (1998) Remodeling of neuronal circuitries in human temporal lobe epilepsy: increased expression of highly polysialylated neural cell adhesion molecule in the hippocampus and the entorhinal cortex. Ann. Neurol., 44: 923–934. Mohapel, P., Armitage, L.L., Gilbert, T.H., Hannesson, D.K., Teskey, G.C. and Corcoran, M.E. (2000) Mossy fiber sprouting is dissociated from kindling of generalized seizures in the guinea-pig. Neuroreport, 11: 2897–2901. Molnar, P. and Nadler, J.V. (1999) Mossy fiber-granule cell synapses in the normal and epileptic rat dentate gyrus studied with minimal laser photostimulation. J. Neurophysiol., 82: 1883–1894. Nadler, J.V., Perry, B.W. and Cotman, C.W. (1980a) Selective reinnervation of hippocampal area CA1 and the fascia dentata after destruction of CA3-CA4 afferents with kainic acid. Brain Res., 182: 1–9. Nadler, J.V., Perry, B.W., Gentry, C. and Cotman, C.W. (1980b) Degeneration of hippocampal CA3 pyramidal cells induced by intraventricular kainic acid. J. Comp. Neurol., 192: 333–359. Nadler, J.V., Perry, B.W., Gentry, C. and Cotman, C.W. (1981) Fate of the hippocampal mossy fiber projection after destruction of its postsynaptic targets with intraventricular kainic acid. J. Comp. Neurol., 196: 549–569. Niquet, J., Jorquera, I., Faissner, A., Ben-Ari, Y. and Represa, A. (1995) Gliosis and axonal sprouting in the hippocampus of epileptic rats are associated with an increase of tenascin-C immunoreactivity. J. Neurocytol., 24: 611–624. Nissinen, J., Lukasiuk, K. and Pitkanen, A. (2001) Is mossy fiber sprouting present at the time of the first spontaneous seizures in rat experimental temporal lobe epilepsy? Hippocampus, 11: 299–310. Nusser, Z., Hajos, N., Somogyi, P. and Mody, I. (1998) Increased number of synaptic GABA(A) receptors underlies potentiation at hippocampal inhibitory synapses. Nature, 395: 172–177. Okazaki, M.M., Evenson, D.A. and Nadler, J.V. (1995) Hippocampal mossy fiber sprouting and synapse formation after

562 status epilepticus in rats: visualization after retrograde transport of biocytin. J. Comp. Neurol., 352: 515–534. Otis, T.S., De Koninck, Y. and Mody, I. (1994) Lasting potentiation of inhibition is associated with an increased number of gamma-aminobutyric acid type A receptors activated during miniature inhibitory postsynaptic currents. Proc. Natl. Acad. Sci. U.S.A., 91: 7698–7702. Parent, J.M., Janumpalli, S., McNamara, J.O. and Lowenstein, D.H. (1998) Increased dentate granule cell neurogenesis following amygdala kindling in the adult rat. Neurosci. Lett., 247: 9–12. Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking of new glutamatergic pathways in the dentate gyrus of hippocampal slices from kainate-induced epileptic rats. J. Neurophysiol., 79: 418–429. Patrylo, P.R., Schweitzer, J.S. and Dudek, F.E. (1999) Abnormal responses to perforant path stimulation in the dentate gyrus of slices from rats with kainate-induced epilepsy and mossy fiber reorganization. Epilepsy Res., 36: 31–42. Peng, Z. and Houser, C.R. (2005) Temporal patterns of fos expression in the dentate gyrus after spontaneous seizures in a mouse model of temporal lobe epilepsy. J. Neurosci., 25: 7210–7220. Peng, Z., Huang, C.S., Stell, B.M., Mody, I. and Houser, C.R. (2004) Altered expression of the delta subunit of the GABA(A) receptor in a mouse model of temporal lobe epilepsy. J. Neurosci., 24: 8629–8639. Pierce, J.P. and Milner, T.A. (2001) Parallel increases in the synaptic and surface areas of mossy fiber terminals following seizure induction. Synapse, 39: 249–256. Pitkanen, A., Nissinen, J., Lukasiuk, K., Jutila, L., Paljarvi, L., Salmenpera, T., Karkola, K., Vapalahti, M. and Ylinen, A. (2000) Association between the density of mossy fiber sprouting and seizure frequency in experimental and human temporal lobe epilepsy. Epilepsia, 41: S24–S29. Pretel, S., Applegate, C.D. and Piekut, D. (1997) Apoptotic and necrotic cell death following kindling induced seizures. Acta Histochem., 99: 71–79. Qiao, X. and Noebels, J.L. (1993) Developmental analysis of hippocampal mossy fiber outgrowth in a mutant mouse with inherited spike-wave seizures. J. Neurosci., 13: 4622–4635. Ramirez-Amaya, V., Escobar, M.L., Chao, V. and BermudezRattoni, F. (1999) Synaptogenesis of mossy fibers induced by spatial water maze overtraining. Hippocampus, 9: 631–636. Represa, A. and Ben-Ari, Y. (1992) Kindling is associated with the formation of novel mossy fibre synapses in the CA3 region. Exp. Brain Res., 92: 69–78. Represa, A., Le Gall La Salle, G. and Ben-Ari, Y. (1989a) Hippocampal plasticity in the kindling model of epilepsy in rats. Neurosci. Lett., 99: 345–350. Represa, A., Robain, O., Tremblay, E. and Ben-Ari, Y. (1989b) Hippocampal plasticity in childhood epilepsy. Neurosci. Lett., 99: 351–355. Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy fibers in rats and gerbils form synapses with the somata and proximal dendrites of basket cells in the dentate gyrus. Hippocampus, 1: 355–364.

Ribak, C.E., Seress, L., Weber, P., Epstein, C.M., Henry, T.R. and Bakay, R.A. (1998) Alumina gel injections into the temporal lobe of rhesus monkeys cause complex partial seizures and morphological changes found in human temporal lobe epilepsy. J. Comp. Neurol., 401: 266–290. Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and Nadler, J.V. (2000) Status epilepticus-induced hilar basal dendrites on rodent granule cells contribute to recurrent excitatory circuitry. J. Comp. Neurol., 428: 240–253. Romcy-Pereira, R.N. and Garcia-Cairasco, N. (2003) Hippocampal cell proliferation and epileptogenesis after audiogenic kindling are not accompanied by mossy fiber sprouting or Fluoro-Jade staining. Neuroscience, 119: 533–546. van Rooyen, F., Young, N.A., Larson, S.E. and Teskey, G.C. (2006) Hippocampal kindling leads to motor map expansion. Epilepsia, 47: 1383–1391. Salin, P., Tseng, G.F., Hoffman, S., Parada, I. and Prince, D.A. (1995) Axonal sprouting in layer V pyramidal neurons of chronically injured cerebral cortex. J. Neurosci., 15: 8234–8245. Sayin, U., Osting, S., Hagen, J., Rutecki, P. and Sutula, T. (2003) Spontaneous seizures and loss of axo-axonic and axosomatic inhibition induced by repeated brief seizures in kindled rats. J. Neurosci., 23: 2759–2768. Sayin, U., Rutecki, P. and Sutula, T. (1999) NMDA-dependent currents in granule cells of the dentate gyrus contribute to induction but not permanence of kindling. J. Neurophysiol., 81: 564–574. Scharfman, H.E., Sollas, A.L., Berger, R.E. and Goodman, J.H. (2003) Electrophysiological evidence of monosynaptic excitatory transmission between granule cells after seizure-induced mossy fiber sprouting. J. Neurophysiol., 90: 2536–2547. Scharfman, H.E., Sollas, A.L. and Goodman, J.H. (2002) Spontaneous recurrent seizures after pilocarpine-induced status epilepticus activate calbindin-immunoreactive hilar cells of the rat dentate gyrus. Neuroscience, 111: 71–81. Sloviter, R.S., Dichter, M.A., Rachinsky, T.L., Dean, E., Goodman, J.H., Sollas, A.L. and Martin, D.L. (1996) Basal expression and induction of glutamate decarboxylase and GABA in excitatory granule cells of the rat and monkey hippocampal dentate gyrus. J. Comp. Neurol., 373: 593–618. Sloviter, R.S., Zappone, C.A., Harvey, B.D. and Frotscher, M. (2006) Kainic acid-induced recurrent mossy fiber innervation of dentate gyrus inhibitory interneurons: possible anatomical substrate of granule cell hyper-inhibition in chronically epileptic rats. J. Comp. Neurol., 494: 944–960. Smith, B.N. and Dudek, F.E. (2001) Short- and long-term changes in CA1 network excitability after kainate treatment in rats. J. Neurophysiol., 85: 1–9. Sole, R. and Goodwin, B. (2000) Signs of Life — How Complexity Invades Biology. Basic Books, New York. Staley, K.J., Soldo, B.L. and Proctor, W.R. (1995) Ionic mechanisms of neuronal excitation by inhibitory GABA(A) receptors. Science, 269: 977–981. Stanfield, B.B. (1989) Excessive intra- and supragranular mossy fibers in the dentate gyrus of tottering (tg/tg) mice. Brain Res., 480: 294–299.

563 Staubli, U., Gall, C. and Lynch, G. (1984) The distribution of the commissural-associational afferents of the dentate gyrus after perforant path lesions in one-day-old rats. Brain Res., 292: 156–159. Steward, O. (1976) Reinnervation of dentate gyrus by homologous afferents following entorhinal cortical lesions in adult rats. Science, 194: 426–428. Steward, O. (1992) Lesion-induced synapse reorganization in the hippocampus of cats: sprouting of entorhinal, commissural/ associational, and mossy fiber projections after unilateral entorhinal cortex lesions, with comments on the normal organization of these pathways. Hippocampus, 2: 247–268. Steward, O., Cotman, C. and Lynch, G. (1976) A quantitative autoradiographic and electrophysiological study of the reinnervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral entorhinal lesions. Brain Res., 114: 181–200. Steward, O., Cotman, C.W. and Lynch, G.S. (1973) Re-establishment of electrophysiologically functional entorhinal cortical input to the dentate gyrus deafferented by ipsilateral entorhinal lesions: innervation by the contralateral entorhinal cortex. Exp. Brain Res., 18: 396–414. Steward, O., Cotman, C.W. and Lynch, G.S. (1974) Growth of a new fiber projection in the brain of adult rats: re-innervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral entorhinal lesions. Exp. Brain Res., 20: 45–66. Steward, O. and Messenheimer, J.A. (1978) Histochemical evidence for a post-lesion reorganization of cholinergic afferents in the hippocampal formation of the mature cat. J. Comp. Neurol., 178: 697–709. Stringer, J.L. (1996) Repeated seizures increase GFAP and vimentin in the hippocampus. Brain Res., 717: 147–153. Stringer, J.L. and Lothman, E.W. (1989) Repetitive seizures cause an increase in paired-pulse inhibition in the dentate gyrus. Neurosci. Lett., 105: 91–95. Sur, M. and Rubenstein, J.L. (2005) Patterning and plasticity of the cerebral cortex. Science, 310: 805–810. Sutula, T. (2002) Seizure-induced axonal sprouting: assessing connections between injury, local circuits, and epileptogenesis. Epilepsy Curr., 2: 86–91. Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L. (1989) Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol., 26: 321–330. Sutula, T., Harrison, C. and Steward, O. (1986) Chronic epileptogenesis induced by kindling of the entorhinal cortex: the role of the dentate gyrus. Brain Res., 385: 291–299. Sutula, T., He, X.X., Cavazos, J. and Scott, G. (1988) Synaptic reorganization in the hippocampus induced by abnormal functional activity. Science, 239: 1147–1150. Sutula, T., Zhang, P., Lynch, M., Sayin, U., Golarai, G. and Rod, R. (1998) Synaptic and axonal remodeling of mossy fibers in the hilus and supragranular region of the dentate

gyrus in kainate-treated rats. J. Comp. Neurol., 390: 578–594. Tauck, D.L. and Nadler, J.V. (1985) Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid-treated rats. J. Neurosci., 5: 1016–1022. Timofeeva, O.A. and Peterson, G.M. (1999) Dissociation of mossy fiber sprouting and electrically-induced seizure sensitivity: rapid kindling versus adaptation. Epilepsy Res., 33: 99–115. Toyoda, I. and Buckmaster, P.S. (2005) Prolonged infusion of cycloheximide does not block mossy fiber sprouting in a model of temporal lobe epilepsy. Epilepsia, 46: 1017–1020. Tuff, L.P., Racine, R.J. and Adamec, R. (1983a) The effects of kindling on GABA-mediated inhibition in the dentate gyrus of the rat. I. Paired-pulse depression. Brain Res., 277: 79–90. Tuff, L.P., Racine, R.J. and Mishra, R.K. (1983b) The effects of kindling on GABA-mediated inhibition in the dentate gyrus of the rat. II. Receptor binding. Brain Res., 277: 91–98. Vaidya, V.A., Siuciak, J.A., Du, F. and Duman, R.S. (1999) Hippocampal mossy fiber sprouting induced by chronic electroconvulsive seizures. Neuroscience, 89: 157–166. Wenzel, H.J., Woolley, C.S., Robbins, C.A. and Schwartzkroin, P.A. (2000) Kainic acid-induced mossy fiber sprouting and synapse formation in the dentate gyrus of rats. Hippocampus, 10: 244–260. Williams, P.A., Wuarin, J.P., Dou, P., Ferraro, D.J. and Dudek, F.E. (2002) Reassessment of the effects of cycloheximide on mossy fiber sprouting and epileptogenesis in the pilocarpine model of temporal lobe epilepsy. J. Neurophysiol., 88: 2075–2087. Worrell, G.A., Parish, L., Cranstoun, S.D., Jonas, R., Baltuch, G. and Litt, B. (2004) High-frequency oscillations and seizure generation in neocortical epilepsy. Brain, 127: 1496–1506. Wuarin, J.P. and Dudek, F.E. (1996) Electrographic seizures and new recurrent excitatory circuits in the dentate gyrus of hippocampal slices from kainate-treated epileptic rats. J. Neurosci., 16: 4438–4448. Wuarin, J.P. and Dudek, F.E. (2001) Excitatory synaptic input to granule cells increases with time after kainate treatment. J. Neurophysiol., 85: 1067–1077. Wuarin, J.P., Kim, Y.I., Cepeda, C., Tasker, J.G., Walsh, J.P., Peacock, W.J., Buchwald, N.A. and Dudek, F.E. (1990) Synaptic transmission in human neocortex removed for treatment of intractable epilepsy in children. Ann. Neurol., 28: 503–511. Zhang, L.X., Smith, M.A., Li, X.L., Weiss, S.R.B. and Post, R.M. (1998) Apoptosis of hippocampal neurons after amygdala kindled seizures. Mol. Brain Res., 55: 198–208. Zhang, N. and Houser, C.R. (1999) Ultrastructural localization of dynorphin in the dentate gyrus in human temporal lobe epilepsy: a study of reorganized mossy fiber synapses. J. Comp. Neurol., 405: 472–490.

This page intentionally left blank

SECTION V

The Dentate Gyrus Network

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 30

A behavioral analysis of dentate gyrus function Raymond P. Kesner University of Utah, Department of Psychology, 380 S. 1530 E., Room 502, Salt Lake City, UT 84121, USA

Abstract: Computational models of the dentate gyrus (DG) have suggested based on anatomical, electrophysiological, and computer simulation data that the DG plays an important role in learning and memory by processing and representing spatial information on the basis of conjunctive encoding, pattern separation, and encoding of spatial information in conjunction with the CA3. Behavioral evidence supports a role for the DG in mnemonic processing of spatial information based on the operation of conjunctive encoding of multiple sensory inputs, pattern separation of spatial (especially metric) information, and subsequent encoding in cooperation with CA3. A potential role of the DG in mediating processes, such as recall of sequential information and short-term memory as well as temporal order for remote memory, are also discussed. Keywords: conjunctive encoding; spatial pattern separation; dentate gyrus; encoding; retrieval Introduction

Conjunctive encoding

The anatomy and neural circuitry of the dentate gyrus (DG) has been described in detail in previous chapters in this volume (Chapter 1, Amaral, Scharfman and Lavenex). Based on the anatomy of the DG, its input and output pathways, and the development of a computational model, Rolls (1996) and Rolls and Kesner (2006) have suggested that the DG has three major functions including conjunctive encoding of multiple sensory inputs, spatial pattern separation, and facilitation of encoding of spatial information based on its outputs to CA3. This is accomplished by a competitive learning network with Hebb-like modifiability to remove redundancy from the inputs and produce a more orthogonal, sparse, and categorized set of outputs.

It can be shown that the DG receives multiple sensory inputs including vestibular, olfactory, visual, auditory, and somatosensory from the perirhinal cortex and lateral entorhinal cortex in conjunction with spatially organized grid cells from the medial entorhinal cortex (Hafting et al., 2005) to represent metric spatial representations. The perforant path input in the DG can be divided into a medial and lateral component. The medial component processes spatial information and the lateral component processes non-spatial (e.g. objects, odors) information (Witter et al., 1989a; Hargreaves et al., 2005). Based on the idea that the medial perforant path input into the DG mediates spatial information via activation of NMDA receptors and the lateral perforant path input into the DG mediates visual object information via activation of opioid receptors, the following experiment was conducted. Using a paradigm developed by Poucet (1989), rats were tested for the

Corresponding author. Tel.: +801 581 7430; Fax: +801 581 5841; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63030-1

567

568

detection of a novel spatial change and the detection of a novel visual object change under the influence of direct infusions of AP5 (an NMDA antagonist) or naloxone (a m opiate antagonist) into the DG. The results indicated that naloxone infusions into the DG disrupted both novelty detection of a spatial location and a visual object, whereas AP5 infusions into the DG disrupted only detection of a novel spatial location, but had no effect on detection of a novel object. In contrast, infusions of AP5 into the CA1 region disrupts only the detection of a spatial location change, but not a visual object change, whereas naloxone into the CA1 region disrupts only the detection of a visual object change, but not a spatial location change (Hunsaker et al., 2007). These data suggest that the DG uses conjunctive encoding of visual object and spatial information to provide for a spatial representation, which I will show below might be based on metric information.

Spatial pattern separation It can clearly be demonstrated that single cells within the hippocampus are activated by most sensory inputs, including vestibular, olfactory, visual, auditory, and somatosensory as well as higher order integration of sensory stimuli (Cohen and Eichenbaum, 1993). The question of importance is whether these sensory inputs via conjunctive encoding have a memory representation within the hippocampus. One possible role for the hippocampus in processing all sensory information might be to provide sensory markers to demarcate a spatial location, so that the hippocampus can more efficiently mediate spatial information. Thus, it is possible that one of the main processing functions of the hippocampus is to encode and separate spatial events from each other. This would ensure that new highly processed sensory information is organized within the hippocampus and enhances the possibility of encoding and temporarily remembering one place as separate from another place. This may be accomplished via pattern separation of event information, so that similar spatial events can be separated from each other and spatial interference is reduced. To the

extent that DG acts to produce separate representations of different but similar places, it is predicted that the DG will be especially important when memories must be formed about similar places. Rolls’ model proposes that pattern separation is facilitated by sparse connections in the mossy-fiber system, which connects DG granular cells to CA3 pyramidal neurons. The separation of patterns is accomplished based on the low probability that any two CA3 neurons will receive mossy fiber input synapses from a similar subset of DG cells. The mossy fiber inputs to CA3 from DG are suggested to be essential during learning and may influence which CA3 neurons will fire based on the distributed activity in the DG. The cells of the DG are suggested to act as a competitive learning network with Hebb-like modifiability to reduce redundancy and produce sparse, orthogonal outputs. O’Reilly and McClelland (1994) and Shapiro and Olton (1994) also suggest that the mossy fiber connections between the DG and CA3 may support pattern separation. Rolls (1996) notes additional characteristics of the mossy fiber projection system that may promote pattern separation in the DG-CA3 system and hence, efficient information storage in CA3. First, mossy fiber synapses are very large and terminate close to the soma of the CA3 pyramidal neurons in the pyramidal layer. Therefore, the mossy fiber synapses will be relatively powerful in activating the CA3 cell. The projections from layer II of entorhinal cortex to CA3 terminate in the lacunosum moleculare layer of the CA3 pyramidal cells, which is much further from the soma. Second, the firing activity of granule cells within the DG is sparse (Jung and McNaughton, 1992) and coupled with the small number of connections to CA3 cells, which would produce a sparse signal that may be transformed into an even sparser signal in CA3. It also has been demonstrated that the place fields of DG cells (Mizumori et al., 1990) and specifically granular cells (Jung and McNaughton, 1992) are small and highly reliable, which may support the role of DG in pattern separation. In addition, the mossy fibers demonstrate non-associative plasticity (Brown et al., 1989), which may enhance the signalto-noise ratio such that the mossy fiber cell would

569

produce a non-linearly amplified current in the CA3 cell. Due to this type of plasticity in the DG, a particular stimulus such as spatial location would be likely to activate the same population of CA3 neurons across subsequent presentations, which may result in economical storage, because any given CA3 cell could be used in different memories. If disruption of DG function results in inefficient pattern separation, then deficits on spatial tasks may occur when there is increased overlap or similarity among distal cues and presumably increased similarity among representations within the DG. Remembering a specific location in an 8-arm maze, a water maze, or a spatial context in fear conditioning may be influenced by the degree of overlap among critical distal spatial cues. Rats with lesions in DG have been tested on a working memory version of the radial 8-arm maze. The results demonstrated that a lesion of the DG resulted in deficits similar to complete hippocampal lesions (Walsh et al., 1986; Tilson et al., 1987; McLamb et al., 1988; Emerich and Walsh, 1989). In addition, rats with DG lesions were tested on the Morris water maze task and showed deficits comparable to rats with complete hippocampal lesions when the start location varied on each trial (Sutherland et al., 1983; Nanry et al., 1989; Xavier et al., 1999; Jeltsch et al., 2001). Lee and Kesner (2004a) tested rats with DG lesions on acquisition and retrieval of contextual fear conditioning. Rats with DG lesions showed initial impairments in freezing behavior during acquisition of the task but eventually reached the level of freezing of controls with subsequent testing. When retrieval of contextual fear was examined in rats with DG lesions 24 h after acquisition, the animals showed a significant deficit in freezing compared to controls. Based on these studies it is clear that DG lesions impair spatial memory. The DG lesioned animals may be impaired as a result of impaired pattern separation; however, it is difficult to determine if a particular memory process is impaired in these animals using these three paradigms. To examine the contribution of the DG to spatial pattern separation, Gilbert et al. (2001) tested rats with DG lesions using a paradigm that

measured short-term memory for spatial location information as a function of spatial similarity between two spatial locations (Gilbert et al., 1998). Specifically, the study was designed to examine the role of the DG subregion in discriminating spatial locations when rats were required to remember a spatial location based on distal environmental cues and differentiate between the to-be-remembered location and a distractor location with different degrees of similarity or overlap among the distal cues. Animals were tested using a cheeseboard maze apparatus on a delayed-match-to-sample for a spatial location task. Animals were trained to displace an object that was randomly positioned to cover a baited food well in 1 of 15 locations along a row of food wells. Following a short delay, the animals were required to choose between two objects identical to the sample phase object. One object was in the same location as the sample phase object and the second object was in a different location along the row of food wells. An animal was rewarded for displacing the object in the same spatial location as the sample phase object (correct choice) but received no reward for displacing the foil object (incorrect choice). Five spatial separations, from 15 cm to 105 cm, were used to separate the correct object and the foil object during the choice phase. The results showed that rats with DG lesions were significantly impaired at short spatial separations; however, the performance of the DG lesioned animals increased as a function of increased spatial separation between the correct object and the foil on the choice phases. The performance of rats with DG lesioned matched controls at the largest spatial separation. The graded nature of the impairment and the significant linear improvement in performance as a function of increased separation illustrate the deficit in pattern separation. Based on these results, it was concluded that lesions of the DG decrease the efficiency of spatial pattern separation, which resulted in impairments on trials with increased spatial proximity and increased spatial similarity among working memory representations. Based on the evidence that the hippocampus, and especially the DG, receives inputs from all sensory modalities, there is a possibility that the DG uses sensory markers to demarcate a spatial

570

location, allowing the DG to more efficiently represent spatial information. Thus, one function of the DG may be to encode and separate events in space resulting in spatial pattern separation. Spatial pattern separation would ensure that new highly processed sensory information is organized within the hippocampus, which in turn enhances the possibility of encoding and temporarily remembering one spatial location as separate from another location. Based on the observations that cells in the CA3 and CA1 region respond to changes in the metric and topological aspects of the environment (O’Keefe and Burgess, 1996; Jeffery and Anderson, 2003), one can ask the question whether these different features of the spatial environment are processed via the DG and subsequently transferred to the CA3 subregion or are these features communicated via the direct perforant path projection to the CA3 subregion? In both cases the information may then be transferred to the CA1 subregion. To answer this question, Goodrich-Hunsaker et al. (2005) examined the contribution of the DG to memory for metric and topological spatial relationships. Using a modified version of an exploratory paradigm developed by Poucet (1989), rats with DG, CA3, and CA1 lesions and controls were tested on tasks involving a metric spatial manipulation. In this task, a rat was allowed to explore two different visual objects that were separated by a specific distance on a cheeseboard maze. On the initial presentation of the objects, the rat explored each object. However, across subsequent presentations of the objects in the same spatial locations, the rat habituated and eventually spent less time exploring the objects. Once the rat had habituated to the objects in their locations, the metric spatial distance between the two objects was manipulated so the two objects were either closer together or further apart. The time the rat spent exploring each moved object was recorded. The results showed that DG lesions impaired detection of the metric distance change, in that rats with DG lesions spent significantly less time exploring the two objects that were displaced relative to controls. Rats with CA3 or CA1 lesions displayed smaller reductions in re-exploration and the DG was reliably impaired relative to controls and CA1. It is also

possible to examine the detection of a topological spatial change. In the topological manipulation condition, rats were allowed to explore four different visual objects that were positioned in a square on the cheeseboard maze. The rats again were allowed to explore the objects and eventually habituated to the objects with subsequent exposure. However, following habituation, the locations of two of the objects were switched and the time the rat spent exploring each object was recorded. The results showed that rats with CA1 lesions impair, whereas DG or CA3, lesions do not impair the detection of the topological manipulation. The results suggest that neurons in the DG may be critically involved in processing spatial information on a metric scale, but may not be necessary for representing topological space. The results of both experiments provide for empirical validation of the role of DG in spatial pattern separation and support the predictions of computational models (Rolls, 1996; Rolls and Kesner, 2006). There are some studies in the literature that have demonstrated that hippocampal lesions, including the DG, can result in deficits for spatial tasks that can be interpreted to be a function of increased interference and an impairment in the utilization of a spatial pattern separation process. A few examples will suffice. Because rats are started in different locations in the standard water maze task, there is potential interference among similar and overlapping spatial patterns. Thus, the observation that hippocampal lesioned rats are impaired in learning and subsequent consolidation of important spatial information in this task could be due to difficulty to separate spatial patterns resulting in enhanced spatial interference. Support for this idea comes from the observation of Eichenbaum et al. (1990) who demonstrated that when fimbria-fornix lesioned rats are trained on the water maze task from only a single starting position (less spatial interference) there are hardly any learning deficits, whereas training from many different starting points resulted in learning difficulties. In a somewhat similar study, it was shown that total hippocampal lesioned rats learned or consolidated rather readily that only one spatial location was correct on an 8-arm maze (Hunt et al., 1994). In a different study, McDonald and

571

White (1995) used a place preference procedure in an 8-arm maze. In this procedure food is placed at the end of one arm and no food is placed at the end of another arm. In a subsequent preference test normal rats prefer the arm that contained the food. In this study fornix lesioned rats acquired the place preference task as quickly as controls if the arm locations were opposite each other, but the fornix lesioned rats were markedly impaired if the locations were adjacent to each other. Clearly, there would be greater spatial interference when the spatial locations are adjacent to each other rather than far apart. In a different task, rats were trained on a pair-wise visual spatial discrimination of vertical lines that were not very far apart. Compared to controls, rats with hippocampal lesions were not able to learn the task (Lazaris et al., 2006). Thus, spatial pattern separation may play an important role in the acquisition of new spatial information and there is a good possibility that the DG may have been the subregion responsible for the impairments in the various tasks described above. Do other subregions of the hippocampus engage in spatial pattern separation? It appears that the CA3 region may also engage spatial pattern separation processes. For example, Tanila (1999) showed that CA3 place cells were able to maintain distinct representations of two visually identical environments and selectively reactivate either one of the representation patterns depending on the experience of the rat. Also, Leutgeb et al. (2004) recently showed that when rats experienced a completely different environment, CA3 place cells developed orthogonal representations of those different environments by changing their firing rates between the two environments, whereas CA1 place cells maintained similar responses. In a different study Vazdarjanova and Guzowski (2004) placed rats in two different environments separated by 30 min. The two environments differed greatly in that different objects were located in each room. The authors were able to monitor the time course of activations of ensembles of neurons in both CA3 and CA1, using a new immediate-early gene-based brain-imaging method (Arc/H1a catFISH). When the two environments were significantly different, CA3

neurons exhibited lower overlap in their activity between the two environments compared to CA1 neurons. This could be consistent with the computational point that if CA3 is an autoassociator, the pattern representations within it should be as orthogonal as possible to maximize memory capacity and minimize interference. The actual pattern separation may be performed, the computational model holds, as a result of the operation of the dentate granule cells as a competitive net and the nature of the mossy fiber connections to CA3 cells. Thus, CA3 may represent different environments relatively orthogonally. In the computational account, each environment would be a separate chart, and the number of charts that could be stored in CA3 would be high if the representations in each chart were relatively orthogonal to those in other charts and further, charts could operate independently (Stringer et al., 2004). Any one chart of a given spatial environment can be understood as a continuous attractor network with place cells with Gaussian-shaped place fields that overlap continuously with each other. In different charts (different spatial environments) the same neurons may represent very different regions of space, and neurons representing close places in one environment may represent distant locations in another environment (chart). It appears that the DG mediated pattern separation is based on reducing interference between discrete spatial locations based on metric representations of distance between the spatial locations, whereas the CA3 mediated pattern separation is based on reducing interference across global environments based on representation of unique charts. Whether the actual pattern separation in CA3 is performed as a result of the operation of the dentate granule cells as a competitive net and the nature of the mossy fiber connections to CA3 cells remains to be investigated. An alternative possibility is that pattern separation observed for the global environment is separate from pattern separation observed for discrete elements in the DG. The dissociation between these two forms of pattern separation could be achieved by assuming that the perforant path input into CA3 is critical for representing a global environment or context. The perforant path into

572

CA3 can also be divided into a medial and lateral component. Because there is associative LTP in CA3 between the medial or lateral perforant path and the intrinsic commissural/associational-CA3 synapses (Martinez et al., 2002), either place, object or contextual cues can be processed by the associative medial and lateral perforant path connections to CA3 cells. This is consistent with the finding that disruption of the perforant path input impairs the multiple trial acquisition of an objectplace paired associate task [perhaps because retrieval of what has been previously learned is impaired (unpublished observations)]. In addition, Martinez et al. (2002) demonstrated associative (cooperative) LTP between the medial and lateral perforant path inputs to the CA3 neurons. This could provide a mechanism for object (lateral perforant path) — place (medial perforant path) associative learning in conjunction with a global context. Further support for this idea is based on testing rats for the detection of a novel spatial change and the detection of a novel visual object change under the influence of direct infusions of AP5 or naloxone into the CA3. The results indicated that naloxone or AP5 infusions into CA3 disrupted both novelty detection of a spatial location and a visual object, suggesting that the perforant path input into CA3 could provide the context for integration of both spatial and visual object information (Hunsaker et al., 2007). These studies thus indicate that both the DG and CA3 play an important role in spatial pattern separation for spatial memory tasks. Other forms of pattern separation do not involve the DG subregion of the hippocampus. For example, temporal pattern separation is mediated by the CA1, but not the DG (Gilbert et al., 2001). Furthermore, hippocampal lesions including DG do not produce a deficit for pattern separation of reward values, visual objects, or motor responses (Gilbert and Kesner, 2002, 2003b, 2006; Saksida et al., 2006). Instead, the perirhinal cortex subserves pattern separation for visual objects (Bussey et al., 2002; Gilbert and Kesner, 2003), the amygdala subserves pattern separation for reward value (Gilbert and Kesner, 2002) and the caudate nucleus subserves pattern separation for motor responses (Kesner and Gilbert, 2006).

Encoding vs. retrieval of spatial information Rolls’ (1996) computational model postulates that the dentate/mossy fiber system is necessary for setting up the appropriate conditions for optimal storage of new information in the CA3 system (which could be called encoding). In order to test the model rats with DG or CA3 lesions were administered 10 learning trials per day in a HebbWilliams maze. The results based on a within-day analysis indicated that DG and CA3 lesions impaired the acquisition of this task, consistent with an encoding or learning impairment (Lee and Kesner, 2004b; Jerman et al., 2006). However, when tested using a between days analysis, retrieval of what had been learned previously was not impaired by DG lesions (Lee and Kesner, 2004b; Jerman et al., 2006). In the Hebb-Williams learning task, one can measure improvement in performance within a day, reflecting the operation of encoding of new information based in part on short-term memory representations and one can also measure improvement in performance between days, reflecting the operation of retrieval of information either based on intermediate-term memory representations mediated by synaptic consolidation and/or access to stored information. It is recognized that the separation of encoding from retrieval processes is extremely difficult. Therefore, it is assumed that during acquisition within a day, there will be a greater involvement of encoding compared to retrieval processes, and that during retention across days, there will be a greater involvement of retrieval compared to encoding processes. It is also assumed that encoding encompasses spatial pattern separation processes in conjunction with associative processes and representations within short-term memory. Even though there is likely to be some retrieval from short-term memory that may also occur during acquisition, it is assumed not to be the dominant factor governing performance within the first 10 trials on Day 1. Furthermore, it is assumed that retrieval 24 h later encompasses associative processes as well as representations within intermediate-term memory. Even though there is likely to be some encoding that may also occur during retrieval, it is assumed not to be the most critical

573

determinant of performance within the first five trials on Day 2. It is therefore of interest that the results indicate that both DG and CA3 lesions disrupted encoding, but not retrieval, even though CA3 lesions did produce more errors during acquisition than DG lesions. The DG lesion data are consistent with the results of (Lassalle et al., 2000), who showed that in a water maze learning task, lesions of the mossy fibers disrupted encoding, but not retrieval. The CA3 lesion data are consistent with the idea that short-term encoding of new information is mediated by CA3. Other studies involving alteration of CA3 function support this idea (Lee and Kesner, 2002, 2003, 2004a; Nakazawa et al., 2003). Thus, the data suggest a possible cooperation between the DG and CA3 in encoding of information in a Hebb-Williams maze, in that there is a deficit in encoding of information following either a DG or CA3 lesion. The present data appear to be in conflict with the results of Lee and Kesner (2004b), wherein a lesion to the perforant path into CA3 caused a deficit in retrieval but no effect for encoding. There is the possibility that the lesion to the perforant path may have disrupted the Schaffer collateral output from CA3 into CA1. Since the Lee and Kesner paper was published (2004b), HebbWilliams maze data for CA1 lesioned animals from our laboratory has shown that animals with lesions to dorsal CA1 show no deficit in encoding but a significant deficit in retrieval relative to a control group (Vago et al., 2007). If the Schaffer collateral input into CA1 were disrupted by the perforant path lesion performed by Lee and Kesner (2004b), then the effects seen in that study would be more indicative of a CA1 effect rather than a CA3 effect. This would be more consistent with the present results indicating that a lesion to CA3 does not result in a deficit for retrieval of information — only encoding. Given that there was an encoding, but not a retrieval deficit for the DG and CA3, could there be an interaction in terms of the encoding processes between the DG and the CA3? This cooperation between DG and CA3 could theoretically derive from the observation that the DG granule cells project to CA3 pyramidal neurons via mossy fiber

projections forming the primary output of the DG. Based on characteristics of the mossy fiber system, Rolls (1996) suggests that pattern separation may be a function of the DG and its mossy fiber projections to CA3 and thus may facilitate the encoding of spatial information via an interaction between the DG and CA3. However, recent studies have shown that the functions of these two hippocampal subregions can be dissociated using behavioral tasks. For example, Gilbert and Kesner (2003a) demonstrated that DG lesioned animals were able to learn object-place and odor-place paired-associate tasks as quickly as controls. However, rats with CA3 lesions showed significant learning impairments on both tasks. Furthermore, a lesion of the perforant path input into the CA3 also disrupted object-place paired associate learning (unpublished observations). These results could indicate that the CA3, but not the DG, subregion is involved in associative learning. Furthermore, DG, but not CA3, lesioned rats produce a deficit in the acquisition of the standard version water maze (Sutherland et al., 1983; Nanry et al., 1989; Xavier et al., 1999; Lassalle et al., 2000; Jeltsch et al., 2001). The dissociations that arise between DG and CA3 are primarily due to the CA3 subregion of the hippocampus having two major inputs with a direct connection from the DG via the mossy fibers and a direct input from the perforant path that bypasses the DG. An alternate explanation is possible since a lesion of the perforant path disrupts retrieval, but not encoding (Lee and Kesner, 2004b). The differences between DG and CA3 in the above mentioned tasks could be simply a function of differential intrinsic processing of similar spatial information. As an another example Gilbert et al. (2001) and Gilbert and Kesner (2006) tested rats with DG or CA3 lesions using a paradigm that measured one-trial short-term memory for spatial location information as a function of spatial similarity between two spatial locations. The results showed that rats with DG lesions were significantly impaired at short spatial separations; however, the performance of the DG lesioned rats improved as a function of increased spatial separation between the correct object and the foil on the choice phases. In contrast, CA3 lesioned rats were impaired for all spatial

574

separations, suggesting a disruption of a shortterm memory process subserved by CA3 as an intrinsic contribution to the spatial pattern separation task. Therefore, there is no guarantee that any potential cooperation between DG and CA3 in encoding information in the Hebb-Williams maze is due to the direct connection between these two regions. To examine this issue, a disconnection study was carried out using an ipsilateral lesion (DG and CA3 lesion on one side only) group vs. a contralateral lesion (DG on one side and CA3 on the other side) group. It is assumed the right and left hemispheres operate in parallel. Crossed lesions (i.e. unilateral lesions in contralateral hemispheres), therefore, would disrupt communication within each of the two hemispheres, thus functionally disconnecting the two brain regions. If the DG and CA3 subregions of the hippocampus interact, then crossed lesioned rats should be markedly impaired relative to animals with lesions on the same side. If these two regions do not interact, then crossed lesions would not produce a deficit, suggesting that each subregion produces a deficit in encoding for different reasons such as spatial pattern separation problems vs. associative or short-term memory problems. The results indicate that rats with an ipsilateral lesion of both DG and CA3 do not produce a deficit in either encoding or retrieval in the HebbWilliams maze. In contrast, rats with a DG lesion on one side of the brain and a CA3 lesion on the other side of the brain disrupted encoding, but not retrieval, of information on the Hebb-Williams maze. Thus, the DG and CA3 interact to support encoding of spatial information in the present task. There is always the possibility that following a DG lesion on one side and a CA3 lesion on the other side, there would be a significant amount of degeneration within CA3 following the DG lesion, so that there would be a lesion of the CA3 lesion on the other side in effect creating a bilateral lesion of CA3 and similarly a significant amount of degeneration within DG following the CA3 lesion in effect creating a bilateral DG lesion. Against this argument is that upon histological analysis, there is no clear loss of the contralateral CA3 following the DG lesion and no loss of the contralateral DG following the CA3 lesion (see Jerman et al., 2005 for data

concerning subregional specificity of CA3 and DG lesions). Thus, the data suggest that the DG and CA3 cooperate and interact with each other in the encoding of new spatial information in the HebbWilliams maze without affecting retrieval.

Additional potential functions of the dentate gyrus Based on the observation that neurogenesis occurs in the DG and that new DG granule cells can be formed across time, it has been proposed that the DG mediates a temporal pattern separation mechanism that can generate patterns of episodic memories within remote memory (Aimone et al., 2006). There are currently no behavioral data available to assess the proposal. It should be noted that the CA1, but not DG, is involved in temporal ordering of spatial information within an intermediate memory system, but remote memory has not been investigated (Kesner et al., 2004). There are some data that suggest that disruption of neurogenesis using MAM (an anti-mitotic agent) impairs the learning of trace eye-blink conditioning, but has no effect on delay eye-blink conditioning, water maze spatial navigation, or contextual fear conditioning (Shors, 2002). Furthermore, trace eye-lid conditioning prevents the loss of newly developed granule cells (Shors, 2004). These data suggest that the DG may play a role in temporal processing, but it appears to be based on a different mechanism than the one proposed by Aimone et al. (2006). Based on the observation that the DG has a recurrent network system in which DG granule cells excite mossy cells which feed back modifiable excitatory connections to the granule cells and that CA3 pyramidal cells have a feedback connection to the dentate network (see Scharfman’s Chapter 34, this volume), Lisman (1999) has suggested that this recurrent circuit can provide for short-term memory across seconds and thus can play an important role in providing accurate recall of sequences. From a behavioral perspective, there is evidence that the DG contributes to short-term or working memory (Xavier et al., 1999; Lee and Kesner, 2003), which could be a function of both the CA3 and DG recurrent circuits. However, there is

575

currently no behavioral evidence to demonstrate that the DG contributes to sequence recall. In summary, based on the development of computational models of DG and behavioral evidence based on dysfunction of DG, it appears that the DG mediates mnemonic processing of spatial information. The processes subserved by DG include (a) the operation of conjunctive encoding of multiple sensory inputs, implying an integration of sensory inputs to determine a spatial representation, (b) pattern separation of spatial (especially metric) information, involving the reduction of interference between similar spatial locations, and (c) subsequent encoding in cooperation with CA3, suggesting an important role in the acquisition of spatial information. A potential role of the DG in recall of sequential information requires more empirical evidence. References Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for adult neurogenesis in the encoding of time in new memories. Nat. Neurosci., 9: 723–727. Brown, T.H., Ganong, A.H., Kairiss, E.W., Keenan, C.L. and Kelso, S.R. (1989) Long-term potentiation in two synaptic systems of the hippocampal brain slice. In: Byrne J.H. and Berry W.O. (Eds.), Neural Models of Plasticity. Academic Press, San Diego, CA, pp. 266–306. Bussey, T.J., Saksida, L.M. and Murray, E.A. (2002) Perirhinal cortex resolves and feature ambiguity in complex discriminations. Eur. J. Neurosci., 15: 365–374. Cohen, N.J. and Eichenbaum, H.B. (1993) Memory, Amnesia, and Hippocampal Function. MIT Press, Cambridge. Eichenbaum, H., Stewart, C. and Morris, R.G.M. (1990) Hippocampal representation in spatial learning. J. Neurosci., 10: 331–339. Emerich, D.F. and Walsh, T.J. (1989) Selective working memory impairments following intradentate injection of colchicine: attenuation of the behavioral not the neuropathological effects by gangliosides GM1 and AGF2. Physiol. Behav., 45: 93–101. Gilbert, P.E. and Kesner, R.P. (2002) The amygdala but not the hippocampus is involved in pattern separation based on reward value. Neurobiol. Learn. Mem., 77: 338–353. Gilbert, P.E. and Kesner, R.P. (2003a) Localization of function within the dorsal hippocampus: the role of the dorsal CA3 subregion in paired-associate learning. Behav. Neurosci., 117: 1385–1394. Gilbert, P.E. and Kesner, R.P. (2003b) Recognition memory for complex visual discrimination is influenced by stimulus interference in rodents with perirhinal cortex damage. Learn. Mem., 10: 525–530.

Gilbert, P.E. and Kesner, R.P. (2006) The role of dorsal CA3 hippocampal subregion in spatial working memory and pattern separation. Behav. Brain Res., 169: 142–149. Gilbert, P.E., Kesner, R.P. and DeCoteau, W.E. (1998) The role of the hippocampus in mediating spatial pattern separation. J. Neurosci., 18: 804–810. Gilbert, P.E., Kesner, R.P. and Lee, I. (2001) Dissociating hippocampal subregions: a double dissociation between the dentate gyrus and CA1. Hippocampus, 11: 626–636. Goodrich-Hunsaker, N.J., Hunsaker, M.R. and Kesner, R.P. (2005) Effects of hippocampus sub-regional lesions for metric and topological spatial information processing. Society for Neuroscience Abstracts, Number 647.1. Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser, E.I. (2005) Microstructure of a spatial map in the entorhinal cortex. Nature, 436: 801–806. Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J. (2005) Major dissociation between medial and lateral entorhinal input to dorsal hippocampus. Science, 308: 1792–1794. Hunsaker, M.R., Mooy, G.G., Swift, J. and Kesner, R.P. (2007) Dissociations of the role of the medial and lateral perforant path projections into dorsal DG, CA3 and CA1 for spatial and nonspatial (visual object) information processing. Behav. Neuroscience (in press). Hunt, M.E., Kesner, R.P. and Evans, R.B. (1994) Memory for spatial location: functional dissociation of entorhinal cortex and hippocampus. Psychobiology, 22: 186–194. Jeffery, K.J. and Anderson, M.I. (2003) Dissociation of the geometric and contextual influences on place cells. Hippocampus, 13: 868–872. Jeltsch, H., Bertrand, F., Lazarus, C. and Cassel, J.-C. (2001) Cognitive performances and locomotor activity following dentate granule cell damage in rats: role of lesion extent and type of memory tested. Neurobiol. Learn. Mem., 76: 81–105. Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Disconnection analysis of CA3 and DG in mediating encoding but not retrieval in a spatial maze learning task. Learn. Mem., 13: 458–464. Jerman, T.S., Kesner, R.P., Lee, I. and Berman, R.F. (2005) Patterns of hippocampal cell loss based on subregional lesions of the hippocampus. Brain Res., 1065: 1–7. Jung, M.W. and McNaughton, B.L. (1992) Spatial selectivity of unit activity in the hippocampal granular layer. Hippocampus, 3: 165–182. O’Keefe, J. and Burgess, N. (1996) Geometric determinants of the place field of hippocampal neurons. Nature, 381: 425–428. Kesner, R.P. and Gilbert, P.E. (2006) The role of the medial caudate nucleus, but not the hippocampus, in a matching-to sample task for a motor response. Eur. J. Neurosci., 23: 1888–1894. Kesner, R.P., Lee, I. and Gilbert, P. (2004) A behavioral assessment of hippocampal function based on a subregional analysis. Rev. Neurosci., 15: 333–351. Lassalle, J.M., Bataille, T. and Halley, H. (2000) Reversible inactivation of the hippocampal mossy fiber synapses in mice

576 impairs spatial learning, but neither consolidation nor memory retrieval, in the Morris navigation task. Neurobiol. Learn. Mem., 73: 243–257. Lazaris, A., Talpos, J.C., Dias, R., Saksida, L.M. and Bussey, T.J. (2006) Lesions of the hippocampus impair a pair-wise spatial discrimination task in a touchscreen-equipped operant box. 5th Forum of European Neuroscience, Program No. A160.9 Poster board 390. FENS Forum Abstracts, Vienna, Austria. Lee, I. and Kesner, R.P. (2002) Differential contribution of NMDA receptors in hippocampal subregions to spatial working memory. Nat. Neurosci., 5: 162–168. Lee, I. and Kesner, R.P. (2003) Differential role of dorsal hippocampus subregions in spatial working memory with short versus intermediate delay. Behav. Neurosci., 117: 1044–1053. Lee, I. and Kesner, R.P. (2004a) Differential contributions of dorsal hippocampal subregions to memory acquisition and retrieval in contextual fear-conditioning. Hippocampus, 14: 301–310. Lee, I. and Kesner, R.P. (2004b) Encoding versus retrieval of spatial memory: double dissociation between the dentate gyrus and the perforant path inputs into CA3 in the dorsal hippocampus. Hippocampus, 14: 66–76. Leutgeb, S., Leutgeb, J.K., Treves, A., Moser, M.B. and Moser, E.L. (2004) Distinct ensemble codes in hippocampal areas CA3 and CA1. Science, 305: 1295–1298. Lisman, J.E. (1999) Relating hippocampal circuitry to function: recall of memory sequences by reciprocal dentate-CA3 interactions. Neuron, 22: 233–242. Martinez, C.O., Do, V.H., Martinez Jr., J.L. and Derrick, B.E. (2002) Associative long-term potentiation (LTP) among extrinsic afferents of the hippocampal CA3 region in vivo. Brain Res., 940: 86–94. McDonald, R.J. and White, N.M. (1995) Hippocampal and nonhippocampal contributions to place learning in rats. Behav. Neurosci., 109: 579–593. McLamb, R.L., Mundy, W.R. and Tilson, H.A. (1988) Intradentate colchicine disrupts the acquisition and performance of a working memory task in the radial arm maze. Neurotoxicology, 9: 521–528. Mizumori, S.J.Y., Perez, G.M., Alvarado, M.C., Barnes, C.A. and McNaughton, B.L. (1990) Reversible inactivation of the medial septum differentially affects two forms of learning in rats. Brain Res., 528: 12–20. Nakazawa, K., Sun, L.D., Quirk, M.C., Rondi-Reig, L., Wilson, M.A. and Tonegawa, S. (2003) Hippocampal CA3 NMDA receptors are crucial for memory acquisition of onetime experience. Neuron, 38: 305–315. Nanry, K.P., Mundy, W.R. and Tilson, H.A. (1989) Colchicineinduced alternations of reference memory in rats: role of spatial versus non-spatial task components. Behav. Brain Res., 35: 45–53. Poucet, B. (1989) Object exploration, habituation, and response to a spatial change in rats following septal or medial frontal cortical damage. Behav. Neurosci., 103: 1009–1016. O’Reilly, R.C. and McClelland, J.L. (1994) Hippocampal conjunctive encoding, storage, and recall: avoiding a trade- off. Hippocampus, 4: 661–682.

Rolls, E.T. (1996) A theory of hippocampal function in memory. Hippocampus, 6: 601–620. Rolls, E.T. and Kesner, R.P. (2006) A computational theory of hippocampal function, and empirical tests of the theory. Prog. Neurobiol., 79: 1–48. Saksida, L.M., Bussey, T.J., Buckmaster, C.A. and Murray, E.A. (2006) No effect of hippocampal lesions on perirhinal cortex-dependent feature-ambiguous visual discriminations. Hippocampus, 16: 421–430. Shapiro, M.L. and Olton, D.S. (1994) Hippocampal function and interference. In: Schacter D.L. and Tulving E. (Eds.), Memory Systems 1994. MIT Press, London, pp. 141–146. Shors, T.J. (2002) Neurogenesis may relate to some but not all types of hippocampal-dependent learning. Hippocampus, 12: 578–584. Shors, T.J. (2004) Memory traces of trace memories: neurogenesis, synaptogenesis and awareness. Trends Neurosci., 27: 250–256. Stringer, S.M., Rolls, E.T. and Trappenberg, T.P. (2004) Selforganising continuous attractor networks with multiple activity packets, and the representation of space. Neurl. Netwk., 17: 5–27. Sutherland, R.J., Whitshaw, I.Q. and Kolb, B. (1983) A behavioural analysis of spatial localization following electrolytic, kainate- or colchicine-induced damage to the hippocampal formation in the rat. Behav. Brain Res., 7: 133–153. Tanila, H. (1999) Hippocampal place cells can develop distinct representations of two visually identical environments. Hippocampus, 9: 235–246. Tilson, H.A., Rogers, B.S., Crimes, L., Harry, J.G., Peterson, N.J., Hong, J.S. and Dryer, R.S. (1987) Time-dependent neurobiological effects of colchicine administered directly into the hippocampus of rats. Brain Res., 408: 163–172. Vago D.R., Bevan, A. and Kesner, R.P. (2007) The role of the direct perforant path input to the CA1 subregion of the dorsal hippocampus in memory retention and retrieval. Hippocampus (in press). Vazdarjanova, A. and Guzowski, J.F. (2004) Differences in hippocampal neuronal population responses to modifications of an environmental context: evidence for distinct, yet complementary, functions of CA3 and CA1 ensembles. J. Neurosci., 24: 6489–6496. Walsh, T.J., Schulz, D., Tilson, H.A. and Schmechel, D.E. (1986) Colchicine-induced granule cell loss in rat hippocampus: selective behavioral and histological alterations. Brain Res., 389: 23–36. Witter, M.P., Groenewegen, H.J., Lopes da Silva, F.H. and Lohman, A.H. (1989) Functional organization of the extrinsic and intrinsic circuitry of the parahippocampal region. Prog. Neurobiol., 33: 161–253. Xavier, G.F., Oliveira-Filho, F.J.B. and Santos, A.M.G. (1999) Dentate gyrus-selective colchicine lesion and disruption of performance in spatial tasks: difficulties in ‘‘place strategy’’ because of a lack of flexibility in the use of environmental cues? Hippocampus, 9: 668–681.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 31

Models, structure, function: the transformation of cortical signals in the dentate gyrus La´szlo´ Acsa´dy1, and Szabolcs Ka´li1,2 1

Institute of Experimental Medicine, Hungarian Academy of Sciences, PO Box 67, 1450 Budapest, Hungary 2 HAS-PPCU-SU Neurobionics Research Group, Budapest, Hungary

Abstract: Our central question is why the hippocampal CA3 region is the only cortical area capable of forming interference-free representations of complex environmental events (episodes), given that apparently all cortical regions have recurrent excitatory circuits with modifiable synapses, the basic substrate for autoassociative memory networks. We review evidence for the radical (but classic) view that a unique transformation of incoming cortical signals by the dentate gyrus and the subsequent faithful transfer of the resulting code by the mossy fibers are absolutely critical for the appropriate association of memory items by CA3 and, in general, for hippocampal function. In particular, at the gate of the hippocampal formation, the dentate gyrus possesses a set of unusual properties, which selectively evolved for the task of code transformation between cortical afferents and the hippocampus. These evolutionarily conserved anatomical features enable the dentate gyrus to translate the noisy signal of the upstream cortical areas into the sparse and specific code of hippocampal formation, which is indispensable for the efficient storage and recall of multiple, multidimensional memory items. To achieve this goal the mossy fiber pathway maximally utilizes the opportunity to differentially regulate its postsynaptic partners. Selective innervation of CA3 pyramidal cells and interneurons by distinct terminal types creates a favorable condition to differentially regulate the short-term and long-term plasticity and the motility of various mossy terminal types. The utility of this highly dynamic system appears to be the frequency-dependent fine-tuning of excitation and inhibition evoked by the large and the small mossy terminals respectively. This will determine exactly which CA3 cell population is active and induces permanent modification in the autoassociational network of the CA3 region. The learning of arbitrary associations of complex, multimodal items during a single exposure (i.e., episodic memory) is irreversibly compromised following hippocampal lesions, but the acquisition of simple associative pairings (e.g., classical conditioning) and motor learning remain intact (Squire, 1992). Hippocampal-dependent memory traces can be used flexibly, i.e., they can be activated in a context different from the one where they were learned. Clinical studies in humans, as well as a large number of behavioral and physiological

Introduction The hippocampus, a peculiar cortical structure, has long been known to be involved in higher order cognitive functions, most notably, memory formation and spatial navigation (Scoville and Milner, 1957; O’Keefe and Dostrovsky, 1971). Not all kinds of memory depend on the hippocampus. Corresponding author. Tel.: +36-1-210-9413; Fax: +36-1-210-9412; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63031-3

577

578

experiments in other mammalian species (mainly rodents) have also implicated the hippocampus in the formation and flexible use of world-centered (allocentric) spatial representations (O’Keefe and Nadel, 1978). It has been argued repeatedly that these two domains share several important characteristics, and may have fairly similar computational requirements. In particular, both episodic memory and spatial navigation may require the fast storage and subsequent recall of specific conjunctions of environmental stimuli. It has been suggested that the bulk of neocortex, which is assumed to be involved in continuously creating and refining an internal representation of the general structure of the observed world, may not be well suited for the rapid, interference-free acquisition of such specific memory traces (McClelland et al., 1995). On the other hand, the hippocampus might be optimized for exactly this operation, and could thus complement the generic learning capabilities of the neocortex. However, the hippocampus constitutes only a tiny fraction of the cortical areas, and it has a relatively simple structure. It receives input from and sends information back to multimodal associational cortical areas. The question is why neocortex needs the hippocampal loop to implement rapid learning of arbitrary complex associations? Why is it that other cortical areas with multiple cellular layers cannot do the same job? What is so special about hippocampus that enables it to establish the most complex memory traces? In short, what is the ‘‘trick’’ of the hippocampus? For a long time, the central role for memory formation was assigned to the hippocampal CA3 region (Marr, 1971; McNaughton and Morris, 1987; Rolls, 1989). This hippocampal subfield was favored by most computational neuroscientists and electrophysiologists since its principal cells, the CA3 pyramidal cells, form a so-called ‘‘autoassociative memory network’’ with their abundant, local recurrent collaterals (Li et al., 1994). In computational models these types of network were found to be optimal for the efficient storage of a large number of memory items and for reactivation of complete memory traces if only part of the trace was provided, which are key components of episodic memory. Moreover, the synapses among

CA3 pyramidal cells, as well as those between pyramidal cells in CA3 and their primary downstream target CA1, are subject to associative and cooperative long-term potentiation (LTP), a possible molecular mechanism underlying memory formation (Debanne et al., 1998). Interestingly, however, many features of CA3 pyramidal cells are shared by pyramidal cells of other cortical areas. Neocortical pyramidal cells have just as profuse recurrent local collaterals and their synapses are subject to long-term plastic changes. CA3 and layer II–III cells also have many intrinsic electrophysiological characteristics in common. Thus, autoassociative networks with plastic synapses are abundant in the cortex. Therefore, it is not immediately obvious why the formation of complex memory traces is restricted to the hippocampal formation and cannot be performed by other multimodal associational cortical areas. Two features of the CA3 area, however, clearly distinguish it from other cortical regions. CA3 pyramidal cells receive a prominent excitatory input to their proximal apical dendrites (Ramon y Cajal, 1911; Claiborne et al., 1986), which enables an unusually strong spike coupling between an upstream region — the dentate gyrus (DG) — and CA3 (Henze et al., 2002). The faithfulness of this synaptic transmission is unparalleled in excitatory cortical circuits. The second feature (shared by the principal cells of the other hippocampal subregions) is the way in which the firing pattern of pyramidal cell codes the environmentally relevant stimuli. Most pyramidal cells in other cortical regions are characterized by higher spontaneous firing rates, and a lower rate of modulation by the appropriate stimuli. In sharp contrast, background activity in hippocampal principal cells, and granule cells in particular, is very low (Muller et al., 1987; Barnes et al., 1990; Quirk et al., 1992; Jung and McNaughton, 1993). However, when hippocampal principal cells participate in information transfer (e.g., place cells), their activity increases enormously. In computational terms, the hippocampus uses a sparse code, whereas coding in other cortical areas is denser. Apparently, the hippocampus and the other cortical regions ‘‘speak’’ different neuronal languages. Since the environmentally specific information reaches the

579

hippocampus via other cortical areas, the immediate consequence of the difference in coding strategies is that densely encoded cortical information reaching the hippocampus needs to be translated into a sparser hippocampal code. In other words, an interface is needed between the hippocampus and neocortex. The main task of this interface would be the translation of the neocortical code into a hippocampal one, and to transfer the new code as efficiently as possible to the next station of information processing. According to the classic view, this interface is the DG, the first step of the trisynaptic hippocampal loop. We argue that the DG is the ‘‘odd-man-out’’ among cortical regions. In particular, at the gate of the hippocampal formation, the DG possesses a set of unusual properties that selectively evolved for the task of code transformation between cortical afferents and the hippocampus. These evolutionarily conserved anatomical features enable the DG to translate the noisy signal of the upstream cortical areas to the sparse and specific code of hippocampal formation, which is indispensable for the formation of multiple, multidimensional memory items.

Computational requirements for the formation of episodic memories There is a sort of general consensus about how the hippocampus could contribute to cortical memory functions (Alvarez and Squire, 1994; Treves and Rolls, 1994; McClelland et al., 1995; Kali and Dayan, 2004; Rolls and Kesner, 2006). First, the hippocampus is assumed to be capable of rapidly creating and storing a memory trace, which is distinct from all existing traces. The memory trace is associated with a snapshot of activity in medial temporal neocortex (particularly entorhinal cortex), which, in turn, is thought to represent a compressed version of activity in the rest of neocortex. Second, the hippocampus is thought to be capable of retrieving particular stored traces if the entorhinal activity pattern provides only a partial or noisy version of the corresponding original pattern. Using this cue, the hippocampus reinstates the original pattern in the entorhinal cortex and,

subsequently, the rest of neocortex. This operation is referred to as pattern completion, or autoassociative memory function. Finally, the hippocampus may also be capable of autonomously reactivating stored memory traces in the absence of specific retrieval cues, thereby reactivating complete cortical memory representations during ‘‘offline’’ behavioral states (e.g., slow-wave sleep). Such replay may contribute in various ways to the consolidation (transfer to a final repository) and maintenance of episodic and semantic memory (Kali and Dayan, 2004). But do we have any reason to believe that the hippocampus is even capable of carrying out these operations, and if so, that it is in some sense optimized for exactly these tasks? The most widely cited evidence is the existence of the extensive recurrent collateral network of pyramidal neurons in area CA3. Such recurrent networks are the classic examples of autoassociative memory devices, whose properties have been extensively investigated by theoretical means and computer simulations (Marr, 1971; McNaughton and Morris, 1987; Willshaw and Buckingham, 1990; Treves and Rolls, 1992; Samsonovich and McNaughton, 1997; Kali and Dayan, 2000). In particular, the storage capacity of such networks, i.e., the number of patterns they can store and retrieve reliably, has been determined (Treves and Rolls, 1992), and was found to depend substantially on the properties of the set of input patterns that we attempt to store. Capacity is roughly inversely proportional to the sparsity of individual patterns (i.e., the proportion of active units in a pattern), and generally increases as the overlap between different stored patterns decreases. Therefore, it is reasonable to assume that the need to maximize storage capacity is an important reason for the conspicuously low activity levels of principal cell populations in the hippocampus (the proportion of active principal cells in any hippocampal subfield at any given moment is thought to be on the order of a few percent (Barnes et al., 1990; Quirk et al., 1992; Jung and McNaughton, 1993; Treves and Rolls, 1994 #5013). However, activity patterns in most areas of neocortex, including entorhinal cortex, appear to be much denser (involving a larger proportion of

580

neurons at any given time), which is thought to be beneficial for generalization (McClelland et al., 1995), an important characteristic of the kind of representational learning that the neocortex may be engaged in. On the other hand, if the optimal type of activity pattern (dense vs. sparse) is different in the neocortex and the hippocampus, then the information contained in the entorhinal dense code needs to be ‘‘translated’’ into a sparse code when hippocampal representations are created (and vice versa). In principle, such translation could be implemented directly by a single set of projections from the input area to the autoassociative network (i.e., by the direct perforant path input from entorhinal cortex to the CA3 region), without an intervening specific interface for sparsification. However, as argued on theoretical grounds by Treves and Rolls (1992), a projection with the characteristics of the perforant path — a large number of relatively weak, associatively modifiable synapses on each target cell — may be optimal during retrieval, but a different type of input to the CA3 recurrent network — one with a small number of individually strong synapses per cell — is probably required for the storage of new memory traces. The mossy fiber pathway, the projection to CA3 from granule cells, meets the requirements for this second type of input. Additional theoretical and computer simulation studies by O’Reilly and McClelland (1994) indicated that the two-stage pathway from EC to DG to CA3 could perform very effective pattern separation, provided that individual mossy fiber connections were sufficiently strong to transfer the benefits of pattern separation in the DG to CA3. They further argued that the coexistence of this indirect pathway with the direct EC–CA3 connection enables the hippocampus to avoid an inherent conflict between pattern separation and pattern completion, both of which are important for the efficient operation of the hippocampal autoassociator. In summary, these studies suggest that a possible role of the DG is to form sparse, pattern-separated representations of entorhinal activity patterns, and transmit this sparse representation reliably for subsequent storage in the CA3 recurrent network. But how does the DG create sparse representations from the relatively dense entorhinal activity

patterns, which constitute its only major cortical input? Since no direct experimental investigation of this issue has been undertaken, we need to rely on indirect evidence and the results of computational studies to try to answer this question. At an abstract level, well-known computational algorithms exist which create sparse, pattern-separated representations from distributed input patterns. A simple example is the competitive-learning pattern classification device described by Rumelhart and Zipser (1986), which has been shown to be capable of generating hippocampal place field-like activity patterns from input patterns resembling neocortical sensory representations (Sharp, 1991). This algorithm operates on binary input patterns and generates binary output patterns. For each presentation of an input pattern, the current values of the synaptic weights (which are initially set to random values) are used to determine the feedforward activation of units in the output layer. Then the output unit with the highest level of feedforward input is allowed to become active (this step is assumed to reflect the action of feedforward and feedback inhibitory circuits), and the incoming weights of this unit are allowed to change. This weight change is assumed to be Hebbian in nature: weights from active input units are increased, while weights from inactive units are decreased in such a way that the sum of all incoming weights remains constant (and identical to the sum of incoming weights to all other output units). This way, the ‘‘winning’’ unit will have an even higher level of feedforward activation the next time the same input pattern appears. It will also show an increased response to other similar input patterns; on the other hand, its response to patterns that are very different (with a low degree of overlap) will diminish. As a result, different output units will eventually respond to different kinds of patterns, and end up partitioning the input space among themselves into non-overlapping groups of similar patterns. The algorithm performs both sparsification — since only a single output unit becomes active for any given (distributed) input pattern — and orthogonalization (pattern separation) — since relatively dissimilar, but still overlapping input patterns end up activating different output units (zero overlap).

581

However, is there any experimental evidence that sparsification requires a separate relay station, and that the DG has unique features — besides its well-known ‘‘detonator’’ type of terminals — for code conversion and reliable transmission? In the following pages we review behavioral, morphological, and physiological data relevant to these questions.

Lesion studies Let us first examine whether the putative role of the DG as described above is consistent with behavioral data on the effects of specific lesions. In general, we expect that some, but not all tasks that are sensitive to global hippocampal lesions will also be sensitive to more specific lesions of the DG, and might hope that the pattern of impairments that occur after selective DG lesions sheds light on the role of the DG in hippocampal processing. Most lesion studies have taken advantage of the fact that intrahippocampal injections of colchicine cause a fairly selective destruction of the granule cells of the DG, but other techniques, such as neonatal X-ray irradiation and adrenalectomy, which also cause a similar pattern of damage, have also been used. Perhaps the most consistent finding after DG lesions in rodents has been a severe impairment in the acquisition of the reference memory task in the Morris water maze (Sutherland et al., 1983; McNaughton et al., 1989; Conrad and Roy, 1993; Xavier et al., 1999). Working memory versions of the task are also affected, although perhaps to a lesser extent (Xavier et al., 1999; Jeltsch et al., 2001). Reference and working memory performance in the radial arm maze are also compromised (McNaughton et al., 1989; Jeltsch et al., 2001). In some more recent studies, DG lesions were also found to cause impairment in a delayed-matching-to-place task, in a temporal task (Costa et al., 2005), and in a task which required the detection of metric distance change between objects (Goodrich-Hunsaker et al., 2005). Recently, there have been some attempts to test more directly the proposed contributions of the DG to hippocampal function. In particular, Gilbert et al. (2001) designed a short-term spatial

memory task where the proximity of relevant locations could be varied systematically. They found that DG lesions impaired performance at small, but not at large separations, consistent with the role of the DG in spatial pattern separation. Lassalle et al. (2000) examined the consequences of selective and reversible inactivation of mossy fiber synapses in CA3 in mice during various stages of a reference memory task in the Morris water maze. They found that the mossy fiber input from DG to CA3 was essential during learning, but not during the retrieval phase of the task, or in the period directly following learning (the early stages of consolidation). Similarly, Lee and Kesner (2004) attempted to distinguish encoding and retrieval deficits during the acquisition of a navigation task in the Hebb-Williams maze following lesions of various type. They found that lesions of the DG impaired learning, but not retrieval, in this task; conversely, lesions of the perforant path input to CA3 affected retrieval, but not learning. In summary, the available behavioral data from rodents with lesions to dentate granule cells or their mossy fiber output are generally consistent with a crucial role of the DG in providing input to area CA3 during the acquisition of allocentric spatial information, and provide some support for a more specific function in (spatial) pattern separation. So what are the morphological peculiarities of the DG that support its role in orthogonalization and make it indispensable for proper CA3 function? Morphological arguments — heterogeneous terminal types of the mossy fibers Let’s see first how Cajal described of the unusual terminals types of the mossy fibers: ‘‘ythere arise either short and thick divergent appendages or quite long fine threads that end in a swelling. Thus, we reproduced here the arrangement (although less distinctly) that we described in certain branched fibers of the cerebellum the mossy fibers. Therefore without further ado let us apply the same name to the axons of the granules of the fascia dentata.’’

582

It is clear from the above description that the investigator who named the mossy fibers by examining Golgi-stained material clearly identified that this peculiar fiber system has more than one terminal type. The well-recognizable giant endings (the large mossy terminals) gave rise to thin filamentous structures that ended in terminal-like swellings. Apparently the name ‘‘mossy fiber’’ is based on the presence of these filopodial extensions. Later electron microscopic work identified that filopodial terminals indeed establish asymmetrical synapses (Amaral, 1979; Claiborne et al., 1986) (Fig. 1). In addition, a third terminal type has been described, small ‘‘en passant’’ boutons, which resemble the most the conventional axonal varicosities of cortical pyramidal cells (Claiborne et al., 1986) (Fig. 1). Together the two smaller mossy terminal types far outnumber their larger counterpart, but still, the total number of terminals is actually very low (Claiborne et al., 1986; Acsady et al., 1998). A single granule cell has no more than
200 terminals, which is at least two orders of magnitude less than the number of varicosities along the axonal arbor of a cortical pyramidal cell. How can these few but variable terminals account for the code conversion and efficient transmission as outlined above?

Unique features for code conversion The computational studies summarized above suggested that memory formation in the hippocampus may be a two-step process: the sparsification of the entorhinal signal by granule cells, followed by an association of the now sparse code in the CA3 network. Modeling studies suggest that the ‘‘winner-take-all’’ method of sparsification requires the recruitment of strong feedback inhibition. According to this scheme, the output of a small population of granule cells that become active in a given environmental context (e.g., during a couple of theta cycles as the rat passes through their place fields) activate GABAergic interneurons which exert fast and strong feedback inhibition on the somata and dendrites of the non-coding granule cells, shunting their entorhinal inputs and precluding their firing. As a result, synaptic plasticity only

takes place at the perforant path input of the coding (granule) cells. However, feedback inhibition is well-known in all cortical regions. Is there any reason to suppose that feedback inhibition is more powerful in the DG than in other cortical regions, which makes it especially useful for pattern separation? In cortical regions, interneurons constitute 10–20% of all the neurons. Cortical pyramidal neurons innervate their postsynaptic principal and interneuron targets in a quasi-random manner, i.e., the incidence of the targets is determined by the relative distribution of the neuron types (Gulyas et al., 1993; Sik et al., 1993). Thus, the estimated ratio of interneurons among the postsynaptic targets of cortical pyramidal cells is around 10%. In the DG granule cells have axon collaterals only in the hilus below the granule cell layer not in stratum granulosum or stratum moleculare (Claiborne et al., 1986; Acsady et al., 1998). As a consequence only inhibitory neurons having somata and/or dendrites in the hilus can participate in feedback inhibition. Close to 50% of the hilar neurons are GABAergic (Houser and Esclapez, 1994) which suggests that inhibitory cells may be abundant among the postsynaptic targets of granule cells. The 5–8 hilar collaterals of the granule cells possess around 7–12 large mossy boutons and 102–147 small terminals (filopodial and en passant boutons) (Acsady et al., 1998). The postsynaptic targets of the small terminal types are almost exclusively interneurons, whereas targets of the large mossy terminals are mainly excitatory mossy cells (Acsady et al., 1998) (Fig. 2). Thus, due to the surprising target selectivity of granule cell terminal types, interneurons may constitute up to 90% of the postsynaptic targets of the mossy fibers in the hilus, in contrast to the 10–20% GABAergic targets in other cortical regions. Many of these hilar neurons provide feedback inhibitory control of the granule cells, suggesting that proportionally stronger feedback inhibition is recruited here than in other cortical regions. A characteristic cell type of the hilus, the somatostatin-immunoreactive interneuron, provides a good example of the strong recurrent inhibition operating in this system. The

583

Fig. 1. Electron micrographs of different terminal types along the mossy fibers in the CA3 region (A–C, E) and of a CA3 pyramidal cell terminal (D) for comparison. All electron micrographs have the same magnification. A, B, A small en passant terminal establishes a single asymmetrical synapse on a dendritic shaft, showing the characteristic long perforated postsynaptic density of small terminal types on interneurons (arrows). C, A filopodial extension of a mossy fiber terminal forms a synapse (arrow) with a Substance P receptor-immunoreactive interneuron. D, A CA3 pyramidal cell establishes asymmetrical synapse on a simple spine of a CA1 pyramidal neuron. E, A large, double-headed mossy fiber terminal forms multiple contacts (arrows) with thorny excrescences of a CA3 pyramidal cell. All active zones converged on the same pyramidal cell. The individual release sites are short. Scale bars: A–D, 0.5 mm; E, 1 mm. Reprinted with permission from Acsady et al., 1998, Society for Neuroscience.

somatostatin-containing hilar neurons restrict their entire dendritic arbor to the hilus and innervate the dendritic segment of granule cells in the zone where entorhinal afferents terminate (hence

they are also called HIPP cells, i.e., HIlar Interneurons with Perforant Path associated axon terminal) (Han et al., 1993; Sik et al., 1997). Unlike many other interneuron types that

584

Fig. 2. Filopodial extensions of mossy fiber terminals are specialized to innervate GABAergic cells. Artistic rendition of two large mossy terminals, each equipped with four filopodial extensions (large arrowheads). The mossy fibers were labeled by intracellular injection of biocytin into two neighboring granule cells. All filopodial terminals were examined in the electron microscope (not shown) and all contacted the dendrites or spines of altogether six GABAergic neurons. Four of the GABAergic neurons were identified by their Substance P receptorcontent and two of them by ultrastructural characteristics. Five of the six postsynaptic interneurons were spiny cells. Arrows point to the main axons. Reprinted with permission from Acsady et al., 1998, Society for Neuroscience.

are characterized by a smooth dendritic surface, the dendrites of HIPP cells are densely covered with thousands of long thin, elaborated spines (Baude et al., 1993). In contrast to the compartmentalized spines of the pyramidal cells, which usually receive a single terminal, these spines are contacted by multiple asymmetrical synapses (up to 8–10) (Acsady et al., 1998). Thus, these spines increase the total synaptic input of the HIPP cells enormously. Small terminals of granule cells have been described to contact these spines (at least 30% of the total synaptic output of granule cells contacts HIPP cells) (Acsady et al., 1998), whereas the axons of CA3 pyramidal cells that project back to the hilus selectively avoided HIPP cells (Wittner et al., 2006), suggesting that the majority of the excitatory input of these cells originates from granule cells. Since the number of contact between a granule cell and a HIPP cell is only 1 or 2 these morphological data indicate the convergence of

several thousand granule cells on any one of these interneurons. The axons of HIPP cells terminate in the same layers as the perforant path, and therefore are in a critical position to modulate the information transfer from the entorhinal cortex to the DG (Han et al., 1993). A single HIPP cell may form an unusually large number of axon terminals in the molecular layer (up to 80,000 compared to 5000–10,000 in the case of a basket cell), the vast majority of which innervate granule cells (Sik et al., 1997). The little data available about the activity of HIPP cells suggest that these neurons are not fast-firing cells; rather, their activity is principally driven by the firing pattern of granule cells, a key feature for appropriately timed feedback inhibition (Buckmaster and Schwartzkroin, 1995). Another unique morphological feature of the connectivity in DG is the lack of interaction among the major inhibitory basket cell classes and among basket cells and other interneurons (Acsady et al., 2000). In other hippocampal regions as well as in other cortical regions, basket cells densely innervate other basket- and nonbasket-type interneurons (Sik et al., 1995; Cobb et al., 1997; Tamas et al., 1998) forming an interacting local GABAergic network. For example, in the CA1 region the somatic region of parvalbuminpositive basket cells is contacted by more GABAergic terminals than the somatic region of pyramidal cells (Gulyas et al., 1999; Megias et al., 2001). In sharp contrast, GABAergic cells in the hilus of the DG receive, on average, 15–40 times less input from local basket cells than mossy cells, the principal excitatory cell type of this region (Acsady et al., 2000). Since GABAergic cells inhibit each other, this connectivity pattern suggests minimal disinhibitory influence in the hilus and different GABAergic network dynamics. But what are the physiological properties of the granule cell–interneuron synapses? Examination of synaptic transmission at the granule cell–basket cell synapses demonstrated very fast kinetics: the postsynaptic conductance of the unitary current demonstrated submillisecond rise and decay (Geiger et al., 1997). This effect was largely attributed to the high synchrony of transmitter release and the rapid time course of AMPA receptor

585

deactivation. The fast postsynaptic response allows rapid activation of feedback inhibition, which supports a role in pattern separation. These data suggest that the basic principles of connectivity between excitatory and inhibitory neurons in the DG are significantly different from those in any other cortical region. The peculiar arrangement of excitatory and inhibitory connections in the DG suggests an unusually strong recruitment of inhibition that can be used to suppress the activity of granule cells in a competitive manner. As a result, only granule cells with the strongest entorhinal excitatory drive will participate in the information transfer to CA3. This small population of active granule cells could effectively prevent large granule cell populations from reaching firing threshold via the strong feedback inhibitory system outlined above. Only those entorhinal inputs will be potentiated which contact active granule cells. This would further strengthen the competitive process, which could lead to the sparsification of the entorhinal signal resulting, e.g., in the sharp and focused place fields of the granule cells which are in sharp contrast to the grid-like entorhinal signal (see below). The next step is to faithfully transmit this recoded cortical signal to the associational station, the CA3 region, and to ensure that only the activated subset of CA3 pyramidal cells would be included in the autoassociation network responsible for memory storage.

Unique features for efficient and sparse transmission The giant mossy terminals display all the morphological features of a classic ‘‘detonator’’ or ‘‘driver’’ type terminal. No other cortical synapse is comparable to them, but several subcortical structures (e.g., thalamus, cerebellum) utilize similar synaptic arrangements to secure faithful synaptic transmission (Sherman and Guillery, 1998). A single large mossy terminal may establish up to 30–40 release sites, all converging on the proximal dendrite of a single postsynaptic pyramidal cell (Chicurel and Harris, 1992; Acsady et al., 1998). One would predict that the short electrotonic distance from the soma would maximize the

efficacy of the input in driving the postsynaptic cell to threshold. Recent in vitro and in vivo data confirm this assumption. Monosynaptic AMPA/ kainate receptor-mediated EPSCs from granule cell–pyramidal cell pairs had a mean peak amplitude of 163.0723 pA at 70 mV in organotypic slice cultures which displayed morphological properties similar to the in vivo condition (Mori et al., 2004). Strong excitatory action has been described earlier between granule cells and their excitatory targets in the hilus, the mossy cells (Scharfman et al., 1990). In the in vivo anesthetized preparation, repetitive firing in a single granule cell reliably induced action potentials in monosynaptically connected CA3 pyramidal cells (Henze et al., 2002). It has to be emphasized that such strong coupling is extremely rare among excitatory cells in cortical circuits. The general rule is that a large number of excitatory inputs have to be simultaneously active to reach postsynaptic spike threshold. But how sparse is the detonator signal in morphological terms? The number of large mossy terminals along the single unbranching axon of granule cells within area CA3 is very low (average: 12.3; range: 10–18) (Acsady et al., 1998). If one considers that a single CA3 pyramidal cell may have up to 60,000 terminals, it is straightforward to conclude that granule cells are a specific cortical cell type designed to transfer the sparse code generated in DG very effectively to only a restricted set of postsynaptic pyramidal cells. Two recent studies described additional surprising features of the mossy fibers. These features not only support faithful transmission through this pathway, but also demonstrate the computational power of the axon terminals in unexpected ways. The first study (Engel and Jonas, 2005) describes the active properties of the large mossy terminals, which facilitates reliable transmission of high frequency trains of action potentials. Apparently, mossy terminals have a very high density of specialized Na+ channels with faster activation and inactivation kinetics than somatic Na+ channels. These Na+ channels enable reliable action potential invasion into large mossy terminals and increase presynaptic Ca2+ influx, resulting in up to 16-fold increase of transmitter release. In addition,

586

modeling studies suggest that this active property of the terminals is absolutely necessary to induce Ca2+ influx into the filopodial extensions to trigger glutamate release at the mossy fiber–interneuron synapse. This mechanism may underlie the high release probability of the filopodial connection compared to the mossy fiber–CA3 pyramidal cell synapse (Jonas et al., 1993; Lawrence et al., 2004). The second study (Alle and Geiger, 2006) demonstrates that mossy fibers transmit not only action potentials, but also postsynaptic potentials originating in the soma-dendritic compartment. These ‘‘excitatory presynaptic potentials’’ alter the transmitter release of the subsequent action potential (within a 10–20 ms delay), thus enabling the axon to integrate subthreshold and suprathreshold signals provided they are temporally proximal. In sum, mossy terminals fulfill all criteria for a classical, sparse ‘‘detonator’’ synapse. The strong granule cell–pyramidal cell connection, however, poses two major problems for the autoassociative network of the CA3 region. Problem 1. Spontaneously active granule cells, which represent only background noise, may induce irrelevant spiking of CA3 pyramidal cells, resulting in non-coding CA3 networks which may overlap with the coding networks. Problem 2. Spontaneously active CA3 pyramidal cells may generate action potentials coincident with the CA3 spikes evoked by granule cell firing, which represent the given environment. These noncoding EPSPs will be also strengthened in the autoassociative CA3 network, which would ruin the orthogonalized, sparse code. A first solution to Problem 1 is that granule cells have very low spontaneous firing rates (0.1–0.01 spike/s) (Jung and McNaughton, 1993), probably due to their hyperpolarized resting membrane potential (80 mV in vivo) (Penttonen et al., 1997) and/or strong inhibitory control. Still, if we consider that there are one million granule cells per hippocampus in rodents (Seress, 1988) and we take the low end of their spontaneous discharge frequency (0.01 Hz), we can calculate that at least 10 granule cells will fire in any one millisecond,

activating
120 CA3 pyramidal cells, each of which has around 30,000 terminals within the CA3 (Li et al., 1994). Thus, other solutions are needed to solve Problem 1. Similar to the hilus, the number of small mossy fiber terminals in the CA3 region exceeds the number of large mossy fiber terminals by a factor of at least four (Acsady et al., 1998). As in the hilus, the small terminals selectively innervate inhibitory cells. All examined inhibitory cell classes (perisomatic, dendritic, and interneuron-selective) were among the postsynaptic elements of granule cells. Since a single mossy fiber rarely innervates a postsynaptic interneuron via multiple contacts (Acsady et al., 1998), granule cell firing activates at least four times as many inhibitory as excitatory cells. Physiological data confirm strong activation of this feedforward inhibitory circuit. In paired recordings, a single action potential in the granule cell induced a biphasic response in the postsynaptic CA3 pyramidal cell: a brief EPSC followed by a pronounced IPSC (Mori et al., 2004). The direction of the summated charge transfer was outward, indicating a net inhibitory synaptic response. The authors calculated that approximately four interneurons fired together to evoke the measured inhibitory responses, which corresponds well to the anatomical data. Reliable activation of interneurons was also observed in vivo. Multiple granule cell spikes induced firing in monosynaptically coupled interneurons with quite high probability (Henze et al., 2002) (Fig. 3D), suggesting that although the granule cell–interneuron contact rarely contains multiple release sites, the single active zone of the small terminals is still efficient enough to induce postsynaptic firing of the GABAergic cell. Recruitment of more inhibitory than excitatory cells suggests that the net effect of the excitatory mossy fiber system on the CA3 pyramidal cell population is, counter-intuitively, inhibitory. A peculiar EEG transient, the dentate spike, can be utilized to demonstrate the impact of granule cell activation on the postsynaptic cell populations along the dentate–CA3 axis during a ‘‘natural’’ stimulus. Dentate spikes are short-duration, largeamplitude field potentials caused by synchronous activation of the entorhinal input, which occur during behavioral immobility and slow-wave sleep

587 B

D 0.12

C

Probability

0.03

0.08 0.02 0.04

0.01 0

0.00 0

4

6

8

10

100 Hz

E 1.0

0

2

4

6

8

10

F 0.3

0.6

Probability

60 Hz 40 Hz

0.8

Probability

2

20 Hz

0.4

10 Hz

0.2

0.2

0.1

0.0 0.0 0

0.2

0.4

0.6

0.8

Time (s)

1

1

3

5 7 Spike number

9

Fig. 3. Spike transmission dynamics between a granule cell and its interneuron and pyramidal cell targets in CA3c in vivo. (A) Camera lucida reconstruction of the extracellular electrode track and biocytin-labeled granule cell. Inset, a higher-power view of the mossy fiber axon near the probe track. Arrowheads, mossy fiber boutons. (B) Superimposed (n ¼ 60) intracellularly evoked action potentials in a granule cell (bottom traces) and simultaneously recorded extracellular units (filtered 0.8–8 kHz). Note the time-locked response of a putative pyramidal cell to the granule cell action potentials. (C–D) Cross-correlograms between the evoked granule cell action potentials and the activity of a putative CA3c pyramidal cell (C) or interneuron (D). The values are shuffle corrected and expressed as probability (number of unit spikes per bin/total number of granule cell spikes). Arrowhead, peak time of the granule cell action potential. (E) Representative results of the effect of intratrain frequency on spike transmission probability for a putative pyramidal cell. (F) Spike transmission probability (shuffle-corrected probability of spike in 6 ms following granule cell spike) as a function of spike number in evoked 100 Hz train (7s.e.m). Solid line, putative interneurons (n ¼ 24); dotted line, putative pyramidal cells (n ¼ 21). Scale bars, (A) 50 mm; inset 20 mm; (B) 1 ms, 25 mV, 75 mV. m, molecular layer; g, granule cell layer; h, hilus; IC, intracellular electrode track; EC, extracellular electrode track.

(Bragin et al., 1995). Extracellular recordings during dentate spikes demonstrated increased unit activity in hilar neurons (many of which are GABAergic), but suppressed multiunit activity in the CA3 region (Bragin et al., 1995). Intracellular studies confirmed depolarization of granule cells and hilar interneurons but hyperpolarization in CA3 and CA1 pyramidal cells (Penttonen et al.,

1997). It is worth mentioning that this pattern of activity is in sharp contrast to the neuronal behavior that can be observed during the other major excitatory field transient in the hippocampus, the sharp wave, which originates in CA3. Since in this case there is no inhibitory ‘‘barrier’’ comparable to the dentate GABAergic network which could block the spread of excitation, sharp waves

588

propagate not only to the CA1 region but also to parahippocampal cortical regions via polysynaptic activation (Chrobak and Buzsaki, 1994, 1996). Most recently the efficacy of the dentate inhibitory ‘‘barrier’’ was demonstrated by comparing the correlation of intracellular activity of neurons in various cortical fields with the UP and DOWN states in the EEG during slow cortical oscillation (Isomura et al., 2006). Surprisingly, all cortical regions (including entorhinal cortex and DG) changed intracellular activity coherently with the EEG states with the exception of CA3 region. In contrast to basically all cortical cells CA3 pyramidal cells were not active during the UP states. Apparently the inhibitory network launched by the dentate during the UP states is able to shield CA3 even from the most synchronous excitatory events of the cortical mantle. In summary, anatomical and physiological data provide convergent evidence for the conclusion that the output of the granule cells activates an unusually strong feedforward inhibition to area CA3. Since even a single basket cell is able to block action potential generation in a large number of pyramidal cells (Miles et al., 1996), feedforward inhibition will effectively reduce the number of active CA3 pyramidal cells during dentate activation. In addition, in the CA3 region, specialized interneuron types exist which restrict their dendritic (Gulyas et al., 1991) or axonal (Vida and Frotscher, 2000) arbor to stratum lucidum of CA3, suggesting selective control of this pathway. Thus, the likely solution of Problems 1 and 2 is that ‘‘background’’ or ‘‘non-coding’’ EPSPs and action potentials in CA3 are actively inhibited during dentate–CA3 information transfer by the strong feedforward inhibition. In this way only the small population of CA3 cells receiving the decorrelated sparse dentate signal will be active, and following the Hebbian rule, only the recurrent synapses between the activated CA3 cells will be potentiated. According to the modeling studies (see above) the matrix of potentiated synapses will represent the memory trace. However, what happens to the EPSPs generated by the spontaneous activity of CA3 pyramidal cells immediately before the dentate input arrives? The EPSPs among pyramidal cells are quite slow

(half duration, 27 ms) (Miles and Wong, 1986) and accidentally the mossy fiber input can arrive together with their peak, thus these early background (unwanted) EPSPs can be potentiated before the disynaptic feed forward inhibition arrives. A recent report provides a possible solution for this problem (Kobayashi and Poo, 2004). This study describes LTP at the recurrent synapses of the CA3 network induced by paired stimulation of mossy fiber input and the associational/commissural input. First, this study provides direct and elegant evidence for the role of mossy fibers in changing the synaptic strength of the autoassociative CA3 matrix. Second, an interesting observation from this study helps resolve the problem of early EPSPs. The results demonstrate that the potentiation of associational input depends on the relative timing of mossy fiber and associational spike trains. The potentiation was smaller if additional associational spikes were added before the paired stimulation, compared to the protocol that added spikes after the paired (associational/mossy fiber) pulses. The effect depended on mGluR1 activation. Thus, apparently the system favors the potentiation of CA3 EPSPs arriving coincidentally or after the dentate signal. Since the mossy fiber input is able to induce CA3 spiking in vivo, these EPSPs will mostly represent the spiking of CA3 pyramidal cells evoked by the sparse, orthogonalized dentate input. In this way only the associational EPSPs representing dentate activity will be potentiated, but EPSPs arriving earlier, representing spontaneous CA3 activity, will not. Target-dependent plasticity — the meaning of various terminals types If we consider the strong feedforward inhibition operating along the dentate–CA3 axis the following, third problem arises: Problem 3. How to overcome the strong feedforward inhibition when the specific dentate pattern has to activate the pyramidal cell? Apparently the solution to Problem 3 is that short- and long-term plasticity at large and small types of mossy fiber ending are different.

589

The physiological data clearly demonstrate that the small terminals are not only distinct morphological units, evolved to contact a large number of interneurons, but discrete compartments which harbor distinct molecular machinery for plasticity different from their giant cousins. Small and large mossy terminals display different types of short-term plasticity. The giant mossy terminal–pyramidal cell connection express very strong short-term facilitation (Salin et al., 1996; Toth et al., 2000), reaching up to threefold amplitude increase on average at the fifth EPSC in case of 20 Hz stimulation. (This is highly unusual in the case of giant ‘‘driver’’-like terminals; e.g., in the thalamus excitatory terminals with a similar synaptic arrangement show strong depression in relay cells; Reichova and Sherman, 2004). In contrast the small mossy terminal–interneuron connection shows short-term depression after repetitive mossy fiber stimulation in about 50% of the interneurons. Others show modest facilitation at 20 Hz (Toth et al., 2000), but at 40 Hz, all responses are depressed after the 8th action potential (Mori et al., 2004). The third synapse in the feedforward inhibitory circuit, the interneuron–CA3 pyramidal cell contact, expresses pronounced short-term depression (Mori et al., 2004) at all frequencies tested. This variability in short-term plasticity at different mossy fiber synapses favors conditions for a single granule cell action potential to induce weaker excitation, but relatively strong feedforward inhibition. Repetitive firing, however, rapidly increases the effect of excitation and at the same time decreases the efficacy of inhibition. Thus, the net effect of mossy fiber activation on CA3 pyramidal cells depends heavily on the frequency of granule cell firing. The system appears to act as a high-pass filter, where spike transmission from granule cell to pyramidal cell is blocked at low frequencies but favored when the frequency of granule cell discharge increases. Recently this assumption was tested directly in in vitro paired recordings of granule cells and pyramidal cells (Mori et al., 2004). The postsynaptic potentials evoked in pyramidal cells by granule cell stimulation were measured at increasing frequencies (10–40 Hz). The inhibitory dominant PSPs

observed during low frequency trains switched to excitatory dominant PSPs at high frequencies. At 40 Hz EPSPs dominated the response already after the third granule cell action potentials, whereas at 10 Hz the response remained inhibitory even at the 15th action potential. Frequency-dependent facilitation of mossy fiber transmission was demonstrated in monosynaptically coupled granule cell — pyramidal cell pairs in vivo as well (Henze et al., 2002) (Fig. 3). The probability of postsynaptic pyramidal cell spikes rapidly increased with increasing granule cell firing and reached 0.8 at 100 Hz (Fig. 3E), which is an extremely high value in cortical circuits and results in an almost one-to-one relay of the granule cell activity. Within the spike train the probability of CA3 pyramidal spikes increased with the number of presynaptic granule cell spikes. The maximum spike transmission probability was reached after the 4–5th spike (Fig. 3F). In contrast, in the case of interneurons, the probability of transmission did not increase with an increasing number of presynaptic spikes at 100 Hz and remained lower than that of the pyramidal cell (Fig. 3F). These in vitro and in vivo data indicate that differential shortterm plasticity in this feed-forward circuit result in a frequency-dependent shift of the polarity of postsynaptic response. In the freely moving condition, granule cells have very low spontaneous firing rates, which can rapidly increase to 40 Hz (Jung and McNaughton, 1993) as the animal enters the place field of the neuron. As a consequence of the frequencydependent switch from excitation to inhibition described above, the probability of spike transmission between the DG and CA3 is very low at low firing rates, despite the ‘‘detonator’’ nature of the synapse. When granule cells code the specific information of the environment, the probability of spike transmission to CA3 becomes very high. Thus, granule cells can act as a ‘‘conditional detonator’’ as suggested by Henze et al. (2002). This mechanism solves both the problem of filtering out non-coding spontaneous activity (Problem 1), and resolves the issue that strong feedforward inhibition must be overcome when specific information must be transmitted from dentate to CA3 (Problem 3).

590

The delicate frequency-dependent balance between excitation and inhibition substantially increases the computational power of the mossy fibers. It raises the possibility that they not only participate in the faithful transmission of an orthogonalized dentate signal but they themselves participate in creating the non-overlapping representations in the CA3 region. Recent data suggest (Leutgeb et al., 2007, see below) that the representation of the environment is more orthogonalized in CA3 than in the DG, which questions the role of dentate as the sole contributor to this process. Due to its high-pass filter nature, the dentate–CA3 circuit may refine the dentate code, and, e.g., participate in creating the single receptive field observed in CA3 place cells as opposed to the multiple place fields of dentate place cells. The critical variable in this process will be the exact firing pattern of the dentate granule cell. For instance, assume that a granule cell fires at 40 Hz in one of its receptive fields but only 10 Hz in the other. Based on the data discussed above, only the higher firing rate will induce firing in the CA3 pyramidal cell, the lower activity will be filtered out, and the CA3 pyramidal cell will display a single receptive field as a result. What about long-term changes? One presentation of the dentate code may not be sufficient to induce long-term changes in the CA3 recurrent network. Multiple presentations (e.g., crossing the place field several times) or autonomous replay of the memory trace in absence of the original condition, may be necessary to induce long-term plasticity. Replay of a given firing pattern may occur during different EEG states, as has been shown for theta and sharp wave activity (Nadasdy et al., 1999), or during the subsequent sleep episodes (Wilson and Mcnaughton, 1993). Tetanic stimulation of mossy fibers induces LTP in pyramidal neurons, but is either without effect, or it induces depression at synapses onto interneurons (Maccaferri et al., 1998). Since the mossy fiber–LTP onto pyramidal cells critically depends on cAMP, the effect can be explained by the absence of the adenylyl cyclase-cAMP cascade from the filopodial and ‘‘en passant’’ small terminals. As a result of this differential LTP, the critical frequency at which inhibition switches to

excitation may change and/or the steepness of the high-pass filter cutoff may increase. In this way, presynaptic mossy fiber–LTP may help the orthogonalization process performed by the dentate–CA3 circuit and creates a good opportunity for the faithful activation of the same CA3 circuit in case of repetitive presentation. Finally, morphological plasticity of the small terminals provides yet another way to fine-tune the balance of excitation and inhibition in the mossy fiber pathway. The structure of the filopodial terminals strongly resembles rapidly advancing and retracting axonal filopodia observed during axonal development and synapse formation. Recent studies of slice cultures indeed demonstrated that over one-third of the filopodia are highly active (De Paola et al., 2003; Tashiro et al., 2003). In addition, in mature slices, approximately 9% of the small en passant boutons were also labile (halflife, approximately 1 day), in contrast to the stability of large terminals. Interestingly, in mature cultures, the total number of synapses remained stable in the presence of substantial turnover of individual terminal structures. Motility of the filopodial extensions was observed not only in slice cultures but also in the acute whole-mount hippocampal preparation and acute hippocampal slices. This actin-based motility can be regulated by brain-derived neurotrophic factor (BDNF), AMPA and/or kainate receptors, in a cAMP dependent manner (De Paola et al., 2003; Tashiro et al., 2003). These data strongly suggest that target selectivity of the small terminal types of the mossy fiber is based on the dynamic morphological properties of these axonal elements. In addition, they suggest that even in mature animals activity change, traumatic injury or cell loss may induce rapid remodeling of the mossy fiber-tointerneuron connections from granule cells. Indeed, ischemic damage, which results in the loss of hilar and stratum lucidum interneurons, dramatically reduces the number of filopodia (Arabadzisz and Freund, 1999). In summary, the mossy fiber pathway maximally utilizes the opportunity to differentially regulate its postsynaptic partners. Selective innervation of pyramidal cells and interneurons by distinct terminal types creates a favorable condition to

591

differentially regulate short-term and long-term plasticity and the motility of various mossy terminal types. The ‘‘bottom-line’’ of this highly dynamic system appears to be the fine-tuning of the balance between the excitation evoked by the large ‘‘detonator’’ terminals and the feedforward inhibition activated by the small terminals. This will determine exactly which dentate firing patterns will induce permanent modification of the autoassociational network of the CA3 region. The final test for all predictions regarding dentate function is the examination of firing activity in the freely moving animal.

Neuronal activity patterns in vivo and the processing of entorhinal input by the dentate gyrus Compared to the vast amount of data available on spatial (and non-spatial) representations in area CA1 of the hippocampus, relatively little is known about the in vivo firing properties of neurons in the DG under various behavioral conditions. During exploration, granule cells display spatially selective activity, and their firing behavior is, at least under most conditions studied so far, qualitatively quite similar to that of pyramidal cells (place cells) in areas CA1 and CA3. In particular, in the radial arm maze, granule cells have directional place fields, although the average number of subfields is somewhat higher and the average size of these subfields is slightly smaller than in CA3 place cells (Jung and McNaughton, 1993). The firing activity of dentate granule cells is modulated by local field potential oscillations in the theta frequency range, and the timing of individual action potentials changes during traversals of the place field similar to phase precession observed in CA1 pyramidal cells (Skaggs et al., 1996). When different spatial cues were put in conflict by manipulating the environment, sudden coherent transitions (known as reference frame shifts) could be observed in the activity of the set of simultaneously recorded granule cells, also analogous to the behavior of the CA1 cell population under these conditions (Gothard et al., 2001). In an experiment where an explicit attempt was made to identify non-spatial as well as spatial responses, a subpopulation of

granule cells showed position-selective or positionindependent reward site responses, whereas another population showed pure place responses (Tabuchi et al., 2003). When rats are allowed to explore a novel environment for the first time, both dentate granule cells and CA1 pyramidal cells acquire distinct spatial preferences within the first few minutes (Nitz and McNaughton, 2004). However, concurrently recorded interneurons showed very different behavior in the two regions: while most CA1 interneurons transiently decreased their activity while the animal was exploring a novel environment, the majority of interneurons in the DG significantly increased their activity during the same epoch. These data are congruent with the role of strong feedback inhibition in establishing sparse and decorrelated output in the DG. In order to understand the nature of the computations performed by the DG, it is essential to have an accurate description of the cortical input it receives. The last few years have witnessed a major revolution in our understanding of spatial representations in entorhinal cortex. Until a few years ago, existing data on EC representations indicated that, although neurons in EC were spatially selective, their place fields were much larger and less clearly defined than those in hippocampal areas (Quirk et al., 1992; Frank et al., 2000). Indeed, this was perhaps the most direct experimental evidence for the assumption that sparse, orthogonal, ‘‘hippocampal-type’’ representations are created first in the DG. However, when spatial firing patterns were measured in the part of medial entorhinal cortex (mEC) which projects to the dorsal HC (where place fields are normally recorded), much smaller place fields, similar in size to corresponding hippocampal place fields, were found (Fyhn et al., 2004), while areas of mEC with large place fields projected to more ventral parts of the HC, which itself was found to have large place fields (Maurer et al., 2005). From these data, it appeared that there was in fact no major transformation of spatial representations from EC to the DG (and the rest of the hippocampus), although subtler differences (especially in response to environmental manipulations) could not be ruled out. However, it soon emerged that if spatial firing patterns

592

are recorded over a larger spatial scale, entorhinal and hippocampal representations are again fundamentally different. In particular, EC place fields were found to repeat periodically at the vertices of a regular hexagonal lattice, and different EC ‘‘grid cells’’ differ in the center location (phase), orientation, and scale of the grid (Hafting et al., 2005). In contrast, hippocampal place cells have at most a few discrete place fields even in these larger environments, and their fields typically do not have any special geometrical relationship. Thus, it currently appears that hippocampal spatial representations are in fact different in nature, and in particular, much sparser over a large environment than representations in the (medial) entorhinal cortex, and a transformation (in fact, a pattern separation operation) by the DG is still required. Indeed, if we consider two locations in the environment which are separated by the grid period (assumed to be relatively invariable) in the area of entorhinal cortex which projects to a given part of the HC, the activity of the grid cell population will be quite similar, while the hippocampal activity pattern will be distinct, reflecting the outcome of some kind of pattern separation process. In fact, the need for pattern separation becomes even more obvious if we consider two environments. The root of the problem is the observation that the relative grid parameters (phase and orientation) of different grid cells appear to be fixed in all environments, so, in principle, there must be corresponding locations (and orientations) in two distinct environments where the activity of the entire grid cell population is identical. This raises an obvious question: how distinct spatial representations of the two environments could be formed in the HC based on a single entorhinal grid cell representation. One possible answer to this question is based on the fact that the entorhinal grids are not perfectly regular; for instance, peaks in the grid vary in amplitude, so that, in principle, there could be sufficient spatial information present in the amplitudes to distinguish different environments. Another, perhaps more plausible explanation is that the hippocampus (and, in particular, the DG) receives, in addition to input from medial EC, where grid cells are located, input from lateral EC, where neuronal firing patterns have a lower spatial

information content, but instead, carry more information about relevant objects (Hargreaves et al., 2005). Such object-based information is probably sufficient to disambiguate different environments, and create distinct codes in the HC. Hippocampal pattern separation between environments of varying degrees of similarity has recently been investigated in a series of experiments (Leutgeb et al., 2004, 2005a,b), and the initial analysis in areas CA3 and CA1 has now been partially extended to mEC and the DG (Leutgeb et al., 2007; Hafting et al., 2006). By analyzing the spatial firing fields of neurons within environments of varying degrees of similarity (manipulations included switching between different recording locations, as well as changes in the shape and/or the color of the enclosure), Moser and colleagues found that firing patterns of neurons in all hippocampal areas distinguished between different environments much better than those of grid cells in mEC. Essentially no pattern separation was detected in EC, as population firing patterns in enclosures of different shapes, colors, or even in different rooms were not significantly different (any observed changes occurred coherently in all recorded neurons; Hafting et al., 2006). However, they also found that the basic properties of pattern separation were different between different subfields of the hippocampus. In area CA3, the phenomenon termed ‘‘rate remapping’’ was observed when the shape of the enclosure was varied continuously: the firing rate pattern of the active cell population changed continuously as the environment was gradually transformed, while place field locations remained constant (Leutgeb et al., 2005a). In the end, environments of clearly distinct shapes (e.g., circle vs. square) activated CA3 pyramidal cell populations with relatively little overlap — indeed, when recordings were made in two different rooms, the two populations of active CA3 neurons appeared to be chosen independently, a phenomenon referred to as ‘‘global remapping’’ (Leutgeb et al., 2005b). Interestingly, preliminary data from recent experiments have revealed a radically different type of pattern separation in the DG. First, unlike in CA3, the same dentate cells were found to be active in different environments (Leutgeb et al.,

593

2007). Second, DG neurons typically had multiple place fields even in a single environment, consistent with earlier data (Jung and McNaughton, 1993). Finally, the peak firing rate within any one place field of a single cell varied with even small changes in the shape of the enclosure, independently of rate changes in other fields of the same cell (Leutgeb et al., 2007). These results show that the DG performs a pattern separation operation on its entorhinal inputs, but pattern separation appears to work differently from what was previously assumed. The observation that the proportion of simultaneously active cells is much lower in the DG than in EC (a fact that was confirmed by the data described above, and independently by another recent study, which measured immediate early gene expression in the DG following spatial experience; Chawla et al., 2005) suggested that different environments might activate different sets of neurons in the DG, thereby implementing a particularly efficient type of pattern separation. The results of Moser and colleagues now suggest that this might not be the case, and different environments (as well as different locations within these environments) might be encoded by different firing rate patterns in essentially the same population of active DG neurons. This latter scheme potentially also allows a fine discrimination of different environments (and locations) based on the DG population activity pattern (especially since the firing rates of DG neurons appear to be rather sensitive to changes in environmental features). However, since the CA3 code appears to be more efficiently orthogonalized than the DG code, an additional processing step may be needed. The physiological data reviewed above indicate that the mossy fiber projection may be utilized to arrive at the sparser, more completely pattern-separated representation recorded in area CA3, but further contributions from other sources (temporo-ammonic pathway and local network connections) cannot be excluded. The different behavior of spatial representations at different stages of hippocampal processing clearly argues against the simple view that all properties of hippocampal place cells originate in the DG, and downstream areas simply inherit these properties. A more complex view of the

formation of hippocampal representations is also indicated by recordings of neuronal activity patterns following selective lesions. In particular, following colchicine lesions of dentate granule cells as described above, place cell representations could still be observed in area CA1 (McNaughton et al., 1989). Similarly, the spatial representation in area CA1 was largely intact after surgical separation from all other hippocampal areas, which left it with the direct projection from entorhinal cortex as its sole cortical input (Brun et al., 2002). Therefore, hippocampal areas other than the DG must be capable of creating sparse, distributed, placefield-like representations on their own under some circumstances, probably based on their direct entorhinal inputs. However, it has to be kept in mind that despite the presence of a proper place cell representation in CA1, navigation memory was compromised following both types of lesion, suggesting that utilization of the spatial code at the behavioral level requires intact dentate–CA3 interaction. In summary, these data suggest that the DG does perform a pattern separation of the entorhinal signal, which, however, needs further processing to achieve the sparse and decorrelated activity pattern observed in the CA3 region.

The dentate gyrus: an evolutionary-developmental perspective Apparently the basic organization of the dentate–CA3 network has the deepest phylogenetic root among cortical regions. Before the divergence of reptilian-mammalian lineages which apparently preceded the mass extinction of the Permian period (
250 million years ago) neurons in all cortical areas were most probably packed in a single cellular layer. During the ontogenesis they likely followed an outside-in pattern of histogenesis, where newly generated neurons settle below the older ones, like in extant reptiles (Goffinet et al., 1986). Their main excitatory afferents entered and terminated in the embryonic marginal zone, above the cortical plate, i.e., in the same zone where the apical dendrites of their main targets (excitatory principal cells) were present (Ten Donkelaar,

594

1998). As suggested recently, this arrangement does not support the evolution of a cortical structure with multiple cellular layers and distinct areas (Super and Uylings, 2001). As the mammalian nervous system evolved, the dorsal and lateral cortex of the ancient reptilian cortex underwent a significant modification that opened up tremendous opportunities for areal and cellular diversification (Super and Uylings, 2001). This included changing the pattern of histogenesis to the inside-out pattern, where newly generated neurons settle above the older ones, and changing the pattern of axonal ingrowth (for a review, see Super and Uylings, 2001). In the mammalian neocortex, most of the afferents fibers enter not above but below the cortical plate, via the subplate (Allendoerfer and Shatz, 1994; Molnar, 2000). This new developmental scheme allowed the development of multilayered cortical structures with highly variable cell types, rich reciprocal interaction with the thalamus and the establishment of new cortical regions. Little developmental change occurred, however, in the medial and dorsomedial cortex, which became the mammalian allocortex. In the dorsomedial cortex that is homologous with the Cornu Ammonis, the pattern of histogenesis changed to the inside-out pattern, but the pattern of axonal ingrowth did not (Super et al., 1998). However, in the most ‘‘stubborn’’ structure, the DG, the basic reptilian developmental pattern was retained (Bayer, 1980) histogenesis follows the outside-in pattern and the main excitatory afferents enter above the cortical plate. Neocortex has evolved rapidly to become a highly complex multilayered structure, with extensive intracortical and thalamo-cortical reciprocal connections (Butler, 1994; Nieuwenhuys, 1994; Rakic, 1995; Northcutt and Kaas, 1995). Allocortex remained a unilayered structure with a basically unidirectional information flow, DG being the only cortical structure without thalamic input (Amaral and Witter, 1989). Neocortex has been significantly expanded laterally and was parceled into numerous functionally segregated areas, but the hippocampus retained the original two major subfields, the DG and Cornu Ammonis. For these two regions, only the Cornu Ammonis showed

some areal segregation, and the original reptilian dorsomedial cortex became CA3, CA2, CA1, and subiculum (Amaral et al., 1990). Again, the DG showed no areal segregation. But why did the DG–CA3 connection remain essentially unchanged during the course of evolution, when the rest of the cortical mantle underwent a dramatic reorganization? Here we would like to propose that the reason behind the protracted evolutionary pattern of the hippocampus, and especially the DG, is the structural constraints of hippocampal function. Apparently, the formation of freely accessible multidimensional memory traces can only be performed in a two-step process that includes a segregation of input followed by an associative step. The reciprocally coupled, multilayered cortical structures evolved for a different role (for a more complex interaction with the environment). Apparently, the basic plan of the two-step information processing through the hippocampal formation remained essentially the same from lizard to human. Similarly to mammals, in reptiles, the dentate-equivalent medial cortex receives the cortical input. This cortical region lacks an autoassociative network and projects the recoded information unidirectionally, to the reptilian analogue of the Cornu Ammonis, the dorsomedial cortex (Lopez-Garcia and MartinezGuijarro, 1988; Martinez-Guijarro et al., 1991; de la Iglesia et al., 1994). Association functions in the reptile may take place in the next step here in the dorsomedial cortex, where an extensive recurrent collateral system exists, similar to the CA3 region in mammals (Martinez-Guijarro et al., 1984). Interestingly, damage to the reptilian homologue of hippocampus causes similar learning problems as hippocampal lesions do in mammals (Rodriguez et al., 2002), suggesting an analogous structuralfunctional relationship. Has anything changed in DG during the mammalian evolution? In primates, the volumetric ratio of the DG and CA3 has changed in favor of the first. Indeed, DG is ‘‘dentate’’ sensu stricto only in primates, where it includes numerous infoldings. Among these areas the hilus showed the largest relative increase in volume and cell number (Seress, 1988) underlying the importance of the region, where most of the peculiarities in

595

microcircuits have been noticed. Apparently, the feed back regulation of granule cells became more elaborated with the increasing complexity of information to be categorized by the system.

Conclusions, unresolved questions The DG appears to be an ancient cortical structure from the phylogenetic perspective, yet it displays a number of unique features not found in other cortical regions. Its morphological and physiological properties are highly specialized, and these can explain, at least in part, the role assigned to the DG and CA3 by computational theories. In particular, mossy fibers are characterized by distinct mechanisms of signal transfer at excitatory vs. inhibitory targets, and unusually strong activation of GABAergic circuits. This arrangement allows a delicate balance of excitation and inhibition, which is utilized for code conversion and sparsification at the entorhinal–dentate connection and for frequency-dependent spike transfer at the dentate–CA3 connection. Several issues, however, remain unresolved. More comparable data are needed from freely moving animals to understand the precise computation that takes place in the DG and in CA3. Since DG–CA3 transmission appears to be dependent on the frequency of granule cell discharge, DG firing patterns should be carefully analyzed, and particularly with respect to multiple receptive fields. The role of two major excitatory inputs of the granule cells, not discussed here, the mossy cell input and the supramammillary afferents, is necessary to clarify DG information processing comprehensively. Both mossy cells and supramammillary afferents contact the proximal dendrites of granule cells, and therefore are likely to exert a powerful influence. Mossy cells of the hilus have highly divergent axons, and thus may link distant DG populations involved in coding similar environmental events, whereas the supramammillary input may mediate the modulation of granule cells by the theta rhythm. The role of rhythmic EEG activities (theta, gamma) in mediating signal transfer, code conversion, and the short- and long-term plasticity is unclear at the present time. Similarly,

the role of dentate spikes is not explored. It is tempting to speculate that they may participate in memory replay, like the sharp wave in the CA3–CA1 network (Buzsaki, 1989), but definitive proof is currently unavailable. From the computational perspective, it is not clear how the lack of connectivity among granule cells and the relative paucity of the direct backprojection from the CA3 to the DG (Li et al., 1994) helps the categorization function in DG. In conclusion, this peculiar neocortex–archicortex interface will likely keep us busy for a long time.

Acknowledgment This work was supported by the Wellcome Trust (A.L. is the recipient of a Wellcome Trust International Senior Fellowship), the Institut de Cerveau et de la Moelle e´piniere, the Hungarian Scientific Research Fund (OTKA T 049100) and the EU Framework 6.

References Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G. and Freund, T.F. (2000) Unusual target selectivity of perisomatic inhibitory cells in the hilar region of the rat hippocampus. J. Neurosci., 20: 6907–6919. Alle, H. and Geiger, J.R. (2006) Combined analog and action potential coding in hippocampal mossy fibers. Science, 311: 1290–1293. Allendoerfer, K.L. and Shatz, C.J. (1994) The subplate, a transient neocortical structure—its role in the development of connections between thalamus and cortex. Annu. Rev. Neurosci., 17: 185–218. Alvarez, P. and Squire, L.R. (1994) Memory consolidation and the medial temporal lobe: a simple network model. Proc. Natl. Acad. Sci. U.S.A., 91: 7041–7045. Amaral, D.G. (1979) Synaptic extensions from the mossy fibers of the fascia dentata. Anat. Embryol. (Berl), 155: 241–251. Amaral, D.G., Ishizuka, N. and Claiborne, B. (1990) Neurons, numbers and the hippocampal network. Understanding the Brain Through the Hippocampus, 83. Amaral, D.G. and Witter, M.P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31: 571–591.

596 Arabadzisz, D. and Freund, T.F. (1999) Changes in excitatory and inhibitory circuits of the rat hippocampus 12–14 months after complete forebrain ischemia [In Process Citation]. Neuroscience, 92: 27–45. Barnes, C.A., McNaughton, B.L., Mizumori, S.J., Leonard, B.W. and Lin, L.H. (1990) Comparison of spatial and temporal characteristics of neuronal activity in sequential stages of hippocampal processing. Prog. Brain Res., 83: 287–300. Baude, A., Nusser, Z., David, J., Roberts, B., Mulvihill, E., McIlhinney, R.A.J. and Somogyi, P. (1993) The metabotropic glutamate receptor (mGluR1O´) is concentrated at perisynaptic membrane of neuronal subpopulations as detected by immunogold reaction. Neuron, 11: 771–787. Bayer, S.A. (1980) Development of the hippocampal region in the rat. I. Neurogenesis examined with 3H-thymidine autoradiography. J. Comp. Neurol., 190: 87–114. Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M. and Buzsaki, G. (1995) Dentate EEG spikes and associated interneuronal population bursts in the hippocampal hilar region of the rat. J. Neurophysiol., 73: 1691–1705. Brun, V.H., Otnass, M.K., Molden, S., Steffenach, H.A., Witter, M.P., Moser, M.B. and Moser, E.I. (2002) Place cells and place recognition maintained by direct entorhinalhippocampal circuitry. Science, 296: 2243–2246. Buckmaster, P.S. and Schwartzkroin, P.A. (1995) Interneurons and inhibition in the dentate gyrus of the rat in vivo. J. Neurosci., 15: 774–789. Butler, A.B. (1994) The evolution of the dorsal thalamus of jawed vertebrates, including mammals: cladistic analysis and a new hypothesis. Brain Res. Brain Res. Rev., 19: 29–65. Buzsaki, G. (1989) Two-stage model of memory trace formation: a role for ‘noisy’ brain states. Neuroscience, 310: 551–570. Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P., Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton, B.L. and Barnes, C.A. (2005) Sparse, environmentally selective expression of Arc RNA in the upper blade of the rodent fascia dentata by brief spatial experience. Hippocampus, 15: 579–586. Chicurel, M.E. and Harris, K.M. (1992) 3-dimensional analysis of the structure and composition of CA3 branched dendritic spines and their synaptic relationships with mossy fiber boutons in the rat hippocampus. J. Comp. Neurol., 325: 169–182. Chrobak, J.J. and Buzsaki, G. (1994) Selective activation of deep layer (V–VI) retrohippocampal cortical neurons during hippocampal sharp waves in the behaving rat. J. Neurosci., 14: 6160–6170. Chrobak, J.J. and Buzsaki, G. (1996) High-frequency oscillations in the output networks of the hippocampal-entorhinal axis of the freely behaving rat. J. Neurosci., 16: 3056–3066. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1986) A light and electron microscopic analysis of the mossy fibers of the rat dentate gyrus. J. Comp. Neurol., 246: 435–458. Cobb, S.R., Halasy, K., Vida, I., Nyiri, G., Tamas, G., Buhl, E.H. and Somogyi, P. (1997) Synaptic effects of identified interneurons innervating both interneurons and pyramidal

cells in the rat hippocampus [published erratum appears in Neuroscience 1997 Oct; 80(3): 971]. Neuroscience, 79: 629–648. Conrad, C.D. and Roy, E.J. (1993) Selective loss of hippocampal granule cells following adrenalectomy: implications for spatial memory. J. Neurosci., 13: 2582–2590. Costa, V.C., Bueno, J.L. and Xavier, G.F. (2005) Dentate gyrus-selective colchicine lesion and performance in temporal and spatial tasks. Behav. Brain Res., 160: 286–303. Debanne, D., Gahwiler, B.H. and Thompson, S.M. (1998) Long-term synaptic plasticity between pairs of individual CA3 pyramidal cells in rat hippocampal slice cultures. J. Physiol., 507(Pt 1): 237–247. De Paola, V., Arber, S. and Caroni, P. (2003) AMPA receptors regulate dynamic equilibrium of presynaptic terminals in mature hippocampal networks. Nat. Neurosci., 6: 491–500. Engel, D. and Jonas, P. (2005) Presynaptic action potential amplification by voltage-gated Na+ channels in hippocampal mossy fiber boutons. Neuron, 45: 405–417. Frank, L.M., Brown, E.N. and Wilson, M. (2000) Trajectory encoding in the hippocampus and entorhinal cortex. Neuron, 27: 169–178. Fyhn, M., Molden, S., Witter, M.P., Moser, E.I. and Moser, M.B. (2004) Spatial representation in the entorhinal cortex. Science, 305: 1258–1264. Geiger, J.R., Lubke, J., Roth, A., Frotscher, M. and Jonas, P. (1997) Submillisecond AMPA receptor-mediated signaling at a principal neuron-interneuron synapse. Neuron, 18: 1009–1023. Gilbert, P.E., Kesner, R.P. and Lee, I. (2001) Dissociating hippocampal subregions: double dissociation between dentate gyrus and CA1. Hippocampus, 11: 626–636. Goffinet, A.M., Daumerie, C., Langerwerf, B. and Pieau, C. (1986) Neurogenesis in reptilian cortical structures: 3Hthymidine autoradiographic analysis. J. Comp. Neurol., 243: 106–116. Goodrich-Hunsaker, N.J., Hunsaker, M.R. and Kesner, R.P. (2005) Dissociating the role of the parietal cortex and dorsal hippocampus for spatial information processing. Behav. Neurosci., 119: 1307–1315. Gothard, K.M., Hoffman, K.L., Battaglia, F.P. and McNaughton, B.L. (2001) Dentate gyrus and ca1 ensemble activity during spatial reference frame shifts in the presence and absence of visual input. J. Neurosci., 21: 7284–7292. Gulyas, A.I., Megias, M., Emri, Z. and Freund, T.F. (1999) Total number and ratio of excitatory and inhibitory synapses converging onto single interneurons of different types in the CA1 area of the rat hippocampus. J. Neurosci., 19: 10082–10097. Gulyas, A.I., Miettinen, R., Jacobowitz, D.M. and Freund, T.F. (1991) Calretinin-immunoreacti cells in the rat hippocampus. I. A new type of neuron — specifically associated with the mossy fibre system — revealed. Eur. J. Neurosci. Suppl., 4: 158. Gulyas, A.I., Miles, R., Ha´jos, N. and Freund, T.F. (1993) Precision and variability in postsynaptic target selection of inhibitory cells in the hippocampal CA3 region. Eur. J. Neurosci., 5: 1729–1751.

597 Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser, E.I. (2005) Microstructure of a spatial map in the entorhinal cortex. Nature, 436: 801–806. Hafting, T., Fyhn, M., Treves, A., Moser, E.I. and Moser, M.B. (2006) Coherent reallignment of entorhinal grid cells coincides global remapping in the hippocampus. FENS, Vienna. Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A high degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 395–410. Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J. (2005) Major dissociation between medial and lateral entorhinal input to dorsal hippocampus. Science, 308: 1792–1794. Henze, D.A., Wittner, L. and Buzsaki, G. (2002) Single granule cells reliably discharge targets in the hippocampal CA3 network in vivo. Nat. Neurosci., 5: 790–795. Houser, C.R. and Esclapez, M. (1994) Localization of mRNAs encoding two forms of glutamic acid decarboxylase in the rat hippocampal formation. Hippocampus, 4: 530–545. de la Iglesia, J.A., Martinez-Guijarro, F.I. and Lopez-Garcia, C. (1994) Neurons of the medial cortex outer plexiform layer of the lizard Podarcis hispanica: Golgi and immunocytochemical studies. J. Comp. Neurol., 341: 184–203. Isomura, Y., Sirota, A., Ozen, S., Montgomery, S., Mizuseki, K., Henze, D.A. and Buzsaki, G. (2006) Integration and segregation of activity in entorhinal-hippocampal subregions by neocortical slow oscillations. Neuron, 52: 871–882. Jeltsch, H., Bertrand, F., Lazarus, C. and Cassel, J.C. (2001) Cognitive performances and locomotor activity following dentate granule cell damage in rats: role of lesion extent and type of memory tested. Neurobiol. Learn. Mem., 76: 81–105. Jonas, P., Major, G. and Sakmann, B. (1993) Quantal components of unitary EPSCs at the mossy fibre synapse on CA3 pyramidal cells of rat hippocampus. J. Physiol. Paris London, 472: 615–663. Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of unit activity in the hippocampal granular layer. Hippocampus, 3: 165–182. Kali, S. and Dayan, P. (2000) The involvement of recurrent connections in area CA3 in establishing the properties of place fields: a model. J. Neurosci., 20: 7463–7477. Kali, S. and Dayan, P. (2004) Off-line replay maintains declarative memories in a model of hippocampal-neocortical interactions. Nat. Neurosci., 7: 286–294. O’Keefe, J. and Dostrovsky, J. (1971) The hippocampus as a spatial map. Preliminary evidence from unit activity in the freely-moving rat. Brain Res., 34: 171–175. O’Keefe, J. and Nadel, L. (1978) The hippocampus as a cognitive map. Clarendon, Oxford. Kobayashi, K. and Poo, M.M. (2004) Spike train timing-dependent associative modification of hippocampal CA3 recurrent synapses by mossy fibers. Neuron, 41: 445–454. Lassalle, J.M., Bataille, T. and Halley, H. (2000) Reversible inactivation of the hippocampal mossy fiber synapses in mice impairs spatial learning, but neither consolidation nor

memory retrieval, in the Morris navigation task. Neurobiol. Learn. Mem., 73: 243–257. Lawrence, J.J., Grinspan, Z.M. and McBain, C.J. (2004) Quantal transmission at mossy fibre targets in the CA3 region of the rat hippocampus. J. Physiol., 554: 175–193. Lee, I. and Kesner, R.P. (2004) Encoding versus retrieval of spatial memory: double dissociation between the dentate gyrus and the perforant path inputs into CA3 in the dorsal hippocampus. Hippocampus, 14: 66–76. Leutgeb, J.K., Leutgeb, S., Moser, M.B. and Moser, E.I. (2007) Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science, 315: 961–966. Leutgeb, J.K., Leutgeb, S., Treves, A., Meyer, R., Barnes, C.A., McNaughton, B.L., Moser, M.B. and Moser, E.I. (2005a) Progressive transformation of hippocampal neuronal representations in ‘‘morphed’’ environments. Neuron, 48: 345–358. Leutgeb, S., Leutgeb, J.K., Barnes, C.A., Moser, E.I., McNaughton, B.L. and Moser, M.B. (2005b) Independent codes for spatial and episodic memory in hippocampal neuronal ensembles. Science, 309: 619–623. Leutgeb, S., Leutgeb, J.K., Treves, A., Moser, M.B. and Moser, E.I. (2004) Distinct ensemble codes in hippocampal areas CA3 and CA1. Science, 305: 1295–1298. Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The hippocampal ca3 network — an in vivo intracellular labeling study. J. Comp. Neurol., 339: 181–208. Lopez-Garcia, C. and Martinez-Guijarro, F.J. (1988) Neurons in the medial cortex give rise to Timm-positive boutons in the cerebral cortex of lizards. Brain Res., 463: 205–217. Maccaferri, G., Toth, K. and McBain, C.J. (1998) Target-specific expression of presynaptic mossy fiber plasticity. Science, 279: 1368–1370. Marr, D. (1971) Simple memory: a theory for archicortex. Philos. Trans. R. Soc. Lond. B Biol. Sci., 262: 23–81. Martinez Guijarro, F.J., Berbel, P.J., Molowny, A. and Lopez Garcia, C. (1984) Apical dendritic spines and axonic terminals in the bipyramidal neurons of the dorsomedial cortex of lizards (Lacerta). Anat. Embryol., 170: 321–326. Martinez-Guijarro, F.J., Soriano, E., Del Rio, J.A. and LopezGarcia, C. (1991) Zinc-positive boutons in the cerebral cortex of lizards show glutamate immunoreactivity. J. Neurocytol., 20: 834–843. Maurer, A.P., Vanrhoads, S.R., Sutherland, G.R., Lipa, P. and McNaughton, B.L. (2005) Self-motion and the origin of differential spatial scaling along the septo-temporal axis of the hippocampus. Hippocampus, 15: 841–852. McClelland, J.L., McNaughton, B.L. and O’Reilly, R.C. (1995) Why there are complementary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory. Psychol. Rev., 102: 419–457. McNaughton, B.L., Barnes, C.A., Meltzer, J. and Sutherland, R.J. (1989) Hippocampal granule cells are necessary for normal spatial learning but not for spatially-selective pyramidal cell discharge. Experimental brain research. Experimentelle Hirnforschung, 76: 485–496.

598 McNaughton, B.L. and Morris, N.G. (1987) Hippocampal synaptic enhancement and information storage within a distributed memory system. TINS, 10: 408–415. Megias, M., Emri, Z., Freund, T.F. and Gulyas, A.I. (2001) Total number and distribution of inhibitory and excitatory synapses on hippocampal CA1 pyramidal cells. Neuroscience, 102: 527–540. Miles, R., Toth, K., Gulyas, A.I., Ha´jos, N. and Freund, T.F. (1996) Differences between somatic and dendritic inhibition in the hippocampus. Neuron, 16: 815–823. Miles, R. and Wong, R.K. (1986) Excitatory synaptic interactions between CA3 neurones in the guinea-pig hippocampus. J. Physiol., 373: 397–418. Molnar, Z. (2000) Development and evolution of thalamocortical interactions. Eur. J. Morphol., 38: 313–320. Mori, M., Abegg, M.H., Gahwiler, B.H. and Gerber, U. (2004) A frequency-dependent switch from inhibition to excitation in a hippocampal unitary circuit. Nature, 431: 453–456. Muller, R.U., Kubie, J.L. and Ranck Jr., J.B. (1987) Spatial firing patterns of hippocampal complex-spike cells in a fixed environment. J. Neurosci., 7: 1935–1950. Nadasdy, Z., Hirase, H., Czurko, A., Csicsvari, J. and Buzsaki, G. (1999) Replay and time compression of recurring spike sequences in the hippocampus [in process citation]. J. Neurosci., 19: 9497–9507. Nieuwenhuys, R. (1994) The neocortex. An overview of its evolutionary development, structural organization and synaptology. Anat. Embryol., 190: 307–337. Nitz, D. and McNaughton, B. (2004) Differential modulation of CA1 and dentate gyrus interneurons during exploration of novel environments. J. Neurophysiol., 91: 863–872. Northcutt, R.G. and Kaas, J.H. (1995) The emergence and evolution of mammalian neocortex. Trends Neurosci., 18: 373–379. Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki, G. (1997) Feed-forward and feed-back activation of the dentate gyrus in vivo during dentate spikes and sharp wave bursts. Hippocampus, 7: 437–450. Quirk, G.J., Muller, R.U., Kubie, J.L. and Ranck Jr., J.B. (1992) The positional firing properties of medial entorhinal neurons: description and comparison with hippocampal place cells. J. Neurosci., 12: 1945–1963. Rakic, P. (1995) A small step for the cell, a giant leap for mankind: a hypothesis of neocortical expansion during evolution. Trends Neurosci., 18: 383–388. Ramon y Cajal, S. (1911) Histologie de systeme nerveux de l’homme et des vertebres tomme II. Maloine, Paris. Reichova, I. and Sherman, S.M. (2004) Somatosensory corticothalamic projections: distinguishing drivers from modulators. J. Neurophysiol., 92: 2185–2197. O’Reilly, R.C. and McClelland, J.L. (1994) Hippocampal conjunctive encoding, storage, and recall: avoiding a trade-off. Hippocampus, 4: 661–682. Rodriguez, F., Lopez, J.C., Vargas, J.P., Gomez, Y., Broglio, C. and Salas, C. (2002) Conservation of spatial memory function in the pallial forebrain of reptiles and ray-finned fishes. J. Neurosci., 22: 2894–2903.

Rolls, E.T. (1989) Functions of neuronal networks in the hippocampus and cerebral cortex in memory. In: Cotterill R.M.J. (Ed.), Models of Brain Function. Cambridge University Press, Cambridge, pp. 15–33. Rolls, E.T. and Kesner, R.P. (2006) A computational theory of hippocampal function, and empirical tests of the theory. Prog. Neurobiol., 79: 1–48. Rumelhart, D.E. and Zipser, D. (1986) Feature discovery by competitive learning. In: Rumelhart D.E., McClelland J. and Group. P.R. (Eds.), Parallel Distributed Processing: Explorations in the Microstructure of Cognition. Vol. 1. Foundations. MIT Press, Cambridge, MA, pp. 151–193. Salin, P.A., Scanziani, M., Malenka, R.C. and Nicoll, R.A. (1996) Distinct short-term plasticity at two excitatory synapses in the hippocampus. Proc. Natl. Acad. Sci. U.S.A., 93: 13304–13309. Samsonovich, A. and McNaughton, B.L. (1997) Path integration and cognitive mapping in a continuous attractor neural network model. J. Neurosci., 17: 5900–5920. Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A. (1990) Synaptic Connections of Dentate Granule Cells and Hilar Neurons — Results of Paired Intracellular Recordings and Intracellular Horseradish Peroxidase Injections. Neuroscience, 37: 693–707. Scoville, W.B. and Milner, B. (1957) Loss of recent memory after bilateral hippocampal lesions. J. Neurol. Neurosurg. Psychiatry, 20: 11–21. Seress, L. (1988) Interspecies comparison of the hippocampal formation shows increased emphasis on the regio superior in the Ammon’s horn of the human brain. Journal fur Hirnforschung, 29: 335–340. Sharp, P.E. (1991) Computer simulation of hippocampal place cells. Psychobiology, 19: 103–115. Sherman, S.M. and Guillery, R.W. (1998) On the actions that one nerve cell can have on another: distinguishing ‘‘drivers’’ from ‘‘modulators.’’ Proc. Natl. Acad. Sci. U.S.A., 95: 7121–7126. Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in the hippocampal dentate gyrus: an in vivo intracellular study. Eur. J. Neurosci., 9: 573–588. Sik, A., Penttonen, M., Ylinen, A. and Buzsaki, G. (1995) Hippocampal CA1 interneurons: an in vivo intracellular labeling study. J. Neurosci., 15(10): 6651–6665. Sik, A., Tamamaki, N. and Freund, T.F. (1993) Complete axon arborization of a single CA3 pyramidal cell in the rat hippocampus, and its relationship with postsynaptic parvalbumin-containing interneurons. Eur. J. Neurosci., 5: 1719–1728. Skaggs, W.E., McNaughton, B.L., Wilson, M.A. and Barnes, C.A. (1996) Theta phase precession in hippocampal neuronal populations and the compression of temporal sequences. Hippocampus, 6: 149–172. Squire, L.R. (1992) Memory and the hippocampus: a synthesis from findings with rats, monkeys, and humans. Psychol. Rev., 99: 195–231. Super, H., Martinez, A., Del Rio, J.A. and Soriano, E. (1998) Involvement of distinct pioneer neurons in the formation of

599 layer-specific connections in the hippocampus. J. Neurosci., 18: 4616–4626. Super, H. and Uylings, H.B. (2001) The early differentiation of the neocortex: a hypothesis on neocortical evolution. Cerebral cortex New York, N.Y., 11: 1101–1109. Sutherland, R.J., Whishaw, I.Q. and Kolb, B. (1983) A behavioural analysis of spatial localization following electrolytic, kainate- or colchicine-induced damage to the hippocampal formation in the rat. Behav. Brain Res., 7: 133–153. Tabuchi, E., Mulder, A.B. and Wiener, S.I. (2003) Reward value invariant place responses and reward site associated activity in hippocampal neurons of behaving rats. Hippocampus, 13: 117–132. Tamas, G., Somogyi, P. and Buhl, E.H. (1998) Differentially interconnected networks of GABAergic interneurons in the visual cortex of the cat. J. Neurosci., 18: 4255–4270. Tashiro, A., Dunaevsky, A., Blazeski, R., Mason, C.A. and Yuste, R. (2003) Bidirectional regulation of hippocampal mossy fiber filopodial motility by kainate receptors: a twostep model of synaptogenesis. Neuron, 38: 773–784. Ten Donkelaar, H. (1998) Reptiles. In: Nieuwenhuys R., Ten Donkelaar H.J. and Nicholson C. (Eds.), The Central Nervous System of Vertebrates. Springer, Berlin. Toth, K., Suares, G., Lawrence, J.J., Philips-Tansey, E. and McBain, C.J. (2000) Differential mechanisms of transmission

at three types of mossy fiber synapse. J. Neurosci., 20: 8279–8289. Treves, A. and Rolls, E.T. (1992) Computational constraints suggest the need for two distinct input systems to the hippocampal CA3 network. Hippocampus, 2: 189–199. Treves, A. and Rolls, E.T. (1994) Computational analysis of the role of the hippocampus in memory. Hippocampus, 4: 374–391. Vida, I. and Frotscher, M. (2000) A hippocampal interneuron associated with the mossy fiber system. Proc. Natl. Acad. Sci. U.S.A., 97: 1275–1280. Willshaw, D.J. and Buckingham, J.T. (1990) An assessment of Marr theory of the hippocampus as a temporary memory store. Phil. Trans. R. Soc. London, 329: 205–215. Wilson, M.A. and Mcnaughton, B.L. (1993) Dynamics of the hippocampal ensemble code for space. Science, 261: 1055–1058. Wittner, L., Henze, D.A., Zaborszky, L. and Buzsaki, G. (2006) Hippocampal CA3 pyramidal cells selectively innervate aspiny interneurons. Eur. J. Neurosci., 24: 1286–1298. Xavier, G.F., Oliveira-Filho, F.J. and Santos, A.M. (1999) Dentate gyrus-selective colchicine lesion and disruption of performance in spatial tasks: difficulties in ‘‘place strategy’’ because of a lack of flexibility in the use of environmental cues? Hippocampus, 9, 668–681.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 32

The dentate gyrus as a filter or gate: a look back and a look ahead David Hsu Department of Neurology, University of Wisconsin, 600 Highland Avenue, H6/526, Madison, WI 53792, USA

Abstract: The idea of the dentate gyrus as a gate or filter at the entrance to the hippocampus, blocking or filtering incoming excitation from the entorhinal cortex, has been an intriguing one. Here we review the historical development of the idea, and discuss whether it may be possible to be more specific in defining this gate. We propose that dentate function can be understood within a context of Hebbian association and competition: hilar mossy cells help the dentate granule cells to recognize incoming entorhinal patterns of activity (Hebbian association), after which patterns that are consistently and repetitively presented to the dentate gyrus are passed through, while random, more transient patterns are blocked (non-associative Hebbian competition). Translamellar inhibition as well as translamellar potentiation can be understood in this context. The dentate-hilar complex thus plays the role of a ‘‘pattern excluder’’, not a pattern completer. The unique role of pattern exclusion may explain the peculiar qualities of dentate granule cells and hilar mossy cells. Keywords: dentate gate; dentate filter; pattern excluder 1992; Staley et al., 1992; Williamson et al., 1993), and strong GABA receptor-mediated inhibition (Mody et al., 1992; Coulter, 1999; Nusser and Mody, 2002; Stell and Mody, 2002; Cohen et al., 2003; Mody, 2005).

Introduction The dentate gyrus, sitting between the entorhinal cortex and area CA3, is both anatomically well positioned and physiologically predisposed to play the role of a gate, blocking or filtering excitatory activity from the entorhinal cortex and controlling the amount of excitation that gets through to the hippocampus. Normal adult granule cells rarely generate action potentials. In part this is because there is little direct interconnectivity between dentate granule cells under normal conditions (reviewed in Chapter 1 of this volume). In addition, granule cells have a high resting membrane potential (Fricke and Prince, 1984; Scharfman,

History of the idea Data that the dentate gyrus may serve as a gate appeared in at least as early as 1966 in work by Andersen et al. (1966). In these experiments, the hippocampal formation of adult rabbits was exposed by removal of the overlying neocortex and corpus callosum. Stimulating electrodes were placed into entorhinal cortex or in the perforant pathway. Extracellular as well as intracellular recordings were made from electrodes placed in the

Corresponding author. Tel.: +1 608 263 8551; Fax: +1 608 263 0412; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63032-5

601

602

transverse plane of the hippocampus, piercing CA1 and both blades of the dentate gyrus. Perforant pathway stimulation resulted in a negative wave reflecting granule cell excitatory postsynaptic potentials (EPSPs) generated in the middle third of the dentate gyrus molecular layer. The EPSP was followed by a large and slow inhibitory postsynaptic potential (IPSP), persisting for 100–150 ms. That EPSPs were followed by IPSPs persisting for 100–150 ms suggested that perforant pathway stimulation at frequencies higher than 10 Hz should result in a decremental granule cell population spike response (habituation). For pulse stimulus durations of less than 1 s, this decremental response was indeed observed. However, for longer pulse durations of 6 and 9 s, there was a gradual incremental granule cell response (facilitation), beginning after a few seconds of repetitive stimulation. The degree of facilitation increased with increasing stimulation frequency up to a maximum of 10 Hz, and was abolished by a stimulation frequency higher than 20 Hz (Andersen et al., 1966). Although not stated explicitly by the authors, these experiments established one way that the dentate gyrus appears to act as a gate. Short-duration stimuli carried at a certain frequency are blocked, but longer duration stimuli carried at the same frequency are facilitated. Later studies demonstrated additional ways the dentate gyrus acts as a gate. In 1976, Alger and Teyler applied repetitive perforant pathway stimulation to rat hippocampal slices at 1 Hz for a total duration of 10 s (Alger and Teyler, 1976; Teyler and Alger, 1976). They found an incremental EPSP and population spike response with each succeeding stimulus (facilitation) in CA3 and CA1 but a decremental response (habituation) in the dentate gyrus. In contrast, stimulation of the perforant pathway at 15 Hz for 15 s produced potentiation of EPSP and population spike responses of dentate gyrus, CA3 and CA1. These results are similar to those of Andersen et al. (1966), showing that the dentate gyrus suppresses flow of excitation from perforant pathway input but that this suppression can be reversed with longer duration stimuli in an appropriate frequency range. Winson and Abzug (1977, 1978) placed stimulating electrodes into the angular bundle of the

perforant pathway in behaving rats, and compared dentate granular layer population spike amplitudes and molecular layer EPSPs in slow wave and rapid eye movement (REM) sleep vs. the alert and still state. The authors found that slow wave and REM sleep states are associated with larger population spikes but smaller EPSPs compared to the still, alert state. To explain these findings, the authors hypothesized that there was relative hyperpolarization of the dentate granule cell membrane potential during the still, alert state compared to slow-wave sleep. The relative hyperpolarization was interpreted as a gating mechanism. Collins et al. (1983) studied the role of the dentate gyrus in seizure propagation in behaving rats. Stimulation with either focal chemoconvulsant injection into or electrical stimulation of entorhinal cortex produced a graded response. If focal chemoconvulsant injection induced fewer than 10 spikes per minute in entorhinal cortex, no or minimal behavioral changes were noted, and no metabolic changes were noted on later sectioning and deoxyglucose autoradiography. If 10–30 spikes per minute are induced in entorhinal cortex, then there is increased deoxyglucose uptake in the entorhinal cortex and in the dentate gyrus, but restricted in dentate gyrus to the molecular layer. Behaviors associated with weak seizure activity were observed. However, if greater than 40 spikes per minute were induced, moderate seizures developed. Metabolic changes then spread to the entire dentate gyrus, areas CA3 and CA1, the septal nucleus, and occasionally to the amygdala, nucleus accumbens, and ventral pallidum-lateral preoptic area. Spread to the contralateral side also occurred. Because metabolic changes were initially restricted to the molecular layer of the dentate gyrus, and only with further increases of chemical or electrical stimulation were these changes able to spread to other parts of the hippocampus and to extrahippocampal structures, the authors concluded that ‘‘the sequential changes in [deoxyglucose] metabolism suggest that the dentate gyrus acts as a restrictive gateway for seizure spread from entorhinal cortex to the rest of the limbic systemy.’’ (Collins et al., 1983). Further work by Lothman and coworkers was reviewed in 1992 (Lothman et al., 1992). Working

603

in vivo with unanesthetized rats, Stringer et al. (1989) and Stringer and Lothman (1992) developed the concept of maximal dentate activation (MDA). Their experiments involved stimulation of the perforant pathway or area CA3 until a maximal, apparently saturating level of activity was recorded in the dentate gyrus. This state was defined as MDA. MDA was most easily obtained with a stimulating frequency between 10–40 Hz. The onset of MDA occurred with a pronounced negative shift of the DC potential, reflecting a depolarization of the granule cell layer. In addition, there was an abrupt rise in extracellular K+ concentration, and the appearance of bursts of large amplitude population spikes. If stimulation was triggered above the threshold for MDA, afterdischarges were observed, which could persist for a short time after stimulation ceased. MDA in the intact rat was always bilateral and associated with synchronous epileptiform discharges in bilateral CA3, CA1, subiculum, and entorhinal cortex. Lesioning the entorhinal cortex on one side can block MDA on that side, but the contralateral side retained the ability to reach MDA. If stimulation were applied to the perforant pathway, MDA occurred in the dentate gyrus before activity was recorded in CA1. Thus, it appeared that transmission flowed from entorhinal cortex to dentate gyrus to CA3 and CA1, and not by the direct entorhinal to CA1 (temporoammonic) pathway. Lothman et al. (1992) and Stringer and Lothman (1992) interpreted their data in the following way: ‘‘(1) MDA serves to initiate and sustain reverberatory seizure activity in [the] hippocampal–parahippocampal loop; (2) this reverberatory seizure activity bombards extrahippocampal structures; (3) MDA can directly (within the hippocampal–parahippocampal loop) and indirectly (by influencing sites outside the hippocampal–parahippocampal loop) modulate the length of electrographic seizures and, in turn, their propagation, thereby affecting the expression of various types of behavioral seizures; (4) MDA can be accessed from any point within the hippocampal–parahippocampal loop; (5) MDA can also be accessed from points outside this loop....’’. Thus the dentate gyrus, when its function as a control

point is breached, actually acts as a ‘‘promoter’’ or ‘‘amplifier’’ of seizural discharges. Walther et al. (1986) studied rat brain slices in superfusate containing a low concentration of magnesium, which was shown to induce repetitive burst discharges in their slices. The entorhinal cortex demonstrated prolonged epilepiform discharges, lasting minutes at a time. The subiculum was also capable of spontaneous discharges, lasting up to 9 s. In contrast, isolated minislices of area CA3 were only capable of brief spontaneous transients, and the dentate gyrus demonstrated no activity at all when connections to the entorhinal cortex were disrupted. Because even prolonged epileptiform discharges in the entorhinal cortex elicited only brief transient activity in the dentate gyrus, the authors suggested that ‘‘the dentate gyrus y may serve as a filter which reduces the excitatory load into CA3 and hence into CA1’’ (Walther et al., 1986). Further details of the pharmacology and electrophysiology of these slices were reviewed in Heinemann et al. (1992). A striking visual demonstration of dentate gating was presented by Iijima et al. (1996) using optical imaging with fluorescence voltage-sensitive dye in rat brain slices. After superfusing their slices with an antagonist of GABA-A receptors, electrical stimulation in the superficial layers of the entorhinal cortex led to signals reflecting robust excitation of the entorhinal cortex. This first spread throughout the superficial layers of the entorhinal cortex, then involved the deep layers. At 33.6 ms after stimulation, excitation invaded the hippocampus, lasting until 151.2 ms after stimulation. The entorhinal cortex remained active through this time, and activity reverberated within the entorhinal cortex for the next 200 ms. In this period, the hippocampus showed only weak and partial activation. A second stimulus, delivered 352.8 ms after the first, caused further reverberatory activity in the entorhinal cortex, which then again penetrated to the hippocampus and led to hippocampal excitation lasting about 70 ms. A second experiment was also performed in normal solution (without the GABA-A receptor antagonist), using 1 Hz repetitive stimulation instead of single stimuli. Each stimulation resulted in increased activity in the entorhinal cortex, but it

604

was not until the seventh stimulus that activity penetrated to the hippocampus. These results demonstrated that entorhinal cortex activation, even when robust, does not easily penetrate into the hippocampus, presumably due to gating at the level of the entrance to the hippocampus, i.e., at the dentate gyrus. Behr et al. (1998) used kindled animals to demonstrate the dentate gate and its ‘‘breakdown’’. The authors first superfused entorhinal-hippocampal brain slices in low-magnesium solution. Spontaneous seizure-like activity in both entorhinal cortex and in area CA3 was then recorded, but activity was significantly larger in amplitude and longer in duration in area CA3 of kindled slices compared to control slices. Transecting the perforant pathway greatly diminished epileptiform activity in area CA3, but transecting the subiculumto-entorhinal pathway did not. These results demonstrated that epileptiform activity can be passed from entorhinal cortex through the dentate gyrus into area CA3, and that passage is made more likely after kindling. The authors then devised an experiment so that only the entorhinal cortex was locally perfused with both a GABA-A receptor antagonist and elevated K+ to induce epileptiform activity in a spatially-specific manner. Electrical stimulation of the entorhinal cortex resulted in a much stronger response in the dentate gyrus of kindled slices. Furthermore, in a separate experiment, spontaneous entorhinal interictal activity failed to trigger epileptiform discharges in dentate gyrus in 8 out of 8 control slices, but did trigger them in 7 out of 9 kindled slices. Taken together, these results show that, in normal brain, the dentate gyrus appears to prevent epileptiform activity in the entorhinal cortex from reaching the hippocampus, but after epileptogenesis (exemplified by kindling), the dentate gyrus no longer functions as a gate. Evidence for dentate gating of seizures was also found by monitoring the level of c-fos protein expression after the development of spontaneous seizures in a pilocarpine model of epilepsy in mice (Peng and Houser, 2005). The expression of c-fos protein is a marker for neuronal activity. Increased c-fos protein levels are evident 20–40 min after c-fos activation, at least in many neuronal types

where it has been studied. At 15 min after a 1–2 min spontaneous behavioral seizure in epileptic mice, c-fos labeling appeared in dentate granule cells, spread throughout the entire extent of the dentate gyrus but not involving the interneurons of the dentate-hilar border or the dentate molecular layer. At 30 min, c-fos staining was intense in the dentate gyrus, involving both granule cells and dentate-hilar border interneurons, and increased c-fos staining also spread to the rest of the hippocampus. At 1–2 h, c-fos staining began to fade in the dentate granule cells but was intense in interneurons of the dentate-hilar border and the dentate molecular layer. At 4 h, c-fos staining was lighter throughout the hippocampus, including the dentate gyrus, compared to controls. These data suggested that dentate granule cell activation was likely to have been an early event in spontaneous seizures. The authors commented that the rate of c-fos expression varies between cell types, so that c-fos expression occurred first in dentate granule cells and later in dentate interneurons is suggestive, but not proof that activity in granule cells preceded that in dentate interneurons. The breakdown of dentate gating and its presumed relationship to epileptogenesis motivated much of the research described above, at least as early as the work by Collins et al. (1983). It was recognized that the dentate gyrus is normally resistant to the propagation of discharges from the entorhinal cortex. In the setting of limbic epilepsy (i.e., temporal lobe epilepsy), the gate is thought to be compromised, so that seizure activity from entorhinal cortex is allowed into the hippocampus, and propagated in a reverberatory cycle back to entorhinal cortex again (Stringer and Lothman, 1992). This suggestion has led to many studies, which have focused on reasons why the dentate gyrus gate may ‘‘breakdown’’ in temporal lobe epilepsy. Based primarily on animal models of temporal lobe epilepsy, the results have suggested that the dentate gyrus gate may breakdown because of a change in the balance of excitation and inhibition of dentate gyrus granule cells. Decreased inhibition of granule cells could develop because of seizure-induced loss of GABAergic neurons or altered expression of GABAA receptors, among many other reasons (Mody et al.,

605

1992; Mody, 2005). Increased excitation may develop because of mossy fiber sprouting, as well as other factors (Stringer and Lothman, 1992; Sutula et al., 1992; Jackson and Scharfman, 1996; Scharfman, 2004). A more complex interplay between initial hyperexcitability followed by chronic hyperinhibition has also been suggested (Sloviter et al., 2006).

The idea of dentate gate vs. filter In summary, there is now a series of studies, which suggest that activity in the entorhinal cortex is often halted, delayed, or diminished at the dentate gyrus. The decreased excitability appears to be related at least in part to the strong, prolonged dentate granule cell IPSP first described by Andersen et al. (1966). Further, as also found in that study, repetitive perforant pathway stimulation for a prolonged period of time (a few seconds or longer) results in facilitation of succeeding stimulations. Thus if the dentate gyrus is a gate, it is a gate that can be opened if one is persistent, i.e., if one keeps ‘‘knocking’’ on it. Why does the dentate gate open with repeated stimulation? Meticulous simultaneous intracellular recordings seem to show that dentate and hilar interneurons respond strongly and faithfully to dentate granule cell discharges (Scharfman et al., 1990). However, with repeated granule cell discharge, the interneuronal response switch from action potentials to EPSPs — the interneurons still hear the command to fire but stop firing. Conversely, dentate and hilar interneuronal discharges produce IPSPs in dentate granule cells, but with many failures. Interestingly, failures are more likely after a large IPSP. These results suggest that at least part of the gating function resides in the local granule cell and interneuron circuitry, and that this inhibitory circuit is tuned down in efficacy with repeated activation (Scharfman et al., 1990). Such activity-dependent disinhibition, involving principal neurons with their local interneuronal circuitry, appears to be an important recurring theme in other brain areas as well (Ben-Ari et al., 1980; Wong and Watkins, 1982; McCarren and

Alger, 1985; Deisz and Prince, 1989; Thompson and Gahwiler, 1989a, b, c; Scanziani et al., 1991; Mott and Lewis, 1992; Thomson et al., 1993). For instance, in area CA3, repeated stimulation of pyramidal neurons leads to decreased IPSCs via two mechanisms: (a) prolonged activation of GABA-A receptors, which leads to a chloride influx into the principal neuron, which leads to a decrease in driving force for chloride-mediated GABAergic inhibition, and (b) presynaptic negative feedback of GABA onto GABA-B receptors, which leads to decreased presynaptic GABA release (Thompson and Gahwiler, 1989a, b, c). Further details on complex GABAergic responses have been studied and reviewed (Kaila, 1994; Kaila et al., 1997; Staley, 2004). Given that the dentate gyrus can function as a gate, is there also a way in which the dentate gyrus might act as a filter? The prolonged IPSP in effect acts as a high-frequency filter, but is there a more specific way in which the dentate gyrus might act as a filter of information? We would like to suggest that the dentate gyrus does indeed filter information in a specific way. We prepare for discussion of dentate filtering function with the following comments on hippocampal anatomy, mossy cell function, the relation of translamellar inhibition and potentiation to associative Hebbian learning, and the role of synaptic scaling in non-associative Hebbian competition. A detailed review of hippocampal anatomy appears in other chapters of this volume. We note here only that the hippocampus has a striking lamellar structure, with the projections of the perforant pathway, the mossy fibers, the Schaffer collaterals and the alvear pathway, all appearing to be on nearly a plane (‘‘lamella’’) perpendicular to the longitudinal axis of the hippocampus (Andersen et al., 1969, 1971; Amaral and Witter, 1989). A point to be emphasized here for the discussion below is that hilar mossy cells project maximally onto granule cells that are septally and temporally displaced from the mossy cells of origin, not onto the granule cells from which the mossy cells receive input. That is, mossy cells projections are unique in being preferentially perpendicular to the plane of the lamellae (Amaral and Witter, 1989).

606

Why do mossy cells project in this way? A detailed discussion of mossy cell function appears in a separate chapter of this volume. We summarize the principal features and suggest its role in dentate-hilar function as follows: (1) mossy cells can mediate both translamellar inhibition as well as potentiation; (2) translamellar potentiation as mediated by mossy cells is weak, and an individual mossy cell, by itself, cannot cause a dentate granule cell to discharge; (3) optimal potentiation of dentate granule cells requires near-simultaneous stimulation of both the perforant and association pathways (which we refer to as ‘‘double input’’), to within a time interval on the order of about 5 ms. A caveat is that the functional width of a lamella at the level of the dentate gyrus may be as thick as 2.5 mm (Zappone and Sloviter, 2004). We also propose that granule cells scale their activity to pass only the most favored or most potentiated dentate patterns. That is, individual granule cells evaluate not only whether individual synapses are favorable or not (the traditional, associative Hebbian LTP or LTD), but also whether the number of action potentials averaged over some period of time is too low or too high. If the average activity of one particular neuron is too low, all synaptic strengths of this neuron are scaled up; and if too high, all synaptic strengths are scaled down, so as to preserve the average activity within some characteristic range (LeMasson et al., 1993; Turrigiano and Nelson, 2000). Experimental evidence for this kind of activity-dependent homeostatic synaptic scaling in other brain areas exists (Royer and Pare, 2003; Wierenga et al., 2005). Hebbian systems that are not capable of similar homeostasis of activity evolve inevitably into a state of tonic hyperactivity or global silence (Miller, 1996, Marder and Prinz, 2002). As a consequence of activity-dependent synaptic scaling, the establishment of potentiated input patterns causes the response of a neural system to non-potentiated patterns to be scaled down, even in the absence of specific LTD mechanisms for the non-potentiated patterns. That is, for synapses to survive in competition with other synapses, it is not enough that they not be specifically identified as being unfavorable; synapses will nonetheless be scaled down in strength if there are other synapses

that are systematically scaled up or potentiated. We refer to synaptic scaling as being representative of a type of non-associative Hebbian competition. Thus we suggest that translamellar potentiation be viewed in terms of Hebbian associative learning, with the additional twist that double input from both the perforant and associative pathways results in more effective potentiation. Indeed, input from only one source, e.g., the association pathway only, may actually result in depotentiation through Hebbian competition. To see this, consider repetitive perforant pathway input that arrives at dentate granule cells in a certain number of lamellae (Fig. 1). The dentate granule cells in these lamellae fire multiple action potentials. These granule cells cause mossy cells downstream in the same lamellae to fire multiple action potentials as well. These mossy cells then send signals to many other lamellae. Some of these other lamellae receive near-simultaneous perforant and association pathway input, and some do not. For lamellae that do receive near-simultaneous perforant and association pathway input, one expects the mossy

EC

DG

MC Fig. 1. Dentate-hilar potentiation is mediated by double input from both entorhinal cortex neurons and from hilar mossy cells. EC ¼ entorhinal cortex; DG ¼ dentate gyrus granule cells; MC ¼ mossy cells of hilus. Repetitive stimulation of dentate granule cells by entorhinal cortex neurons causes transmission of excitation to mossy cells in the hilus. The mossy cells then stimulate extralamellar granule cells (plus interneurons near those granule cells). Those granule cells that receive input from both entorhinal cortex and from mossy cells become potentiated, while those that do not, become depotentiated. Potentiated connections are represented by thick arrows. Depotentiated connections are represented by dashed arrows.

607

cell-to-granule cell connection to be potentiated. However, Hebbian competition then requires that non-potentiated connections be scaled down in strength. Thus lamellae that do not receive simultaneous perforant and association pathway input will find their mossy cell input weakened. What is the timescale for translamellar inhibition? A scaling mechanism should take place on a timescale that is much longer than the baseline dentate granule cell firing interval, because a scaling mechanism requires monitoring and averaging the firing rate over some period of time. One set of experiments (Zappone and Sloviter, 2004) found translamellar inhibition to appear on a timescale of 200 s. A timescale this long is consistent with a scaling mechanism, and is not consistent with direct connectivity-related effects (i.e., disynaptic mossy cell to basket cell to granule cell transmissions). Translamellar potentiation and inhibition can now be put together in a consistent scheme for system learning. Translamellar potentiation allows associative learning, while translamellar inhibition helps maintain dynamical system stability. Consistently successful double input from both perforant and association pathways results in translamellar potentiation (Steward et al., 1977; Buzsaki and Eidelberg, 1982; Strowbridge et al., 1992; Hetherington et al., 1994; Strowbridge and Schwartzkroin, 1996; Kleschevnikov and Routtenberg, 2003), while multiple inputs from mossy cells without concomitant perforant pathway input result in translamellar inhibition (Zappone and Sloviter, 2004). Translamellar inhibition and potentiation are thus complementary mechanisms, both necessary for a stable system capable of continual learning.

Dentate-hilar filtering function: a hypothesis Various ways in which mossy cells can help the dentate gyrus function as a filter have been proposed. Buckmaster and Schwartzkroin (1994) have suggested a granule cell association hypothesis, wherein the mossy cells help to link subpopulations of granule cells. They suggested that the dentate-hilar role is one of pattern recognition, where

the role of the mossy cells is to fill in missing components of perforant pathway input. For example, if the dentate-hilar complex learns to recognize a pattern involving co-activation of granule cells in lamellae A, B, and C, but later receives perforant pathway input only at lamellae A and B, then the mossy cells via association pathway potentiation will nonetheless stimulate granule cells in lamella C to fire. This type of pattern recognition is often referred to as ‘‘pattern completion’’. It is tuned to be sensitive but not specific. The mossy cells will cause granule cell co-activation in a remembered pattern, if the input pattern is ‘‘close enough’’ to the remembered pattern. That mossy cell projections are perpendicular to the perforant pathway and project preferentially to distant granule cells, sets up an ideal geometry for the mossy cells to play an associative role. It was not clear to these authors why associations between distant granule cells should be mediated by a separate cell population (the mossy cells), but it was conjectured that this arrangement allowed for independent influences to act on granule cells and mossy cells separately, and that nearest-neighbor granule cell co-activations may be discouraged, thus preventing a dangerous accretion of co-localized excitation (Buckmaster and Schwartzkroin, 1994). Alternatively, considering that granule cells are difficult to activate while mossy cells are easily activated, Jackson and Scharfman (1996) proposed that mossy cells act as a switch: ‘‘By keeping activity in the granule cells either above or below a threshold for potentiation of synapses on pyramidal cells, mossy cells could create a bistable system, and thus form a gate to control whether or not information will be stored in the downstream elements of the trisynaptic circuit’’. Other influences on the mossy cells could presumably determine whether the switch is turned on or off. We offer yet another explanation of the dentate filter, closer to that of Buckmaster and Schwartzkroin (1994) but differing in an important way. Mossy cells can help bring granule cells closer to threshold but rarely trigger granule cell action potentials by themselves (Hetherington et al., 1994; Scharfman, 1995; Kleschevnikov and

608

Routtenberg, 2003). Such weak mossy cell association does not lend itself to high-sensitivity pattern recognition, as missing components of a perforant pathway pattern are not likely to be filled in by mossy cell collateral input. The finding that temporal to septal association is absent or very weak in rats (Hetherington et al., 1994) would also lead to very poor pattern completion capabilities, as essentially there is no pattern completion capability in the temporal to septal direction. Furthermore, association pathway potentiation, as discussed in the previous section, is best triggered with double input from both the perforant and association pathways (Fig. 1). Inconsistent pattern input may not simply be ignored, but may lead to loss of potentiation and possibly even inhibition of granule cell response. Thus, we agree that the dentate-hilar complex is a pattern recognition complex, but we propose that it represents an unforgiving pattern recognizer, one that is specific but not necessarily sensitive. We propose that the dentate-hilar complex is not a pattern completer, but a pattern excluder. Its job is to exclude input patterns that are not exactly right. Dentate-hilar function, in our conjecture, thus consists of the following steps: (1) patterns are presented to the dentate-hilar complex via repetitive perforant pathway input, (2) mossy cells strengthen granule cell responses to patterns that are repeated in a consistent and persistent way (translamellar potentiation), and weaken random or erratically presented patterns (translamellar inhibition), (3) Hebbian competition, through non-associative mechanisms, scales granule cell firing thresholds to fire only with the most highly potentiated pattern or patterns, (4) with future repetitions, the most highly potentiated pattern or patterns are allowed to pass through, while more random patterns are blocked. In this model, the dentate gyrus functions as both gate and filter. It is a gate that can be opened by persistent, repetitive stimulation, and it is a filter in that it prefers that the stimulation be consistent, in terms of the pattern of the stimulation as distributed along the longitudinal axis. That is, one must ‘‘knock’’ on the gate not only many times, but in nearly exactly the same way each time. Random knocks are ignored.

What is the optimal frequency for opening the dentate gate? Comparing repetitive stimulation at 0.5, 4, 7, 10, and 20 Hz, Andersen et al. (1966) found that 10 Hz was optimal. Comparing repetitive stimulation at 0.5, 5, and 100 Hz, Mott and Lewis (1992) found that 5 Hz was optimal. The frequencies 5–10 Hz fall in the theta–alpha band, which is known to be prominent in limbic structures including the hippocampus (Bland, 1986; Freund and Buzsaki, 1996; Buzsaki, 2002; Buzsaki et al., 2003). We therefore conjecture that theta oscillations indicate activity requiring opening of the dentate gate. We comment that the dentate-hilar complex combines a sluggish but powerful excitatory source (the dentate granule cells) with a highly labile but weaker component (the mossy cells). The sluggishness of the granule cells and the weakness of mossy cell output allow the granule cells to maintain high specificity, but the lability of the mossy cells nonetheless allows highly responsive associative learning. This dual need for specificity and association may explain why dentate-hilar association is mediated by a specialized cell population (the mossy cells), while in CA3 and in the neocortex, association occurs directly between principal cells. The dentate-hilar complex is unique in being tuned for specificity.

A look ahead Our conjecture for dentate-hilar function is speculative, but testable. The simplest type of experiment would be to monitor all EPSPs from the dentate granular layer in one lamella, and to calculate the initial slope of each EPSP, denoted E in units of volts per second. One would also keep track of which E’s result in population spikes. Then a distribution function can be constructed, G(E), giving the probability of observing each value of E (Fig. 2, filled squares). One can also determine the threshold E0, defined as the smallest E above which a population spike is likely (e.g., with a likelihood greater than 95%). Alternatively, one can define various threshold functions with parameters that can be extracted from experiment. For instance, one can hypothesize a threshold

609

0.030

0.025

G(E) and Eo

0.020

0.015

0.010

0.005

0.000 0

2

4

6

8

10

E Fig. 2. Hypothetical distributions of G(E) and E0 with E in arbitrary units. Filled squares: baseline G(E). The single solid square at E ¼ 5.28 represents the threshold E0 such that 95% of the distribution has EoE0. Open triangles: hypothetical G(E) after potentiation. The threshold is at E ¼ 6.72. Open circles: hypothetical G(E) in chronic epilepsy. The threshold is at E ¼ 6.00.

function of the form PðEÞ ¼ A expðBðE
E 0 ÞÞ, where P(E) is the probability that a given E results in a population spike. Here A is a normalization constant, E0 is the firing threshold for E, and B controls how sharp is the transition to firing. The parameters B and E0 can be extracted from experiment, by plotting log P(E) vs. E. The idea of an adjustable firing threshold E0 is key to our model for dentate-hilar function. What happens to G(E) and E0 after potentiation? If one stimulates the perforant pathway repetitively at a frequency fS with a certain number of repetitions NR, one expects potentiation of the dentate response to future stimulations. The ideal frequency for potentiation may be 5–10 Hz for 6–9 s (Andersen et al., 1966; Mott and Lewis, 1992), as discussed above. The relevant E to monitor may be the median or maximal value over the NR repetitions. After potentiation, one may repeat the procedure for constructing G(E), and see how this distribution function has changed. One may hypothesize that G(E) develops a new peak at

higher E, representing potentiated EPSPs, with threshold E0 between this new peak and the old peak (Fig. 2, open triangles). EPSPs from the new high-E peak are passed by the dentate gyrus, while those from the old peak are blocked. The more distinct is this new peak, the easier it becomes to exclude incorrect patterns. Small fluctuations in the value of E0 would not greatly affect specificity. What happens to G(E) and E0 in chronic epilepsy? With loss of mossy cells, one might hypothesize that it becomes more difficult to create a distinct high-E peak. One might hypothesize, for instance, that only a high-E shoulder is created (Fig. 2, open circles). The function of the dentatehilar complex as a pattern excluder would thus be degraded. The threshold E0, furthermore, would have to be placed on a steeper part of the curve for G(E). Any slight fluctuation of E0 would cause dentate activity either to be overly inhibited (E0 too high) or overly excitable (E0 too low). The effect of recurrent excitatory mossy fiber collaterals (reviewed in Chapter 29 by Sutula and Dudek in this volume) on G(E) and E0 should also be very interesting. One might expect either a

610

high-E shoulder or a separate higher-E peak, similar to that seen due to potentiation in the normal state discussed above. However, unlike the high-E contribution mediated by mossy cells, the high-E contribution from recurrent mossy fiber collaterals carries no useful information, because the associated EPSPs are the result of local recurrent excitations. Furthermore, if the high-E contributions from recurrent mossy fiber collaterals and from mossy cells overlap, then the filtering function of the dentate-hilar complex may be severely degraded. A more ambitious experimental goal would be to determine the functional width of a lamella, and to develop techniques to stimulate and to record from individual lamellae reliably. This goal is likely to be technically challenging. The simplest alternative would be to place a single stimulating electrode into the angular bundle, one in each hemisphere, and a single recording electrode into the dentate granular layer, again one in each hemisphere. Presumably, one is guaranteed in this arrangement to have one pair of stimulating and recording electrodes in each of two distinct, nonoverlapping lamellae (one in each hemisphere). However many distinct lamellae are accessible to experiment, one may then stimulate a subset of them simultaneously and repetitively, at a certain frequency of repetition for a certain number of repetitions, NR. This frequency may again be taken in the range of 5–10 Hz (Andersen et al., 1966; Mott and Lewis, 1992). The spatial pattern of the stimuli represents the pattern to be learned. A distribution function G(E) can then be constructed for this system, with one G(E) for each lamella. Of interest would be the number of repetitions, NR, needed to teach a given target pattern, and whether the train of NR repetitions need to be repeated a certain number of times. After potentiation of the target pattern is achieved, a test of sensitivity would be to present, simultaneously, the target spatial pattern plus a random pattern of variable amplitude, and then see if the random component is blocked while the target pattern is allowed through. A test of specificity would be to present only a part of the target spatial pattern, and see how close the presented pattern has to be to the target pattern to be passed through.

The stimulation strength necessary to produce EPSPs and population spikes is also of interest. One expects a threshold effect for the stimulation strength S (in units of volts), wherein a minimal value of this necessary before population spikes are seen. A distribution function can be defined, D(E,S), giving the probability of observing a given E and S. It would be of interest to know what happens to D(E,S), E0 and S0 after potentiation, and in the context of chronic epilepsy. Finally, if it turns out to be true that the dentate-hilar complex is a pattern excluder, one may then employ similar arguments as developed in this chapter to speculate on the function of the auxiliary pathways, e.g., the direct pathways from entorhinal cortex to area CA3 (Hjorth-Simonsen and Jeune, 1972; Steward and Scoville, 1976; Witter and Amaral, 1991). One possible function for the auxiliary pathways may be to help validate signal passed into the hippocampus. We discuss this point at a little greater length below. By the activity-dependent homeostasis hypothesis (LeMasson et al., 1993; Turrigiano and Nelson, 2000), all principal neurons have a preferred target firing rate, with homeostatic mechanisms to return to this rate over some period of time if perturbed away from it. If we assume that this hypothesis applies to granule cells, then there must be some rate at which granule cells fire action potentials spontaneously, even in the absence of perforant pathway stimulation. This rate is low but cannot be zero. How can CA3 principal neurons downstream from granule cells know if the signals they receive from granule cells are due to spontaneous granule cell activity or due to perforant pathway stimulated activity? There should be some way to ignore the inevitable (if rare) spontaneous granule cell discharge, while not compromising a faithful response to legitimate, stimulated granule cell discharge. The answer may be that CA3 principal neurons have their own G(E) distribution function, and they scale their firing thresholds to fire only with the most highly potentiated inputs. Thus if a set of CA3 principal neurons have been trained to expect ‘‘double input’’ from granule cells and from entorhinal neurons via the auxiliary entorhinal-toCA3 pathway, then these CA3 principal neurons

611

are less likely to fire if they receive input only from granule cells. Thus, auxiliary pathways may play a crosschecking role, validating information arriving via the main pathways. If a confirmatory signal does not arrive via an auxiliary pathway, then information arriving via the main pathway may be ignored. The additional input from the auxiliary pathway may be needed to push a principal neuron above the firing threshold. In summary, the current wealth of experimental data on the dentate gyrus shows that the dentate gyrus does function as a gate. We further conjecture that it also functions as a highly specific pattern recognizer, or filter. The data to date do not directly address this conjecture. We suggest future experiments that may help to prove or disprove the filtering conjecture. Even if the specifics of our conjecture are wrong, we hope these experiments will deepen our insight into the structure and function of the dentate gyrus and hippocampus.

Acknowledgments I am grateful to Drs. Helen Scharfman and Thomas Sutula for helpful advice and for the opportunity to write this chapter. I also thank the American Epilepsy Society for support, and the National Institutes of Health and National Center for Research Resources K12 Roadmap, Project number 8K12RR023268-02.

References Alger, B.E. and Teyler, T.J. (1976) Long-term and short-term plasticity in the CA1, CA3, and dentate regions of the rat hippocampal slice. Brain Res., 110(3): 463–480. Amaral, D.G. and Witter, M.P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31(3): 571–591. Andersen, P., Bliss, T.V., et al. (1969) Lamellar organization of hippocampal excitatory pathways. Acta Physiol. Scand., 76(1): 4A–5A. Andersen, P., Bliss, T.V., et al. (1971) Lamellar organization of hippocampal pathways. Exp. Brain Res., 13(2): 222–238. Andersen, P., Holmqvist, B., et al. (1966) Entorhinal activation of dentate granule cells. Acta Physiol. Scand., 66(4): 448–460. Behr, J., Lyson, K.J., et al. (1998) Enhanced propagation of epileptiform activity through the kindled dentate gyrus. J. Neurophysiol., 79(4): 1726–1732.

Ben-Ari, Y., Krnjevic, K., et al. (1980) Lability of synaptic inhibition of hippocampal pyramidal cells. J. Physiol., 298: 36P–37P [proceedings]. Bland, B.H. (1986) The physiology and pharmacology of hippocampal formation theta rhythms. Prog. Neurobiol., 26(1): 1–54. Buckmaster, P.S. and Schwartzkroin, P.A. (1994) Hippocampal mossy cell function: a speculative view. Hippocampus, 4(4): 393–402. Buzsaki, G. (2002) Theta oscillations in the hippocampus. Neuron, 33(3): 325–340. Buzsaki, G., Buhl, D.L., et al. (2003) Hippocampal network patterns of activity in the mouse. Neuroscience, 116(1): 201–211. Buzsaki, G. and Eidelberg, E. (1982) Direct afferent excitation and long-term potentiation of hippocampal interneurons. J. Neurophysiol., 48(3): 597–607. Cohen, A.S., Lin, D.D., et al. (2003) Dentate granule cell GABA(A) receptors in epileptic hippocampus: enhanced synaptic efficacy and altered pharmacology. Eur. J. Neurosci., 17(8): 1607–1616. Collins, R.C., Tearse, R.G., et al. (1983) Functional anatomy of limbic seizures: focal discharges from medial entorhinal cortex in rat. Brain Res., 280(1): 25–40. Coulter, D.A. (1999) Chronic epileptogenic cellular alterations in the limbic system after status epilepticus. Epilepsia, 40(Suppl 1): S23–S33 discussion S40–S41. Deisz, R.A. and Prince, D.A. (1989) Frequency-dependent depression of inhibition in guinea-pig neocortex in vitro by GABAB receptor feed-back on GABA release. J. Physiol., 412: 513–541. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6(4): 347–470. Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of dentate gyrus granule cells. J. Neurophysiol., 51(2): 195–209. Heinemann, U., Beck, H., et al. (1992) The dentate gyrus as a regulated gate for the propagation of epileptiform activity. Epilepsy Res. Suppl., 7: 273–280. Hetherington, P.A., Austin, K.B., et al. (1994) Ipsilateral associational pathway in the dentate gyrus: an excitatory feedback system that supports N-methyl-D-aspartate-dependent long-term potentiation. Hippocampus, 4(4): 422–438. Hjorth-Simonsen, A. and Jeune, B. (1972) Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. J. Comp. Neurol., 144(2): 215–232. Iijima, T., Witter, M.P., et al. (1996) Entorhinal-hippocampal interactions revealed by real-time imaging. Science, 272(5265): 1176–1179. Jackson, M.B. and Scharfman, H.E. (1996) Positive feedback from hilar mossy cells to granule cells in the dentate gyrus revealed by voltage-sensitive dye and microelectrode recording. J. Neurophysiol., 76(1): 601–616. Kaila, K. (1994) Ionic basis of GABAA receptor channel function in the nervous system. Prog. Neurobiol., 42(4): 489–537. Kaila, K., Lamsa, K., et al. (1997) Long-lasting GABA-mediated depolarization evoked by high-frequency stimulation in pyramidal neurons of rat hippocampal slice is attributable to

612 a network-driven, bicarbonate-dependent K+ transient. J. Neurosci., 17(20): 7662–7672. Kleschevnikov, A.M. and Routtenberg, A. (2003) Long-term potentiation recruits a trisynaptic excitatory associative network within the mouse dentate gyrus. Eur. J. Neurosci., 17(12): 2690–2702. LeMasson, G., Marder, E., et al. (1993) Activity-dependent regulation of conductances in model neurons. Science, 259(5103): 1915–1917. Lothman, E.W., Stringer, J.L., et al. (1992) The dentate gyrus as a control point for seizures in the hippocampus and beyond. Epilepsy Res. Suppl., 7: 301–313. Marder, E. and Prinz, A.A. (2002) Modeling stability in neuron and network function: the role of activity in homeostasis. Bioessays, 24: 1145–1154. McCarren, M. and Alger, B.E. (1985) Use-dependent depression of IPSPs in rat hippocampal pyramidal cells in vitro. J. Neurophysiol., 53(2): 557–571. Miller, K.D. (1996) Synaptic economics: competition and cooperation in synaptic plasticity. Neuron, 17: 371–374. Mody, I. (2005) Aspects of the homeostaic plasticity of GABAA receptor-mediated inhibition. J. Physiol., 562(Pt 1): 37–46. Mody, I., Otis, T.S., et al. (1992) The balance between excitation and inhibition in dentate granule cells and its role in epilepsy. Epilepsy Res. Suppl., 9: 331–339. Mott, D.D. and Lewis, D.V. (1992) GABAB receptors mediate disinhibition and facilitate long-term potentiation in the dentate gyrus. Epilepsy Res. Suppl., 7: 119–134. Nusser, Z. and Mody, I. (2002) Selective modulation of tonic and phasic inhibitions in dentate gyrus granule cells. J. Neurophysiol., 87(5): 2624–2628. Peng, Z. and Houser, C.R. (2005) Temporal patterns of fos expression in the dentate gyrus after spontaneous seizures in a mouse model of temporal lobe epilepsy. J. Neurosci., 25(31): 7210–7220. Royer, S. and Pare, D. (2003) Conservation of total synaptic weight through balanced synaptic depression and potentiation. Nature, 422(6931): 518–522. Scanziani, M., Gahwiler, B.H., et al. (1991) Paroxysmal inhibitory potentials mediated by GABAB receptors in partially disinhibited rat hippocampal slice cultures. J. Physiol., 444: 375–396. Scharfman, H.E. (1992) Differentiation of rat dentate neurons by morphology and electrophysiology in hippocampal slices: granule cells, spiny hilar cells and aspiny ‘fast-spiking’ cells. Epilepsy Res. Suppl., 7: 93–109. Scharfman, H.E. (1995) Electrophysiological evidence that dentate hilar mossy cells are excitatory and innervate both granule cells and interneurons. J. Neurophysiol., 74(1): 179–194. Scharfman, H.E. (2004) Functional implications of seizure-induced neurogenesis. Adv. Exp. Med. Biol., 548: 192–212. Scharfman, H.E., Kunkel, D.D., et al. (1990) Synaptic connections of dentate granule cells and hilar neurons: results of paired intracellular recordings and intracellular horseradish peroxidase injections. Neuroscience, 37(3): 693–707.

Sloviter, R.S., Zappone, C.A., et al. (2006) Kainic acid-induced recurrent mossy fiber innervation of dentate gyrus inhibitory interneurons: possible anatomical substrate of granule cell hyperinhibition in chronically epileptic rats. J. Comp. Neurol., 494(6): 944–960. Staley, K.J. (2004) Role of the depolarizing GABA response in epilepsy. Adv. Exp. Med. Biol., 548: 104–109. Staley, K.J., Otis, T.S., et al. (1992) Membrane properties of dentate gyrus granule cells: comparison of sharp microelectrode and whole-cell recordings. J. Neurophysiol., 67(5): 1346–1358. Stell, B.M. and Mody, I. (2002) Receptors with different affinities mediate phasic and tonic GABA(A) conductances in hippocampal neurons. J. Neurosci., 22(10): RC223. Steward, O. and Scoville, S.A. (1976) Cells of origin of entorhinal cortical afferents to the hippocampus and fascia dentata of the rat. J. Comp. Neurol., 169(3): 347–370. Steward, O., White, W.F., et al. (1977) Potentiation of the excitatory synaptic action of commissural, associational and entorhinal afferents to dentate granule cells. Brain Res., 134(3): 551–560. Stringer, J.L. and Lothman, E.W. (1992) Reverberatory seizure discharges in hippocampal-parahippocampal circuits. Exp. Neurol., 116(2): 198–203. Stringer, J.L., Williamson, J.M., et al. (1989) Induction of paroxysmal discharges in the dentate gyrus: frequency dependence and relationship to afterdischarge production. J. Neurophysiol., 62(1): 126–135. Strowbridge, B.W., Buckmaster, P.S., et al. (1992) Potentiation of spontaneous synaptic activity in rat mossy cells. Neurosci. Lett., 142(2): 205–210. Strowbridge, B.W. and Schwartzkroin, P.A. (1996) Transient potentiation of spontaneous EPSPs in rat mossy cells induced by depolarization of a single neurone. J. Physiol., 494(Pt 2): 493–510. Sutula, T.P., Golarai, G., et al. (1992) Assessing the functional significance of mossy fiber sprouting. Epilepsy Res. Suppl., 7: 251–259. Teyler, T.J. and Alger, B.E. (1976) Monosynaptic habituation in the vertebrate forebrain: the dentate gyrus examined in vitro. Brain Res., 115(3): 413–425. Thompson, S.M. and Gahwiler, B.H. (1989a) Activity-dependent disinhibition. I. Repetitive stimulation reduces IPSP driving force and conductance in the hippocampus in vitro. J. Neurophysiol., 61(3): 501–511. Thompson, S.M. and Gahwiler, B.H. (1989b) Activity-dependent disinhibition. II. Effects of extracellular potassium, furosemide, and membrane potential on ECl- in hippocampal CA3 neurons. J. Neurophysiol., 61(3): 512–523. Thompson, S.M. and Gahwiler, B.H. (1989c) Activity-dependent disinhibition. III. Desensitization and GABAB receptormediated presynaptic inhibition in the hippocampus in vitro. J. Neurophysiol., 61(3): 524–533. Thomson, A.M., Deuchars, J., et al. (1993) Single axon excitatory postsynaptic potentials in neocortical interneurons exhibit pronounced paired pulse facilitation. Neuroscience, 54(2): 347–360.

613 Turrigiano, G.G. and Nelson, S.B. (2000) Hebb and homeostasis in neuronal plasticity. Curr. Opin. Neurobiol., 10(3): 358–364. Walther, H., Lambert, J.D., et al. (1986) Epileptiform activity in combined slices of the hippocampus, subiculum and entorhinal cortex during perfusion with low magnesium medium. Neurosci. Lett., 69(2): 156–161. Wierenga, C.J., Ibata, K., et al. (2005) Postsynaptic expression of homeostatic plasticity at neocortical synapses. J. Neurosci., 25(11): 2895–2905. Williamson, A., Spencer, D.D., et al. (1993) Comparison between the membrane and synaptic properties of human and rodent dentate granule cells. Brain Res., 622(1-2): 194–202. Winson, J. and Abzug, C. (1977) Gating of neuronal transmission in the hippocampus: efficacy of transmis-

sion varies with behavioral state. Science, 196(4295): 1223–1225. Winson, J. and Abzug, C. (1978) Neuronal transmission through hippocampal pathways dependent on behavior. J. Neurophysiol., 41(3): 716–732. Witter, M.P. and Amaral, D.G. (1991) Entorhinal cortex of the monkey: V. Projections to the dentate gyrus, hippocampus, and subicular complex. J. Comp. Neurol., 307(3): 437–459. Wong, R.K. and Watkins, D.J. (1982) Cellular factors influencing GABA response in hippocampal pyramidal cells. J. Neurophysiol., 48(4): 938–951. Zappone, C.A. and Sloviter, R.S. (2004) Translamellar disinhibition in the rat hippocampal dentate gyrus after seizure-induced degeneration of vulnerable hilar neurons. J. Neurosci., 24(4): 853–864.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 33

Role of the dual entorhinal inputs to hippocampus: a hypothesis based on cue/action (non-self/self) couplets John E. Lisman Department of Biology and Volen Center for Complex Systems, Brandeis University, Waltham, MA 02454, USA

Abstract: The hippocampus sits at the highest level of memory processing circuits and receives two major inputs, one coming from the lateral entorhinal cortex and one coming from the medial entorhinal cortex. This duality must be of fundamental importance, but its functional meaning remains unclear. A computational model used for robot navigation (Verschure, P.F., et al. (2003). Nature, 425: 620–624) has a dual information structure that may provide insight. In this model, information is stored as couplets consisting of information about the current sensory cues and information about the current action of the robot. Sequences of such couplets are stored in a short-term memory buffer and transferred to a long-term memory store whenever a goal is found. The overall system enhances the ability of the robot to find reward sites because stored sequences enable the robot to retrace the path to a goal site whenever any of the cues along the path to a goal is subsequently encountered. A review of the literature suggests that the idea of cue/ action couplets can be usefully mapped onto the function of the entorhinal cortex. Cue information may be supplied by the lateral entorhinal cortex whereas action (motor) information may be supplied by the medial entorhinal cortex. However, given that self-position information is prominent in the medial pathway and that this is not directly related to action, a modified formulation of the duality is proposed in which the fundamental distinction is between information about non-self vs. information about self. According to this view, the lateral entorhinal pathway carries information about external (non-self) cues and their positions (in egocentric coordinates) whereas the medial entorhinal pathway carries information about the organism itself, including its position (in allocentric coordinates), motor actions and goals. Keywords: short-term memory; reward; navigation; sequence memory; dopamine and medial regions. These provide excitatory input to several parts of the hippocampus, but the duality of their structure is most obvious in the dentate gyrus. The outermost dendritic region of dentate granule cells receives exclusive input from the lateral entorhinal cortex whereas the middle dendritic region receives exclusive input from the medial entorhinal cortex. Each granule cell receives both types of inputs, thereby integrating the

Overview The duality (Fig. 1) of the connections from the entorhinal cortex to the hippocampus is anatomically striking (Burwell et al., 1995). The entorhinal cortex contains two distinct regions, the lateral Corresponding author. Tel.: +1 781 736 3145 or 3148; Fax: +1 781 736 3107; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63033-7

615

616

Fig. 1. Wiring diagram of the hippocampus (interneurons excluded) showing dual inputs from the lateral and medial entorhinal cortex. The layer 2 cortical inputs to the dentate and CA3 diverge fan out (f) widely over these networks and then provide convergent input to individual dentate granule cells. In contrast, the layer 3 cortical inputs are specialized for individual subregions of CA1. This is an example of a point-to-point (p-p) connection. Within the hippocampus, granule cells provide input, via mossy fibers to the mossy cells of the dentate and to CA3 cells. CA3 cells make feedback connections to themselves and to the mossy cells of the dentate. These, in turn, provide excitatory input to granule cells in the inner third of the granule cell dendritic tree. (See Color Plate 33.1 in color plate section.)

two lines of cortical information. The purpose of this review is to discuss the possible functional basis of this mysterious duality. A priori, what are the grand dualities that might be considered? Here is a list of some possibilities: what/where; specific/context; sensory/motor; past/ present; conscious/unconscious; rewarded/punished; stimulus/response. My interest in the last of these, stimulus/response, was stimulated by a paper on robotic control (Verschure et al., 2003). In that paper, the authors describe a ‘‘brain-like’’ computer program that enables a robot to efficiently find sites at which reward is located. The authors postulated several levels of control. At the lowest levels, circuits support classical conditioning. In this way, the robot learns to associate a previously neutral stimulus with a reward that is close to the robot.

But the robot becomes much more efficient at finding reward sites if there are additional circuits. These make it possible for a cue that is far away from a reward site to specify a complex path that, according to previous experience, led to the reward site. The circuitry that allows the robot to do this (Fig. 2) involves a ‘‘cortical’’ multi-item ‘‘circular’’ short-term memory (STM) buffer that stores the last five salient events, and a long-term ‘‘hippocampal’’ memory that stores sequences of events in long-term memory (LTM). When the robot forages and accidentally finds a reward, it incorporates the entire content of the buffer into its LTM store. Importantly, each event is defined as a stimulus/response couplet (Verschure and Voegtlin, 1998; Verschure et al., 2003). The ‘‘stimulus’’ part of the couplet corresponds to a prototype of the sensations from the external world at a particular

617

Fig. 2. Diagram of circuits responsible for storing events that led to the discovery of goals according to the model of Verschure et al. (2003). Stimulus (C) and motor action (A) for each event are stored in a circular (first in, first out) short-term memory buffer. This multi-item buffer can store five C/A couplets. When a goal is reached, the sequence in the short-term memory buffer is transferred to the long-term memory buffer together with a last item representing the goal (G). As illustrated, the long-term memory store has stored six sequences.

location. The response part of the couplet corresponds to the motor act that was executed to get from that cue to the next location (cue) in the sequence. ‘‘Stimulus’’ and ‘‘response’’ have a particular meaning in classical conditioning that is not appropriate in the current context; I will therefore use the terminology cue/action (C/A) to describe a couplet. Thus the utilization of stored couplets can be described as follows. If the robot comes across a previously encountered cue, Cn, it can get to the reward site using a simple algorithm: execute the action An, that is associated with Cn in LTM; when Cn+1 is found, execute An+1, etc. Several comments are in order. First, Verschure (personal communication) was not aware of the dual inputs to the hippocampus; the dual representation was utilized simply because it enhanced the robot’s efficiency at finding goals. Second the use of a sequence of cue-driven actions to find a goal site is an experimentally observed mode of animal navigation (Collett et al., 2003) and is related to ideas incorporated into previous reinforcement driven models of behavior (Barto and

Sutton, 1981; Hasselmo, 2005). This form of navigation is less flexible than map-based navigation, which may have developed later in the evolution. Importantly, the use of C/A sequences to specify the route to an accidentally discovered reward site absolutely relies on one-trial learning, a property that remains a defining property of human hippocampal episodic memory. Indeed, the development of a multi-item STM buffer may have been the crucial evolutionary change that made one-trial learning possible rather than a change in synaptic plasticity mechanisms themselves. This is because such buffers capture one-trial, brief events and then, upon command, provide the repetitive activity patterns to LTM networks that are necessary to produce stable synaptic modifications. In the following sections I will address two questions: 1. Is the idea of storing C/A couplets potentially applicable to understanding the dual cortical inputs to the hippocampus? In particular, is there any way of mapping this duality onto

618

the known properties of the medial and lateral entorhinal inputs to the hippocampus? 2. What is the status of the evidence for the various building blocks postulated in the model of Fig. 2?

Evaluation of the cue/action (C/A) couplet hypothesis Evidence that the hippocampus receives action (motor) and cue (sensory) information The idea that the hippocampus might be influenced by the motor system is one that is not commonly discussed, but support for this idea is actually quite long-standing. Vanderwolf (1969) observed that the frequency and amplitude of hippocampal theta oscillations is modulated by the speed of a rat [reviewed in Bland and Oddie (2001)]. Other work shows the influence of motor set and running speed on place cell firing (McNaughton et al., 1983; Foster et al., 1989; Wiener et al., 1989). Ranck (1973) found hippocampal neurons that fired on ‘‘approach’’ to objects. More recent work provides direct evidence that ‘‘motor intent’’ influences hippocampal function. Specifically, when rats run in a T-maze, the firing of neurons that occurs when the rat is in the vertical stem of the T can be different depending on which direction the rat will turn at the top of the stem (Frank et al., 2000; Wood et al., 2000; Bower et al., 2005). Because the sensory stimulation in the stem is always identical, it is hard to escape the conclusion that goals or action are what cause the difference in firing. It might be argued that what is represented is the goal, in some abstract sense not related to the motor action needed to get to goal, but this is not the case. In experiments using a plus maze, rats were started in the north or south arm and had to reach a goal in the east arm; when the rat reached the east arm, many cells fired differentially depending on whether the approach was from the north or south (Ferbinteanu and Shapiro, 2003). Thus, although a common goal was involved, information about the specific path taken was encoded. Recent results emphasize that pathway-specific firing may occur

only when the behavioral paradigm has consequences for reward (Smith and Mizumori, 2006). The influence of goals on hippocampal responses is not limited to rats, but is also evident in humans (Ekstrom et al., 2003). There is also strong experimental support for the entry of sensory-specific information into the hippocampus. Odor-specific cells are found in the rat hippocampus (Wood et al., 1999), and neurons that are scene-dependent have been observed in monkey hippocampus (Wirth et al., 2003). Rat hippocampal neurons become sensitive to tones after aversive conditioning (Berger et al., 1976; Moita et al., 2003). Particularly clear evidence for cells responsive to particular faces has been obtained in humans (Quiroga et al., 2005). It thus seems clear that both motor and sensory information can get to the hippocampus. Moreover, particularly important from the perspective of the hypothesis developed here is the finding of cells in both rat (Ranck, 1973) and monkey (Wirth et al., 2003) that jointly encode sensory and motor information. For instance, in the monkey work, particular pictures are rewarded for movements in a particular direction; as learning becomes apparent, some hippocampal neurons fire selectively when a particular picture is presented and the monkey performs the correct learned movement.

Evidence that the hippocampus is necessary for sensory/motor associations An important series of experiments has addressed whether the hippocampus is necessary for learning the association between sensory cues and motor responses. It was found that lesions to the fornix, a major input/output tract of the hippocampus, prevent a monkey from learning which direction to move in response to a sensory cue (Rupniak and Gaffan, 1987), and interferes with the ability of a monkey to report the previous movement that it made (Gaffan, 1985). Recent work (Brasted et al., 2005) extends these observations by showing that fornix lesions prevent learning of arbitrary sensory/motor tasks in single trials, the fast learning that is required to enhance foraging (see above). Related work on instrumental conditioning in rats

619

demonstrates a similar requirement for hippocampal involvement (Corbit and Balleine, 2000). Taken together, these observations indicate that hippocampal cells code for much more than position and are influenced by the goal and actions required to get there. Moreover, the association of these two types of information in the hippocampus appears necessary for behavioral responses based on this association. I now turn to the issue of which region of the entorhinal cortex supplies each type of information.

Sensory properties of the lateral entorhinal cortex The perirhinal cortex, often termed inferotemporal cortex, is involved in high-level aspects of object recognition (Brown and Aggleton, 2001). This region provides strong input to the lateral entorhinal cortex, but little to the medial entorhinal (Burwell, 2000). Physiological evidence for sensory input to the lateral entorhinal cortex is its responsiveness to stimulation of olfactory cortex in rat (Gnatkovsky et al., 2004) and the presence of novelty-sensitive responses to visually presented objects in monkey (Fahy et al., 1993). Thus, the idea that the lateral perforant pathway is sensory driven has experimental support. Consistent with a sensory role, lesions of the hippocampal inputs from the lateral entorhinal cortex produces decreased investigation of novel objects (Myhrer, 1988).

Motor/goal properties of the medial entorhinal cortex In contrast to the cells of the lateral entorhinal cortex, which do not have spatial properties (Hargreaves et al., 2005), layers 2 and 3 of the medial entorhinal cortex contain grid cells with robust spatial properties (Fyhn et al., 2004; Hafting et al., 2005; Sargolini et al., 2006). These cells are coded in an allocentric coordinate system (i.e. with respect to an absolute reference frame in the environment, as e.g. north/south). Consistent with the presence of such allocentric spatial information, lesions of the hippocampal inputs from the medial entorhinal (but not the lateral entorhinal cortex)

interfere with place learning (Ferbinteanu et al., 1999). Validity of the C/A duality In mapping the C/A duality onto the entorhinal cortex, we would have to suppose that the lateral region is sensory (cue) and that the medial region represents ‘‘action’’. However, the most abundant cells in the medial region are grid cells, representing the position of the rat in the environment, and such information does not seem directly tied to action. At best one could argue that motor information may be used to compute position (McNaughton et al., 2006); the computation is not dramatically altered when important sensory cues, such as vision, are taken away, probably because a path integration mechanism can compute position using head direction and velocity information, both of which are motor-related. A significant number of layer 3 cells are directiondependent, providing additional suggestive evidence that goal/motor information is present in the medial entorhinal cortex (Sargolini et al., 2006). Thus, to some extent, positional cells could be described as related to action, but this argument is not compelling. Another major theory of the duality (what/where) also faces difficulties. If one views grid cells as primarily representing ‘‘where’’, it at first seems sensible to conclude (Hargreaves et al., 2005) that the dual inputs to the hippocampus form a what/where couplet, not a C/A couplet. However, does the ‘‘where’’ in this formulation refer to the organism or to objects in the environment? The brain needs to keep track of both, and it is unclear how this would be done in the context of a what/where duality. Cue/action as a special case of a non-self/self duality To deal with these difficulties, I suggest a reformulation of the duality as non-self information vs. self-information. According to this view, the lateral entorhinal pathway would carry information about what is in the environment (non-self) and the ‘‘where’’ of those objects. The coding of the

620

‘‘where’’ component would presumably be done in the egocentric (relative to the observer) coordinates of the parietal lobe. The medial entorhinal pathway would encode information about various aspects of self that includes a general specification of where one is in the environment (in allocentric coordinates — relative to a fixed map of the environment) and the action being taken towards achieving goals. Importantly, the allocentric coding evident in the medial entorhinal cortex could provide an excellent way of specifying directional action (e.g. turn North) because the correct action can be triggered by a cue approached from any direction. In contrast, egocentric specification (e.g. turn Right) would work only if the cue is approached from the same angle as during initial learning. To give a specific example of how non-self/self information might be encoded by a lateral/medial entorhinal couplet, consider the following description of a moment in my morning commute. When I see the following non-self information: [McDonald’s and, to its right, Burger King (complex cue specified in egocentric coordinates)], this would be coupled to the following self information: [given that I am on my way to work (my goal), when I’m near Brandeis (my place approximated in allocentric coordinates), I turn West (my action in allocentric coordinates)]. Parenthetically, one can see from this example that purely landmark-based navigation would fail if there are many similar landmarks, as is the case for fast food establishments, and that even a crude allocentric system (‘‘near Brandeis’’) can therefore be helpful in disambiguating landmark cues. It would seem that the non-self/self formulation is a plausible one and fits the data somewhat better than the C/A formulation. In the final sections I will turn to the question of how this formulation might be tested.

Evidence for the building blocks required by the model (Fig. 2) The general model of Verschure et al. requires several important building blocks: a circular, multiitem STM buffer, a LTM capable of holding couplet sequences, and a control system that would

transfer the buffer content to the LTM store when a reward site is found. I will now review what is known about the existence of such components. Evidence for the circular multi-item STM buffer in cortex A requirement of reward-dependent storage of long paths leading to the reward is the existence of a multi-item ‘‘circular’’ buffer. When the reward site is found, the contents of the buffer (the C/A sequences leading to the reward site) are transferred into the LTM store. The existence of a multi-item STM in humans has been deduced from psychophysical studies (Atkinson and Shiffrin, 1968). This buffer is thought to have limited capacity (772 items) and to be ‘‘circular’’. Here ‘‘circular’’ means that when the buffer is full, the next arriving item knocks out the item that has been in the buffer the longest (i.e. first in, first out). In the context of the model shown in Fig. 2, the circular property ensures that when the goal is reached, transfer of buffer information into LTM will incorporate information about the last five events that occurred before the goal was found. A physiologically plausible model of a multiitem STM buffer has been developed (Lisman and Idiart, 1995) and a recent variant has first in, first out properties (Koene and Hasselmo, 2006). fMRI evidence points to the temporal lobe as a site of a multi-item working memory buffer, as evidenced by the load dependence of the fMRI signal (Fiebach et al., 2006). There is evidence that the entorhinal cortex can maintain STM information (Otto and Eichenbaum, 1992) [reviewed in Jensen and Lisman (2005); Hasselmo and Stern (2006)]. Recent experiments indicate that the persistent firing is due to a cholinergically enhanced intrinsic conductance that produces an after depolarization (Fransen et al., 2006), as postulated in the theoretical models (Lisman and Idiart, 1995). Evidence for reward-mediated fixation of STM content into hippocampal LTM It is generally thought that reward signaling is mediated by the dopaminergic cells of the ventral

621

tegmental area and substantia nigra. It was originally thought that dopamine did not affect the hippocampus, but recent work indicates that there is dopaminergic innervation of the hippocampus and that it can profoundly affect transmission and enhance LTP [reviewed in Lisman and Grace (2005)]. In the context of the model of Fig. 2, one might imagine that it is the dopamine reward signal that stimulates reward-dependent transfer of information from STM to LTM, but no experiments have addressed this directly, so not much can be said about this possibility at the present time. One relevant observation (Foster and Wilson, 2006) is that when rats reach a reward site, a replay of the sequence of events (places) that led the rat to the reward occurs in the hippocampus (the replay occurs in reverse order). Although it is suspected that this replay occurs as a result of processes within the hippocampus itself, the possibility that it actually occurs because of input from the cortex needs to be directly tested. Although the role of dopamine in transfer, is uncertain, there is quite strong evidence that dopamine is necessary for the late stage of LTP (reviewed in Lisman and Grace, 2005). In this way, dopamine may contribute to the reward-dependent storage of sequences envisioned in Fig. 2.

Evidence for hippocampal LTM stores capable of storing sequences There have been many computational models of the hippocampus (Burgess et al., 2001; Kunec et al., 2005; Rolls and Kesner, 2006), but most deal exclusively with autoassociational memories that encode events at a given moment, and so cannot be used to account for sequences of events. Models developed in my laboratory, which have been improved over time (Jensen and Lisman, 1996; Lisman, 1999; Lisman et al., 2005), deal with how the hippocampus can store and recall sequences (see also Levy (1996)]. This class of models shows how gamma and theta oscillations, standard NMDA receptor-mediated plasticity processes and known anatomical connections (including the feedback connections from CA3 to dentate) can mediate

successful transfer of sequences from a multi-item cortical STM buffer to hippocampal LTM. It may be instructive to consider how such processes might transfer C/A sequences from STM to LTM and later use the stored information to recall them. During learning, the elements of C from lateral entorhinal cortex and the elements of A from medial entorhinal cortex converge on subsets of granule cells; each granule cell that fires forms an element of the C/A representation. The output of active granule cells excites a subset of CA3 pyramidal cells, and these are combined into an associative representation of the C/A couplet by LTP in the recurrent connections of the active CA3 cells. The CA3 output is then sent back to the dentate, where the synapses onto granule cells representing elements of the next C/A couplet in the sequence become potentiated. These feedback synapses thus form linkages between sequential events. In this way, the entire sequence can be transferred from a cortical buffer to the hippocampal memory [see Lisman et al. (2005) for details]. Now let us assume that the organism subsequently comes across C2, and the corresponding input is supplied to the dentate, which then provides input to CA3. Through autoassociative memory of the stored couplet, the A2 component of the couplet will be evoked. The completed C2/A2 couplet is sent back to the dentate, where it evokes the C3/A3 couplet (this is termed a chaining step). The vagaries of synaptic transmission will, however, lead to a slightly corrupted representation of the couplet, which, if used to chain to the next element, would lead to an even more corrupted version of C4/A4. However, the reciprocal connections between the dentate and CA3 can serve to prevent such concatenation of errors (Lisman, 1999). Specifically, C3A3 sent to CA3 where the autoassociative properties of CA3 correct the errors in the C3/A3 representation and it is this corrected version that is sent from CA3 to the dentate, where it triggers C4/A4. Thus, through sequential chaining and correction processes, the dentate and CA3 can lead to accurate recall of the entire sequence of couplets leading to a goal. Other brain circuits could then use this stored information to make a decision about whether this goal is worthwhile in the current context and, if so, to

622

execute the specific motor plans stored in the couplet sequence. Although most computational models of the hippocampus posit that the sole function of the dentate is to produce orthogonalization (the generation of very different firing patterns to slightly different input patterns), it is clear from the anatomy that a second function, the mixing of lateral and medial inputs, must also be important. If the ideas proposed here are correct, dentate and CA3 cells should display a unique code that is a mixture of sensory cues and action. Importantly, the mixing of cortical inputs to CA1 is already quite different; the region close to CA3 gets exclusive input from the medial entorhinal cortex whereas the region closer to the subiculum gets exclusive input from the lateral entorhinal input (Fig. 1). The same segregation holds for the CA1 projections back to cortex. This pattern of segregation is different from dentate/ CA3, where all cells receive both lateral and medial information, thus presumably mixing them into a joint representation. CA1 does not appear to be part of this coding system; it is thus likely to have a role in converting the dentate/CA3 code back into the separate codes of the lateral and medial entorhinal cortex (Lisman, 1999).

Concluding remarks and predictions of the model My goal in this discussion has been to examine the ideas incorporated into the model of Fig. 2 and to assess whether they can be usefully mapped onto the architecture of the hippocampus (Fig. 1). My general conclusion is that the correspondence is promising. The basic building blocks utilized in the model have a reasonable correspondence to the capabilities of the cortex and hippocampus. The more specific question, whether any of the functional dualities inspired by the model (C/A, non-self/self) underlie the structural duality of entorhinal inputs, cannot yet be answered with any certainty. The idea that the medial entorhinal encodes aspects of self whereas the lateral entorhinal encodes the outside world (non-self) leads to experimentally testable predictions, as outlined in the following paragraphs.

Representation of self-action on the T-maze in the medial entorhinal cortex It should be possible to test the hypothesis proposed here by recording from the entorhinal cortex during the T- and plus-maze paradigms that demonstrate motor/action information in the hippocampus. The specific prediction is that the medial, but not the lateral entorhinal cortex should encode motor/goal information. Preliminary evidence indicates that this prediction is correct (Lipton et al., 2006).

Unique conjunctions of non-self/self in the dentate/ CA3; cue information in the lateral entorhinal cortex A second set of predictions has to do with the changes in representation. The dentate and CA3 should have cells that represent conjunction of non-self and self (e.g. cue and action) and this conjunction should not be present in layers 2 and 3 of the entorhinal cortex. An example of such a conjunction would be a cell that responded to scene, but only in conjunction with a cued action (Wirth et al., 2003). In these experiments, the scene and cue information should be encoded in the lateral entorhinal; the action taken (which may or may not be consistent with the cue) should be encoded in the medial entorhinal.

Conditioned fearful stimuli vs. the emotion of fear The output of the basolateral amygdala represents a sensory signal, the conditioned stimulus (CS), and may well be source of the hippocampal response to the CS that develops after conditioning (Berger et al., 1976). Because these signals are nonself similar signals would be predicted in the lateral entorhinal cortex. In contrast the further processing in the amygdala that occurs in the central nucleus (Pare et al., 2004; Balleine and Killcross, 2006) is thought to represent the emotion of fear and to evoke fear-related responses in the hypothalamus and brainstem. Such emotional responses are a property of self and would be predicted to occur in the medial entorhinal cortex.

623

Connectivity with areas mediating personal/extrapersonal attention If there is a fundamental segregation of information about self and non-self in the entorhinal cortex, might one expect to find functionally related subdivisions at earlier stages of cortical processing? In this regard it is interesting that very recent work (Committeri et al., 2007; Ortigue et al., 2006) points to two very different forms of hemi-neglect due to brain injuries. One form, termed ‘‘extra-personal’’, leads to neglect of one half of the external world; the other form, termed ‘‘personal’’ leads to neglect of half of the body. These forms can be doubly dissociated and involve injuries to different brain regions. It will be of interest to determine whether these regions have selective connections to the medial and lateral entorhinal cortex. Note added in proof I have recently become aware of other works on cortical and subcortical processing that posits a separate neural system for self-referential processing (Buckner and Carroll, 2007; Northoff et al., 2006). Acknowledgments I thank Andre Fenton, Menno Witter, Bruce McNaughton, Edvard Moser, James Knierim, Steven Wise, Howard Eichenbaum, David Redish, and Matthew Shapiro for useful conversations. I thank Paul Verschure for comments on the manuscript. This work was supported by NIH grants R01 NH027337, P50 MH060450-07A1, and R01 MH61975-01A2. Special thanks also to the Dart Foundation and the Marine Biological Laboratory. References Atkinson, R. and Shiffrin, R. (1968) Human memory: a proposed system and its control processes. In: Spence K. (Ed.), The psychology of learning and motivation: advances in research and theory, Vol. 2. Academic Press, New York. Balleine, B.W. and Killcross, S. (2006) Parallel incentive processing: an integrated view of amygdala function. Trends Neurosci., 29: 272–279.

Barto, A.G. and Sutton, R.S. (1981) Landmark learning: an illustration of associative search. Biol. Cybern., 42: 1–8. Berger, T.W., Alger, B. and Thompson, R.F. (1976) Neuronal substrate of classical conditioning in the hippocampus. Science, 192: 483–485. Bland, B.H. and Oddie, S.D. (2001) Theta band oscillation and synchrony in the hippocampal formation and associated structures: the case for its role in sensorimotor integration. Behav. Brain Res., 127: 119–136. Bower, M.R., Euston, D.R. and McNaughton, B.L. (2005) Sequential-context-dependent hippocampal activity is not necessary to learn sequences with repeated elements. J. Neurosci., 25: 1313–1323. Brasted, P.J., Bussey, T.J., Murray, E.A. and Wise, S.P. (2005) Conditional motor learning in the nonspatial domain: effects of errorless learning and the contribution of the fornix to one-trial learning. Behav. Neurosci., 119: 662–676. Brown, M.W. and Aggleton, J.P. (2001) Recognition memory: what are the roles of the perirhinal cortex and hippocampus? Nat. Rev. Neurosci., 2: 51–61. Buckner, R.L. and Carroll, D.C. (2007) Self-projection and the brain. Trends Cogn. Sci., 11: 49–57. Burgess, N., Becker, S., King, J.A. and O’Keefe, J. (2001) Memory for events and their spatial context: models and experiments. Philos. Trans. R. Soc. Lond. B Biol. Sci., 356: 1493–1503. Burwell, R.D. (2000) The parahippocampal region: corticocortical connectivity. Ann. N.Y. Acad. Sci., 911: 25–42. Burwell, R.D., Witter, M.P. and Amaral, D.G. (1995) Perirhinal and postrhinal cortices of the rat: a review of the neuroanatomical literature and comparison with findings from the monkey brain. Hippocampus, 5: 390–408. Collett, T.S., Graham, P. and Durier, V. (2003) Route learning by insects. Curr. Opin. Neurobiol., 13: 718–725. Committeri, G., Pitzalis, S., Galati, G., Patria, F., Pelle, G., Sabatini, U., Castriota-Scanderbeg, A., Piccardi, L., Guariglia, C. and Pizzamiglio, L. (2007) Neural bases of personal and extrapersonal neglect in humans. Brain, 130: 431–441. Corbit, L.H. and Balleine, B.W. (2000) The role of the hippocampus in instrumental conditioning. J. Neurosci., 20: 4233–4239. Ekstrom, A.D., Kahana, M.J., Caplan, J.B., Fields, T.A., Isham, E.A., Newman, E.L. and Fried, I. (2003) Cellular networks underlying human spatial navigation. Nature, 425: 184–188. Fahy, F.L., Riches, I.P. and Brown, M.W. (1993) Neuronal activity related to visual recognition memory: long-term memory and the encoding of recency and familiarity information in the primate anterior and medial inferior temporal and rhinal cortex. Exp. Brain Res., 96: 457–472. Ferbinteanu, J., Holsinger, R.M. and McDonald, R.J. (1999) Lesions of the medial or lateral perforant path have different effects on hippocampal contributions to place learning and on fear conditioning to context. Behav. Brain Res., 101: 65–84.

624 Ferbinteanu, J. and Shapiro, M.L. (2003) Prospective and retrospective memory coding in the hippocampus. Neuron, 40: 1227–1239. Fiebach, C.J., Rissman, J. and D’Esposito, M. (2006) Modulation of inferotemporal cortex activation during verbal working memory maintenance. Neuron, 51: 251–261. Foster, D.J. and Wilson, M.A. (2006) Reverse replay of behavioural sequences in hippocampal place cells during the awake state. Nature, 440: 680–683. Foster, T.C., Castro, C.A. and McNaughton, B.L. (1989) Spatial selectivity of rat hippocampal neurons: dependence on preparedness for movement. Science, 244: 1580–1582. Frank, L.M., Brown, E.N. and Wilson, M. (2000) Trajectory encoding in the hippocampus and entorhinal cortex. Neuron, 27: 169–178. Fransen, E., Tahvildari, B., Egorov, A.V., Hasselmo, M.E. and Alonso, A.A. (2006) Mechanism of graded persistent cellular activity of entorhinal cortex layer v neurons. Neuron, 49: 735–746. Fyhn, M., Molden, S., Witter, M.P., Moser, E.I. and Moser, M.B. (2004) Spatial representation in the entorhinal cortex. Science, 305: 1258–1264. Gaffan, D. (1985) Hippocampus: memory, habit and voluntary movement. Philos. Trans. R. Soc. Lond. B Biol. Sci., 308: 87–99. Gnatkovsky, V., Uva, L. and de Curtis, M. (2004) Topographic distribution of direct and hippocampus-mediated entorhinal cortex activity evoked by olfactory tract stimulation. Eur. J. Neurosci., 20: 1897–1905. Hafting, T., Fyhn, M., Molden, S., Moser, M.B. and Moser, E.I. (2005) Microstructure of a spatial map in the entorhinal cortex. Nature, 436: 801–806. Hargreaves, E.L., Rao, G., Lee, I. and Knierim, J.J. (2005) Major dissociation between medial and lateral entorhinal input to dorsal hippocampus. Science, 308: 1792–1794. Hasselmo, M.E. (2005) A model of prefrontal cortical mechanisms for goal-directed behavior. J. Cogn. Neurosci., 17: 1115–1129. Hasselmo, M.E. and Stern, C.E. (2006) Mechanisms underlying working memory for novel information. Trends. Cogn. Sci., 10: 487–493. Jensen, O. and Lisman, J.E. (1996) Hippocampal CA3 region predicts memory sequences: accounting for the phase precession of place cells. Learn. Mem., 3: 279–287. Jensen, O. and Lisman, J.E. (2005) Hippocampal sequence-encoding driven by a cortical multi-item working memory buffer. Trends Neurosci., 28: 67–72. Koene, R.A. and Hasselmo, M.E. (2006) First-in-first-out item replacement in a model of short-term memory based on persistent spiking. Cereb. Cortex (epub ahead of print). Kunec, S., Hasselmo, M.E. and Kopell, N. (2005) Encoding and retrieval in the CA3 region of the hippocampus: a model of theta-phase separation. J. Neurophysiol., 94: 70–82.

Levy, W.B. (1996) A sequence predicting CA3 is a flexible associator that learns and uses context to solve hippocampallike tasks. Hippocampus, 6: 579–590. Lipton, P.A., White, J.A. and Eichenbaum, H. (2006) Disambiguation of overlapping spatial sequences in medial entorhinal cortex. In: Annual Society for Neuroscience Meeting. Atlanta, GA: Society for Neuroscience, Program #66.6/ Poster #Z23. Lisman, J.E. (1999) Relating hippocampal circuitry to function: recall of memory sequences by reciprocal dentate-CA3 interactions. Neuron, 22: 233–242. Lisman, J.E. and Grace, A.A. (2005) The hippocampal-VTA loop: controlling the entry of information into long-term memory. Neuron, 46: 703–713. Lisman, J.E. and Idiart, M.A. (1995) Storage of 7+/ 2 shortterm memories in oscillatory subcycles. Science, 267: 1512–1515. Lisman, J.E., Talamini, L.M. and Raffone, A. (2005) Recall of memory sequences by interaction of the dentate and CA3: a revised model of the phase precession. Neural Netw., 18: 1191–1201. McNaughton, B.L., Barnes, C.A. and O’Keefe, J. (1983) The contributions of position, direction, and velocity to single unit activity in the hippocampus of freely-moving rats. Exp. Brain Res., 52: 41–49. McNaughton, B.L., Battaglia, F.P., Jensen, O., Moser, E.I. and Moser, M.B. (2006) Path integration and the neural basis of the ‘cognitive map’. Nat. Rev. Neurosci., 7: 663–678. Moita, MA., Rosis, S., Zhou, Y., LeDoux, J.E. and Blair, H.T. (2003) Hippocampal place cells acquire location-specific responses to the conditioned stimulus during auditory fear conditioning. Neuron, 37: 485–497. Myhrer, T. (1988) Exploratory behavior and reaction to novelty in rats with hippocampal perforant path systems disrupted. Behav. Neurosci., 102: 356–362. Northoff, G., Heinzel, A., de Greck, M., Bermpohl, F., Dobrowolny, H. and Panksepp, J. (2006) Self-referential processing in our brain – a meta-analysis of imaging studies on the self. Neuroimage, 31: 440–457. Ortigue, S., Megevand, P., Perren, F., Landis, T. and Blanke, O. (2006) Double dissociation between representational personal and extrapersonal neglect. Neurology, 66: 1414–1417. Otto, T. and Eichenbaum, H. (1992) Complementary roles of the orbital prefrontal cortex and the perirhinal-entorhinal cortices in an odor-guided delayed-nonmatching-to-sample task. Behav. Neurosci., 106: 762–775. Pare, D., Quirk, G.J. and Ledoux, J.E. (2004) New vistas on amygdala networks in conditioned fear. J. Neurophysiol., 92: 1–9. Quiroga, R.Q., Reddy, L., Kreiman, G., Koch, C. and Fried, I. (2005) Invariant visual representation by single neurons in the human brain. Nature, 435: 1102–1107. Ranck Jr., J.B. (1973) Studies on single neurons in dorsal hippocampal formation and septum in unrestrained rats. I. Behavioral correlates and firing repertoires. Exp. Neurol., 41: 461–531.

625 Rolls, E.T. and Kesner, R.P. (2006) A computational theory of hippocampal function, and empirical tests of the theory. Prog. Neurobiol., 79: 1–48. Rupniak, N.M. and Gaffan, D. (1987) Monkey hippocampus and learning about spatially directed movements. J. Neurosci., 7: 2331–2337. Sargolini, F., Fyhn, M., Hafting, T., McNaughton, B.L., Witter, M.P., Moser, M.B. and Moser, E.I. (2006) Conjunctive representation of position, direction, and velocity in entorhinal cortex. Science, 312: 758–762. Smith, D.M. and Mizumori, S.J. (2006) Learning-related development of context-specific neuronal responses to places and events: the hippocampal role in context processing. J. Neurosci., 26: 3154–3163. Vanderwolf, C.H. (1969) Hippocampal electrical activity and voluntary movement in the rat. Electroencephalogr. Clin. Neurophysiol., 26: 407–418. Verschure, P.F. and Voegtlin, T. (1998) A bottom up approach towards the acquisition and expression of sequential

representations applied to a behaving real-world device: Distributed Adaptive Control III. Neural Netw., 11: 1531–1549. Verschure, P.F., Voegtlin, T. and Douglas, R.J. (2003) Environmentally mediated synergy between perception and behaviour in mobile robots. Nature, 425: 620–624. Wiener, S.I., Paul, C.A. and Eichenbaum, H. (1989) Spatial and behavioral correlates of hippocampal neuronal activity. J. Neurosci., 9: 2737–2763. Wirth, S., Yanike, M., Frank, L.M., Smith, A.C., Brown, E.N. and Suzuki, W.A. (2003) Single neurons in the monkey hippocampus and learning of new associations. Science, 300: 1578–1581. Wood, E.R., Dudchenko, P.A. and Eichenbaum, H. (1999) The global record of memory in hippocampal neuronal activity. Nature, 397: 613–616. Wood, E.R., Dudchenko, P.A., Robitsek, R.J. and Eichenbaum, H. (2000) Hippocampal neurons encode information about different types of memory episodes occurring in the same location. Neuron, 27: 623–633.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 34

The CA3 ‘‘backprojection’’ to the dentate gyrus Helen E. Scharfman Department of Pharmacology and Neurology, Columbia University, New York, NY 10032, USA and the Center for Neural Recovery and Rehabilitation Research, Helen Hayes Hospital, New York State Department of Health, Rte 9W, West Haverstraw, NY 10993-1195, USA

Abstract: The hippocampus is typically described in the context of the trisynaptic circuit, a pathway that relays information from the perforant path to the dentate gyrus, dentate to area CA3, and CA3 to area CA1. Associated with this concept is the assumption that most hippocampal information processing occurs along the trisynaptic circuit. However, the entorhinal cortex may not be the only major extrinsic input to consider, and the trisynaptic circuit may not be the only way information is processed in hippocampus. Area CA3 receives input from a variety of sources, and may be as much of an ‘‘entry point’’ to hippocampus as the dentate gyrus. The axon of CA3 pyramidal cells targets diverse cell types, and has commissural projections, which together make it able to send information to much more of the hippocampus than granule cells. Therefore, CA3 pyramidal cells seem better designed to spread information through hippocampus than the granule cells. From this perspective, CA3 may be a point of entry that receives information which needs to be ‘‘broadcasted,’’ whereas the dentate gyrus may be a point of entry that receives information with more selective needs for hippocampal processing. One aspect of the argument that CA3 pyramidal cells have a widespread projection is based on a part of its axonal arbor that has received relatively little attention, the collaterals that project in the opposite direction to the trisynaptic circuit, ‘‘back’’ to the dentate gyrus. The evidence for this ‘‘backprojection’’ to the dentate gyrus is strong, particularly in area CA3c, the region closest to the dentate gyrus, and in temporal hippocampus. The influence on granule cells is indirect, through hilar mossy cells and GABAergic neurons of the dentate gyrus, and appears to include direct projections in the case of CA3c pyramidal cells of ventral hippocampus. Physiological studies suggest that normally area CA3 does not have a robust excitatory influence on granule cells, but serves instead to inhibit it by activating dentate gyrus GABAergic neurons. Thus, GABAergic inhibition normally controls the backprojection to dentate granule cells, analogous to the way GABAergic inhibition appears to control the perforant path input to granule cells. From this perspective, the dentate gyrus has two robust glutamatergic inputs, entorhinal cortex and CA3, and two ‘‘gates,’’ or inhibitory filters that reduce the efficacy of both inputs, keeping granule cells relatively quiescent. When GABAergic inhibition is reduced experimentally, or under pathological conditions, CA3 pyramidal cells activate granule cells reliably, and do so primarily by disynaptic excitation that is mediated by mossy cells. We suggest that the backprojection has important functions normally that are dynamically regulated by nonprincipal cells of the dentate gyrus. Slightly reduced GABAergic input would lead to increased polysynaptic associative processing between CA3 and the dentate gyrus. Under pathological conditions associated with loss of GABAergic interneurons, the backprojection may support reverberatory excitatory Corresponding author. Tel.: +1 845 398 5427; Fax: +1 845 398 5422; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63034-9

627

628

activity between CA3, mossy cells, and granule cells, possibly enhanced by mossy fiber sprouting. In this case, the backprojection could be important to seizure activity originating in hippocampus, and help explain the seizure susceptibility of ventral hippocampus. Keywords: pyramidal cell; dentate gyrus; hilus; mossy cell; interneuron; granule cell; recurrent excitation Introduction Most explanations of hippocampal circuitry begin with the identification of the major subfields and the trisynaptic circuit. As shown in Fig. 1, the hippocampus of mammals is composed of three subfields: area CA1, containing primarily CA1 pyramidal cells, area CA3, also composed primarily of pyramidal cells, and the dentate gyrus, where the primary principal cell is the granule cell. The trisynaptic circuit is composed of three sequential glutamatergic synapses: 1) perforant path axons of layer II neurons in entorhinal cortex project to the outer twothirds of the dentate gyrus molecular layer, the location of the distal granule cell dendrites, 2) mossy fiber axons of granule cells project to proximal dendrites of area CA3 pyramidal cells, and 3) the Schaffer collateral axons of CA3 pyramidal cells project to stratum radiatum of CA3, where the apical dendrites of area CA1 pyramidal cells are located (Amaral and Witter, 1989). This projection is clearly fundamental, and its importance to information processing in

hippocampus is not to be doubted here. The point of emphasis here is, instead, that the trisynaptic pathway is not the only circuit to consider. One reason for hesitation is that a considerable number of studies have now shown that nonprincipal cells of hippocampus play critical roles in modulating the trisynaptic pathway. They are innervated by each of the fiber pathways mentioned above, and have diverse projections to all hippocampal principal cells. For example, within the dentate gyrus, the perforant path and the mossy fibers innervate numerous types of GABAergic neurons, as well as glutamatergic mossy cells of the hilus. The GABAergic ‘‘interneurons’’ and mossy cells project to granule cells at the level of the soma, axon hillock, and at every part of the dendritic tree. Indeed, it has been proposed that the primary targets of mossy fibers are GABAergic neurons, and the primary effect of mossy fiber transmission under normal conditions is inhibitory (Acsady et al., 1998). This appears to be the case in recordings of CA3 neurons in slices, where they dominant response to mossy fiber stimulation is an inhibitory postsynaptic potential (IPSP) (Scharfman, 1993a, b), despite the large unitary excitatory postsynaptic potential (EPSP) produced by granule cells (Scharfman et al., 1990) (for review, see Jaffe and

Fig. 1. The ‘‘trisynapto-centric’’ vs. ‘‘CA3-centric’’ view of hippocampal information processing. (A) The trisynaptic circuit is diagrammed schematically for a horizontal section through the adult rodent hippocampus. Cell layers are in gray. (B) The axonal arbor of CA3 pyramidal cells is schematically presented to illustrate the perspective that these neurons may be a central point of information processing in the hippocampus.

629

Gutierrez, this volume). It has been suggested that other characteristics of mossy fiber transmission may be even more important than their ability to depolarize postsynaptic pyramidal cells (Urban et al., 2001). In this chapter an alternative to the trisynaptocentric view is emphasized. It involves the axons of CA3 pyramidal cells that project in the opposite direction of the trisynaptic circuit (‘‘back projection’’), into the dentate gyrus (Fig. 1B).

Anatomical evidence for a ‘‘backprojection’’ What is the evidence for a ‘‘backprojection?’’ First we will consider anatomical arguments and then physiological data. Historically, there have been only rare suggestions that CA3 pyramidal cells might project into the dentate gyrus (Zimmer, 1971; Schwerdtfeger and Sarvey, 1983). These did not seem to attract much attention, possibly because some of the experimental approaches had technical limitations. For example, the most common approach for tract tracing at the time relied on tracer injection, but CA3 is difficult to label selectively with this technique. Newer approaches provided the requisite selectivity. These techniques involved dye injection, typically biocytin, into individual CA3 pyramidal cells using intracellular microelectrodes, followed by axon reconstruction to evaluate the axon arbor. Using this approach, Ishizuka et al. (1990) filled individual CA3 neurons in hippocampal slices of the rat, and emphasized several aspects of the axon of CA3 neurons. They found evidence for axon collateralization within the hilus in many of their sampled cells, particularly those with a cell body in ‘‘proximal’’ CA3 (near the dentate gyrus, i.e., CA3b and CA3c). Li et al. (1994) injected biocytin into CA3 pyramidal cells in vivo, also using adult rats, and found even greater evidence for a backprojection. They found that individual CA3 neurons had collaterals in the hilus whether the cell body of origin was in CA3a, b, or c. Like Ishizuka et al. (1990), they found the most hilar collateralization from cells within CA3c, and they specifically identified the ventral portion of CA3c region as the area with the greatest collateralization in the hilus.

Ventral CA3c pyramidal cells collateralized approximately 3–4  as much in the hilus as dorsal CA3c pyramidal cells, or CA3b cells. Li et al. (1994) also noted that CA3 pyramidal cell axons entered the granule cell layer and collateralized in the inner molecular layer, and this was predominantly a characteristic of ventral CA3c pyramidal cells. These axons exhibited numerous varicosities, suggesting that they innervated dentate gyrus granule cells, and possibly other cell types. Therefore, the CA3c population of ventral hippocampus appears to be a subset of CA3 pyramidal cells with the highest potential for dentate gyrus interactions. These studies were the first to define hilar collateralization of CA3 neurons unequivocally. However, one caveat was that the sample sizes were small, and it remains unclear exactly how representative the data were of the entire CA3 population. It also is unclear whether the data from adult rat might vary across age and species. Due to the labor-intensive nature of the experiments, these questions remain unanswered. Comparisons of CA3a, b, and c pyramidal cells are interesting to review in light of the findings that their axons appear to have a different preference for the dentate gyrus. Anatomically, CA3c pyramidal cells are heterogeneous neurons (Scharfman, 1993; Turner et al., 1995), sometimes reflecting stereotypical pyramidal cell morphology, but in other cases they may appear more similar to nonpyramidal cells. Some of the electrophysiological characteristics of CA3c pyramidal cells appear to be distinct from CA3a and b, and this may be important because it could lead to unique physiological attributes to the backprojection. CA3c neurons are not as easily activated by mossy fiber stimulation as CA3b neurons in slices, suggesting that the GABAergic neurons which are targets of mossy fibers innervate CA3c more than CA3b. All else being equal, such preferential inhibition might quiet the backprojection normally, relative to other CA3 projection systems. In comparisons of CA3b and CA3c pyramidal cells of the adult male rat, almost all CA3c neurons demonstrated burst (phasic) firing in response to current injection, whereas CA3b pyramidal cells were much more heterogeneous, often exhibiting

630

trains of action potential (tonic firing) only. Variation in CA3a and b in this respect has been reported (Bilkey and Schwartzkroin, 1990). If CA3c neurons do have a greater percentage of neurons which intrinsically discharge in bursts, the information processed by backprojecting axons may differ in the way it influences granule cells from the way information is transferred to other targets of CA3 neurons. CA3c neurons demonstrated the longest time constant, which is noteworthy because this characteristic might lead to a greater capacity for integration of synaptic inputs from distinct locations along the dendrites (Scharfman, 1993b). However, CA3c neurons appear to have a relatively short electrotonic length, which would lead to rapid decay before integration of dendritic inputs would reach the soma. Together the data would suggest a specialization of the dendrite as a separable processing unit, perhaps more than other CA3 neurons. It also would suggest that somatic inputs would have a relatively greater influence on action potential output than dendritic inputs. CA3c is more vulnerable to certain types of excitotoxic stimulation than CA3a or b pyramidal cells (Chang and Dyer, 1985; Freund et al., 1992; Miettinen et al., 1998; Haas et al., 2001). It was proposed that CA3 may be an area where burst generation occurs in models of epilepsy, but others have suggested CA3a/b is that location (Knowles et al., 1987; Colom and Saggau, 1994). Selective lesions to CA3a vs. b vs. c are difficult to make without injuring adjacent areas, so the functional relevance of the differences between CA3a, b, and c remain unclear. In summary, the anatomical data suggest that ventral hippocampus and CA3c neurons make the most robust backprojection in normal adult rats, and there is collateralization in the hilus, granule cell layer and inner molecular layer. Specific physiological characteristics of CA3c neurons may impart unique functional consequences to the backprojection.

Physiological evidence for a ‘‘backprojection’’ Some of the first physiological evidence that CA3 axons innervated the dentate gyrus came from

studies in hippocampal slices, which showed that a stimulating electrode placed in the hilus could activate an antidromic population spike, recorded extracellularly in the CA3 pyramidal cell layer. Subsequent studies identified what cell types in the dentate gyrus could be influenced by the CA3 projection. Several approaches established that hilar neurons could be activated readily, but granule cells could not. First, fimbria stimulation was used to activate CA3 pyramidal cells instead of the hilus, because antidromic action potentials are easily elicited in CA3 neurons by stimulation at this site, with no evidence for direct activation of hilar neurons and granule cells. Direct activation would be expected because hilar cells that project commissurally have an axon that descends into stratum oriens of CA3b, presumably on its way to the fimbria, but in our slices, we have found that it is severed before reaching the fimbria. Direct activation of granule cells would only be possible if current spread entered stratum lucidum, which recordings show does not occur after fimbria stimulation using the methods we have employed. Extracellular recordings showed that such fimbria stimulation led to antidromic and orthodromic activation of CA3 b and c pyramidal cells, followed by an extracellular negative wave in the granule cell layer and inner molecular layer that reversed polarity in the middle molecular layer (Fig. 2). Due to the negative resting potential of granule cells, the negative potential could reflect an EPSP or depolarized IPSP, and it was shown subsequently by intracellular recording that it was a depolarized GABAA receptor-mediated IPSP (Fig. 3; Scharfman, 1994a, b, c). This fimbria stimulation site activated hilar mossy cells and hilar GABAergic neurons disynaptically, and was sensitive to the AMPA receptor antagonist CNQX but not the muscarinic antagonist atropine (Scharfman, 1993a). At a longer latency, the granule cell IPSP began. These data suggesting that fimbria activation of CA3 pyramidal cells led to hilar neuron activation, presumably by hilar collaterals of CA3 pyramidal cells. Comparing the latency to the response of CA3 pyramidal cells and hilar cells supported this hypothesis, because CA3 pyramidal cells were

631

Fig. 2. Recordings in hippocampal slices show evidence of trisynaptic and backprojecting pathways. (A) Extracellular recordings of evoked responses to five stimulation sites (1–5) and four recording sites (A–D). Recordings were made sequentially in the same slice in response to a fixed stimulus for each stimulus site. (B) A schematic is used to allow better comparisons of fimbria-evoked and outer molecular layer-evoked responses recorded at four sites in the same slices: the outer molecular layer, granule cell layer, CA3c cell layer, and CA3b cell layer. X marks recording site locations. (C) Comparison of evoked responses to fimbria stimulation in the same slice demonstrate the responses in the granule cell layer following pyramidal cell activation by the fimbria, but do so with a delay. Superimposition of responses illustrates that the onset of the granule cell layer field potential begins approximately 1–2 ms after the antidromic population spike recorded in area CA3b. This delay suggests an intermediary synapse, probably in CA3c or the hilus, between CA3b and granule cells. Indeed, the granule cell field potential begins immediately after the CA3c orthodromic population spike, which probably is due to CA3b recurrent excitation of CA3c pyramidal cells. At the same time, hilar cells are also activated and innervate granule cells (see later figures), although CA3c pyramidal cells could innervate granule cells also (Li et al., 1994). The bottom trace is an IPSP recorded intracellularly from a granule cell in the same slice in response to the same stimulus. It shows that the onset of the IPSP is similar to the onset of the granule cell layer field potential, which likely reflects the average of many IPSPs in granule cells situated around the extracellular electrode.

activated at short latency and hilar cells approximately 2 ms later. The same stimulus evoked a long latency IPSP in granule cells, consistent with a CA3-hilar neuron-granule cell pathway. In the presence of bicuculline, a GABAA receptor antagonist, fimbria stimulation evoked EPSPs in granule cells, and did so with a similar latency as the IPSPs, suggesting a CA3-mossy cell-granule cell pathway is normally masked by the activation of GABAergic interneurons by CA3 hilar collaterals (Fig. 3). Additional studies revealed that the

GABAergic neurons may not only involve those in the hilus, but also the ‘‘basket cells’’ of the granule cell layer (Scharfman, 1994a, c; Kneisler and Dingledine, 1995). Use of GABAA receptor antagonists also allowed another approach to examine the effects of CA3 on the dentate gyrus, one that involved no stimulation to activate CA3 neurons. In the presence of GABAergic receptor antagonists such as penicillin, picrotoxin, or bicuculline, CA3 neurons discharge rhythmically in bursts of action

632

Fig. 3. Physiological evidence for a CA3 backprojection mediated by hilar neurons. (A) A schematic illustrates the backprojection supported by physiological evidence collected to date. It shows that CA3 pyramidal cells innervate GABAergic and glutamatergic mossy cells of the hilus, which in turn innervate granule cells. (B) Recordings illustrate the physiological correlates of the schematic in A. Intracellular recordings from granule cells in response to fimbria stimulation in an adult male rat slice illustrate an evoked IPSP that reverses at 70 mV, indicated a GABAA receptor-mediated IPSP is primarily evoked under normal conditions. After the GABAA receptor antagonist bicuculline was bath-applied, the evoked response was an EPSP followed by an IPSP that reversed at approximately 80 mV, suggesting an EPSP followed by a GABAB receptor-mediated IPSP is normally masked by GABAA receptormediated inhibition. From Scharfman (1994a). Reprinted with permission from the Society for Neuroscience.

potentials (Schwartzkroin and Prince, 1977; Knowles et al., 1987; Scharfman, 1994a, c). Therefore, one can examine simultaneously any other cell type in the slice to determine when and if they are concurrently influenced. Simultaneous recordings showed that immediately after the onset of spontaneous bursts in CA3 pyramidal cells, hilar mossy cells and GABAergic interneurons also demonstrated bursts (Scharfman, 1994a, c). Granule cells depolarized and could reach threshold, but did so at a greater delay from the CA3 burst than hilar cells. The hilar activity and granule cell excitation, but not CA3 bursts, were blocked by transecting the border between the dentate gyrus and CA3b, or by application of CNQX (Fig. 4; Scharfman, 1994a). When the timing of the burst discharges was examined more closely, the onset of the depolarization of CA3 neurons immediately preceded the onset in the mossy cells and GABAergic neurons, and the onset of hilar neuron bursts immediately preceded granule cell depolarization, suggesting single synaptic delays of a CA3hilus-granule cell pathway (Scharfman, 1994a, c).

The most parsimonious explanation was that CA3 pyramidal cells innervated both types of hilar neurons, and evidence was obtained for this by simultaneous intracellular recordings from monosynaptically connected pyramidal cells and mossy cells (Scharfman, 1994b). The same approach showed that mossy cells innervated granule cells and interneurons in the hilus (Scharfman, 1995), suggesting that multiple polysynaptic circuits could be set in motion by CA3 pyramidal cells, and ultimately influence dentate granule cells. Current source density has also provided evidence of CA3–dentate gyrus interactions similar to those discussed above (Wu et al., 1998). In summary, there is strong evidence from studies in hippocampal slices that the CA3 backprojection innervates hilar neurons, both mossy cells and GABAergic neurons. Currently there is no physiological evidence that CA3 directly innervates the granule cells of the dentate gyrus, but absence of evidence is not a proof of absence. At the present time, the evidence suggests that the granule cells, even if they receive a direct

633

Fig. 4. The CA3 excitatory backprojection is dependent on hilar mossy cells. (A) Schematic illustrating the experimental approach for results shown in B–C. A site in CA3c was used for extracellular recording while recording intracellularly from a granule cell. Pressure application of CNQX was used at two distinct sites in the hilus or in the inner molecular layer. CNQX was applied in microdrops that were barely detectable by eye, allowing preferential application to select areas of the slice. (B) In the presence of bicuculline, CA3 epileptiform discharges were evoked by fimbria stimulation, and following the onset of the burst discharge, a granule cell depolarized and discharged two action potentials. After CNQX was pressure-applied to the hilus, the granule cell response decreased but the CA3 discharge remained unaffected. (C) In a different slice, CNQX pressure-application to the inner molecular layer, near the recorded granule cell, reversibly decreased the EPSP of the granule cell, suggesting an inner molecular layer glutamatergic synapse was necessary for the EPSP. From Scharfman (1994a). Reprinted with permission from the Society for Neuroscience.

projection from CA3 neurons, are primarily inhibited by the backprojection, not activated. However, once they are relieved of GABAA receptormediated inhibition, a strong excitatory circuit is unmasked. The excitatory circuit appears to be primarily derived from the mossy cell projection to the inner molecular layer. Functional implications of the CA3 backprojection Hippocampal information processing Is CA3 the point of entry to hippocampus? Given the projection of CA3 pyramidal cells innervates CA1, the dentate gyrus, other CA3 neurons by recurrent collaterals, and sends information both ipsilaterally and contralaterally, the argument can be made that it is a cell type that may be central to hippocampal information processing (Fig. 1B). In light of this, it is interesting that the general assumption is ‘‘trisynapto-centric,’’ that granule cells receive the primary input, and the mossy fiber control how the hippocampus processes the incoming input. This view might be correct in cases that

involve layer II of entorhinal cortex specifically, but it may be rare that layer II is ever activated in isolation, given its broad input from other layers within the entorhinal cortex, and the diverse inputs to entorhinal cortex. In actuality, it is highly likely that the sensory and other cortical input that comes to entorhinal cortex would also involve subcortical structures that at the same time might reach CA3 through the fimbria. Furthermore, the temporoammonic pathway may activate CA3 neurons at lower threshold than granule cells (see Chapter by Derrick, this volume), and this further supports the tenet that incoming information might first activate the CA3 region of hippocampus, rather than do so via dentate granule cells. Given the data that mossy fibers primarily inhibit CA3 pyramidal cells, the dentate may actually function to keep CA3 from being activated excessively by the concomitant inputs from the cortex and subcortical zones. The dentate gyrus has two ‘‘gates’’ rather than one Another implication of the discussion above, particularly the physiological studies’ is that the dentate gyrus actually has two ‘‘gates’’ (Fig. 5). The

634

Fig. 5. The trisynaptic circuit and CA3 backprojection supports a bi-directional gate to the dentate gyrus granule cells.

first is a gate that prevents activation by the entorhinal cortex (see Chapter by Hsu, this volume). This may not be an all-or-none function, so the word ‘‘filter’’ may be better (see Chapter by Dudek and Sutula, this volume). However, the concept is appropriate regardless of the semantics: cortical input is not very effective in activating granule cells above their threshold. It appears that a second ‘‘gate’’ or ‘‘filter’’ also exists, from CA3 pyramidal cells back to the dentate gyrus. Thus, the backprojection does not appear to be effective in depolarizing granule cells, and it appears that the reason is the divergence of CA3 hilar collaterals to both excitatory hilar mossy cells, and inhibitory GABAergic interneurons. Evidently, the inhibitory neurons exert a stronger net influence, and it is important to note that in vivo it may be different, because mossy cell axons are truncated in the slice much more than GABAergic neurons. In vivo, one would predict that CA3 would inhibit granule cells within the same lamella, but activate them in distal sections of hippocampus because of the long translamellar mossy cell axon projection. When inhibition is reduced experimentally, CA3 discharge is extremely robust in depolarizing granule cells, and it appears to be primarily due to inner molecular synapses of mossy cells. Thus, the hilar neurons modulate the CA3 backprojection under normal conditions, and any condition that would reduce GABAergic inhibition would be expected to facilitate CA3 activation of dentate granule cells.

If this occurs, it would be predicted that associative processes would be enhanced. In other words, not only could there be recurrent excitatory circuits among CA3 pyramidal cells that perform autoassociative functions, but polysynaptic autoassociation would also potentially develop between CA3 and the granule cells when GABAergic inhibition is weak. One argument that makes this even more compelling is that granule cells do not necessarily need to reach threshold in order for mossy fiber transmission to occur (Alle and Geiger, 2006). Subthreshold depolarizations in granule cells can lead to effects in granule cell targets. An important implication is that autoassociative networks, particularly the complex CA3–dentate network discussed here, is likely to be controlled by GABAergic hilar neurons and mossy cells.

Pathological conditions The CA3 backprojection may subserve an important role in conditions like ischemia or epilepsy, when hilar GABAergic neurons are injured or killed (Johansen et al., 1987; Sloviter, 1987; see Chapter by Tallent, this volume). In both conditions, the somatostatin-immunoreactive hilar neurons that project preferentially to the outer molecular layer are lost. In animal models of epilepsy that use status epilepticus as the initial experimental manipulation to induce an epileptic state, there are many types of GABAergic neurons that may be affected, but most laboratories suggest a relative preservation of parvalbumin-immunoreactive basket cells. The animal models of epilepsy are also interesting because both the granule cell axons and the pyramidal cell axons may sprout new collaterals in response to injury adjacent to them. Granule cell axons sprout into the inner molecular layer and hilus, as well as stratum oriens in CA3. It would be of interest to determine whether the backprojection increases in response to injury. Regardless of these changes, it does appear that the backprojection is robust under epileptic conditions. The data are from hippocampal slices from animals that had status epilepticus and became epileptic, i.e., they developed recurrent spontaneous seizures for the rest of their lifespan. In most of

635

Fig. 6. Evidence for associative networks and reverberatory circuits in the dentate gyrus under control of GABAergic inhibition. (A) A schematic illustrates electrode locations for the recordings shown in B. (B) Stimulation of specific sites in the slice evoked bursts of action potentials in CA3 pyramidal cells, and simultaneous intracellular recordings from a granule cell show bi-directional activation of CA3 and the granule cell. (C) In a different slice from B where CA3 epileptiform bursts were followed by numerous after discharges, CA3 appeared to precede activation of the simultaneously-recorded granule cell, but during the after discharges, the converse appeared to develop. From Scharfman (1994a). Reprinted with permission from the Society for Neuroscience.

these animals, CA3 pyramidal cell burst discharges occur when slices are prepared, even when bicuculline is absent (Scharfman et al., 2001). This affords an opportunity to examine the backprojection by simply examining concurrent activity in hilar neurons and granule cells. When this has been done, it is clear that the backprojection to mossy cells still exists, because mossy cells exhibit synchronized burst discharges with CA3 pyramidal cells in slices from rats with recurrent seizures (Scharfman et al., 2001). Interneurons in the hilus do as well, although the population of both mossy cells and interneurons is reduced relative to the normal rat, and it is not yet clear which GABAergic neurons are involved of the many types that exist. Granule cells usually appear to be relatively quiet when CA3 burst discharges occur (Scharfman and McCloskey, 2007) but this is not always the case: some granule cells have depolarizations during the CA3 burst discharges. In this case, bicuculline may not be required because of a loss of GABAergic inhibition when hilar inhibitory neurons are killed after status epilepticus. This pathway may be relevant to the seizures that occur in the whole animal, because the CA3 discharges appear to generate reverberatory activity between the dentate and CA3 pyramidal cells

(Fig. 6). During burst discharges in CA3, we have found that the onset of the burst discharge occurs in pyramidal cells first, and then in granule cells, but during after discharges that follow the primary burst, granule cell discharge may precede the pyramidal cells. These data suggest CA3 may be a generator of activity initially, but granule cells can initiate further excitation. In the whole animal, such reverberatory excitation could lead to substantial network activity of CA3/dentate neurons, and possibly trigger seizures after a certain threshold of network activity is reached. One would predict that this might occur in more ventral hippocampus than dorsal, given that the backprojection is more prevalent in ventral hippocampus. This prediction is supported by findings in slices of temporal hippocampus, where seizure activity is most robust (Bragdon et al., 1986; Borck and Jefferys, 1999; Scharfman et al., 2002).

Conclusions The CA3 region is often considered to be the second subfield that receives information from extrinsic afferents to hippocampus, while the dentate gyrus is the initial entry point. Here we suggest

636

that CA3 may be a major gateway to hippocampus, particularly with respect to subcortical input. One reason to think of CA3 as a central access point is that it has a projection that reaches the vast majority of hippocampal neurons, both ipsilaterally and contralaterally, including both CA1 and the dentate gyrus. The projection to the dentate gyrus, a ‘backprojection,’ has not been studied in as much detail as the Schaffer collateral system, leading to more appreciation of CA1–CA3 interactions when CA3–dentate networks may actually be just as important to hippocampal function. Physiological data suggest that the CA3 backprojection engages intermediary hilar neurons to exert its primary effect on granule cells, although direct projections to granule cells in a subset of CA3 neurons in ventral hippocampus have been shown by anatomical methods. GABAergic inhibition appears to regulate the ability of CA3 pyramidal cells to exert a primarily inhibitory or excitatory effect on granule cells, and to control the possible reverberatory activity between the two subfields. CA3–dentate interactions may subserve an important associative function in the behaving animal, influenced by factors that depress GABAergic neuron activity or GABAergic transmission; this may also apply under pathological conditions when GABAergic neurons are damaged or lost, such as temporal lobe epilepsy.

Acknowledgments NIH (NINDS) and the New York State Department of Health.

References Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Alle, H. and Geiger, J.R. (2006) Combined analog and action potential coding in hippocampal mossy fibers. Science, 311: 1290–1293. Amaral, D.G. and Witter, M.P. (1989) The three-dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience, 31: 571–591.

Bilkey, D.K. and Schwartzkroin, P.A. (1990) Variation in electrophysiology and morphology of hippocampal CA3 pyramidal cells. Brain Res., 514: 77–83. Borck, C. and Jefferys, J.G. (1999) Seizure-like events in disinhibited ventral slices of adult rat hippocampus. J. Neurophysiol., 82: 2130–2142. Bragdon, A.C., Taylor, D.M. and Wilson, W.A. (1986) Potassium-induced epileptiform activity in area CA3 varies markedly along the septotemporal axis of the rat hippocampus. Brain Res., 378: 169–173. Chang, L.W. and Dyer, R.S. (1985) Septotemporal gradients of trimethyltin-induced hippocampal lesions. Neurobehav. Toxicol. Teratol., 7: 43–49. Colom, L.V. and Saggau, P. (1994) Spontaneous interictal-like activity originates in multiple areas of the CA2-CA3 region of hippocampal slices. J. Neurophysiol., 71: 1574–1585. Freund, T.F., Ylinen, A., Miettinen, R., Pitkanen, A., Lahtinen, H., Baimbridge, K.G. and Riekkinen, P.J. (1992) Pattern of neuronal death in the rat hippocampus after status epilepticus. Relationship to calcium binding protein content and ischemic vulnerability. Brain Res. Bull., 28: 27–38. Haas, K.Z., Sperber, E.F., Opanashuk, L.A., Stanton, P.K. and Moshe, S.L. (2001) Resistance of immature hippocampus to morphologic and physiologic alterations following status epilepticus or kindling. Hippocampus, 11: 615–625. Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization of intrahippocampal projections originating from CA3 pyramidal cells in the rat. J. Comp. Neurol., 295: 580–623. Johansen, F.F., Zimmer, J. and Diemer, N.H. (1987) Early loss of somatostatin neurons in dentate hilus after cerebral ischemia in the rat precedes CA-1 pyramidal cell loss. Acta Neuropathol. (Berl.), 73: 110–114. Kneisler, T.B. and Dingledine, R. (1995) Synaptic input from CA3 pyramidal cells to dentate basket cells in rat hippocampus. J. Physiol., 487: 125–146. Knowles, W.D., Traub, R.D. and Strowbridge, B.W. (1987) The initiation and spread of epileptiform bursts in the in vitro hippocampal slice. Neuroscience, 21: 441–455. Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The hippocampal CA3 network: an in vivo intracellular labeling study. J. Comp. Neurol., 339: 181–208. Miettinen, R., Kotti, T., Tuunanen, J., Toppinen, A., Riekkinen Sr., P. and Halonen, T. (1998) Hippocampal damage after injection of kainic acid into the rat entorhinal cortex. Brain Res., 813: 9–17. Scharfman, H.E. (1993a) Activation of dentate hilar neurons by stimulation of the fimbria in rat hippocampal slices. Neurosci Lett., 156: 61–66. Scharfman, H.E. (1993b) Spiny neurons of area CA3c in rat hippocampal slices have similar electrophysiological characteristics and synaptic responses despite morphological variation. Hippocampus, 3: 9–28. Scharfman, H.E. (1994a) EPSPs of dentate gyrus granule cells during epileptiform bursts of dentate hilar ‘‘mossy’’ cells and area CA3 pyramidal cells in disinhibited rat hippocampal slices. J. Neurosci., 14: 6041–6057.

637 Scharfman, H.E. (1994b) Evidence from simultaneous intracellular recordings in rat hippocampal slices that area CA3 pyramidal cells innervate dentate hilar mossy cells. J. Neurophysiol., 72: 2167–2180. Scharfman, H.E. (1994c) Synchronization of area CA3 hippocampal pyramidal cells and non-granule cells of the dentate gyrus in bicuculline-treated rat hippocampal slices. Neuroscience, 59: 245–257. Scharfman, H.E. (1995) Electrophysiological evidence that dentate hilar mossy cells are excitatory and innervate both granule cells and interneurons. J. Neurophysiol., 74: 179–194. Scharfman, H.E., Kunkel, D.D. and Schwartzkroin, P.A. (1990) Synaptic connections of dentate granule cells and hilar neurons: results of paired intracellular recordings and intracellular horseradish peroxidase injections. Neuroscience, 37: 693–707. Scharfman, H.E., Goodman, J.H. and McCloskey, D.P. (2007) Ectopic granule cells of the rat dentate gyrus. Dev. Neurosci., 29: 14–27. Scharfman, H.E., Smith, K.L., Goodman, J.H. and Sollas, A.L. (2001) Survival of dentate hilar mossy cells after pilocarpineinduced seizures and their synchronized burst discharges with area CA3 pyramidal cells. Neuroscience, 104: 741–759. Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and Goodman, J.H. (2002) Structural and functional asymmetry

in the normal and epileptic rat dentate gyrus. J. Comp. Neurol., 454: 424–439. Schwartzkroin, P.A. and Prince, D.A. (1977) Penicillin-induced epileptiform activity in the hippocampal in vitro preparation. Ann. Neurol., 1: 463–469. Schwerdtfeger, W.K. and Sarvey, J.M. (1983) Connectivity of the hilar region of the hippocampal formation in the rat. J. Hirnforschung, 24: 201–207. Sloviter, R.S. (1987) Decreased hippocampal inhibition and a selective loss of interneurons in experimental epilepsy. Science, 235: 73–76. Turner, D.A., Li, X.G., Pyapali, G.K., Ylinen, A. and Buzsaki, G. (1995) Morphometric and electrical properties of reconstructed hippocampal CA3 neurons recorded in vivo. J. Comp. Neurol., 356: 580–594. Urban, N.N., Henze, D.A. and Barrionuevo, G. (2001) Revisiting the role of the hippocampal mossy fiber synapse. Hippocampus, 11: 408–417. Wu, K., Canning, K.J. and Leung, L.S. (1998) Functional interconnections between CA3 and the dentate gyrus revealed by current source density analysis. Hippocampus, 8: 217–230. Zimmer, J. (1971) Ipsilateral afferents to the commissural zone of the fascia dentata, demonstrated in decommissurated rats by silver impregnation. J. Comp. Neurol., 142: 393–416.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 35

Modeling the dentate gyrus Robert J. Morgan1,, Vijayalakshmi Santhakumar2 and Ivan Soltesz1 1

Department of Anatomy and Neurobiology, 193 Irvine Hall, University of California, Irvine, CA 92697, USA Department of Neurology, David Geffen School of Medicine, University of California, Los Angeles, CA, USA

2

Abstract: Computational modeling has become an increasingly useful tool for studying complex neuronal circuits such as the dentate gyrus. In order to effectively apply computational techniques and theories to answer pressing biological questions, however, it is necessary to develop detailed, data-driven models. Development of such models is a complicated process, akin to putting together a jigsaw puzzle with the pieces being such things as cell types, cell numbers, and specific connectivity. This chapter provides a walkthrough for the development of a very large-scale, biophysically realistic model of the dentate gyrus. Subsequently, it demonstrates the utility of a modeling approach in asking and answering questions about both healthy and pathological states involving the modeled brain region. Finally, this chapter discusses some predictions that come directly from the model that can be tested in future experimental approaches. Keywords: computational model; epilepsy; granule cell; interneuron; mossy cell; sclerosis (Bartos et al., 2001; Vida et al., 2006), homeostatic plasticity (Howard et al., 2007), and topological alterations during epileptogenesis (Dyhrfjeld-Johnsen et al., 2007). These studies demonstrate the importance and utility of computational modeling in contributing to our understanding of the dentate gyrus both in healthy animals and in pathological states such as epilepsy. In this chapter, we will use the analogy of a jigsaw puzzle to walk through the process of modeling the dentate gyrus neural network. In the first and second parts, ‘‘Finding the right pieces’’ and ‘‘Putting the pieces together,’’ we will discuss what types of considerations are necessary when building a large-scale, data-driven model of a normal, healthy rat dentate gyrus (all data presented in this chapter comes from the rat dentate). Then, in ‘‘The big picture’’ we will examine how such a model is useful for studying the effects of structural and functional network alterations that occur during epileptogenesis. Finally, we will conclude

The dentate gyrus in the mammalian brain is a complex neuronal circuit that is thought to serve as a gate for activity propagation throughout the limbic system (Heinemann et al., 1992; Lothman et al., 1992). Scientific understanding of such complex systems has traditionally been obtained through use of animal models, culture systems, and numerous other experimental approaches. In the past several decades, however, computational modeling has become an increasingly useful tool for studying and understanding neuronal network structure and function. In fact, the dentate gyrus has been the subject of a number of recent computational models looking at such issues as mossy fiber sprouting (Lytton et al., 1998; Santhakumar et al., 2005; Dyhrfield-Johnsen et al., 2007) and hilar cell loss (Santhakumar et al., 2005; Dyhrfjeld-Johnsen et al., 2007), interneurons and shunting inhibition Corresponding author. Tel.: +1 (949) 824-3967; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63035-0

639

640

with ‘‘Tackling future puzzles and problems’’ by discussing future directions for dentate modeling and some predictions that can be tested experimentally.

Finding the right pieces Assembling a computational model is much like putting together a puzzle. The pieces include cell types, cell numbers, synapses, and other cellular interactions such as gap junctions, and distributions of axons and dendrites. These are assembled in a specific fashion upon a backdrop that includes the software necessary for implementing the model and the hardware, the physical computer and memory used for the model calculations. Just as the number and size of pieces in a puzzle varies, the number of components and the details of those components vary dramatically in computational models. The precise amount of detail necessary in a model depends directly on the question the model is designed to answer. Examining the effects of the loss of specific subgroups of cell types, for example, requires the model to capture the relative changes in cell densities, necessitating the development of largescale models. The first two sections of this chapter will provide an overview of the general approach for developing such a large-scale model to study both the healthy rat dentate gyrus and the effects of epileptogenesis with emphasis on mossy fiber sprouting and hilar cell loss. With over a million cells and a billion synapses amongst at least eight distinct cell types in the dentate gyrus, the first step in developing a model is to determine the anatomy of the dentate network. This can be accomplished by creating a connection matrix of the dentate gyrus (Patton and Mcnaughton, 1995), which can be thought of as the border or backbone of our puzzle. The matrix itself contains a number of distinct pieces that are outlined here and described in detail below. The first piece is the cell type. Eight types of dentate cells can be identified as anatomically well-described: granule cells (GCs), mossy cells (MCs), basket cells (BCs), axo-axonic cells, molecular layer cells with axonal projections to the perforant path (MOPP cells), hilar cells with axonal projections to the perforant path (HIPP

cells), hilar cells with axonal projections to the commissural-associational pathway (HICAP cells), and interneuron-specific (IS) cells (Fig. 1) (Santhakumar and Soltesz, 2004). The second piece is the number of cells of each type, data that can be estimated from published data (Table 1). The third piece is the specific connectivity between cell types. This quantifies precisely how many postsynaptic cells among each of the eight cell types a single presynaptic neuron of a given type innervates (e.g., from the third row, second column in Table 1; a single basket cell innervates approximately 1,250 GCs; mean and ranges are indicated, with references). The final piece required for constructing the connection matrix of the dentate is the spatial constraint in cell-to-cell connectivity. For each cell type, the extent of the axons of single cells along the septo-temporal axis of the dentate gyrus can be determined from in vivo single cell fills published in the literature (Fig. 2). The first piece: cell types Determination of cell type is based on several factors including: (1) the nature of the neurotransmitter released; (2) location in the laminar structure of the dentate gyrus; (3) morphological features such as shape of the soma, dendritic arbor, and axonal distribution pattern; and (4) presence of specific markers such as calcium-binding proteins and neuropeptides (for detailed description of cell type discrimination see Freund and Buzsaki, 1996). According to these criteria, two excitatory cell types (GCs and MCs) and at least six distinct inhibitory interneurons can be identified in the dentate gyrus (Fig. 1). The second piece: cell numbers A practical approach for determining cell numbers for various cell types involves first identifying the cell layers and the total number of neurons in each layer. Then, it is possible to estimate subtype numbers by estimating the percentage of neurons that express specific markers (Buckmaster et al., 1996, 2002a; Nomura et al., 1997a, b). The dentate gyrus is composed of three cell layers: the molecular layer,

641

Fig. 1. Schematic of the basic circuitry of the dentate gyrus. Relational representation of the healthy dentate gyrus illustrating the network connections between the eight major cell types (GC, granule cell; BC, basket cell; MC, mossy cell; AAC, axo-axonic cells; MOPP, molecular layer interneurons with axons in perforant-path termination zone; HIPP, hilar interneurons with axons in perforantpath termination zone; HICAP, hilar interneurons with axons in the commissural/associational pathway termination zone; IS, interneuron selective cells). The schematic shows the characteristic location of the various cell types within the three layers of the dentate gyrus. Note, however, that this diagram does not indicate the topography of axonal connectivity or the somato-dendritic location of the synapses. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

granule cell layer, and the hilus (Fig. 1). The GCs are the primary projection cells of the dentate gyrus. These are small, densely packed cells, typically located in the granule cell layer of the dentate gyrus. The number of GCs in the rat dentate gyrus is approximately 1,000,000 (Gaarskjaer, 1978; Boss et al., 1985; West, 1990; Patton and Mcnaughton, 1995; Freund and Buzsaki, 1996). The other major excitatory cell type, the MC, is located in the hilus and does not express the GABA synthetic enzyme glutamic acid decarboxylase (GAD). Buckmaster and Jongen-Relo (1999) estimated the number of GAD-mRNA negative neurons in the dentate hilus

(presumed MCs) as approximately 30,000. The GABAergic interneurons are located in all three layers of the dentate gyrus. These include the BCs, parvalbumin-positive (PV) axo-axonic cells (whose axon terminals project exclusively to the axon initial segment of excitatory cells), somatostatin-positive HIPP cells (Freund and Buzsaki, 1996; Katona et al., 1999), nitric oxide synthase-positive HICAP cells (Freund and Buzsaki, 1996), and aspiny and calretinin-positive hilar IS cells (Gulyas et al., 1996). The number of GABAergic cell types along the granule cell-hilar border and in the hilus can be determined based on published histochemical data

642 Table 1. Connectivity matrix for the neuronal network of the control dentate gyrus Granule cells

Mossy cells

Basket cells

Axo-axonic ells MOPP cells

HIPP cells

HICAP cells

IS cells

Granule cells (1,000,000) Ref. [1–5] Mossy cells (30,000) Ref. [11] Basket cells (10,000) Ref. [16,17]

X X Ref. [6] 32,500 30,000–35,000 Ref. [4,11–13] 1250 1000–1500 Ref. [4,16–19]

X X Ref. [6] 5 5 Ref. [14] X X Ref. [18]

110 100–120 Ref. [4,10,11] 600 600 Ref. [12,13] 0.5 0–1 Ref. [18]

40 30–50 Ref. [4,7,10,11] 200 200 Ref. [12,13] X X Ref. [18]

20 10–30 Ref. [7] X X Ref. [15] X X Ref. [10,20]

3000

15 10–20 Ref. [6–9] 7.5 5–10 Ref. [13] 35 20–50 Ref. [16,17,20,21] X

3 1–5 Ref. [6,7,9] 7.5 5–10 Ref. [13] X X Ref. [18]

Axo-axonic cells (2000) Ref. [4,22]

9.5 7–12 Ref. [7] 350 200–500 Ref. [12,13] 75 50–100 Ref. [11,16,17,19] 150

X

X

X

X

X

2000–4000 Ref. [4,18,22]

X Ref. [5,18]

X Ref. [5,18]

X Ref. [5,18]

X Ref. [5,18]

X Ref. [5,18]

X Ref. [5,18,19]

MOPP cells (4000) Ref. [11,14] HIPP cells (12,000) Ref. [11]

7500 5000–10,000 Ref. [14] 1550 1500–1600 Ref. [4,11,20]

40 30–50 Ref. [14,25] 450 400–500 Ref. [4,11,20]

1.5 1–2 Ref. [14,26] 30 20–40 Ref. [20,25]

7.5 5–10 Ref. [14,25] 15 10–20 Ref. [25]

X X Ref. [14,20,25] X X Ref. [14,20,25]

7.5 5–10 Ref. [14,25] 15 10–20 Ref. [25]

X X Ref. [14,15] X X Ref. [15,20]

HICAP cells (3000) Ref. [5,29,30] IS cells (3000) Ref. [15,29,30]

700 700 Ref. [4,11,20] X X Ref. [15]

100–200 Ref. [4,5,11,14,23] X X Ref. [14,24] 35 20–50 Ref. [4,11,12,27,28] 35 30–40 Ref. [20] X X Ref. [15]

175 150–200 Ref. [4,11,20] 7.5 5–10 Ref. [15,19]

X X Ref. [20] X X Ref. [15]

15 10–20 Ref. [14,20] X X

50 50 Ref. [20] 7.5 5–10 Ref. [19]

50 50 Ref. [20] 7.5 5–10 Ref. [19]

X X 450 100–800 Ref. [15]

Notes: Cell numbers and connectivity values were estimated from published data for granule cells, mossy cells, basket cells, axo-axonic cells, molecular layer interneurons with axonal projections to the perforant path (MOPP), hilar interneurons with axonal projections to the perforant path (HIPP), hilar interneurons with axonal projections to the commissural/associational pathway (HICAP) and interneuron specific cells (IS). Connectivity is given as number of connections to a postsynaptic population (row 1) from a single presynaptic neuron (column 1). The average number of connections used in the 1:1 structural model is given in bold. Note, however, that the small world structure (see Section ‘‘Structural alterations in the dentate during epileptogenesis’’) was preserved even if only the extreme low or the extreme high estimates were used for the calculation of L and C. References given in table correspond to: 1Gaarskjaer (1978), 2Boss et al. (1985), 3West (1990), 4Patton and McNaughton (1995), 5Freund and Buzsaki (1996), 6Buckmaster and Dudek (1999), 7Acsady et al. (1998), 8Geiger et al. (1997), 9Blasco-Ibanez et al. (2000), 10Gulyas et al. (1992), 11Buckmaster and Jongen-Relo (1999), 12 Buckmaster et al. (1996), 13Wenzel et al. (1997), 14Han et al. (1993), 15Gulyas et al. (1996), 16Babb et al. (1988), 17Woodson et al. (1989), 18Halasy and Somogyi (1993), 19Acsady et al. (2000), 20Sik et al. (1997), 21Bartos et al. (2001), 22Li et al. (1992), 23Ribak et al. (1985), 24Frotscher et al. (1991), 25Katona et al. (1999), 26Soriano et al. (1990), 27Claiborne et al. (1990), 28Buckmaster et al. (2002a), 29Nomura et al. (1997a), 30Nomura et al. (1997b). Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

(Buckmaster et al., 1996, 2002a; Nomura et al., 1997a, b). Buckmaster and Jongen-Relo (1999) estimated the total number of GABAergic neurons in the molecular layer of the dentate gyrus as approximately 10,000. An even distribution between inner, medial, and outer molecular layers is assumed, yielding 4,000 MOPP cells with somata located in the inner molecular layer (Han et al., 1993). Note that this estimate excludes the molecular layer interneurons that project primarily outside of the dentate gyrus such as the outer molecular layer interneurons projecting to the subiculum (Ceranik et al., 1997).

The third piece: connectivity between cell types In the past decade, a vast amount of high-quality data about the connectivity of the dentate gyrus has been collected, and this data serves as an invaluable resource in constructing the cell-type specific connectivity matrix for the dentate gyrus. The connectivity for each cell type is summarized in Table 1 and described below. Each value was determined by a detailed survey of the anatomical literature and was based on an assumption of uniform bouton density along the axon of the presynaptic cell. This assumption is in agreement with

643

Fig. 2. Gaussian fits to experimentally determined distributions of axonal branch length used in construction of the models of the dentate gyrus. (A) Plot shows the averaged axonal distribution of 13 granule cells (Buckmaster and Dudek, 1999) and the corresponding Gaussian fit. (B) Fit to the septo-temporal distribution of axonal lengths of a filled and reconstructed basket cell (Sik et al., 1997). (C) Fit to the axonal distribution of a CA1 axo-axonic cell (Li et al., 1992). (D) Gaussian fit to the averaged axonal distributions of three HIPP cells from gerbil (Buckmaster et al., 2002a, b). (E) Fit to averaged axonal distributions of three mossy cells illustrates the characteristic bimodal pattern of distribution (Buckmaster et al., 1996). (F) Histogram of the axonal lengths of a HICAP cell along the long axis of the dentate gyrus (Sik et al., 1997) and the Gaussian fit to the distribution. All distributions were based on axonal reconstruction of cells filled in vivo. In all plots, the septal end of the dentate gyrus is on the left (indicated by negative coordinates) and the soma is located at zero. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

the in vivo data of Sik et al. (1997), and it greatly simplifies the estimation of connectivity from anatomical data on axonal length and synapse density per unit length of axon. Granule cells Mossy fibers (GC axons) in a healthy rat dentate gyrus are primarily restricted to the hilus (97%),

with few collaterals (3%) in the GC layer (Buckmaster and Dudek, 1999). In addition to MCs (Acsady et al., 1998), mossy fibers have also been shown to contact BCs (Buckmaster and Schwartzkroin, 1994; Geiger et al., 1997) and PV interneurons (Blasco-Ibanez et al., 2000). With a total of 400–500 synaptic contacts made by a single mossy fiber (Acsady et al., 1998), the 3% of each axon located in the GC layer (Buckmaster and

644

Dudek, 1999) is estimated to contact 15 BCs and 3 axo-axonic cells. In the hilus, a single GC forms large, complex mossy fiber boutons that innervate 7–12 MCs (Acsady et al., 1998), while an estimated 100–150 mossy fiber terminals target hilar interneurons with approximately one synapse per postsynaptic interneuron (Acsady et al., 1998). Gulyas et al. (1992) estimated that a single spiny calretinin-positive cell (presumed HIPP cell) is contacted by approximately 9,000 GCs. With 12,000 HIPP cells and 1,000,000 GCs, each GC can be estimated to contact approximately 110 HIPP cells and 40 HICAP cells. Additionally, in agreement with the presence of mossy fiber terminals on aspiny calretinin-positive interneurons (Acsady et al., 1998), 15 mossy fibers are expected to synapse onto IS cells. Since mossy fibers avoid the molecular layer (Buckmaster and Dudek, 1999) in the healthy dentate gyrus, it is assumed that they do not contact MOPP cells. However, in animal models of temporal lobe epilepsy, sprouted mossy fibers have been shown to contact up to 500 postsynaptic GCs (Buckmaster et al., 2002b).

Mossy cells A single filled mossy cell axon has been reported to make 35,000 synapses in the inner molecular layer (Buckmaster et al., 1996; Wenzel et al., 1997). Assuming a single synapse per postsynaptic cell, a single mossy cell is estimated to contact 30,000–35,000 GCs. Of the 2700 synapses made by a single MC axon in the hilus, approximately 40% target GABA-negative neurons (Wenzel et al., 1997). As each mossy cell is estimated to make 1–5 synaptic contacts on a single postsynaptic mossy cell (Buckmaster et al., 1996), it is estimated that each mossy cell contacts approximately 350 other MCs (this is likely to be a generous estimate since it is based on the assumption that all GAD negative somata in the hilus represent MCs). The remaining 60% of the hilar mossy cell axons target GABA-positive cells (Buckmaster et al., 1996; Wenzel et al., 1997), with no reports demonstrating mossy cell synapses onto IS cells. Assuming that there is no preferential target selectivity between HIPP and HICAP cells and that each

postsynaptic hilar interneuron receives two synaptic contacts from a single mossy cell axon (Buckmaster et al., 1996), each mossy cell is estimated to contact 600 HIPP and 200 HICAP cells. There is very low mossy cell to interneuron connectivity in the inner molecular layer (Wenzel et al., 1997); MCs could contact 5–10 basket and axo-axonic cells and approximately 5 MOPP cells with somata in the inner molecular layer (Han et al., 1993). Basket cells In the CA3 region of the rat hippocampus, each principal cell is contacted by approximately 200 BCs (Halasy and Somogyi, 1993), but a GC in the dentate could be contacted by as few as 30 BCs. Assuming that each of the 1,000,000 GCs is contacted by 115 BCs each making 1–20 synaptic connections (Halasy and Somogyi, 1993; Acsady et al., 2000), it can be estimated that each BC contacts approximately 1,250 GCs. MCs receive 10–15 basket cell synapses (Acsady et al., 2000), resulting in an estimate of 75 BC to mossy cell synapses per BC. Approximately 1% of the 11,000 synapses made by a single basket cell axon in the GC layer are onto other BCs (Sik et al., 1997) with 3–7 synapses per postsynaptic cell (Bartos et al., 2001). Consequently, each BC in the dentate gyrus contacts approximately 35 other BCs. Since hilar and molecular layer interneurons are not a major target of basket cells (Halasy and Somogyi, 1993), a single BC may contact 0–1 HIPP cells. Similarly, the BC synapses onto axo-axonic cells, HICAP, and MOPP cells are assumed to be negligible. As PV cells preferentially contact other PV-positive cells in the hilus (Acsady et al., 2000), we assume that BCs do not contact calretinin-positive IS cells (Gulyas et al., 1992). Axo-axonic cells Most synapses made by axo-axonic cell axons are thought to target GC axon initial segments (Halasy and Somogyi, 1993). However, a small fraction of axon collaterals also descend into the superficial and deep hilus (Han et al., 1993; Freund and Buzsaki, 1996). In neocortex, an axo-axonic

645

cell makes 4–10 synapses on the axon intial segment of a postsynaptic cell (Li et al., 1992). With 22,000,000 estimated axon initial segment synapses in the GC layer (Halasy and Somogyi, 1993) and 4 synapses per postsynaptic cell (based on the data from the neocortex from Li et al., 1992), each of the 2,000 axo-axonic cells are estimated to target approximately 3,000 GCs. MCs receive axo-axonic cell input (Ribak et al., 1985), and with the comparatively small fraction of axons from axo-axonic cells in the hilus (Han et al., 1993; Freund and Buzsaki, 1996), it can be estimated that axo-axonic cells target a number of MCs equal to approximately 5% of their GC targets, which results in approximately 150 MCs. Since axo-axonic cells primarily target the axon initial segment of non-GABAergic cells (Halasy and Somogyi, 1993; Freund and Buzsaki, 1996), it may be assumed that these cells do not project to interneurons.

HIPP cells HIPP cells have been estimated to contact about 1,600 GCs and 450 BCs with 1–5 synapses per postsynaptic cell (Sik et al., 1997). MCs can have one dendrite in the molecular layer (Amaral, 1978; Ribak et al., 1985; Scharfman, 1991) which can be targeted by HIPP cell axons, whereas GCs have two primary dendrites (Claiborne et al., 1990; Lubke et al., 1998). Since the mossy cell to GC ratio is approximately 1:30, the mossy cell dendrites constitute only approximately 1/60 of the targets for HIPP cells. Increasing this fraction to approximately 1/45 to account for the presence of a few HIPP cell contacts on MC in the hilus (Buckmaster et al., 2002a) results in an estimate that each HIPP cell contacts approximately 35 MCs. HIPP cell axonal divergence onto HICAP and MOPP cells in the molecular layer can be assumed to be similar to that of somatostatin-positive cells in CA1 (Katona et al., 1999) and estimated to be 15 connections onto each population. The HIPP cell axonal divergence to axo-axonic cells is estimated to be between the divergence to basket and HICAP cells; therefore the HIPP cell axon likely contacts 30 axo-axonic cells.

MOPP cells MOPP cells target an estimated 7,500 GCs in the rat dentate gyrus. While MOPP cell axons project in the horizontal axis to a similar extent as HIPP cells, they show considerably less collateralization (Han et al., 1993), resulting in an estimate of half as many synapses onto MOPP and HICAP cells as HIPP cells make. As MOPP cell axons are restricted to the molecular layer (Han et al., 1993) and do not target the basal dendrites of BCs, they are assumed to contact less than 1/10 the number of BCs targeted by HIPP cells. Likewise, MOPP cells with axons restricted to the outer and middle molecular layers (Han et al., 1993) would not target the hilar dendrites of axo-axonic cells (Soriano et al., 1990) or the axo-axonic cells with somata and proximal dendrites in the hilus (Han et al., 1993). It is estimated that MOPP cells only contact 1–2 axo-axonic cells. As the MOPP cell axonal arbors in the molecular layer (Han et al., 1993) do not overlap with major parts of the dendritic arborizations of MCs (Frotscher et al., 1991), HIPP cells (Han et al., 1993; Sik et al., 1997; Katona et al., 1999) or IS cells (Gulyas et al., 1996), the connectivity to these cells is negligible.

HICAP cells Sik et al. (1997) estimated that the septo-temporal extent and bouton density of HICAP cell axons is similar to the HIPP cell axons, whereas the estimated axonal length of HICAP cells is approximately half of the HIPP cell axonal length. Thus, it is estimated that HICAP cells contact about half the number of GCs contacted by HIPP cells. However, since HICAP cells have an additional 3% of axon collaterals in the hilus (Sik et al., 1997), their number of postsynaptic MCs can be assumed to be the same as that of the HIPP cells. HICAP cells are assumed to contact less than half the number of BCs targeted by HIPP cells (
175) and a negligible number of axo-axonic cells. With a total of 26,000 synapses from a single HICAP cell axon (Sik et al., 1997), approximately 700 synapses should be present in the hilus. Assuming 2–5 synapses per postsynaptic cell, each HICAP

646

cell could contact 100–300 hilar cells. Each HICAP cell is assumed to target 50 HIPP and HICAP cells, which, along with 35 synapses on MCs, is in the estimated range. Although the total axonal length of HICAP cells is only about half of that of HIPP cells, the number of MOPP cells targeted is assumed to be the same (
10–20), as the HICAP cell axons primarily project to the inner molecular layer where both cell bodies and proximal dendrites of MOPP cells are located (Han et al., 1993). IS cells IS cells contact an estimated 100–800 other IS cells and 5–10 (presumably CCK positive) BCs (Gulyas et al., 1996). Acsady et al. (2000) suggested that CCK cells would include both basket cell and HICAP morphologies and that IS cells also project to somatostatin-positive HIPP cells. These data result in an estimate that IS cells project to 5–10 HICAP cells and HIPP cells. The fourth piece: spatial constraints and axonal distributions In addition to the number of connections made by any given cell, the distribution of these connections along the septo-temporal axis of the dentate gyrus is very important. Fortunately, in vivo single cell fills have been performed which provide a very complete description of the axonal extent of the eight cell types discussed above. Averages of these in vivo axonal distributions can be fit with either a single or double Gaussian for each cell type, which then defines the model axonal distribution for each cell of that type (Fig. 2). These Gaussians then describe the probability of any given cell connecting to another cell based on the distance between the two. Putting the pieces together Once the border of the puzzle is complete (the connection matrix in our analogy), it is possible to start assembling the bulk of the picture. In the case

of modeling, this corresponds to a number of distinct obstacles. Single cell models must be created with sufficient detail to function appropriately in the network; network scaling must be implemented (i.e., what percentage of the actual cell numbers, connections, etc., will be represented in the network); receptor types and synaptic data must be included; and network topology must be determined (will the network be a strip or a ring, e.g., and how will connection probabilities be implemented to account for the shape and size of the network). Finally, the stimulation paradigm must be determined so that the fully connected network will actually start to produce data. These obstacles, and ways to overcome them, are discussed in detail in this section.

Multicompartmental single cell models In a network model that proposes to simulate both somatic and dendritic inputs and model synaptic contacts on specific dendritic compartments, it is necessary to develop multicompartmental, biophysically realistic models of individual cell types. It is not always necessary to develop these models from scratch, however, as ModelDB (http://senselab.med.yale.edu/senselab/modeldb/) is a searchable resource that makes published models of several cell types including dentate GCs available for download (Aradi and Holmes, 1999a; Santhakumar et al., 2005). These models can be readily adapted to a wide variety of networks. Morphological properties for multicompartmental models of other cell types not available on ModelDB, such as some of the other dentate cell types, can be developed based on the data reported in the literature (Buckmaster et al., 1993, 2002b; Freund and Buzsaki, 1996; Geiger et al., 1997; Bartos et al., 2001). The intrinsic properties of cell types can be modeled based on data from whole cell physiological experiments obtained in the presence of blockers of synaptic activity (Lubke et al., 1998). Detailed descriptions of how to model passive and active properties of individual cells can be found in The Neuron Book (Carnevale and Hines, 2006), but we will discuss a couple of relevant issues here

647

briefly. First, since GCs (Desmond and Levy, 1985) and MCs (Amaral, 1978) are rich in dendritic spines, it is necessary to account for the membrane area contribution of the spines (Rall et al., 1992). This can be accomplished by decreasing membrane resistivity (increasing leak conductance) and increasing membrane capacitance (Aradi and Holmes, 1999b). An additional consideration is the presence of spontaneous activity in some cell types such as MCs (Ishizuka et al., 1995; Ratzliff et al., 2004), as well as their lower input resistance in the presence of background synaptic activity (Ratzliff et al., 2004), compared to the input resistance in ionotropic glutamate and GABA receptor antagonists (Lubke et al., 1998). Spontaneous firing rate can be simulated via a constant direct current injection (Santhakumar et al., 2005), and the synaptic background activity can be simulated as a fluctuating point conductance with balanced excitation and inhibition, as described in Destexhe et al. (2001). Both excitatory cell types (GCs and MCs) are well defined and have a vast amount of electrophysiological data available for the construction of single cell models. However, the wealth of information available for the interneurons is substantially less for many cell types. Additionally, the role of interneurons in the network must be considered with regard to the question being asked. For example, in a model network that is pursuing the question of the role of mossy fiber sprouting and hilar cell loss in epileptogenesis (Santhakumar et al., 2005; Dyhrfjeld-Johnsen et al., 2007), several factors must be considered. First, inhibition can operate in two principal modes. One mode is feedforward inhibition, where an interneuron is activated by the same afferent input as the excitatory neuron that it contacts. The basket cell is the primary source of feed-forward inhibition in the dentate gurus (Freund and Buzsaki, 1996). The second mode is feedback inhibition, where interneurons are activated by excitatory cells which they subsequently inhibit. The BCs and HIPP cells are the two key feedback inhibitory cells in the dentate (Freund and Buzsaki, 1996; Buckmaster et al., 2002a). Another important aspect of inhibition is the location of synaptic contact between interneuron and principal cell. Perisomatic inhibition

by BCs is considered crucial in maintaining weak GC activity in response to afferent input. The HIPP cells, on the other hand, synapse on the dendritic region near the afferent inputs, and they are likely to modulate integration of dendritic inputs. The loss of this dendritic inhibition with an essentially intact or increased somatic inhibition has been proposed to be particularly relevant to epileptogenesis (Cossart et al., 2001). Furthermore, BCs are resistant to cell death whereas HIPP cells are extremely vulnerable in epileptic tissue (Buckmaster et al., 2002a). These features make the basket and HIPP cells essential constituents of the dentate model. The remaining four interneurons also possess unique features of inhibition. The MOPP cells operate by providing GCs with feed-forward dendritic inhibition (Halasy and Somogyi, 1993). The HICAP cells synapse on the proximal dendrites of GCs, near where mossy cell axons terminate and provide feedback inhibition (Freund and Buzsaki, 1996). The axo-axonic cells provide GABAergic input at the axon initial segments of GCs and MCs and potentially play powerful roles in modifying spike initiation (Freund and Buzsaki, 1996; Howard et al., 2005). The IS cells connect exclusively to other interneurons and form a well connected network that could modulate excitability and synchrony of the network. Unfortunately, detailed physiological data for the development of realistic models of MOPP, HICAP, axo-axonic, and IS cells are not readily available. Therefore, in an approach that avoids errors of commission one could choose to include only basket and HIPP cells in the network. However, this makes it essential to perform control simulations to examine whether augmenting inhibition (in order to compensate for excluded cell types) influences the overall outcome of the network simulations.

Network scaling Once the single cell models are complete and ready to go, it is time to start thinking about how big the network is going to be. With the immense increases in computational power over the last

648

couple of decades, very large networks are quite feasible. However, some questions may not require such large networks, and often computational resources will be limited. Also important in determining what scale a network should be is whether the network includes functional model cells that require a great deal of computational power to process. Purely structural network models (that do not use multicompartmental functional single cell models), also called network graphs, can provide a great deal of useful information about neuronal circuits. Indeed, Dyhrfjeld-Johnsen et al. (2007) created an extraordinarily large 1:1 scale structural model of the dentate gyrus that yielded a great deal of new data and will be discussed briefly in Section ‘‘The big picture’’. Putting purely structural graphs aside for now though, and focusing on the functional model that we are building in this chapter, we can determine how to scale the cell numbers. We know that the GC:MC:BC:HIPP ratio is 1000:30:10:12. In agreement with these ratios, the Lytton et al. (1998) model that focused on the network effects of sprouting included 50 GCs, 2 MCs, and 2 interneurons. Using this relatively small network representation of the dentate they demonstrated that the effect of sprouting on enhancing network excitability was strongly limited by the intrinsic firing properties of GCs (e.g., strong spike frequency adaptation). Although this model made the prediction that mossy fiber sprouting could lead to hyperexcitability, the network was highly interconnected as a consequence of the small size. This network architecture, combined with a very strong perforant path synaptic conductance, could have resulted in artificial synchronous activity in the network. Increasing the size of the Lytton et al. (1998) model by 10-fold entails building a 1:2000 scale dentate with 500 GCs, 15 MCs, 6 BCs and 6 HIPP cells. This 500+ cell model can be readily simulated on a laptop and allows for manipulating hilar cell numbers and connectivity patterns (Santhakumar et al., 2005). However, even in a network of this size, artificially large conductances as well as other factors (e.g., strong edge effects and unrealistic distance dependent cell-type specific connectivity) could influence the results. With the current computational power of state-of-the-art multiprocessor computers,

the size of a realistic dentate model can be scaled up further to a 1:20 scale model with 50,000 GCs (Dyhrfjeld-Johnsen et al., 2007). The specific connection parameters that can be used for a model of this size are shown in Table 2. In Section ‘‘The big picture,’’ we will discuss results obtained from this very large-scale network with direct implications for epileptogenesis.

Receptor types and synapses Building a model of a dentate network with hundreds or thousands of biophysically realistic multicompartmental model cells is only a meaningful venture if those cells have a way to talk to one another. For this reason, we must consider neurotransmitter-receptor types that must be included in the network. Again, the types of receptors, synapses, and other cellular interaction devices (such as gap junctions) to include in the network will depend on the particular question the network is designed to answer. For the example network we have been building throughout this chapter, the goal is to determine the effects of mossy fiber sprouting and hilar cell loss on excitability. To examine these effects, the minimal requirement is to incorporate ionotropic glutamatergic AMPA synapses and GABAA synapses. Experiments from both normal and epileptic animals are constantly ongoing, and as the data become available it will be possible to add other receptors such as NMDA and metabotropic receptors. It would also be interesting to incorporate greater diversity at individual synapses through incorporation of mechanisms such as short-term plasticity. Each addition to the network must be carefully considered, however, as greater complexity necessarily increases the load on the computational resources. Since the basic circuit effects of sprouting and cell loss can be simulated by including AMPA and GABAA synapses, these will be the types included in our example model. Postsynaptic conductances can be represented as a sum of two exponentials (Bartos et al., 2001). The peak conductance (gmax), rise and decay time constants, and the synaptic delays (distinct from the axonal conduction delay described later) for each network connection can be estimated from

649 Table 2. Parameters of the 50,000+ neuron functional model network Functional model network parameters From

To –>

Granule cellsa (50,000)

Convergence Divergence Synapse weight (nS)

Mossy cells (1500)

GC

MC

BC

HC

68.03 68.03 1.00

78.05 2.34 0.20

370.95 3.71 0.94

2266.64 27.19 0.10

Convergence Divergence Synapse weight (nS)

243.62 8120.82 0.30

87.23 87.23 0.50

5.59 1.86 0.30

375.53 150.21 0.20

Basket cells (500)

Convergence Divergence Synapse weight (nS)

3.11 313.22 1.60

6.31 18.93 1.50

8.98 8.98 7.60

n/a n/a n/a

HIPP cells (600)

Convergence Divergence Synapse weight (nS)

4.82 401.86 0.50

3.76 9.39 1.00

140.13 116.77 0.50

n/a n/a n/a

Perforant pathb

Synapse weight (nS)

20

17.5

10

n/a

Notes: The cell numbers (column 1) and synaptic connectivity values and strengths in the functional model network used for the activity simulations in Fig. 4. Note that this network is smaller (50,000+ cells) than the full-scale dentate gyrus (over 1 million cells) model described in Section ‘‘Structural alterations in the dentate during epileptogenesis’’. Therefore, the connectivity had to be adjusted from what is shown in Table 1. Convergence is given as number of connections onto a single postsynaptic neuron (row 1) from a presynaptic neuronal population (column 1). For example, 243 mossy cells converge on a single granule cell in this network. Divergence is given as the number of connections to a postsynaptic population (row 1) from a single presynaptic neuron (column 1). For example, a single mossy cell makes synapses on 8120 postsynaptic granule cells in this network. The strengths of the connections are given in nS. For example, the strength of the excitatory synapse formed by a single mossy cell on a single granule cell is 0.3 nS. a Granule cell to granule cell connections represent values at 100% sprouting. b Perforant path input to 5000 granule cells (2 synapses each), 50 basket cells (2 synapses each) and 10 mossy cells (1 synapse each) in the central 10 bins of the network model. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

experimental data (Kneisler and Dingledine, 1995; Geiger et al., 1997; Bartos et al., 2001; Santhakumar et al., 2005) and details of all of the specific cellular and synaptic properties can be found in Santhakumar et al. (2005). When making such estimates, care must be taken to ensure that comparable conditions were used in the experiments that provide the data. For example, since the data for synaptic conductances is typically obtained by paired recording experiments from several different groups, the temperature at which the recordings were performed must be accounted for. Q10 estimates can be used to convert kinetic data from room temperature recordings to the appropriate values at physiological temperature. Axons can be modeled implicitly by activating a given synapse after either a static or distance-dependent delay once a presynaptic cell crosses a preset membrane potential threshold (Bartos et al., 2001).

Network topology The shape of the network and the way in which connections are made depend strongly on the size of the network. For instance, a network of approximately 50 or 500 cells (Lytton et al., 1998; Santhakumar et al., 2005, respectively) may require a ring structure to avoid edge effects (i.e., cells on the edge of a network have as few as half as many pre- and postsynaptic targets compared to cells in the middle). A network of 50,000+ cells (Dyhrfjeld-Johnsen et al., 2007), on the other hand can be set up as a linear strip, more similar to the actual topology of the biological dentate. Similarly, small networks may require non-topographic connectivity. That is, the postsynaptic targets of each cell are selected at random from the pool of potential target neurons while maintaining the cell type specific divergence and

650

convergence. Larger networks can incorporate the spatial rules discussed above in Section 7 The fourth piece: Spatial constraints and axonal distributions; the connectivity between cells can be distributed according to the experimentally derived axonal distributions of the various cell types (Fig. 2).

Network stimulation A network such as the dentate gyrus, in which most cells are not designed to be spontaneously active, requires some afferent input to initiate network activity. This input can come in a number of forms, again dependent on the specific question as well as the particular network being modeled. For the dentate gyrus as a whole, it makes sense to simulate a perforant path input from the entorhinal cortex. This input to the network is located on both dendrites of all GCs and the apical dendrites of all BCs. Since 15% of MCs have also been shown to receive direct perforant path input (Scharfman, 1991; Buckmaster et al., 1992), the number of MCs that should receive input in a given model can be easily calculated. Mass stimulation of the perforant path can be simulated by activating a maximum peak AMPA conductance in cells postsynaptic to the stimulation (Santhakumar et al., 2000). The strength of the perforant path input in the 50,000+ cell functional model from Dyhrfjeld-Johnsen et al. (2007) is given in Table 2 (note the very large synaptic weight for perforant path input compared to the small weights for cell to cell synapses). In the biological network and in physiological studies it is unlikely that the entire network is activated simultaneously. Rather, focal activation can be simulated by activating the input to GCs and BCs located in a model hippocampal ‘‘lamella’’. If focal network activation is undesirable, spontaneous network activity can be simulated instead. This can be accomplished by providing uncorrelated activation (perforant path inputs with Poisson distributed interspike intervals) to each granule cell, basket cell, and 15% of the MCs. Other methods for initiation of network activity can be devised as necessary.

The big picture Following essentially the strategy outlined in the previous two sections, dentate gyrus modelers have assembled a number of different puzzles, producing dentate models of a wide range of sizes aimed at answering a large number of important questions. We have walked through the creation of a very large-scale dentate model capable of handling functional network simulations with greater than 50,000 cells. However, we have not yet discussed how a model of this type can help us gain a better understanding of disease processes and hopefully how to treat them. Without the ability to provide information of this nature, models are simply incomplete representations of reality and of little use. Fortunately, models can provide a great deal of information because they give scientists complete control over variables that cannot be isolated in typical experimental situations. In this section, we will briefly describe an example in which using large scale dentate gyrus modeling allowed Dyhrfjeld-Johnsen et al. (2007) to isolate structural changes in the dentate during epileptogenesis and conclude that structural changes alone can lead to hyperexcitability.

Structural alterations in the dentate during epileptogenesis As mentioned throughout this chapter, two primary structural alterations occur in the dentate gyrus during epileptogenesis: mossy fiber sprouting and hilar cell death (including hilar interneurons and MC). These two processes (which we will refer to simply as ‘‘sclerosis,’’ defined here as the two common structural changes in the dentate associated with mesial temporal lobe sclerosis) can be implemented in a dentate gyrus model according to the data driven approach given above. Detailed descriptions of how to implement both of these processes can be found in Santhakumar et al. (2005) and Dyhrfjeld-Johnsen et al. (2007). Briefly, a maximal level of sclerosis can be simulated by adding an average of approximately 275 granule cell to granule cell connections (Buckmaster et al., 2002b) and by removing all hilar cells from the

651

network. Using this maximum value as a point of reference, the severity of sclerosis can be approximated by simulating percentages of the maximum (e.g., at 50% sclerosis, simulating epileptogenesis following a moderate head injury (Howard et al., 2007), each granule cell would contact approximately 137 other GCs, and 50% of hilar cells would be lost). A 1:1 scale structural model of the dentate gyrus can be constructed where each cell is simply a node in the computer (nodes have no functional properties and are therefore much less computationally demanding than the multicompartmental cell models described previously) and synapses between cells are represented as non-functioning directed links (Dyhrfjeld-Johnsen et al., 2007). This structural network can then be utilized to understand the basic structure of the healthy dentate gyrus and how that structure changes during epileptogenesis. In order to understand how network structure can be measured and later how structural changes have functional implications, it is necessary to explain the parameters that define network topology. Two measures that Watts and Strogatz (1998) originally employed to assess the structure of the nervous system of the worm C. elegans can be used to analyze the structure of the dentate network: the average path length, L, and the average clustering coefficient, C. The average path length is defined as the average number of steps required to move from any node to any other node in the network, and it is therefore a measure of how well connected the network is globally. The average clustering coefficient, on the other hand, is a measure of local connectivity. It is defined as the fraction of all possible connections between ‘‘postsynaptic’’ nodes of a given node that are actually formed. The average clustering coefficient for an entire network is simply the average of the clustering coefficients for each node in the network. The topological measures, C and L, can be best thought of in terms of a social network, or a network of friends. C measures the probability that any two friends of a given person also know each other. L measures how many steps in a chain of friends or acquaintances would be necessary to connect two random people. A common theory

that utilizes this parameter is that anyone on the planet is separated from anyone else by up to six degrees of separation, or six people. While this theory has never been satisfactorily proven, it has led to interesting experiments and even a party game called the Six Degrees of Kevin Bacon, in which the participants have to connect any actor or actress to Kevin Bacon through their film roles in as few steps as possible. Based on these two topological measures, networks can be divided into three classes: (1) Regular, with a long path length and large clustering coefficient; (2) Random, with a short path length and small clustering coefficient; and (3) Small World, with a short path length and large clustering coefficient. Representative graphs of these types of networks are displayed in Fig. 3. As shown, the regular network (Fig. 3A) is characterized by a high local connectivity, but the distance (via connected nodes) between any two random nodes, especially nodes at the ends of the graph, can be quite large. On the other hand, the random network (Fig. 3B) contains many long distance connections, contributing to a very short average path length. There is little local connectivity however, as nodes are just as likely to connect to far away nodes as they are to their neighbors. The small world network (Fig. 3C) acts as a sort of compromise between the random and regular graphs. Indeed, there is a high degree of local connectivity combined with a number of long-range connections, yielding a both locally and globally well-connected network. Small world networks can also be thought of in terms of social networking. Humans often form a cohesive group of friends who are well connected locally due to various constraints such as geography, common activities and so on. Additionally, any given person in the group may know some people who are totally unrelated to the local group, but who each have cohesive local groups of their own, thus creating numerous long distance connections between the locally well-connected networks. So how do these L and C values play a role in understanding the dentate gyrus? Analysis of the structural model of the dentate gyrus in the healthy state, without any mossy fiber sprouting or hilar cell loss, indicates that the dentate gyrus is a small

652

Fig. 3. Schematics of three basic network topologies. (A) Regular network topology. The nodes in a regular network are connected to their nearest neighbors, resulting in a high degree of local interconnectedness (high clustering coefficient C), but also requiring a large number of steps to reach other nodes in the network from a given starting point (high average path length L). (B) Random network topology. In a random network, there is no spatial restriction on the connectivity of the individual nodes, resulting in a network with a low average path length L but also a low clustering coefficient C. (C) Small world network topology. Reconnection of even a few of the local connections in a regular network to distal nodes in a random manner results in the emergence of a small world network with a high clustering coefficient C but a low average path length L. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

world network. It has a very high clustering coefficient and a path length that is only 2.68, approximately the same as the path length of the C. elegans nervous system, a nervous system with only 282 cells! This finding means that any cell is connected to any of the other one million cells in the dentate network by less than three synapses, and cells are highly connected locally, resulting in fast local computations and the ability to efficiently relay signals to distant parts of the network. While the analysis of the healthy dentate gyrus provides useful and novel information about its structure, perhaps the most interesting and counterintuitive structural information comes from the analysis of networks undergoing sclerosis. These networks demonstrate that the small world characteristics of the dentate gyrus are actually enhanced up to approximately 80% sclerosis, despite a massive loss of connections (see Section ‘‘Mossy cells’’ for MC connectivity; note the massive divergence onto GCs that is lost as sclerosis progresses and MCs die). In fact, the total number of links in the network decreases by 74% while the total number

of nodes lost due to hilar cell death is only 4.5% at maximal sclerosis, yet the network becomes seemingly more locally and globally well connected. This finding predicts that during epileptogenesis, the dentate gyrus may become more readily able to transmit information throughout the network and likely increase synchronous firing as well, phenomena that could very well contribute to a seizure phenotype. It is only at nearly 100% sclerosis that the dentate gyrus network transforms into a more regular network structure, predicting a counterintuitive decrease in hyperexcitability at maximal sclerosis.

Functional implications of structural alterations Now that we have discussed the structural analysis of the healthy and sclerotic dentate gyrus, it remains to be seen whether network activity is actually influenced by the structural changes in the ways we might expect based on the predictions in the preceding section. As in the 1:1 scale structural

653

model, sclerosis can be implemented in a functional model of the dentate gyrus. This model contains approximately 50,000 cells, owing to the increased need for computational resources to perform the calculations necessary with multicompartmental biophysically realistic cells (see Section ‘‘Multicompartmental single cell models’’). Simulations can be performed as sclerosis progresses in steps (e.g., perform simulations at 20 then 40%, and so on), and activity in the network can be quantified by a number of different measures. Four different measures were used in DyhrfjeldJohnsen et al. (2007): (1) duration of granule cell firing (the time of the last granule cell action potential in the network minus the time of the first granule cell action potential); (2) average number of action potentials per granule cell; (3) time until activity of the most distant GCs from the stimulation point (i.e., latency to full network activity); and (4) synchrony of granule cell action potentials. Results of the functional simulations reveal that network activity very closely parallels structural alterations in the dentate gyrus. The healthy non-sclerotic dentate model shows very limited firing in response to a single perforant path input (Fig. 4A), in accordance with the biological data (Santhakumar et al., 2001). Additionally, as sclerosis progresses, the functional network becomes increasingly hyperexcitable up to approximately 80% sclerosis (Figs 4B–E), again agreeing with in vitro measures of epileptiform activity (Rafiq et al., 1995) and in accordance with the enhanced small world features revealed by the structural analysis. A very interesting phenomenon then occurs at levels of sclerosis exceeding 80%. Hyperexcitability of the functional model network actually decreases (Fig. 4F)! The level of sclerosis at which hyperexcitability diminishes corresponds perfectly to the point at which the network transforms from a small world structure to a more regular network. Thus, the functional network follows the same pattern that is seen in the structural network, and the functional effects are solely due to the structural network alterations, as no intrinsic cellular or synaptic properties are altered in the model as sclerosis progresses. Interestingly, network dynamics are also affected by structural changes; a relatively uniform pattern of granule cell activity (from 40 to 60% sclerosis; Fig. 4C, D)

transforms into a pattern with distinct waves of activity (from 80 to 100% sclerosis; Fig. 4E, F) that can collide and mutually annihilate (Netoff et al., 2004; Roxin et al., 2004). These findings not only support the important role of structural alterations in determining functional network activity, but they are in agreement with experimental observations in both epileptic animals and humans. Indeed, no studies that have quantified hilar cell loss in animal models of epilepsy ever reported 100% cell loss (Cavazos and Sutula, 1990; Cavazos et al., 1994; Leite et al., 1996; Buckmaster and Dudek, 1997; Mathern et al., 1997; Buckmaster and Jongen-Relo, 1999; Gorter et al., 2001; van Vliet et al., 2004; Zappone and Sloviter, 2004). Additionally, in surgically removed specimens from pharmacologically intractable human temporal lobe epilepsy patients (Gabriel et al., 2004), cell counts showed that only approximately 80% of hilar cells were lost, even in patients with severe sclerosis (Blumcke et al., 2000). This finding coincides perfectly with the period of maximal hyperexcitability in the functional model networks, when sclerosis reaches 80%.

Understanding the limitations of models It is important to consider that no model can ever fully replicate the system that it is modeling. As a result, it is crucial to be certain that results and conclusions derived from models are robust under a number of potentially confounding situations. The modeling strategy presented in this chapter attempts to ensure that available biological data is represented and accounted for. However, a number of specific components of the dentate circuit and connectivity often cannot be modeled due to lack of precise data. Additionally, depending on the particular question the model is designed to answer, several factors may be purposely omitted in order to best focus on the relevant issues. For the model presented in this section, for example, changes to intrinsic cellular and synaptic properties during sclerosis were omitted since the experiments were designed to study structural alterations in isolation. For these reasons, an extensive series of control simulations must be

654

Fig. 4. Effects of the sclerosis-related topological changes on granule cell activity in functional model networks. A–F. Raster plots of the first 300 ms of action potential discharges of granule cells in the 50,000+ neuron functional model network (granule cells number 1–50,000, plotted on the y-axis; each dot corresponds to an action potential fired by a granule cell) at increasing degrees of sclerosis. Network activity was initiated by a single stimulation of the perforant path input to granule cells number 22,500–27,499 and to 10 mossy cells and 50 basket cells (distributed in the same area as the stimulated granule cells) at t ¼ 5 ms (as in Santhakumar et al., 2005). Note that the most pronounced hyperactivity was observed at sub-maximal (80%) sclerosis. Reproduced with permission from Dyhrfjeld-Johnsen et al. (2007).

performed to test the effects of a wide variety of conditions that are not well constrained by the available experimental data. Controls will vary from model to model, but many are needed in any situation. Those performed for the structural model presented here were the following: 1) Variations in cell numbers. 2) Variations in connectivity estimates. 3) Inhomogeneous distribution of neuron densities along the septo-temporal axis. 4) Inhomogeneity in connectivity along the transverse axis. 5) Altered axonal distributions at the septal and temporal poles (the anatomical boundaries of the dentate gyrus). 6) Offset degrees of sprouting and hilar neuron loss.

7) Implementation of a bilateral model of the dentate gyrus including commissural projections. All of the controls employed for the structural model were also used as controls for the functional model. In addition, the functional model had its own set of controls, including: 1) Double inhibitory synaptic strengths. 2) Axonal conduction delays. 3) Spontaneous instead of stimulation-evoked activity. All of the results from these control simulations supported the basic findings of the original models, thus greatly strengthening the primary conclusions.

655

Tackling future puzzles and problems Modeling has come a long way in the last several decades. Even 10 years ago, large-scale realistic modeling was nearly impossible due to computational restrictions and the requirement of vast supercomputers to perform necessary calculations. Today, however, it is possible to perform quite indepth modeling studies on a home desktop. Models such as the 50,000 cell network in Dyhrfjeld-Johnsen et al. (2007) can be realistically run on a dual processor system (albeit with 16–32 GB of RAM). Smaller networks such as the 500 cell model in Santhakumar et al. (2005) can be run in a few minutes on a typical laptop. These are wonderful advancements, as the development of large-scale models that can simulate brain circuits with greater realism is critical to our understanding of the experimentally determined cellular and molecular underpinnings of diseases like epilepsy. Additionally, large-scale realistic networks are necessary for our understanding of structure–function relationships, especially in a highly plastic system like the dentate gyrus. The exact structural changes that occur in many pathological states such as epilepsy are as yet unknown. For example, recent evidence indicates that local connection probability in various brain areas may be modified by intra-class correlations (Yoshimura and Callaway, 2005; Yoshimura et al., 2005) and over-representation of small network motifs (Milo et al., 2002; Reigl et al., 2004; Sporns and Kotter, 2004; Song et al., 2005). While no direct evidence is present to implicate such factors in the dentate gyrus, it is likely that unknown structural alterations have a significant impact on network activity. Computational modeling studies provide an excellent tool for determining what the probable result of a wide variety of structural alterations could be. Subsequent animal model experiments can then attempt to visualize the structural alterations predicted by the computation models. In an example of this type of approach, Morgan and Soltesz (2006) have studied probable connectivity patterns within the recurrent GC network that results from mossy fiber sprouting. Their results indicate that the presence of a small number (at most 5%) of GCs with increased connectivity

(on average, 5–6 times the average number of connections) could serve as hubs and strongly promote hyperexcitability within the sclerotic dentate gyrus network. Interestingly, this finding is supported by the presence, in epileptic rats, of a small percentage of GCs that have basal dendrites (Spigelman et al., 1998) and which receive a vast number of additional mossy fiber contacts (Buckmaster and Thind, 2005). Computational modeling studies provide the best way of isolating structural changes from intrinsic cellular and synaptic changes that occur simultaneously in virtually all pathological states. Large-scale simulations will also make it possible to hold structure constant and test the circuit effects of specific therapeutic interventions, thereby allowing fast, cheap, efficient, and specific analysis of treatments for a variety of pathological processes. The fast pace at which scientists are gleaning knowledge about the dentate from animal models will allow for inclusion of a complete complement of interneuronal subtypes, various receptor and modulatory systems, realistic channel distributions, dendritic integration processes, and synaptic plasticity into computational models. These additions will greatly broaden the scope of questions that a model can address and the model’s predictive power. In the coming years, realistic modeling is only going to become more important in bridging the gaps between animals and humans, molecules and behavior.

Acknowledgements The authors would like to thank Dr. Jonas Dyhrfjeld-Johnsen for his contributions to the research discussed in this chapter and his helpful comments on the chapter. This work was funded by NIH grant NS35915 to IS and UCI MSTP to RM.

References Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Acsady, L., Katona, I., Martinez-Guijarro, F.J., Buzsaki, G. and Freund, T.F. (2000) Unusual target selectivity of

656 perisomatic inhibitory cells in the hilar region of the rat hippocampus. J. Neurosci., 20: 6907–6919. Amaral, D.G. (1978) A Golgi study of cell types in the hilar region of the hippocampus in the rat. J. Comp. Neurol., 182: 851–914. Aradi, I. and Holmes, W.R. (1999a) Role of multiple calcium and calcium-dependent conductances in regulation of hippocampal dentate granule cell excitability. J. Comput. Neurosci., 6: 215–235. Aradi, I. and Holmes, W.R. (1999b) Active dendrites regulate spatio-temporal synaptic integration in hippocampal dentate granule cells. Neurocomputing, 26–27: 45–51. Babb, T.L., Pretorius, J.K., Kupfer, W.R. and Brown, W.J. (1988) Distribution of glutamate-decarboxylase-immunoreactive neurons and synapses in the rat and monkey hippocampus: light and electron-microscopy. J. Comp. Neurol., 278: 121–138. Bartos, M., Vida, I., Frotscher, M., Geiger, J.R.P. and Jonas, P. (2001) Rapid signaling at inhibitory synapses in a dentate gyrus interneuron network. J. Neurosci., 21: 2687–2698. Blasco-Ibanez, J.M., Martinez-Guijarro, F.J. and Freund, T.F. (2000) Recurrent mossy fibers preferentially innervate parvalbumin-immunoreactive interneurons in the granule cell layer of the rat dentate gyrus. Neuroreport, 11: 3219–3225. Blumcke, I., Suter, B., Behle, K., Kuhn, R., Schramm, J., Elger, C.E. and Wiestler, O.D. (2000) Loss of hilar mossy cells in Ammon’s horn sclerosis. Epilepsia 41 Suppl 6: S174–180. Boss, B.D., Peterson, G.M. and Cowan, W.M. (1985) On the number of neurons in the dentate gyrus of the rat. Brain Res., 338: 144–150. Buckmaster, P.S. and Dudek, F.E. (1997) Network properties of the dentate gyrus in epileptic rats with hilar neuron loss and granule cell axon reorganization. J. Neurophysiol., 77: 2685–2696. Buckmaster, P.S. and Dudek, F.E. (1999) In vivo intracellular analysis of granule cell axon reorganization in epileptic rats. J. Neurophysiol., 81: 712–721. Buckmaster, P.S. and Jongen-Relo, A.L. (1999) Highly specific neuron loss preserves lateral inhibitory circuits in the dentate gyrus of kainate-induced epileptic rats. J. Neurosci., 19: 9519–9529. Buckmaster, P.S., Strowbridge, B.W., Kunkel, D.D., Schmiege, D.L. and Schwartzkroin, P.A. (1992) Mossy cell axonal projections to the dentate gyrus molecular layer in the rat hippocampal slice. Hippocampus, 2: 349–362. Buckmaster, P.S., Strowbridge, B.W. and Schwartzkroin, P.A. (1993) A comparison of rat hippocampal mossy cells and CA3c pyramidal cells. J. Neurophysiol., 70: 1281–1299. Buckmaster, P.S. and Schwartzkroin, P.A. (1994) Hippocampal mossy cell function: a speculative view. Hippocampus, 4: 393–402. Buckmaster, P.S. and Thind, K. (2005) American epilepsy society meeting abstract. Epilepsia, 46: 91–131. Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwartzkroin, P.A. (1996) Axon arbors and synaptic connections of hippocampal mossy cells in the rat in vivo. J. Comp. Neurol., 366: 270–292.

Buckmaster, P.S., Yamawaki, R. and Zhang, G.F. (2002a) Axon arbors and synaptic connections of a vulnerable population of interneurons in the dentate gyrus in vivo. J. Comp. Neurol., 445: 360–373. Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002b) Axon sprouting in a model of temporal lobe epilepsy creates a predominantly excitatory feedback circuit. J. Neurosci., 22: 6650–6658. Carnevale, N.T. and Hines, M.L. (2006) The Neuron Book. Cambridge University Press, Cambridge, UK. Cavazos, J.E., Das, I. and Sutula, T.P. (1994) Neuronal loss induced in limbic pathways by kindling: evidence for induction of hippocampal sclerosis by repeated brief seizures. J. Neurosci., 14: 3106–3121. Cavazos, J.E. and Sutula, T.P. (1990) Progressive neuronal loss induced by kindling: a possible mechanism for mossy fiber synaptic reorganization and hippocampal sclerosis. Brain Res., 527: 1–6. Ceranik, K., Bender, R., Geiger, J.R.P., Monyer, H., Jonas, P., Frotscher, M. and Lubke, J. (1997) A novel type of GABAergic interneuron connecting the input and the output regions of the hippocampus. J. Neurosci., 17: 5380–5394. Claiborne, B.J., Amaral, D.G. and Cowan, W.M. (1990) Quantitative, three-dimensional analysis of granule cell dendrites in the rat dentate gyrus. J. Comp. Neurol., 302: 206–219. Cossart, R., Dinocourt, C., Hirsch, J.C., Merchan-Perez, A., De Felipe, J., Ben-Ari, Y., Esclapez, M. and Bernard, C. (2001) Dendritic but not somatic GABAergic inhibition is decreased in experimental epilepsy. Nat. Neurosci., 4: 52–62. Desmond, N.L. and Levy, W.B. (1985) Granule cell dendritic spine density in the rat hippocampus varies with spine shape and location. Neurosci. Lett., 54: 219–224. Destexhe, A., Rudolph, M., Fellous, J.M. and Sejnowski, T.J. (2001) Fluctuating synaptic conductances recreate in vivolike activity in neocortical neurons. Neuroscience, 107: 13–24. Dyhrfield-Johnsen, J., Santhakumar, V., Morgan, R.J., Huerta, R., Tsimring, L. and Soltesz, I. (2007) Topological determinants of epileptogenesis in large-scale structural and functional models of the dentate gyrus derived from experimental data. J. Neurophysiol., 97(2): 1566–1587. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Frotscher, M., Seress, L., Schwerdtfeger, W.K. and Buhl, E. (1991) The mossy cells of the fascia dentata: a comparative study of their fine structure and synaptic connections in rodents and primates. J. Comp. Neurol., 312: 145–163. Gaarskjaer, F.B. (1978) Organization of mossy fiber system of rat studied in extended hippocampi.1. Terminal area related to number of granule and pyramidal cells. J. Comp. Neurol., 178: 49–71. Gabriel, S., Njunting, M., Pomper, J.K., Merschhemke, M., Sanabria, E.R.G., Eilers, A., Kivi, A., Zeller, M., Meencke, H.-J., Cavalheiro, E.A., Heinemann, U. and Lehmann, T.-N. (2004) Stimulus and potassium-induced epileptiform activity in the human dentate gyrus from patients with and without hippocampal sclerosis. J. Neurosci., 24: 10416–10430.

657 Gorter, J.A., van Vliet, E.A., Aronica, E. and Lopes da Silva, F.H. (2001) Progression of spontaneous seizures after status epilepticus is associated with mossy fibre sprouting and extensive bilateral loss of hilar parvalbumin and somatostatin-immunoreactive neurons. Eur. J. Neurosci., 13: 657–669. Geiger, J.R.P., Lubke, J., Roth, A., Frotscher, M. and Jonas, P. (1997) Submillisecond AMPA receptor-mediated signaling at a principal neuron-interneuron synapse. Neuron, 18: 1009–1023. Gulyas, A.I., Hajos, N. and Freund, T.F. (1996) Interneurons containing calretinin are specialized to control other interneurons in the rat hippocampus. J. Neurosci., 16: 3397–3411. Gulyas, A.I., Miettinen, R., Jacobowitz, D.M. and Freund, T.F. (1992) Calretinin is present in nonpyramidal cells of the rat hippocampus .1. A new type of neuron specifically associated with the mossy fiber system. Neuroscience, 48: 1–27. Halasy, K. and Somogyi, P. (1993) Subdivisions in the multiple gabaergic innervation of granule cells in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 411–429. Han, Z.S., Buhl, E.H., Lorinczi, Z. and Somogyi, P. (1993) A high-degree of spatial selectivity in the axonal and dendritic domains of physiologically identified local-circuit neurons in the dentate gyrus of the rat hippocampus. Eur. J. Neurosci., 5: 395–410. Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and Zhang, C.L. (1992) The dentate gyrus as a regulated gate for the propagation of epileptiform activity. Epilepsy Res. Suppl., 7: 273–280. Howard, A., Tamas, G. and Soltesz, I. (2005) Lighting the chandelier: new vistas for axo-axonic cells. Trends Neurosci., 28: 310–316. Howard, A.L., Neu, A., Morgan, R.J., Echegoyen, J.C. and Soltesz, I. (2007) Opposing Modifications in Intrinsic Currents and Synaptic Inputs in Post-Traumatic Mossy Cells: Evidence for Single-Cell Homeostasis in a Hyperexcitable Network. J. Neurophysiol., 97(3): 2394–2409. Ishizuka, N., Cowan, W.M. and Amaral, D.G. (1995) A quantitative analysis of the dendritic organization of pyramidal cells in the rat hippocampus. J. Comp. Neurol., 362: 17–45. Katona, I., Acsady, L. and Freund, T.F. (1999) Postsynaptic targets of somatostatin-immunoreactive interneurons in the rat hippocampus. Neuroscience, 88: 37–55. Kneisler, T.B. and Dingledine, R. (1995) Synaptic input from CA3 pyramidal cells to dentate basket cells in rat hippocampus. J. Physiol., 487(Pt 1): 125–146. Leite, J.P., Babb, T.L., Pretorius, J.K., Kuhlman, P.A., Yeoman, K.M. and Mathern, G.W. (1996) Neuron loss, mossy fiber sprouting, and interictal spikes after intrahippocampal kainate in developing rats. Epilepsy Res., 26: 219–231. Li, X.G., Somogyi, P., Tepper, J.M. and Buzsaki, G. (1992) Axonal and dendritic arborization of an intracellularly labeled chandelier cell in the ca1 region of rat hippocampus. Exp. Brain Res., 90: 519–525. Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The dentate gyrus as a control point for seizures in the hippocampus and beyond. Epilepsy Res. Suppl., 7: 301–313.

Lubke, J., Frotscher, M. and Spruston, N. (1998) Specialized electrophysiological properties of anatomically identified neurons in the hilar region of the rat fascia dentata. J. Neurophysiol., 79: 1518–1534. Lytton, W.W., Hellman, K.M. and Sutula, T.P. (1998) Computer models of hippocampal circuit changes of the kindling model of epilepsy. Artif. Intell. Med., 13: 81–97. Mathern, G.W., Bertram III, E.H., Babb, T.L., Pretorius, J.K., Kuhlman, P.A., Spradlin, S. and Mendoza, D. (1997) In contrast to kindled seizures, the frequency of spontaneous epilepsy in the limbic status model correlates with greater aberrant fascia dentata excitatory and inhibitory axon sprouting, and increased staining for N-methyl-D-aspartate, AMPA and GABA(A) receptors. Neuroscience, 77: 1003–1019. Milo, R., Shen-Orr, S., Itzkovitz, S., Kashtan, N., Chklovskii, D. and Alon, U. (2002) Network motifs: simple building blocks of complex networks. Science, 298: 824–827. Morgan, R.J. and Soltesz, I. (2006) American epilepsy society meeting abstract. Epilepsia, 47: 18–28. Netoff, T.I., Clewley, R., Arno, S., Keck, T. and White, J.A. (2004) Epilepsy in small-world networks. J. Neurosci., 24: 8075–8083. Nomura, T., Fukuda, T., Aika, Y., Heizmann, C.W., Emson, P.C., Kobayashi, T. and Kosaka, T. (1997a) Laminar distribution of non-principal neurons in the rat hippocampus, with special reference to their compositional difference among layers. Brain Res., 764: 197–204. Nomura, T., Fukuda, T., Aika, Y., Heizmann, C.W., Emson, P.C., Kobayashi, T. and Kosaka, T. (1997b) Distribution of nonprincipal neurons in the rat hippocampus, with special reference to their dorsoventral difference. Brain Res., 751: 64–80. Patton, P.E. and Mcnaughton, B. (1995) Connection matrix of the hippocampal-formation .1. The dentate gyrus. Hippocampus, 5: 245–286. Rafiq, A., Zhang, Y.F., DeLorenzo, R.J. and Coulter, D.A. (1995) Long-duration self-sustained epileptiform activity in the hippocampal-parahippocampal slice: a model of status epilepticus. J. Neurophysiol., 74: 2028–2042. Rall, W., Burke, R.E., Holmes, W.R., Jack, J.J., Redman, S.J. and Segev, I. (1992) Matching dendritic neuron models to experimental data. Physiol. Rev., 72: S159–S186. Ratzliff, A.H., Howard, A.L., Santhakumar, V., Osapay, I. and Soltesz, I. (2004) Rapid deletion of mossy cells does not result in a hyperexcitable dentate gyrus: implications for epileptogenesis. J. Neurosci., 24: 2259–2269. Reigl, M., Alon, U. and Chklovskii, D.B. (2004) Search for computational modules in the C. elegans brain. BMC Biol., 2: 25. Ribak, C.E., Seress, L. and Amaral, D.G. (1985) The development, ultrastructure and synaptic connections of the mossy cells of the dentate gyrus. J. Neurocytol., 14: 835–857. Roxin, A., Riecke, H. and Solla, S.A. (2004) Self-sustained activity in a small-world network of excitable neurons. Phys. Rev. Lett., 92: 198101. Santhakumar, V., Aradi, I. and Soltesz, I. (2005) Role of mossy fiber sprouting and mossy cell loss in hyperexcitability: a

658 network model of the dentate gyrus incorporating cell types and axonal topography. J. Neurophysiol., 93: 437–453. Santhakumar, V., Bender, R., Frotscher, M., Ross, S.T., Hollrigel, G.S., Toth, Z. and Soltesz, I. (2000) Granule cell hyperexcitability in the early post-traumatic rat dentate gyrus: the ‘irritable mossy cell’ hypothesis. J. Physiol., 524(Pt 1): 117–134. Santhakumar, V., Ratzliff, A.D., Jeng, J., Toth, Z. and Soltesz, I. (2001) Long-term hyperexcitability in the hippocampus after experimental head trauma. Ann. Neurol., 50: 708–717. Santhakumar, V. and Soltesz, I. (2004) Plasticity of interneuronal species diversity and parameter variance in neurological diseases. Trends Neurosci., 27: 504–510. Scharfman, H.E. (1991) Dentate hilar cells with dendrites in the molecular layer have lower thresholds for synaptic activation by perforant path than granule cells. J. Neurosci., 11: 1660–1673. Sik, A., Penttonen, M. and Buzsaki, G. (1997) Interneurons in the hippocampal dentate gyrus: an in vivo intracellular study. Eur. J. Neurosci., 9: 573–588. Song, S., Sjostrom, P.J., Reigl, M., Nelson, S. and Chklovskii, D.B. (2005) Highly nonrandom features of synaptic connectivity in local cortical circuits. PLoS Biol., 3: e68. Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic chandelier cells in the rat fascia dentata: Golgi-electron microscopy and immunocytochemical studies. J. Comp. Neurol., 293: 1–25. Spigelman, I., Yan, X.X., Obenaus, A., Lee, E.Y., Wasterlain, C.G. and Ribak, C.E. (1998) Dentate granule cells form novel basal dendrites in a rat model of temporal lobe epilepsy. Neuroscience, 86: 109–120. Sporns, O. and Kotter, R. (2004) Motifs in brain networks. PLoS Biol., 2: e369.

Vida, I., Bartos, M. and Jonas, P. (2006) Shunting inhibition improves robustness of gamma oscillations in hippocampal interneuron networks by homogenizing firing rates. Neuron, 49: 107–117. van Vliet, E.A., Aronica, E., Tolner, E.A., Lopes da Silva, F.H. and Gorter, J.A. (2004) Progression of temporal lobe epilepsy in the rat is associated with immunocytochemical changes in inhibitory interneurons in specific regions of the hippocampal formation. Exp. Neurol., 187: 367–379. Watts, D.J. and Strogatz, S.H. (1998) Collective dynamics of ‘small-world’ networks. Nature, 393: 440–442. Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E. and Schwartzkroin, P.A. (1997) Ultrastructural localization of neurotransmitter immunoreactivity in mossy cell axons and their synaptic targets in the rat dentate gyrus. Hippocampus, 7: 559–570. West, M.J. (1990) Stereological studies of the hippocampus: a comparison of the hippocampal subdivisions of diverse species including hedgehogs, laboratory rodents, wild mice and men. Prog. Brain Res., 83: 13–36. Woodson, W., Nitecka, L. and Benari, Y. (1989) Organization of the gabaergic system in the rat hippocampal-formation: a quantitative immunocytochemical study. J. Comp. Neurol., 280: 254–271. Yoshimura, Y. and Callaway, E.M. (2005) Fine-scale specificity of cortical networks depends on inhibitory cell type and connectivity. Nat. Neurosci., 8: 1552–1559. Yoshimura, Y., Dantzker, J.L. and Callaway, E.M. (2005) Excitatory cortical neurons form fine-scale functional networks. Nature, 433: 868–873. Zappone, C.A. and Sloviter, R.S. (2004) Translamellar disinhibition in the rat hippocampal dentate gyrus after seizureinduced degeneration of vulnerable hilar neurons. J. Neurosci., 24: 853–864.

SECTION VI

The Dentate Gyrus in Aging and Disease

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 36

Hippocampal granule cells in normal aging: insights from electrophysiological and functional imaging experiments Monica K. Chawla1,2 and Carol A. Barnes1,2,3, 1

Arizona Research Laboratories Division of Neural Systems, Memory and Aging, University of Arizona, Life Sciences North, Room 384, P.O. Box 245115, Tucson, AZ 85724, USA 2 Evelyn F. McKnight Brain Institute, University of Arizona, Tucson, AZ 85724, USA 3 Departments of Psychology and Neurology, University of Arizona, Tucson, AZ 85724, USA

Abstract: Normal aging, in the absence of neurodegenerative disease, can provide important insights into the mechanisms by which the brain can maintain cognitive abilities across the lifespan. Hippocampaldependent memory processes can become vulnerable as age advances. The focus of this chapter is the contribution of hippocampal granule cells to cognitive impairments that are observed during aging. A number of alterations in structure, function, and gene expression have been observed in aged granule cells, any of which may lead to adaptive, compensatory or detrimental consequences to hippocampal function. As the average life span of humans continues to increase, those who reach 100 years or beyond is more common. Individuals that have aged successfully, and exhibit high levels of cognitive ability can provide useful clues into the enormous potential possessed by the mammalian brain. Keywords: fascia dentata; dentate gyrus; LTP; fluorescence in situ hybridization; FMRI

appear to be involved in cell maintenance include cytokine molecules, expression of antiapoptotic proteins, and generation of various survival-related proteins like chaperones and antioxidants. As human populations continue to shift towards older average ages, it is becoming increasingly important to understand the factors that can ‘‘tip the balance’’ so that more individuals age successfully and fewer develop pathological conditions associated with aging, including Alzheimer’s and Parkinson’s diseases. Normal aging in the absence of disease can provide insight into the stable, plastic, and adaptive strategies adopted by the brain in these high functioning aged individuals. Indeed, an important goal for research in the neurobiology of aging is to identify

Introduction Aging is accompanied by numerous molecular, cellular, and structural changes that have functional consequences. Successful aging depends upon multiple factors including genetic background, and positive or negative environmental or social conditions. The observation that the brain is able to adapt to a variety of insults, suggests a large degree of heterogeneity in the expression of processes that can be employed by neurons or glia to optimize function. Among the physiological factors that Corresponding author. Tel.: +1 520-626-2616; Fax: +1 520-626-2618; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63036-2

661

662

mechanisms by which some aged individuals show minimal decrement in cognitive, motor and sensory skills across their lifespan. There is now a general consensus that some aspects of learning and memory functions decline with advancing age (for reviews see, Landfield, 1988; Rosenzweig and Barnes, 2003; Driscoll and Sutherland, 2005; Burke and Barnes, 2006). Memory processes that rely on hippocampal integrity are clearly vulnerable to advancing age in humans (e.g., Cabeza et al., 2005), monkeys (e.g., Rapp and Amaral, 1991; Keuker et al., 2000) and in rodent models of aging (e.g., Barnes, 1979; Geinisman et al., 1986; Barnes, 1988, 1994; Kadar et al., 1994; Gallagher and Rapp, 1997; Bach et al., 1999; Rapp et al., 1999; Driscoll et al., 2006). This chapter focuses on the contribution of hippocampal granule cells to cognitive impairments of aged animals. First, the basic anatomical, biochemical and biophysical properties of young granule cells will be compared with older granule cells. Whether these modifications in aging are adaptive, compensatory or detrimental to the hippocampus will be discussed. Next, functional connectivity, inferred by extracellular electrophysiological recording methods, is examined in an attempt to relate functional synaptic changes to cognitive decline in aged rats. Additionally, since aging is a multidimensional process that can affect many complex cellular mechanisms, transcriptional regulation of immediate early genomic responses are also examined here. The triggers and consequences of altered gene responses are poorly understood at present. Recent advances in imaging technologies have enabled both the identification of regions of brain that are activated by specific behavioral experiences in monkeys and humans (Small et al., 2002, 2004) as well as specific cells that are actively engaged in a particular behavioral experience (Guzowski et al., 1999). The examples that follow will illustrate the complex pattern of changes that occur during aging with a focus on granule cells. These patterns are not only subtle but highly sub-region specific and point towards compensatory mechanisms engaged by the brain to adapt to the complexities of the multidimensional, enigmatic process called normal aging.

Granule cell numbers in aging Early ideas suggested that cognitive decline observed during aging is due to global loss of neurons in rats, monkeys, and humans is now known to be incorrect. Brody (1955) first reported that the reduction in brain weight observed in older individuals was due to massive neuronal loss across the cortex. While these observations were later replicated in rats and humans in other laboratories (Ball, 1977; Coleman and Flood, 1987; Coleman et al., 1987), these early studies used cell counting methods without stereological techniques, and were found to be incorrect in later experiments that used unbiased cell counting methods (Sterio, 1984; West, 1993a; West et al., 1994; Morrison and Hof, 1997). It is now generally accepted that age-related cognitive decline is not due to massive neuronal loss. Using very detailed electron microscopic methods and serial reconstructions Geinisman and Bondareff (1976) first reported that granule cell numbers are not decreased in aged rat, but rather, they lose about a third of their medial entorhinal synaptic input. A number of more recent investigations have replicated these results using modern counting methods and have found no age-related neuronal loss of CA pyramidal cells or dentate gyrus granule cells (Rapp and Gallagher, 1996; Rasmussen et al., 1996) in rats. Similarly, in aged monkeys (Peters et al., 1996), and in humans (West, 1993b), there appears to be no hippocampal granule or CA pyramidal cell loss in aging. Moreover, entorhinal cortical layers II and III cell numbers do not change during aging in rats or monkeys (Merrill et al., 2000, 2001; Rapp et al., 2002).

Granule cell neurogenesis and aging Granule cells are unusual, in that they are among the few neuronal cells that undergo neurogenesis in adulthood (Altman and Das, 1965; Gould and Cameron, 1996; Eriksson, 2003). Neurogenesis has also been observed in olfactory bulb neurons across the lifespan of mammals. The newborn granule cells originate from the subgranular zone of the dentate gyrus and migrate into the granular layer (Kuhn et al., 1996), develop dendritic

663

processes (Ribak et al., 2004) and mature into functional neurons displaying some of the same synaptic responsiveness and other electrophysiological properties as do existing granule cells (van Praag et al., 2002; Song et al., 2005). Continous neurogenesis in the dentate gyrus throughout the life span of mammals may be of great functional significance. For example, Shors et al. (2001) have shown that numbers of newborn granule cells correlate with effective hippocampal-dependent memory function. Whether, newly born granule cells are integrated into functional networks mediating spatial information processing was recently investigated by Ramirez-Amaya et al. (2006). Using triple immunohistochemistry and confocal microscopy, they were able to show that granule cells born in the mature dentate gyrus (5-month old), newborn granule cells demonstrated by the mitotic marker-BrdU express the plasticity-related immediate-early gene Arc protein in response to spatial exploration. Additionally, the proportion of Arc protein-containing cells was significantly higher in the newly born granule cell population compared to the preexisting population of granule cells (2.8% vs. 1.6%, respectively). Recently it was shown that exercise-induced increases in neurogenesis are associated with better performance on the Morris water maze (van Praag et al., 2005), and a correlation between higher levels of neurogenesis and preserved spatial memory in old rats has been observed by Drapeau et al. (2003). Other studies, however, have not found correlations between neurogenesis and behavior (Bizon and Gallagher, 2003), or in some instances a negative correlation (Bizon et al., 2004) in the aged rodent. An interesting hypothesis recently proposed for the role of adult neurogenesis, suggests that the newborn granule cells participate in forming time-tagged associations for laying down new memories (Aimone et al., 2006). Aged rats do show granule cell neurogenesis and old granule cell survival rates can be modified by enriched environments (Kempermann and Gage, 1999; Kempermann et al., 2002). Generation of newly born granule cells from progenitor cells, however, show a marked decline beginning in middle age (Kuhn et al., 1996; Rao et al., 2005), but the numbers of newly born granule cells that become

neurons remain stable from middle age into older ages (Rao et al., 2005). Moreover, the increase in neurogenesis that is induced by specific types of learning (contextual fear conditioning) is reduced in aged rats, as well as the survival of new granule cells (Wati et al., 2006). The reduction in integration of newborn granule cells into hippocampal circuitry with age may contribute to the cognitive impairment in aged animals.

Age-related changes in granule cell synaptic contacts The constancy in granule cell numbers (Geinisman and Bondareff, 1976) and the volume of the granule cell layer (Coleman and Flood, 1987) with increasing age in rodents stands in contrast to reports of a decrease in synapse numbers. As was the case with cell counts in early literature, synapse counts conducted across age groups produced variable results, possibly due to methods that were less accurate than current techniques (Geinisman and Bondareff, 1976; Geinisman et al., 1977; Geinisman, 1979; Hoff et al., 1982; McWilliams and Lynch, 1984; Bertoni-Freddari et al., 1986; Badiali de Giorgi et al., 1987). Using unbiased stereological techniques, Geinisman et al. (1992) demonstrated a loss of axospinous synapse numbers in the middle third of the molecular layer of the dentate gyrus in memory-impaired old rats. When the entire hippocampus was homogenized in preparation for protein analysis to examine possible age-related differences in levels of synapserelated proteins (synaptogamin-, synaptophysin-, or synaptosomal-associated protein 25), no age effects were observed (Nicolle et al., 1999). However, with the use of methods that can detect circuit-specific differences in synaptic protein levels, age-related differences are revealed. Smith et al. (2000) examined the layer II entorhinal cortical projection that inervates both CA3 stratum lacunosum moleculare and dentate gyrus molecular layer. There was a significant loss of synaptophysin in both areas in memory-impaired old rats. Studies measuring synaptic markers, however, cannot elucidate the possible functional changes that result from alterations in these proteins, such as synaptic strength.

664

The answer to these kinds of functional questions, require electrophysiological investigation (see Synaptic Response section below). Granule cell dendrites in aging As discussed above, since numbers of granule cells do not change with age, it follows that the numbers of primary dendritic shafts should also remain unchanged. One study, however, did report a reduction in numbers of granule cell shafts in old rats (Geinisman et al., 1978). Because nonstereological methods were used in that study, the results may most easily be explained by sampling bias. An investigation in young and old rhesus monkeys, using stereological techniques, revealed no significant difference in total dendritic length, number of dendritic branch points and total segment number between young and old granule cells (Luebke and Rosene, 2003). They did, however, find that the molecular layer was decreased in width (between the granule cell layer of the upper blade and the hippocampal fissure) in old (>24 years) vs. young (o11 years) monkeys. Can spatial memory deficits with advancing age be explained in part by a decrease in spine numbers? von Bohlen und Halbach et al., 2006 recently reported no significant reduction in spine density in the apical dendrites of the upper blade DG and area CA1 of old mice, although the mean spine length did show a reduction in both areas in young (6–7 months) compared to old (21–22 months) mice. Interestingly, Samsonovich and Ascoli (2006) have recently concluded, using statistical analysis of digitally reconstructed neurons (from several different brain regions and different experimental manipulations) that there is morphological homeostasis in apical and basal dendritic processes of neurons. Therefore, dendritic changes in one area may be compensated by changes in other areas in order to preserve the dendritic tree, spine numbers, and other characteristics of granule cells. Biophysical properties of aged granule cells Does the absence of gross granule cell loss in aging mean a general preservation of function?

Electrophysiological comparisons of granule cells of young and old rats have been made using both in vivo and in vitro preparations. These studies reveal that most of the basic granule cell electrical properties remain constant over the lifespan of rodents, mice, and rabbits (Barnes et al., 1994). For example, studies investigating resting membrane potential, membrane time constant, the relation between applied current and input resistance, the amplitude and duration of the Na+-dependent action potential, threshold of action potential following depolarizing current injection, and rise time and half width of the intracellular EPSP (Barnes and McNaughton, 1979, 1980a; Barnes et al., 1987; Baskys et al., 1987; Niesen et al., 1988; Foster et al., 1991) have not revealed age-associated changes (for review, see Barnes, 1994; Rosenzweig and Barnes, 2003). Interestingly, the action potential threshold to orthodromic stimulation is decreased in aged granule cells (Barnes and McNaughton, 1980a). The observation that threshold to action potential discharge to direct current does not change during aging, while the action potential threshold to orthodromic stimulation does change, might be explained in several ways. Among these was the hypothesis that there is increased electrotonic coupling between granule cells with age (Barnes et al., 1987). If there were more gap junctional connections between granule cells, current would flow from the granule cell with the lowest threshold into adjacent cells that had not reached threshold. Overall, this would result in a lowering of the threshold for the entire population resulting from stimulation of the perforant path. Previously, it had been shown by MacVicar and Dudek (1982) that 60% of granule cells were electrotonically coupled, using Lucifer yellow injections and freeze fracture methods. In the Barnes et al. (1987), study a low molecular weight fluorescence dye, 5,6-carboxyfluorescein, was used to explore intercellular connections via gap junctions in in vitro slice preparations. Increased dye coupling was observed in all three hippocampal subregions with age. In fact, granule cells showed the most extensive coupling (i.e., 64% of old granule cells were coupled to at least one other cell compared to 49% of young granule cells). Accompanying this increased

665

Fig. 1. (A) Two representative granule cells filled with 5,6 carboxyfluorescein from the dentate gyrus of a 24-month-old rat. (B) Histograms showing the numbers of carboxyfluorescein injections that resulted in single, double, triple, or greater numbers of granule cells filled with dye. Aged rats showed significantly increased electrotonic coupling compared to young animals. This may account for the increased excitability of old granule cells. Adapted with modifications from Barnes et al. (1987). (See Color Plate 36.1 in color plate section.)

electrotonic coupling, old granule cells also exhibited an apparent increase in postsynaptic excitability, as assessed by the ratio of the population spike area to EPSP amplitude (26%). Figure 1A shows an example of two granule cells filled with 5,6-carboxyfluorescein taken from the dentate gyrus of a 24-month-old rat. The bar graph shown in Fig. 1B shows that older animals have significantly higher electrotonic coupling probabilities compared to young granule cells. The increase in electrotonic coupling and excitability may contribute to compensatory processes acting to keep the probability of discharge of individual granule cells constant in spite of a considerable loss (20–30%) of entorhinal cortical afferents to the hippocampal dentate gyrus.

Single unit recordings in granule cells Very few studies have investigated the spontaneous firing characteristics of granule cells in freely behaving animals. This is partially due to the fact that few granule cells fire during any given behavioral

condition in the dentate gyrus (Jung and McNaughton, 1993; Shen et al., 1998; Gothard et al., 2001). Additionally, granule cells were originally misclassified as having ‘‘theta-cell-like’’ firing characteristics (Deadwyler et al., 1976; Rose et al., 1983). It is now known, however, that granule cells have low firing rates and selective place fields in environments, similar to hippocampal pyramidal cell place fields. For example, Jung and McNaughton (1993) recorded single unit activity from granule cells while rats performed a spatial task on an eight-arm radial maze. It was found that granule cells showed both spatial and direction specificity in their firing properties, although their place fields were smaller and more ‘‘discontigous’’ than pyramidal cell place fields. Additionally, single granule cells have multiple place fields that were stable over at least several days. It was also observed that fewer granule cells, compared with hippocampal pyramidal cells, showed place-specific firing within any given environment (Barnes et al., 1990). These electrophysiological data are consistent with theoretical ideas suggesting that it is advantageous for hippocampal autoassociative memory

666

processing for information to be coded sparsely (Marr, 1971; McNaughton and Morris, 1987; Rolls, 1990). If the dentate gyrus performs the function of a pattern separator, the formation of distinct representations would be facilitated by amplification of small differences in the input to the granule cell network. Although few granule cells may fire in any given situation, consistent with a pattern separation mechanism, granule cell firing patterns undergo substantial modification after only minimal changes in an environment (Leutgeb et al., 2007). As discussed above, earlier studies mistakenly identified granule cells to possess firing characteristics similar to inhibitory interneurons. This may be explained in part, because of the ease with which higher firing rate cells are encountered along a recording tract trajectory. Mizumori et al. (1989) were able to identify characteristics that could differentiate between interneurons (basket cells) and granule cells. For example, interneurons have high spontaneous discharge rates and a relative absence of spike accommodation, while granule cells have lower firing rates. There have been no systematic age comparisons of the firing characteristics of granule cells; however, a radial maze spatial working memory task was used to assess the possible contribution of age-related changes in the firing characteristics of theta cells (putative interneurons) (Mizumori et al., 1992). The mean discharge rates of theta cells were significantly higher in the stratum granulosum in old rats compared to young animals. This region-specific change in the firing characteristics of theta cells may have important consequences for information processing with increasing age.

Perforant path granule cell synaptic responses in vivo and in vitro Depending on the location of termination of activated perforant path synaptic contacts onto the granule cell dendritic tree, the resulting field potential responses have distinctive rise time and shape characteristics. Because of the curved geometry of the dentate gyrus, field potentials recorded within the hilar region are nearly proportional to the

current transients at the granule cell bodies. Thus, the waveform of the population synaptic responses recorded below the granule cell layer, should reflect the passive electrical properties of the dendrites, and location of the active synapses as in an equivalent cylinder model. Because a square root relationship between synaptic distance and the rise time is empirically robust (Jefferys, 1975), the square root of the rise time of the extracellular synaptic response recorded from the hilus was calculated and used as an electrotonic location parameter for comparison across the lifespan of the rat (Barnes and McNaughton, 1979). By selectively stimulating small subsets of perforant path fibers along its medio-lateral axis, synapses at increasing distances from the soma were activated and the response waveform assessed (as illustrated in Fig. 2). When the mean electrotonic location parameter was compared across age groups, the rise time of the extracellular synaptic response was similar across the lifespan. When the medians of the frequency distributions (as shown in Fig. 3) of the electrotonic location parameter were compared in three age groups of rats, there was a small shift towards lower values with advancing age. This small shift in the distribution of extracellularly recorded synaptic response rise times may suggest atrophy of finer dendritic branches, which was, in fact an observation made by Geinisman et al. (1978) in the supragranular zone of old rats. Barnes (1979) first noted, using extracellular field potential recordings in young and old rats, that, for a given stimulus intensity delivered to the perforant path afferent fibers, old Long Evans rats showed reduced field EPSP amplitudes compared with young Long Evans rats. Using both in vivo and in vitro preparations to record extracellularly in young and old rats, Barnes and McNaughton (1980a) found that while there was no difference in the EPSP size at threshold stimulation intensities, at higher stimulus strengths, the EPSP was consistently smaller in old rats, over a wide range of intensities (see Fig. 4). It is also possible to record the compound action potentials of incoming perforant path fibers (i.e., the presynaptic fiber potential) to estimate the numbers of perforant path axon collaterals that terminate in the granule cell molecular layer.

667

Fig. 2. (A) Schematic diagram of a granule cell in the dentate gyrus showing the extracellular recording electrode placed slightly below the cell body layer (in the hilus). (B) Representative extracellularly recorded synaptic responses evoked from stimulation of different subpopulations of perforant path fibers are shown adjacent to their approximately corresponding input location on the granule cell dendrites. Note that the rise time characteristics of the synaptic responses vary as a function of the dendritic location of the activated synapses. Adapted with modifications from Barnes and McNaughton (1979).

When the presynaptic fiber potential was measured in young and old rats, again there was no change in the response amplitude at threshold stimulus intensity levels, however, above threshold, younger animals showed larger fiber potential response amplitudes. This finding was consistent with the reported decline in anatomical synaptic contacts, as shown by Geinisman (1979). A smaller presynaptic fiber potential implies that fewer axon collaterals are activated by the stimulus, and is therefore a possible explanation for the smaller extracellular field EPSP observed in old rats. When field EPSP amplitudes were plotted against fiber potential amplitudes, however, the synaptic response was larger for a given input size. Thus, smaller fiber potentials gave rise to larger synaptic responses in old rats, inspite of the fact that at stimulus intensities above threshold fiber potential amplitudes were smaller. One explanation for this result would be that individual synapses that survive in the old rat would be more effective in depolarizing the

granule cell, i.e., have stronger weights. The later observation is presumed to be due to axon collateral pruning in the perforant path projection from the entorhinal cortex. This hypothesis was tested in vitro, using minimal stimulation methods and quantal analysis techniques while recording intracellularly from granule cells of aged rats (Foster et al., 1991). Foster et al. tested rats of three different ages (5, 9, and 24 months) and found that the increased EPSP amplitude of granule cells in old rats was due to an increase in quantal size, which implicates an increase in postsynaptic sensitivity. Since the magnitude of paired-pulse facilitation was unchanged between age groups, the increase in synaptic strength could not be explained by an increase in the probability or number of transmitter quanta released. These post synaptic changes in synaptic strength have important functional implications for maintaining granule cell activity within a functional range and may suggest a compensatory mechanism employed by the aging dentate gyrus.

668

Fig. 3. (A) The frequency distributions of the electrotonic location parameter from adolescent (2 months), middle-aged (12 months), or old (24 months) rats as percentage of number of observations for each age group. (B) Bar graph shows the medians of the frequency distributions of the electrotonic parameter in adolescent, middle-aged, or old rats. Note the small shift towards lower values with advancing age. This small shift in the distribution of extracellularly recorded synaptic response rise times towards lower values with advancing age may suggest atrophy of the finer dendritic branches. Adapted and modified from Barnes and McNaughton (1979).

Cholinergic responses in vitro in aging granule cells When cholinergic afferent fibers are stimulated at 40 Hz, a slow cholinergic (atropine-sensitive) EPSP can be elicited in all three hippocampal subregions. When the response to this type of stimulus, delivered to stratum granulosum, was evaluated in animals from three different age groups (3, 9, or 24 months) there was a significant age-related decline (50%) in the amplitude of the slow cholinergic EPSP (Shen and Barnes, 1996). There was also a similar decine in cholinergic transmission in the CA3 and CA1 regions of old rats. Therefore, the neuromodulatory role of the cholinergic system in aging may alter

excitability of granule cells, which could significantly influence hippocampal network behavior (Huerta and Lisman, 1993; Hasselmo et al., 1996; Hasselmo, 1999; Barnes et al., 2000b; Ikonen et al., 2002).

Modification of synaptic weights at the perforant path-granule cell synapse The classic studies published in 1973 by Bliss and Gardner-Medwin (Bliss and Gardner-Medwin, 1973; Bliss and Lomo, 1973) demonstrate that synaptic projections from the entorhinal cortex could be modified, both in vivo (Bliss and

669

Fig. 4. The relation between field EPSP amplitude and perforant path stimulus intensity is shown in young (K) and old (J) rats for in vivo (A) and in vitro (D) preparations. The response waveforms (averaged across age groups) are shown for young (B) and old (C) animals from an in vivo experiment. Superimposed traces at various stimulus levels recorded simultaneously from the granule cell layer (E) and the molecular layer (F) are shown for an in vitro slice experiment. The dashed lines in B and C and E indicate the time of measurement of the field EPSP (2 ms after stimulus onset). The sharp negative deflections in B, C, and E are population spikes (asterisks) whereas the early negative deflection in F represents the presynaptic fiber response (indicated by arrow). Adapted from Barnes and McNaughton (1980a).

Gardner-Medwin, 1973) and in vitro (Bliss and Lomo, 1973) by patterned activation of perforant path afferents to hippocampal granule cells. This phenomenon, now referred to as long-term potentiation or LTP (Douglas and Goddard, 1975), required convergence of inputs onto granule cells or ‘‘cooperativity’’ (McNaughton et al., 1978), consistent with Hebb’s neural postulate for association (Hebb, 1949). It was subsequently found that this electrophysiological demonstration of coincidence detection worked through a novel glutamatergic receptor, the NMDA receptor (Collingridge, 1985). The NMDA receptor was found to be voltage-dependent and to allow influx of calcium, presumed at that time to be necessary for initiating durable forms of plasticity. In fact, Morris et al. (1986) showed that NMDA-receptor antagonism (via APV administration) was able to disrupt performance of the spatial, hippocampaldependent version of the Morris swim task, while leaving the nonspatial version intact.

The relationship between the durability of LTP and spatial memory was first reported in 1979 (Barnes, 1979), with the demonstration that spatial performance accuracy on the circular platform task was correlated with the durablility of LTP, as assessed over a number of days. This experiment was conducted in young and old Long Evans rats, trained on the circular platform spatial memory task, where the goal was to escape from a brightly illuminated open surface by finding the one hole (identified by the distal visual cues in the room) that would allow the rat to escape into a dark box. Following behavioral testing, 32 young and 32 old animals were implanted bilaterally with indwelling electrodes that could monitor field potentials in the dentate gyrus elicited by stimulation of the medial perforant path to granule cells. The results of this experiment were able to show, for the first time, that within each age group there was a significant correlation between the decay time constant of late phase LTP and spatial memory. The same

670

relationship between spatial behavior and LTP maintenance has been found in young and old mice at the Schaffer collateral — CA1 synapse in vitro (Bach et al., 1999). In these experiments a nonspatial version of the circular platform task was also administered, and the performance on the cued task did not correlate with LTP durability. To examine the LTP induction process over days, as well as its maintenance after induction levels were saturated, young and old rats were prepared for long-term field potential recordings from the dentate gyrus. After baseline evoked response stability to low frequency test pulses had been established, bursts of high frequency stimulation were delivered once per day for 7 consecutive days. Although the old rats reached asymptotic levels of LTP following this treatment more slowly than did the younger rats, they did achieve the same proportional levels of synaptic strengthening. Decay of LTP was then monitored for several weeks. Old rats showed faster decay over days than did young rats (time constants of 17 and 37 days, respectively); Barnes and McNaughton (1980a). As discussed above, if LTP is induced using robust stimulus parameters at the perforant pathgranule cell synapse, there the degree of LTP is equivalent between age groups (Barnes, 1979; Barnes and McNaughton, 1980a). One question is whether there is an altered threshold for induction of LTP at this synapse. Certainly there are fewer synaptic contacts in the termination zone of the medial perforant path synapses (Geinisman et al., 1992b) and thus, less synaptic convergence from perforant path stimulation would be achieved in older rats. To address the possibility that the threshold for LTP was altered in old rats, Barnes et al. (2000a) directly depolarized the granule cells with intracellular current injection paired with orthodromic stimulation, thereby eliminating the necessity for synaptic convergence, which is typically required to induce LTP (McNaughton et al., 1978). Using this paradigm more current was required to induce LTP in aged animals. One possible mechanism for this apparent deficit in LTP induction in the aged dentate gyrus could be alterations in NMDA receptors. The biochemical data reported for the

dentate gyrus are complex during aging (Magnusson et al., 2002, 2003), and it is not clear whether changes in receptor-related functions would result in the changes in plasticity (e.g., Wenk and Barnes, 2000). While, the examination of spatial memory and LTP in the hippocampus of old rats has provided important insights into age-related functional decline, it has also raised a number of questions. Among these is, what causes the faster decay in LTP in old animals? If LTP decay could be manipulated, memory function might be facilitated in older animals. The mechanisms involved in the maintenance of LTP are, unfortunately, less well understood than those involved in induction of LTP. As discussed in the following section, some insight may be gained through an examination of the role played by immediate early genes in durable forms of synaptic changes that may underlie memory changes in old mammals.

Immediate early genes The induction and translation of immediate early genes, lead to changes in multiple intracellular signaling cascades, and may play an important role in changes observed in age-related plasticity mechanisms. A number of IEGs are powerfully activated by stimulation parameters that lead to LTP induction. As discussed in the previous section, old animals, in addition to changes in LTP induction thresholds, also exhibit decreases in LTP maintenance that are not detectable until several days after robust induction protocols have been applied. Studies examining the basis of the maintenance phase of long-term plasticity have demonstrated the requirement for RNA and protein synthesis following the conditioning stimulus (Montarolo et al., 1986; Nguyen et al., 1994; Lanahan et al., 1997). Therefore, it is plausible that age-dependent memory decline could result from alterations in pathways subsequent to transcriptional responses mediated by IEGs. Cole et al. (1989) established a relationship between LTP induction and activation of several IEGs (including, zif268, c-fos, c-jun, jun-B, and Arc). In that study, when LTP at the perforant

671

path-granule cell synapse was induced by high frequency stimulation, levels of zif 268 mRNA increased significantly. This increase in zif 268 mRNA could be blocked both by NMDA-receptor antagonists and by convergent synaptic inhibitory inputs known to block LTP, thus establishing a correlation between IEG expression and LTP. Similar findings were reported in un-anesthetized rats by Dragunow et al. (1989). The early studies that linked IEG expression in granule cells with LTP, were however, called into question by suggestions that the thresholds for LTP induction and gene activation are different. To address this issue, Worley et al. (1993) used a chronic in vivo recording approach to monitor two different patterns of LTP-inducing stimuli to perforant path afferents. The RNA expression pattern and the protein products of several IEGs (zif 268, jun-B, c-jun, and c-fos) were monitored in young and old animals, following the two different electrical stimulation protocols (10 pulse train or 50 pulse train, delivered at 400 Hz). Although, different patterns of IEG expression emerged, only zif 268 was detectable at thresholds that induced LTP. Additionally, when the in situ autoradiograms were quantified using densitometry, there were no significant agerelated differences in zif 268 mRNA levels between groups. Using the more sensitive reverse Northern method, Lanahan et al. (1997), examined a panel of several genes (c-fos, jun B, c-jun, fos B, fra1, and zif 268) known to be induced by LTP. In that study, an LTP induction protocol (high frequency stimulation of the perforant fibers) known to result in LTP maintenance deficits in old rats (Barnes and McNaughton, 1980b) was used. The dentate gyrus was extracted from young (9 month) and old (26 month) animals and reverse Northern blots were quantified using a phosphoimager. Of all the genes examined (including zif 268) only c-fos RNA level differed between young and old rats. Interestingly, while both young and old rats showed elevated c-fos expression following LTP induction, the older rats showed greater levels of expression than did the young animals. This indicated that changes in c-fos, may contribute to altered signaling mechanisms that are involved in age-related hippocampal plasticity deficits observed in aging.

As the studies of specific IEGs in activity-dependent synaptic plasticity continued, an increase in a unique gene, activity-regulated cytoskeleton-associated gene (Arc, also known as Arg 3.1) was identified, following seizures, or LTP induced at the perforant path granule cell synapse (Link et al., 1995; Lyford et al., 1995). It seemed logical to investigate whether Arc had a role in the maintenance phase of LTP. Guzowski et al. (2000) used intrahippocampal infusions of antisense oligodeoxynucleotides to block the synthesis of Arc protein and monitored the effect on LTP and spatial memory. Blocking the synthesis of Arc protein in the dorsal hippocampus impaired the maintenance phase of LTP induced at the perforant path-granule cell synapse, and the long-term consolidation of memory for a spatial location (Morris water maze) without interfering with either the induction phase of LTP or water maze acquisition. Thus, a role for Arc in hippocampal-dependent plasticity was established. The IEG Arc proved to be a useful tool to clarify behavior-related synaptic plasticity as well, as it is rapidly induced when animals explore their environment. Guzowski et al. (1999) showed that Arc can be induced in hippocampal CA1 neurons following two episodes of spatial experience and that the proportions of cells expressing Arc matched predictions derived from electrophysiological recordings. In that investigation it was also shown that Arc RNA is transcribed within 1–2 min following behavior, and that it appears as discrete transcriptional foci in the nucleus (for 2–15 min). Within 15–30 min the RNA translocates to the cytoplasm, and can be subsequently found in the dendrites (45 min). This unique property of Arc formed the basis for the development of a powerful new technique called compartment analysis of temporal activity by fluorescence in situ hybridization or catFISH. With this technique, the behavior-induced activity history of neuronal ensembles could be inferred at two different time points, throughout the brain (e.g., Guzowski et al., 1999; Chawla et al., 2004). Although, many IEGs can be used in the catFISH technique, Arc provides distinct advantages in that it is dynamically regulated in many brain regions. When the pattern of spatial exploration-induced Arc expression was examined in hippocampal

672

granule cells of young rats, it was found to be predominantly located in the upper blade of the dentate gyrus (Chawla et al., 2005). No major anatomical differences in afferents to the upper vs. lower blade easily account for this interesting observation, since only minor differences in anatomical connectivity exist between these regions. The proportion of active cells following spatial experience is low for granule cells compared to CA1 pyramidal cells (4% vs. 40%), and is consistent with the sparse firing pattern observed in the dentate gyrus in electrophysiological studies (Barnes et al., 1990; Jung and McNaughton, 1993; Shen et al., 1997; Gothard et al., 2001). Figure 5 illustrates the expression of behaviorally induced Arc in young rats (Fig. 5A), which is significantly elevated compared to baseline levels of expression in young animals that have not performed the spatial foraging task (Fig. 5B). The difference in upper and lower blade Arc expression cannot be explained by the inability of the lower blade to express Arc, because as early as 5 min following maximal electroconvulsive shock (MECS) treatment, both the upper and lower blades show robust Arc expression. It is possible that granule cells in the lower blade are largely inactive for a variety of experiences.

Vulnerability of dentate gyrus in aging The information gathered over the past several decades from electrophysiological recording and anatomical studies, has been critical in furthering our understanding of the kinds of learning and memory deficits that should be expected in normal aging. Anatomical and physiological studies can be complimented by functional imaging approaches that can directly identify behavioral correlates of neuronal function. Several imaging techniques are now available that provide cellular and temporal resolution in this regard. Among them is the functional imaging method based on measurement of cerebral blood flow volumes using magnetic resonance imaging (MRI) that have been shown to correlate with regional energy metabolism (e.g., Gonzalez et al., 1995; Hyder et al., 2001). Recently, Small et al. (2004) measured CBV

Fig. 5. Image of granule cells (blue) following in situ hybridization using a c-RNA probe for the IEG Arc. (A) Arc positive cells (shown in red) in the granule cell layer from the upper blade of a young rat following two 5 min (separated by a 20 min rest period in the home cage) episodes of spatial exploratory behavior. (B) Granule cells from a caged control animal of the same age, that did not perform the exploratory behavior, are shown, illustrating baseline or negligible expression of Arc mRNA. (See Color Plate 36.5 in color plate section.)

values in 14 young and old rhesus monkeys (7–31 years) that also performed a memory assessment task (delayed nonmatching-to-sample, DNMS). Representative CBV meaures from a young and old monkey is shown in Fig. 6A. The CBV measures indicated a significant inverse correlation with age only in the dentate gyrus (Fig. 6B). In addition, when CBV measurements were plotted

673

Fig. 6. (A) Examples of gadolinium-induced change in MRI signal, (a measure of cerebral blood flow volume, CBV, that is a correlate of brain metabolism) from a young and old rhesus monkey. Warm colors indicate higher CBV values. The white circle, identified in precontrast images, overlies the dentate gyrus. (B) CBV measurements plotted vs. age in the dentate gyrus indicates a significant decline with advancing age. (C) CBV measurements in the dentate gyrus of old monkeys plotted against memory performance on the delayed nonmatching-to-sample test. A significant relationship between CBV and memory performance is observed, selectively in the old animals. Adapted with modifications from Small et al. (2004). (See Color Plate 36.6 in color plate section.)

against memory performance on a delayed nonmatching-to-sample task; again, only the dentate gyrus in old monkeys showed a significant correlation with CBV activity (Fig. 6C). As discussed previously, catFISH can provide both temporal and cellular resolution in determining the activity history of individual neurons. When cellular Arc RNA was determined using catFISH combined with high-resolution confocal microscopy in young, middle-aged, and old rats, only the dentate gyrus showed age-related changes. As shown in Fig. 7B, a significant age-related decline in ArcRNA-containing granule cells was observed in aged animals. Importantly, as was seen in monkeys with MRI imaging, the pyramidal subregion in rats was unaffected by advancing age, i.e., there was no change in Arc-expressing pyramidal cell numbers. Furthermore, using MRI studies in humans, Small et al. (2002) have also reported that

subiculum and dentate gyrus show selective changes in normal aging that could be correlated with memory decline in these individuals. Thus, a cross species consensus is now emerging in that the dentate gyrus in rats, monkeys, and humans is the hippocampal subregion that is particularly vulnerable to advancing age.

Conclusions and summary Aging is accompanied by selective memory changes that depend on proper hippocampal function. Investigation of hippocampal anatomy, physiology, and gene expression in normal aging provides a foundation for understanding age-related cognitive decline. A common theme that emerges from examination of the neurobiological changes in aging granule cells, is that cell loss cannot account for the cognitive decline, as granule

674

Fig. 7. Behavior-induced Arc mRNA expression pattern in young and old rat granule cells of the dentate gyrus. (A) Individual examples of behaviorally induced Arc RNA. Fluorescence-tagged CY3 Arc RNA containing cells are green, and granule cell nuclei are shown in red (counterstained with propydium iodide, a nucleic acid stain). (B) The average proportion of Arc-positive neurons (% of total cells) measured from the three different age groups (9, 15, and 24 months) is shown for the dentate gyrus. Note that the dentate gyrus of old rats contains a significantly lower proportion of behaviorally induced Arc-positive granule cells compared to young or middle-aged animals. (See Color Plate 36.7 in color plate section.)

cell numbers are preserved during aging. Although there is an absence of neuronal degeneration and a general preservation of basic biophysical properties, there is a significant loss of axodentritic synapses onto granule cells from entorhinal afferents. In addition to the synapse loss, there are a number of other changes that occur in aged animals. Among them, is an altered threshold for LTP induction, an increase in strength of the surviving synapses, and increased electrotonic coupling between the old granule cells. Together, these factors may result in an attenuation of the deleterious impact of aging on cognition. In spite of the compensatory changes mentioned above, granule cells appear to be particularly vulnerable to the effects of aging compared to hippocampal pyramidal cells. For example, there is a reduction in the numbers of granule cells that show behavior-induced gene expression in old rats. Furthermore, older monkeys that have difficulty performing a hippocampal-dependent learning task also have lower functional activity in the dentate gyrus as assessed by cerebral blood volume measures using MRI techniques. Additionally, MRI studies in humans also indicate that dentate gyrus is also a hippocampal subregion with selective changes in normal aging that could be correlated with memory decline in these individuals.

With the human population living well into their eighties or nineties, it is becoming increasingly important to find ways to preserve cognitive function, and maintain an optimal quality of life during the last decades. New imaging techniques, pharmacological interventions or genetic manipulations, can advance our knowledge regarding normal aging. A multidisciplinary effort will be essential for identifying early interventions to ameliorate declining cognition with advancing age.

Acknowledgements We thank Michelle Carroll for her excellent technical assistance in preparing the figures. The authors and some of this work were supported by grants from the National Institute of Health AG009219 and AG003376, the state of Arizona and ADHS and the McKnight Brain Research Foundation.

References Aimone, J.B., Wiles, J. and Gage, F.H. (2006) Potential role for adult neurogenesis in the encoding of time in new memories. Nat. Neurosci., 9: 723–727. Altman, J. and Das, G.D. (1965) Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol., 124: 319–335.

675 Bach, M.E., Barad, M., Son, H., Zhuo, M., Lu, Y.F., Shih, R., Mansuy, I., Hawkins, R.D. and Kandel, E.R. (1999) Agerelated defects in spatial memory are correlated with defects in the late phase of hippocampal long-term potentiation in vitro and are attenuated by drugs that enhance the cAMP signaling pathway. Proc. Natl. Acad. Sci. U.S.A., 96: 5280–5285. Badiali de Giorgi, L., Bonvicini, F., Bianchi, D., Bossoni, G. and Laschi, R. (1987) Ultrastructural aspects of ageing rat hippocampus and effects of L-acetyl-carnitine treatment. Drugs Exp. Clin. Res., 13: 185–189. Ball, M.J. (1977) Neuronal loss, neurofibrillary tangles and granulovacuolar degeneration in the hippocampus with ageing and dementia. A quantitative study. Acta Neuropathol. (Berl.), 37: 111–118. Barnes, C.A. (1979) Memory deficits associated with senescence: a neurophysiological and behavioral study in the rat. J. Comp. Physiol. Psychol., 93: 74–104. Barnes, C.A. (1988) Spatial learning and memory processes: the search for their neurobiological mechanisms in the rat. Trends Neurosci., 11: 163–169. Barnes, C.A. (1994) Normal aging: regionally specific changes in hippocampal synaptic transmission. Trends Neurosci., 17: 13–18. Barnes, C.A. and McNaughton, B.L. (1979) Neurophysiological comparison of dendritic cable properties in adolescent, middle-aged, and senescent rats. Exp. Aging Res., 5: 195–206. Barnes, C.A. and McNaughton, B.L. (1980a) Physiological compensation for loss of afferent synapses in rat hippocampal granule cells during senescence. J. Physiol., 309: 473–485. Barnes, C.A. and McNaughton, B.L. (1980b) Spatial memory and hippocampal synaptic plasticity in senescent and middleaged rats. In: Stein D. (Ed.), The Psychobiology of Aging: Problems and Perspectives. Elsevier Press, New York, pp. 253–272. Barnes, C.A., McNaughton, B.L., Mizumori, S.J., Leonard, B.W. and Lin, L.H. (1990) Comparison of spatial and temporal characteristics of neuronal activity in sequential stages of hippocampal processing. Prog. Brain Res., 83: 287–300. Barnes, C.A., Meltzer, J., Houston, F., Orr, G., McGann, K. and Wenk, G.L. (2000b) Chronic treatment of old rats with donepezil or galantamine: effects on memory, hippocampal plasticity and nicotinic receptors. Neuroscience, 99: 17–23. Barnes, C.A., Rao, G. and Houston, F.P. (2000a) LTP induction threshold change in old rats at the perforant path — granule cell synapse. Neurobiol. Aging, 21: 613–620. Barnes, C.A., Rao, G. and McNaughton, B.L. (1987) Increased electrotonic coupling in aged rat hippocampus: a possible mechanism for cellular excitability changes. J. Comp. Neurol., 259: 549–558. Barnes, C.A., Treves, A., Rao, G. and Shen, J. (1994) Electrophysiological markers of cognitive aging: region specificity and computational consequences. Semin. Neurosci., 6: 359–367. Baskys, A., Niesen, C.E. and Carlen, P.L. (1987) Altered modulatory actions of serotonin on dentate granule cells of aged rats. Brain Res., 419: 112–118. Bertoni-Freddari, C., Giuli, C., Pieri, C. and Paci, D. (1986) Quantitative investigation of the morphological plasticity of

synaptic junctions in rat dentate gyrus during aging. Brain Res., 366: 187–192. Bizon, J.L. and Gallagher, M. (2003) Production of new cells in the rat dentate gyrus over the lifespan: relation to cognitive decline. Eur. J. Neurosci., 18: 215–219. Bizon, J.L., Lee, H.J. and Gallagher, M. (2004) Neurogenesis in a rat model of age-related cognitive decline. Aging Cell, 3: 227–234. Bliss, T.V.P. and Gardner-Medwin, A.R. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the unanaesthetised rabbit following stimulation of the perforant path. J. Physiol., 232: 357–374. Bliss, T.V.P. and Lomo, T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anesthetized rabbit following stimulus of perforant path. J. Physiol., 232: 331–356. von Bohlen und Halbach, O., Zacher, C., Gass, P. and Unsicker, K. (2006) Age-related alterations in hippocampal spines and deficiencies in spatial memory in mice. J. Neurosci. Res., 83: 525–531. Brody, H. (1955) Organization of the cerebral cortex. III. A study of aging in the human cerebral cortex. J. Comp. Neurol., 102: 511–556. Burke, S.N. and Barnes, C.A. (2006) Neural plasticity in the ageing brain. Nat. Rev. Neurosci., 7: 30–40. Cabeza, R., Nyberg, L. and Park, D. (2005) Cognitive Neuroscience of Aging. Oxford University Press, New York. Chawla, M.K., Guzowski, J.F., Ramirez-Amaya, V., Lipa, P., Hoffman, K.L., Marriott, L.K., Worley, P.F., McNaughton, B.L. and Barnes, C.A. (2005) Sparse, environmentally selective expression of Arc RNA in the upper blade of the rodent fascia dentata by brief spatial experience. Hippocampus, 15: 579–586. Chawla, M.K., Lin, G., Olson, K., Vazdarjanova, A., Burke, S.N., McNaughton, B.L., Worley, P.F., Guzowski, J.F., Roysam, B. and Barnes, C.A. (2004) 3D-catFISH: a system for automated quantitative three-dimensional compartmental analysis of temporal gene transcription activity imaged by fluorescence in situ hybridization. J. Neurosci. Methods, 139: 13–24. Cole, A.J., Saffen, D.W., Baraban, J.M. and Worley, P.F. (1989) Rapid increase of an immediate early gene messenger RNA in hippocampal neurons by synaptic NMDA receptor activation. Nature, 340: 474–476. Coleman, P.D. and Flood, D.G. (1987) Neuron numbers and dendritic extent in normal aging and Alzheimer’s disease. Neurobiol. Aging, 8: 521–545. Coleman, P.D., Flood, D.G. and West, M.J. (1987) Volumes of the components of the hippocampus in the aging F344 rat. J. Comp. Neurol., 266: 300–306. Collingridge, G.L. (1985) Long term potentiation in the hippocampus: Mechanisms of initiation and modulation by neurotransmitters. Trends Pharmacol. Sci., 6: 407–411. Deadwyler, S.A., Gribkoff, V., Cotman, C. and Lynch, G. (1976) Long lasting changes in the spontaneous activity of hippocampal neurons following stimulation of the entorhinal cortex. Brain. Res. Bull., 1: 1–7.

676 Douglas, R.M. and Goddard, G.V. (1975) Long-term potentiation of the perforant path-granule cell synapse in the rat hippocampus. Brain Res., 86: 205–215. Dragunow, M., Abraham, W.C., Goulding, M., Mason, S.E., Robertson, H.A. and Faull, R.L. (1989) Long-term potentiation and the induction of c-fos mRNA and proteins in the dentate gyrus of unanesthetized rats. Neurosci. Lett., 101: 274–280. Drapeau, E., Mayo, W., Aurousseau, C., Le Moal, M., Piazza, P.V. and Abrous, D.N. (2003) Spatial memory performances of aged rats in the water maze predict levels of hippocampal neurogenesis. Proc. Natl. Acad. Sci. U.S.A., 100: 14385–14390. Driscoll, I., Howard, S.R., Stone, J.C., Monfils, M.H., Tomanek, B., Brooks, W.M. and Sutherland, R.J. (2006) The aging hippocampus: a multi-level analysis in the rat. Neuroscience, 139: 1173–1185. Driscoll, I. and Sutherland, R.J. (2005) The aging hippocampus: navigating between rat and human experiments. Rev. Neurosci., 16: 87–121. Eriksson, P.S. (2003) Neurogenesis and its implications for regeneration in the adult brain. J. Rehabil. Med., 41(Suppl.): 17–19. Foster, T.C., Barnes, C.A., Rao, G. and McNaughton, B.L. (1991) Increase in perforant path quantal size in aged F-344 rats. Neurobiol. Aging, 12: 441–448. Gallagher, M. and Rapp, P.R. (1997) The use of animal models to study the effects of aging on cognition. Annu. Rev. Psychol., 48: 339–370. Geinisman, Y. (1979) Loss of axosomatic synapses in the dentate gyrus of aged rats. Brain Res., 168: 485–492. Geinisman, Y. and Bondareff, W. (1976) Decrease in the number of synapses in the senescent brain: A quantitative electron microscopic analysis of the dentate gyrus molecular layer in the rat. Mech. Aging Dev., 5: 11–23. Geinisman, Y., Bondareff, W. and Dodge, J.T. (1977) Partial deafferntation of neurons in the dentate gyrus of the senescent rat. Brain Res., 134: 541–545. Geinisman, Y., Bondareff, W. and Dodge, J.T. (1978) Dendritic atrophy in the dentate gyrus of the senescent rat. Am. J. Anat., 152: 321–329. Geinisman, Y., de Toledo-Morrell, L. and Morrell, F. (1986) Loss of perforated synapses in the dentate gyrus: morphological substrate of memory deficit in aged rats. Proc. Natl. Acad. Sci. U.S.A., 83: 3027–3031. Geinisman, Y., de Toledo-Morrell, L., Morrell, F., Persina, I.S. and Rossi, M. (1992) Age-related loss of axospinous synapses formed by two afferent systems in the rat dentate gyrus as revealed by the unbiased stereological dissector technique. Hippocampus, 2: 437–444. Gonzalez, R.G., Fischman, A.J., Guimaraes, A.R., Carr, C.A., Stern, C.E., Halpern, E.F., Growdon, J.H. and Rosen, B.R. (1995) Functional MR in the evaluation of dementia: correlation of abnormal dynamic cerebral blood volume measurements with changes in cerebral metabolism on positron emission tomography with fludeoxyglucose F 18. AJNR Am. J. Neuroradiol., 16: 1763–1770.

Gothard, K.M., Hoffman, K.L., Battaglia, F.P. and McNaughton, B.L. (2001) Dentate gyrus and ca1 ensemble activity during spatial reference frame shifts in the presence and absence of visual input. J. Neurosci., 21: 7284–7292. Gould, E. and Cameron, H.A. (1996) Regulation of neuronal birth, migration and death in the rat dentate gyrus. Dev. Neurosci., 18: 22–35. Guzowski, J.F., Lyford, G.L., Stevenson, G.D., Houston, F.P., McGaugh, J.L., Worley, P.F. and Barnes, C.A. (2000) Inhibition of activity-dependent arc protein expression in the rat hippocampus impairs the maintenance of long-term potentiation and the consolidation of long-term memory. J. Neurosci., 20: 3993–4001. Guzowski, J.F., McNaughton, B.L., Barnes, C.A. and Worley, P.F. (1999) Environment-specific expression of the immediate-early gene Arc in hippocampal neuronal ensembles. Nat. Neurosci., 2: 1120–1124. Hasselmo, M.E. (1999) Neuromodulation: acetylcholine and memory consolidation. Trends Cogn. Sci., 3: 351–359. Hasselmo, M.E., Wyble, B.P. and Wallenstein, G.V. (1996) Encoding and retrieval of episodic memories: role of cholinergic and GABAergic modulation in the hippocampus. Hippocampus, 6: 693–708. Hebb, D.O. (1949) The Organization of Behavior. Wiley, New York. Hoff, S.F., Scheff, S.W., Bernardo, L.S. and Cotman, C.W. (1982) Lesion-induced synaptogenesis in the dentate gyrus of aged rats: I. Loss and reaquisition of normal synaptic density. J. Comp. Neurol., 205: 246–252. Huerta, P.T. and Lisman, J.E. (1993) Heightened synaptic plasticity of hippocampal CA1 neurons during a cholinergically induced rhythmic state. Nature, 364: 723–725. Hyder, F., Kida, I., Behar, K.L., Kennan, R.P., Maciejewski, P.K. and Rothman, D.L. (2001) Quantitative functional imaging of the brain: towards mapping neuronal activity by BOLD fMRI. NMR Biomed., 14: 413–431. Ikonen, S., McMahan, R., Gallagher, M., Eichenbaum, H. and Tanila, H. (2002) Cholinergic system regulation of spatial representation by the hippocampus. Hippocampus, 12: 386–397. Jefferys, J.G.R. (1975) Propagation of action potentials into the dendrite of hippocampal granule cells in vitro. J. Physiol., 249: 16–18. Jung, M.W. and McNaughton, B.L. (1993) Spatial selectivity of unit activity in the hippocampal granular layer. Hippocampus, 3:165–182. Kadar, T., Arbel, I., Silbermann, M. and Levy, A. (1994) Morphological hippocampal changes during normal aging and their relation to cognitive deterioration. J. Neural. Transm. Suppl., 44: 133–143. Kempermann, G. and Gage, F.H. (1999) Experience-dependent regulation of adult hippocampal neurogenesis: effects of long-term stimulation and stimulus withdrawal. Hippocampus, 9: 321–332. Kempermann, G., Gast, D. and Gage, F.H. (2002) Neuroplasticity in old age: sustained fivefold induction of hippocampal

677 neurogenesis by long-term environmental enrichment. Ann. Neurol., 52: 135–143. Keuker, J.I., Luiten, P.G. and Fuchs, E. (2000) Capillary changes in hippocampal CA1 and CA3 areas of the aging rhesus monkey. Acta Neuropathol. (Berl.), 100: 665–672. Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Neurogenesis in the dentate gyrus of the adult rat: age-related decrease of neuronal progenitor proliferation. J. Neurosci., 16: 2027–2033. Lanahan, A., Lyford, G., Stevenson, G.S., Worley, P.F. and Barnes, C.A. (1997) Selective alteration of long-term potentiation-induced transcriptional response in hippocampus of aged, memory-impaired rats. J. Neurosci., 17: 2876–2885. Landfield, P.W. (1988) Hippocampal neurobiological mechanisms of age-related memory dysfunction. Neurobiol. Aging, 9: 571–579. Leutgeb, J.K., Leutgeb, S., Moser, M.B. and Moser, E.I. (2007) Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science, 315: 961–966. Link, W., Konietzko, U., Kauselmann, G., Krug, M., Schwanke, B., Frey, U. and Kuhl, D. (1995) Somatodendritic expression of an immediate early gene is regulated by synaptic activity. Proc. Natl. Acad. Sci. U.S.A., 92: 5734–5738. Luebke, J.I. and Rosene, D.L. (2003) Aging alters dendritic morphology, input resistance, and inhibitory signaling in dentate granule cells of the rhesus monkey. J. Comp. Neurol., 460: 573–584. Lyford, G.L., Yamagata, K., Kaufmann, W.E., Barnes, C.A., Sanders, L.K., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., Lanahan, A.A. and Worley, P.F. (1995) Arc, a growth factor and activity-regulated gene, encodes a novel cytoskeletonassociated protein that is enriched in neuronal dendrites. Neuron, 14: 433–445. MacVicar, B.A. and Dudek, F.E. (1982) Electrotonic coupling between granule cells of rat dentate gyrus: Physiological and anatomical evidence. J. Neurophysiol., 47: 579–592. Magnusson, K.R., Nelson, S.E. and Young, A.B. (2002) Agerelated changes in the protein expression of subunits of the NMDA receptor. Brain Res. Mol. Brain Res., 99: 40–45. Magnusson, K.R., Scruggs, B., Aniya, J., Wright, K.C., Ontl, T., Xing, Y. and Bai, L. (2003) Age-related deficits in mice performing working memory tasks in a water maze. Behav. Neurosci., 117: 485–495. Marr, D. (1971) Simple memory: a theory for archicortex. Philos. Trans. R. Soc. B, 262: 23–81. McNaughton, B.L., Douglas, R.M. and Goddard, G.V. (1978) Synaptic enhancement in fascia dentata: cooperativity among coactive afferents. Brain Res., 157: 277–293. McNaughton, B.L. and Morris, R.G.M. (1987) Hippocampal synaptic enhancement and information storage within a distributed memory system. Trends Neurosci., 10: 408–415. McWilliams, J.R. and Lynch, G. (1984) Synaptic density and axonal aprouting in rat hippocampus: stability in adulthood abd decline in late adulthood. Brain Res., 294: 152–156. Merrill, D.A., Chiba, A.A. and Tuszynski, M.H. (2001) Conservation of neuronal number and size in entorhinal cortex of

behaviorally characterized aged rats. J. Comp. Neurol., 438: 445–456. Merrill, D.A., Roberts, J.A. and Tuszynski, M.H. (2000) Conservation of neuron number and size in entorhinal cortex layers, I.I., III, and V/VI of aged primates. J. Comp. Neurol., 422: 396–401. Mizumori, S.J., Barnes, C.A. and McNaughton, B.L. (1989) Reversible inactivation of the medial septum: selective effects on the spontaneous unit activity of different hippocampal cell types. Brain Res., 500: 99–106. Mizumori, S.J., Barnes, C.A. and McNaughton, B.L. (1992) Differential effects of age on subpopulations of hippocampal theta cells. Neurobiol. Aging, 13: 673–679. Montarolo, P.G., Goelet, P., Castellucci, V.F., Morgan, J., Kandel, E.R. and Schacher, S. (1986) A critical period for macromolecular synthesis in long-term heterosynaptic facilitation in Aplysia. Science, 234: 1249–1254. Morris, R.G.M., Anderson, E., Lynch, G.S. and Baudry, M. (1986) Selective impairment of learning and blockade of longterm potentiation by an N-methyl-D-aspartate receptor antagonist, AP5. Nature, 319: 774–776. Morrison, J.H. and Hof, P.R. (1997) Life and death of neurons in the aging brain. Science, 278: 412–419. Nguyen, P.V., Abel, T. and Kandel, E.R. (1994) Requirement of a critical period of transcription for induction of a late phase of LTP. Science, 265: 1104–1107. Nicolle, M.M., Gallagher, M. and McKinney, M. (1999) No loss of synaptic proteins in the hippocampus of aged, behaviorally impaired rats. Neurobiol. Aging, 20: 343–348. Niesen, C.E., Baskys, A. and Carlen, P.L. (1988) Reversed ethanol effects on potassium conductances in aged hippocampal dentate granule neurons. Brain Res., 445: 137–141. Peters, A., Rosene, D.L., Moss, M.B., Kemper, T.L., Abraham, C.R., Tigges, J. and Albert, M.S. (1996) Neurobiological bases of age-related cognitive decline in the rhesus monkey. J. Neuropathol. Exp. Neurol., 55: 861–874. van Praag, H., Schinder, A.F., Christie, B.R., Toni, N., Palmer, T.D. and Gage, F.H. (2002) Functional neurogenesis in the adult hippocampus. Nature, 415: 1030–1034. van Praag, H., Shubert, T., Zhao, C. and Gage, F.H. (2005) Exercise enhances learning and hippocampal neurogenesis in aged mice. J. Neurosci., 25: 8680–8685. Ramirez-Amaya, V., Gage, F.H., Worley, P.F. and Barnes, C.A. (2006) Integration of new neurons into functional neural networks. J. Neurosci., 26: 12237–12241. Rao, M.S., Hattiangady, B., Abdel-Rahman, A., Stanley, D.P. and Shetty, A.K. (2005) Newly born cells in the ageing dentate gyrus display normal migration, survival and neuronal fate choice but endure retarded early maturation. Eur. J. Neurosci., 21: 464–476. Rapp, P.R. and Amaral, D.G. (1991) Recognition memory deficits in a subpopulation of aged monkeys resemble the effects of medial temporal lobe damage. Neurobiol. Aging, 12: 481–486. Rapp, P.R., Deroche, P.S., Mao, Y. and Burwell, R.D. (2002) Neuron number in the parahippocampal region is preserved

678 in aged rats with spatial learning deficits. Cereb. Cortex, 12: 1171–1179. Rapp, P.R. and Gallagher, M. (1996) Preserved neuron number in the hippocampus of aged rats with spatial learning deficits. Proc. Natl. Acad. Sci. U.S.A., 93: 9926–9930. Rapp, P.R., Stack, E.C. and Gallagher, M. (1999) Morphometric studies of the aged hippocampus: I. Volumetric analysis in behaviorally characterized rats. J. Comp. Neurol., 403: 459–470. Rasmussen, T., Schliemann, T., Sorensen, J.C., Zimmer, J. and West, M.J. (1996) Memory impaired aged rats: no loss of principal hippocampal and subicular neurons. Neurobiol. Aging, 17: 143–147. Ribak, C.E., Korn, M.J., Shan, Z. and Obenaus, A. (2004) Dendritic growth cones and recurrent basal dendrites are typical features of newly generated dentate granule cells in the adult hippocampus. Brain Res., 1000: 195–199. Rolls, E.T. (1990) Functions of the primate hippocampus in spatial processing and memory. In: Kesner R.P. and Olton D.S. (Eds.), Neurobiology of Comparative Cognition. Comparative Cognition and Neuroscience. Lawrence Erlbaum Associates, Inc., Hillsdale, MI, pp. 339–362. Rose, G., Diamond, D. and Lynch, G.S. (1983) Dentate granule cells have the behavioral characteristics of theta neurons. Brain Res., 266: 29–37. Rosenzweig, E.S. and Barnes, C.A. (2003) Impact of aging on hippocampal function: plasticity, network dynamics, and cognition. Prog. Neurobiol., 69: 143–179. Samsonovich, A.V. and Ascoli, G.A. (2006) Morphological homeostasis in cortical dendrites. Proc. Natl. Acad. Sci. U.S.A., 103: 1569–1574. Shen, J. and Barnes, C.A. (1996) Age-related decrease in cholinergic synaptic transmission in three hippocampal subfields. Neurobiol. Aging, 17: 439–451. Shen, J., Barnes, C.A., McNaughton, B.L., Skaggs, W.E. and Weaver, K.L. (1997) The effect of aging on experience-dependent plasticity of hippocampal place cells. J. Neurosci., 17: 6769–6782. Shen, J., Kudrimoti, H.S., McNaughton, B.L. and Barnes, C.A. (1998) Reactivation of neuronal ensembles in hippocampal dentate gyrus during sleep after spatial experience. J. Sleep Res., 7(Suppl 1): 6–16.

Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and Gould, E. (2001) Neurogenesis in the adult is involved in the formation of trace memories. Nature, 410: 372–376. Small, S.A., Chawla, M.K., Buonocore, M., Rapp, P.R. and Barnes, C.A. (2004) Imaging correlates of brain function in monkeys and rats isolates a hippocampal subregion differentially vulnerable to aging. Proc. Natl. Acad. Sci. U.S.A., 101: 7181–7186. Small, S.A., Tsai, W.Y., DeLaPaz, R., Mayeux, R. and Stern, Y. (2002) Imaging hippocampal function across the human life span: is memory decline normal or not? Ann. Neurol., 51: 290–295. Smith, T.D., Adams, M.M., Gallagher, M., Morrison, J.H. and Rapp, P.R. (2000) Circuit-specific alterations in hippocampal synaptophysin immunoreactivity predict spatial learning impairment in aged rats. J. Neurosci., 20: 6587–6593. Song, H., Kempermann, G., Overstreet Wadiche, L., Zhao, C., Schinder, A.F. and Bischofberger, J. (2005) New neurons in the adult mammalian brain: synaptogenesis and functional integration. J. Neurosci., 25: 10366–10368. Sterio, D.C. (1984) The unbiased estimation of number and sizes of arbitrary particles using the disector. J. Microsc., 134: 127–136. Wati, H., Kudo, K., Qiao, C., Kuroki, T. and Kanba, S. (2006) A decreased survival of proliferated cells in the hippocampus is associated with a decline in spatial memory in aged rats. Neurosci Lett., 399: 171–174. Wenk, G.L. and Barnes, C.A. (2000) Regional changes in the hippocampal density of AMPA and NMDA receptors across the lifespan of the rat. Brain Res., 885: 1–5. West, M.J. (1993a) New stereological methods for counting neurons. Neurobiol. Aging, 14: 275–285. West, M.J. (1993b) Regionally specific loss of neurons in the aging human hippocampus. Neurobiol. Aging, 14: 287–293. West, M.J., Coleman, P.D., Flood, D.G. and Troncoso, J.C. (1994) Differences in the pattern of hippocampal neuronal loss in normal ageing and Alzheimer’s disease. Lancet, 344: 769–772. Worley, P.F., Bhat, R.V., Baraban, J.M., Erickson, C.A., McNaughton, B.L. and Barnes, C.A. (1993) Thresholds for synaptic activation of transcription factors in hippocampus: correlation with long-term enhancement. J. Neurosci., 13: 4776–4786.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 37

The effects of aging on dentate circuitry and function Peter R. Patrylo1, and Anne Williamson2 1

Department of Physiology, Southern Illinois University School of Medicine Carbondale, IL 62901, USA 2 Department of Neurosurgery, Yale University School of Medicine New Haven, CT 06520, USA

Abstract: The central nervous system (CNS) undergoes a variety of anatomic, physiologic, and behavioral changes during aging. One region that has received a great deal of attention is the hippocampal formation due to the increased incidence of impaired spatial learning and memory with age. The hippocampal formation is also highly susceptible to Alzheimer’s disease, ischemia/hypoxia, and seizure generation, the three most common aging-related neurological disorders. While data reveal that the dentate gyrus plays a key role in hippocampal function and dysfunction, the majority of electrophysiological studies that have examined the effects of age on the hippocampal formation have focused on CA3 and CA1. We perceive this to be an oversight and consequently will highlight data in this review which demonstrate an age-related disruption in dentate circuitry and function, and propose that these changes contribute to the decline in hippocampal-dependent behavior seen with ‘‘normal’’ aging. Keywords: dentate gating; mossy fibers; recurrent excitation; IPSP; epileptogenicity; spatial learning and memory; aging function comes from several experimental findings. First, 2-deoxyglucose studies and electrophysiological recordings have shown that seizures are unlikely to invade the hippocampus unless epileptiform activity is generated within the dentate gyrus itself (Heinemann et al., 1992; Lothman et al., 1992). Second, low frequency orthodromic activation of dentate granule cells normally suppresses CA3 pyramidal cell activity (Bragin et al., 1995; Penttonen et al., 1997) since granule cell axons preferentially form synapses with inhibitory interneurons in CA3 (Acsady et al., 1998). If granule cell activity is increased, however, glutamatergic excitation can be evoked in CA3 due to greater frequency facilitation at synapses between granule cell axons (mossy fibers) and pyramidal neurons (Henze et al., 2002; Lawrence and McBain, 2003). Third, several groups have shown that, if the intrinsic properties or connectivity of the dentate gyrus are altered

Dentate filter function ‘‘Normal’’ and pathologic hippocampal function (e.g., spatial learning and memory, and epileptiform activity, respectively) are both associated with an increase in excitatory activity between principal cells. Determining whether ‘‘normal’’ or pathological function occurs has been proposed to depend, in part, on the dentate gyrus due to it’s strategic position within the tri-synaptic circuit of the hippocampal formation that allows it to ‘‘filter/gate’’ the majority of excitatory input entering hippocampal region CA3, a region highly susceptible to epileptiform activity (Johnson and Brown, 1984; Miles and Wong, 1987; Christian and Dudek, 1988). Evidence that the dentate gyrus subserves a gating/filter Corresponding author. Tel.: +1618 453 6743; Fax: +1618 453 1517; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63037-4

679

680

experimentally or naturally (e.g., insult or genetic variability), hippocampal-dependent function is affected adversely including, learning and memory (Lipp et al., 1984; Crusio et al., 1987; McNaughton et al., 1989; Nanry et al., 1989; Bernasconi-Guastalla et al., 1994; Schuster et al., 1997), synaptic plasticity (Beck et al., 2000), and seizure susceptibility (Grimes et al., 1990; Cronin et al., 1992; Grecksch et al., 1995; Wuarin and Dudek, 1996; Patrylo and Dudek, 1998). The cellular and circuit properties that bestow the dentate gyrus with the capacity to ‘‘filter/gate’’ excitatory activity are unknown. However this region does exhibit several unique physiological properties that are likely to contribute to this function. Among these properties are: (1) the principal cells (granule cells) having a relatively hyperpolarized resting membrane potential and strong spike frequency adaptation (Fricke and Prince, 1984; Staley et al., 1992); (2) the capacity of dentate granule cells to fire normally in single spike mode following disinhibition versus a graded burst discharge (Schwartzkroin and Prince, 1978; Fricke and Prince, 1984; Lynch et al., 2000); (3) the presence of strong feedfoward and feedback inhibitory circuits (Andersen, 1975; Freund and Buzsa´ki, 1996); (4) a paucity of recurrent excitatory interconnections (Schwartzkroin and Prince, 1978; Fricke and Prince, 1984; Wuarin and Dudek, 1996; Nadler, 2003); and (5) the dynamic properties of mossy fiber — CA3 synaptic transmission (e.g., Lawrence and McBain, 2003).

Aging and dentate filter function Spatial learning and memory, synaptic plasticity, and seizure susceptibility are all affected adversely with age. Moreover, available data suggest that several of the cellular and circuit properties of the dentate gyrus are altered during aging (see below) and consequently it raises the question whether dentate filter function is also comprised. To begin to determine the effects of age on dentate filter function, Patrylo et al. (2007) examined field potential activity elicited in hippocampal region CA3 with dentate molecular layer stimulation in brain slices from adult (2–11 months) and aged Fisher

344 (21–32 months) rats. The Fischer 344 rat is one of the most commonly used strain of rodents in neurobiology of aging research, and the ages used reflects the convention of classifying rodents as aged when approximately 50% of their cohorts have perished. To ensure that the activity elicited in CA3 was disynaptic in nature, and thus a reflection of dentate filtering, only responses with a delay Z2 ms were examined (monosynaptic interconnections normally exhibit a delay r2 ms). These experiments revealed that during a train of 5 Hz molecular layer stimulation (theta frequency) hyperexcitable activity (multiple population spikes) occurs in area CA3 in approximately 30% of aged slices (Fig. 1C). These slices also frequently exhibited spontaneous interictal-like activity in CA3 following termination of the train. In contrast, 100% of adult slices and the remaining 70% of aged slices exhibited either suppression or no change in the amplitude of the activity evoked during the train and no subsequent epileptiform activity. Further, this compromise in dentate filter function appeared to be associated with an impaired capacity of the dentate gyrus to respond to afferent input, since 5 Hz molecular layer stimulation evoked hyperexcitable activity in the dentate gyrus in approximately the same percentage of aged slices as were noted to generate hyperexcitable activity in CA3 (Patrylo et al., 2007). These data were also proposed to reflect an aging-related compromise in inhibitory activity in CA3 since modeling studies indicate that epileptiform activity is generated in hippocampal region CA3 if granule cell activity is increased and inhibitory tone is compromised (Lawrence and McBain, 2003). This aging-related breakdown of dentate filter function poses two questions. (1) What are the underlying mechanisms for this decline in function? (2) Does this disruption in dentate function contribute to aging-related hippocampal dysfunction (e.g., impaired spatial learning and memory)? While, the answers to these questions are currently unclear, in the subsequent sections we will highlight aging-related changes that are likely to contribute to this decline in dentate filter function, and propose that this change in function contributes to the decline in spatial learning and memory and enhanced susceptibility for complex partial seizures during aging.

681

Fig. 1. Dentate filter function is impaired in a subpopulation of aged rodents. (A) Schematic diagram illustrating the position of the recording and stimulating electrodes used to assess dentate filter function. Note that the recordings were performed in CA3. (B) The experimental paradigm illustrated schematically. (C) In adult slices, 5 Hz molecular layer stimulation elicited minimal change in the evoked response, recorded from CA3. In contrast, similar stimulation could elicit hyperexcitability in approximately 30% of aged slices. (D) The majority of aged slices exhibiting multiple population spikes during 5 Hz molecular layer stimulation developed spontaneous epileptiform activity subsequently.

Potential mechanisms contributing to the agingrelated decline in dentate filter function Numerous changes have been reported to occur in the dentate gyrus during aging and a partial listing

is illustrated in Table 1. To summarize briefly, aging is associated with: (1) a change in synaptic connectivity (e.g., Geinisman et al., 1992; Patrylo et al., 2007); (2) an increase in dye coupling between granule cells (Barnes et al., 1987); (3) a

682

change in receptor and channel properties (e.g., Gutierrez et al., 1996; Magnusson et al., 2006); (4) a compromise in synaptic plasticity (e.g., Barnes and McNaughton, 1980b; de Toledo-Morrell et al., 1988; Barnes, 2000; Barnes et al., 2000); (5) a decrease in extracellular volume (Sykova et al., 1998); and (6) a change in the levels of growth factors that can affect neuronal survival, metabolism, and activity (e.g., Katoh-Semba et al., 1998; Sonntag et al., 1999). While all of these changes would be expected to alter dentate network activity and thus filter function, it is beyond the scope of this article to provide a detailed review of each change. Consequently, we will focus on data demonstrating an aging-related change in local inhibitory and excitatory circuitry (GABAergic and glutamatergic respectively) since similar alterations have been reported in models of epilepsy and have been associated with a breakdown of dentate filter function (e.g., Behr et al., 1998; Finnerty et al., 2001; Nadler, 2003).

GABAergic circuitry Interneurons help regulate principal cell excitability, receptive field size, and plasticity, and consequently play a critical role in information processing as well as pathophysiology (Freund and Buzsa´ki, 1996; Paulsen and Moser, 1998). In the dentate gyrus a tremendous diversity of interneurons are found (Freund and Buzsa´ki, 1996) that can be divided into two main categories based on the distribution of their axonal arbor and subsequently physiological function. The first group exhibits an axonal arbor that forms synapses with the somata, initial dendritic tree, and axon hillock of granule cells and are involved with regulating the capacity of granule cells to discharge. The second group exhibits axonal projections that form synapses with the distal segments of granule cell dendrites and subsequently help regulate afferent input (i.e., perforant path input). These two types of interneurons play distinct roles in ‘‘normal’’ hippocampal function (Austin et al., 1989; Moser, 1996) and are affected differentially in specific neuropathological conditions (e.g., Kobayashi and Buckmaster, 2003).

During aging, the inhibitory system in the dentate gyrus exhibits a variety of anatomic and biochemical changes that suggest a compromise in function. Among these changes are: (1) a decrease in glutamic acid decarboxylase (GAD; Shetty and Turner, 1998; Vela et al., 2003; Ling et al., 2005) and extracellular GABA levels (Almaguer-Melian et al., 2005); (2) an increase in GABA transaminase (Lumeng and Li, 1974; Kugler and Baier, 1992; Hwang et al., 2004); and (3) a decrease in the number of neurons immunopositive for GAD, and/or other peptides/neuromodulators that colocalize in inhibitory neurons (Shetty and Turner, 1998; Cadacio et al., 2003; Vela et al., 2003; Stanley and Shetty, 2004). It is important to note however that several other changes occur during aging that have been suggested to be compensatory. For example, an increase in the levels of the GABA a1 subunit has been noted in the hippocampus and dentate gyrus of aged rodents and has been proposed to be an adaptive response to the decrease in GAD levels (e.g., Gutierrez et al., 1996) since incorporation of the a1 subunit into a GABA receptor/channel complex results in a current that has a prolonged decay constant (e.g., Fisher, 2004). Examination of the anatomical data on the distribution and number of GAD immunoreactive neurons suggest that not all inhibitory circuits are affected equally with age. Specifically, Shetty and Turner (1998) reported that the number of GAD-immunoreactive neurons within and subjacent to the granule cell layer is preserved with age, while GAD immunopositive neuron number in the hilus is decreased. Based on the characteristic laminar location of specific types of interneurons (e.g., Freund and Buzsaki, 1996), these findings suggest that the basket and chandelier cells, those primarily responsible for somatic and axonal inhibition, are preserved with age and subsequently suggest that feedback inhibition should be preserved during aging. In agreement with this suggestion, the number of parvalbuminimmunoreactive neurons within and adjacent to the granule cell layer is preserved with age (Shetty and Turner, 1998) and basket and chandelier cells co-localize this calcium binding protein (e.g., Soriano et al., 1990; Ribak, 1992; Ribak et al., 1993; Freund and Buzsa´ki, 1996). Further,

Table 1. Aging-related changes in the dentate gyrus

Glutamatergic NT Entorhinal

Assoc/Com Mossy fibers

AMPA receptors

KA receptors

NMDA receptors

Miscellaneous

Findings

Species

Exemplary reference

Decreased axospinous synapse number in the MML Decreased ratio of Timm staining volume in the MML/ ML Decreased pre-synaptic volley Decreased axospinous synapse number in the IML Expanded zone of Timm staining in the IML Elaboration of Timm stained mossy fibers into the GCL and IML — can decrease slightly in very old subjects Elaboration of Timm stained mossy fibers into the IML Decreased GluR1 and GluR2 mRNA levels No significant change in [3H] AMPA binding density Decreased [3H] AMPA binding density Decreased [3H]KA binding density in the hilus Expanded zone, but not density, of [3H]KA binding in the IML Non-significant trend for decreased [3H]NMDA receptor binding in ML No significant change in [3H] glutamate binding (AP-5 subtraction) Non-significant trend for decreased NR2B mRNA Decreased NR2B mRNA; decreased NR1 mRNA in the ventral DG Decreased [3H] glutamate and [3H] CPP binding density Decreased NR2B, but not NR2A and NR1, binding density; NR2B decreased preferentially in the intermediate DG Decreased NR1-IR Increased probability for prolonged EPSPs in BIC

Rat – F344 Rat – Long Evans

Geinisman et al. (1992) Rapp et al. (1999)

Rat – F344 Rat – F344 Rat – Long Evans Guinea pig

Barnes and McNaughton (1980a) Geinisman et al. (1992) Rapp et al. (1999) Wolfer and Lipp (1995)

Human Rat – Wistar Rat – Long Evans Rat – SD Rat – Long Evans Rat – Long Evans Rat – Long Evans

Cassell and Brown (1984) Pagliusi et al. (1994) Nicolle et al. (1996) Wenk and Barnes (2000) Nagahara et al. (1993) Nagahara et al. (1993); Nicolle et al. (1996) Nicolle et al. (1996)

Rat – SD

Wenk and Barnes (2000)

Rhesus monkey Mice – C57Bl/6

Bai et al. (2004) Magnusson (2000)

Mice – C57Bl/6 Mice – C57BL/6J

Magnusson (2000) Magnusson et al. (2006)

Rhesus monkey Rat – F344

Gazzaley et al. (1996) Patrylo et al. (2007)

683

684

Table 1 (continued ) Findings

Species

Exemplary reference

Decreased number of GAD-immunopositive neurons Decreased axodendritic synapse number in OML (on shaft) Preserved a1 receptor subunit IR Increased a1 (but not b2, b3, and g2) mRNA levels and IR Increased [3H] zolpidem binding Preserved paired-pulse inhibition and polysynaptic evoked IPSPs (conductances and reversal potentials) Decreased sIPSP frequency Increased mIPSC decay time constant Increased benzodiazepine potentiation of mIPSC decay constant

Rat – F344 Rhesus monkey

Shetty and Turner, (1998) Tigges et al. (1995)

Rhesus monkey Rat – F344 & SD Rat – Wistar Rat – F344

Rissman et al. (2006) Gutierrez et al. (1996) Ruano et al. (1995) Barnes (1979); Patrylo et al. (2007)

Rat – F344 Rhesus monkey Rhesus monkey

Patrylo et al. (2007) Luebke and Rosene (2003) Luebke and Rosene (2003)

Decreased LTP maintenance

Rat – F344

Decreased capacity for LTP induction at low, but not high, stimulus intensities Decreased capacity for perforant path — kindling Decreased BDNF-induced LTP

Rat – F344 Rat – F344 Rat – Wistar

Barnes (1979); Barnes and McNaughton (1980b) Barnes (1979); Barnes et al. (2000); Orr et al. (2001) de Toledo-Morell et al. (1984) Gooney et al. (2004)

Ion channels and intrinsic properties Ca2+

Decreased calcium currents (e.g., L-type)

Rat – F344

Intrinsic membrane properties

No change in RMP, Rin, spike width, AP amplitude, and time constant

Rat – F344

No change in AHP and spike frequency adaptation

Rat – F344

GABAergic NT Cell number Synapse number Receptors Binding Electrophysiology

Plasticity

Reynolds and Carlen (1989) (however see Herman et al., 1998; Thibault et al., 2001, regarding CA1) Barnes and McNaughton (1980a); Niesen et al. (1988); Patrylo et al. (2007) Niesen et al. (1988)

Other NTs and neuropeptides NPY Somatostatin ACh

Growth factors BDNF

Insulin and IGF NGF and NT3

Decreased number of NPY immunopositive neurons

Rat – F344

Decreased somatostatin mRNA levels and IR Decreased vesicular ACh transporter IR Decreased ChAT IR Decreased amplitude of the slow cholinergic EPSP Decreased M2 and M1 receptor binding; trend for increased M3 receptor binding Increased [3H]AF-DX 384 binding (increased M2 levels) Decreased nAChRa4 IR in polymorphic region

Rat – Wistar Rhesus monkey Rat – Wistar Rat – F344 Rat – F344

Hattiangady et al. (2005); Cadacio et al. (2003) Vela et al. (2003) Calhoun et al. (2004) Lukoyanov et al. (1999) Shen and Barnes (1996) Tayebati et al. (2002)

Rat – Long Evans Mouse – CBA/J

Aubert et al. (1995) Rogers et al. (1998)

Decreased BDNF activity and IR Decreased BDNF IR Increased BDNF IR in the hilus No change in binding density for the IGF1, IGF2, and insulin receptors Unknown whether changes are specifically seen within the dentate gyrus

Rat – F344 Rhesus monkey Rat – SD Rat – Long Evans

Hattiangady et al. (2005) Hayashi et al. (2001) Katoh-Semba et al. (1998) Dore et al. (1997)

Decreased rate of basal neurogenesis

Rat – F344

Maturation of newly generated neurons retarded

Rat – F344

Kuhn et al. (1996); Merrill et al. (2003) Rao et al. (2005)

Decreased extracellular volume Increased dye coupling

Rat – Wistar Rat – F344

Sykova et al. (1998) Barnes et al. (1987)

Neurogenesis

Miscellaneous

ACh, acetylcholine; AF-DX, 5,11-dihydro-11-[[(2-(2-[(dipropylamino) methyl]-1-piperidinyl) ethyl) amino] carbonyl]-6H-pyrido [2, 3-b] [1,4]-benzodiazepin-6-one; AHP, after hyperpolarization; AMPA, alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; AP, action potential; AP-5, amino phosphonovalerate; Assoc/Com, associational/commissural; BDNF, brain-derived neurotrophic factor; BIC, bicuculline methiodide; CPP, [(7)-2-carboxypiperazin-4-yl] propyl-1-phosphonic acid; DG, dentate gyrus; EPSP, excitatory post-synaptic potential; F344, Fischer 344; GAD, glutamic acid decarboxylase; GCL, granule cell layer; IGF1, insulin-like growth factor-1; IGF2, insulin-like growth factor-2; IML, inner molecular layer; IPSP, inhibitory post-synaptic potential; IR, immunoreactivity; KA, kainate; LTP, long term potentiation; mIPSC, miniature inhibitory post-synaptic current; ML, molecular layer; MML, middle molecular layer; NGF, nerve growth factor; NMDA, n-methyl-d-aspartate; NPY, neuropeptide Y; NT, neurotransmission; NT3, neurotrophin-3; OML, outer molecular layer; Rin, input resistance; RMP, resting membrane potential; SD, Sprague-Dawleys; IPSP, Spontaneous IPSP.

685

686

electrophysiological data from hippocampal slices from aged rodents have shown that paired-pulse inhibition is preserved in the aged dentate gyrus (Barnes 1979; Patrylo et al., 2007), and whole cell current clamp data reveal no significant change in the reversal potentials and conductances of fast and slow inhibitory post-synaptic potentials (IPSPs) evoked in granule cells from aged versus adult rodents (Patrylo et al., 2007). Regarding the aging-related decrease in number of GAD immunopositive neurons within the hilus, anatomic and electrophysiological data suggest that these interneurons are likely to be those involved with inhibiting the distal dendrites of granule cells and subsequently modulating afferent input. Specifically, anatomic studies have shown that the number of somatostatin (SOM) and neuropeptide-Y (NPY) immunoreactive interneurons are decreased within the hilus with age (Cadacio et al., 2003; Vela et al., 2003) and these types of interneurons are known to have an axonal arborization that targets the distal dendrites of granule cells (e.g., Leranth et al., 1990; Freund and Buzsa´ki, 1996; Katona et al., 1999; Maccaferri et al., 2000). Further, whole cell patch recordings from hippocampal slices from adult and aged Fisher 344 rats have shown that the overall frequency of spontaneous IPSPs (sIPSPs) is decreased in aged granule cells, although the frequency of sIPSPs with a large amplitude and prolonged duration is preserved (Patrylo et al., 2007). Since large amplitude composite inhibitory events are believed to arise from input impinging on the soma and proximal dendrites of granule cells, while smaller unitary events are believed to result from more distal input (Solte`sz et al., 1995; Williams et al., 1998), this aging-related decrease in sIPSP frequency is likely to reflect a reduction of inhibitory input onto the distal dendrites of granule cells (Patrylo et al., 2007). In vitro studies on hippocampal slices from aged primates have also noted a trend for reduced IPSC frequency, although this change did not reach significance (Luebke and Rosene, 2003). It is interesting to note that afferent input is suppressed during theta wave activity (Moser, 1996), while feedback inhibition onto the soma/ axon of granule cells is decreased (Austin et al., 1989; Moser, 1996). Consequently, it has been proposed that inhibitory input onto distal

dendrites and presynaptic afferent fibers play a critical role in dentate filter function and encoding of new spatial information due to increasing the signal to noise ratio of incoming cortical input (Paulsen and Moser, 1998). These findings also suggest that if such dendritic/afferent inhibition is compromised, both functions would be affected adversely. Based on the data presented above, we propose that inhibitory input onto the distal dendrites of granule cells, and potentially afferent terminals, is decreased during aging and results in abnormal processing of afferent input during spatial exploration, and subsequently decreases the signal to noise ratio below the level required for transmitting new spatial information. Experiments examining dendritic inhibition in behaviorally characterized aged animals are required to address this possibility, however. Whether the decrease in the number of GAD-IR neurons within the hilus of the dentate gyrus of aged rodents reflects a loss of SOM and NPY containing interneurons, a change in their properties (decreased protein levels), or both, is unclear with currently available data supporting either possibility. For example, studies examining total hilar neuron number reveal that there is no aging-related change (Rasmussen et al., 1996), while studies examining the number of NPY immunoreactive neurons suggest that a relatively small subpopulation of neurons may be lost (Cadacio et al., 2003; Vela et al., 2003). It should be noted, however, that attention to accurate immunohistochemical cell quantification is critical, since the level of an immunohistochemical marker may simply decrease below the level required for visual detection without a loss of the cell expressing the protein. In this regard, it is interesting to note that Stanley and Shetty (2004) have reported that the loss of GAD-IR neurons in hippocampal regions CA3 and CA1 does not correspond with a decrease in total neuron number, and that several groups have shown that GAD, SOM, and NPY protein expression can be regulated by afferent input and neurotrophic levels (Akhtar and Land, 1991; Hyman et al., 1994; Carnahan and Nawa, 1995; Patel et al., 1995; Aamodt et al., 2000; Marty, 2000; Reibel et al., 2000; de Almeida et al., 2002) which may be altered with age (e.g., Geinisman et al., 1992; Katoh-Semba et al., 1998; Patrylo

687

et al., 2007). Consequently, additional studies are required to determine whether aging is associated with a loss of interneurons in the dentate gyrus, or a change in protein expression and/or function. Regardless, these aging-related changes in inhibitory efficacy are likely to affect granule cell activity and synchronization, and consequently their capacity to respond to afferent input. If this decline in inhibitory efficacy is further coupled with an enhancement of NMDA-receptor mediated activity, or an increase in recurrent excitatory interconnections, one would predict a breakdown of the dentate gate (e.g., Lynch et al., 2000; Finnerty et al., 2001; Nadler, 2003). During aging it is unlikely that an enhancement of NMDA receptormediated activity in the dentate gyrus contributes to the decline in filter function, since decreases in the protein levels of certain NMDA receptor subunits (Gazzaley et al., 1996; Nicolle et al., 1996; Magnusson, 2000; Bai et al., 2004; Magnusson et al., 2006), the strength of NMDA-receptor mediated activity (Rao et al., 1994), and NMDA-receptor mediated plasticity have been reported with age (Barnes and McNaughton, 1980b; de ToledoMorrell et al., 1988; Barnes et al., 2000). As for a change (increase) in local excitatory interconnectivity contributing to the compromise in dentate filter function, we will review data in the next section that address this possibility.

Glutamatergic circuitry In the early 1990s, Geinisman et al. (1992) reported that axospinous synapses number is decreased in the molecular layer (middle and inner molecular layers) of the dentate gyrus in aged rodents and is associated with the cognitive status of the animal (loss occurred in cognitively impaired subjects). This loss of axospinous synapses is believed to reflect a decrease in entorhinal and associational/commissural input onto granule cell dendrites, because electrophysiological studies indicate that the amplitude of the pre-synaptic volley and resulting EPSP are decreased in the dentate gyrus of aged rats (e.g., Barnes and McNaughton, 1980a, b). Further, a loss of presynaptic neurons (layer II neurons in the entorhinal cortex) is

unlikely to contribute to this change since neuron number in the entorhinal cortex is preserved in aged animals (Merrill et al., 2001; Rapp et al., 2002). Subsequent studies using synaptophysin immunohistochemistry (Smith et al., 2000) have provided corroboratory findings. Specifically, Rapp and colleagues noted that there is a decrease in the level of synaptophysin immunostaining in the molecular layer of the dentate gyrus, as well as stratum lacunosum moleculare of CA3 in aged rodents with cognitive impairment. Although this study did not determine whether this change in synaptophysin immunohistochemistry reflects a loss or modification of synapses, the results do reveal an aging-related change in a pre-synaptic marker that is found preferentially in cognitively impaired rodents. Based on these results, Rapp and colleagues proposed that the decline in spatial learning and memory during aging results from the disruption in information flow from the entorhinal cortex to the hippocampus. It should be noted however, that the decrease in axospinous synapse number in the molecular layer of aged rodents could also reflect a decrease in excitatory input onto inhibitory interneurons with dendritic spines (Soriano and Frotscher, 1993; Freund and Buzsa´ki, 1996; Mott et al., 1997), although one could argue that these are few in number compared to granule cell dendrites. Further, a decrease of inhibitory input onto granule cell dendritic spines is unlikely to be a significant contributor because GABAergic synapses are not typically axospinous (Halasy and Somogyi, 1993; Soriano and Frotscher, 1993; Freund and Buzsa´ki, 1996; Katona et al., 1999). In the adult CNS, a loss of entorhinal input into the dentate gyrus (e.g., unilateral lesion) results in a transient compromise in cognitive function and an increased propensity for epileptiform activity (e.g., Steward, 1982; Kelley and Steward, 1996). These effects eventually reverse, however, due to reorganization of the remaining excitatory circuits, in particular the temporodentate pathway arising from the contralateral entorhinal cortex (e.g., Steward, 1982; Reeves and Smith, 1987). Associational/commissural fibers and the mossy fibers have also been shown to reorganize following a loss of entorhinal input (e.g., Lynch et al., 1976;

688

Steward et al., 1976; Steward, 1982, 1992; West and Dewey, 1984; Reeves and Smith, 1987). During aging, the loss of entorhinal input is bilateral in nature, and therefore suggests that the capacity for restorative reorganization is likely to be reduced. This raises the question whether there is evidence for reorganization of the associational/commissural fibers and/or mossy fibers in the aged dentate gyrus. Reorganization of these pathways would increase recurrent excitatory interconnections within the hippocampus and thus potentially affect dentate filter function as has been suggested in models of temporal lobe epilepsy (e.g., Behr et al., 1998; Finnerty et al., 2001; Nadler, 2003). Electrophysiological studies suggest that local excitatory circuits do reorganize within the dentate gyrus with age. Current source density (CSD) studies have shown that antidromic stimulation elicits an additional peak of net inward current in the granule cell and inner molecular layer of the dentate gyrus in slices from aged, but not adult, rodents (Barnes and McNaughton, 1983). While the reason for this additional current sink is unknown, it is consistent with the presence of additional synaptic input onto the soma and proximal dendrites of granule cells. Furthermore, whole cell patch clamp recordings from transverse hippocampal slices have demonstrated that granule cells from aged rodents have an increased propensity for generating spontaneous, prolonged, large amplitude EPSPs following exposure to the GABAA receptor antagonist bicuculline methiodide (Patrylo et al., 2007). These events resemble the giant synaptic potentials that underlie paroxysmal depolarizing shifts, and suggest that aberrant excitatory circuits are present in the aged dentate gyrus, since granule cells do not typically exhibit reverberatory activity following disinhibition (Schwartzkroin and Prince, 1978; Fricke and Prince, 1984; Wuarin and Dudek, 1996). Although the cellular player(s) that contribute to this aberrant excitatory circuit are unclear, anatomic data suggest several possibilities. First, the volume of [3H] kainate binding (Nicolle et al., 1996) and Timm staining (Rapp et al., 1999) within the inner molecular layer relative to the middle/outer molecular layers is increased in the aged dentate gyrus, and thus it could reflect an expansion of the associational/commissural fibers.

With regard to the CSD studies (Barnes and McNaughton, 1983) however, this expansion is unlikely to produce the novel net inward current in the granule cell and inner molecular layer of the aged dentate gyrus since even though granule cells form a feedback circuit with hilar mossy cells and CA3 pyramidal neurons in the in vivo dentate gyrus (Ishizuka et al., 1990; Li et al., 1994; Buckmaster et al., 1996; Wenzel et al., 1997) the majority of these interconnections are severed during the preparation of transverse hippocampal slices (Ishizuka et al., 1990; however see Scharfman, 1994; Buckmaster et al., 1996; Wenzel et al., 1997). Thus, unless the distribution of the association/commissural fibers is altered along the longitudinal axis of the hippocampus during aging, it is unlikely that this reorganization accounts for the increase in excitatory activity in hippocampal slices from aged rodents. Moreover, Geinisman et al. (1992) suggest that there is a decrease in associational/commissural — granule cell axospinous synapses within the inner molecular layer. An alternate explanation for the net inward current in the granule cell and inner molecular layer is that the distribution of mossy fibers changes with age. In this regard, studies that have labeled the mossy fibers selectively in tissue from humans and rodents using the Timm stain have shown that mossy fibers can extend up into the granule cell and inner molecular layers in the transverse plane of the hippocampus in aged subjects (Cassell and Brown, 1984; Wolfer and Lipp, 1995). While this distribution pattern of mossy fibers corresponds with the location of the novel net current flux in the aged dentate gyrus (Barnes and McNaughton, 1983), no data are currently available to verify that they lead to increased granule cell to granule cell synapses within the aged dentate gyrus. Recent data suggest an additional mechanism that could contribute to the aberrant excitatory interconnections in the aged dentate gyrus (Rao et al., 2005). Specifically, it was demonstrated that newly generated granule cells in middle-aged and aged rodents exhibit a slowing of developmental maturation and an increased incidence of basal dendrites relative to granule cells born in younger animals. While it is unclear whether these basal dendrites persist in the newly generated granule

689

cells, it is interesting to note that GABAA receptor-mediated inhibition plays a critical role in integrating newly generated granule cells into dentate circuitry (Ge et al., 2006) and inhibitory tone is disrupted during aging (see above). Several groups have reported basal dendrites in animal models of temporal lobe epilepsy (Spigelman et al., 1998; Buckmaster and Dudek, 1999; Ribak et al., 2000) and have provided evidence that these aberrant processes can create anomalous excitatory circuits capable of positive feedback (e.g., Ribak et al., 2000). In summary, the efficacy of the inhibitory system is compromised within the dentate gyrus during aging, and excitatory interconnections appear to be increased. This disruption in the balance of inhibition relative to excitation could impair dentate filter function, as has been shown in animal models of epilepsy that exhibit similar, although more severe, changes in dentate circuitry.

The dentate gyrus, spatial learning and memory, and aging One of the most common aging-related neurological dysfunctions is a decline in spatial learning and memory, which has been reported to occur in humans, non-human primates, and rodents (Barnes 1979; Gage et al., 1984; de Toledo-Morrell et al., 1984; Barnes and McNaughton, 1985; Gallagher and Pelleymonter, 1988; Uttl and Graf, 1993; Rapp and Gallagher, 1996; Wilkniss et al., 1997). While the dentate gyrus exhibits several physiological characteristics that are believed to be critical for spatial learning and memory, including burst discharges during theta (Rose et al., 1983; Buzsaki and Czeh, 1992), phase precession (Skaggs et al., 1996), and reactivation during sleep (Shen et al., 1998), and data indicate that alternations of dentate intrinsic properties or connectivity can have detrimental effects on learning and memory, it is not entirely clear whether aging-related changes in dentate circuitry and function contribute to the decline in cognitive function seen with age. There is good reason to define the potential role that the dentate gyrus plays in aging-related cognitive decline, because recent data suggest that aging

is not only associated with a decline in memory retention, but also a reduced capacity to encode new information rapidly (Tanila et al., 1997; Rosenzweig and Barnes, 2003; Wilson et al., 2005), and encoding depends on the entorhinal–dentate gyrus–CA3 pathway (e.g., Lee and Kesner, 2004; Jerman et al., 2006; Rolls and Kesner, 2006). Recent data reveal that aged rodents with a decreased ability to encode new spatial information exhibit an increase in spontaneous activity in CA3 (Wilson et al., 2005). It has been suggested that this increase in basal activity in CA3 decreases the signal-to-noise ratio by increasing the ‘‘noise’’, and thus obscures the influx of new pertinent spatial information. The hyperexcitability observed in CA3 in aged rodents during and after 5 Hz perforant path stimulation (Patrylo et al., 2007) could be an in vitro equivalent of the increase in basal CA3 activity reported in vivo by Wilson et al. (2005). Consequently, we postulate that the aging-related disruption in dentate gating, an effect that occurs in a fraction of aged subjects, contributes to a decreased ability to encode new information and thus the compromise in spatial learning and memory seen in a subpopulation of aged subjects. This increase in basal activity could also affect cognitive function negatively by changing the capacity for synaptic plasticity, due to a metaplastic shift in network dynamics (Abraham and Bear, 1996).

The dentate gyrus, seizure susceptibility, and aging Epidemiologic studies indicate that aging is also associated with an increase in the incidence and prevalence for seizure disorders (Loiseau et al., 1990; Hauser et al., 1991, 1996; Hauser, 1992). Indeed, seizure disorders are the third most common neurological disorder seen in the elderly, with up to 5% of individuals Z65 years of age affected (Tallis et al., 1991). The most common type of seizure observed in this age group is complex-partial in nature (Hauser, 1997; Rowan et al., 2005), and it is well established that limbic structures (i.e., hippocampus) are frequently involved. Furthermore, 50% of newly identified seizure disorders in the elderly have no identifiable antecedent (i.e., are cryptogenic), which has led to the suggestion that

690

aging itself may be epileptogenic (Ng et al., 1985; Hauser et al., 1996; Hauser, 1997). Data from laboratory animals support and extend these clinical observations with an enhanced susceptibility for hippocampal-dependent seizures seen in aged rodents. For example, aged rodents exhibit longer duration afterdischarges during kindling, and faster seizure propagation to the contralateral hemisphere (Chiba et al., 1992). The rate of kindling is slower in aged rodents however (de Toledo-Morrell et al., 1984; Chiba et al., 1992), although this is likely to reflect a decreased capacity for plasticity during aging, rather than a decreased susceptibility to seizures, since the progression of kindling is dependent on NMDA receptor-mediated plasticity (Mody et al., 1988; Sayin et al., 1999) and decreases in NMDA receptormediated properties occur with age (e.g., Rao et al., 1994; Barnes et al., 2000; Magnusson et al., 2006). Aged rodents also demonstrate increased seizure susceptibility when tested with the convulsant kainic acid. Systemic treatment with kainic acid results in exacerbated pre-seizure behavioral manifestations in aged rodents, compared to adults (e.g., wet-dog shakes), and a decreased latency to onset for limbic seizures and sustained seizures (Darbin et al., 2004). Since the dentate gyrus plays a role in generating wet-dog shakes (Damiano and Connor, 1984; Frush and McNamara, 1986; Barnes and Mitchell, 1990; Grimes et al., 1990; Mitchell et al., 1990), and regulating limbic seizures (Dasheiff and McNamara, 1982; Lothman et al., 1992; Grecksch et al., 1995), it is plausible that a breakdown of dentate filter function results in an enhanced propensity for dentate–CA3 activation and thus an increased susceptibility to seizures. Support for this hypothesis comes from two sources. First, the mossy fiber–CA3 synapse is preferentially susceptible to kainic acid (Ben-Ari and Cossart, 2000), and preliminary data suggest that aged hippocampal slices exhibit a lower threshold for kainic acid-induced epileptiform activity (Willingham et al., 2006). Second, a similar, although more severe, disruption of dentate circuitry and gating has been observed in the dentate gyrus in models of epilepsy that are associated with hippocampal activation (e.g., Behr et al., 1998; Finnerty et al., 2001; Nadler, 2003). Taken together, the data from

kindling and kainic acid models of epilepsy suggest that aging-related changes in dentate circuitry and function are likely to contribute to the increased incidence for complex-partial seizures with age. Conclusions In conclusion, extensive changes occur in the dentate gyrus at the structural and functional levels during aging, and we suggest that one critical consequence is the breakdown in filter function. Changes in dentate gyrus gating could explain the increase in hippocampal-dependent seizure susceptibility that occurs during the aging process. These, and additional effects in associated structures such as area CA3, may also contribute to the decline in cognitive function with age. However, the precise role of the dentate gyrus in age-related functional changes (i.e., whether it is the cause or the effect of the increase in seizures susceptibility or cognitive decline) remains to be addressed completely. Acknowledgments Supported by a grant from the NIH (AG00795) and the Epilepsy Foundation through the generous support of the CDC (PRP). References Aamodt, S.M., Shi, J., Colonnese, M.T., Veras, W. and Constantine-Paton, M. (2000) Chronic NMDA exposure accelerates development of GABAergic inhibition in the superior colliculus. J. Neurophysiol., 83: 1580–1591. Abraham, W.C. and Bear, M.F. (1996) Metaplasticity: the plasticity of synaptic plasticity. Trends Neurosci., 19: 126–130. Acsady, L., Kamondi, A., Sik, A., Freund, T. and Buzsaki, G. (1998) GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci., 18: 3386–3403. Akhtar, N.D. and Land, P.W. (1991) Activity-dependent regulation of glutamic acid decarboxylase in the rat barrel cortex: effects of neonatal versus adult sensory deprivation. J. Comp. Neurol., 307: 200–213. Almaguer-Melian, W., Cruz-Aguado, R., Riva Cde, L., Kendrick, K.M., Frey, J.U. and Bergado, J. (2005) Effect of LTPreinforcing paradigms on neurotransmitter release in the dentate gyrus of young and aged rats. Biochem. Biophys. Res. Commun., 327: 877–883.

691 de Almeida, O.M., Gardino, P.F., Loureiro dos Santos, N.E., Yamasaki, E.N., de Mello, M.C., Hokoc, J.N. and de Mello, F.G. (2002) Opposite roles of GABA and excitatory amino acids on the control of GAD expression in cultured retina cells. Brain Res., 925: 89–99. Andersen, P. (1975) Organization of hippocampal neurons and their interconnections. In: Isaacson R.L. and Pribram K.H. (Eds.), The Hippocampus, Vol. 1. Plenum Press, New York, pp. 155–175. Aubert, I., Rowe, W., Meaney, M.J., Gauthier, S. and Quirion, R. (1995) Cholinergic markers in aged cognitively impaired Long-Evans rats. Neuroscience, 67: 277–292. Austin, K.B., Bronzino, J.D. and Morgane, P.J. (1989) Pairedpulse facilitation and inhibition in the dentate gyrus is dependent on behavioral state. Exp. Brain Res., 77(3): 594–604. Bai, L., Hof, P.R., Standaert, D.G., Xing, Y., Nelson, S.E., Young, A.B. and Magnusson, K.R. (2004) Changes in the expression of the NR2B subunit during aging in macaque monkeys. Neurobiol. Aging, 25: 201–208. Barnes, C.A. (1979) Memory deficits associated with senescence: a neurophysiological and behavioral study in the rat. J. Comp. Psychol., 93: 74–104. Barnes, C.A. (2000) Plasticity in the aging central nervous system. Int. Rev. Neurobiol., 45: 339–354. Barnes, C.A. and McNaughton, B.L. (1980a) Physiological compensation for loss of afferent synapses in rat hippocampal granule cells during senescence. J. Physiol. Lond., 309: 473–485. Barnes, C.A. and McNaughton, B.L. (1980b) Spatial memory and hippocampal synaptic plasticity in middle-aged and senescent rats. In: Stein D. (Ed.), Psychobiology of Aging: Problems and Perspectives. Elsevier/North-Holland, New York, pp. 253–272. Barnes, C.A. and McNaughton, B.L. (1983) Excitability changes in hippocampal granule cells of senescent rat. In: Changeux J.P., Glowinski J., Imbert M. and Bloom F.E. (Eds.), Molecular and Cellular Interactions Underlying Higher Brain Function, Progress in Brain Research. Elsevier, New York, pp. 445–451. Barnes, C.A. and McNaughton, B.L. (1985) An age-comparison of the rates of acquisition and forgetting of spatial information in relation to long-term enhancement of hippocampal synapses. Behav. Neurosci., 6: 563–571. Barnes, C.A., Rao, G. and Houston, F.P. (2000) LTP induction threshold change in old rats at the perforant path — granule cell synapse. Neurobiol. Aging, 21: 613–620. Barnes, C.A., Rao, G. and McNaughton, B.L. (1987) Increased electrotonic coupling in aged rat hippocampus: a possible mechanism for cellular excitability changes. J. Comp. Neurol., 259: 549–558. Barnes, M.I. and Mitchell, C.L. (1990) Differential effects of colchicine lesions of dentate granule cells on wet dog shakes and seizures elicited by direct hippocampal stimulation. Physiol. Behav., 48: 131–138. Beck, H., Goussakov, I.V., Lie, A., Helmstaedter, C. and Elger, C.E. (2000) Synaptic plasticity in the human dentate gyrus. J. Neurosci., 20: 7080–7086.

Behr, J., Lyson, K.J. and Mody, I. (1998) Enhanced propagation of epileptiform activity through the kindled dentate gyrus. J. Neurophysiol., 79: 1726–1732. Ben-Ari, Y. and Cossart, R. (2000) Kainate, a double agent that generates seizures: two decades of progress. Trends Neurosci., 23: 580–587. Bernasconi-Guastalla, S., Wolfer, D.P. and Lipp, H.P. (1994) Hippocampal mossy fibers and swimming navigation in mice: correlations with size and left-right asymmetries. Hippocampus, 4: 53–63. Bragin, A., Jando, G., Nadasdy, Z., van Landeghem, M. and Buzsaki, G. (1995) Dentate EEG spikes and associated interneuronal population bursts in the hippocampal hilar region of the rat. J. Neurophysiol., 73: 1691–1705. Buckmaster, P.S. and Dudek, F.E. (1999) In vivo intracellular analysis of granule cell axon reorganization in epileptic rats. J. Neurophysiol., 81: 712–721. Buckmaster, P.S., Wenzel, H.J., Kunkel, D.D. and Schwartzkroin, PA. (1996) Axon arbors and synaptic connections of hippocampal mossy cells in the rat in vivo. J. Comp. Neurol., 366: 271–292. Buzsaki, G. and Czeh, G. (1992) Physiological function of granule cells: a hypothesis. Epilepsy Res. Suppl., 7: 281–290. Cadacio, C.L., Milner, T.A., Gallagher, M. and Pierce, J.P. (2003) Hilar neuropeptide Y interneuron loss in the aged rat hippocampal formation. Exp. Neurol., 183: 147–158. Calhoun, M.E., Mao, Y., Roberts, J.A. and Rapp, P.R. (2004) Reduction in hippocampal cholinergic innervation is unrelated to recognition memory impairment in aged rhesus monkeys. J. Comp. Neurol., 475: 238–246. Carnahan, J. and Nawa, H. (1995) Regulation of neuropeptide expression in the brain by neurotrophins. Potential role in vivo. Mol. Neurobiol., 10: 135–149. Cassell, M.D. and Brown, M.W. (1984) The distribution of Timm’s stain in the nonsulphide-perfused human hippocampal formation. J. Comp. Neurol., 222: 461–471. Chiba, S., Muneoka, Y., Sato, Y. and Miyagishi, T. (1992) Seizure susceptibility in aged rats — pentylenetetrazol-induced seizures and amygdaloid kindling seizures. No To Shinkei [Brain Nerve], 44: 559–564. Christian, E.P. and Dudek, F.E. (1988) Characteristics of local excitatory circuits studied with glutamate microapplication in the CA3 area of rat hippocampal slices. J. Neurophysiol., 59: 90–109. Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E. (1992) Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers. Brain Res., 573: 305–310. Crusio, W.E., Schwegler, H. and Lipp, H.P. (1987) Radialmaze performance and structural variation of the hippocampus in mice: a correlation with mossy fibre distribution. Brain Res., 425: 182–185. Damiano, B.P. and Connor, J.D. (1984) Hippocampal mediation of shaking behavior induced by electrical stimulation of the perforant path in the rat. Brain Res., 308: 383–386.

692 Darbin, O., Naritoku, D. and Patrylo, P.R. (2004) Aging alters electroencephalographic and clinical manifestations of kainate-induced status epilepticus. Epilepsia, 45: 1219–1227. Dasheiff, R.M. and McNamara, J.O. (1982) Intradentate colchicine retards the developments of amygdala kindling. Ann. Neurol., 11: 347–352. Dore, S., Kar, S., Rowe, W. and Quirion, R. (1997) Distribution and levels of [125I]IGF-I, [125I]IGF-II and [125I]insulin receptor binding sites in the hippocampus of aged memoryunimpaired and -impaired rats. Neuroscience, 80: 1033–1040. Finnerty, G.T., Whittington, M.A. and Jefferys, J.G. (2001) Altered dentate filtering during the transition to seizure in the rat tetanus toxin model of epilepsy. J. Neurophysiol., 86: 2748–2753. Fisher, J.L. (2004) The alpha 1 and alpha 6 subunit subtypes of the mammalian GABA(A) receptor confer distinct channel gating kinetics. J. Physiol., 561: 433–448. Freund, T.F. and Buzsa´ki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of dentate gyrus granule cells. J. Neurophysiol., 51: 195–209. Frush, D.P. and McNamara, J.O. (1986) Evidence implicating dentate granule cells in wet dog shakes produced by kindling stimulations of entorhinal cortex. Exp. Neurol., 92: 102–113. Gage, F.H., Bjorklund, A., Stenevi, U., Dunnett, S.B. and Kelly, P.A.T. (1984) Intrahippocampal septal grafts ameliorate learning impairments in aged rats. Science, 225: 533–536. Gallagher, M. and Pelleymonter, M.A. (1988) Spatial learning deficits in old rats: a model for memory decline in the aged. Neurobiol. Aging, 9: 549–556. Gazzaley, A.H., Siegel, S.J., Kordower, J.H., Mufson, E.J. and Morrison, J.H. (1996) Circuit-specific alterations of N-methyl-D-aspartate receptor subunit 1 in the dentate gyrus of aged monkeys. Proc. Natl. Acad. Sci. U.S.A., 93: 3121–3125. Ge, S., Goh, E.L., Sailor, K.A., Kitabatake, Y., Ming, G.L. and Song, H. (2006) GABA regulates synaptic integration of newly generated neurons in the adult brain. Nature, 439: 589–593. Geinisman, Y., de Toledo-Morrell, L., Morrell, F., Persina, I.S. and Rossi, M. (1992) Age-related loss of axospinous synapses formed by two afferent systems in the rat dentate gyrus as revealed by the unbiased stereological dissector technique. Hippocampus, 2: 437–444. Gooney, M., Messaoudi, E., Maher, F.O., Bramham, C.R. and Lynch, M.A. (2004) BDNF-induced LTP in dentate gyrus is impaired with age: analysis of changes in cell signaling events. Neurobiol. Aging, 25: 1323–1331. Grecksch, G., Ruethrich, H., Bernstein, H.G. and Becker, A. (1995) PTZ-kindling after colchicine lesion in the dentate gyrus of the rat hippocampus. Physiol. Behav., 58: 695–698. Grimes, L.M., Earnhardt, T.S., Mitchell, C.L., Tilson, H.A. and Hong, J.S. (1990) Granule cells in the ventral, but not dorsal, dentate gyrus are essential for kainic acid-induced wet dog shakes. Brain Res., 514: 167–170. Gutierrez, A., Khan, Z.U., Ruano, D., Miralles, C.P., Vitorica, J. and De Blas, A.L. (1996) Aging-related subunit expression

changes of the GABAA receptor in the rat hippocampus. Neuroscience, 74: 341–348. Halasy, K. and Somogyi, P. (1993) Distribution of GABAergic synapses and their targets in the dentate gyrus of rat: a quantitative immunoelectron microscopic analysis. J. Hirnforsch., 34: 299–308. Hattiangady, B., Rao, M.S., Shetty, G.A. and Shetty, A.K. (2005) Brain-derived neurotrophic factor, phosphorylated cyclic AMP response element binding protein and neuropeptide Y decline as early as middle age in the dentate gyrus and CA1 and CA3 subfields of the hippocampus. Exp. Neurol., 195: 353–371. Hauser, W.A. (1992) Seizure disorders: the changes with age. Epilepsia, 33: S6–S14. Hauser, W.A. (1997) Epidemiology of seizures and epilepsy in the elderly. In: Rowan A.J. and Ramsay R.E. (Eds.), Seizures and Epilepsy in the Elderly. Butterworth-Heinemann, Boston, MA, pp. 7–18. Hauser, W.A., Annegers, J.F. and Kurland, L.T. (1991) Prevalence of epilepsy in Rochester, Minnesota. Epilepsia, 32: 429–445. Hauser, W.A., Annegers, J.F. and Rocca, W.A. (1996) Descriptive epidemiology of epilepsy: contributions of population-based studies from Rochester, Minnesota. Mayo Clin. Proc., 71: 576–586. Hayashi, M., Mistunaga, F., Ohira, K. and Shimizu, K. (2001) Changes in BDNF-immunoreactive structures in the hippocampal formation of the aged macaque monkey. Brain Res., 918: 191–196. Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and Zhang, C.L. (1992) The dentate gyrus as a regulated gate for the propagation of epileptiform activity. Epilepsy Res. Suppl., 7: 273–280. Henze, D.A., Wittner, L. and Buzsaki, G. (2002) Single granule cells reliably discharge targets in the hippocampal CA3 network in vivo. Nat. Neurosci., 5: 790–795. Herman, J.P., Chen, K.C., Booze, R. and Landfield, P.W. (1998) Up-regulation of alpha1D Ca2+ channel subunit mRNA expression in the hippocampus of aged F344 rats. Neurobiol. Aging, 19: 581–587. Hwang, I.K., Kim, D.W., Yoo, K.Y., Kim, D.S., Kim, K.S., Kang, J.H., Choi, S.Y., Kim, Y.S., Kang, T.C. and Won, M.H. (2004) Age-related changes of the gamma-aminobutyric acid transaminase immunoreactivity in the hippocampus and dentate gyrus of the Mongolian gerbil. Brain Res., 1017: 77–84. Hyman, C., Juhasz, M., Jackson, C. and Wright, P. (1994) Ip NY, Lindsay RM. Overlapping and distinct actions of the neurotrophins BDNF, NT-3, and NT-4/5 on cultured dopaminergic and GABAergic neurons of the ventral mesencephalon. J. Neurosci., 14: 335–347. Ishizuka, N., Weber, J. and Amaral, D.G. (1990) Organization of intrahippocampal projections originating from CA3 pyramidal cells in the rat. J. Comp. Neurol., 295: 580–623. Jerman, T., Kesner, R.P. and Hunsaker, M.R. (2006) Disconnection analysis of CA3 and DG in mediating encoding but

693 not retrieval in a spatial maze learning task. Learn. Mem., 13: 458–464. Johnson, D. and Brown, T.H. (1984) Mechanisms of neuronal burst generation. In: Schwartzkroin P.A. and Wheal H.V. (Eds.), Electrophysiology of Epilepsy. Academic Press, New York, pp. 277–301. Katoh-Semba, R., Semba, R., Takeuchi, I.K. and Kato, K. (1998) Age-related changes in levels of brain-derived neurotrophic factor in selected brain regions of rats, normal mice and senescence-accelerated mice: a comparison to those of nerve growth factor and neurotrophin-3. Neurosci. Res., 31: 227–234. Katona, I., Acsady, L. and Freund, T.F. (1999) Postsynaptic targets of somatostatin-immunoreactive interneurons in the rat hippocampus. Neuroscience, 88: 37–55. Kelley, M.S. and Steward, O. (1996) The process of reinnervation in the dentate gyrus of adult rats: physiological events at the time of the lesion and during the early postlesion period. Exp. Neurol., 139: 73–82. Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition of dentate granule cells in a model of temporal lob epilepsy. J. Neurosci., 23: 2440–2452. Kugler, P. and Baier, G. (1992) Mitochondrial enzymes related to glutamate and GABA metabolism in the hippocampus of young and aged rats: a quantitative histochemical study. Neurochem. Res., 17: 179–185. Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Neurogenesis in the dentate gyrus of the adult rat: age-related decrease of neuronal progenitor proliferation. J. Neurosci., 16: 2027–2033. Lawrence, J.J. and McBain, C.J. (2003) Interneuron diversity series: containing the detonation — feedforward inhibition in the CA3 hippocampus. Trends Neurosci., 26: 631–640. Lee, I. and Kesner, R.P. (2004) Encoding versus retrieval of spatial memory: double dissociation between the dentate gyrus and the perforant path inputs into CA3 in the dorsal hippocampus. Hippocampus, 14: 66–76. Leranth, C., Malcolm, A.J. and Frotscher, M. (1990) Afferent and efferent synaptic connections of somatostatin-immunoreactive neurons in the rat fascia dentate. J. Comp. Neurol., 295: 111–122. Li, X.G., Somogyi, P., Ylinen, A. and Buzsaki, G. (1994) The hippocampal CA3 network: an in vivo intracellular labeling study. J. Comp. Neurol., 339: 181–208. Ling, L.L., Hughes, L.F. and Caspary, D.M. (2005) Age-related loss of the GABA synthetic enzyme glutamic acid decarboxylase in rat primary auditory cortex. Neuroscience, 132: 1103–1113. Lipp, H.-P., Schwegler, H. and Driscoll, P. (1984) Postnatal modification of hippocampal circuitry alters avoidance learning in adult rats. Science, 225: 80–82. Loiseau, J., Louiseau, P., Duche, B., Guyot, M., Dartigues, J.F. and Aublet, B. (1990) A survey of epileptic disorders in southwest France: seizures in elderly patients. Ann. Neurol., 27: 232–237.

Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The dentate gyrus as a control point for seizures in the hippocampus and beyond. Epilepsy Res. Suppl., 7: 301–313. Luebke, J.I. and Rosene, D.L. (2003) Aging alters dendritic morphology, input resistance, and inhibitory signaling in dentate granule cells of the rhesus monkey. J. Comp. Neurol., 460: 573–584. Lukoyanov, N.V., Andrade, J.P., Dulce Madeira, M. and Paula-Barbosa, M.M. (1999) Effects of age and sex on the water maze performance and hippocampal cholinergic fibers in rats. Neurosci. Lett., 269: 141–144. Lumeng, L. and Li, T.K. (1974) Vitamin B6 metabolism in chronic alcohol abuse. Pyridoxal phosphate levels in plasma and the effects of acetaldehyde on pyridoxal phosphate synthesis and degradation in human erythrocytes. J. Clin. Invest., 53: 693–704. Lynch, G., Gall, C., Rose, G. and Cotman, C. (1976) Changes in the distribution of the dentate gyrus associational system following unilateral or bilateral entorhinal lesions in the adult rat. Brain Res., 110: 57–71. Lynch, M., Sayin, U., Golarai, G. and Sutula, T. (2000) NMDA receptor-dependent plasticity of granule cell spiking in the dentate gyrus of normal and epileptic rats. J. Neurophysiol., 84: 2868–2879. Maccaferri, G., Roberts, J.D., Szucs, P., Cottingham, C.A. and Somogyi, P. (2000) Cell surface domain specific postsynaptic currents evoked by identified GABAergic neurons in rat hippocampus in vitro. J. Physiol., 524: 91–116. Magnusson, K.R. (2000) Declines in mRNA expression of different subunits may account for differential effects of aging on agonist and antagonist binding to the NMDA receptor. J. Neurosci., 20: 1666–1674. Magnusson, K.R., Kresge, D. and Supon, J. (2006) Differential effects of aging on NMDA receptors in the intermediate versus the dorsal hippocampus. Neurobiol. Aging, 27: 324–333. Marty, S. (2000) Differences in the regulation of neuropeptide Y, somatostatin and parvalbumin levels in hippocampal interneurons by neuronal activity and BDNF. Prog. Brain Res., 128: 1. McNaughton, B.L., Barnes, C.A., Meltzer, J. and Sutherland, R.J. (1989) Hippocampal granule cells are necessary for normal spatial learning but not for spatially-selective pyramidal cell discharge. Exp. Brain Res., 76: 485–496. Merrill, D.A., Chiba, A.A. and Tuszynski, M.H. (2001) Conservation of neuronal number and size in the entorhinal cortex of behaviorally characterized aged rats. J. Comp. Neurol., 438: 445–456. Merrill, D.A., Karim, R., Darraq, M., Chiba, A.A. and Tuszynski, M.H. (2003) Hippocampal cell genesis does not correlate with spatial learning ability in aged rats. J. Comp. Neurol., 459: 201–207. Miles, R. and Wong, R.K.S. (1987) Inhibitory control of local excitatory circuits in the guinea pig hippocampus. J. Physiol., 388: 611–629. Mitchell, C.L., Grimes, L.M. and Hong, J.S. (1990) Granule cells in the ventral dentate gyrus are essential for kainic

694 acid-induced wet dog shakes but not those induced by precipitated abstinence in morphine-dependent rats. Brain Res., 511(2): 338–340. Mody, I., Stanton, P.K. and Heinemann, U. (1988) Activation of N-methyl-D-aspartate receptors parallels changes in cellular and synaptic properties of dentate gyrus granule cells after kindling. J. Neurophysiol., 59: 1033–1054. Moser, E.I. (1996) Altered inhibition of dentate granule cells during spatial learning in an exploration task. J. Neurosci., 16: 1247–1259. Mott, D.D., Turner, D.A., Okazaki, M.M. and Lewis, D.V. (1997) Interneurons of the dentate-hilus border of the rat dentate gyrus: morphological and electrophysiological heterogeneity. J. Neurosci., 17: 3990–4005. Nadler, J.V. (2003) The recurrent mossy fiber pathway of the epileptic brain. Neurochem. Res., 28: 1649–1658. Nagahara, A.H., Nicolle, M.M. and Gallagher, M. (1993) Alterations in [3H]-kainate receptor binding in the hippocampal formation of aged Long-Evans rats. Hippocampus, 3: 269–277. Nanry, K.P., Mundy, W.R. and Tilson, H.A. (1989) Colchicineinduced alterations of reference memory in rats: role of spatial versus non-spatial task components. Behav. Brain Res., 35: 45–53. Ng, S.K., Hauser, W.A., Brust, J.C.M., Healton, E.B. and Susser, M.W. (1985) Risk factors for adult-onset first seizures. Ann. Neurol., 18: 153. Nicolle, M.M., Bizon, J.L. and Gallagher, M. (1996) In vitro autoradiography of ionotropic glutamate receptors in hippocampus and striatum of aged Long-Evans rats: relationship to spatial learning. Neuroscience, 74: 741–756. Niesen, C.E., Baskys, A. and Carlen, P.L. (1988) Reversed ethanol effects on potassium conductances in aged hippocampal dentate granule neurons. Brain Res., 445: 137–141. Orr, G., Rao, G., Houston, F.P., McNaughton, B.L. and Barnes, C.A. (2001) Hippocampal synaptic plasticity is modulated by theta rhythm in the fascia dentata of adult and aged freely behaving rats. Hippocampus, 11: 647–654. Pagliusi, S.R., Gerrard, P., Abdallah, M., Talabot, D. and Catsicas, S. (1994) Age-related changes in expression of AMPA-selective glutamate receptor subunits: is calcium-permeability altered in hippocampal neurons? Neuroscience, 61: 429–433. Patel, Y.C., Liu, J.L., Warszynska, A., Kent, G., Papachristou, D.N. and Patel, S.C. (1995) Differential stimulation of somatostatin but not neuropeptide Y gene expression by quinolinic acid in cultured cortical neurons. J. Neurochem., 65: 998–1006. Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking of new glutamatergic pathways in the dentate gyrus of hippocampal slices from kainate-induced epileptic rats. J. Neurophysiol., 79: 418–429. Patrylo, P.R., Tyagi, I., Willingham, A.L., Lee, S. and Williamson, A. (2007) Dentate filter function is altered in a proepileptic fashion during aging. Epilepsia (in press) doi: 10.1111/j.1528-1167.2007.01139.X

Paulsen, O. and Moser, E.I. (1998) A model of hippocampal memory encoding and retrieval: GABAergic control of synaptic plasticity. Trends Neurosci., 21: 273–278. Penttonen, M., Kamondi, A., Sik, A., Acsady, L. and Buzsaki, G. (1997) Feed-forward and feed-back activation of the dentate gyrus in vivo during dentate spikes and sharp wave bursts. Hippocampus, 7: 437–450. Rao, G., Barnes, C.A., McNaughton, B.L. and Shen, J. (1994) Age-related decrease in the NMDA-mediated EPSP in FD. Soc. Neurosci. Abstr., 20: 1207. Rao, M.S., Hattiangady, B., Abdel-Rahman, A., Stanley, D.P. and Shetty, A.K. (2005) Newly born cells in the ageing dentate gyrus display normal migration, survival and neuronal fate choice but endure retarded early maturation. Eur. J. Neurosci., 21: 464–476. Rapp, P.R., Deroche, P.S., Mao, Y. and Burwell, R.D. (2002) Neuron number in the parahippocampal region is preserved in aged rats with spatial learning deficits. Cereb. Cortex, 12: 1171–1179. Rapp, P.R. and Gallagher, M. (1996) Preserved neuron number in the hippocampus of aged rats with spatial learning deficits. Proc. Natl. Acad. Sci., 93: 9926–9930. Rapp, P.R., Stack, E.C. and Gallagher, M. (1999) Morphometric studies of the aged hippocampus: I. Volumetric analysis in behaviorally characterized rats. J. Comp. Neurol., 403: 459–470. Rasmussen, T., Schliemann, T., Sorensen, J.C., Zimmer, J. and West, M.J. (1996) Memory impaired aged rats: no loss of principal hippocampal and subicular neurons. Neurobiol. Aging, 17: 143–147. Reeves, T.M. and Smith, D.C. (1987) Reinnervation of the dentate gyrus and recovery of alternation behavior following entorhinal cortex lesions. Behav. Neurosci., 101: 179–186. Reibel, S., Vivien-Roels, B., Le, B.T., Larmet, Y., Carnahan, J., Marescaux, C. and Depaulis, A. (2000) Overexpression of neuropeptide Y induced by brain-derived neurotrophic factor in the rat hippocampus is long lasting. Eur. J. Neurosci., 12: 595–605. Reynolds, J.N. and Carlen, P.L. (1989) Diminished calcium currents in aged hippocampal dentate gyrus granule neurones. Brain Res., 479: 384–390. Ribak, C.E. (1992) Local circuitry of GABAergic basket cells in the dentate gyrus. Epilepsy Res. Suppl., 7: 29–47. Ribak, C.E., Seress, L. and Leranth, C. (1993) Electron microscopic immunocytochemical study of the distribution of parvalbumin-containing neurons and axon terminals in the primate dentate gyrus and Ammon’s horn. J. Comp. Neurol., 327: 298–321. Ribak, C.E., Tran, P.H., Spigelman, I., Okazaki, M.M. and Nadler, J.V. (2000) Status epilepticus-induced hilar basal dendrites on rodent granule cells contribute to recurrent excitatory circuitry. J. Comp. Neurol., 428: 240–253. Rissman, R.A., Nocera, R., Fuller, L.M., Kordower, J.H. and Armstrong, D.M. (2006) Age-related alterations in GABA(A) receptor subunits in the nonhuman primate hippocampus. Brain Res., 1073-1074: 120–130.

695 Rogers, S.W., Gahring, L.C., Collins, A.C. and Marks, M. (1998) Age-related changes in neuronal nicotinic acetylcholine receptor subunit alpha4 expression are modified by longterm nicotine administration. J. Neurosci., 18: 4825–4832. Rolls, E.T. and Kesner, R.P. (2006) A computational theory of hippocampal function, and empirical tests of the theory. Prog. Neurobiol., 79: 1–48. Rose, G., Diamond, D. and Lynch, G.S. (1983) Dentate granule cells in the rat hippocampal formation have the behavioral characteristics of theta neurons. Brain Res., 266: 29–37. Rosenzweig, E.S. and Barnes, C.A. (2003) Impact of aging on hippocampal function: plasticity, network dynamics, and cognition. Prog. Neurobiol., 69: 143–179. Rowan, A.J., Ramsay, R.E., Collins, J.F., Pryor, F., Boardman, K.D., Uthman, B.M., Spitz, M., Frederick, T., Towne, A., Carter, G.S., Marks, W., Felicetta, J., Tomyanovich, M.L. and VA Cooperative Study 428 Group. (2005) New onset geriatric epilepsy: a randomized study of gabapentin, lamotrigine, and carbamazepine. Neurology, 64: 1868–1873. Ruano, D., Benavides, J., Machado, A. and Vitorica, J. (1995) Aging-associated changes in the pharmacological properties of the benzodiazepine (omega) receptor isotypes in the rat hippocampus. J. Neurochem., 64: 867–873. Sayin, U¨., Rutecki, P. and Sutula, T. (1999) NMDA-dependent currents in granule cells of the dentate gyrus contribute to induction but not permanence of kindling. J. Neurophysiol., 81: 564–574. Scharfman, H.E. (1994) Synchronization of area CA3 hippocampal pyramidal cells and non-granule cells of the dentate gyrus in bicuculline-treated rat hippocampal slices. Neuroscience, 59: 245–257. Schuster, G., Cassel, J.C. and Will, B. (1997) Comparison of the behavioral and morphological effects of colchicine- or neutral fluid-induced destruction of granule cells in the dentate gyrus of the rat. Neurobiol. Learn. Mem., 68: 86–91. Schwartzkroin, P.A. and Prince, D.A. (1978) Cellular and field potential properties of epileptogenic hippocampal slices. Brain Res., 147: 117–130. Shen, J. and Barnes, C.A. (1996) Age-related decrease in cholinergic synaptic transmission in three hippocampal subfields. Neurobiol. Aging, 17: 439–451. Shen, J., Kudrimoti, H.S., McNaughton, B.L. and Barnes, C.A. (1998) Reactivation of neuronal ensembles in hippocampal dentate gyrus during sleep after spatial experience. J. Sleep Res., 7: 6–16. Shetty, A.K. and Turner, D.A. (1998) Hippocampal interneurons expressing glutamic acid decarboxylase and calciumbinding proteins decrease with aging in Fischer 344 rats. J. Comp. Neurol., 394: 252–269. Skaggs, W.E., McNaughton, B.L., Wilson, M.A. and Barnes, C.A. (1996) Theta phase precession in hippocampal neuronal populations and the compression of temporal sequences. Hippocampus, 6: 149–172. Smith, T.D., Adams, M.M., Gallagher, M., Morrison, J.H. and Rapp, P.R. (2000) Circuit-specific alterations in hippocampal

synaptophysin immunoreactivity predict spatial learning impairment in aged rats. J. Neurosci., 20: 6587–6593. Solte`sz, I., Smetters, D.K. and Mody, I. (1995) Tonic inhibition originates from synapses close to the soma. Neuron, 14: 1273–1283. Sonntag, W.E., Lynch, C.D., Bennett, S.A., Khan, A.S., Thornton, P.L., Cooney, P.T., Ingram, R.L., McShane, T. and Brunso-Bechtold, J.K. (1999) Alterations in insulin-like growth factor-1 gene and protein expression and type 1 insulin-like growth factor receptors in the brains of ageing rats. Neuroscience, 88: 269–279. Soriano, E. and Frotscher, M. (1993) GABAergic innervation of the rat fascia dentata: a novel type of interneuron in the granule cell layer with extensive axonal arborization in the molecular layer. J. Comp. Neurol., 334: 385–396. Soriano, E., Nitsch, R. and Frotscher, M. (1990) Axo-axonic chandelier cells in the rat fascia dentata: golgi-electron microscopy and immunocytochemical studies. J. Comp. Neurol., 293: 1–25. Spigelman, I., Yan, X.X., Obenaus, A., Lee, E.Y., Wasterlain, C.G. and Ribak, C.E. (1998) Dentate granule cells form novel basal dendrites in a rat model of temporal lobe epilepsy. Neuroscience, 86: 109–120. Staley, K.J., Otis, T.S. and Mody, I. (1992) Membrane properties of dentate gyrus granule cells: comparison of sharp microelectrode and whole-cell recordings. J. Neurophysiol., 67: 1346–1358. Stanley, D.P. and Shetty, A.K. (2004) Aging in the rat hippocampus is associated with widespread reductions in the number of glutamate decarboxylase-67 positive interneurons but not interneuron degeneration. J. Neurochem., 89: 204–216. Steward, O. (1982) Assessing the functional significance of lesion-induced neuronal plasticity. Int. Rev. Neurobiol., 23: 197–254. Steward, O. (1992) Lesion-induced synapse reorganization in the hippocampus of cats: sprouting of entorhinal, commissural/associational, and mossy fiber projections after unilateral entorhinal cortex lesions, with comments on the normal organization of these pathways. Hippocampus, 2: 247–268. Steward, O., Loesche, J. and Horton, W.C. (1976) Behavioral correlates of denervation and reinnervation of the hippocampal formation of the rat: open field activity and cue utilization following bilateral entorhinal cortex lesions. Brain Res. Bull., 2: 41–48. Sykova, E., Mazel, T. and Simonova, Z. (1998) Diffusion constraints and neuron-glia interaction during aging. Exp. Gerontol., 33: 837–851. Tallis, R., Hall, G., Craig, I. and Dean, A. (1991) How common are epileptic seizures in old age? Age Ageing, 20: 442–448. Tanila, H., Shapiro, M., Gallagher, M. and Eichenbaum, H. (1997) Brain aging: changes in the nature of information coding by the hippocampus. J. Neurosci., 17: 5155–5166. Tayebati, S.K., Amenta, F., El-Assouad, D. and Zaccheo, D. (2002) Muscarinic cholinergic receptor subtypes in the hippocampus of aged rats. Mech. Ageing Dev., 123: 521–528.

696 Thibault, O., Hadley, R. and Landfield, P.W. (2001) Elevated postsynaptic [Ca2+]i and L-type calcium channel activity in aged hippocampal neurons: relationship to impaired synaptic plasticity. J. Neurosci., 21: 9744–9756. Tigges, J., Herndon, J.G. and Rosene, D.L. (1995) Mild agerelated changes in the dentate gyrus of adult rhesus monkeys. Acta Anat. (Basel), 153: 39–48. de Toledo-Morrell, L., Geinisman, Y. and Morrell, F. (1988) Age-dependent alterations in hippocampal synaptic plasticity: relation to memory disorders. Neurobiol. Aging, 9: 581–590. de Toledo-Morrell, L., Morrell, F. and Fleming, S. (1984) Agedependent deficits in spatial memory are related to impaired hippocampal kindling. Behav. Neurosci., 98: 902–907. Uttl, B. and Graf, P. (1993) Episodic spatial memory in adulthood. Psychol. Aging, 8: 257–273. Vela, J., Gutierrez, A., Vitorica, J. and Ruano, D. (2003) Rat hippocampal GABAergic molecular markers are differentially affected by ageing. J. Neurochem., 85: 368–377. Wenk, G.L. and Barnes, C.A. (2000) Regional changes in the hippocampal density of AMPA and NMDA receptors across the lifespan of the rat. Brain Res., 885: 1–5. Wenzel, H.J., Buckmaster, P.S., Anderson, N.L., Wenzel, M.E. and Schwartzkroin, P.A. (1997) Ultrastructural localization of neurotransmitter immunoreactivity in mossy cell axons and their synaptic targets in the rat dentate gyrus. Hippocampus, 7: 559–570.

West, J.R. and Dewey, S.L. (1984) Mossy fiber sprouting in the fascia dentata after unilateral entorhinal lesions: quantitative analysis using computer-assisted image processing. Neuroscience, 13: 377–384. Wilkniss, S.M., Jones, M.G., Korol, D.L., Gold, P.E. and Manning, C.A. (1997) Age-related differences in an ecologically based study of route learning. Psychol. Aging, 12: 372–375. Williams, S.R., Buhl, E.H. and Mody, I. (1998) The dynamics of synchronized neurotransmitter release determined from compound spontaneous IPSCs in rat dentate granule neurones in vitro. J. Physiol., 510: 477–497. Willingham, A., Tokhi, A., Jones, R. and Patrylo, P.R. (2006) Hippocampal slices from aged rats have an enhanced susceptibility to kainate-induced epileptiform activity. Abstract # 3.102, 60th annual meeting of the American Epilspy Society. Wilson, I.A., Ikonen, S., Gallagher, M., Eichenbaum, H. and Tanila, H. (2005) Age-associated alterations of hippocampal place cells are subregion specific. J. Neurosci., 25: 6877–6886. Wolfer, D.P. and Lipp, H.P. (1995) Evidence for physiological growth of hippocampal mossy fiber collaterals in the guinea pig during puberty and adulthood. Hippocampus, 5: 329–340. Wuarin, J.P. and Dudek, F.E. (1996) Electrographic seizures and new recurrent excitatory circuits in the dentate gyrus of hippocampal slices from kaintae-treated epileptic rats. J. Neurosci., 16: 4438–4448.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 38

Dentate gyrus neurogenesis and depression Amar Sahay1,2,, Michael R. Drew1,2 and Rene Hen1,2,3, 1

Departments of Neuroscience and Psychiatry, Columbia University, New York, NY 10032, USA 2 Division of Integrative Neuroscience, Columbia University, New York, NY 10032, USA 3 Department of Pharmacology, Columbia University, New York, NY 10032, USA

Abstract: Major depressive disorder (MDD) is a debilitating and complex psychiatric disorder that involves multiple neural circuits and genetic and non-genetic risk factors. In the quest for elucidating the neurobiological basis of MDD, hippocampal neurogenesis has emerged as a candidate substrate, both for the etiology as well as treatment of MDD. This chapter critiques the advances made in the study of hippocampal neurogenesis as they relate to the neurogenic hypothesis of MDD. While an involvement of neurogenesis in the etiology of depression remains highly speculative, preclinical studies have revealed a novel and previously unrecognized role for hippocampal neurogenesis in mediating some of the behavioral effects of antidepressants. The implications of these findings are discussed to reevaluate the role of hippocampal neurogenesis in MDD. Keywords: dentate gyrus; depression; neurogenesis; serotonin; antidepressants; hippocampus

ability to concentrate, diminished interest in pleasurable activities, daily insomnia or hypersomnia, weight loss or gain, and recurrent suicidal ideation. The diagnostic criteria for MDD convey the complexity of the disease and suggest that multiple neural circuits subserving distinct cognitive and affective processes are likely to be involved. Our comprehension of the mechanisms underlying the pathogenesis of MDD has evolved considerably since the formulation of the monoamine hypothesis (Bunney and Davis, 1965; Schildkraut, 1965; Nestler et al., 2002). The recent emphasis on neural circuits as opposed to a chemical imbalance catalyzed a fundamental shift in our conceptualization of MDD and psychiatric disorders. It provided a framework to understand how genes, through their effects on neural circuits, influence our ability to encode experience and adapt to environmental stimuli and stressors. Implicit in this idea is that genes moderate vulnerability to the

Introduction Understanding the neurobiological basis of major depressive disorder (MDD) is one of the most pressing challenges for today’s society. Severe forms of depression affect 2–5% of the U.S. population, and mood disorders impact 7% of the world’s population and rank among the top ten causes of disability (Murray and Lopez, 1996). The diagnosis of MDD based on the criteria established by the Diagnostics and Statistical Manual of Mental Disorders (American Psychological Association, 2000) includes the persistence of depressed mood, low self esteem, feelings of hopelessness, decreased Corresponding author: Tel.: +1 212-543-5477; Fax: +1 212-543-5074; E-mail: [email protected] (A. Sahay) Corresponding author: Tel: +1 212-543-5328; Fax: +1 212-543-5074; E-mail: [email protected] (R. Hen) DOI: 10.1016/S0079-6123(07)63038-6

697

698

effects of environmental stress during particularly sensitive or critical periods in brain development by determining the optimal range of neuronal circuit function for the organism. Indeed, the neurotrophic, neuroplasticity and network hypotheses of MDD all reflect the biology of gene products in the context of synaptic and structural plasticity of neural circuitry (Duman et al., 1997; Duman, 2002; Nestler et al., 2002; Castren, 2005). Dentate gyrus neurogenesis has gained considerable attention as both a form of structural plasticity and as a neural substrate for the pathophysiology of MDD. The neurogenic hypothesis posits that a decrease in the production of newborn dentate granule cells in the hippocampus causally relates to the pathogenesis and pathophysiology of MDD and that enhanced neurogenesis is necessary for treatment of depression (Duman et al., 2000; Jacobs et al., 2000). The hypothesis, when first proposed, was predicated on the following observations, which are reviewed in greater detail in subsequent sections. First, stress, which is widely recognized as a major causal factor in MDD, is known to suppress neurogenesis. Second, most antidepressant (AD) treatments increase hippocampal neurogenesis. Third, imbalance in the serotonin system influences hippocampal neurogenesis. Fourth, the induction of neurogenesis is contingent upon chronic but not subchronic (acute) selective serotonin reuptake inhibitor (SSRI) treatment, paralleling the time course for therapeutic actions of ADs. Finally, the therapeutic lag in the response to SSRIs in patients with MDD mirrors the timeline of maturation and integration of newborn dentate granule cells. Consequently, the dentate gyrus and neurogenesis therein are potential substrates for the AD response. Central to the neurogenic hypothesis is the assumption that the dentate gyrus plays an important role in mediating cognitive and affective processes. Moreover, since levels of neurogenesis change during the lifetime of the organism, changes in dentate neurogenesis may contribute to dentate gyrus function in different ways. Another assumption is that neurogenesis represents a potentially adaptive mechanism or form of plasticity. Deficits in neurogenesis during critical

periods in brain development could, therefore, be pathogenic in that they profoundly impact the trajectory of emotional development. Deficits in adult hippocampal neurogenesis could compromise hippocampal-dependent functions and contribute to the pathophysiology of MDD. In this chapter, we focus on hippocampal neurogenesis as it relates to MDD. Our aim is to distill the observations made in the rapidly growing field of hippocampal neurogenesis and to critically assess the putative role of neurogenesis in the etiology and treatment of MDD. We begin by defining a framework for the reader to understand how neurogenesis can contribute to dentate gyrus function. Within this framework, we will first evaluate evidence for deficits in hippocampal neurogenesis in patients with MDD. We will then examine the role of the serotonergic system in hippocampal neurogenesis because the best characterized genetic risk alleles for MDD encode components of this system. Because susceptibility to MDD conferred by genes is likely to be revealed by environmental risk factors such as stress, we will discuss the relationship between stress and neurogenesis. We then review the considerable evidence linking the effects of ADs with increased hippocampal neurogenesis. Finally, we will turn to evidence provided by studies using preclinical models that attempt to establish a causal link between hippocampal neurogenesis and the etiology, and pathophysiology of MDD and the requirement for neurogenesis in mediating the behavioral effects of ADs.

Neurogenesis and MDD A general framework for neurogenesis and dentate gyrus function Since the seminal findings of Altman and Das in 1965, it is now well accepted that the adult hippocampus is host to the birth and integration of newborn dentate granule cells in the dentate gyrus (Altman and Das, 1965). In the rat, the species for which the best data are available, it is estimated that 9000 new cells are born each day in the DG, and, of these, approximately 50% go on to express

699

neuron-specific markers. At this rate, the number of new granule neurons born each month is equal to 6% of the mature granule cell population (Cameron and McKay, 2001). In non-human primates, the rate of neurogenesis may be lower than the rate documented in rodents (Kornack and Rakic, 1999; Gould et al., 1999b). One should bear in mind that these data reflect neurogenesis under laboratory housing conditions, and given the increase in neurogenesis with environmental enrichment (Kempermann et al., 1997; Gould et al., 1999a), could underestimate the rates of neurogenesis in the normal habitat. Likewise, rates of neurogenesis in man maybe underestimated by available data, because human data were based on a single study in which tissue samples were taken from cancer patients injected with a mitotic marker, bromo-deoxyuridine (BrdU) before death (Eriksson et al., 1998). The number of BrdU-labeled neurons entering the neuronal lineage was lower than that reported for marmosets and rodents, but the age of subjects could explain the difference because they were old, and neurogenesis declines with age (Seki and Arai, 1995; Kuhn et al., 1996; Rao and Shetty, 2004). Therefore, it is unclear to what extent the relatively low level of neurogenesis observed in the human subjects was due to real species difference. The study of adult hippocampal neurogenesis has revealed it to be a robust phenomenon that is capable of conferring previously unrecognized forms of plasticity to the dentate gyrus. For example, it is clear that both net addition of newly generated neurons and replacement of mature cells occur in the adult dentate gyrus and that the extent to which these processes occur may vary with the animals age, and environmental and physiological parameters (Bayer et al., 1982; West, 1993; Kempermann et al., 1998; Nottebohm, 2002; Amrein et al., 2004; Wiskott et al., 2006). Modeling and computational approaches have revealed merits of both net addition and replacement in optimizing hippocampal network function (Chambers et al., 2004; Becker, 2005; Meltzer et al., 2005; Wiskott et al., 2006). Figure 1 illustrates the distinct, but potentially interrelated, ways by which neurogenesis can modify the cellular composition of the dentate gyrus.

1. Increase the number of mature dentate granule cells The integration of newborn neurons can result in an increase in the granule cell layer of the dentate gyrus. It is conceivable that a net increase in size is possible only within a certain period in an animal’s life. An increase in cell number can result from an enhancement in the rate of proliferation or the percentage of newborn neurons that survive. 2. Provide a reservoir of highly plastic immature neurons in the adult dentate gyrus Newly generated dentate granule cells also exhibit forms of synaptic plasticity distinct from those of mature cells in the adult hippocampus. Newborn dentate granule cells show unique physiological properties such as lower thresholds for induction of long-term potentiation and long-term depression than do mature neurons (Schmidt-Hieber et al., 2004; Song et al., 2005). Moreover, newborn dentate granule cells, unlike mature granule cells, are able to undergo LTP under conditions of increased GABAergic inhibition (Wang et al., 2000; Snyder et al., 2001; Saxe et al., 2006). Thus, in addition to conferring structural plasticity to the dentate gyrus, neurogenesis also creates a transient reservoir of excitable, highly plastic cells that may serve a unique biological function, distinct from that of mature granule neurons. The size of such a reservoir can by influenced by numerous physiological and environmental factors and contingencies. 3. Generate multiple cell types in the dentate gyrus While it is widely agreed that neurogenesis in the subgranular zone (SGZ) results in generation of dentate granule cells, there is one report showing that GABAergic basket cells in the dentate gyrus incorporate BrdU and form functional inhibitory synapses with dentate granule cells (Liu et al., 2003). Thus, it is plausible that the generation of interneurons may occur under certain conditions to influence network activity. Clearly more evidence is needed to support this possibility. In addition to the generation of neurons in

700

Fig. 1. A schematic of the dentate gyrus granule cell layer (GCL) illustrating the different ways by which neurogenesis can influence its structure and function. Boxed panel reveals a cross section of the dentate GCL with the different populations that reside within it: mature granule cells born during development (light blue), adult-generated mature granule cells (dark blue), adult-born immature neurons (red) and interneurons (green). Over the lifespan, the GCL may increase in size due to a net addition of new neurons (A) or may remain unchanged due to a net replacement of developmentally generated granule cells (B). Changes in neurogenesis can result in increased representation of interneurons (C), a larger pool of adult generated immature neurons (D) or the generation of mature neurons with distinct physiological and biochemical properties (E). Conceivably, neurogenesis may be altered in any one of these ways in MDD. Conversely, AD drugs may influence DG function in more than one way to exert their behavioral effects. (See Color Plate 38.1 in color plate section.)

the dentate gyrus, proliferation in the SGZ also generates glial cells. The emerging role for glial cells in modulating synaptic function in health and disease underscores the need to understand how newly generated glial cells contribute to hippocampal physiology and function (Ma et al., 2005; Haydon and Carmignoto, 2006). 4. Drive turnover and replacement of mature dentate granule cells The integration of newborn neurons can occur to replace the death of mature neurons.

Such a mechanism, when predominant, would not increase the size of the dentate gyrus but ensure replacement of cells, whose functions are impaired, and rejuvenate the network with new cells (Nottebohm, 2002). There is some evidence to suggest that adultgenerated mature dentate granule cells, while sharing electrophysiological properties with their early-development-born counterparts, exhibit greater plasticity in response to behaviorally relevant stimuli (Laplagne et al., 2006; Ramirez-Amaya et al., 2006).

701

Independent of the balance between the integration of new neurons and the death of mature neurons, it is conceivable that a specific form of experience can result in a larger representation of specific granule cells selected for by that kind of experience. Such a representation may manifest in a distinct pattern of biochemical and electrophysiological properties found in one cohort of newborn cells versus another. Functional heterogeneity within the hippocampus and dentate gyrus supports the possibility that subsets of neurons within different regions of the dentate gyrus could reflect distinct experiences (Moser and Moser, 1998; Scharfman et al., 2002; Silva et al., 2006). The aforementioned ways by which neurogenesis contributes to the structure and function of the dentate gyrus convey the complexity of the phenomenon of hippocampal neurogenesis. They also remind us of the many ways by which neurogenesis may be altered in pathological conditions.

Hippocampal dysfunction and atrophy in MDD Hippocampal dysfunction in MDD is well supported by clinical studies which have shown that MDD is often accompanied by deficits in declarative learning and memory and diminished cognitive flexibility that are dissociable from changes in motivation (Austin et al., 2001; Fossati et al., 2002). The anatomical and functional segregation of the hippocampus along its septotemporal axis suggests roles for the hippocampus in both cognitive and emotional processes (Moser and Moser, 1998; Strange and Dolan, 1999; Strange et al., 1999; Bannerman et al., 2003). As a first step to thinking about the contribution of the dentate gyrus and dentate neurogenesis to the pathophysiology of MDD, one must consider the direct evidence for hippocampal dysfunction and atrophy in MDD. While longitudinal studies linking changes in hippocampal structure and function with the etiology of depression are lacking, we have made progress in identifying changes in hippocampal function, cellular structure and volume that are associated with pathophysiology of depression. Direct measurements of hippocampal function in

the depressed brain are made using neuroimaging techniques such as positron emission tomography (PET), a powerful way of identifying neural structures with altered metabolic activity. Studies on patients with MDD have revealed alterations, but only in a very small number of studies and with conflicting results. This is partly due to the limited resolution of PET. Using cerebral blood flow PET, one group reported increased blood flow in the hippocampus of acutely depressed patients with a short duration of illness (Videbech et al., 2001, 2002; Videbech and Ravnkilde, 2004). By contrast, two other studies have shown either a decrease or no change in metabolism in the hippocampus of patients with MDD using fluorodeoxyglucose (FDG)–PET imaging (Saxena et al., 2001; Drevets et al., 2002; Kimbrell et al., 2002). Differences in patient profile with regards to severity and duration of illness and treatment could explain these differences. Alternatively, it has been suggested that increased activity, if untreated, may result in hippocampal atrophy and decreased metabolism. Atrophy would occlude detection of changes in activity. A consistent finding that has emerged from magnetic resonance imaging (MRI) studies on patients with MDD is a reduction in hippocampal volume. Despite a few studies that failed to report any differences between patients with MDD and control groups using MRI, there is consensus for reduced hippocampal volume in MDD (Sheline, 1996; Sheline et al., 1999; Bremner et al., 2000; von Gunten et al., 2000; Vakili et al., 2000; Neumeister et al., 2005). Two recent meta-analyses of studies measuring temporal lobe structures in MDD compellingly demonstrate a reduction in hippocampus in people with recurrent depression relative to ageand sex-matched controls (Campbell et al., 2004; Videbech and Ravnkilde, 2004). Interestingly, frequency of depressive episodes and the duration for which depression is untreated correlate with magnitude of reduction in hippocampal volume (MacQueen et al., 2003; Sheline et al., 2003). Taken together, the evidence argues for reduced hippocampal volume in MDD and that such changes are likely to be a result of depression rather than a cause. However, it is worth mentioning here that a smaller hippocampus is

702

thought to be a predisposing factor for, rather than a consequence of, post-traumatic stress disorder (Gilbertson et al., 2002). The significance of hippocampal volume change in the context of cognitive deficits is conveyed by a recent study that shows that healthy individuals who complained of memory impairments had smaller hippocampal volumes than non-impaired controls (van der Flier et al., 2004). Another study showed a correlation between deficits in recollection memory performance and reduced hippocampal volume in elderly depressed patients (von Gunten and Ron, 2004). However, these studies are limited in number and comprised elderly individuals who are likely to have other brain changes that may contribute to the memory impairments. Changes in hippocampal volume can be explained by several different mechanisms that may operate in concert at the cellular and circuit levels including: (i) Increased apoptosis of mature neurons or glial cells. (ii) A loss of neuropil which may involve changes in dendritic complexity, spine density, and number and size of afferent and efferent axonal projections. (iii) Reduced neurogenesis or gliogenesis in the SGZ of dentate gyrus. The link between neurogenesis and hippocampal volume has been addressed in histological analyses of postmortem tissue obtained from brains of patients with MDD, and will be discussed next.

Neurogenesis and cell death in MDD Pathohistological studies of postmortem tissue of patients, while small in number, have provided some clues about the nature of cellular changes in the hippocampus of a depressed individual. One study examined synaptic density and glial cell number using synaptophysin and GFAP-immunoreactivity, respectively, and found no differences in the hippocampus of medicated patients with MDD relative to controls (Muller et al., 2001). Another study revealed low levels of apoptosis in the dentate gyrus, CA1, CA4, subiculum and entorhinal cortex of patients with MDD (Lucassen et al., 2001). A third study showed a significant increase in cell density of granule and glial cells in the

dentate gyrus and pyramidal neurons and glial cells in the CA fields (Stockmeier et al., 2004). In addition, the authors reported a reduction in soma size of pyramidal neurons and a trend towards the same in dentate granule cells. Finally, one group directly examined the proliferation of cells in the adult dentate gyrus of MDD patients using an M-phase marker, Ki-67 (Reif et al., 2006). Their results showed no changes in Ki-67 immunopositive cells in hippocampus of depressed patients. While this study is informative and is the first to estimate levels of proliferation in the MDD brain, it must be interpreted with several caveats in mind. First, a reduction in neurogenesis in patients with MDD could be masked by AD-mediated increase in cell proliferation. Second, and for obvious reasons, the study could not measure changes in survival of newborn cells or examine the kinetics of turnover and maturation of newborn dentate granule cells. The pathohistological analyses suggest that changes in neuropil, rather than neurogenesis, may account for reductions in hippocampal volume. Indeed, the effects of stress on hippocampal white matter are well documented. Preclinical studies have shown that volumetric changes result from reduced dendritic complexity and not ablation of hippocampal neurogenesis (Santarelli et al., 2003; McEwen, 2005). It should be noted that while these data argue against the possibility that reduced cell number owing to extensive cell death or decreased neurogenesis is a primary mediator of hippocampal volume change, they do not directly address the possibility that neurogenesis is altered in MDD. Since most of the patients were on medication at the time of death, and since AD drugs potently upregulate hippocampal neurogenesis, it is possible that depression-related alterations in neurogenesis could have been masked in these studies. Moreover, since hippocampal neurogenesis in humans is likely to change with age, small differences in proliferation or survival of newborn neurons in postmortem analyses of older patients could be difficult to detect. Further studies on postmortem tissue of medicated and non-medicated individuals are needed to identify specific changes in hippocampal neurogenesis associated with MDD.

703

Genes, environment and MDD Depression is a complex and multifactorial illness with genetic and non-genetic underpinnings. The heritability of MDD is likely to be in the range of 40–50% and there is substantial evidence to suggest that the phenotypic expression of MDD is contingent upon interactions between the genetic make-up of the individual and environmental factors, an interaction that has a dramatic effect on the formation and functioning of neural circuitry (Sullivan et al., 2000; Kendler et al., 2001; Caspi and Moffitt, 2006; Leonardo and Hen, 2006; Levinson, 2006). Here, it must be emphasized that human susceptibility to MDD as revealed by environmental factors is tremendously magnified in early life. For example, adults who had experienced four out of seven traumatic events in early life had a 4.6-fold increased risk of developing depressive symptoms later in life and were 12.2-fold more likely to commit suicide (Felitti et al., 1998; Chapman et al., 2004). These studies underscore the idea of a critical period in ‘‘emotional development’’ when mechanisms mediating neural circuit synaptic- and structural plasticity are particularly susceptible to environmental input, and if compromised by a vulnerability conferred by genes, can result in maladaptive alterations in neural circuit function and pathological behavior. While the search for candidate genes for MDD has yielded some convergence from linkage studies that certain genetic loci are involved, more data are clearly needed to conclusively implicate a specific gene in the etiology of MDD. A notable exception is the serotonin transporter (5-HTT) polymorphism. In a landmark study, Caspi and colleagues found that a functional polymorphism in the 5-HTT gene moderated the sensitivity of individuals to the depressogenic effects of early life stress, a finding recently replicated by Kendler and colleagues (Caspi et al., 2003; Kendler et al., 2005). Caspi and colleagues found that people who carried one or two copies of the ‘‘short’’ allele of 5-HTT, associated with lower levels of 5-HTT and impaired reuptake of serotonin (5-HT) at synapses, had more depressive symptoms and suicidal behavior in relation to stressful life events than did people who had the ‘‘long allele’’.

Importantly, the association between the 5-HTT polymorphism and depression is only observed in individuals who had experienced significant stressful life events. These findings argue that the serotonergic system has a critical influence on neurodevelopmental processes that lead to MDD. If hippocampal neurogenesis is to be a considered as a candidate neural substrate for depression, then the effects of serotonergic dysregulation on it warrants comment. The following section addresses the role of the serotonergic system in modulating dentate gyrus structure and function.

The 5-HT system and hippocampal neurogenesis Serotonergic terminals originating from the dorsal raphe nucleus (DRN) and median raphe nuclei (MRN) diffusely innervate multiple structures in the vertebrate forebrain and reach the ventricles via the supra-ependymal plexus (Azmitia and Segal, 1978; Freund et al., 1990). The serotonergic innervation of the hilus, molecular layers of the dentate gyrus and the SGZ supports the possibility that 5-HT signaling may influence adult neurogenesis (Oleskevich et al., 1991). The idea that serotonin can influence neurogenesis was first proposed three decades ago (Lauder and Krebs, 1978). However, the specific ways by which the 5-HT system influences adult dentate neurogenesis was established only relatively recently. The first studies to address the role of 5-HT in adult hippocampal neurogenesis used serotonin depletion analyses. Injection of the serotonin neurotoxin 5,7-dihydroneurotoxin (5,7-DHT) or 5-HT synthesis inhibitor parachlorophenylalanine (PCPA) into the raphe of young female rats resulted in a reduction of dentate granule cell proliferation and the number of immature neurons as assessed by BrdU uptake and PSA-NCAM immunostaining, respectively (Brezun and Daszuta, 1999, 2000). Moreover, the same group showed that they could rescue the deficit in hippocampal proliferation in these rats following intrahippocampal grafts of embryonic 5-HT neurons (Brezun and Daszuta, 2000). A potential role for 5-HT in influencing the maturation of dentate granule cells comes from studies in which rat pups were treated

704

with phenylchloromethamphetamine (PCA) or 5,7-DHT to reduce serotonergic innervation of the forebrain. Analysis of dentate granule cells in rodents with reduced serotonergic innervation revealed fewer dendritic spines and synapses, but otherwise normal dendritic complexity (Yan et al., 1997; Faber and Haring, 1999), suggesting that serotonergic signaling is important for selected, and not all, aspects of the neuronal maturation process. While the limitations inherent to these pharmacological lesion studies must be acknowledged, these studies illustrate how deficits in the 5-HT system can have consequences for the maturation of dentate granule cells. Given the striking recapitulation in adult neurogenesis of the earlydevelopmental neuronal maturation process, it is likely that changes in 5-HT signaling have similar consequences on neurons born in adulthood (Esposito et al., 2005; Laplagne et al., 2006; Overstreet-Wadiche and Westbrook, 2006). Indeed, the well-characterized effects of 5-HT receptor agonist and antagonists and SSRIs on adult neurogenesis (Malberg et al., 2000) solidify the link between 5-HT and adult hippocampal neurogenesis. Importantly, studies on 5-HT receptors, which are discussed next, offer a glimpse into the ways by which altered serotonin levels as a consequence of a genetic polymorphism, such as the 5-HTT polymorphism, can influence the birth and maturation of newborn dentate granule cells. The effects of 5-HT levels on neurogenesis reflect the sum of interactions between the synthesis of 5-HT, its release and its actions at different 5-HT postsynaptic receptors acting in both a cell autonomous and non-cell autonomous manner. The effects of 5-HT on a newborn neuron depend on the repertoire of 5-HT receptors that it expresses. Since the maturation of newborn neurons is intimately connected with the activity of the network, it is also influenced by the actions of different 5-HT receptors expressed within the hippocampal formation in interneurons, mature dentate granule cells and afferent projections arising in the entorhinal cortex. The 5-HT1A receptor (5-HT1AR) is the best-studied 5-HT receptor in the context of adult hippocampal neurogenesis. Acute administration of 5-HT1A antagonists results in decreased cell proliferation in the adult

dentate gyrus (Radley and Jacobs, 2002). Consistent with these findings, acute or chronic treatment with the 5-HT1A agonist 8-OH DPAT increases proliferation in the SGZ and the number of adult born neurons (Santarelli et al., 2003; Banasr et al., 2004). The effects of activating 5-HT1AR appear to be restricted to proliferation and do not affect the differentiation of newborn progenitors into neurons or glial cells. The increase in proliferation could reflect a change in rate of progression through the cell cycle or an increase in the size of the proliferative pool in the SGZ. It is unclear, given the experimental design employed in these studies, whether activation of 5-HT1AR also influences the survival of newborn neurons. Interestingly, mice lacking the 5-HT1AR fail to respond to the neurogenic effects of chronic fluoxetine (Santarelli et al., 2003). Two other 5-HT receptors, the 5-HT2A and 5-HT1B receptors, have also been implicated in cell proliferation in the adult SGZ. While neither 5-HT1B agonists nor antagonists affect baseline cell proliferation, the former can, however, restore normal levels of proliferation in PCPA pretreated rats. These data suggest that effects of the 5-HT1B receptor on cell proliferation are small under physiological conditions, but can become important when 5-HT levels are decreased. The pharmacology of the 5-HT2A receptor is also complex. 5-HT2A antagonists decrease cell proliferation but agonists have no effect (Banasr et al., 2004), suggesting that under physiological conditions, the 5-HT2A-dependent signaling pathways that modulate neurogenesis may be saturated. Taken together, these observations reveal the differential effects of recruiting different postsynaptic 5-HT receptors on hippocampal neurogenesis. Central to understanding how changes in 5-HT levels influence neurogenesis is knowledge of the expression of different 5-HT receptors in neural progenitors and at different stages of their maturation. Conspicuously absent from the pharmacological studies is precise information for 5-HT receptor distribution in the adult SGZ. The 5-HT1AR is an exception to the rule. We know that the 5-HT1AR is expressed at very low levels, if any, in the SGZ of the rodent hippocampus.

705

In both rat and mouse, 5-HT1AR expression is restricted to the mature dentate granule cells rather than the immature population of cells during development of the dentate gyrus (Patel and Zhou, 2005; Sahay and Hen, unpublished data). The effect of 5-HT1AR agonists on cell proliferation is, therefore, likely to be non-cell autonomous. It is possible that the 5-HT1AR is required in hilar interneurons or mature dentate granule cells to mediate the effects of 5-HT on cell proliferation. Cell-type specific ablation and overexpression of the 5-HT1AR will reveal its precise contribution in different cell types to adult hippocampal neurogenesis. The specific role of different 5-HT receptors in the maturation and integration of newborn neurons is still to be elucidated. It is also unclear how 5-HT may impact the turnover of mature dentate granule cells or influence the survival of newborn dentate granule cells. The role for distinct 5-HT receptors in regulation of developmental processes such as dendritic development, synaptogenesis and glutamate receptor trafficking (Kondoh et al., 2004; Kvachnina et al., 2005; Yuen et al., 2005a, b) suggests that 5-HT receptor-dependent mechanisms during development may be conserved in neurogenesis in the adult brain depending on which 5-HT receptors are expressed and when during neuronal maturation. In addition, 5-HT signaling can induce the production of neurotrophins and growth factors known to regulate hippocampal neurogenesis.

The role of stress in MDD and its effects on neurogenesis Stressful life experiences play a pivotal role in development of MDD in individuals with a genetic vulnerability (Holsboer and Barden, 1996; Gold and Chrousos, 2002). Major stressors precede the appearance of the first symptoms, and dysregulation of the hypothalamic-pituitary-adenocortical (HPA) system is often observed in patients with MDD (Carroll et al., 1968a). An optimally functional HPA system enables the organism to respond appropriately to stressful stimuli by controlling the production of adrenal steroids,

the glucocorticoids, which mobilize energy, increase cardiovascular tone and influence immune and nervous system functions. In response to a stressful event, for example, a neuroendocrine cascade is initiated in which corticotropin releasing hormone (CRH) is released from neurons in the paraventricular nucleus (PVN) in the hypothalamus, which then triggers release of corticotropin by the anterior pituitary to stimulate glucorticoid secretion by the adrenal cortex. A hyperactive HPA, on the other hand, results in the oversecretion of glucocorticoids with deleterious consequences for the physiology of the viscera and brain (Sapolsky, 2000). In addition, increased glucocorticoid levels can impair serotonergic signaling (Joels and van Riel, 2004). It is therefore not surprising that the stress response is tightly regulated by efferents from multiple brain regions that converge onto the PVN, where incoming information is integrated to elicit an adaptive response. Afferent inputs of the PVN are inhibitory or facilitative in nature and arise in brain stem nuclei, amygdala, cortex, septum and the hippocampus, and are in turn regulated by a negative feedback system. The hippocampus, for example, exercises a powerful inhibitory influence on HPA function to terminate the stress response, and is in turn regulated by glucocorticoids acting on cognate receptors expressed within the hippocampal formation. Dysregulation of such control can, therefore, result in hypersecretion of glucocorticoids and an exaggerated stress response that can severely impact neural circuit function. Consistent with preclinical findings, about half of all depressed patients show a blunted response to the dexamethasone suppression test (Carroll et al., 1968b; Holsboer et al., 1982), which measures the ability of an exogenous glucocorticoid receptor agonist to suppress endogenous stress hormone release. Moreover, patients with MDD show elevated levels of CRH in the cerebrospinal fluid, increased numbers of CRH containing cells in the PVN, and decreased CRH binding in the prefrontal cortex. Thus, the optimal function of the hippocampal formation is a critical factor in modulation of the stress response, and an exaggerated stress response can, in turn, negatively impact hippocampal function.

706

One effect of stress on neural circuitry within the hippocampal formation is the suppression of neurogenesis by glucocorticoids. Stress-induced suppression of cell proliferation in the DG has been reported in several different mammalian species and there is considerable evidence arguing for a role for glucocorticoids as the mediators of the stress response. Elimination of circulating adrenal steroids by adrenalectomy, for example, increases cell proliferation and neurogenesis in the adult dentate gyrus (Cameron and McKay, 1999). Exogenous administration of corticosterone, on the other hand, suppresses proliferation (Cameron and Gould, 1994). Glucocorticoids have been shown to inhibit the proliferation and differentiation of neural progenitors, and also the survival of young neurons (Wong and Herbert, 2004, 2006). These effects are likely to be mediated directly through high affinity mineralocorticoid receptors (MR) and low affinity glucocorticoid receptors (GR) that are expressed at various stages of maturation and also indirectly through changes in network activity of the hippocampal formation (see chapter by Joels in this volume). Analysis of receptor distribution in the dentate gyrus of rodents reveals that GRs are expressed in both neural progenitors and mature dentate granule cells, while MRs are expressed only in the latter. Throughout most of adulthood, neither GRs nor MRs are expressed in immature neurons (Cameron et al., 1993; Garcia et al., 2004a). However, in aged rodents GR and MR expression is seen in immature neurons, suggesting that immature neurons at this stage, but not earlier in the animal’s life, may show increased sensitivity to corticosterone action. The dynamic expression of GR and MR during neurogenesis and the ability of corticosteroids to regulate the expression of growth factors such as IGF-1, BDNF and EGF (Islam et al., 1998; Schaaf et al., 2000), which have distinct effects on proliferation and survival (Kuhn et al., 1997; Aberg et al., 2000; Sairanen et al., 2005), illustrate the cell autonomous and non-cell autonomous ways by which corticosteroids can affect neurogenesis. It should be noted that the sensitivity of neurons or neural progenitors to corticosteroids must be studied with the state of the neuron or network in

mind. For example, glucocorticoids have deleterious effects on neurogenesis during stressful episodes but not during physical activity. This may be due to the fact that physical activity, unlike stress, elicits the production of growth factors that may buffer the effects of corticosteroids on neurons. Likewise, given the differences in the developing and adult brain, an increase in glucocorticoids during early postnatal life may have profoundly different effects from those in adulthood. A recent study modeling early life stress using a maternal separation paradigm in rodents supports this idea (Mirescu et al., 2004). Maternal separation during the early postnatal period in rodents leads to persistent changes in the HPA axis, protracted release of corticosterone in response to mild stressors, increased anxiety behavior (Huot et al., 2001), impaired maternal care (Lovic et al., 2001) and impaired spatial navigation (Huot et al., 2002). Rats subjected to prolonged maternal deprivation during the early postnatal period also showed reduced proliferation in the dentate gyrus of neural progenitors in adulthood. Interestingly, the early life stressor did not affect the number of mature neurons, suggesting a compensatory increase in survival of newly generated neurons. The authors in this study showed that the change in proliferation could be rescued in adulthood by adrenalectomy and by reducing corticosteroid production. Without adrenalectomy, the corticosterone levels were normal in stressed animals, not only under baseline conditions but also in response to a stressor, suggesting that the suppressed proliferation was likely to be result of increased sensitivity to corticosteroids rather than increased levels (Mirescu et al., 2004). It is tempting to speculate that a shift in the temporal pattern of MR and GR distribution during neurogenesis may contribute to this increased sensitivity. It is possible that changes in neurogenesis as a result of HPA axis dysregulation contribute to the pathophysiology of MDD by affecting the role of the hippocampus in learning and other cognitiveemotional processes. Preclinical studies (discussed later) have proven tremendously informative in defining the physiological relevance of adult hippocampal neurogenesis to these processes, and have greatly facilitated our understanding of how

707

stress-mediated suppression of neurogenesis may be relevant to the pathophysiology of MDD. A second possible consequence of the stress-mediated suppression of neurogenesis is that the ability of the hippocampus to regulate HPA activity becomes compromised. However, evidence directly linking changes in dentate gyrus function with HPA activity is scarce. Lesions of the hippocampus and fimbria-fornix transections reduce the ability of dexamethasone to inhibit stress induced adrenocortical responses and results in hypersecretion of glucocorticoids and ACTH following stressful stimuli (Knigge, 1961; Sapolsky et al., 1989; Herman et al., 1992; Feldman and Weidenfeld, 1993; Bratt et al., 2001; Goursaud et al., 2006). Antagonism of GR in the rat hippocampus results in hypersecretion of ACTH and corticosterone following stressful stimuli (Sapolsky, 1994; Feldman and Weidenfeld, 1999). Analysis of subfields within the hippocampal formation confirms the differential contributions of CA fields and the dentate gyrus in the ventral hippocampus in regulating HPA reactivity. Intriguingly, lesion studies indicate that damage of ventral subiculum but not ventral CA1 or dentate gyrus results in prolonged glucocorticoid responses (Herman et al., 1992, 1995). By contrast, electrical stimulation of DG in anesthesized animals inhibits corticosteroid secretion (Dunn and Orr, 1984). These studies suggest that lesions of DG in adulthood may be compensated for by activity in structures downstream such as the subiculum. Genetic manipulations that specifically impair glucocorticoid signaling in the DG both in adulthood and in early postnatal life will prove critical in establishing the link between hippocampal neurogenesis and HPA reactivity. In sum, the well-documented effects of stress on neurogenesis suggest that patients with MDD are likely to show reductions in neuronal proliferation. While it is plausible that a reduction in neurogenesis in turn contributes to HPA dysregulation, there is as yet no evidence for this. Neurogenesis and antidepressants It is well recognized that all of the major classes of ADs are associated with a several week delay in

onset. This delay is likely to reflect changes in structural and synaptic plasticity in the brain mediated by multiple mechanisms involving monoaminergic signaling and neurotrophins. PET imaging studies on MDD patients treated with SSRIs such as paroxetine and fluoxetine have helped define a neuroanatomical basis comprising corticolimbic circuits (Seminowicz et al., 2004). Structures that showed changes in metabolic activity included the subgenual cingulate, hippocampus and prefrontal cortex (Mayberg et al., 2000; Kennedy et al., 2001). One form of structural plasticity within the hippocampus that is consistent with the delayed onset of ADs is the birth and subsequent integration of newborn dentate granule cells in the adult dentate gyrus. Moreover, almost all ADs known to date increase adult neurogenesis. Therefore, the idea that ADs may work through enhancing neurogenesis has received abundant attention and is now considered central to the neurogenic hypothesis of MDD (Malberg and Schechter, 2005).

Neurogenic effects of antidepressant treatments Numerous groups have shown that different classes of ADs including 5-HT and norepinephrine selective reuptake inhibitors, tricyclics, monoamine oxidase inhibitors, phosphodiesterase inhibitors and electroconvulsive shock therapy increase neurogenesis (Madsen et al., 2000; Malberg et al., 2000; Manev et al., 2001; Nakagawa et al., 2002b). AD treatment does not appear to affect the ratio of newly generated neurons to glial cells, with the majority of newborn cells adopting the neuronal fate. The neurogenic effects of ADs are specific to the SGZ, and are not observed in other components of the ventricular system such as the lateral ventricles or the subventricular zone. Moreover, administration of non-AD psychotropic drugs such as haloperidol does not increase hippocampal neurogenesis (Eisch, 2002). Other treatments reported to have AD effects, including exercise (Babyak et al., 2000; Singh et al., 2001; Motl et al., 2004), environmental enrichment and estrogen, have also been shown to increase neurogenesis (van Praag et al., 1999; Tanapat et al., 1999;

708

Rhodes et al., 2003; Meshi et al., 2006). In addition, also lithium, which is used in the treatment of bipolar disorder increases neurogenesis (Chen et al., 2000). The one AD treatment that has not been shown to enhance neurogenesis is repetitive transcranial magnetic stimulation or rTMS, which is still awaiting FDA approval (Loo and Mitchell, 2005). While rTMS has been shown to reverse the effects of chronic psychosocial stress on stress hormone levels, it does not upregulate neurogenesis (Czeh et al., 2002). That ADs block the behavioral effects of stress and restore normal levels of neurogenesis in the adult hippocampus lends further credence to the possibility that ADs may work by increasing neurogenesis to exert their behavioral effects. In tree shrews, chronic exposure to psychosocial conflict results in a decrease in cell proliferation, which is blocked by treatment with the atypical AD tianeptine (Czeh et al., 2001). In another model of depression, the learned helplessness (LH) paradigm, exposure to inescapable shock engenders prodepressive behavior and a reduction in hippocampal cell proliferation, both of which, are reversed by AD treatment (Cryan et al., 2002; Malberg and Duman, 2003). In addition, ECS enhances cell proliferation after chronic corticosterone treatment (Hellsten et al., 2002). It is well known that ADs have pleiotropic effects on neuronal circuits. That a diverse range of ADs appears to enhance neurogenesis indicates that the dentate gyrus may be a neuroanatomical substrate to target for the development of novel AD treatments. In the next section, we describe recent preclinical studies elucidating how SSRIs modulate adult hippocampal neurogenesis, and then in the following sections we describe work from animal models aimed at identifying whether neurogenesis is necessary for the behavioral effects of AD treatments.

Serotonin-dependent ADs and hippocampal neurogenesis SSRIs represent the most successful class of ADs identified to date. Based on our understanding of neurogenesis in the SGZ, it is conceivable that

SSRIs act directly on progenitors or immature neurons to influence processes such as proliferation, differentiation, maturation and survival. In addition, SSRIs are also likely to modulate network activity within the dentate gyrus and as a result, regulate neurogenesis indirectly. By virtue of their effects on neurogenesis, SSRIs may be capable of driving replacement or turnover within the dentate gyrus and catalyzing the insertion of newly generated neurons with distinct electrophysiological and biochemical properties. The bestcharacterized effect of SSRIs to date is the increase in proliferation of neural progenitors in the SGZ. Studies in rodents indicate that a 14-day, but not shorter-term, administration of fluoxetine (1–5 days) is sufficient to upregulate cell proliferation (Malberg et al., 2000). By contrast, a longer treatment regimen is required to enhance survival of newly generated neuroblasts. Fluoxetine treatment for 28 days, but not 14 days, following BrdU injection resulted in an increase in cell survival (Malberg et al., 2000; Nakagawa et al., 2002a). The delay with which the neurogenic effects emerge after the initiation of SSRI treatment provides a potential mechanism to explain the therapeutic lag in the effects of these drugs. Namely, it could be that the therapeutic effects depend on the increase in proliferation, which itself requires several weeks of treatment. However, closer inspection of this hypothesis reveals several shortcomings. It is unlikely that a boost in proliferation would produce an immediate psychological effect. Rodent studies indicate that newborn neurons do not become functionally integrated until approximately 2–4 weeks after exiting the cell cycle, so it would seem that any functional effects elicited by the increase in proliferation would not appear until that time (Esposito et al., 2005). This means that behavioral effects of SSRIs mediated by increased proliferation should not manifest until about 4 weeks after treatment initiation. In contrast, some behavioral effects of AD drugs in animal models begin immediately following acute treatment, and virtually all the behavioral effects manifest within 4 weeks of SSRI treatment. In primates the time required for maturation of new neurons is greater than in rodents (Kohler et al.,

709

2006), so the predicted therapeutic lag would be even longer. A recent meta-analysis of clinical data indicates that some of the therapeutic effects of SSRIs may commence very rapidly after treatment initiation (within 1–2 weeks) (Taylor et al., 2006). Thus, increases in proliferation are unlikely to underlie the early onset effects of these drugs, but it remains plausible that neurogenesis contributes to the more long-term effects. Consistent with this interpretation, remarkable preliminary data from Meltzer, Deisseroth and colleagues suggest that mature adult-born neurons contribute to the behavioral effects of SSRIs (Meltzer et al., 2006). In this study, rats were given 7 days of fluoxetine treatment and then tested behaviorally 1 month later. At that time point, the previously treated rats showed enhanced performance in the forced swim test, and histology revealed an increase in the number of newly generated neurons. The results suggest that the proliferative effects of SSRIs may appear sooner than once thought (after 1 week rather than 2 weeks of treatment), and that some behavioral effects of these drugs depend not on the acute presence of the drug but rather on slow, time-dependent processes initiated by drug treatment, such as neurogenesis and circuit reorganization. The studies that examined the effects of fluoxetine on proliferation do not identify the specific types of neural progenitors that respond to changes in 5-HT levels. A recent study addressed this question using transgenic mice in which expression of a fluorescent reporter gene was regulated by a nestin promoter fragment. Because nestin is expressed in multiple progenitor cell types, this approach allowed the visualization and quantification of the distinct sub types of neural progenitors that reside within the SGZ (Encinas et al., 2006). The results of this study showed that only a specific proliferative cell type, the transiently amplifying neural progenitor (ANP) that exists as an intermediate between the type I radial glial-like neural progenitor and the type II cell (Filippov et al., 2003; Tozuka et al., 2005; Encinas et al., 2006), directly responds to fluoxetine. Moreover, the study confirmed that once exposure to fluoxetine ends, the rate of progenitor cell division is restored to baseline and that

the increase in ANPs translates into a net increase in the number of new neurons. A net increase in the number of mature neurons implies that SSRIs also increases the population of immature neurons. It is presently not clear how SSRIs influence the maturation of immature neurons with regards to their physiological properties, synaptic connectivity and dendritic complexity. What is well appreciated, however, is that SSRI treatment results in the induction of growth factors and neurotrophins whose effects on maturation and survival are well understood (Carlezon et al., 2005; Duman and Monteggia, 2006) and, importantly, whose receptors are expressed in immature neurons. It is also possible that SSRIs may induce the secretion of growth factors from neural progenitors, which then influence the function of neighboring mature granule cells in a paracrine manner. There is no evidence for this as yet. The consequences of increased proliferation and survival of newly generated cells following fluoxetine treatment for the net size of the granule cell population of the dentate gyrus have not been ascertained. One study suggested that there may not be a net change in the size of the dentate gyrus because the increase in newly generated neurons is offset by death of previously born mature granule cells (Sairanen et al., 2005). Based on their data showing that chronic fluoxetine treatment increases not only proliferation and survival of newly generated neurons but also the rate of apoptosis, the authors argue that the net size of the dentate gyrus does not change with chronic AD treatment. Finally, it is plausible that SSRI treatment alters the physiological properties of newly generated neurons, generating a cohort of cells within the dentate that are unique. These unique cells may confer greater adaptive potential to the dentate gyrus than naı¨ ve newly generated neurons. Much work remains to be done in this area to address the different ways by which SSRIs influence neurogenesis. Delineating the pattern of expression of distinct 5-HT receptor subtypes within different cell types in the SGZ and the DG will shed light on how changes in 5-HT can elicit the diverse range of effects discussed here. It follows that the

710

identification of genes downstream of 5-HT receptors that mediate the behavioral effects of ADs will pave the way for developing neurogenic nonmonoamine based therapeutics. Preclinical studies have proven invaluable in defining the contribution of neurogenesis to the etiology and treatment of MDD as they allow for discernment between correlation and causality. Preclinical studies: in the search for causality Ultimately, whether neurogenesis is causally related to the etiology or treatment of depression requires the use of animal models in which neurogenesis and emotional state can be experimentally manipulated. If reduced neurogenesis contributes to depression, it should be possible to produce a depressive phenotype by experimentally reducing neurogenesis. Conversely, behavioral manipulations that produce a depressive phenotype should reduce neurogenesis and do so before the behavioral manifestations of depression develop. Of course, the predictive value of such experiments will depend on the specificity of the methods for manipulating neurogenesis and the validity of the animal models of depression. This section will review recent work that addresses the role of neurogenesis in the etiology and treatment of MDD assessed in preclinical models. Lessons from stress based depression paradigms and neurogenesis One prediction drawn from the neurogenic hypothesis of MDD is that behavioral manipulations that produce depressed behavior in animals should produce a concomitant decrease in neurogenesis. Several methods have been used to produce depression-like behavior in rodents, all involving the exposure to inescapable stress. One such procedure is LH, originally developed by Seligman and colleagues (Overmier and Seligman, 1967; Seligman and Beagley, 1975). LH is a relatively short-term procedure in which animals are exposed to inescapable shock inside a conditioning chamber. Subjects are then given an escape/avoidance task in which shock is controllable (e.g., shuttling from

one side of the chamber to the other cancels or terminates the shock). Exposure to inescapable shock impairs subsequent acquisition of the escape/avoidance task (relative to naı¨ ve subjects), arguably because the animal has learned it is helpless (Willner and Mitchell, 2002). LH training also reduces cell proliferation in the DG. Treatment with AD drugs alleviates both the behavioral helplessness (Malberg and Duman, 2003; Chourbaji et al., 2005) and the reduction in cell proliferation (Malberg and Duman, 2003). However, there are several reasons why reductions in proliferation are not a likely mechanism of the behavioral helplessness. First, behavioral helplessness manifests immediately with exposure to inescapable shock, and it is unlikely that a reduction in proliferation could so rapidly give rise to a behavioral effect. If the reduction in proliferation were to impact behavior, the effects would more likely manifest 1–3 weeks after the onset of the reduction in proliferation, at the time when the newborn cells would be becoming functionally integrated into DG circuits. Similarly, acute dosing with AD drugs is sufficient to alleviate behavioral helplessness (Malberg and Duman, 2003; Chourbaji et al., 2005), but the neurogenic effect of these drugs requires chronic treatment (Malberg et al., 2000). Finally, a recent study has demonstrated that LH training in rats reduces cell proliferation in all subjects, but only a subset of subjects display behavioral helplessness (Vollmayr et al., 2003). One unexplored possibility invoked by these studies is that the extant population of immature neurons, rather than the proliferative pool, is a substrate for the depressogenic effects of stress. Studies are underway to test this hypothesis. Neurogenesis can more plausibly be linked to the effects of chronic exposure to stress, which have been studied extensively in animals. This work typically involves exposing animals to variety of mild stressors over a period of several weeks. Stressors include food and water deprivation, temperature changes, restraint and tail suspension (Strekalova et al., 2004; Willner, 2005; Mineur et al., 2006). There are variations in the methods used for chronic stress in rats (Willner et al., 1987, 1992; Willner, 2005) and mice (Mineur et al., 2003, 2006; Strekalova et al., 2004). The effects of

711

chronic stress include a reduction in sucrose preference, which has been interpreted as anhedonia (Willner et al., 1987; Strekalova et al., 2004), alterations in sleep–wake cycle, reduced sexual and self-care behavior, and increases in anxiety-like behavior in traditional tests such as the elevatedplus maze (Willner, 2005). Unlike the effects of LH, the behavioral effects of chronic stress are ameliorated by chronic but not acute treatment with ADs (Willner et al., 1987; Yalcin et al., 2005), suggesting that chronic stress may be a better model of depression because it captures the therapeutic lag seen in human patients. An abundance of research demonstrates that hippocampal neurogenesis is reduced by a variety of chronic stress procedures (for review, see Duman, 2004), and this reduction in neurogenesis is blocked by chronic treatment with AD drugs (Alonso et al., 2004). Moreover, a recent study has shown that among rats exposed to chronic stress, only a subset respond behaviorally to SSRI treatment (Jayatissa et al., 2006). Interestingly, neurogenesis was restored to normal levels only in the behaviorally identified responders.

Does blockade of neurogenesis produce symptoms of depression? The animal research on chronic stress evidences an intriguing correlation between the rate of neurogenesis and emotional state. Chronic stress produces a behavioral phenotype analogous to aspects of depression while reducing neurogenesis. AD drugs restore neurogenesis and ameliorate the behavioral phenotype. Does this correlation reflect that neurogenesis is a causal mechanism for these behavioral changes? Or is neurogenesis simply a marker for changes in other biological pathways or network activity? We have begun to address this question experimentally by blocking hippocampal neurogenesis and then examining the behavioral consequences. One method of arresting neurogenesis is to subject the brain to low doses of X-irradiation. Irradiation kills mitotic cells such that neurogenesis is arrested virtually completely (Monje et al., 2002, 2003; Wojtowicz, 2006). In our laboratory, a shield is

used to target X-rays specifically over the hippocampus, while protecting regions anterior and posterior, including the subventricular zone. Blocking neurogenesis with this procedure has no effect in a number of relevant behavioral tasks, including the novelty-suppressed feeding test (Santarelli et al., 2003) and the novelty-induced hypophagia test (our unpublished data). Both of these tests are AD screens that measure the latency of a mouse to venture into the center of an open field or novel context to obtain food. Latency-tofeed is decreased by chronic AD drug treatment and acute anxiolytic treatment, but not by acute AD drug treatment (Bodnoff et al., 1989; Dulawa et al., 2004). Irradiation does not affect behavior in two traditional anxiety tests, the elevated-plus maze and light-dark choice test (our unpublished data), nor does it increase the susceptibility of mice to the effects of chronic stress (Santarelli et al., 2003). We have also examined some of these behaviors in a transgenic mouse line in which neuronal proliferation can be blocked conditionally. The mouse line expresses herpes-simplex virus thymidine kinase (HSV-TK) under control of the GFAP promoter (GFAP-TK) (Garcia et al., 2004b). Mitotic cells expressing HSV-TK are killed by the antiviral drug ganciclovir. Thus, in this mouse, the dividing neuronal progenitors, which express GFAP, are killed after ganciclovir is administered, and as a result, neurogenesis is reduced to low levels (Garcia et al., 2004b; Saxe et al., 2006). As with irradiation, blocking neurogenesis in this mouse line had no effect on anxiety-like behavior in several tests, including the open field, NSF test, or light-dark choice test (Saxe et al., 2006). Thus, we have not found any evidence that blocking neurogenesis in adult mice produces a depression-like phenotype.

A role for hippocampal neurogenesis in learning The only domain in which direct behavioral effects of arresting neurogenesis have been reported is in learning and memory. Indeed, the very first evidence for a behavioral function of neurogenesis in mammals came from studies of learning and

712

memory. These studies showed that participating in a hippocampus-dependent classical conditioning procedure (trace eyeblink conditioning) enhances the survival of newborn neurons in the SGZ (Gould et al., 1999a) and that reducing neurogenesis with the anti-mitotic compound methylazoxymethanol acetate (MAM) impairs acquisition of the same task (Shors et al., 2001). The use of MAM is encumbered by deleterious side effects and moreover, it does not completely block neurogenesis (Dupret et al., 2005). Subsequent experiments using more targeted methods for arresting neurogenesis have confirmed that neurogenesis is required for some hippocampusdependent learning tasks (Snyder et al., 2005; Saxe et al., 2006; Winocur et al., 2006). A recent study from our laboratory has also revealed a paradoxical role for hippocampal neurogenesis in a hippocampus-dependent working memory paradigm, where ablation of newly generated neurons results in improved performance (Saxe et al., 2007). Perhaps the most compelling of these findings is the requirement of neurogenesis for contextual fear conditioning, which has been demonstrated in two species (rats and mice) using two different methods for arresting neurogenesis (Saxe et al., 2006; Winocur et al., 2006). Contextual fear conditioning is a form of Pavlovian conditioning produced by pairing a distinctive context (spatial location) with footshock. As a result of the pairing, animals exhibit characteristic fear responses (e.g., freezing, defecation, potentiation of the startle response) when re-exposed to the context. Lesions to the hippocampus often impair this form of learning (Phillips and LeDoux, 1992; Matus-Amat et al., 2004), presumably because the hippocampus participates in mnemonic encoding of the spatial context. Blocking neurogenesis prior to training in this procedure reduces the amount of the contextual fear expressed when rodents are reexposed to the training context. Importantly, blocking neurogenesis does not impair fear conditioning to a discrete tone stimulus, indicating that shock sensitivity and motor control of the fear response are not impaired. The impairment of contextual fear conditioning may thus reflect a

requirement of neurogenesis for the encoding of novel contexts and/or for assigning emotional valence to contexts. How (and whether) these learning impairments relate to depression is unclear. Cognitive impairments have been reported in depression, but cognitive impairment is not a cardinal feature of the disease, in contrast to some other psychiatric illnesses, namely schizophrenia. Still, it is certainly the case that depressed patients have an impaired ability to ‘‘contextualize’’ negative emotions, in that these emotions are overgeneralized across experiences and situations. Much more research into the putative role of neurogenesis in contextual learning will be needed to determine whether reduced neurogenesis could give rise to these features of depression.

Neurogenesis is required for some behavioral effects of AD drugs Although animal models have not provided evidence that reduced neurogenesis causes depression-like symptoms, these models have produced evidence that neurogenesis is involved in the therapeutic effects of AD drugs. Three recent studies have used the targeted irradiation procedure described above to test whether DG neurogenesis is required for the behavioral effects of AD drugs in rodent models. A study conducted in our laboratory demonstrated that neurogenesis is required for the effects of both imipramine, a classic tricyclic AD, and fluoxetine in two mouse behavioral screens for AD activity (Santarelli et al., 2003), the NSF test and a chronic stress procedure. Importantly, this study has been replicated in our laboratory using the GFAP-TK mice (unpublished results). In a separate series of experiments conducted by another laboratory, the synthetic cannabinoid HU210 was shown to have AD-like effects in the NSF paradigm following 10 days of treatment, and, interestingly, this effect was blocked by X-ray irradiation (Jiang et al., 2005). In addition, there is a recent preliminary report using rats (Meltzer et al., 2006) that irradiation blocks the behavioral effects of fluoxetine in the

713

forced swim test. Thus, a neurogenic dependence for the behavioral effects of ADs has been revealed for three different drugs in three different AD screens and using two different ways to ablate neurogenesis. The above work suggests that neurogenesis may be a critical substrate for AD efficacy. However, an important limitation of this work is the reliance on a very limited number of animal models and AD treatments. It is unlikely that the three behavioral assays used in these experiments capture all the clinically relevant features of AD treatments, and consequently it remains possible that some clinically important features of these treatments are neurogenesis-independent. Indeed, two more recent studies from our lab confirm that this caveat is valid. One study from our laboratory (Holick et al., 2007) examined the effects of AD drugs on behavior and DG neurogenesis in the Balb-c mouse strain, a strain that exhibits high anxiety in behavioral tests. In this strain, chronic fluoxetine treatment reduced anxiety-like and depressive behavior in the novelty-induced hypophagia and forced-swim paradigms but failed to increase neuronal proliferation. Not surprisingly, the behavioral effects of these drugs are not blocked by irradiation in this strain. A second study found that the anxiolytic effects of environmental enrichment do not require neurogenesis (Meshi et al., 2006). In sum, these studies suggest that AD-like effects can be achieved through at least two different pathways, one that is neurogenesis-dependent and one that is neurogenesis-independent. AD drugs and cannabinoids appear to use a neurogenesis-dependent pathway in some circumstances but not in others; chronic stress may be one factor that governs this dichotomy. Enrichment appears to use a neurogenesis-independent pathway either alone or in combination with the neurogenesisdependent pathway. An important outstanding question not addressed by these experiments is whether the upregulation of neurogenesis is sufficient to produce AD effects. Testing this hypothesis will require new methods for very specifically upregulating aspects of neurogenesis.

Summary General insights: is neurogenesis a missing link or is the link still missing? Research on MDD in the last decade has led to considerable maturation of our understanding of how different neural circuits function in a normal brain and in the context of pathology. Several notable findings have emerged from studies in humans with MDD and preclinical models of MDD. First, the AD response and the pathogenesis of MDD may have different neural substrates. Second, the pathogenic mechanisms may differ from those that underlie the pathophysiology of MDD. Third, any model explaining the etiology of MDD must incorporate genetic vulnerability, stressors, critical periods and multiple neural circuits. In other words, MDD is more likely than not, a result of multiple ‘‘hits’’ to the brain (deficits in multiple neural substrates). This last finding underscores the need for more refined preclinical models for depression. In this section, we revisit the neurogenic hypothesis of MDD with attention to the evidence discussed thus far in the context of etiology, pathophysiology and treatment.

The neurogenic hypothesis and the etiology of MDD Only recently the neurogenic hypothesis of MDD joined the neurotrophic and network hypotheses as a candidate to explain the neurobiological basis of MDD. Unlike the neurotrophic and network hypothesis, however, the neurogenic hypothesis implicates only one brain region, the dentate gyrus, as the primary neural substrate for the etiology of MDD and the AD response. As this review makes clear, the evidence for neurogenesis as an etiological factor in MDD is scarce at best. Pathological analyses of postmortem tissue obtained from patients with MDD have not revealed alterations in the size of the dentate gyrus, decreased number of proliferative cells, or changes in the degree of ongoing apoptosis. The data on apoptosis do not indicate the age of neurons that

714

Fig. 2. A model to evaluate the role of neurogenesis as a substrate in the etiology, pathophysiology and treatment of MDD. MDD is likely to arise from the synergistic effects of stress, biological vulnerability conferred by risk alleles and deficits in multiple neural circuits. Whether hippocampal neurogenesis is involved in the etiology of MDD is at present unclear. Altered hippocampal neurogenesis may result as a consequence of pathogenic mechanisms and contribute to the pathophysiology of MDD. The treatment of some, but not all, symptoms of MDD may rely on neurogenesis.

are dying and therefore, preclude an assessment of changes in rates of turnover or survival. As noted earlier in this chapter, more studies on postmortem tissue obtained from non-medicated patients are required to conclusively characterize dentate gyrus neurogenesis in depressed individuals. Nevertheless, these data establish that changes in hippocampal volume are unlikely to result from changes in hippocampal neurogenesis. Preclinical studies have yielded data more directly controverting the role of neurogenesis in the etiology of MDD: ablation of neurogenesis in adult otherwise normal animals does not engender a depression-like phenotype. However, there are some important caveats to these preclinical studies. It is almost certainly the case that MDD involves the simultaneous presence of multiple risk factors or ‘‘hits’’ (Fig. 2). If this were the case, ablation of neurogenesis in wild-type adult animals would not be sufficient to elicit a depression-like phenotype. The ablation of neurogenesis might create a diathesis for depression that would only be revealed in the presence of other genetic or environmental insults. Moreover, the timing with respect to when the brain experiences an insult, whether it be genetic or environmental, is also critical. Clearly, more studies are needed that incorporate the effects of stressors and assess the consequences of altering neurogenesis during the early postnatal period in rodents with genetic backgrounds that harbor different risk alleles. These studies will reveal whether changes in neurogenesis lend a diathesis for MDD, and inform us about the complex interplay between multiple neural circuits in

directing the emotional trajectory. Finally, longitudinal neuroimaging studies in humans are needed to reveal how the hippocampal landscape changes with the appearance of the first symptoms of MDD and over time.

The neurogenic hypothesis and the pathophysiology of MDD The increasingly appreciated role for neurogenesis in hippocampus-dependent learning as defined by work from several different laboratories using rodents provides evidence that neurogenesis makes a functionally significant contribution to hippocampal circuits. It is thus plausible that the putative reductions in neurogenesis associated with MDD have important psychological implications. These could include cognitive deficits, which are a possible mechanism of the emotional symptoms of the disease (Beck, 2005). In addition, it is now appreciated that the hippocampus, particularly its ventral (anterior, in humans) extent, has a central role in emotional regulation. The development of higher resolution neuroimaging techniques will enable us to visualize changes in dentate gyrus activity in patients with MDD and during learning. The use of novel genetic approaches to selectively manipulate the maturation of newborn neurons influence their survival, and drive turnover of mature dentate granule cells will undoubtedly enhance our understanding of how neurogenesis contributes to dentate function and to the deficits seen in MDD.

715

The neurogenic hypothesis and the treatment of MDD The finding that neurogenesis is one mechanism used by ADs to exert their behavioral effects has now been repeated by several different laboratories using rodents. This is in striking contrast to the absence of data implicating neurogenesis in the pathogenesis of MDD. One interpretation of this apparent paradox (besides the limitations of current preclinical models highlighted in section ‘‘The neurogenic hypothesis and the etiology of MDD’’) is that the etiology and treatment of MDD may have different neural substrates. Such a dissociation is suggested by a recent association study that looked at 21 candidate polymorphisms and showed that the genetic basis for the capacity to respond to monoamine-based ADs differed from that of susceptibility to MDD (Garriock et al., 2006). An interesting parallel was demonstrated by preclinical studies on BDNF and depression, which showed that increased BDNF signaling in the hippocampus is sufficient to induce AD-like effects, but genetic ablation of BDNF on its own does not elicit a depression-like phenotype (Duman and Monteggia, 2006). While the data from preclinical studies and ADs are encouraging, more work is needed to solidify the requirement for neurogenesis in mediating the behavioral effects of ADs. Conspicuously absent from the roster of experiments are those that show that increased neurogenesis is sufficient for the behavioral effects of ADs. Experiments along these lines using genetic approaches to specifically increase the number of newly generated neurons are currently underway in our laboratory. Moreover, given the pleiotropic effects of ADs on neurogenesis, it at present unclear whether immature neurons or the adult-generated mature dentate granule cells are required to induce behavioral change. Inducible genetic approaches using promoters specific to these different cell types will allow for cell type specific manipulations to unequivocally identify the cellular substrates and mechanisms underlying the neurogenic-dependent AD response. In interpreting the data on the neurogenic dependence of ADs, we must remind ourselves of the

potentially different ways by which ADs may rely on neurogenesis for their behavioral effects (Fig. 1, Drew and Hen, in press). It is plausible that the short-term effects of ADs, for example, are mediated by ADs acting on the extant reservoir of adult-generated immature neurons or by accelerating the maturation process of an extant population of adult-generated immature neurons with concomitant replacement/addition to the dentate gyrus. Likewise, the rapid effects of ADs in reversing behavioral changes induced by stress in paradigms such as LH discussed earlier could also be mediated through the extant population of immature neurons or through adult-generated mature dentate granule cells. In this regard, it is worthy to note that inducing BDNF and CREB expression in the DG but not other hippocampal subfields (or other brain regions) is sufficient to elicit an AD response (Chen et al., 2001; Shirayama et al., 2002; Duman and Monteggia, 2006). Thus, the DG is an attractive neural substrate for the AD response and these studies highlight a previously unrecognized function for the dentate gyrus. Given our present understanding of DG function (see chapter by Kesner in this volume), it is unclear how enhancement in processes such as pattern separation and conjunctive encoding contribute to AD-like behavior assessed in these depression paradigms. Whether cognitive behavior therapy, an endorsed line of treatment often used in parallel with AD drugs, increases activity in the DG is yet to be determined. In conclusion, there is much work to be done on hippocampal neurogenesis to ascertain its role in the brain and still much more to establish a role for it in MDD. The convergence of insights from preclinical studies and neuroimaging studies and identifi cation of novel genetic risk alleles will undoubtedly help establish whether the neurogenic hypothesis can explain facets of this complex and debilitating psychiatric disorder. At the very least, the neurogenic hypothesis, like any elegant hypothesis, has succeeded in generating the momentum needed to rigorously test its tenets.

716

Acknowledgments The authors would like to thank members of the Hen laboratory for helpful discussions. Funding support was provided by NARSAD (A.S, M.R.D and R.H) and by Charles H. Revson Foundation Senior Fellowship in Biomedical Science (M.R.D).

References Aberg, M.A., Aberg, N.D., Hedbacker, H., Oscarsson, J. and Eriksson, P.S. (2000) Peripheral infusion of IGF-I selectively induces neurogenesis in the adult rat hippocampus. J. Neurosci., 20: 2896–2903. Alonso, R., Griebel, G., Pavone, G., Stemmelin, J., Le Fur, G. and Soubrie, P. (2004) Blockade of CRF(1) or V(1b) receptors reverses stress-induced suppression of neurogenesis in a mouse model of depression. Mol. Psychiatry, 9: 278–286 224. Altman, J. and Das, G.D. (1965) Autoradiographic and histological evidence of postnatal hippocampal neurogenesis in rats. J. Comp. Neurol., 124: 319–335. Amrein, I., Slomianka, L. and Lipp, H.P. (2004) Granule cell number, cell death and cell proliferation in the dentate gyrus of wild-living rodents. Eur. J. Neurosci., 20: 3342–3350. American Psychiatric Association (2000) Diagnostic and Statistical Manual of Mental Disorders. Fourth Edition—Text Revision (DSMIV-TR). Austin, M.P., Mitchell, P. and Goodwin, G.M. (2001) Cognitive deficits in depression: possible implications for functional neuropathology. Br. J. Psychiatry, 178: 200–206. Azmitia, E.C. and Segal, M. (1978) An autoradiographic analysis of the differential ascending projections of the dorsal and median raphe nuclei in the rat. J. Comp. Neurol., 179: 641–667. Babyak, M., Blumenthal, J.A., Herman, S., Khatri, P., Doraiswamy, M., Moore, K., Craighead, W.E., Baldewicz, T.T. and Krishnan, K.R. (2000) Exercise treatment for major depression: maintenance of therapeutic benefit at 10 months. Psychosom. Med., 62: 633–638. Banasr, M., Hery, M., Printemps, R. and Daszuta, A. (2004) Serotonin-induced increases in adult cell proliferation and neurogenesis are mediated through different and common 5HT receptor subtypes in the dentate gyrus and the subventricular zone. Neuropsychopharmacology, 29: 450–460. Bannerman, D.M., Grubb, M., Deacon, R.M., Yee, B.K., Feldon, J. and Rawlins, J.N. (2003) Ventral hippocampal lesions affect anxiety but not spatial learning. Behav. Brain Res., 139: 197–213. Bayer, S.A., Yackel, J.W. and Puri, P.S. (1982) Neurons in the rat dentate gyrus granular layer substantially increase during juvenile and adult life. Science, 216: 890–892. Beck, A.T. (2005) The current state of cognitive therapy: a 40-year retrospective. Arch. Gen. Psychiatry, 62: 953–959.

Becker, S. (2005) A computational principle for hippocampal learning and neurogenesis. Hippocampus, 15: 722–738. Bodnoff, S.R., Suranyi-Cadotte, B., Quirion, R. and Meaney, M.J. (1989) A comparison of the effects of diazepam versus several typical and atypical anti-depressant drugs in an animal model of anxiety. Psychopharmacology (Berl.), 97: 277–279. Bratt, A.M., Kelley, S.P., Knowles, J.P., Barrett, J., Davis, K., Davis, M. and Mittleman, G. (2001) Long term modulation of the HPA axis by the hippocampus. Behavioral, biochemical and immunological endpoints in rats exposed to chronic mild stress. Psychoneuroendocrinology, 26: 121–145. Bremner, J.D., Narayan, M., Anderson, E.R., Staib, L.H., Miller, H.L. and Charney, D.S. (2000) Hippocampal volume reduction in major depression. Am. J. Psychiatry, 157: 115–118. Brezun, J.M. and Daszuta, A. (1999) Depletion in serotonin decreases neurogenesis in the dentate gyrus and the subventricular zone of adult rats. Neuroscience, 89: 999–1002. Brezun, J.M. and Daszuta, A. (2000) Serotonin may stimulate granule cell proliferation in the adult hippocampus, as observed in rats grafted with foetal raphe neurons. Eur. J. Neurosci., 12: 391–396. Bunney Jr., W.E. and Davis, J.M. (1965) Norepinephrine in depressive reactions. A review. Arch. Gen. Psychiatry, 13: 483–494. Cameron, H.A. and Gould, E. (1994) Adult neurogenesis is regulated by adrenal steroids in the dentate gyrus. Neuroscience, 61: 203–209. Cameron, H.A. and McKay, R.D. (1999) Restoring production of hippocampal neurons in old age. Nat. Neurosci., 2: 894–897. Cameron, H.A. and McKay, R.D. (2001) Adult neurogenesis produces a large pool of new granule cells in the dentate gyrus. J. Comp. Neurol., 435: 406–417. Cameron, H.A., Woolley, C.S. and Gould, E. (1993) Adrenal steroid receptor immunoreactivity in cells born in the adult rat dentate gyrus. Brain Res., 611: 342–346. Campbell, S., Marriott, M., Nahmias, C. and Macqueen, G.M. (2004) Lower hippocampal volume in patients suffering from depression: a meta-analysis. Am. J. Psychiatry, 161: 598–607. Carlezon Jr., W.A., Duman, R.S. and Nestler, E.J. (2005) The many faces of CREB. Trends Neurosci., 28: 436–445. Carroll, B.J., Martin, F.I. and Davies, B. (1968a) Pituitaryadrenal function in depression. Lancet, 1: 1373–1374. Carroll, B.J., Martin, F.I. and Davies, B. (1968b) Resistance to suppression by dexamethasone of plasma 11-O. H.C.S. levels in severe depressive illness. Br. Med. J., 3: 285–287. Caspi, A. and Moffitt, T.E. (2006) Gene-environment interactions in psychiatry: joining forces with neuroscience. Nat. Rev. Neurosci., 7: 583–590. Caspi, A., Sugden, K., Moffitt, T.E., Taylor, A., Craig, I.W., Harrington, H., Mcclay, J., Mill, J., Martin, J., Braithwaite, A. and Poulton, R. (2003) Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science, 301: 386–389.

717 Castren, E. (2005) Is mood chemistry? Nat. Rev. Neurosci., 6: 241–246. Chambers, R.A., Potenza, M.N., Hoffman, R.E. and Miranker, W. (2004) Simulated apoptosis/neurogenesis regulates learning and memory capabilities of adaptive neural networks. Neuropsychopharmacology, 29: 747–758. Chapman, D.P., Whitfield, C.L., Felitti, V.J., Dube, S.R., Edwards, V.J. and Anda, R.F. (2004) Adverse childhood experiences and the risk of depressive disorders in adulthood. J. Affect. Disord., 82: 217–225. Chen, A.C., Shirayama, Y., Shin, K.H., Neve, R.L. and Duman, R.S. (2001) Expression of the cAMP response element binding protein (CREB) in hippocampus produces an antidepressant effect. Biol. Psychiatry, 49: 753–762. Chen, G., Rajkowska, G., Du, F., Seraji-Bozorgzad, N. and Manji, H.K. (2000) Enhancement of hippocampal neurogenesis by lithium. J. Neurochem., 75: 1729–1734. Chourbaji, S., Zacher, C., Sanchis-Segura, C., Dormann, C., Vollmayr, B. and Gass, P. (2005) Learned helplessness: validity and reliability of depressive-like states in mice. Brain Res. Brain Res. Protoc., 16: 70–78. Cryan, J.F., Markou, A. and Lucki, I. (2002) Assessing antidepressant activity in rodents: recent developments and future needs. Trends Pharmacol. Sci., 23: 238–245. Czeh, B., Michaelis, T., Watanabe, T., Frahm, J., de Biurrun, G., van Kampen, M., Bartolomucci, A. and Fuchs, E. (2001) Stress-induced changes in cerebral metabolites, hippocampal volume, and cell proliferation are prevented by antidepressant treatment with tianeptine. Proc. Natl. Acad. Sci. U.S.A., 98: 12796–12801. Czeh, B., Welt, T., Fischer, A.K., Erhardt, A., Schmitt, W., Muller, M.B., Toschi, N., Fuchs, E. and Keck, M.E. (2002) Chronic psychosocial stress and concomitant repetitive transcranial magnetic stimulation: effects on stress hormone levels and adult hippocampal neurogenesis. Biol. Psychiatry, 52: 1057–1065. Drevets, W.C., Bogers, W. and Raichle, M.E. (2002) Functional anatomical correlates of antidepressant drug treatment assessed using PET measures of regional glucose metabolism. Eur. Neuropsychopharmacol., 12: 527–544. Drew, M.R. and Hen, R. (in press) Adult hippocampal neurogenesis as a target for the treatment of depression. CNS Neurol. Disord. Drug Targets. Dulawa, S.C., Holick, K.A., Gundersen, B. and Hen, R. (2004) Effects of chronic fluoxetine in animal models of anxiety and depression. Neuropsychopharmacology, 29: 1321–1330. Duman, R.S. (2002) Structural alterations in depression: cellular mechanisms underlying pathology and treatment of mood disorders. CNS Spectrosc., 7: 140–142 144–147. Duman, R.S. (2004) Depression: a case of neuronal life and death? Biol. Psychiatry, 56: 140–145. Duman, R.S., Heninger, G.R. and Nestler, E.J. (1997) A molecular and cellular theory of depression. Arch. Gen. Psychiatry, 54: 597–606. Duman, R.S., Malberg, J., Nakagawa, S. and D’sa, C. (2000) Neuronal plasticity and survival in mood disorders. Biol. Psychiatry, 48: 732–739.

Duman, R.S. and Monteggia, L.M. (2006) A neurotrophic model for stress-related mood disorders. Biol. Psychiatry, 59: 1116–1127. Dunn, J.D. and Orr, S.E. (1984) Differential plasma corticosterone responses to hippocampal stimulation. Exp. Brain Res., 54: 1–6. Dupret, D., Montaron, M.F., Drapeau, E., Aurousseau, C., Le Moal, M., Piazza, P.V. and Abrous, D.N. (2005) Methylazoxymethanol acetate does not fully block cell genesis in the young and aged dentate gyrus. Eur. J. Neurosci., 22: 778–783. Eisch, A.J. (2002) Adult neurogenesis: implications for psychiatry. Prog. Brain Res., 138: 315–342. Encinas, J.M., Vaahtokari, A. and Enikolopov, G. (2006) Fluoxetine targets early progenitor cells in the adult brain. Proc. Natl. Acad. Sci. U.S.A., 103: 8233–8238. Eriksson, P.S., Perfilieva, E., Bjork-Eriksson, T., Alborn, A.M., Nordborg, C., Peterson, D.A. and Gage, F.H. (1998) Neurogenesis in the adult human hippocampus. Nat. Med., 4: 1313–1317. Esposito, M.S., Piatti, V.C., Laplagne, D.A., Morgenstern, N.A., Ferrari, C.C., Pitossi, F.J. and Schinder, A.F. (2005) Neuronal differentiation in the adult hippocampus recapitulates embryonic development. J. Neurosci., 25: 10074–10086. Faber, K.M. and Haring, J.H. (1999) Synaptogenesis in the postnatal rat fascia dentata is influenced by 5-HT1a receptor activation. Brain Res. Dev. Brain Res., 114: 245–252. Feldman, S. and Weidenfeld, J. (1993) The dorsal hippocampus modifies the negative feedback effect of glucocorticoids on the adrenocortical and median eminence CRF-41 responses to photic stimulation. Brain Res., 614: 227–232. Feldman, S. and Weidenfeld, J. (1999) Glucocorticoid receptor antagonists in the hippocampus modify the negative feedback following neural stimuli. Brain Res., 821: 33–37. Felitti, V.J., Anda, R.F., Nordenberg, D., Williamson, D.F., Spitz, A.M., Edwards, V., Koss, M.P. and Marks, J.S. (1998) Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. Am. J. Prev. Med., 14: 245–258. Filippov, V., Kronenberg, G., Pivneva, T., Reuter, K., Steiner, B., Wang, L.P., Yamaguchi, M., Kettenmann, H. and Kempermann, G. (2003) Subpopulation of nestin-expressing progenitor cells in the adult murine hippocampus shows electrophysiological and morphological characteristics of astrocytes. Mol. Cell Neurosci., 23: 373–382. van der Flier, W.M., van Buchem, M.A., Weverling-Rijnsburger, A.W., Mutsaers, E.R., Bollen, E.L., Admiraal-Behloul, F., Westendorp, R.G. and Middelkoop, H.A. (2004) Memory complaints in patients with normal cognition are associated with smaller hippocampal volumes. J. Neurol., 251: 671–675. Fossati, P., Coyette, F., Ergis, A.M. and Allilaire, J.F. (2002) Influence of age and executive functioning on verbal memory of inpatients with depression. J. Affect. Disord., 68: 261–271.

718 Freund, T.F., Gulyas, A.I., Acsady, L., Gorcs, T. and Toth, K. (1990) Serotonergic control of the hippocampus via local inhibitory interneurons. Proc. Natl. Acad. Sci. U.S.A., 87: 8501–8505. Garcia, A., Steiner, B., Kronenberg, G., Bick-Sander, A. and Kempermann, G. (2004a) Age-dependent expression of glucocorticoid- and mineralocorticoid receptors on neural precursor cell populations in the adult murine hippocampus. Aging Cell, 3: 363–371. Garcia, A.D., Doan, N.B., Imura, T., Bush, T.G. and Sofroniew, M.V. (2004b) GFAP-expressing progenitors are the principal source of constitutive neurogenesis in adult mouse forebrain. Nat. Neurosci., 7: 1233–1241. Garriock, H.A., Delgado, P., Kling, M.A., Carpenter, L.L., Burke, M., Burke, W.J., Schwartz, T., Marangell, L.B., Husain, M., Erickson, R.P. and Moreno, F.A. (2006) Number of risk genotypes is a risk factor for major depressive disorder: a case control study. Behav. Brain Funct., 2: 24. Gilbertson, M.W., Shenton, M.E., Ciszewski, A., Kasai, K., Lasko, N.B., Orr, S.P. and Pitman, R.K. (2002) Smaller hippocampal volume predicts pathologic vulnerability to psychological trauma. Nat. Neurosci., 5: 1242–1247. Gold, P.W. and Chrousos, G.P. (2002) Organization of the stress system and its dysregulation in melancholic and atypical depression: high vs. low CRH/NE states. Mol. Psychiatry, 7: 254–275. Gould, E., Beylin, A., Tanapat, P., Reeves, A. and Shors, T.J. (1999a) Learning enhances adult neurogenesis in the hippocampal formation. Nat. Neurosci., 2: 260–265. Gould, E., Reeves, A.J., Fallah, M., Tanapat, P., Gross, C.G. and Fuchs, E. (1999b) Hippocampal neurogenesis in adult Old World primates. Proc. Natl. Acad. Sci. U.S.A., 96: 5263–5267. Goursaud, A.P., Mendoza, S.P. and Capitanio, J.P. (2006) Do neonatal bilateral ibotenic acid lesions of the hippocampal formation or of the amygdala impair HPA axis responsiveness and regulation in infant rhesus macaques (Macaca mulatta)? Brain Res., 1071: 97–104. von Gunten, A., Fox, N.C., Cipolotti, L. and Ron, M.A. (2000) A volumetric study of hippocampus and amygdala in depressed patients with subjective memory problems. J. Neuropsychiatry Clin. Neurosci., 12: 493–498. von Gunten, A. and Ron, M.A. (2004) Hippocampal volume and subjective memory impairment in depressed patients. Eur. Psychiatry, 19: 438–440. Haydon, P.G. and Carmignoto, G. (2006) Astrocyte control of synaptic transmission and neurovascular coupling. Physiol. Rev., 86: 1009–1031. Hellsten, J., Wennstrom, M., Mohapel, P., Ekdahl, C.T., Bengzon, J. and Tingstrom, A. (2002) Electroconvulsive seizures increase hippocampal neurogenesis after chronic corticosterone treatment. Eur. J. Neurosci., 16: 283–290. Herman, J.P., Cullinan, W.E., Morano, M.I., Akil, H. and Watson, S.J. (1995) Contribution of the ventral subiculum to inhibitory regulation of the hypothalamo-pituitary-adrenocortical axis. J. Neuroendocrinol., 7: 475–482.

Herman, J.P., Cullinan, W.E., Young, E.A., Akil, H. and Watson, S.J. (1992) Selective forebrain fiber tract lesions implicate ventral hippocampal structures in tonic regulation of paraventricular nucleus corticotropin-releasing hormone (CRH) and arginine vasopressin (AVP) mRNA expression. Brain Res., 592: 228–238. Holick, K.A., Lee, D.C., Hen, R. and Dulawa, S.C. (2007) Behavioral effects of chronic fluoxetine in BALB/cJ mice do not require adult hippocampal neurogenesis or the serotonin 1A receptor. Neuropsychopharmacology. Holsboer, F. and Barden, N. (1996) Antidepressants and hypothalamic-pituitary-adrenocortical regulation. Endocr. Rev., 17: 187–205. Holsboer, F., Liebl, R. and Hofschuster, E. (1982) Repeated dexamethasone suppression test during depressive illness. Normalisation of test result compared with clinical improvement. J. Affect. Disord., 4: 93–101. Huot, R.L., Plotsky, P.M., Lenox, R.H. and Mcnamara, R.K. (2002) Neonatal maternal separation reduces hippocampal mossy fiber density in adult Long Evans rats. Brain Res., 950: 52–63. Huot, R.L., Thrivikraman, K.V., Meaney, M.J. and Plotsky, P.M. (2001) Development of adult ethanol preference and anxiety as a consequence of neonatal maternal separation in Long Evans rats and reversal with antidepressant treatment. Psychopharmacology (Berl.), 158: 366–373. Islam, A., Ayer-Lelievre, C., Heigenskold, C., Bogdanovic, N., Winblad, B. and Adem, A. (1998) Changes in IGF-1 receptors in the hippocampus of adult rats after long-term adrenalectomy: receptor autoradiography and in situ hybridization histochemistry. Brain Res., 797: 342–346. Jacobs, B.L., Praag, H. and Gage, F.H. (2000) Adult brain neurogenesis and psychiatry: a novel theory of depression. Mol. Psychiatry, 5: 262–269. Jayatissa, M.N., Bisgaard, C., Tingstrom, A., Papp, M. and Wiborg, O. (2006) Hippocampal cytogenesis correlates to escitalopram-mediated recovery in a chronic mild stress rat model of depression. Neuropsychopharmacology, 31: 2395–2404. Jiang, W., Zhang, Y., Xiao, L., van Cleemput, J., Ji, S.P., Bai, G. and Zhang, X. (2005) Cannabinoids promote embryonic and adult hippocampus neurogenesis and produce anxiolyticand antidepressant-like effects. J. Clin. Invest., 115: 3104–3116. Joels, M. and van Riel, E. (2004) Mineralocorticoid and glucocorticoid receptor-mediated effects on serotonergic transmission in health and disease. Ann. N.Y. Acad. Sci., 1032: 301–303. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1997) More hippocampal neurons in adult mice living in an enriched environment. Nature, 386: 493–495. Kempermann, G., Kuhn, H.G. and Gage, F.H. (1998) Experience-induced neurogenesis in the senescent dentate gyrus. J. Neurosci., 18: 3206–3212. Kendler, K.S., Kuhn, J.W., Vittum, J., Prescott, C.A. and Riley, B. (2005) The interaction of stressful life events and a serotonin transporter polymorphism in the prediction of

719 episodes of major depression: a replication. Arch. Gen. Psychiatry, 62: 529–535. Kendler, K.S., Thornton, L.M. and Gardner, C.O. (2001) Genetic risk, number of previous depressive episodes, and stressful life events in predicting onset of major depression. Am. J. Psychiatry, 158: 582–586. Kennedy, S.H., Evans, K.R., Kruger, S., Mayberg, H.S., Meyer, J.H., Mccann, S., Arifuzzman, A.I., Houle, S. and Vaccarino, F.J. (2001) Changes in regional brain glucose metabolism measured with positron emission tomography after paroxetine treatment of major depression. Am. J. Psychiatry, 158: 899–905. Kimbrell, T.A., Ketter, T.A., George, M.S., Little, J.T., Benson, B.E., Willis, M.W., Herscovitch, P. and Post, R.M. (2002) Regional cerebral glucose utilization in patients with a range of severities of unipolar depression. Biol. Psychiatry, 51: 237–252. Knigge, K.M. (1961) Adrenocortical response to stress in rats with lesions in hippocampus and amygdala. Proc. Soc. Exp. Biol. Med., 108: 18–21. Kohler, S.J., Jennings, V., Boklewski, J.L., Degrush, B.J., Williams, N.I., Rockcastle, N.J., Cameron, J.L. and Greenough, W.T. (2006) Maturation of new granule cells in the dentate gyrus of adult macaque monkeys. Soc. Neurosci. Abstr. Kondoh, M., Shiga, T. and Okado, N. (2004) Regulation of dendrite formation of Purkinje cells by serotonin through serotonin1A and serotonin2A receptors in culture. Neurosci. Res., 48: 101–109. Kornack, D.R. and Rakic, P. (1999) Continuation of neurogenesis in the hippocampus of the adult macaque monkey. Proc. Natl. Acad. Sci. U.S.A., 96: 5768–5773. Kuhn, H.G., Dickinson-Anson, H. and Gage, F.H. (1996) Neurogenesis in the dentate gyrus of the adult rat: age-related decrease of neuronal progenitor proliferation. J. Neurosci., 16: 2027–2033. Kuhn, H.G., Winkler, J., Kempermann, G., Thal, L.J. and Gage, F.H. (1997) Epidermal growth factor and fibroblast growth factor-2 have different effects on neural progenitors in the adult rat brain. J. Neurosci., 17: 5820–5829. Kvachnina, E., Liu, G., Dityatev, A., Renner, U., Dumuis, A., Richter, D.W., Dityateva, G., Schachner, M., VoynoYasenetskaya, T.A. and Ponimaskin, E.G. (2005) 5-HT7 receptor is coupled to G alpha subunits of heterotrimeric G12-protein to regulate gene transcription and neuronal morphology. J. Neurosci., 25: 7821–7830. Laplagne, D.A., Esposito, M.S., Piatti, V.C., Morgenstern, N.A., Zhao, C., van Praag, H., Gage, F.H. and Schinder, A.F. (2006) Functional convergence of neurons generated in the developing and adult hippocampus. PLoS Biol., 4: e409. Lauder, J.M. and Krebs, H. (1978) Serotonin as a differentiation signal in early neurogenesis. Dev. Neurosci., 1: 15–30. Leonardo, E.D. and Hen, R. (2006) Genetics of affective and anxiety disorders. Annu. Rev. Psychol., 57: 117–137. Levinson, D.F. (2006) The genetics of depression: a review. Biol. Psychiatry, 60: 84–92.

Liu, S., Wang, J., Zhu, D., Fu, Y., Lukowiak, K. and Lu, Y.M. (2003) Generation of functional inhibitory neurons in the adult rat hippocampus. J. Neurosci., 23: 732–736. Loo, C.K. and Mitchell, P.B. (2005) A review of the efficacy of transcranial magnetic stimulation (TMS) treatment for depression, and current and future strategies to optimize efficacy. J. Affect. Disord., 88: 255–267. Lovic, V., Gonzalez, A. and Fleming, A.S. (2001) Maternally separated rats show deficits in maternal care in adulthood. Dev. Psychobiol., 39: 19–33. Lucassen, P.J., Muller, M.B., Holsboer, F., Bauer, J., Holtrop, A., Wouda, J., Hoogendijk, W.J., De Kloet, E.R. and Swaab, D.F. (2001) Hippocampal apoptosis in major depression is a minor event and absent from subareas at risk for glucocorticoid overexposure. Am. J. Pathol., 158: 453–468. Ma, D.K., Ming, G.L. and Song, H. (2005) Glial influences on neural stem cell development: cellular niches for adult neurogenesis. Curr. Opin. Neurobiol., 15: 514–520. Macqueen, G.M., Campbell, S., Mcewen, B.S., Macdonald, K., Amano, S., Joffe, R.T., Nahmias, C. and Young, L.T. (2003) Course of illness, hippocampal function, and hippocampal volume in major depression. Proc. Natl. Acad. Sci. U.S.A., 100: 1387–1392. Madsen, T.M., Treschow, A., Bengzon, J., Bolwig, T.G., Lindvall, O. and Tingstrom, A. (2000) Increased neurogenesis in a model of electroconvulsive therapy. Biol. Psychiatry, 47: 1043–1049. Malberg, J.E. and Duman, R.S. (2003) Cell proliferation in adult hippocampus is decreased by inescapable stress: reversal by fluoxetine treatment. Neuropsychopharmacology, 28: 1562–1571. Malberg, J.E., Eisch, A.J., Nestler, E.J. and Duman, R.S. (2000) Chronic antidepressant treatment increases neurogenesis in adult rat hippocampus. J. Neurosci., 20: 9104–9110. Malberg, J.E. and Schechter, L.E. (2005) Increasing hippocampal neurogenesis: a novel mechanism for antidepressant drugs. Curr. Pharm. Des., 11: 145–155. Manev, H., Uz, T., Smalheiser, N.R. and Manev, R. (2001) Antidepressants alter cell proliferation in the adult brain in vivo and in neural cultures in vitro. Eur. J. Pharmacol., 411: 67–70. Matus-Amat, P., Higgins, E.A., Barrientos, R.M. and Rudy, J.W. (2004) The role of the dorsal hippocampus in the acquisition and retrieval of context memory representations. J. Neurosci., 24: 2431–2439. Mayberg, H.S., Brannan, S.K., Tekell, J.L., Silva, J.A., Mahurin, R.K., Mcginnis, S. and Jerabek, P.A. (2000) Regional metabolic effects of fluoxetine in major depression: serial changes and relationship to clinical response. Biol. Psychiatry, 48: 830–843. Mcewen, B.S. (2005) Glucocorticoids, depression, and mood disorders: structural remodeling in the brain. Metabolism, 54: 20–23. Meltzer, L.A., Roy, M., Parente, V. and Deisseroth, K. (2006) Hippocampal neurogenesis is required for lasting antidepressant effects of fluoxetine. Soc. Neurosci. Abstr.

720 Meltzer, L.A., Yabaluri, R. and Deisseroth, K. (2005) A role for circuit homeostasis in adult neurogenesis. Trends Neurosci., 28: 653–660. Meshi, D., Drew, M.R., Saxe, M., Ansorge, M.S., David, D., Santarelli, L., Malapani, C., Moore, H. and Hen, R. (2006) Hippocampal neurogenesis is not required for behavioral effects of environmental enrichment. Nat. Neurosci., 9: 729–731. Mineur, Y.S., Belzung, C. and Crusio, W.E. (2006) Effects of unpredictable chronic mild stress on anxiety and depressionlike behavior in mice. Behav. Brain Res., 175: 43–50. Mineur, Y.S., Prasol, D.J., Belzung, C. and Crusio, W.E. (2003) Agonistic behavior and unpredictable chronic mild stress in mice. Behav. Genet., 33: 513–519. Mirescu, C., Peters, J.D. and Gould, E. (2004) Early life experience alters response of adult neurogenesis to stress. Nat. Neurosci., 7: 841–846. Monje, M.L., Mizumatsu, S., Fike, J.R. and Palmer, T.D. (2002) Irradiation induces neural precursor-cell dysfunction. Nat. Med., 8: 955–962. Monje, M.L., Toda, H. and Palmer, T.D. (2003) Inflammatory blockade restores adult hippocampal neurogenesis. Science, 302: 1760–1765. Moser, M.B. and Moser, E.I. (1998) Functional differentiation in the hippocampus. Hippocampus, 8: 608–619. Motl, R.W., Birnbaum, A.S., Kubik, M.Y. and Dishman, R.K. (2004) Naturally occurring changes in physical activity are inversely related to depressive symptoms during early adolescence. Psychosom. Med., 66: 336–342. Muller, M.B., Lucassen, P.J., Yassouridis, A., Hoogendijk, W.J., Holsboer, F. and Swaab, D.F. (2001) Neither major depression nor glucocorticoid treatment affects the cellular integrity of the human hippocampus. Eur. J. Neurosci., 14: 1603–1612. Murray, C.J. and Lopez, A.D. (1996) Evidence-based health policy — lessons from the Global Burden of Disease Study. Science, 274: 740–743. Nakagawa, S., Kim, J.E., Lee, R., Chen, J., Fujioka, T., Malberg, J., Tsuji, S. and Duman, R.S. (2002a) Localization of phosphorylated cAMP response element-binding protein in immature neurons of adult hippocampus. J. Neurosci., 22: 9868–9876. Nakagawa, S., Kim, J.E., Lee, R., Malberg, J.E., Chen, J., Steffen, C., Zhang, Y.J., Nestler, E.J. and Duman, R.S. (2002b) Regulation of neurogenesis in adult mouse hippocampus by cAMP and the cAMP response element-binding protein. J. Neurosci., 22: 3673–3682. Nestler, E.J., Barrot, M., Dileone, R.J., Eisch, A.J., Gold, S.J. and Monteggia, L.M. (2002) Neurobiology of depression. Neuron, 34: 13–25. Neumeister, A., Wood, S., Bonne, O., Nugent, A.C., Luckenbaugh, D.A., Young, T., Bain, E.E., Charney, D.S. and Drevets, W.C. (2005) Reduced hippocampal volume in unmedicated, remitted patients with major depression versus control subjects. Biol. Psychiatry, 57: 935–937. Nottebohm, F. (2002) Neuronal replacement in adult brain. Brain Res. Bull., 57: 737–749.

Oleskevich, S., Descarries, L., Watkins, K.C., Seguela, P. and Daszuta, A. (1991) Ultrastructural features of the serotonin innervation in adult rat hippocampus: an immunocytochemical description in single and serial thin sections. Neuroscience, 42: 777–791. Overmier, J.B. and Seligman, M.E. (1967) Effects of inescapable shock upon subsequent escape and avoidance responding. J. Comp. Physiol. Psychol., 63: 28–33. Overstreet-Wadiche, L.S. and Westbrook, G.L. (2006) Functional maturation of adult-generated granule cells. Hippocampus, 16: 208–215. Patel, T.D. and Zhou, F.C. (2005) Ontogeny of 5-HT1A receptor expression in the developing hippocampus. Brain Res. Dev Brain Res., 157: 42–57. Phillips, R.G. and LeDoux, J.E. (1992) Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning. Behav. Neurosci., 106: 274–285. van Praag, H., Kempermann, G. and Gage, F.H. (1999) Running increases cell proliferation and neurogenesis in the adult mouse dentate gyrus. Nat. Neurosci., 2: 266–270. Radley, J.J. and Jacobs, B.L. (2002) 5-HT1A receptor antagonist administration decreases cell proliferation in the dentate gyrus. Brain Res., 955: 264–267. Ramirez-Amaya, V., Marrone, D.F., Gage, F.H., Worley, P.F. and Barnes, C.A. (2006) Integration of new neurons into functional neural networks. J. Neurosci., 26: 12237–12241. Rao, M.S. and Shetty, A.K. (2004) Efficacy of doublecortin as a marker to analyse the absolute number and dendritic growth of newly generated neurons in the adult dentate gyrus. Eur. J. Neurosci., 19: 234–246. Reif, A., Fritzen, S., Finger, M., Strobel, A., Lauer, M., Schmitt, A. and Lesch, K.P. (2006) Neural stem cell proliferation is decreased in schizophrenia, but not in depression. Mol. Psychiatry, 11: 514–522. Rhodes, J.S., van Praag, H., Jeffrey, S., Girard, I., Mitchell, G.S., Garland Jr., T. and Gage, F.H. (2003) Exercise increases hippocampal neurogenesis to high levels but does not improve spatial learning in mice bred for increased voluntary wheel running. Behav. Neurosci., 117: 1006–1016. Sairanen, M., Lucas, G., Ernfors, P., Castren, M. and Castren, E. (2005) Brain-derived neurotrophic factor and antidepressant drugs have different but coordinated effects on neuronal turnover, proliferation, and survival in the adult dentate gyrus. J. Neurosci., 25: 1089–1094. Santarelli, L., Saxe, M., Gross, C., Surget, A., Battaglia, F., Dulawa, S., Weisstaub, N., Lee, J., Duman, R., Arancio, O., Belzung, C. and Hen, R. (2003) Requirement of hippocampal neurogenesis for the behavioral effects of antidepressants. Science, 301: 805–809. Sapolsky, R.M. (1994) The physiological relevance of glucocorticoid endangerment of the hippocampus. Ann. N.Y. Acad. Sci., 746: 294–304 discussion 304–307. Sapolsky, R.M. (2000) Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders. Arch. Gen. Psychiatry, 57: 925–935. Sapolsky, R.M., Armanini, M.P., Sutton, S.W. and Plotsky, P.M. (1989) Elevation of hypophysial portal concentrations

721 of adrenocorticotropin secretagogues after fornix transection. Endocrinology, 125: 2881–2887. Saxe, M.D., Malleret, G., Vronskaya, S., Mendez, I., Garcia, A.D., Sofroniew, M.V., Kandel, E.R. and Hen, R. (2007) Paradoxical influence of hippocampal neurogenesis on working memory. Proc. Natl. Acad. Sci. U.S.A., 104: 4642–4646. Saxe, M.D., Battaglia, F., Wang, J.W., Malleret, G., David, D.J., Monckton, J.E., Garcia, A.D., Sofroniew, M.V., Kandel, E.R., Santarelli, L., Hen, R. and Drew, M.R. (2006) Ablation of hippocampal neurogenesis impairs contextual fear conditioning and synaptic plasticity in the dentate gyrus. Proc. Natl. Acad. Sci. U.S.A., 103: 17501–17506. Saxena, S., Brody, A.L., Ho, M.L., Alborzian, S., Ho, M.K., Maidment, K.M., Huang, S.C., Wu, H.M., Au, S.C. and Baxter Jr., L.R. (2001) Cerebral metabolism in major depression and obsessive-compulsive disorder occurring separately and concurrently. Biol. Psychiatry, 50: 159–170. Schaaf, M.J., de Kloet, E.R. and Vreugdenhil, E. (2000) Corticosterone effects on BDNF expression in the hippocampus. Implications for memory formation. Stress, 3: 201–208. Scharfman, H.E., Sollas, A.L., Smith, K.L., Jackson, M.B. and Goodman, J.H. (2002) Structural and functional asymmetry in the normal and epileptic rat dentate gyrus. J. Comp. Neurol., 454: 424–439. Schildkraut, J.J. (1965) The catecholamine hypothesis of affective disorders: a review of supporting evidence. Am. J. Psychiatry, 122: 509–522. Schmidt-Hieber, C., Jonas, P. and Bischofberger, J. (2004) Enhanced synaptic plasticity in newly generated granule cells of the adult hippocampus. Nature, 429: 184–187. Seki, T. and Arai, Y. (1995) Age-related production of new granule cells in the adult dentate gyrus. Neuroreport, 6: 2479–2482. Seligman, M.E. and Beagley, G. (1975) Learned helplessness in the rat. J. Comp. Physiol. Psychol., 88: 534–541. Seminowicz, D.A., Mayberg, H.S., Mcintosh, A.R., Goldapple, K., Kennedy, S., Segal, Z. and Rafi-Tari, S. (2004) Limbicfrontal circuitry in major depression: a path modeling metanalysis. Neuroimage, 22: 409–418. Sheline, Y.I. (1996) Hippocampal atrophy in major depression: a result of depression-induced neurotoxicity? Mol. Psychiatry, 1: 298–299. Sheline, Y.I., Gado, M.H. and Kraemer, H.C. (2003) Untreated depression and hippocampal volume loss. Am. J. Psychiatry, 160: 1516–1518. Sheline, Y.I., Sanghavi, M., Mintun, M.A. and Gado, M.H. (1999) Depression duration but not age predicts hippocampal volume loss in medically healthy women with recurrent major depression. J. Neurosci., 19: 5034–5043. Shirayama, Y., Chen, A.C., Nakagawa, S., Russell, D.S. and Duman, R.S. (2002) Brain-derived neurotrophic factor produces antidepressant effects in behavioral models of depression. J. Neurosci., 22: 3251–3261. Shors, T.J., Miesegaes, G., Beylin, A., Zhao, M., Rydel, T. and Gould, E. (2001) Neurogenesis in the adult is involved in the formation of trace memories. Nature, 410: 372–376.

Silva, R., Lu, J., Wu, Y., Martins, L., Almeida, O.F. and Sousa, N. (2006) Mapping cellular gains and losses in the postnatal dentate gyrus: implications for psychiatric disorders. Exp. Neurol., 200: 321–331. Singh, N.A., Clements, K.M. and Singh, M.A. (2001) The efficacy of exercise as a long-term antidepressant in elderly subjects: a randomized, controlled trial. J. Gerontol. A Biol. Sci. Med. Sci., 56: M497–M504. Snyder, J.S., Hong, N.S., Mcdonald, R.J. and Wojtowicz, J.M. (2005) A role for adult neurogenesis in spatial long-term memory. Neuroscience, 130: 843–852. Snyder, J.S., Kee, N. and Wojtowicz, J.M. (2001) Effects of adult neurogenesis on synaptic plasticity in the rat dentate gyrus. J. Neurophysiol., 85: 2423–2431. Song, H., Kempermann, G., Overstreet Wadiche, L., Zhao, C., Schinder, A.F. and Bischofberger, J. (2005) New neurons in the adult mammalian brain: synaptogenesis and functional integration. J. Neurosci., 25: 10366–10368. Stockmeier, C.A., Mahajan, G.J., Konick, L.C., Overholser, J.C., Jurjus, G.J., Meltzer, H.Y., Uylings, H.B., Friedman, L. and Rajkowska, G. (2004) Cellular changes in the postmortem hippocampus in major depression. Biol. Psychiatry, 56: 640–650. Strange, B. and Dolan, R. (1999) Functional segregation within the human hippocampus. Mol. Psychiatry, 4: 508–511. Strange, B.A., Fletcher, P.C., Henson, R.N., Friston, K.J. and Dolan, R.J. (1999) Segregating the functions of human hippocampus. Proc. Natl. Acad. Sci. U.S.A., 96: 4034–4039. Strekalova, T., Spanagel, R., Bartsch, D., Henn, F.A. and Gass, P. (2004) Stress-induced anhedonia in mice is associated with deficits in forced swimming and exploration. Neuropsychopharmacology, 29: 2007–2017. Sullivan, P.F., Neale, M.C. and Kendler, K.S. (2000) Genetic epidemiology of major depression: review and meta-analysis. Am. J. Psychiatry, 157: 1552–1562. Tanapat, P., Hastings, N.B., Reeves, A.J. and Gould, E. (1999) Estrogen stimulates a transient increase in the number of new neurons in the dentate gyrus of the adult female rat. J. Neurosci., 19: 5792–5801. Taylor, M.J., Freemantle, N., Geddes, J.R. and Bhagwagar, Z. (2006) Early onset of selective serotonin reuptake inhibitor antidepressant action: systematic review and meta-analysis. Arch. Gen. Psychiatry, 63: 1217–1223. Tozuka, Y., Fukuda, S., Namba, T., Seki, T. and Hisatsune, T. (2005) GABAergic excitation promotes neuronal differentiation in adult hippocampal progenitor cells. Neuron, 47: 803–815. Vakili, K., Pillay, S.S., Lafer, B., Fava, M., Renshaw, P.F., Bonello-Cintron, C.M. and Yurgelun-Todd, D.A. (2000) Hippocampal volume in primary unipolar major depression: a magnetic resonance imaging study. Biol. Psychiatry, 47: 1087–1090. Videbech, P. and Ravnkilde, B. (2004) Hippocampal volume and depression: a meta-analysis of MRI studies. Am. J. Psychiatry, 161: 1957–1966. Videbech, P., Ravnkilde, B., Pedersen, A.R., Egander, A., Landbo, B., Rasmussen, N.A., Andersen, F., Stodkilde-

722 Jorgensen, H., Gjedde, A. and Rosenberg, R. (2001) The Danish PET/depression project: PET findings in patients with major depression. Psychol. Med., 31: 1147–1158. Videbech, P., Ravnkilde, B., Pedersen, T.H., Hartvig, H., Egander, A., Clemmensen, K., Rasmussen, N.A., Andersen, F., Gjedde, A. and Rosenberg, R. (2002) The Danish PET/ depression project: clinical symptoms and cerebral blood flow. A regions-of-interest analysis. Acta Psychiatr. Scand., 106: 35–44. Vollmayr, B., Simonis, C., Weber, S., Gass, P. and Henn, F. (2003) Reduced cell proliferation in the dentate gyrus is not correlated with the development of learned helplessness. Biol. Psychiatry, 54: 1035–1040. Wang, S., Scott, B.W. and Wojtowicz, J.M. (2000) Heterogenous properties of dentate granule neurons in the adult rat. J. Neurobiol., 42: 248–257. West, M.J. (1993) Regionally specific loss of neurons in the aging human hippocampus. Neurobiol. Aging, 14: 287–293. Willner, P. (2005) Chronic mild stress (CMS) revisited: consistency and behavioural-neurobiological concordance in the effects of CMS. Neuropsychobiology, 52: 90–110. Willner, P. and Mitchell, P.J. (2002) The validity of animal models of predisposition to depression. Behav. Pharmacol., 13: 169–188. Willner, P., Muscat, R. and Papp, M. (1992) Chronic mild stress-induced anhedonia: a realistic animal model of depression. Neurosci. Biobehav. Rev., 16: 525–534. Willner, P., Towell, A., Sampson, D., Sophokleous, S. and Muscat, R. (1987) Reduction of sucrose preference by chronic unpredictable mild stress, and its restoration by a tricyclic antidepressant. Psychopharmacology (Berl.), 93: 358–364.

Winocur, G., Wojtowicz, J.M., Sekeres, M., Snyder, J.S. and Wang, S. (2006) Inhibition of neurogenesis interferes with hippocampus-dependent memory function. Hippocampus, 16: 296–304. Wiskott, L., Rasch, M.J. and Kempermann, G. (2006) A functional hypothesis for adult hippocampal neurogenesis: avoidance of catastrophic interference in the dentate gyrus. Hippocampus, 16: 329–343. Wojtowicz, J.M. (2006) Irradiation as an experimental tool in studies of adult neurogenesis. Hippocampus, 16: 261–266. Wong, E.Y. and Herbert, J. (2004) The corticoid environment: a determining factor for neural progenitors’ survival in the adult hippocampus. Eur. J. Neurosci., 20: 2491–2498. Wong, E.Y. and Herbert, J. (2006) Raised circulating corticosterone inhibits neuronal differentiation of progenitor cells in the adult hippocampus. Neuroscience, 137: 83–92. Yalcin, I., Aksu, F. and Belzung, C. (2005) Effects of desipramine and tramadol in a chronic mild stress model in mice are altered by yohimbine but not by pindolol. Eur. J. Pharmacol., 514: 165–174. Yan, W., Wilson, C.C. and Haring, J.H. (1997) Effects of neonatal serotonin depletion on the development of rat dentate granule cells. Brain Res. Dev. Brain Res., 98: 177–184. Yuen, E.Y., Jiang, Q., Chen, P., Gu, Z., Feng, J. and Yan, Z. (2005a) Serotonin 5-HT1A receptors regulate NMDA receptor channels through a microtubule-dependent mechanism. J. Neurosci., 25: 5488–5501. Yuen, E.Y., Jiang, Q., Feng, J. and Yan, Z. (2005b) Microtubule regulation of N-methyl-D-aspartate receptor channels in neurons. J. Biol. Chem., 280: 29420–29427.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 39

The dentate gyrus in Alzheimer’s disease Thomas G. Ohm Institute of Integrative Neuroanatomy, Department of Clinical Cell and Neurobiology, Charite´ CCM, 10098 Berlin, Germany

Abstract: As part of the hippocampus, the dentate gyrus is considered to play a crucial role in associative memory. The reviewed data suggest that the dentate gyrus withstands the formation of plaques, tangles and neuronal death until late stages of Alzheimer’s disease (AD). However, changes related to a disconnecting process, and more subtle intrinsic alterations, may contribute to disturbances in memory and learning observed in early stages of AD. Keywords: hippocampus; granule cells; interneuron; Alzheimer’s disease; dentate gyrus; review originating from 650,000 to 1,100,000 (Gomez Isla et al., 1996; West and Slomianka, 1998; von Gunten et al., 2006) excitatory entorhinal projection neurons located in the pre-a layer or — in a different nomenclature — layer II. Pre-a belongs to the lamina principalis externa, the superimposed concept for the superficial layers, which is separated by the stratum dissecans from the deeper layers, together called lamina principalis interna. Perforant path axons end preferentially within the outer two-thirds of the superficial molecular layer mainly on the (apical) dendrites of the granule cells, but also on dendrites of interneurons such as the basket cells. Origins of other important afferents are cholinergic neurons of the basal forebrain mainly located in the medial septal and diagonal band nucleus and association fibres from the mossy cells of sector CA4. These afferents are also attracted to the inner molecular layer where they terminate together with axons stemming from the reunion nucleus of the thalamus. In humans, the relative proportion of the respective input is not exactly known. Furthermore, the wide-spreading projection from noradrenergic and serotoninergic nuclei of the brain stem (nucleus coeruleus and

Basic anatomy The human dentate gyrus belongs to the heterogeneously composed allocortex, which is located mainly in the antero-medial temporal lobe. As part of the hippocampus, the dentate gyrus is considered to play a crucial role in associative memory, especially with respect to events (‘what happens’) (Morris, 2006). To perform this task, the dentate gyrus is highly organized and numerous synaptic interactions take place involving at least 11 types of interneurons, which form an intrinsic network (Freund and Buzsaki, 1996). In cognitively normal humans,
9–22 million (mean: 14.6) (Seress, 1988; West and Gundersen, 1990; West et al., 1994; Bobinski et al., 1996a, 1997; Simic et al., 1997; Harding et al., 1998; Korbo et al., 2004) densely packed principal (projection) neurons, the granule cells, form the narrow granule cell layer. The major input of the dentate gyrus comes through the perforant path representing axons Corresponding author. Tel.: +49 (0)30 450 528202; Fax: +49 (0)30 450 528913; E-mail: thomas_georg.ohm@ charite.de DOI: 10.1016/S0079-6123(07)63039-8

723

724

oral raphe nuclei), along with some hypothalamic (supra- and/or tuberomamillary) projections form a profuse input to the molecular layer of the dentate gyrus. Neurons from the pre-b layer (also termed layer III) uses the perforant path for a direct projection to the sector CA1 of the Ammon’s horn (probably encoding information used for spatial memory (‘where is it happening’) (Morris, 2006). Here the NMDA receptor-mediated synaptic plasticity integrates the spatial- and event-related associative memory components. Several findings suggest that the CA3 pyramidal cells, which receive the granule cell output, aid at ‘pattern completion’ if recall of spatial information is challenged under situations of incomplete data about the environment (Nakazawa et al., 2002). A hippocampal commissural system (psalterium) is virtually absent (Amaral et al., 1984; Demeter et al., 1985) and only a scanty callosal input arrives from the contralateral hippocampus. This is in contrast to a rodent brain and the classical descriptions made by, e.g., Ko¨lliker or de No´. The granule cells convey the processed information via the mossy fibers, i.e., the granule cells’ axons, to the cornu ammonis (CA). Here they terminate with unusually large and zinc-rich synaptic boutons on proximal dendrites of pyramidal cells of the sector CA3 of the hippocampus, thereby forming the stratum lucidum. Mossy fiber collaterals are abound, terminating on inhibitory local circuit neurons of the dentate gyrus’s polymorphic layer and on the glutamatergic mossy cells ( ¼ modified hilar/CA4 pyramidal cells with ostentatious dendritic excrescences). In monkey, mossy fiber collaterals terminate on basal dendrites of granule cells, which may be the morphological basis for a recurrent excitatory feedback loop (Seress and Frotscher, 1990). Others, however, when performing biocytin-tracing on human hippocampal tissue obtained from surgery, did not find evidence for this (Lim et al., 1997). The local interneuron’s axon terminates with inhibitory synapses on the dendrites of granule cells or on other interneurons. The granule cells mostly direct their dendrites into the molecular layer (apical dendrites) but in one-third (Lim et al., 1997), — much more abundant than in rodents — also into the polymorphic or plexiform layer (basal dendrites) (Seress and Mrzljak, 1987). The axon of

the mossy cell synapses with the dendrite of the basket cell, a pyramid-shaped GABAergic, parvalbumin-containing interneuron. Their short axons profusely give rise of collaterals in the granule cell layer where these branches form inhibitory axosomatic synapses with the granule cells. Common to both projection neurons and interneurons is their richness in calcium-binding proteins. Granule cells contain calbindin and calmyrin, whereas calbindin, parvalbumin or calretinin is found in interneurons.

Alzheimer’s disease-related histopathology The formation and deposition of certain aggregated proteins is most striking among the several histological changes tied up to Alzheimer’s disease (AD). Normally, aggregation of these proteins does not occur, and even more surprisingly, this material is only rarely found in animals. Two proteins seem to be central: tau and Ab. Tau is a microtubule associated, primarily neuronal-axonal protein involved in microtubule assembly and stabilisation, processes controlled by the degree of tau’s phosphorylation (normal 2–3 mol phosphate per mol tau) and OGlcNAcetylation. In humans, alternative splicing of a single gene generates six isoforms, all of which are normally among the most soluble proteins known so far, and all of which are involved in AD (Spillantini and Goedert, 1998). After hyperphosphorylation of some of the
30 potential phosphorylation sites tau tends to self-polymerise. Moderate hyperphosphorylation (4–6 mol phosphate per mol tau) results in sequestration of normal tau whereas higher phosphorylation (10 mol phosphate per mol tau) is associated with filament formation. In AD tau is hyperphosphorylated and forms highly characteristic paired helical filaments (PHF; interval between two twists 80 nm, closest width 10 nm and broadest width 20 nm) and — to a lesser amount — also straight filaments. These filaments represent the protein backbone of intraneuronal histopathological features, i.e., tangles and neuropil threads (AD-related ‘tau-pathology’). These structures can be stained selectively with highly sensitive silver stains such as the Gallyas

725

stain that was proven to be as sensitive and selective as immunostains (Scho¨nheit et al., 2004). Immunostains, however, allow the determination of the hierarchically occurring phosphorylation and thereby the staging of tangles (Uchihara et al., 2001; Augustinack et al., 2002) unveiling that changes in tau phosphorylation precede tangle formation (Bancher et al., 1989; Braak et al., 1994). Tangles occur in somata of neurons, which may live for years (Bobinski et al., 1998) or even decades with a tangle (Morsch et al., 1999). When the tangle-bearing neuron has deceased, inert remnants remain for quite some time, termed ‘ghost tangles’. Eventually, they become cleared by astrocytes. Ghost tangles are usually less-densely twisted, less argyrophilic, affine for acid dyes and Congo red, ubiquitinated and even immunoreactive against anti-Ab-antibodies. Neuropil threads are likewise silver-stainable filaments but primarily located within the dendrites of neurons, often before the respective soma has formed a tangle (Braak et al., 1986). The Alzheimer-related tau pathology shows a regular, highly area-specific, lamina-specific and cell type-specific spread through the cortex allowing their classification into six stages. Neither neuritic plaques nor any other plaque subtype shows a similar degree of reliability in the distribution and spread. The Braak-staging [for more details see: (Braak and Braak, 1991)], developed on Gallyasstained tissue, distinguishes two ‘entorhinal’ (stages I and II), two ‘limbic’ (stages III and IV) and two ‘isocortical stages’ (stages V and VI). The stages I–II do not show clinical or insipient signs of AD (Bancher et al., 1993; Braak et al., 1993). The stage III displays a severe involvement of layers pre-a of the transentorhinal and entorhinal cortices including first occurrence of ghost tangles as sign of beginning neuronal loss and deafferentiation of the dentate gyrus. Stage IV is marked off by an abundance of ghost tangles in layer pre-a, severer tangle formation in layer pri-a and now commencing tangle formation in layer pre-b. Stages V and VI meet the classical neuropathological criteria of AD (Khatchaturian, 1985; Jellinger, 1997; Jellinger and Bancher., 1997) and generally are associated with dementia. At stage VI the entorhinal cortex may have an almost denuded layer pre-a, due to

the death of more than 90% of the neurons (Gomez Isla et al., 1996). Ghost tangles are numerous and may even be cleared in considerable numbers from the neuropil by astrocytes. This may feign a non-tangle related neuron loss (Kril et al., 2002). In terms of time, the development of these six stages was estimated to take almost five decades (Ohm et al., 1995). Although many neurons of the cortex may have developed a tangle in these late stages, it is to note that only few of the many nerve cell types are vulnerable (Braak and Del Tredici, 2004). Ab is physiologically formed by proteolytic processing of a receptor-like transmembrane protein (APP, amyloid precursor protein) (Dyrks et al., 1988) through a sequential action of socalled beta- and gamma-secretases (Haass, 2004). APP is member of a larger superfamily including several splice-variants of APP and APP-like proteins (APPLP1 and 2) and which may share common functions (Bush et al., 1994). Apparently the APP, APPLP1 and APPLP2 showed a similar pattern of expression in the hippocampus with both mRNA and protein in granule cells and pyramidal cells of all sectors of the Ammon’s horn (McNamara et al., 1998). Ab is extracellularly deposited in a variety of states and shapes, collectively termed ‘plaques’. When Ab adopts a b-sheet state, it can be stained as an amyloid-like material (‘b-amyloid-plaque’) with classical amyloid stains such as Congo red or thioflavine. In many cases, a core of b-amyloid is surrounded by a surface of dystrophic neurites and swollen glial processes, and forms by this the so-called ‘neuritic plaque’. b-pleated Ab, however, represents only a limited amount of all extracellularly deposited Ab-peptide. This other Ab forms the ‘diffuse’ or ‘pre-amyloid’ plaques and stains either with immunocytochemical tools or a few specific silver stains. The often-used term ‘senile plaque’ refers frequently to an undefined melange of Ab-states and shapes. The distribution pattern of diffuse and amyloid plaques is not congruent. Ab deposits in phases that are characterized by a hierarchical neuroanatomical pattern (Thal et al., 2002). The relationship between plaques and tangles is a matter of debate since long. For many years the so-called amyloid cascade hypothesis has dominated. The hypothesis

726

postulates that Ab-plaques are formed before tangles and may even cause their formation. Only recently, work using the well-defined circuitry of the hippocampal formation provided strong evidence that tangles develop either before or independent from plaques (Scho¨nheit et al., 2004) and follows an anterograde pattern (Thal et al., 2002; Scho¨nheit et al., 2004). Apart from the tau-pathology and the Abpathology, there is an inconspicuous pathology, which, not surprisingly, correlates best with the degree of cognitive impairment. This is the loss of synapses reported by Davies and colleagues (Davies et al., 1987), later confirmed in many subsequent electron-microscopical studies and (indirectly) by semi-quantitative immunolabelling of synaptic marker proteins (Scheff and Price, 2003). Synapse loss takes place in terminal zones of neurons prone to develop tangles. Thus, it is not surprising that number of tangles also correlate strongly with the cognitive decline. Tangle-bearing neurons express baser levels of the message for the synaptic marker protein synaptophysin (Callahan and Coleman, 1995), suggesting that synapse dysfunction and loss is not only the self-evident consequence of neuronal loss but may occur in still living neurons. This is further indicated by a reduction in the synapse to neuron ratio in dentate gyrus (48%) (BertoniFreddari et al., 1996). Interestingly, dendrites which lay within plaques show similar numbers of spines compared to adjacent dendrites or dendritic segments which lay in the inter-plaque region of the neuropil (Einstein et al., 1994). Apart from numeric loss of synapses there are morphological changes, which may differ depending on the nature of the participating fibers and cells. From mossy fiber terminals decreases in number and area of spines and post-synaptic thickenings are reported, whereas nonmossy fiber terminals seem to behave in an opposite way, i.e., displaying more synapses, spines and postsynaptic thickenings (Kiktenko et al., 1997). AD related histopathology in the dentate gyrus Tau-pathology Densely packed tangles with a globose shape (in contrast to the torch-like shaped tangles, e.g., in

pyramidal cells of the Ammon’s horn) built-up by PHFs occur only in late stages of AD, i.e., stage VI of Braaks’ classification (Braak and Braak, 1991). The percentage of tangle-bearing neurons among all neurons of the granule cell layer is 1.7–4.2 in severe AD (Bobinski et al., 1997). Spherical tauaggregates consisting of straight filaments (18–25 nm in diameter), however, are seen in many granule cells when probing severe Alzheimer cases with TG3 (Wakabayashi et al., 1997), a monoclonal antibody binding a phosphate-dependent epitope in PHFs. Interneurons seem to be very resistant, perhaps because of their high content of calcium-binding proteins such as parvalbumin or calretinin (Braak et al., 1991; Nitsch and Ohm, 1995; Freund and Buzsaki, 1996). However, an obligatory absence cannot be presumed because in frontal cortex 4–6% parvalbumin-containing neurons also reacted with an antibody raised against tau from tangles (Iwamoto and Emson, 1991). Others have reported only a 0–0.7% subpopulation of parvalbumin-containing neurons of the superior frontal gyrus which may have tangles (Sampson et al., 1997). In the hippocampus, however, only two cells out of 1950 examined parvalbumin-containing interneurons showed — an even questionable — tangle formation (Ohm et al., 2002). Also for NOS-containing neurons, it is shown that they resist tangle formation. The relative scarceness of NOS neurons in the dentate gyrus does not allow forming a definite opinion about this region. However, double-labelling of other hippocampal neurons has revealed that more than 3300 tangle-bearing neurons are negative for NOS. Starting with Braak-stage IV, some large multipolar neurons of the plexiform layer and adjacent rim of the sector CA4 develop tangles extending into proximal dendrites. In stage V, a small number of closely arranged tangles are seen in modified pyramidal cells, thereby easily distinguishable from the ones in the multipolar neurons described above. In stage VI, no new type accrues but the number of afflicted neurons increases considerably. Neuropil threads are mainly located in dendrites of neurons that undergo hyperphosphorylation and, subsequently, formation of argyrophilic aggregation of tau (Braak and Braak, 1991). This may indicate that a neuron’s main

727

receptive structure, the dendrite, has lost normal structure and function. Beginning with stage III of Braaks’ classification, hyperphosphorylated tau diffusely stains the outer molecular layer of the dentate gyrus. At stages V and particularly at stage VI, argyrophilic neuropil threads are seen (Thal et al., 2000). This corresponds to the beginning of tangle formation (coarse granular material) in the granule cells at stage V, and marked numbers of tangles exhibiting a globose shape in stage VI (Braak and Braak, 1991). Plaques Ab-deposits in CA1 occur in phase 2 of Ab deposition, the dentate gyrus is involved later, namely in phase 3 (Thal et al., 2002). A correlative analysis suggests that phases 1 and 2 are associated with CDR 0, i.e., no signs of cognitive deterioration (Thal et al., 2002). Neuritic plaques are seen in a zone between the outer two and the inner third of the molecular layer, i.e., within the terminal zone of the perforant path (Crain and Burger, 1988). They form later than the tangles in the layer pre-a of the entorhinal cortex, namely not before Braak-stage IV (Braak and Braak, 1991). The caveat that spot check-like examinations might overlook changes is ruled out in studies using close-meshed serial sections throughout the whole hippocampal formation (Scho¨nheit et al., 2004). The limbic stage III with hyperphosphorylated tau highlighting the outer molecular layer and neuritic plaques formed at stage IV probably represents neuropathological signs of insipient AD (Bancher et al., 1993; Braak et al., 1993). At stage V, the numbers increase and a row-like appearance becomes clear. A dense ribbon of neuritic plaques (Braak and Braak, 1991) characterizes the stage VI. Interestingly, the terminal zone of the mossy fibers is almost devoid of neuritic plaques even in late and severe stages of AD. Synaptic changes In 1988, the first report on AD-related synaptic loss quantitatively assessed by electron microscopy was published (Bertoni-Freddari et al., 1988).

Bertoni-Freddari and colleagues reported highly significant reductions in synaptic length, surface density and numeric density determined in the supragranular band (innermost part of the inner molecular layer). Later ultrastructural examinations have confirmed this (Scheff and Price, 1998), and also assigned it to the outer molecular layer (Dekosky et al., 1996; Scheff et al., 1996). The reduction has been calculated to a 21% decrease in numeric synaptic densities and a 26% decrease in the thickness of the outer molecular layer, and a 27% decrease in numeric synaptic densities and 26% reduction in thickness of the inner molecular layer vs. control, respectively. None of the studies, however, has analysed Braak-staged material, which would have allowed making estimates about the dynamic of the process. Given that synaptophysin represents synapses, a significant reduction was visualized correlating to the cognitive decline. Early, mild and severe Alzheimer cases are accompanied by a 23, 49 and 67% reduction in the outer third of the molecular layer, and a 30, 39 and 62% loss in the middle third, respectively. No significant changes are seen in the inner third of the molecular layer (Masliah et al., 1994). At the time being, we have no information about the relative changes in synapse subtypes, e.g., excitatory vs. inhibitory forms. Many transmitters coexist within the same terminal (Torrealba and Carrasco, 2004). It is, therefore, not insentient to consider the possibility of selective defects associated with only one of the transmitters leaving the other transmitter and the ‘synapse’ intact. Because synapses often terminate on dendritic spines, changes in spine number, spine density and spine shape have been determined as well. Spine numbers of the dentate gyrus are reduced in AD cases (Einstein et al., 1994) by 35–44%, depending on the topographical relationship between plaques and dendrites. Dendrites crossing plaques when drawing through the molecular layer show 9% fewer spines than those traversing in plaque-free regions. This difference, however, is statistically not significant. As there are no data concerning the numbers of axons within plaque-free and plaque regions, the putative role of plaques (be it Abpeptide or Ab-amyloid) on spine maintenance or plasticity in AD remains penumbral. Others have

728

found a reduction in granule cell spine density by 60% (Williams and Matthysse, 1986), or 40% (De Ruiter and Uylings, 1987) in Golgi-stained material and of 23% in DiO-labelled neurons (Ji et al., 2003). One should, in this context, keep in mind that determinations of spine densities not only depend on post- and peri-mortem factors more than other morphometrical measures (Uylings et al., 1986), but also on shrinkage of the dendrite or the selection of an appropriate or representative dendritic intercept. In addition, Golgi stains do show only a small proportion of neurons and they may not be representative. Determinations of the post-synaptic density protein SAP97 has unveiled no significant decrease in AD dentate gyrus (Wakabayashi et al., 1999). Detailed data on intrinsic connectivity in AD is apparently lacking. It is, however, conceivable, that certain subtypes of dentate interneurons react with adaptive changes analogous to those seen under other pathological circumstances [e.g., epilepsy (Magloczky et al., 2000)]. It is to note in this context, however, that epilepsy may be a symptom of advanced AD (Lozsadi and Larner, 2006). With respect to output synapses, mossy fibre terminals are the sole ones. They are marked by zinc and dynorphin. Densitometrical analysis of Timm-stained zinc has found an increase in AD (Goldsmith and Joyce, 1995). This may reflect either an increase in zinc levels per terminal or an increase in terminal with unchanged levels of zinc, or higher shrinkage in Alzheimer tissue with more zinc-rich terminal per tissue volume. The copper– zinc superoxide dismutase, expressed in granule cells, might be unaltered on protein and message level (Ceballos et al., 1991), if the low numbers of investigated cases were representative. However, a 35% reduction is found in the stratum lucidum of CA3 for hMTH1, another neuronal protein suggested to protect from zinc-mediated oxidative damage (Furuta et al., 2001). Decreases in number and area of spines and post-synaptic thickenings are reported with respect to mossy fiber terminals on CA3 dendrites (Kiktenko et al., 1997). Chromogranin B, a large dense core vesicle protein and within the hippocampus particularly rich in mossy fibres, is found reduced (46%) in immunocytostaining intensity (Marksteiner et al.,

2000). 3[H] kainate, which labels binding sites on mossy fibre terminals, is found reduced in autoradiography of the stratum lucidum (34%) (Represa et al., 1988). The dendritic tree (basal as well as apical) of the CA3 pyramids, however, is unchanged (Flood et al., 1987a, b) suggesting no overall loss of synaptic contacts. In line herewith, non-mossy fibre terminals (of unknown origin) form more synapses and are associated with a larger area of spines and form post-synaptic thickenings (Kiktenko et al., 1997).

Neuronal loss In cognitively normal individuals,
9–22 million granule cells are counted with stereological techniques (Seress, 1988; West and Gundersen, 1990; West et al., 1994; Bobinski et al., 1996a, 1997; Simic et al., 1997; Harding et al., 1998; Korbo et al., 2004). The broad range (>factor 2) is probably not only due to a large inter-individual variation but also to differences in the sampling scheme or anatomical delineation in regions where the dentate gyrus does not display a C-shaped band. Thus, relative changes seem to be more informative. Unfortunately, the picture remains somewhat blurred. The reported neuronal cell loss in the granule cell layer ranges from 0% (Korbo et al., 2004), 8% (n.s.) (Bobinski et al., 1997), around 11% (n.s.) (West et al., 1994, 2004), 25% (n.s) (Simic et al., 1997) to 43% (po0.05) (Bobinski et al., 1996b). Curiously, researchers who found also the 43% loss described the 8% loss. However, the fact that five out of six studies have failed to show a statistically significant change suggests that putative dysfunction of the dentate gyrus is not caused by simple loss of granule cells. Interestingly to note is a transient tendency for more granule cells in pre-clinical Alzheimer (+13%) of Braak stages II–IV (West et al., 2004). This may be the result of an increase in neurogenesis, as it appears to occur in some Alzheimer brains (Jin et al., 2004). However, granule cell dispersion, occurring in epilepsy, and suggestive for substantial reactive neurogenesis, does apparently not take place in AD. Because most granule cells normally express calbindin, calbindin expression and protein levels

729

are of interest. A
25% reduction (not statistically significant) is found in a radioactive in situ hybridisation analysis (Maguire-Zeiss et al., 1995) and a 85% reduction mRNA level is seen in hippocampal homogenate determinations (Iacopino and Christakos, 1990). The calbindin protein level in whole hippocampal homogenates is reported to be 67% (Iacopino and Christakos, 1990). The reductions may reflect a loss of calbindin-mRNA and subsequently also calbindin occurring before granule cells fade away. Others, pondered that loss of calbindin in many of the granule cells might relate to the cognitive decline in AD despite of the fact that the total number of granule cells might be unaltered (Greene et al., 2001; Palop et al., 2003). Calbindin loss, however, is a well-known feature of dentate granule cells in epilepsy (Magloczky et al., 1997) where it not necessarily associates with cognitive decline or even dementia. In epilepsy, a group with severe hippocampal cell loss (mainly in CA1), displaying the so-called Ammon’s horn sclerosis, shows a marked reduction in granule cell calbindin levels. A second group with medial temporal lobe epilepsy (but not with Ammon’s horn sclerosis) displays normal calbindin levels in their granule cells. Thus, alternatively to a putative causal role in cognitive decline, depletion of calbindin may increase the resistance of granule cells (Nagerl et al., 2000). Interestingly, immunocytochemical demonstration of calmyrin, another granule cell-associated calcium-binding protein, and which is lacking in interneurons, showed no differences between controls and Alzheimer cases in stages IV–V (Bernstein et al., 2005). Also neuron-specific enolase is found unaltered (Wakabayashi et al., 1999). Together this further supports the notion that granule cells withstand neuronal death in AD. Apart from loss of granule cells, local interneurons have to be considered, too. Many of these are characterized by the presence of certain calciumbinding protein: parvalbumin, calretinin and calbindin. It seems that at least some of the interneuron types (Freund and Buzsaki, 1996) have lost the parvalbumin-phenotype or are even numerically reduced. Data reported, however, are not unanimous. A 60% reduction in neuronal density of parvalbumin-containing neurons is found for

the dentate gyrus and sector CA4 together (Brady and Mufson, 1997), whereas others have found a slight increase in the dentate gyrus (Satoh et al., 1991). The latter study is flawed by the fact that the total number of cases and immunoreactive cells is low. In other brain areas, a reduction is found, ranging between 20 and 72%, depending on brain region and layer (Arai et al., 1987; Solodkin et al., 1996; Mikkonen et al., 1999). Again, others differ when reporting no loss (Hof et al., 1991; Fonseca et al., 1993; Sampson et al., 1997; Leuba et al., 1998). In general, however, the loss of parvalbumin-containing neurons is seen only in late Alzheimer and especially in those regions, which undergo massive differentiation or loss of principal neurons. Thus, it is likely that at least those GABAergic interneurons, which contain parvalbumin, are primarily resistant against the tau pathology and untimely cell death. Another interneuron-associated calcium-binding protein is calretinin, which is found in differentially shaped neurons of the human dentate gyrus (Nitsch and Ohm, 1995). With respect to the AD hippocampus no indication for a loss in the numbers of immunostained neurons is reported (Brion and Resibois, 1994). In line herewith, also other brain areas display an unchanged overall neuronal density of calretinin-positive neurons (Hof et al., 1993; Fonseca and Soriano, 1995; Leuba et al., 1998; Mikkonen et al., 1999). The observed layer-specific increase of small neurons that is concomitantly found with a decrease in large calretinin-containing neurons (Sampson et al., 1997; Leuba et al., 1998; Mikkonen et al., 1999) supports the notion that more subtle plastic–adaptive changes occur. In this context it is also to note that the intensity of the calretinin immunoreaction with a staining gradient (apical dendrites stronger, basal weaker) is inverted in the entorhinal cortex of AD patients (Mikkonen et al., 1999). This suggests a plastic and adaptive process. Coincided with calretinin, the innermost part of the molecular layer is normally strongly immunolabelled by secretoneurin, a member of the chromogranin family. In AD, this band is significantly reduced (52%), whereas the intensity of the calretinin immunostaining is only slightly diminished (16%) (Kaufmann et al., 1998). The rest of the inner molecular layer is

730

unchanged, but a 40% decrease in secretoneurin is found in the outer molecular layer. The third calcium binding protein of various interneurons is calbindin, which, however, is also typical for granule cells (Sloviter et al., 1991; Seress et al., 1993). This implies that determination of protein or message levels represent the contribution of both neuronal classes. The immunocytochemical analysis of the dentate gyrus reveals only few calbindin-containing granule cells and interneurons in Alzheimer brains (Iritani et al., 2001). No changes in the number and distribution of GABAergic neurons are seen in the hippocampus between controls and AD (Yew et al., 1999). Some interneurons of the dentate gyrus, mostly located just underneath the granule cell layer, contain neuropeptide Y (NPY). These peptidergic large polymorphic neurons are reduced in numbers, to only a mild degree in the rostral and middle part of the hippocampus, but evident caudally. Although there are neurons with intense immunostaining and a preserved dendritic tree, most others, however, show a loss or regression of their dendritic tree (Chan-Palay et al., 1986). A further interneuron subgroup is identified by its content of somatostatin. These neurons are evenly spread throughout the polymorphic layer, lacking in the granule cell layer, and are only occasionally found in the molecular layer. In AD, only vestiges remain with greatly diminished dendritic processes (Chan-Palay, 1987). This is in line with biochemical findings, which have determined a 70% loss of somatostatin in hippocampal tissue (Davies and Terry, 1981). Also in other cortex areas, somatostatin-containing interneurons show a decreased numeric density (>70%) (Kumar, 2005) and/or cytoskeletal changes reminiscent of tau-pathology (van de Nes et al., 2002). Together this suggests that this population is more vulnerable than other interneurons. Multipolar CCK-8containing neurons are occasionally found in the molecular layer and numerous polymorphic ones in the anterior part of the dentate gyrus, mainly in the subgranular zone of the polymorphic layer (Lotstra and Vanderhaeghen, 1987). In contrast to animals, CCK-8-positive fibers seem to emanate almost exclusively from intrahippocampal neurons, because fimbria and alveus were found

devoid of CCK-8-positive profiles and only rarely they were seen in the angular bundle. The only exception seems to be a projection via the perforant path probably originating from pre-a neurons. Changes in CCK are not reported for AD hippocampus. Some larger enkephalin-containing neurons are found in the molecular layer of the dentate gyrus (Kulmala, 1985). No changes are seen in AD (Kulmala, 1985) when examining the dentate gyrus or the whole hippocampus (Yew et al., 1999). However, a Leu- and Met-encephalinreceptor binding loss is found in hippocampal homogenates (Rinne et al., 1993). Others have found a statistically significant 48% loss of m- and a 36% loss of k-opiod-binding sites whereas no changes have been seen for d-binding sites in quantitative autoradiography (Mathieu-Kia et al., 2001). Prominent terminal-like Substance P staining is seen in the dentate gyrus and in CA1–4 fields, and multipolar immunolabelled neurons are located in the polymorphic layer of the dentate gyrus as well as in CA1–4 (Kowall et al., 1993). These neurons are NADPH-negative, in contrast to most of the Substance P-containing neurons of the isocortex. Substance P-receptor immunocytochemistry shows a light marking of CA2 pyramidal cells and occasional labelling of basket cells in CA1–4. In AD, staining intensity is reduced in the dentate gyrus but sparing the hilar neurons (Kowall et al., 1993). Nitric oxide is generated in neurons with NADPH-diaphorase activity (NOS-neurons). The distribution and morphological types does not differ significantly between controls and Alzheimer’ disease patients in all sectors of the Ammon’s horn and subiculum (Hyman et al., 1992). Also the total number of NOS neurons is unaltered. However, whereas in CA4 and CA3 a considerable decline (though not statistically significant) in the numbers is found, the subiculum shows a light increase. Only few NOS-neurons are occasionally found in the dentate gyrus, i.e., in 2 of 20 examined hippocampi. Double labelling has revealed that more than 3300 tangle-bearing neurons are negative for NOS. The planimetrically determined size of the neurons does not show changes (exception CA1) but regressive changes are found in the form of foreshortened dendrites and distorted and beaded axons. NADPH-diaphorase activity is

731

histochemically visualized and displays a striking loss in the terminal zone of the perforant path in AD (Rebeck et al., 1993). In controls, this zone stands out clear due to the high level of reaction product and gives the molecular layer a sublaminated aspect. As with NOS antibody staining, only few neurons are stained with NADPHdiaphorase activity in the granule cell layer. CA4 appeared to be unaltered.

The excitatory–inhibitory neurochemistry Because the dentate gyrus receives a glutamatergic excitatory input from the layer pre-a and transmits processed information to CA3, the glutamatergic system has attracted researchers and numerous neurochemical studies have been performed. Early studies using microdissected specimens of the termination zone of the perforant path, and subsequent determinations of the glutamate content therein, have demonstrated an 83% reduction in AD (Hyman et al., 1987). More recent studies have determined possible changes at a higher definition in terms of receptor subtypes and (sub)regional distribution. In contrast to significant decreases in other parts of the hippocampus (or the hippocampus as a whole), the dentate gyrus shows no significant changes for NMDA-R1, 2A and 2B message or protein (despite of trends toward decreased levels) as determined by Western blotting of micropunches (Wakabayashi et al., 1999; MishizenEberz et al., 2004), radioactive in situ hybridisation (Ulas and Cotman, 1997) and immunocytochemistry (Aronica et al., 1998). An immunocytochemical study with Braak-staged tissue has found no changes in early stages but an increase in staining intensity at stages IV and later in all CA fields concomittant with a decrease in staining of the outer molecular layer (Ikonomovic et al., 1999). NMDA-receptor message and protein is located on neurons, i.e., granule cells and pyramidal cells, showing expression differences between the hippocampal subfields. NMDA-R1, -2A and -2B message is highest in granule cells, followed by CA2 and 3, protein highest in CA1 for NMDA-R1 and -2A and granule cells for 2B (Mishizen-Eberz et al., 2004). Likewise AMPA receptors have been examined and

no changes are seen in the dentate gyrus in quantitative autoradiography (Dewar et al., 1991) or a reduction in the outer molecular layer is found, together with an increase in the polymorphic layer (Geddes et al., 1992). However, Kd and Bmax may have changes, making an interpretation difficult. Again, more recent studies give a more defined picture. Glu-R1 is found unaltered in an in situ hybridisation study (Pellegrini-Giampietro et al., 1994) and in immunocytochemical investigations (Aronica et al., 1998). Others, however, find a subtly modified picture. Glu-R1 immunoreactive neurons are apparently unaltered in number (Ikonomovic et al., 1995; Aronica et al., 1998; Wakabayashi et al., 1999) but show an increase in staining intensity in the molecular layer (Hyman et al., 1994; Ikonomovic et al., 1995; Wakabayashi et al., 1999) especially in the inner (Hyman et al., 1994) and supragranular part (Ikonomovic et al., 1995), and in the plexiform layer (Ikonomovic et al., 1995). This increase is on one hand attributed to an increase in the number of immunoreactive fibers but on the other hand due to an increase of immunoreactivity. The latter may also explain why the proportion of Alzheimer cases displaying a signal (specific immunoreactivity over background) is about twice as high as of controls (Hyman et al., 1994). With respect to CA3 (and CA4), i.e., the terminal zone of the granule cells, an increase is found (Aronica et al., 1998), which seems to follow the progression of the Braak-stages (though not statistically significant) (Carter et al., 2004). Curiously, Western blotting shows a significant 70% reduction in the dentate gyrus of AD patients, which is accompanied by virtually identical levels of neuron-specific enolase (Wakabayashi et al., 1999). The latter suggests that no major granule cell loss has taken place and is in line with most of the cell counting studies. The apparent contradiction between the Western blot and immunocytochemical data obtained from identical cases (Wakabayashi et al., 1999) may be due to unmasking of epitopes or increasing of affinity, which is not observed under denaturation conditions in Western blotting. Glu-R2 and Glu-R2/3 immunoreactivity is found to be unaltered (Hyman et al., 1994) or increased in the (inner) molecular layer and plexiform layer of the dentate gyrus (Ikonomovic et al., 1995; Aronica

732

et al., 1998). The number of Glu-R2 immunoreactive neurons in the dentate gyrus is reduced (Ikonomovic et al., 1995). Microdissected punches of hippocampi, grouped according to the Braakclassification into stages 0–II, stages III–IV and stages V–VI, and subjected to Western blotting, has contained comparable amounts of immunoreactive protein (Carter et al., 2004). Interestingly, however, double-labelling of Glu-R2 and tangles in the entorhinal cortex and subiculum/CA1 by the MC1 antibody has resulted in a complementary pattern: number of Glu-R2 immunoreactive cells diminished when the number of MC1 positive (i.e., tangle-bearing cells) increased (Ikonomovic et al., 1997). Others, however, report similar immunoreactivity in both tangle-bearing and tangle-free neurons for any of the studied Glu-R subtypes (Hyman et al., 1994). Glu-R4, though relatively sparse in the hippocampus, is found unaltered in terms of staining pattern and intensity. In sharp contrast to staining with Glu-R1 and Glu-R2 antibodies, the immunoreaction is not localised to dendrites or axons but highlighted neuronal somata including CA3 pyramidal cells and granule cells (Hyman et al., 1994). Quantitative 3[H] kainate autoradiography reveals a statistically significant 54% reduction in the supragranular layer (Represa et al., 1988). Kainate receptors can be visualised with antibodies raised against Glu-R5/6/7. Immunoreactive neurons are granule cells, and mossy cells as well as pyramidal cells of the CA sectors. The somatodendritic compartment is labelled. Other neurons marked are non-pyramidal cells of the stratum oriens and some glial cells in the alveus and fimbria. No significant changes are seen in Alzheimer brains (Aronica et al., 1998). With respect to the inhibitory GABAergic system autoradiographic studies show a similar Kd for 3 [H] GABA hippocampal binding sites between controls and Alzheimer cases (Chu et al., 1987). GABAA receptors are found unaltered, whereas GABAB binding sites are significantly reduced in the molecular layer of the dentate gyrus, the stratum pyramidale and lacunosum-moleculare of the sector CA1 (Chu et al., 1987). A similar picture is seen with 3[H] flunitrazepame except for the reduction in the molecular layer of the gyrus dentatus (Jansen et al., 1990; Penney et al., 1990), which is

found unaltered or even displaying a slight increase. Micropunches of hippocampi subjected to Western blotting of b1, b2, a1 and a5 subunits of GABAA receptors show no differences between controls and Alzheimer individuals. A significant decrease is seen only for a5 in CA3 (20%) (Rissman et al., 2003), whereas a5-ligand binding show only in CA1 a significant 27% decrease (Howell et al., 2000). Immunocytochemical analyses give similar results for GABAA-R-b2/3 (Mizukami et al., 1997) and GABAA-R-a1 (Singer et al., 2006) when comparing cases with Braak-stages I–VI. In severe stages, the GABAA-R-a1 is found decreased locally on soma and dendritic surface of mossy cells. Labelling of interneurons with GABAA-R-a1 or GABAA-R-b2/3 remains unaltered though they display shrunken processes. GABAB-R1, however, show a transient but significant increase in stages II and IV in sectors CA4 and CA3, whereas a 64% loss of immunoreactive neurons of sector CA1 takes place between stages I–II and II–IV which further proceeds to 70% in stages V–VI (Iwakiri et al., 2005). The finding that between stages I–II and III—IV, only 12% of Nissl-stained neurons are lost (and 39% in stages V–VI) suggests that receptor loss precedes neuronal death. In situ hybridisation for GABAA-R-b2 in cases with Braak stages I–VI shows — in line with the protein findings — no changes. In contrast, GABAA-R-b3 message is reduced in all hippocampal sectors but not in the CA4 (Mizukami et al., 1998). In a parenthetical note GABAA-R-a1 message localisation and signal intensity is reported to be apparently unaltered in Alzheimer hippocampus (Pellegrini-Giampietro et al., 1994).

Disconnected dentate gyrus From the above it becomes conceivable that the classical histopathological signs of AD, which correlate strongly with the degree of dementia (tangles), are late events. This also holds true for neuronal cell loss. In contrast, clinical signs such as dementia may develop much earlier. In terms of the Braak classification tangles and neuropil threads evolve in stage V or later, whereas loss of cognitive and memory function as assessed by the mini

733

mental state examination (MMSE) test may occur already in stage III (Bancher et al., 1993; Braak et al., 1993). The dentate gyrus is considered to play a crucial role in associative memory, especially with respect to events (‘what happens’) (Morris, 2006). This raises the possibility that changes which lead to interruption or perturbation of the granule cells’ input isolates the dentate gyrus and causes impairment of learning and memory (Hyman et al., 1984). To discharge one single granule cell,
400 axons of the perforant path have to convey the excitatory input from the pre-a neurons (McNaughton et al., 1991). This suggests that loss of entorhinal pre-a neurons may be sufficient to cause learning and memory defects. Support for this concept comes from several lines of evidence. Correlative clinico-neuropathological studies are, unfortunately, not unambiguous. In general, the numbers of tangles in entorhinal neurons correlate with both Braak staging and deterioration of cognition (Garcia-Sierra et al., 2000, 2001). An early study in 1996 indicated that a 57% loss of layer pre-a neurons (
650,000 in CDR 0 control cases vs. 285,000 in CDR 0.5 individuals) is associated with a Clinical Dementia Rating (CDR) score of 0.5, which represents beginning dementia (Gomez Isla et al., 1996). At that time, the total entorhinal neuron loss is only
30%. This is confirmed by a
35 and 50% pre-a layer neuron loss in CDR 0.5 cases (Kordower et al., 2001; Price et al., 2001), respectively, but not in a newer study (only 1–2% loss) (Hof et al., 2003). Although there is no direct translating between CDR and Braak scores, it is suggested that a CDR 0.5 may represent Braakstage III and a MMSE range of 26–29 (Hof et al., 2003). Severe AD (CDR 3–5, MMSE 0 or Braak stage VI) is found associated with a
90% loss of pre-a neurons (Gomez Isla et al., 1996) and a laminar-specific spongiosis in the perforant path termination zone (Duyckaerts et al., 1998) as well as ALZ50 immunoreactivity highlights the perforant path terminal zone (Hyman et al., 1988). The immunocytochemical detection of the intrinsic presynaptic vesicle protein synaptophysin, suggests a functional or terminal loss in the termination zone of the perforant path (Cabalka et al., 1992; Masliah et al., 1994). Demonstration of the myelinated fibres by the Weil-stain has indicated a

marked rarefaction of white matter along the course of the perforant tract together with a loss of oligodendrocytes thereabout (Morys et al., 1994). In vivo imaging by MRI scans in psychometrically assessed individuals supports this. A decrease in white matter volume in the parahippocampal gyrus that includes the perforant path is found already in amnesic mild cognitive impaired individuals (the MMSE scores suggest that this might represent very mild AD) (Petersen et al., 2006; Stoub et al., 2006). Post-synaptically, the loss of perforant path input seems to induce changes in granule cells. Granule cells remodel their dendritic tree in AD as determined in quantitative morphometrical analyses on Golgi stained (Flood et al., 1985, 1987a, b) and with Lucifer yellow intracellularly filled (Einstein et al., 1994) neurons. In the Golgi studies, the granule cells’ apical dendrites of severe Alzheimer cases are compared to controls. The controls consist of three differentially aged groups of nondemented individuals. The middle-aged group (mean of age 52.2972.60 SEM), the old-aged group (comparable to the Alzheimer group) (73.4072.36 years) and the very old-aged group (90.2071.85 years), however, have various degrees of AD-related histopathology. Flood and colleagues have found a significantly reduced total dendritic length (37%) and a significantly reduced average segment length (dendritic length/number of segments; 34%) but no significant changes in total segment numbers or width of the fascia dentata (granule cell layer + molecular layer) when comparing the old-aged group with the Alzheimer group. Interestingly, the Alzheimer group has only a 16% shorter total dendritic length than the middle-aged group, which has the lowest degree of AD-related neuropathology (no changes or only ‘very slight or slight changes’). This difference as well that between the very old group and the Alzheimer group is not statistically significant in post-hoc tests. It is tempting to speculate that the difference in the dendritic tree measures between the middle-aged group (with no or only very few AD-related neuropathological changes) and the old group (with some more changes but no clinical signs of AD) might represent an early stage in the evolution of AD where regenerative plasticity is

734

dominant (Flood et al., 1985). Later, when differentiation proceeds the balance between regenerative (progressive) and degenerative (regressive) plasticity should shift towards the formation of a reduced dendritic arborisation. Similar to the Golgi studies, the analysis of Lucifer yellow-filled apical granule cell dendrites unveils a statistically significant reduction in dendritic length (by 50%) and also a shorter average segment length (19%) (Einstein et al., 1994). Proliferative signs, however, such as increased number of dendritic twigs or unusual excrescences are not observed. Spine numbers are reduced in AD cases by 35–60% (Williams and Matthysse, 1986; De Ruiter and Uylings, 1987; Einstein et al., 1994; Ji et al., 2003). However, it is likely that spine plasticity will also occur. In early stages, first loss of afferents may induce signals, which allow the local sprouting in order to attract ‘new’ synaptic input from neighboured afferents. Given that
400 axons (synapses) may be necessary to discharge a granule cell (McNaughton et al., 1991), this may be an attempt to keep the neuronal net properly working. Additionally, the existing terminals are statistically significantly enlarged in terms of apposition length (+18% in outer molecular layer and +12% in the inner molecular layer) and — as a mild trend — as well as total synaptic contact area [in outer and inner molecular layer +4% (n.s.)] (Scheff et al., 1996; Scheff and Price, 1998) probably to serve the same purpose. Later, when differentiation proceeds, the balance between regenerative (progressive) and degenerative (regressive) plasticity shifts towards the formation of synaptically stripped dendrites (Scheff and Price, 2003). AP180, a protein considered to be crucially involved in neuronal clathrin-mediated synaptic recycling, is expressed in granule cells and stains the molecular layer of cognitive normal humans homogeneously. In AD, a loss of granule cell body staining but an increase in the staining intensity and width of the inner molecular layer is reported (Yao et al., 1999). This suggests that granule cells show plastic adaptation probably responding to challenges on their dendritic input. Apart from disconnection of principal cells, interneurons may also become disconnected from the entorhinal input. This may alter the balance between excitatory and inhibitory effects. A

comparative morphometrical analysis of apical and basal dendrites of parvalbumin-containing GABAergic interneurons located within or directly at the border of the granule cell layer has been made in Braak-staged cases (Ohm et al., 2002). A non-significant and transient increase in dendritic length, branch order and number of segments of the apical dendrites is reported for an early Braakstage (stage II) followed by a statistically significant decline of the respective measures in stages IV and V. Basal dendrites, in contrast, remained stable. Because apical dendrites receive entorhinal input from pre-a neurons whereas basal dendrites do not (Zipp et al., 1989), the dendrite-specific change suggests an input-specific plastic change. A 35–50% pre-a layer neuron loss is reported for mild cognitive decline (Kordower et al., 2001; Price et al., 2001), which may represent Braak-stage III (Hof et al., 2003). It is therefore likely that a substantial loss of input (>50%) is required to induce the regressive changes which may eventually lead to the interneuron’s own death. It may also play an important role how fast this input decreases. Interestingly, the transient increase in dendritic length, segment numbers and branches (Ohm et al., 2002) parallels a transient tendency for more granule cells in pre-clinical Alzheimer (+13%) of Braak stages II–IV (West et al., 2004). The latter may be the result of an increase in neurogenesis, as it appears to occur in some Alzheimer brain (Jin et al., 2004), the former as an attempt to sprout. Closing remark The available data suggest that the dentate gyrus withstands the formation of plaques, tangles and neuronal death until late stages of AD. However, changes related to a disconnecting process, mainly from loss of entorhinal input, and more subtle intrinsic alterations may contribute to disturbances in memory and learning observed already in early stages of AD. References Amaral, D.G., Insausti, R. and Cowan, W.M. (1984) The commissural connections of the monkey hippocampal formation. J. Comp. Neurol., 224: 307–336.

735 Arai, H., Emson, P.C., Mountjoy, C.Q., Carassco, L.H. and Heizmann, C.W. (1987) Loss of parvalbumin-immunoreactive neurones from cortex in Alzheimer-type dementia. Brain Res., 418: 164–169. Aronica, E., Dickson, D.W., Kress, Y., Morrison, J.H. and Zukin, R.S. (1998) Non-plaque dystrophic dendrites in Alzheimer hippocampus: a new pathological structure revealed by glutamate receptor immunocytochemistry. Neuroscience, 82: 979–991. Augustinack, J.C., Schneider, A., Mandelkow, E.M. and Hyman, B.T. (2002) Specific tau phosphorylation sites correlate with severity of neuronal cytopathology in Alzheimer’s disease. Acta Neuropathol. (Berl.), 103: 26–35. Bancher, C., Braak, H., Fischer, P. and Jellinger, K.A. (1993) Neuropathological staging of Alzheimer lesions and intellectual status in Alzheimer’s and Parkinson’s disease patients. Neurosci. Lett., 162: 179–182. Bancher, C., Brunner, C., Lassmann, H., Budka, H., Jellinger, K., Wiche, G., Seitelberger, F., Grundke Iqbal, I., Iqbal, K. and Wisniewski, H.M. (1989) Accumulation of abnormally phosphorylated tau precedes the formation of neurofibrillary tangles in Alzheimer’s disease. Brain Res., 477: 90–99. Bernstein, H.G., Blazejczyk, M., Rudka, T., Gundelfinger, E.D., Dobrowolny, H., Bogerts, B., Kreutz, M.R., Kuznicki, J. and Wojda, U. (2005) The Alzheimer disease-related calcium-binding protein Calmyrin is present in human forebrain with an altered distribution in Alzheimer’s as compared to normal ageing brains. Neuropathol. Appl. Neurobiol., 31: 314–324. Bertoni-Freddari, C., Fattoretti, P., Casoli, T., Caselli, U. and Meier-Ruge, W. (1996) Deterioration threshold of synaptic morphology in aging and senile dementia of Alzheimer’s type. Anal. Quant. Cytol. Histol., 18: 209–213. Bertoni-Freddari, C., Meier-Ruge, W. and Ulrich, J. (1988) Quantitative morphology of synaptic plasticity in the aging brain. Scanning Microsc., 2: 1027–1034. Bobinski, M., Wegiel, J., Tarnawski, M., de Leon, M.J., Reisberg, B., Miller, D.C. and Wisniewski, H.M. (1998) Duration of neurofibrillary changes in the hippocampal pyramidal neurons. Brain Res., 799: 156–158. Bobinski, M., Wegiel, J., Tarnawski, M., Reisberg, B., de Leon, M.J., Miller, D.C. and Wisniewski, H.M. (1997) Relationships between regional neuronal loss and neurofibrillary changes in the hippocampal formation and duration and severity of Alzheimer disease. J. Neuropathol. Exp. Neurol., 56: 414–420. Bobinski, M., Wegiel, J., Wisniewski, H.M., Tarnawski, M., Reisberg, B., de Leon, M.J. and Miller, D.C. (1996a) Neurofibrillary pathology — correlation with hippocampal formation atrophy in Alzheimer disease. Neurobiol. Aging, 17: 909–919. Bobinski, M., Wegiel, J., Wisniewski, H.M., Tarnawski, M., Reisberg, B., de Leon, M.J. and Miller, D.C. (1996b) Neurofibrillary pathology — correlation with hippocampal formation atrophy in Alzheimer disease. Neurobiol. Aging, 17: 909–919.

Braak, E., Braak, H. and Mandelkow, E.M. (1994) A sequence of cytoskeleton changes related to the formation of neurofibrillary tangles and neuropil threads. Acta Neuropathol. Berl., 87: 554–567. Braak, E., Strotkamp, B. and Braak, H. (1991) Parvalbuminimmunoreactive structures in the hippocampus of the human adult. Cell Tissue Res., 264: 33–48. Braak, H. and Braak, E. (1991) Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol. (Berl.), 82: 239–259. Braak, H., Braak, E., Grundke-Iqbal, I. and Iqbal, K. (1986) Occurrence of neuropil threads in the senile human brain and in Alzheimer’s disease: a third location of paired helical filaments outside of neurofibrillary tangles and neuritic plaques. Neurosci. Lett., 65: 351–355. Braak, H. and Del Tredici, K. (2004) Poor and protracted myelination as a contributory factor to neurodegenerative disorders. Neurobiol. Aging, 25: 19–23. Braak, H., Duyckaerts, C., Braak, E. and Piette, F. (1993) Neuropathological staging of Alzheimer-related changes correlates with psychometrically assessed intellectual status. In: Corain B., Iqbal K., Nicolini M., Winblad B., Wisniewski H. and Zatta P. (Eds.), Alzheimer’s Disease: Advances in Clinical and Basic Research. Wiley, Chicester, pp. 131–137. Brady, D.R. and Mufson, E.J. (1997) Parvalbumin-immunoreactive neurons in the hippocampal formation of Alzheimer’s diseased brain. Neuroscience, 80: 1113–1125. Brion, J.P. and Resibois, A. (1994) A subset of calretininpositive neurons are abnormal in Alzheimer’s disease. Acta Neuropathol. (Berl.), 88: 33–43. Bush, A.I., Pettingell, W.-H.J., de Paradis, M., Tanzi, R.E. and Wasco, W. (1994) The amyloid beta-protein precursor and its mammalian homologues. Evidence for a zinc-modulated heparin-binding superfamily. J. Biol. Chem., 269: 26618–26621. Cabalka, L.M., Hyman, B.T., Goodlett, C.R., Ritchie, T.C. and Vanhoesen, G.W. (1992) Alteration in the pattern of nerve terminal protein immunoreactivity in the perforant pathway in Alzheimer’s disease and in rats after entorhinal lesions. Neurobiol. Aging, 13: 283–291. Callahan, L.M. and Coleman, P.D. (1995) Neurons bearing neurofibrillary tangles are responsible for selected synaptic deficits in Alzheimer’s disease. Neurobiol. Aging, 16: 311–314. Carter, T.L., Rissman, R.A., Mishizen-Eberz, A.J., Wolfe, B.B., Hamilton, R.L., Gandy, S. and Armstrong, D.M. (2004) Differential preservation of AMPA receptor subunits in the hippocampi of Alzheimer’s disease patients according to Braak stage. Exp. Neurol., 187: 299–309. Ceballos, I., Javoy-Agid, F., Delacourte, A., Defossez, A., Lafon, M., Hirsch, E., Nicole, A., Sinet, P.M. and Agid, Y. (1991) Neuronal localization of copper-zinc superoxide dismutase protein and mRNA within the human hippocampus from control and Alzheimer’s disease brains. Free Radic. Res. Commun., 12–13(Pt 2): 571–580. Chan-Palay, V. (1987) Somatostatin immunoreactive neurons in the human hippocampus and cortex shown by immunogold/silver intensification on vibratome sections: coexistence

736 with neuropeptide Y neurons, and effects in Alzheimer-type dementia. J. Comp. Neurol., 260: 201–223. Chan-Palay, V., Lang, W., Haesler, U., Ko¨hler, C. and Yasargil, G. (1986) Distribution of altered hippocampal neurons and axons immunoreactive with antisera against neuropeptide Y in Alzheimer’s-type dementia. J. Comp. Neurol., 248: 376–394. Chu, D.C.M., Penney Jr., J.B. and Young, A.B. (1987) Quantitative autoradiography of hippocampal GABAB and GABAA receptor changes in Alzheimer’s disease. Neurosci. Lett., 82: 246–252. Crain, B.J. and Burger, P.C. (1988) The laminar distribution of neuritic plaques in the fascia dentata of patients with Alzheimer’s disease. Acta Neuropathol. (Berl.), 76: 87–93. Davies, C.A., Mann, D.M.A., Sumpter, P.Q. and Yates, P.O. (1987) A quantitative morphometric analysis of the neuronal and synaptic content of the frontal and temporal cortex in patients with Alzheimer’s disease. J. Neurol. Sci., 78: 151–164. Davies, P. and Terry, R.D. (1981) Cortical somatostatin-like immunoreactivity in cases of Alzheimer’s disease and senile dementia of the Alzheimer type. Neurobiol. Aging, 2: 9–14. De Ruiter, J.P. and Uylings, H.B.M. (1987) Morphometric and dendritic analysis of fascia dentata granule cells in human aging and senile dementia. Brain Res., 402: 217–229. Dekosky, S.T., Scheff, S.W. and Styren, S.D. (1996) Structural correlates of cognition in dementia: quantification and assessment of synapse change. Neurodegeneration, 5: 417–421. Demeter, S., Rosene, D.L. and van Hoesen, G.W. (1985) Interhemispheric pathways of the hippocampal formation, presubiculum, and entorhinal and posterior parahippocampal cortex in the rhesus monkey: the structure and organization of the hippocampal commissures. J. Comp. Neurol., 233: 30–47. Dewar, D., Chalmers, D.T., Graham, D.I. and McCulloch, J. (1991) Glutamate metabotropic and AMPA binding sites are reduced in Alzheimer’s disease — an autoradiographic study of the hippocampus. Brain Res., 553: 58–64. Duyckaerts, C., Colle, M.A., Seilhean, D. and Hauw, J.J. (1998) Laminar spongiosis of the dentate gyrus: a sign of disconnection, present in cases of severe Alzheimer’s disease. Acta Neuropathol. (Berl.), 95: 413–420. Dyrks, T., Weidemann, A., Multhaup, G., Salbaum, J.M., Lemaire, H.G., Kang, J., Mu¨ller-Hill, B., Masters, C.L. and Beyreuther, K. (1988) Identification, transmembrane orientation and biogenesis of the amyloid A4 precursor of Alzheimer’s disease. EMBO J., 7: 949–957. Einstein, G., Buranosky, R. and Crain, B.J. (1994) Dendritic pathology of granule cells in Alzheimer’s disease is unrelated to neuritic plaques. J. Neurosci., 14: 5077–5088. Flood, D.G., Buell, S.J., Defiore, C.H., Horwitz, G.J. and Coleman, P.D. (1985) Age-related dendritic growth in dentate gyrus of human brain is followed by regression in the ‘oldest old’. Brain Res., 345: 366–368. Flood, D.G., Buell, S.J., Horwitz, G.J. and Coleman, P.D. (1987a) Dendritic extent in human dentate gyrus granule cells in normal aging and senile dementia. Brain Res., 402: 205–216.

Flood, D.G., Guarnaccia, M. and Coleman, P.D. (1987b) Dendritic extent in human CA2-3 hippocampal pyramidal neurons in normal aging and senile dementia. Brain Res., 409: 88–96. Fonseca, M. and Soriano, E. (1995) Calretinin-immunoreactive neurons in the normal human temporal cortex and in Alzheimer’s disease. Brain Res., 691: 83–91. Fonseca, M., Soriano, E., Ferrer, I., Martinez, A. and Tunon, T. (1993) Chandelier cell axons identified by parvalbuminimmunoreactivity in the normal human temporal cortex and in Alzheimer’s disease. Neuroscience, 55: 1107–1116. Freund, T.F. and Buzsaki, G. (1996) Interneurons of the hippocampus. Hippocampus, 6: 347–470. Furuta, A., Iida, T., Nakabeppu, Y. and Iwaki, T. (2001) Expression of hMTH1 in the hippocampi of control and Alzheimer’s disease. Neuroreport, 12: 2895–2899. Garcia-Sierra, F., Hauw, J.J., Duyckaerts, C., Wischik, C.M., Munoz, J. and Mena, R. (2000) The extent of neurofibrillary pathology in perforant pathway neurons is the key determinant of dementia in the very old. Acta Neuropathol. (Berl.), 100: 29–35. Garcia-Sierra, F., Wischik, C.M., Harrington, C.R., Luna-Munoz, J. and Mena, R. (2001) Accumulation of C-terminally truncated tau protein associated with vulnerability of the perforant pathway in early stages of neurofibrillary pathology in Alzheimer’s disease. J. Chem. Neuroanat., 22: 65–77. Geddes, J.W., Ulas, J., Brunner, L.C., Choe, W. and Cotman, C.W. (1992) Hippocampal excitatory amino acid receptors in elderly, normal individuals and those with Alzheimer’s disease: non-N-methyl-D-aspartate receptors. Neuroscience, 50: 23–34. Goldsmith, S.K. and Joyce, J.N. (1995) Alterations in hippocampal mossy fiber pathway in schizophrenia and Alzheimer’s disease. Biol. Psychiatry, 37: 122–126. Gomez Isla, T., Price, J.L., McKeel Jr., D.W., Morris, J.C., Growdon, J.H. and Hyman, B.T. (1996) Profound loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer’s disease. J. Neurosci., 16: 4491–4500. Greene, J.R., Radenahmad, N., Wilcock, G.K., Neal, J.W. and Pearson, R.C. (2001) Accumulation of calbindin in cortical pyramidal cells with ageing; a putative protective mechanism which fails in Alzheimer’s disease. Neuropathol. Appl. Neurobiol., 27: 339–342. von Gunten, A., Kovari, E., Bussiere, T., Rivara, C.B., Gold, G., Bouras, C., Hof, P.R. and Giannakopoulos, P. (2006) Cognitive impact of neuronal pathology in the entorhinal cortex and CA1 field in Alzheimer’s disease. Neurobiol. Aging, 27: 270–277. Haass, C. (2004) Take five — BACE and the gamma-secretase quartet conduct Alzheimer’s amyloid beta-peptide generation. EMBO J., 23: 483–488. Harding, A.J., Halliday, G.M. and Kril, J.J. (1998) Variation in hippocampal neuron number with age and brain volume. Cereb. Cortex, 8: 710–718. Hof, P.R., Bussiere, T., Gold, G., Kovari, E., Giannakopoulos, P., Bouras, C., Perl, D.P. and Morrison, J.H. (2003) Stereologic evidence for persistence of viable neurons in layer II of

737 the entorhinal cortex and the CA1 field in Alzheimer disease. J. Neuropathol. Exp. Neurol., 62: 55–67. Hof, P.R., Cox, K., Young, W.G., Celio, M.R., Rogers, J. and Morrison, J.H. (1991) Parvalbumin-immunoreactive neurons in the neocortex are resistant to degeneration in Alzheimer’s disease. J. Neuropathol. Exp. Neurol., 50: 451–462. Hof, P.R., Nimchinsky, E.A., Celio, M.R., Bouras, C. and Morrison, J.H. (1993) Calretinin-immunoreactive neocortical interneurons are unaffected in Alzheimer’s disease. Neurosci. Lett., 152: 145–148. Howell, O., Atack, J.R., Dewar, D., Mckernan, R.M. and Sur, C. (2000) Density and pharmacology of alpha5 subunit-containing GABA(A) receptors are preserved in hippocampus of Alzheimer’s disease patients. Neuroscience, 98: 669–675. Hyman, B.T., van Hoesen, G.W. and Damasio, A.R. (1987) Alzheimer’s disease: glutamate depletion in the hippocampal perforanth pathway zone. Ann. Neurol., 22: 37–40. Hyman, B.T., Van Horsen, G.W., Damasio, A.R. and Barnes, C.L. (1984) Alzheimer’s disease: cell-specific pathology isolates the hippocampal formation. Science, 225: 1168–1170. Hyman, B.T., Kromer, L.J. and van Hoesen, G.W. (1988) A direct demonstration of the perforath pathway terminal zone in Alzheimer’s disease using the monoclonal antibody Alz-50. Brain Res., 450: 392–397. Hyman, B.T., Marzloff, K., Wenniger, J.J., Dawson, T.M., Bredt, D.S. and Snyder, S. (1992) Relative sparing of nitric oxide synthase-containing neurons in the hippocampal formation in Alzheimer’s disease. Ann. Neurol., 32: 818–820. Hyman, B.T., Penney, J.-B.J., Blackstone, C.D. and Young, A.B. (1994) Localization of non-N-methyl-D-aspartate glutamate receptors in normal and Alzheimer hippocampal formation. Ann. Neurol., 35: 31–37. Iacopino, A.M. and Christakos, S. (1990) Specific reduction of calcium-binding protein (28-kilodalton calbindin-D) gene expression in aging and neurodegenerative diseases. Proc. Natl. Acad. Sci. U.S.A., 87: 4078–4082. Ikonomovic, M.D., Mizukami, K., Davies, P., Hamilton, R., Sheffield, R. and Armstrong, D.M. (1997) The loss of GluR2(3) immunoreactivity precedes neurofibrillary tangle formation in the entorhinal cortex and hippocampus of Alzheimer brains. J. Neuropathol. Exp. Neurol., 56: 1018–1027. Ikonomovic, M.D., Mizukami, K., Warde, D., Sheffield, R., Hamilton, R., Wenthold, R.J. and Armstrong, D.M. (1999) Distribution of glutamate receptor subunit NMDAR1 in the hippocampus of normal elderly and patients with Alzheimer’s disease. Exp. Neurol., 160: 194–204. Ikonomovic, M.D., Sheffield, R. and Armstrong, D.M. (1995) AMPA-selective glutamate receptor subtype immunoreactivity in the hippocampal formation of patients with Alzheimer’s disease. Hippocampus, 5: 469–486. Iritani, S., Niizato, K. and Emson, P.C. (2001) Relationship of calbindin D28K-immunoreactive cells and neuropathological changes in the hippocampal formation of Alzheimer’s disease. Neuropathology, 21: 162–167. Iwakiri, M., Mizukami, K., Ikonomovic, M.D., Ishikawa, M., Hidaka, S., Abrahamson, E.E., Dekosky, S.T. and Asada, T. (2005) Changes in hippocampal GABABR1 subunit

expression in Alzheimer’s patients: association with Braak staging. Acta Neuropathol. (Berl.), 109: 467–474. Iwamoto, N. and Emson, P.C. (1991) Demonstration of neurofibrillary tangles in parvalbumin-immunoreactive interneurones in the cerebral cortex of Alzheimer-type dementia brain. Neurosci. Lett., 128: 81–84. Jansen, K.L.R., Faull, R.L.M., Dragunow, M. and Synek, B.L. (1990) Alzheimer’s disease — changes in hippocampal Nmethyl-D-aspartate, quisqualate, neurotensin, adenosine, benzodiazepine, serotonin and opioid receptors — an autoradiographic study. Neuroscience, 39: 613–627. Jellinger, K.A. (1997) Neuropathological staging of Alzheimerrelated lesions: the challenge of establishing relations to age. Neurobiol. Aging, 18: 369–375. Jellinger, K.A. and Bancher, C. (1997) Proposals for re-evaluation of current autopsy criteria for the diagnosis of Alzheimer’s disease. Neurobiol. Aging, 18: S55–S65. Ji, Y., Gong, Y., Gan, W., Beach, T., Holtzman, D.W. and Wisniewski, T. (2003) Apolipoprotein E isoform-specific regulation of dendritic spine morphology in apolipoprotein E transgenic mice and Alzheimer’s disease patients. Neuroscience, 122: 305–315. Jin, K., Peel, A.L., Mao, X.O., Xie, L., Cottrell, B.A., Henshall, D.C. and Greenberg, D.A. (2004) Increased hippocampal neurogenesis in Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A., 101: 343–347. Kaufmann, W.A., Barnas, U., Humpel, C., Nowakowski, K., DeCol, C., Gurka, P., Ransmayr, G., Hinterhuber, H., Winkler, H. and Marksteiner, J. (1998) Synaptic loss reflected by secretoneurin-like immunoreactivity in the human hippocampus in Alzheimer’s disease. Eur. J. Neurosci., 10: 1084–1094. Khatchaturian, Z.S. (1985) Diagnosis of Alzheimer’s disease. Arch. Neurol., 31: 545–548. Kiktenko, A.I., Uranova, N.A. and Denisov, D.V. (1997) Quantitative characteristics of changes in synaptic contacts in the hippocampus in Alzheimer’s disease. Neurosci. Behav. Physiol., 27: 681–682. Korbo, L., Amrein, I., Lipp, H.P., Wolfer, D., Regeur, L., Oster, S. and Pakkenberg, B. (2004) No evidence for loss of hippocampal neurons in non-Alzheimer dementia patients. Acta Neurol. Scand., 109: 132–139. Kordower, J.H., Chu, Y., Stebbins, G.T., Dekosky, S.T., Cochran, E.J., Bennett, D. and Mufson, E.J. (2001) Loss and atrophy of layer II entorhinal cortex neurons in elderly people with mild cognitive impairment. Ann. Neurol., 49: 202–213. Kowall, N.W., Quigley, B.-J.J., Krause, J.E., Lu, F., Kosofsky, B.E. and Ferrante, R.J. (1993) Substance P and substance P receptor histochemistry in human neurodegenerative diseases. Regul. Pept., 46: 174–185. Kril, J.J., Patel, S., Harding, A.J. and Halliday, G.M. (2002) Neuron loss from the hippocampus of Alzheimer’s disease exceeds extracellular neurofibrillary tangle formation. Acta Neuropathol. (Berl.), 103: 370–376. Kulmala, H.K. (1985) Immunocytochemical localization of enkephalin-like immunoreactivity in neurons of human

738 hippocampal formation: effects of ageing and Alzheimer’s disease. Neuropathol. Appl. Neurobiol., 11: 105–115. Kumar, U. (2005) Expression of somatostatin receptor subtypes (SSTR1-5) in Alzheimer’s disease brain: an immunohistochemical analysis. Neuroscience, 134: 525–538. Leuba, G., Kraftsik, R. and Saini, K. (1998) Quantitative distribution of parvalbumin, calretinin, and calbindin D-28k immunoreactive neurons in the visual cortex of normal and Alzheimer cases. Exp. Neurol., 152: 278–291. Lim, C., Blume, H.W., Madsen, J.R. and Saper, C.B. (1997) Connections of the hippocampal formation in humans: I. The mossy fiber pathway. J. Comp. Neurol., 385: 325–351. Lotstra, F. and Vanderhaeghen, J.J. (1987) Distribution of immunoreactive cholecystokinin in the human hippocampus. Peptides, 8: 911–920. Lozsadi, D.A. and Larner, A.J. (2006) Prevalence and causes of seizures at the time of diagnosis of probable Alzheimer’s disease. Dement. Geriatr. Cogn. Disord., 22: 121–124. Magloczky, Z., Halasz, P., Vajda, J., Czirjak, S. and Freund, T.F. (1997) Loss of Calbindin-D28K immunoreactivity from dentate granule cells in human temporal lobe epilepsy. Neuroscience, 76: 377–385. Magloczky, Z., Wittner, L., Borhegyi, Z., Halasz, P., Vajda, J., Czirjak, S. and Freund, T.F. (2000) Changes in the distribution and connectivity of interneurons in the epileptic human dentate gyrus. Neuroscience, 96: 7–25. Maguire-Zeiss, K.A., Li, Z.W., Shimoda, L.M. and Hamill, R.W. (1995) Calbindin D28k mRNA in hippocampus, superior temporal gyrus and cerebellum: comparison between control and Alzheimer disease subjects. Brain Res. Mol. Brain Res., 30: 362–366. Marksteiner, J., Lechner, T., Kaufmann, W.A., Gurka, P., Humpel, C., Nowakowski, C., Maier, H. and Jellinger, K.A. (2000) Distribution of chromogranin B-like immunoreactivity in the human hippocampus and its changes in Alzheimer’s disease. Acta Neuropathol. (Berl.), 100: 205–212. Masliah, E., Mallory, M., Hansen, L., Deteresa, R., Alford, M. and Terry, R. (1994) Synaptic and neuritic alterations during the progression of Alzheimer’s disease. Neurosci. Lett., 174: 67–72. Mathieu-Kia, A.M., Fan, L.Q., Kreek, M.J., Simon, E.J. and Hiller, J.M. (2001) Mu-, delta- and kappa-opioid receptor populations are differentially altered in distinct areas of postmortem brains of Alzheimer’s disease patients. Brain Res., 893: 121–134. McNamara, M.J., Ruff, C.T., Wasco, W., Tanzi, R.E., Thinakaran, G. and Hyman, B.T. (1998) Immunohistochemical and in situ analysis of amyloid precursor-like protein-1 and amyloid precursor-like protein-2 expression in Alzheimer disease and aged control brains. Brain Res., 804: 45–51. McNaughton, B.L., Barnes, C.A., Mizumori, S.J.Y., Green, E.J. and Sharp, P.E. (1991) In: Morell F. (Ed.), Kindling and Synaptic Plasticity. Birkha¨user, Basel, pp. 110–123. Mikkonen, M., Alafuzoff, I., Tapiola, T., Soininen, H. and Miettinen, R. (1999) Subfield- and layer-specific changes in parvalbumin, calretinin and calbindin-D28K immunoreactivity in

the entorhinal cortex in Alzheimer’s disease. Neuroscience, 92: 515–532. Mishizen-Eberz, A.J., Rissman, R.A., Carter, T.L., Ikonomovic, M.D., Wolfe, B.B. and Armstrong, D.M. (2004) Biochemical and molecular studies of NMDA receptor subunits NR1/2A/2B in hippocampal subregions throughout progression of Alzheimer’s disease pathology. Neurobiol. Dis., 15: 80–92. Mizukami, K., Grayson, D.R., Ikonomovic, M.D., Sheffield, R. and Armstrong, D.M. (1998) GABAA receptor beta 2 and beta 3 subunits mRNA in the hippocampal formation of aged human brain with Alzheimer-related neuropathology. Brain Res. Mol. Brain Res., 56: 268–272. Mizukami, K., Ikonomovic, M.D., Grayson, D.R., Rubin, R.T., Warde, D., Sheffield, R., Hamilton, R.L., Davies, P. and Armstrong, D.M. (1997) Immunohistochemical study of GABA(A) receptor beta2/3 subunits in the hippocampal formation of aged brains with Alzheimer-related neuropathologic changes. Exp. Neurol., 147: 333–345. Morris, R.G.M. (2006) Elements of a neurobiological theory of hippocampal function: the role of synaptic plasticity, synaptic tagging and schemas. Eur. J. Neurosci., 23: 2829–2846. Morsch, R., Simon, W. and Coleman, P.D. (1999) Neurons may live for decades with neurofibrillary tangles. J. Neuropathol. Exp. Neurol., 58: 188–197. Morys, J., Sadowski, M., Barcikowska, M., Maciejewska, B. and Narkiewicz, O. (1994) The second layer neurones of the entorhinal cortex and the perforant path in physiological ageing and Alzheimer’s disease. Acta Neurobiol. Exp. (Wars.), 54: 47–53. Nagerl, U.V., Mody, I., Jeub, M., Lie, A.A., Elger, C.E. and Beck, H. (2000) Surviving granule cells of the sclerotic human hippocampus have reduced Ca(2+) influx because of a loss of calbindin-D(28k) in temporal lobe epilepsy. J. Neurosci., 20: 1831–1836. Nakazawa, K., Quirk, M.C., Chitwood, R.A., Watanabe, M., Yeckel, M.F., Sun, L.D., Kato, A., Carr, C.A., Johnston, D., Wilson, M.A. and Tonegawa, S. (2002) Requirement for hippocampal CA3 NMDA receptors in associative memory recall. Science, 297: 211–218. van de Nes, J.A., Sandmann-Keil, D. and Braak, H. (2002) Interstitial cells subjacent to the entorhinal region expressing somatostatin-28 immunoreactivity are susceptible to development of Alzheimer’s disease-related cytoskeletal changes. Acta Neuropathol. (Berl.), 104: 351–356. Nitsch, R. and Ohm, T.G. (1995) Calretinin immunoreactive structures in the human hippocampal formation. J. Comp. Neurol., 360: 475–487. Ohm, T.G., Mu¨ller, H., Braak, H. and Bohl, J. (1995) Closemeshed prevalence rates of different stages as a tool to uncover the rate of Alzheimer’s disease-related neurofibrillary changes. Neuroscience, 64: 209–217. Ohm, T.G., Mu¨nch, S., Scho¨nheit, B., Zarski, R. and Nitsch, R. (2002) Transneuronally altered dendritic processing of tangle-free neurons in Alzheimer’s disease. Acta Neuropathol. (Berl.), 103: 437–443.

739 Palop, J.J., Jones, B., Kekonius, L., Chin, J., Yu, G.Q., Raber, J., Masliah, E. and Mucke, L. (2003) Neuronal depletion of calcium-dependent proteins in the dentate gyrus is tightly linked to Alzheimer’s disease-related cognitive deficits. Proc. Natl. Acad. Sci. U.S.A., 100: 9572–9577. Pellegrini-Giampietro, D.E., Bennett, M.V. and Zukin, R.S. (1994) AMPA/kainate receptor gene expression in normal and Alzheimer’s disease hippocampus. Neuroscience, 61: 41–49. Penney, J.B., Maragos, W.F., Greenamyre, J., Debowey, D.L., Hollingsworth, Z. and Young, A.B. (1990) Excitatory amino acid binding sites in the hippocampal region of Alzheimer’s disease and other dementias. J. Neurol. Neurosurg. Psychiatry, 53: 314–320. Petersen, R.C., Parisi, J.E., Dickson, D.W., Johnson, K.A., Knopman, D.S., Boeve, B.F., Jicha, G.A., Ivnik, R.J., Smith, G.E., Tangalso, E.G., Braak, H. and Kokmen, E. (2006) Neuropathologic features of amnestic mild cognitive impairment. Arch. Neurol., 63: 645–646. Price, J.L., Ko, A.I., Wade, M.J., Tsou, S.K., Mckeel, D.W. and Morris, J.C. (2001) Neuron number in the entorhinal cortex and CA1 in preclinical Alzheimer disease. Arch. Neurol., 58: 1395–1402. Rebeck, G.W., Marzloff, K. and Hyman, B.T. (1993) The pattern of NADPH-diaphorase staining, a marker of nitric oxide synthase activity, is altered in the perforant pathway terminal zone in Alzheimer’s disease. Neurosci. Lett., 152: 165–168. Represa, A., Duyckaerts, C., Tremblay, E., Hauw, J.J. and Ben Ari, Y. (1988) Is senile dementia of the Alzheimer type associated with hippocampal plasticity? Brain Res., 457: 355–359. Rinne, J.O., Lonnberg, P., Marjamaki, P., Molsa, P., Sako, E. and Paljarvi, L. (1993) Brain methionine- and leucineenkephalin receptors in patients with dementia. Neurosci. Lett., 161: 77–80. Rissman, R.A., Mishizen-Eberz, A.J., Carter, .L., Wolfe, B.B., De Blas, A.L., Miralles, C.P., Ikonomovic, M.D. and Armstrong, D.M. (2003) Biochemical analysis of GABA(A) receptor subunits alpha 1, alpha 5, beta 1, beta 2 in the hippocampus of patients with Alzheimer’s disease neuropathology. Neuroscience, 120: 695–704. Sampson, V.L., Morrison, J.H. and Vickers, J.C. (1997) The cellular basis for the relative resistance of parvalbumin and calretinin immunoreactive neocortical neurons to the pathology of Alzheimer’s disease. Exp. Neurol., 145: 295–302. Satoh, J., Tabira, T., Sano, M., Nakayama, H. and Tateishi, J. (1991) Parvalbumin-immunoreactive neurons in the human central nervous system are decreased in Alzheimer’s disease. Acta Neuropathol. (Berl.), 81: 388–395. Scheff, S.W. and Price, D.A. (1998) Synaptic density in the inner molecular layer of the hippocampal dentate gyrus in Alzheimer disease. J. Neuropathol. Exp. Neurol., 57: 1146–1153. Scheff, S.W. and Price, D.A. (2003) Synaptic pathology in Alzheimer’s disease: a review of ultrastructural studies. Neurobiol. Aging, 24: 1029–1046.

Scheff, S.W., Sparks, D.L. and Price, D.A. (1996) Quantitative assessment of synaptic density in the outer molecular layer of the hippocampal dentate gyrus in Alzheimer’s disease. Dementia, 7: 226–232. Scho¨nheit, B., Zarski, R. and Ohm, T.G. (2004) Spatial and temporal relationships between plaques and tangles in Alzheimer-pathology. Neurobiol. Aging, 25: 697–711. Seress, L. (1988) Interspecies comparison of the hippocampal formation shows increased on the regio superior in the Ammon’s horn of the human brain. J. Hirnforsch., 29: 35–340. Seress, L. and Frotscher, M. (1990) Morphological variability is a characteristic feature of granule cells in the primate fascia dentata: a combined Golgi/electron microscope study. J. Comp. Neurol., 293: 253. Seress, L., Gulyas, A.I., Ferrer, I., Tunon, T., Soriano, E. and Freund, T. (1993) Distribution, morphological features, and synaptic connections of parvalbumin- and calbindin D28kimmunoreactive neurons in the human hippocampal formation. J. Comp. Neurol., 337: 208–230. Seress, L. and Mrzljak, L. (1987) Basal dendrites of granule cells are normal features of the fetal and adult dentate gyrus of both monkey and human hippocampal formations. Brain Res., 405: 169–174. Simic, G., Kostovic, I., Winblad, B. and Bogdanovic, N. (1997) Volume and number of neurons of the human hippocampal formation in normal aging and Alzheimer’s disease. J. Comp. Neurol., 379: 482–494. Singer, D., Lehmann, J., Hanisch, K., Hartig, W. and Hoffmann, R. (2006) Neighbored phosphorylation sites as PHF-tau specific markers in Alzheimer’s disease. Biochem. Biophys. Res. Commun., 346: 819–828. Sloviter, R.S., Sollas, A.L., Barbaro, N.M. and Laxer, K.D. (1991) Calcium-binding protein (Calbindin-D28K) and parvalbumin immunocytochemistry in the normal and epileptic human hippocampus. J. Comp. Neurol., 308: 381–396. Solodkin, A., Veldhuizen, S.D. and van Hoesen, G.W. (1996) Contingent vulnerability of entorhinal parvalbumin-containing neurons in Alzheimer’s disease. J. Neurosci., 16: 3311–3321. Spillantini, M.G. and Goedert, M. (1998) Tau protein pathology in neurodegenerative diseases. Trends Neurosci., 21: 428–433. Stoub, T.R., Detoledo-Morrell, L., Stebbins, G.T., Leurgans, S., Bennett, D.A. and Shah, R.C. (2006) Hippocampal disconnection contributes to memory dysfunction in individuals at risk for Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A., 103: 10041–10045. Thal, D.R., Holzer, M., Rub, U., Waldmann, G., Gunzel, S., Zedlick, D. and Schober, R. (2000) Alzheimer-related taupathology in the perforant path target zone and in the hippocampal stratum oriens and radiatum correlates with onset and degree of dementia. Exp. Neurol., 163: 98–110. Thal, D.R., Rub, U., Orantes, M. and Braak, H. (2002) Phases of A beta-deposition in the human brain and its relevance for the development of AD. Neurology, 58: 1791–1800. Torrealba, F. and Carrasco, M.A. (2004) A review on electron microscopy and neurotransmitter systems. Brain Res. Brain Res. Rev., 47: 5–17.

740 Uchihara, T., Nakamura, A., Yamazaki, M. and Mori, O. (2001) Evolution from pretangle neurons to neurofibrillary tangles monitored by thiazin red combined with Gallyas method and double immunofluorescence. Acta Neuropathol. (Berl.), 101: 535–539. Ulas, J. and Cotman, C.W. (1997) Decreased expression of N-methyl-D-aspartate receptor 1 messenger RNA in select regions of Alzheimer brain. Neuroscience, 79: 973–982. Uylings, H.B.M., Ruiz-Marcos, A. and van Pelt, J. (1986) The metric analysis of three-dimensional dendritic tree patterns: a methodological review. J. Neurosci. Methods, 18: 127–151. Wakabayashi, K., Hansen, L.A., Vincent, I., Mallory, M. and Masliah, E. (1997) Neurofibrillary tangles in the dentate granule cells of patients with Alzheimer’s disease, Lewy body disease and progressive supranuclear palsy. Acta Neuropathol. (Berl.), 93: 7–12. Wakabayashi, K., Narisawa-Saito, M., Iwakura, Y., Arai, T., Ikeda, K., Takahashi, H. and Nawa, H. (1999) Phenotypic down-regulation of glutamate receptor subunit GluR1 in Alzheimer’s disease. Neurobiol. Aging, 20: 287–295. West, M.J., Coleman, P.D., Flood, D.G. and Troncoso, J.C. (1994) Differences in the pattern of hippocampal neuronal loss in normal ageing and Alzheimer’s disease. Lancet, 344: 769–772.

West, M.J. and Gundersen, H.J.G. (1990) Unbiased stereological estimation of the number of neurons in the human hippocampus. J. Comp. Neurol., 296: 1–22. West, M.J., Kawas, C.H., Stewart, W.F., Rudow, G.L. and Troncoso, J.C. (2004) Hippocampal neurons in pre-clinical Alzheimer’s disease. Neurobiol. Aging, 25: 1205–1212. West, M.J. and Slomianka, L. (1998) Total number of neurons in the layers of the human entorhinal cortex. Hippocampus, 8: 69–82. Williams, R.S. and Matthysse, S. (1986) Age-related changes in Down syndrome brain and the cellular pathology of Alzheimer disease. Prog. Brain Res., 70: 49–67. Yao, P.J., Morsch, R., Callahan, L.M. and Coleman, P.D. (1999) Changes in synaptic expression of clathrin assembly protein AP180 in Alzheimer’s disease analysed by immunohistochemistry. Neuroscience, 94: 389–394. Yew, D.T., Li, W.P., Webb, S.E., Lai, H.W. and Zhang, L. (1999) Neurotransmitters, peptides, and neural cell adhesion molecules in the cortices of normal elderly humans and Alzheimer patients: a comparison. Exp. Gerontol., 34: 117–133. Zipp, F., Nitsch, R., Soriano, E. and Frotscher, M. (1989) Entorhinal fibers form synaptic contacts on parvalbuminimmunoreactive neurons in the rat fascia dentata. Brain Res., 495: 161–166.

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 40

Hippocampal atrophy and disconnection in incipient and mild Alzheimer’s disease Leyla deToledo-Morrell, Travis R. Stoub and Changsheng Wang Department of Neurological Sciences, Rush University Medical Center, Chicago, IL 60612, USA

Abstract: Quantitative imaging techniques allow the in vivo investigation of age and disease related changes in the brain and their relation to cognitive function. In this chapter we review imaging evidence indicating that the entorhinal cortex and hippocampus show atrophy very early in Alzheimer’s disease (AD) and in individuals who are at risk of developing AD compared to age appropriate controls. Furthermore, the extent and rate of atrophy of the entorhinal cortex, a brain region pathologically involved very early in the disease process, can predict who among the elderly will develop AD. Techniques that assess the integrity of white matter further demonstrate that alterations in the parahippocampal white matter in the region that includes the perforant path could partially disconnect the dentate gyrus and other hippocampal subfields from incoming sensory information. Such partial disconnection and degradation in transmission of sensory information in people at risk of AD and in patients with very mild AD could contribute to the memory dysfunction associated with the early stages of the disease. Keywords: entorhinal cortex; perforant path; imaging; aging; memory The hippocampal dentate gyrus has been implicated as the sub-region most sensitive to the effects of advancing age (Small et al., 2004). Although synaptic loss in the dentate molecular layer has been demonstrated in patients with Alzheimer’s disease (AD) (Scheff and Price, 2003; Scheff et al., 2006), cell loss associated with the disease seems to be more prominent in the CA1 region of the hippocampus (West et al., 1994, 2000). In addition, AD-related pathological changes such as neurofibrillary tangles are most notable in the subiculum and the CA1 region (Van Hoesen and Hyman, 1990). Unfortunately, until recently, in vivo structural imaging techniques did not have the resolution to differentiate sub-regions of the hippocampal formation (hippocampus proper and dentate). However, a recently published structural MRI protocol acquired with a high-field (4 Tesla)

Introduction High-resolution structural magnetic resonance imaging (MRI) techniques provide a unique tool for examining alterations in brain anatomy in vivo during healthy aging and in age-related diseases. In addition, such techniques allow us to (a) examine the relation between alterations in given brain regions and the sequential development of behavioral symptoms in degenerative diseases, and (b) delineate the specific role of certain mesial temporal lobe structures, such as the entorhinal cortex and hippocampus in human memory function, because of the age or disease-related degeneration in those structures. Corresponding author. Tel.: +1 (312) 942 5399; Fax: +1 (312) 563 3570; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63040-4

741

742

magnet holds promise for differentiating hippocampal sub-fields in vivo in future investigations (Mueller et al., 2007). In this chapter, we will review the evidence indicating that there is atrophy of both the entorhinal cortex and hippocampus in the very early stages of AD, as well as in individuals with amnestic mild cognitive impairment (MCI) who are at high risk of developing AD (Petersen et al., 1999; Petersen, 2000). These elderly individuals have impairment in memory function, but do not meet criteria for dementia. Furthermore, we will demonstrate that white matter changes in the region of the parahippocampal gyrus that includes the perforant path in people with amnestic MCI and mild AD could exacerbate the memory dysfunction characteristic of the initial stages of AD, by partially disconnecting the hippocampus from incoming sensory information. Since the MRI techniques in the experiments from our laboratory to be described were similar, they will be detailed first.

Acquisition and quantitation of MRI data All MR images were acquired on a 1.5 Tesla General Electric Signa scanner with the manufacturer’s three-dimensional (3D) Fourier transform spoiled gradient recalled (SPGR) pulse sequence. The acquisition parameters for the T1 weighted sequence were as follows: 124 contiguous images acquired in the coronal plane, 1.6 mm-thick sections, matrix ¼ 256
192, field of view ¼ 22 cm, TR/TE ¼ 34/7, flip angle ¼ 351, signals averaged ¼ 1. The analyze software package (Mayo Clinic Foundation, Rochester, MN, USA) was used for determining the volume of regions of interest and for co-registering sequential scans. Both the entorhinal cortex and hippocampal formation were manually segmented as described below. To correct for individual differences in brain size, entorhinal and hippocampal volumes were divided by total intracranial volume (an accepted measure of pre-morbid brain size) derived from sagittaly reformatted 5 mm slices. To compute intracranial volume, the inner table of the cranium was traced in consecutive sagittal sections spanning the whole

brain. At the level of the foramen magnum, a straight line was drawn from the inner surface of the clivus to the occipital bone. Entorhinal cortex volume was quantified with the use of a protocol developed and validated in our laboratory, technical details of which are presented in Goncharova et al. (2001). The advantage of this protocol is that entorhinal volume is measured from the same oblique coronal sections most commonly used for hippocampal volumetry, so that one of these two adjacent structures is not over estimated at the expense of the other. Briefly, both entorhinal and hippocampal volumes were computed separately for the right and left hemispheres from coronal slices reformatted to be perpendicular to the long axis of the hippocampus. For the entorhinal cortex, tracing began with the first section in which the gyrus ambiens, amygdala and the white matter of the parahippocampal gyrus first appeared visible. The dorsomedial border in rostral sections was the sulcus semiannularis and in caudal sections the subiculum. The shoulder of the collateral sulcus was used as the lateral border, a somewhat conservative criterion that allowed consistency in tracings and avoided the use of different lateral borders, depending on individual differences in the depth of the collateral sulcus (Insausti et al., 1998). The last section traced was three 1.6 mm slices rostral to the image in which the lateral geniculate first appeared visible. The protocol and validation procedures used for quantifying hippocampal volume were published previously (Wilson et al., 1996; deToledo-Morrell et al., 1997). For hippocampal volumetry, tracings started with the first section where the hippocampus could be clearly differentiated from the amygdala by the alveus and included the fimbria, the dentate gyrus, the hippocampus proper and the subiculum. Tracings continued on all consecutive slices until the slice before the full appearance of the fornix. Investigators involved in MRI analyses were blinded to clinical information until all such analyses were completed. Figure 1 shows sample tracings of the entorhinal cortex and hippocampal formation. Both regions of interest were normalized using the following

743

Fig. 1. A sample coronal image illustrating the segmentation of the entorhinal cortex (outlined, right side) and hippocampus (outlined, left side).

formula: (absolute volume in mm3/intracranial volume in mm3)
1000. Hippocampal and entorhinal cortex atrophy in incipient and mild AD The entorhinal cortex and hippocampus are part of the mesial temporal lobe memory system (Squire and Zola-Morgan, 1991; Young et al., 1997). The entorhinal cortex receives sensory input from the neocortex and provides the hippocampus with multi-modal sensory information via the perforant path. These brain regions have received special attention in both post mortem and in vivo investigations on the pathophysiology of AD, since memory dysfunction is one of the earliest hallmarks of the disease. Post mortem pathological studies have implicated the entorhinal cortex and the transentorhinal region as sites that are involved in the early stages of AD and in individuals with MCI (Braak and Braak, 1991, 1995; Gomez-Isla et al., 1996; Braak et al., 1998; Kordower et al., 2001). In vivo MRI investigations have also demonstrated atrophy of both the entorhinal cortex and hippocampus not only in patients diagnosed with mild AD, but also in individuals with amnestic MCI and those with subjective cognitive complaints (Convit et al., 1997; Jack et al., 1997, 1999; de Leon et al., 1996; deToledo-Morrell et al., 1997, 2000a, b; Killiany et al., 2000, 2002; Xu et al., 2000; Dickerson et al., 2001; Du et al., 2001; Jessen et al., 2006; Saykin et al., 2006). In an earlier study from our laboratory (Dickerson et al., 2001) we compared MRI-derived

entorhinal and hippocampal volumes in healthy elderly controls with no cognitive impairment, in people who presented at the clinic with cognitive complaints, but who did not meet criteria for dementia and in patients with very mild AD. Interestingly, the two patient groups differed significantly from controls in entorhinal volume, but not from each other; in contrast, they differed from each other as well as from controls in hippocampal volume, with the very mild AD cases showing the greatest atrophy. Recently, interest has increased in directly comparing the accuracy with which the volumes of these two mesial temporal lobe structures can differentiate those who convert from a nondemented status to AD (i.e., those who are at risk of developing AD). The study described below was undertaken to examine this question using an initial baseline MRI in a group of participants diagnosed with amnestic MCI (deToledo-Morrell et al., 2004). The participants consisted of 27 elderly individuals (Z65 years of age) who received a clinical diagnosis of amnestic MCI during their baseline evaluation. They had a significant deficit in the memory domain, but did not meet criteria for dementia. After the baseline diagnosis of MCI, all participants were followed with yearly clinical evaluations. During a 36-month follow-up period, 10 of the 27 participants originally diagnosed with amnestic MCI converted to AD. All clinical evaluations were carried out at the Rush Alzheimer’s Disease Center (RADC, Chicago, IL) as previously described (Wilson et al., 1996; deToledo-Morrell et al., 1997). Briefly, the

744

evaluation incorporated the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD, Morris et al., 1989) procedures and included a medical history, neurological examination, neuropsychological testing, informant interview and blood tests. Exclusion criteria for entry into the study consisted of evidence of other neurologic, psychiatric or systemic conditions that could cause cognitive impairment (e.g., stroke, alcoholism, major depression, a history of temporal lobe seizures). The clinical diagnosis of probable AD in the longitudinal clinical evaluations followed NINCDS/ADRDA guidelines (McKhann et al., 1984); it required a history of cognitive decline and neuropsychological test evidence of impairment in at least two cognitive domains, one of which had to be memory. Demographic data and Mini Mental State Examination (MMSE, Folstein et al., 1975) scores for the MCI-converters and non-converters are presented in Table 1. The maximum score on the MMSE is 30, with scores Z27 being considered within the normal range. The individuals who developed AD (converters) did not differ from the non-converters in age, but there was a significant difference between them in MMSE scores and in years of education. Total (right+left) normalized entorhinal and hippocampal volumes derived from the baseline MRI for MCI converters and non-converters are shown in Fig. 2. Converters differed significantly from non-converters in total entorhinal [t(25) ¼ 3.090, p ¼ 0.005] and hippocampal [t(25) ¼ 2.442, p ¼ 0.022] volume. Multivariate logistic regression analyses adjusted for education were carried out to examine how well baseline entorhinal and hippocampal volumes could

predict conversion to AD. In these analyses, total normalized entorhinal and hippocampal volumes were used as predictors to determine the extent to which each region of interest contributed separately to describing group differences. The results demonstrated that although both total entorhinal and total hippocampal volumes were independent predictors of conversion [w2(1) ¼ 7.3, p ¼ 0.007, odds ratio per 0.1 unit volume change ¼ 2.024 for the entorhinal cortex and w2(1) ¼ 6.2, p ¼ 0.013, odds ratio ¼ 1.318 for the hippocampus], entorhinal volume was a better predictor with a concordance rate of 90.6%. Thus, for every 0.1 unit decrease in entorhinal cortex volume, the odds of conversion doubled, while the same unit decrease in hippocampal volume increased the chances of conversion by only 30%. Similar analyses were carried out on entorhinal and hippocampal volumes separated by hemisphere. In this case, the right hemisphere entorhinal volume best predicted conversion with a concordance rate of 93.5%. Unlike these findings, hemispheric differences in hippocampal volume did not lead to differential predictions regarding conversion. The in vivo anatomical results presented here are in agreement with post mortem pathological findings and underscore the early involvement of the entorhinal cortex in AD. Taken together, the results just described indicate that entorhinal volume is a good predictor of conversion from MCI to AD. In agreement with these results, others have also reported that baseline entorhinal volume differentiated those at risk for developing AD better than hippocampal volume (Killiany et al., 2002). Our results also demonstrated that the right hemisphere entorhinal volume was the best

Table 1. Demographic characteristics of participants

Age (years) Education (years) Female/male MMSE score  Significantly differently from each other (po0.01).

MCI converters (N ¼ 10)

MCI non-converters (N ¼ 17)

82.7 (74.5) (75.6–89.2) 18.4 (72.1) 6/4 26.1 (71.4)

81.1 (78.1) (66.1–98.0) 15.2 (73.1) 9/8 28.0 (71.8)

745

Fig. 2. Box plot comparing MRI-derived normalized entorhinal (left hand side) and hippocampal (right hand side) volumes in participants with MCI who converted to a diagnosis of Alzheimer’s disease, in contrast to those who did not. The central box shows the data between the upper and lower quartiles, with the median represented by the line. The height of the line is the interquartile range (IQR); the ‘‘whiskers’’ extend from the upper and lower quartiles to a distance of 1.5 IQR away or to the most extreme data point within that range, whichever is closer (adapted with permission from deToledo-Morrell et al., 2004).

predictor of risk of AD. Very similar findings have been recently reported by another laboratory (Tapiola et al., 2006). The reason for this hemispheric difference in entorhinal volume in those who are at risk of developing AD is not clear since, in general, post mortem pathological investigations do not carry out hemispheric comparisons. It should be pointed out that changes in the entorhinal cortex in elderly individuals seem to be predictive of conversion not only from MCI to AD, but also of the transition from normal cognition to MCI or AD. In an interesting report, de Leon and his colleagues (de Leon et al., 2001) used MRI guided resting state [2-18F]-2-deoxy-D-glucose/positron-emission tomography to assess metabolic changes in given brain regions of interest. They found that a reduction in entorhinal cortex metabolism was predictive of cognitive decline among cognitively normal elderly subjects. Although obtained with a different imaging modality, these results are very much in line with our findings and provide further evidence for the very early involvement of the entorhinal cortex in the

development of AD. Similarly, Rodrigue and Raz (2004) found that the rate of entorhinal atrophy in healthy elderly people was associated with greater memory decline.

MRI Predictors of risk of incident AD: a longitudinal study As discussed above, a large number of cross-sectional MRI investigations have demonstrated atrophy of both the hippocampus and entorhinal cortex in patients with mild to moderate AD and in people with cognitive impairment who are at risk of developing AD. Longitudinal studies that measured the volumes of these two mesial temporal lobe structures have shown the annual rate of atrophy of both structures to be greater in patients with AD and in those with MCI compared to elderly control subjects (Jack et al., 1998, 2000, 2004; Cardenas et al., 2003; Du et al., 2003, 2004). However, most of these longitudinal investigations used scans from two-time points separated by 1–5

746

years for determining rate of atrophy, rather than multiple yearly scans. The experiment from our laboratory described above examined MRI markers of AD using a baseline scan and longitudinal clinical evaluations. In the next study, we used proportional odds models to assess the relationship between rate of atrophy of mesial temporal lobe structures and risk of incident AD among non-demented elderly individuals (Stoub et al., 2005). Rate of atrophy of regions of interest was derived from yearly scans. Data reported here were obtained from 58 nondemented elderly participants (age 65 or older) who were followed with annual clinical evaluations as well as high-resolution MRI scans for up to 5 years (baseline and 5 years of follow-up). The clinical evaluations and MRI acquisition and analysis protocols were as described above. Twentythree of the 58 participants received a diagnosis of amnestic MCI during the baseline evaluation and 35 were healthy controls with no cognitive impairment. The number of years until a diagnosis of AD was modeled using proportional odds models (Kalbfleish and Prentice, 1980). For this purpose, AD was treated as an all absorbing state, with onset at the first evaluation at which NINCDS/ ADRDA criteria (McKhann et al., 1984) were met. For participants who did not develop AD, the number of years to last clinical diagnosis was the event time, which was coded as censored. Covariates considered were demographic variables (age, sex, years of education), baseline diagnosis (MCI or control), and volumetric measures (normalized total hippocampal and entorhinal volumes). Volumes of regions of interest derived from scans following a diagnosis of AD were not included in the analyses (for further analytic details, see Stoub et al., 2005). Fourteen of the 58 non-demented participants developed AD during the follow-up period; three of these 14 participants entered the study as control subjects with no cognitive impairment and the rest with a diagnosis of amnestic MCI. Demographic characteristics for participants who did and did not develop AD are presented in Table 2. Since education, but not age, was found to be a

significant predictor of event time, all models were education adjusted. The results of the proportional odds models demonstrated that initial diagnosis of MCI was an important predictor of incident AD [w2(1) ¼ 16.569, po0.0001; Fig. 3A]. Furthermore, both baseline entorhinal volume and its rate of atrophy were independent predictors of incident AD; their addition to initial diagnosis improved the model [w2(2) ¼ 10.904, po0.0043]. Lower baseline entorhinal volume was associated with greater risk of AD (see Fig. 3B); for every 0.25 unit of decrease in baseline entorhinal volume, the odds of incident AD increased by a factor of 4.98. Similarly, increased rate of decline in entorhinal volume was associated with greater risk of AD (see Fig. 3C); for every 0.02 unit of increase in rate of entorhinal atrophy, the likelihood of developing AD increased by a factor of 2.18. Surprisingly, hippocampal volume was not an independent predictor of risk of AD; the addition of baseline hippocampal volume and its slope of decline to initial diagnosis did not improve the model based on initial diagnosis alone [w2(2) ¼ 2.701, p ¼ 0.26]. APOE e4 allele status, a risk factor for AD, was available for 53 of 58 participants (see Table 2). When the analyses described above were carried out on data from these 53 participants, results Table 2. Baseline demographic characteristics of participants

N Sex Male Female Age (years) MMSE Education (years) APOE status 4/4 3/4 3/3 2/4 2/3 2/2 Not available

Non-converters

Converters to AD

44

14

16 28 80 (76) 28.5 (71.5) 15 (73)

4 10 81 (76) 26.9 (71.8) 18 (73)

0 8 25 0 7 0 4

0 5 6 2 0 0 1

Significantly different p ¼ 0.002.  Significantly different p ¼ 0.019.

747

remained similar to those already detailed. However, APOE e4 allele status (i.e., the presence or absence of any e4 allele) was not found to be an important predictor of incident AD when added to

models based on initial diagnosis and entorhinal and hippocampal volume [w2(2) ¼ 1.972, p ¼ 0.16], although the presence or absence of any APOE e4 allele by itself was a predictor of event time [w2(1) ¼ 5.00, p ¼ 0.025]. These results are in line with those from a prior report (Bennett et al., 2003) which demonstrated that the effect of the e4 allele, which was strongly associated with the likelihood of clinical AD, was no longer significant after controlling for the effects of AD pathology, suggesting that such pathology mediates the effect of allele status on clinical disease. In our experiments, MRI-derived hippocampal and entorhinal atrophy could be viewed as surrogate markers of the underlying pathology. The major finding of this study is that among non-demented individuals, the initial size of the entorhinal cortex and its rate of atrophy are significant risk factors for AD, with initial small size and steep declines in volume being associated with incident AD. In contrast, baseline size and rate of hippocampal atrophy were not found to be independent risk factors. Other investigators have reported that hippocampal atrophy at baseline in patients with MCI is associated with risk of AD (Jack et al., 1999). Although our results may, at first sight, seem contrary to this report, it is important to point out that in this paper the investigators assessed only MRI-based hippocampal atrophy and did not directly compare hippocampal and entorhinal volumes in their models. Fig. 3. Estimates of the probability of not yet having been diagnosed with Alzheimer’s disease as a function of years of follow-up for two different populations. The dotted line represents the reference population and the solid line depicts the hypothetical population. The reference population in all panels consists of people who have no cognitive impairment (NCI) at baseline, have an average education level of 16 years, whose normalized entorhinal cortex volume is 1.00 (chosen to be close to the overall mean of 0.97027 at baseline), and for whom the slope of the rate of entorhinal atrophy is 0.03 (which is close to the mean slope of 0.03074). For the model shown in A, the hypothetical population is different from the reference only in that the initial diagnosis is MCI, not NCI. In B, the hypothetical population is different from the reference in terms of baseline entorhinal cortex volume, which was 25% smaller than 1 or 0.75. The hypothetical population shown in C is different from the reference population only in that entorhinal cortex volume declines faster with a slope of 0.05 rather than 0.03 (adapted with permission from Stoub et al., 2005).

748

As discussed earlier in this chapter, post mortem studies have suggested that AD-related pathology may start in the transentorhinal and entorhinal regions and then spread to the hippocampus (Braak and Braak, 1991, 1995; Braak et al., 1998). The fact that shrinkage of the entorhinal cortex, but not of the hippocampus was a predictor of AD in our study is in support of these post mortem findings and suggests that the extent of entorhinal involvement in the disease process may precede hippocampal involvement.

Hippocampal disconnection due to parahippocampal white matter changes in incipient AD: relation to memory dysfunction As discussed up to this point, previous histopathological findings (Hyman et al., 1984; Braak and Braak, 1991, 1995; Gomez-Isla et al., 1996; Braak et al., 1998; Mufson et al., 1999; Kordower et al., 2001), as well as in vivo MRI results indicate that the entorhinal cortex and hippocampus are pathologically involved very early in patients with AD and in those with MCI who are at high risk for developing AD. More specifically, two studies (Gomez-Isla et al., 1996; Kordower et al., 2001) found a loss of entorhinal cortex layer II neurons in patients with AD and in those with MCI, compared to elderly controls. These neurons receive multimodal sensory input from primary sensory and association cortices and project this information to the hippocampus via the perforant path, a white matter tract located in the anterior medial portion of the parahippocampal gyrus (Van Hoesen and Pandya, 1975; Van Hoesen et al., 1975; Amaral et al., 1987). The loss of layer II neurons in the entorhinal cortex and their axons is important since it would result in alterations in information flow to the hippocampus. In addition, damage to the white matter of the parahippocampal gyrus could disrupt afferent connections to the entorhinal cortex and ultimately disconnect multi-modal sensory input to the hippocampus, information vital to the formation of new memories. These changes in white matter may be detectable in vivo with imaging techniques, but have received less attention

in investigations on the pathophysiology of incipient or clinically diagnosed AD and the anatomical underpinnings of the memory dysfunction associated with the disease. Up to this point we have concentrated on a description of structural alterations detected in vivo in the entorhinal cortex and hippocampus. In the next experiment (Stoub et al., 2006) we turned our attention to assessing the contribution of alterations in the parahippocampal white matter, in addition to entorhinal and hippocampal atrophy, to the memory dysfunction in people with amnestic MCI who are at risk of developing AD. The participants in the study consisted of 40 older individuals (mean age 77.977.5 years) who met criteria for amnestic MCI and 50 healthy controls (mean age 78.176.0 years) with no cognitive impairment. Clinical evaluations and acquision of MRI scans were as described before. White matter volume changes throughout the brains of individuals with amnestic MCI, compared to agematched healthy controls were assessed with voxel-based morphometry (VBM, Good et al., 2001). Automated techniques such as VBM that provide quick, unbiased means of comparing changes in white matter at the voxel level are especially useful when there is no a priori knowledge of where in the brain such white matter changes may be taking place. In addition to white matter changes, we assessed differences between the groups in entorhinal cortex and hippocampal volume using manual segmentation and volumetric measurement of the two structures since, in our hands, these measures are better at detecting subtle changes in small gray matter regions. Group differences in whole-brain white matter volumes were assessed with the two-sample t statistic within the Statistical Parametric Mapping software (SPM 99); the significance threshold was set at 0.001. Regions of statistically significant difference were identified according to the Montreal Neurological Institute template and then converted to Talariach coordinates (Talairach and Turnoux, 1988) with the use of MNITOTAL software. The Talairach coordinates were transformed to actual brain areas by using the Talairach Daemon system (Lancaster et al., 2000). These coordinates were used to construct regions

749

of interest that were then applied to individual white matter density maps to extract individual volume values. In order to relate structural changes to memory function, individual memory Z scores were calculated based on combined performance on declarative memory tasks using seven declarative memory scores. These consisted of immediate and delayed recall of the East Boston Story (Albert et al., 1991) and of Story A from the Logical Memory of the Wechsler memory scale — Revised (Wechsler, 1987). An additional test involved the learning and retention of a 10-word list from the CERAD battery (Morris et al., 1989). The three scores for this test included Word List Memory (the total number of words immediately recalled after each of three consecutive presentations of the list), Word List Recall (the number of words recalled after a delay) and Word List Recognition (the number of words correctly recognized in a four-alternative, forced-choice format, administered after Word List Recall). The demographic characteristics of all participants are listed in Table 3. The amnestic MCI group did not differ from healthy controls in age or education, but had significantly lower [t(88) ¼ 6.479, po0.001] MMSE scores and memory Z scores [t(88) ¼ 9.921, po0.001], which is not surprising. As shown in Fig. 4, there was a significant decrease (po0.001) in white matter volume in participants with MCI compared to controls in the anterior-medial aspect of both parahippocampal gyri, in the region that includes the perforant path.

Table 3. Demographic characteristics and memory Z scores of participants

Number of participants Age (years) Education (years) Male/female MMSE score Memory Z score  po0.001.

Healthy controls

MCI

50 78.1 (76.0) 15.0 (72.8) 15/35 29.0 (70.9) 0.5337 (70.4681)

40 77.9 (77.5) 16.2 (73.1) 16/24 27.2 (71.6) 0.5948 (70.6111)

In addition to the parahippocampal white matter change, the group of MCI participants also showed significant entorhinal and hippocampal atrophy (po0.001 for both) similar to what our laboratory and others have reported. The atrophy in both structures was bilateral. For predicting memory function from MRI measures, each of the three MRI indices (i.e., total parahippocampal white matter volume, entorhinal cortex volume and hippocampal volume) was entered singly into a regression analysis that used as the dependent variable individual memory Z scores. Each of the three anatomical measures was found to be a significant predictor of memory function [F(1,88) ¼ 24.08, po0.001 for hippocampal volume; F(1,88) ¼ 10.35, po0.002 for entorhinal cortex volume; and F(1,88) ¼ 11.11, po0.001 for parahippocampal white matter volume]. However, when all three MRI measures were entered simultaneously into a multiple regression model, only hippocampal and white matter volume measures were found to be significant predictors of memory function [t(86) ¼ 3.01, p ¼ 0.003 and t(86) ¼ 2.22, po0.029 respectively]. The major contribution of these data is that, in addition to a reduction in the size of the hippocampus, a decrease in white matter volume in the region of the parahippocampal gyrus that includes the perforant path significantly contributes to the memory dysfunction in people with MCI. The perforant path that supplies the hippocampus with multi-modal information is pathologically involved very early in AD (Hyman et al., 1984). The investigation by Hyman and his collaborators was carried out on post mortem tissue and demonstrated that in patients with very mild AD, there is disconnection of the hippocampus from sensory cortical inputs as a result of loss of layer II cells in the entorhinal cortex. The authors hypothesized that such disconnection could contribute to memory dysfunction during the very early stages of AD; however, the report did not include cognitive data proximal to death. Our results provide an in vivo demonstration of hippocampal disconnection resulting from white matter changes in the region of the perforant path that impacts memory function.

750

Fig. 4. Color map showing significant (p ¼ 0.001) voxels of decreased white matter density in participants with amnestic MCI compared with controls, superimposed on coronal and sagittal slices of a template based on data for all subjects in the study. The image is masked to include white matter regions. The colors correspond to the t values shown on the color bar. Note the bilaterality of significant differences in the white matter of the parahippocampal gyrus (adapted with permission from Stoub et al., 2006). (See Color Plate 40.4 in color plate section.)

White matter volume change may reflect not only loss of afferent and efferent fibers in the region of the parahippocampal gyrus, but also partial demyelination in remaining fibers. Recently developed diffusion tensor imaging (DTI) that detects microstructural alterations in normal appearing white matter could aid in determining the microstructural changes in remaining fibers that may further degrade impulse transmission. DTI detects microstructural alterations in white matter by measuring the directionality of molecular diffusion (fractional anisotropy, FA). Highly organized and well myelinated white matter tracts have high FA because diffusion is constrained by the tract’s cellular organization. As white matter is damaged, FA decreases due to decreased anisotropic diffusion. In fact, in the most recent study from our laboratory (Wang et al., 2006), using a high resolution DTI protocol, we demonstrated that in very mild AD cases there is not only loss of white matter volume in the region of the parahippocampal gyrus, but that remaining fibers are not

normal, as indicated by lower FA values compared to age appropriate controls.

Conclusions In this chapter, we reviewed evidence from in vivo MRI investigations that demonstrate significant entorhinal cortex and hippocampal atrophy not only in patients with very mild AD, but also in individuals who are at risk of developing AD. Atrophy in these mesial temporal lobe structures important for memory function can predict who will develop AD a number of years prior to a clinical diagnosis. Recently, there has been increased interest in investigating alterations in white matter that could lead to degradation in information flow due to partial disconnection of different brain regions. We have shown that white matter volume loss in the region of the parahippocampal gyrus contributes significantly to memory dysfunction in individuals with incipient AD. The exact

751

underlying mechanism of the white matter volume loss in the parahippocampal region cannot be determined in vivo with the tools currently available to us. One may speculate that a decrease in white matter fibers in this region, reflected in white matter volume change, may cause a disruption of input to the dentate gyrus and hippocampus proper from the entorhinal cortex. Because these incoming fibers arise from cells in the entorhinal cortex, entorhinal atrophy may, in part, be the origin of the decrease in white matter volume. Future in vivo imaging studies using high field scanners are necessary to fully elaborate the changes postsynaptic to the entorhinal input; such changes in response to entorhinal atrophy are likely to include the dentate gyrus.

Acknowledgments This research was supported by grants P01 AG09466, P30 AG10161 and R01 AG17917 from the National Institute on Aging, National Institutes of Health. References Albert, M., Smith, L., Scherr, P., Taylor, J., Evans, D. and Funkenstein, H. (1991) Use of brief cognitive tests to identify individuals in the community with clinically diagnosed Alzheimer’s disease. Int. J. Neurosci., 57(3–4): 67–178. Amaral, D.G., Insausti, R. and Cowan, W.M. (1987) The entorhinal cortex of the monkey. I. Cytoarchitectonic organization. J. Comp. Neurol., 264(3): 326–355. Bennett, D.A., Wilson, R.S., Schneider, J.A., Evans, D.A., Aggarwal, N.T., Arnold, S.E., Cochran, E.J., Berry-Kravis, E. and Bienias, J.L. (2003) Apolipoprotein E epsilon4 allele, AD pathology, and the clinical expression of Alzheimer’s disease. Neurology, 60(2): 246–252. Braak, H. and Braak, E. (1991) Neuropathological staging of Alzheimer-related changes. Acta Neuropathol., 82(4): 239–259. Braak, H. and Braak, E. (1995) Staging of Alzheimer’s diseaserelated neurofibrillary changes. Neurobiol. Aging, 16(3): 271–288. Braak, H., Braak, E., Bohl, J. and Bratzke, H. (1998) Evolution of Alzheimer’s disease related cortical lesions. J. Neural Trans. Suppl., 54: 97–106. Cardenas, V.A., Du, A.T., Hardin, D., Ezekiel, F., Weber, P., Jagust, W.J., Chui, H.C., Schuff, N. and Weiner, M.W. (2003) Comparison of methods for measuring longitudinal

brain change in cognitive impairment and dementia. Neurobiol. Aging, 24(4): 537–544. Convit, A., de Leon, M.J., Tarshish, C., DeSanti, S., Tsui, W., Rusinek, H. and George, A. (1997) Specific hippocampal volume reductions in individuals at risk for Alzheimer’s disease. Neurobiol. Aging, 18(2): 131–138. Dickerson, B.C., Goncharova, I., Sullivan, M.P., Forchetti, C., Wilson, R.S., Bennett, D.A., Beckett, L.A. and deToledoMorrell, L. (2001) MRI-derived entorhinal and hippocampal atrophy in incipient and very early Alzheimer’s disease. Neurobiol. Aging, 22(5): 747–754. Du, A.T., Schuff, N., Amend, D., Laakso, M.P., Hsu, Y.Y., Jagust, W.J., Yaffe, K., Kramer, J.H., Reed, B., Norman, D., Chui, H.C. and Weiner, M.W. (2001) MRI of entorhinal cortex and hippocampus in mild cognitive impairment and Alzheimer’s disease. J. Neurol. Neurosurg. Psychiatry, 71(4): 441–447. Du, A.T., Schuff, N., Kramer, J.H., Ganzer, S., Zhu, X.P., Jagust, W.J., Miller, B.L., Reed, B.R., Mungas, D., Yaffe, K., Chui, H.C. and Weiner, M.W. (2004) Higher atrophy rate of entorhinal cortex than hippocampus in AD. Neurology, 62(3): 422–427. Du, A.T., Schuff, N., Zhu, X.P., Jagust, W.J., Miller, B.L., Reed, B.R., Kramer, J.H., Mungas, D., Yaffe, K., Chui, H.C. and Weiner, M.W. (2003) Atrophy rates of entorhinal cortex in AD and normal aging. Neurology, 60(3): 481–486. Folstein, M.F., Folstein, S.E. and McHugh, P.R. (1975) ‘‘Minimental state.’’ A practical method for grading the mental status of patients for the clinician. J. Psychiatr. Res., 12(3): 189–192. Gomez-Isla, T., Price, J.L., McKeel, D.W., Morris, J.C., Growdon, J.H. and Hyman, B.T. (1996) Profound loss of layer. II. Entorhinal cortex neurons occurs in very mild Alzheimer’s disease. J. Neurosci., 16(14): 4491–4500. Good, C.D., Johnstrude, I.S., Ashburner, J., Henson, R.N., Friston, K.J. and Frakoviak, R.S. (2001) A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage, 14(1): 21–36. Goncharova, I.I., Dickerson, B.C., Stoub, T.R. and deToledoMorrell, L. (2001) MRI of entorhinal cortex: a reliable protocol for volumetric measurement. Neurobiol. Aging, 22(5): 737–745. Hyman, B.T., Van Hoesen, G.W., Damasio, A.R. and Barnes, C.L. (1984) Alzheimer’s disease: cell specific pathology isolates the hippocampal formation. Science, 225(4667): 1168–1170. Insausti, R., Juottonen, K., Soininen, H., Insausti, A.M., Partanen, K., Vainio, P., Laakso, M.P. and Pitkanen, A. (1998) MR volumetric analysis of the human entorhinal, perirhinal and temporopolar cortices. Am. J. Neuroradiol., 19(4): 659–671. Jack Jr., C.R., Petersen, R.C., Xu, Y., O’Brien, P.C., Smith, G.E., Ivnik, R.J., Boeve, B.F., Tangalos, E.G. and Kokmen, E. (2000) Rates of hippocampal atrophy correlate with change in clinical status in aging and AD. Neurology, 55(4): 484–489.

752 Jack Jr., C.R., Petersen, R.C., Xu, Y., O’Brien, P.C., Smith, G.E., Ivnik, R.J., Tangalos, E.G. and Kokmen, E. (1998) Rate of medial temporal lobe atrophy in typical aging and Alzheimer’s disease. Neurology, 51(4): 993–999. Jack Jr., C.R., Petersen, R.C., Xu, Y.C., O’Brien, P.C., Smith, G.E., Ivnik, R.J., Boeve, B.F., Waring, S.C., Tangalos, E.G. and Kokmen, E. (1999) Prediction of AD with MRI-based hippocampal volume in mild cognitive impairment. Neurology, 52(7): 1397–1403. Jack Jr., C.R., Petersen, R.C., Xu, Y.C., Waring, S.C., O’Brien, P.C., Tangalos, E.G., Smith, G.E., Ivnik, R.J. and Kokmen, E. (1997) Medial temporal atrophy on MRI in normal aging and very mild Alzheimer’s disease. Neurology, 49(3): 786–794. Jack Jr., C.R., Schlung, M.M., Gunter, J.L., O’Brien, P.C., Weigand, S.D., Knopman, D.S., Boeve, B.F., Ivnik, R.J., Smith, G.E., Cha, R.H., Tangalos, E.G. and Petersen, R.C. (2004) Comparison of different MRI brain atrophy measures with clinical disease progression in AD. Neurology, 62(4): 591–600. Jessen, F., Feyen, L., Freymann, K., Tepest, R., Maier, W., Heun, R., Schild, H.H. and Scheef, L. (2006) Volume reduction of the entorhinal cortex in subjective memory impairment. Neurobiol. Aging, 27(12): 1751–1756. Kalbfleish, J.D. and Prentice, R.L. (1980) The Statistical Analysis of Failure Time Data. Wiley, New York. Killiany, R.J., Gomez-Isla, T., Moss, M., Kikinis, R., Sandor, T., Jolesz, F., Tanzi, R., Jones, K., Hyman, B.T. and Albert, M.S. (2000) Use of structural magnetic resonance imaging to predict who will get Alzheimer’s disease. Ann. Neurol., 47(4): 430–439. Killiany, R.J., Hyman, B.T., Gomez-Isla, T., Moss, M.B., Kikinis, R., Jolesz, F., Tanzi, R., Jones, K. and Albert, M.S. (2002) MRI measures of entorhinal cortex vs. hippocampus in preclinical AD. Neurology, 58(8): 1188–1196. Kordower, J.H., Chu, Y., Stebbins, G.T., DeKosky, S.T., Cochran, E.J., Bennett, D. and Mufson, E.J. (2001) Loss and atrophy of layer. II. Entorhinal cortex neurons in elderly people with mild cognitive impairment. Ann. Neurol., 49(2): 202–213. Lancaster, J.L., Woldorff, M.G., Parsons, L.M., Liotti, M., Freitas, C.S., Rainey, L., Kochunov, P.V., Nickerson, D., Mikiten, S.A. and Fox, P.T. (2000) Automated Talairach atlas labels for functional brain mapping. Hum. Brain Mapp., 10(3): 120–131. de Leon, M.J., Convit, A., George, A.E., Golomb, J., DeSanti, S., Tarshish, C., Rusinek, H., Bobinski, M., Ince, C., Miller, D. and Wisniewski, H. (1996) In vivo structural studies of the hippocampus in normal aging and in incipient Alzheimer’s disease. Ann. N.Y. Acad. Sci., 777: 1–13. de Leon, M.J., Convit, A., Wolf, O.T., Tarshish, C.Y., DeSanti, S., Rusinek, H., Tsui, W., Kandil, E., Sherer, A.J., Roche, A., et al. (2001) Prediction of cognitive decline in normal elderly subjects with 2-[(18)F]fluoro-2-deoxy-D-glucose/positron emission tomography (FDG/PET). Proc. Natl. Acad. Sci. U.S.A., 98(19): 10966–10971. McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D. and Stadlan, E.M. (1984) Clinical diagnosis of

Alzheimer’s disease: report of the NINCDS-ADRDA work group under the auspices of department of health and human services task force on Alzheimer’s disease. Neurology, 34(7): 939–944. Morris, J.C., Heyman, A., Mohs, R.C., van Belle, G., Fillenbaum, G., Mellits, E.D. and Clark, C. (1989) The Consortium to Establish a Registry for Alzheimer’s Disease (CERAD). Part I. Clinical and neuropsychological assessment of Alzheimer’s disease. Neurology, 39(9): 1159–1165. Mueller, S.G., Stables, L., Du, A.T., Schuff, N., Truran, D., Cashdollar, N. and Weiner, M.W. (2007) Measurement of hippocampal subfields and age-related changes with high resolution MRI at 4T. Neurobiol. Aging, 28(5): 719–726. Mufson, E.J., Chin, E.Y., Cochran, E.J., Beckett, L.A., Bennett, D.A. and Kordower, J.H. (1999) Entorhinal cortex beta-amyloid load in individual with mild cognitive impairment. Exp. Neurol., 158(2): 469–490. Petersen, R.C. (2000) Aging, mild cognitive impairment and Alzheimer’s disease. Dementia, 18: 789–805. Petersen, R.C., Smith, G.E., Waring, S.C., Ivnik, R.J., Tangalos, E.G. and Kokmen, E. (1999) Mild cognitive impairment: clinical characterization and outcome. Arch. Neurol., 56(3): 303–308. Rodrigue, K.M. and Raz, N. (2004) Shrinkage of the entorhinal cortex over five years predicts memory performance in healthy adults. J. Neurosci., 24(4): 956–963. Saykin, A.J., Wishart, H.A., Rabin, L.A., Flashman, L.A., West, J.D., McHugh, T.L. and Mamourian, A.C. (2006) Older adults with cognitive complaints show brain atrophy similar to that of amnestic MCI. Neurology, 67(5): 834–842. Scheff, S.W. and Price, D.A. (2003) Synaptic pathology in Alzheimer’s disease: a review of ultrastructural studies. Neurobiol. Aging, 24(8): 1029–1046. Scheff, S.W., Price, D.A., Schmitt, F.A. and Mufson, E.J. (2006) Hippocampal synaptic loss in early Alzheimer’s disease and mild cognitive impairment. Neurobiol. Aging, 27(10): 1372–1384. Small, S.A., Chawla, M.K., Buonocore, M., Rapp, P.R. and Barnes, C.A. (2004) Imaging correlates of brain function in monkeys and rats isolates a hippocampal subregion differentially vulnerable to aging. Proc. Natl. Acad. Sci. U.S.A., 101(18): 7181–7186. Stoub, T.R., Bulgakova, M., Leurgans, S., Bennett, D.A., Fleischman, D., Turner, D.A. and deToledo-Morrell, L. (2005) MRI predictors of risk of Alzheimer’s disease: a longitudinal study. Neurology, 64(9): 1520–1524. Stoub, T.R., deToledo-Morrell, L., Stebbins, G.T., Leurgans, S., Bennett, D.A. and Shah, R.C. (2006) Hippocampal disconnection contributes to memory dysfunction in individuals at risk for Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A., 103(26): 10041–10045. Squire, L.R. and Zola-Morgan, S. (1991) The medial temporal lobe memory system. Science, 253(5026): 1380–1386. Talairach, J. and Turnoux, P. (1988) Co-Planar Stereotaxic Atlas of the Human Brain. Thieme, New York. Tapiola, T., Pennanen, C., Tapiola, M., Tervo, S., Kivipelto, M., Ha¨nninen, T. and Pihlajama¨ki, M., et al. (2006) MRI of

753 hippocampus and entorhinal cortex in mild cognitive impairment: a follow-up study. Neurobiol. Aging, (in press). deToledo-Morrell, L., Dickerson, B., Sullivan, M.P., Spanovic, C., Wilson, R.S. and Bennett, D. (2000a) Hemispheric differences in hippocampal volume predict verbal and spatial memory performance in patients with Alzheimer’s disease. Hippocampus, 10(2): 136–142. deToledo-Morrell, L., Goncharova, I., Dickerson, B., Wilson, R.S. and Bennett, D.A. (2000b) From healthy aging to Alzheimer’s disease: in vivo detection of entorhinal cortex atrophy. Ann. N.Y. Acad. Sci., 911: 240–253. deToledo-Morrell, L., Sullivan, M.P., Morrell, F., Wilson, R.S., Bennett, D.A. and Spencer, S. (1997) Alzheimer’s disease: in vivo detection of differential vulnerability of brain regions. Neurobiol. Aging, 18(5): 463–468. deToledo-Morrell, L., Stoub, T.R., Bulgakova, M., Wilson, R.S., Bennett, D.A., Leurgans, S., Wuu, J. and Turner, D.A. (2004) MRI-derived entorhinal volume is a good predictor of conversion from MCI to AD. Neurobiol. Aging, 25(9): 1197–1203. Van Hoesen, G.W. and Hyman, B.T. (1990) Hippocampal formation: anatomy and the patterns of pathology in Alzheimer’s disease. Prog. Brain Res., 83: 445–457. Van Hoesen, G.W. and Pandya, D.N. (1975) Some connections of the entorhinal (area 28) and perirhinal (area 35) cortices of the rhesus monkey. III. Efferent connections. Brain Res., 95(1): 39–59. Van Hoesen, G.W., Pandya, D.N. and Butters, N. (1975) Some connections of the entorhinal (area 28) and perirhinal

(area 35) cortices of the rhesus monkey. II. Frontal lobe afferents. Brain Res., 95(1): 25–38. Wang, C., Medina, D., Stebbins, G., Shah, R., Bammer, R., Moseley, M. and deToledo-Morrell, L. (2006) Changes in the microstructural integrity of parahippocampal white matter tracts in Alzheimer’s disease. Soc. Neurosci. Abstr., on line. Wechsler, D. (1987) Wechsler Memory Scale: Revised Manual. Psychological Corporation, San Antonio, TX. West, M.J., Coleman, P.D., Flood, D.G. and Troncoso, J.C. (1994) Differences in the pattern of hippocampal neuronal loss in normal aging and Alzheimer’s disease. Lancet, 344(8925): 769–772. West, M.J., Kawas, C.H., Martin, L.J. and Troncoso, J.C. (2000) The CA1 region of the human hippocampus is a hot spot in Alzheimer’s disease. Ann. N.Y. Acad. Sci., 908: 255–259. Wilson, R.S., Sullivan, M.P., deToledo-Morrell, L., Stebbins, G.T., Bennett, D.A. and Morrell, F. (1996) Association of memory and cognition in Alzheimer’s disease with volumetric estimates of temporal lobe structures. Neuropsychology, 10(4): 459–463. Xu, Y., Jack, C.R., O’Brien, P.C., Kokmen, E., Smith, G.E., Ivnik, R.J., Boeve, B.F., Tangalos, R.G. and Petersen, R.C. (2000) Usefulness of MRI measures of entorhinal cortex versus hippocampus in AD. Neurology, 54(9): 1760–1767. Young, B., Otto, T., Fox, G.D. and Eichenbaum, H. (1997) Memory representation within the parahippocampal region. J. Neurosci., 17(13): 5183–5195.

This page intentionally left blank

H.E. Scharfman (Ed.) Progress in Brain Research, Vol. 163 ISSN 0079-6123 Copyright r 2007 Elsevier B.V. All rights reserved

CHAPTER 41

Epileptogenesis in the dentate gyrus: a critical perspective F. Edward Dudek1, and Thomas P. Sutula2 1

Department of Physiology, University of Utah School of Medicine, 420 Chipeta Way, Suite 1700, Salt Lake City, UT 84108, USA 2 Department of Neurology, University of Wisconsin, Madison, WI, USA

Abstract: The dentate gyrus has long been a focal point for studies on the molecular, cellular, and network mechanisms responsible for epileptogenesis in temporal lobe epilepsy (TLE). Although several hypothetical mechanisms are considered in this chapter, two that have garnered particular interest and experimental support are: (1) the selective loss of vulnerable interneurons in the region of the hilus and (2) the formation of new recurrent excitatory circuits after mossy fiber sprouting. Histopathological data show that specific GABAergic interneurons in the hilus are lost in animal models of TLE, and several lines of electrophysiological evidence, including intracellular analyses of postsynaptic currents, support this hypothesis. In particular, whole-cell recordings have demonstrated a reduction in the frequency of miniature inhibitory postsynaptic currents in the dentate gyrus and other areas (e.g., CA1 pyramidal cells), which provides relatively specific evidence for a reduction in GABAergic input to granule cells. These studies support the viewpoint that modest alterations in GABAergic inhibition can have significant functional impact in the dentate gyrus, and suggest that dynamic activity-dependent mechanisms of GABAergic regulation add complexity to this local synaptic circuitry and to analyses of epileptogenesis. In regard to mossy fiber sprouting, a wide variety of experiments involving intracellular or whole-cell recordings during electrical stimulation of the hilus, glutamate microstimulation, and dual recordings from granule cells support the hypothesis that mossy fiber sprouting forms new recurrent excitatory circuits in the dentate gyrus in animal models of TLE. Similar to previous studies on recurrent excitation in the CA3 area, GABA-mediated inhibition and the intrinsic high threshold of granule cells in the dentate gyrus tends to mask the presence of the new recurrent excitatory circuits and reduce the likelihood that reorganized circuits will generate seizure-like activity. How cellular alterations such as neuron loss in the hilus and mossy fiber sprouting influence functional properties is potentially important for understanding fundamental aspects of epileptogenesis, such as the consequences of primary initial injuries, mechanisms underlying network synchronization, and progression of intractability. The continuous nature of the axonal sprouting and formation of recurrent excitation could account for aspects of the latent period and the progressive nature of the epileptogenesis. Future studies will need to identify

Corresponding author. Tel.: +1 (801)587-5880; Fax: +1 (801)581-8075; E-mail: [email protected] DOI: 10.1016/S0079-6123(07)63041-6

755

756

precisely how these hypothetical mechanisms and others contribute to the process whereby epileptic seizures are initiated or propagated through an area such as the dentate gyrus. Finally, in addition to its unique features and potential importance in epileptogenesis, the dentate gyrus may also serve as a model for other cortical structures in acquired epilepsy. Keywords: epilepsy; seizure; granule cell; hilus; interneuron; axon sprouting

Introduction

A definition of epileptogenesis

Overview

Epileptogenesis is broadly considered to be the collective molecular-, cellular-, and systems-level mechanisms whereby spontaneous recurrent seizures develop after a brain insult in a manner that is time-dependent. For purposes of this review, we will consider epileptogenesis in a more specific and restrictive sense (Stables et al., 2003), rather than the loose and more general meaning often associated with the term. The more restrictive definition of epileptogenesis does not include: (1) acute seizures associated with brain injury, (2) pharmacological treatments in animals or isolated tissue that induce seizure-like activity, or (3) genetic abnormalities that are associated with increased seizure susceptibility. While those processes are of interest for understanding the acute mechanisms of network synchronization and are sometimes included in the broader less restrictive definition of epileptogenesis, we believe for purposes of this critical analysis that focusing on the more restrictive definition has advantages for understanding how chronic epilepsy develops and progresses. We believe the term epileptogenesis also applies to the progressive worsening of the epilepsy — in terms of the increased frequency and/or severity of seizures — after the first of the spontaneous recurrent seizures. A critical debate in the field is whether acquired epilepsy (versus genetic epilepsy) requires actual brain damage and neuronal loss; some researchers believe, particularly in regard to acquired pediatric epilepsy (e.g., after hypoxia, prolonged febrile seizures, or status epilepticus (SE) in children and immature animals), that frank injury with neuronal death does not need to be present for the development of chronic epilepsy (i.e., epileptogenesis) to occur. This issue has been controversial, particularly in research on the dentate gyrus. Many studies have regarded increased

For decades, the dentate gyrus has attracted the attention of epilepsy researchers, and this interest has led to many publications that focus on this particular brain structure. The high level of interest stems from the development of several concepts that have been derived from traditional brain science and from both basic and clinical epilepsy research. The goal of this chapter is to consider some of the key hypotheses, concepts and controversies in the field of epilepsy that are centered on the role of the dentate gyrus in ‘‘epileptogenesis’’ and on how different possible mechanisms in the dentate gyrus may be altered in a manner that might lead to temporal lobe epilepsy (TLE). Epileptogenesis is broadly defined as the biological process responsible for the development of spontaneous recurrent seizures after an acquired brain insult or a primary abnormality such as a mutation or developmental defect. Even focusing on epileptogensis alone, the field is too extensive and complicated to review in this brief chapter. Thus, our aim is to target a few basic concepts in epilepsy research, and provide commentary in those areas where we believe there are particular problems in the field and/or where a different or modified perspective would be useful. We will provide only minimal discussion of neurogenesis and GABA-A receptors as they relate to epileptogenesis in the dentate gyrus because these topics are amply covered by Parent (this volume) and Coulter and Carlson (this volume), respectively. Similarly, a detailed discussion of the electrophysiological studies of the human dentate gyrus in relation to epileptogenesis is provided by Williamson and Patrylo (this volume).

757

seizure susceptibility as a surrogate marker for epileptogenesis, and so another critical issue involves the question: how good are the data that an animal shows epileptogenesis (i.e., becomes ‘‘epileptic’’ with spontaneous recurrent seizures)?

Concepts of epileptogenesis pertinent to the dentate gyrus Several questions about epileptogenesis in the dentate gyrus have captured the attention of the epilepsy research community during the last decade; these concepts will serve as targets for discussion in terms of interpretation of the presently available literature and suggestions for future research. The hippocampus has long been seen as a critical structure in TLE, particularly as an epileptic focus (i.e., presence of interictal spikes), as an epileptogenic zone (site of seizure initiation), and thus, as a primary target for resections aimed at surgical treatment of intractable TLE (see Engel, 1989 for review of the earlier literature). This emphasis on the hippocampus in TLE, and by extrapolation of the dentate gyrus, however, has been challenged from many perspectives (e.g., see Gloor, 1992; Bertram, 1997; Bertram et al., 1998). Because the dentate gyrus has traditionally been viewed as the first stage or ‘‘gate’’ into the classical trisynaptic circuit of the hippocampus, a hippocampo-centric view has led to a dentate-centric focus in the minds of many researchers. This is not to dismiss the concept that the hippocampus and dentate gyrus are likely important in TLE; rather, it is to acknowledge that other areas may be as important (or even more important), and that the mechanisms that occur during epileptogenesis in the dentate gyrus may also be present in other brain regions. The observation of ‘‘maximal dentate activation’’ strengthened the dentate ‘‘gate’’ concept in epilepsy research, and provided a hypothetical electrophysiological basis for how the dentate could serve a ‘‘closed gate’’ versus an ‘‘open gate’’ function (Heinemann et al., 1992; Lothman et al., 1992; Stringer and Lothman, 1992). Another basis for interest in the dentate gyrus by epilepsy researchers has been the hypothesis

(e.g., Sloviter, 1991; Sloviter et al., 2003) that the loss of hilar neurons (or ‘‘endfolium sclerosis’’ — considered to be a subset of the more general histopathological condition of mesial temporal sclerosis; Margerison and Corsellis, 1966) is the central feature or minimum essential substrate of TLE. Several studies on animal models of traumatic brain injury (e.g., Lowenstein et al., 1992; Toth et al., 1997) and ischemia (e.g., Crain et al., 1988; Hsu and Buzsaki, 1993; Williams et al., 2004), which may lead to epilepsy, have found that the hilus is particularly prone to neuronal injury; therefore, interest has focused on the vulnerability of the different types of hilar neurons, and how this might alter the electrophysiological properties of the dentate network. At least two separate concepts or hypotheses have emerged concerning the consequences of hilar neuron loss and the possible development of increased network excitability: (1) a reduction in the number of interneurons (e.g., Obenaus et al., 1993; Houser and Esclapez, 1996) and/or (2) a loss of excitatory input to interneurons (i.e., the ‘‘dormant basket cell hypothesis’’; e.g., Sloviter, 1987; Sloviter et al., 2003). In a separate but related line of research, at least four groups (de Lanerolle et al., 1989; Sutula et al., 1989; Houser et al., 1990; Babb et al., 1991) found that the presence of mesial temporal sclerosis in specimens from surgical resections for intractable TLE was correlated with ‘‘mossy fiber sprouting,’’ where the Timm stain method densely stains the inner molecular layer, which is not present in normal tissue (e.g., Tauck and Nadler, 1985; see Sutula and Dudek, this volume). Although the dentate gyrus is strategically located, and both hilar neuron loss and mossy fiber sprouting are common features of TLE, the dentate gyrus and its granule cells are clearly not required for epileptogenesis and both hilar neuron loss and mossy fiber sprouting in the dentate gyrus do not appear to be necessary or sufficient for spontaneous recurrent seizures to occur. For reasons that we will discuss and that are also developed in another chapter (Sutula and Dudek, this volume), these observations do not diminish the potential importance of understanding how the dentate gyrus contributes to hippocampal epileptogenesis and the syndromes

758

of TLE. In the sections that follow, we will consider how some of the features of the organization and electrophysiology of the dentate gyrus are specifically pertinent to the role of the dentate gyrus in epileptogenesis.

Epileptogenesis in the dentate gyrus: the laminar organization of the hippocampus leads to large field potentials and thus an increased ability to detect epileptiform events While the dentate gyrus and hippocampus are commonly involved in TLE, which is the most common type of intractable epilepsy, the high interest and potential importance of this region of limbic circuitry derives from the frequent observation of interictal discharges and apparent onset of seizures during intrahippocampal depth recordings both clinically and in animal models of TLE, in addition to the characteristic histopathological (and clinical imaging) abnormalities (see Engel, 1989 for classical overview). An important point is that because of the highly organized and laminar structure of the hippocampus, including the dentate gyrus, synchronous activity in any particular part of the hippocampus generates especially large field potentials in the extracellular space. Therefore, compared to other structures with less packed and layered cell bodies, such as the amygdala and entorhinal cortex, epileptiform events (including interictal spikes and seizures) would be expected to be much more readily observed in the hippocampus and dentate gyrus than other parts of the cerebral cortex. Thus, a seizure may start in a nearby structure (e.g., amygdala or entorhinal cortex) with less lamination that generates a smaller and more-difficult-to-detect field potential, and then spread to the dentate gyrus and other hippocampal areas, where the relatively largeamplitude field potentials of the seizure might be first detected. For example, considerable interest has focused on the occurrence of ‘‘fast ripples,’’ which is vernacular that refers to high-frequency electrographic oscillations (i.e., approximately 200–500 Hz), as markers of epileptic seizure onset in the dentate gyrus and elsewhere in the hippocampus and parahippocampal areas. Based on

their time course, these events are almost certainly synchronous action potentials (i.e., small population spikes) that are likely to be much larger in the dentate gyrus and hippocampus than other seizure-prone areas because of their laminar structure (Bragin et al., 2002). As a result of the laminar organization, it is and will continue to be difficult to assess where electrographic seizures actually start, and whether an area in the hippocampus, or in another structure, is the actual initiation site. Nonetheless, the ability to analyze different components of the field potentials in the dentate gyrus (and hippocampus proper) in response to electrical stimulation of synaptic inputs and during spontaneous seizures provides important potential avenues for research.

Epileptogenesis in the dentate gyrus: the normal dentate gyrus has a high threshold for seizure generation, so it may not be particularly prone to chronic epileptogenesis It is widely believed that the most important epileptogenic structures are particularly susceptible to seizure generation in the normal brain, and these structures are even more prone to ‘‘hyperexcitability’’ in the epileptic brain, where the term hyperexcitability describes a general increase in response to a particular stimulus or enhanced tendency to generate repetitive synchronous neuronal discharges manifesting as a burst of population spikes. If this is the case, one could argue that the dentate gyrus — as studied in the normal brain — should be relatively low on the list of candidate structures for epileptogenesis. Although the dentate gyrus is a unique structure with a high degree of neuronal homogeneity among the granule cells and heterogeneity in the hilus, it has long been known that the dentate granule cells are highly resistant to epileptiform activity after treatment with convulsive agents, such as GABA-A receptor antagonists (e.g., Fricke and Prince, 1984). In the normal brain, the dentate gyrus appears to be the least susceptible part of the hippocampus to generate either brief interictal-like bursts or prolonged seizure-like activity, at least when compared to the CA1 and CA3 areas. This

759

resistance to epileptiform activity is widely believed to be due to strong GABAergic inhibition, but the relatively negative resting membrane potential, high threshold for action potential generation, and high degree of spike accommodation to maintained depolarization (e.g., see Staley et al., 1992) are probably equally as important reasons why the dentate granule cells are relatively seizure-resistant under normal conditions. Many studies in epilepsy research have not considered these fundamental properties of the normal dentate gyrus; and on these grounds, the dentate gyrus may be an unlikely area for seizure onset during epileptogenesis (i.e., be an epileptogenic zone), because seizure threshold of the dentate gyrus often appears high compared to other structures even when tested in chronically epileptic animals. The transformation of the synaptic interactions in the dentate during epileptogenesis could, however, be more complex and important than is apparent from the previous discussion. Indeed, if properties such as a more negative resting membrane potential, higher action potential threshold, lower spontaneous firing rate, and resistance to repetitive firing endow the dentate gyrus with resistance to synchronous network discharge and epileptogenesis, then processes such as hilar neuronal loss and formation of recurrent excitatory circuits by mossy fiber sprouting could be especially important in the transformation of the dentate gyrus into a more epileptogenic structure.

Models of epileptogenesis in the dentate gyrus: evidence for filtering and gating properties in the dentate gyrus The perforant path projection from the entorhinal cortex to the dentate gyrus is the first synapse of the classic tri-synaptic hippocampal circuit, and this synapse to the granule cells is generally believed to be relatively resistant to transmission of activity into CA3 (i.e., the dentate can be considered to be a gate that is usually closed; see Hsu, this volume). The ‘‘gate’’ property of the dentate gyrus is often thought to impede the propagation of normal electrical activity and seizures into hippocampus, and when this gating property

fails, the dentate may allow propagation of activity into other structures that are projection targets from the hippocampus (Heinemann et al., 1992; Lothman et al., 1992; Stringer and Lothman, 1992). There is evidence that the ‘‘gate’’ function of the dentate gyrus may operate in an ‘‘all or none’’ fashion, as implied by the observation of ‘‘maximal dentate activation’’ associated with propagation of epileptiform activity. One basis for the ‘‘gate’’ concept is the relatively high threshold for excitation of the dentate, and according to this perspective, reduction in the ‘‘gate’’ function of the dentate gyrus could be an epileptogenic mechanism that would promote increased excitability (i.e., hyperexcitability) to perforant path stimuli, transforming synchronous excitatory postsynaptic potentials (EPSPs) with superimposed action potentials into a large burst of action potentials. This mechanism could apply to synaptic inputs that include interictal spikes and actual seizures; thus, it has been considered that the ‘‘gate’’ to the hippocampus normally restricts or blocks epileptiform activity. While this ‘‘gating’’ property may restrict epileptiform activity, the parallel pathway from the entorhinal cortex projecting directly to the CA3 and CA1 areas may still transmit discharges into the hippocampus regardless of the properties of the dentate gyrus. Other studies suggest that the dentate gyrus can serve as a filter whereby activity is blocked from entering the dentate under some conditions but not others (e.g., see Iijima et al., 1996; Behr et al., 1998; versus Ang et al., 2006). Although the dentate gyrus may not be necessary for transmission of signals into the hippocampus, it is not fully understood how the dentate gyrus processes electrical information in relation to normal hippocampal integration (see Kesner, this volume) or how the synaptically reorganized dentate gyrus affects the transmission of seizures into the hippocampus in animal models of epilepsy or patients with TLE (see Sutula and Dudek, this volume). Thus, the dentate gyrus presumably plays an important role in epileptogenesis, even though the entorhinal cortex appears to project ‘‘downstream’’ to the CA3 and CA1 areas of the hippocampus. The concept that the dentate simply functions as a ‘‘gate’’ that normally blocks seizure activity in the entorhinal

760

cortex from propagating into the hippocampus may, however, be an oversimplification and require a more sophisticated assessment in the complex network of serial and parallel pathways projecting into the hippocampus and to parahippocampal structures, such as the subiculum, which has also been implicated in TLE (Cohen et al., 1998).

potentially important functions regarding spontaneous interictal spikes and seizures versus electrically stimulated events.

Animal models of epileptogenesis during TLE: alterations of the dentate gyrus Models and hypothetical mechanisms of epileptogenesis

Epileptogenesis in the dentate gyrus: repeated activation of the dentate gyrus can promote propagation of seizures into the hippocampus The high seizure threshold of the normal dentate gyrus becomes dramatically reduced after it has experienced a few electrically induced afterdischarges (i.e., maximal dentate activation), and so the dentate appears to be highly sensitive to previous electrical activity (Heinemann et al., 1992; Lothman et al., 1992; Stringer and Lothman, 1992). Even a single afterdischarge in the dentate gyrus increases synaptic transmission for periods of as long as 3 months, and induction of long-term potentiation increases susceptibility to evoked network activity and afterdischarges (Sutula and Steward, 1986, 1987; Sayin et al., 1999). These forms of network plasticity occur over a relatively short time frame and are therefore unlikely to account for chronic epileptogenesis, but relatively long-lasting alterations could promote network synchronization in the dentate gyrus and hippocampal pathways. The potential role of the dentate in modulating, reducing, or filtering some forms of electrical activity and possibly single seizures would be degraded by activity-dependent enhancement of synaptic transmission in granule cells, by allowing more propagation of clusters of seizures into the hippocampus proper. Therefore, although some forms of entorhinal cortical activity can normally bypass the dentate gyrus during propagation into the hippocampus, it remains possible that the altered ability of repetitive seizures or seizure clusters to spread into the hippocampus after maximal dentate activation or kindling may be a critical feature of epileptogenesis during TLE. The present approaches with hippocampal slices, or even intact anaesthetized preparations, might fail to identify

Animal models simulating histopathological features of human TLE in the dentate gyrus have been used for the study of chronic epileptogenesis. The two most widely used models of epileptogenesis in adult animals, SE and kindling, have been used to examine processes of epileptogeneis following an initial injury (e.g., SE) and repeated brief seizures (i.e., kindling). Both of these induction scenarios mimic common but distinctive clinical features of human TLE, namely syndromes that follow an initial precipitating injury and cryptogenic cases that gradually progress to intractability. These models have provided experimental opportunities to assess two general types of hypothetical mechanisms derived from histopathological alterations in the dentate gyrus commonly associated with human TLE that would change the balance of excitation and inhibition, specifically reducing GABAergic inhibition and/or increasing glutamatergic excitation. The hypothetical reduction in inhibition has been proposed to be due to a decrease in inhibitory interneurons and/or a reduction of excitatory synaptic input to inhibitory interneurons (i.e., the ‘‘dormant basket cell’’ hypothesis). Both of these effects could occur as a result of the histopathological observation of a loss of hilar neurons. The hypothesis of enhanced excitation in the dentate gyrus has focused on new recurrent excitatory circuits associated with mossy fiber sprouting, as detected by increased Timm stain product in the inner molecular layer of the dentate gyrus. Although there are many differences between the kindling and SE models, both show some of the histopathological alterations in the dentate gyrus associated with TLE, and have at least some utility to address these hypothetical mechanisms of epileptogenesis.

761

Models of epileptogenesis based on kindling Kindling is a progressive decrease in the threshold for induction of afterdischarges to daily electrical stimulations of the amygdala or other limbic structures. This animal model progressively develops a chronic irreversible state where low-intensity stimuli trigger prolonged afterdischarges and seizures. Kindling initially induces low-level apoptosis in the hilus (Bengzon et al., 1997) and cumulatively results in progressive loss of neurons not only in the hilus, but in CA3 and CA1 in a pattern resembling human hippocampal sclerosis (Cavazos et al., 1994). Kindling also induces mossy fiber sprouting, and the Timm stain in the molecular layer progressively increases over many months with prolonged kindling (Cavazos et al., 1991), although the density of staining in the inner molecular layer is generally less than is typically seen in SE models where epileptogenesis occurs over several weeks and months. Models of epileptogenesis based on SE The two main approaches that have been invoked to induce SE and subsequent epileptogenesis involve injection of chemotoxins or electrical stimulation. Kainic acid (i.e., kainate; Ben-Ari, 1985; Nadler, 1991) and pilocarpine (Turski et al., 1983; with or without lithium pretreatment) are the most commonly used chemotoxins, but repetitive electrical stimulation of certain limbic structures can also induce SE (e.g., Lothman et al., 1990). The repetitive seizures during SE cause neuronal loss in the hilus of the dentate gyrus and in the CA3 and CA1 areas of the hippocampus, plus other brain areas. The neuronal death in the hilus is thought to contribute to mossy fiber sprouting (see Sutula and Dudek, this volume). Specific neuronal injury in the hilus and hippocampus versus more extensive brain damage associated with human TLE The extrahippocampal damage associated with the models based on SE is often severe and extensive, and thus it has been argued that the SE models do

not recapitulate the key histopathological features of TLE (Sloviter, 1991). Based on an interpretation of Margerison and Corsellis (1966) (see also Meldrum and Bruton, 1992), some workers have argued that the SE models have too little damage in areas that are completely damaged in TLE (e.g., CA1 and hilus), but too much damage in extrahippocampal areas. Although the hilus was seen to be the most common area for severe neuronal loss (about 65% of patients) in a histopathological study on autopsy tissue from patients institutionalized due to epilepsy and other potential disorders, Margerison and Corsellis (1966) also emphasized that extrahippocampal damage is commonly observed in many patients with classic features of mesial temporal scerosis. For example, Margerison and Corsellis (1966) state that their study of the whole brain ‘‘has emphasized the need to see beyond the temporal lobes and to take into account the possible importance of damage in other parts of the brain as well.’’ More recent imaging studies have emphasized the more widespread injury to areas that communicate with the hippocampus, such as the amygdala and the entorhinal and perirhinal cortex (see Cascino, 2005), although other brain areas are damaged in some patients (Margerison and Corsellis, 1966). Thus, the hippocampus in general and the dentate gyrus in particular are brain regions of considerable interest in relation to human TLE, but other areas are also involved. Furthermore, it is not clear that the SE models are actually such a poor reflection of TLE, as some would suggest.

Histopathological alterations in the dentate gyrus of animal models of epileptogenesis The models of TLE based on SE often have more damage to other brain areas than those TLE patients with rather focal hippocampal and parahippocampal lesions, but the SE models show many histopathological similarities to human TLE in the dentate gyrus. The SE models typically show extensive histopathological damage to the hilus of the dentate gyrus (Fig. 1) and to the CA3 and/or CA1 areas of the hippocampus that is roughly similar to human TLE. These models have been employed

762

Fig. 1. Nissl staining to show hilar neurons in a control rat versus a rat with kainate-induced epilepsy. (A and B) Sections in the septal hippocampus show some loss of hilar neurons in a kainate-treated rat (B) compared to control rat (A). (C and D) Near the temporal end of the hippocampus, the loss of hilar neurons in the kainate-treated rat was severe (D). m: molecular layer; g: granule layer; h: hilus; CA3: CA3 pyramidal cell layer; bar ¼ 500 mm. Adapted with permission from Buckmaster and Dudek (1997) Fig. 1, p. 388.

extensively, because they are relatively easy to use (although mortality can be a problem), and they show histopathological alterations in the hilus of the dentate gyrus that broadly reflect some of the features of human TLE. In the SE models, spontaneous seizures develop with an apparent latent period and progressively increase in seizure severity and frequency. The SE models provide a contrast to kindling, where neuronal injury is more gradually induced in the hilus and other regions, mossy fiber sprouting gradually progresses as a function of the number of evoked seizures, and spontaneous seizures develop more slowly, typically after 90–100 seizures evoked by daily stimulation (Sayin et al., 2003). When spontaneous seizures are present in more extensively kindled animals, mossy fiber sprouting is more prominent and interneurons are lost in the hilus (Sayin et al., 2003). Thus, the kindling and SE models recapitulate certain features of the histopathological changes and seizure susceptibility characteristic of TLE, but with rates of induction and progression that distinctively mimic

clinical features of some TLE syndromes, including onset with a delay after initial injury and progression to intractability. It should be emphasized that not all cases of human TLE develop after an obvious initial injury and not all cases progress to intractability, so neither the SE models nor kindling are universally applicable to the broad range of syndromes of human TLE. It should be further emphasized that most histopathological studies on humans with TLE are based on patients who have suffered from intractable epilepsy for many years or even decades, while research on animals (e.g., the SE models) does not usually involve periods greater than a few weeks or months. Damage to the hilus is a common characteristic of TLE The dormant basket cell hypothesis Many studies over the last two decades have aimed to address the issue of which types of

763

neurons (see Houser, this volume) are lost in the hilus in human tissue samples and in animal models of TLE (particularly the SE models). Although certain vulnerable GABAergic interneurons are lost, many other GABAergic neurons remain intact after a repetitive stimulation protocol to the perforant path that causes ‘‘hyperexcitability’’ of dentate field potentials (Sloviter, 1987, 1991). Because many inhibitory basket cells were still present within the dentate gyrus in human TLE and in animal models, Sloviter (1987, 1991) proposed the ‘‘dormant basket cell’’ hypothesis, which essentially states that TLE is associated with loss of excitatory mossy cells that normally project to GABAergic basket cell interneurons; therefore; the basket cells in the hilus of TLE patients and animal models of TLE have lost excitatory synaptic input (i.e., become ‘‘dormant’’), thus leading to the hyperexcitability. This hypothesis, however, is based on the relatively acute effects of repetitive extracellular stimulation (i.e., after hours and a few days), when chronic epileptic seizures either do not occur or are quite rare. Furthermore, nearly all of the electrophysiological data presented in support of this hypothesis involve in vivo pairedpulse experiments with extracellular stimulation and recording techniques using a range of interpulse intervals and repetitive stimulation frequencies (e.g., Sloviter, 1987; Sloviter et al., 2003) that are difficult to interpret and are generated by complicated physiological processes at many other cellular sites and levels (e.g., the perforant path-to-granule cell synapse; GABA-A receptors, chloride ion homeostasis, etc). A reduction of afferent input to GABAergic neurons has been reported in the CA1 area (Bekenstein and Lothman, 1993), but this study did not directly assess excitation to interneurons and alterations of a variety of physiological processes are likely to contribute to a net defect in inhibitory mechanisms. For example, Doherty and Dingledine (2001) found a reduced excitatory drive onto interneurons in the dentate gyrus after SE, and this effect was due to a use-dependent mechanism involving group II metabotropic glutamate receptors. Several other studies have addressed the dormant basket cell hypothesis. Based on a series of experiments using hippocampal slices from animals with

chronic epilepsy (i.e., SE models), Bernard et al. (1998) (see also Esclapez et al., 1997) have provided several lines of evidence that basket cells are not ‘‘dormant’’ in TLE. In addition, studies on miniature and spontaneous inhibitory postsynaptic currents (mIPSCs and sIPSCs) of dentate granule cells from the kainate model provide evidence that interneurons are spontaneously active even when iontopic gluatamatergic receptors are blocked. When in vitro electrophysiological experiments were conducted on hippocampal slices after the pharmacological blockade of ionotropic glutamatergic receptors to isolate mIPSCs and sIPSCs from glutamatergic synaptic circuits (Shao and Dudek, 2005), the frequency of sIPSCs in dentate granule cells at o1 week and >3 months (i.e., from chronically epileptic rats) after kainateinduced SE was not significantly different from controls, and was much higher than when the slices were bathed in tetrodotoxin to block sodium-mediated action potentials and isolate mIPSCs. The observation that blocking action potentials with tetrodotoxin greatly reduced the frequency of IPSCs (i.e., the frequency of sIPSCs was much higher than mIPSCs) in all groups indicates that the interneurons that project to granule cells in kainate-treated rats (and controls) are clearly not ‘‘dormant’’ (i.e., they are spontaneously active) either o1 week or >3 months (i.e., from chronically epileptic rats) after kainate-induced SE, even when recorded in isolated slices where GABAergic inhibitory circuits are intact and bathed in pharmacological agents that block glutamatergic EPSCs. Although this approach relies on mIPSCs and sIPSCs in granule cells as a measure of the upstream GABAergic inhibitory network, it is not affected by potential differences in the responses to extracellular stimulation, nor is it influenced by simultaneous alterations in glutamatergic synapses or circuits. This analysis is likely also somewhat of a simplification, since it has focused on somatic inhibition (i.e., the primary target of basket cells), but other studies with combined dendritic and somatic recording in the CA1 area of slices from pilocarpine-treated rats have provided evidence for differential effects of epileptogenesis on dendritic and somatic inhibition (Cossart et al., 2001). It remains possible, however,

764

that basket cells and some other remaining interneurons may experience a significant reduction of excitatory synaptic input due to the loss of input from glutamatergic neurons in the hilus and elsewhere, and this loss of excitatory input could importantly affect the ‘‘balance’’ of synaptic excitation and inhibition. Experiments with paired-pulse stimulation, particularly in intact animals, lack the mechanistic resolution to provide evidence supporting the dormant basket cell hypothesis. Given the numerous studies at the cellular and synaptic levels indicating that surviving interneurons have significant spontaneous and evoked activity in SE models, and the interpretive ambiguity and limitations of paired-pulse measurements using a range of interpulse intervals and repetitive stimulation, overall the evidence for dormancy of basket cells as a mechanism of epileptogenesis is weak and indirect. Quantitative electrophysiological studies with whole-cell recordings of EPSCs to assess changes in the excitatory synaptic inputs to the different types of GABAergic interneurons during epileptogenesis, preferably using several different animal models of TLE (i.e., including but not limited to kindling and SE models), could be useful to further evaluate this hypothesis.

Loss of interneurons Whereas the dormant basket cell hypothesis has not garnered support from recent electrophysiological experiments using modern recording techniques, data with a wide range of morphological and physiological techniques from many laboratories in human tissue and animal models directly support the hypothesis that GABAergic inhibition undergoes alterations in association with modest loss of specific interneurons in TLE. Early studies in a variety of hippocampal and cortical areas using various immunohistochemical and in situ hybridization techniques have shown that some GABAergic interneurons (i.e., a relatively small number, and only specific types) are lost in TLE (Fig. 2; e.g., Sloviter, 1987; Obenaus et al., 1993; Buckmaster and Dudek, 1997; Gorter et al., 2001). Immunohistochemical data and silver stains after

forebrain ischemia have shown that selectively vulnerable neurons, including cells that are likely GABAergic interneurons, are damaged. Similar studies, focusing on immunohistochemical techniques, identified several different types of GABAergic interneurons that were damaged after traumatic brain injury, and many of the injured interneuron types corresponded to those lost in TLE. These histopathological alterations in interneuron populations may contribute to the hyperexcitability often seen shortly after experimental SE and other brain insults (e.g., Lowenstein et al., 1992). Several other mechanisms, however, can hypothetically be responsible for or contribute to the hyperactive responses to extracellular stimulation within days after experimental SE, including alterations arising from direct SE-induced changes in chloride distribution and other GABA-A receptor-mediated mechanisms (e.g., Kapur et al., 1999). Several laboratories have used high-resolution whole-cell recordings of mIPSCs to test more directly the hypothesis of a loss of interneuron input to granule cells (Kobayashi and Buckmaster, 2003; Shao and Dudek, 2005) and CA1 pyramidal cells (Wierenga and Wadman, 1999), and they have found that the frequency of mIPSCs is reduced in these models, which is consistent with a loss of inhibitory GABAergic terminals (i.e., death of GABAergic interneurons) in TLE. Similar data showing a reduction in the frequency of mIPSCs have been reported after lateral fluid percussion injury (Toth et al., 1997). An important advantage of this approach is that it is independent of the complexities and uncertainties of extracellular stimulation within a multisynaptic circuit. Because all of the experimental recordings were conducted in tetrodotoxin to block action potential-mediated activity (and in some cases, ionotropic glutamatergic antagonists to block fast EPSPs), the number of other potential confounding mechanisms and difficult-to-control variables was reduced. In contrast to the lack of evidence for dormancy of basket cells, the evidence from direct experiments on several animal models from several laboratories that epileptogenesis is associated with a modest but measurable reduction in inhibition to dentate granule

765

Fig. 2. GAD mRNA-containing neurons in the dentate gyrus of a control (A) and pilocarpine-treated (B) rat. (A) In the normal dentate gyrus, numerous GAD mRNA-labeled neurons are present in the hilus (H). Many labeled neurons are also located along the inner border of the granule cell layer (G) and in the molecular layer (M). (B) In a pilocarpine-treated animal, the number of labeled neurons throughout the hilus (H) is substantially reduced. Some GAD mRNA-containing neurons persist along the base of the granule cell layer (G) and in the molecular layer (M); scale bar ¼ 200 mm. Adapted with permission from Obenaus et al. (1993) Fig. 4, p. 4476.

cells associated with a loss of specific types of vulnerable interneurons in the hilus of the dentate gyrus (and also in CA1) is quite strong. Because this reduction in the number of specific interneurons ultimately would be expected to enhance the effects of excitatory synaptic input to the granule cell population, and because assessments of GABAergic inhibition using extracellular stimulation depend heavily on the intensity of the stimulation, the types of analyses that employ electrical stimulation of an afferent input such as the perforant path are problematic. Furthermore, the presence of fewer interneurons will lead to complex effects during the repetitive activation expected to occur as multiple interictal spikes and seizures begin to occur in the entorhinal cortex and dentate gyrus.

Mossy fiber sprouting is also a common feature of TLE Axon sprouting and increased recurrent excitation Mossy fiber sprouting has attracted considerable attention because the Timm stain shows a dramatic and reproducible change in the distribution of the mossy fiber axons in the inner molecular layer of the dentate gyrus associated with TLE (Fig. 3). The ease and relative simplicity of this staining technique and the apparent clarity of the results have drawn attention to the dentate gyrus and this particular form of synaptic reorganization of a local-circuit pathway. The fact that numerous laboratories have independently observed Timm stain in the inner molecular layer of human tissue

766

Fig. 3. The Timm stain showed dark labeling in the granule cell layer and inner molecular layer in rats with kainate-induced epilepsy versus control rats. (A) Hippocampal sections from a control rat had relatively little Timm stain (with light cresyl violet counterstaining) in the granule cell and inner molecular layers. (B–E) Sections of the middle region along the septotemporal axis of the hippocampus from rats with kainate-induced epilepsy displayed different degrees of abnormal Timm staining in the granule cell layer (B shows the least and E the most). Arrows indicate regions of Timm staining in the inner molecular layer. M: molecular layer; g: granule layer; h: hilus; CA3: CA3 pyramidal cell layer; bar ¼ 500 mm. Adapted with permission from Buckmaster and Dudek (1997) Fig. 7, p. 395.

from patients with intractable epilepsy (de Lanerolle et al., 1989; Sutula et al., 1989; Houser et al., 1990; Babb et al., 1991) supports the concept that it is one of the changes likely to be associated with the broadly defined concept of synaptic reorganization. Animal models, ranging from pilocarpineto kainate-treated rats to the kindling model, have all revealed a time-dependent increase in Timm stain in the inner molecular layer of the dentate gyrus. As reviewed elsewhere in this volume (Sutula and Dudek, this volume), several lines of electrophysiological and ultrastructural data suggest that most if not all of the new synapses in the ‘‘reorganized’’ dentate gyrus are excitatory. For example, several laboratories have shown that hilar (and perforant path) stimulation can lead to delayed or long-latency EPSPs and prolonged

spike bursts several weeks or months after kainate treatment, particularly when slices from rats with chronic epilepsy are bathed in GABA-A receptor antagonists and/or high potassium, versus similar experiments on slices from control animals where hilar stimulation evokes one antidromic action potential and sometimes an EPSP (Tauck and Nadler, 1985; Cronin et al., 1992; Wuarin and Dudek, 1996; Patrylo and Dudek, 1998; Hardison et al., 2000, Lynch and Sutula, 2000). The advantage of hilar over perforant path stimulation in these earlier studies was that electrical stimulation of the hilus would be expected to activate the mossy fiber axons of the granule cells themselves versus the powerful monosynaptic inputs of the perforant path, but in either case, the key result is that the EPSPs in slices from rats

767

with robust mossy fiber sprouting occurred with a longer and more variable latency than would be expected from a direct monosynaptic input. These experiments were founded conceptually on the studies of recurrent excitation in the CA3 area (Traub and Wong, 1982, 1983; Miles and Wong, 1987), where low-intensity stimulation triggered all-or-none synaptic bursts with a long–and variable latency and paired intracellular recordings showed multisynaptic interactions, only when GABA-A mediated inhibition was blocked. The advantage of performing these experiments during blockade of GABA-A mediate inhibition is based not only on the previous work showing that GABA-mediated inhibition has a ‘‘masking’’ effect on recurrent excitation (Christian and Dudek, 1988; Traub and Wong, 1982, 1983; Miles and Wong, 1987), but also because it controls at least partially for epileptogenesis-associated differences in GABA-mediated inhibition (see above), particularly since the experimental design in these studies compared slices from the animal model of epilepsy with exactly equivalent slices from control animals. Electrical stimulation, however, activates axons of passage and can thus be nonspecific. More direct experiments using glutamate microstimulation techniques that do not activate fibers of passage, including glutamate microdrops (Wuarin and Dudek, 1996; Lynch and Sutula, 2000) and focal photoactivation of caged glutamate (Molnar and Nadler, 1999; Wuarin and Dudek, 2001), showed that (1) relatively selective stimulation of dentate granule cells could cause EPSCs in other granule cells in normal medium, (2) these recurrent excitatory circuits increased in density and effectiveness over the ensuing months after experimental SE, and (3) when the slices were bathed in bicuculline, the new local excitatory circuits of the granule cells could generate novel network bursting to focal uncaging of glutamate in the granule cell layer (but not elsewhere) several months after SE, but not at earlier times and not in slices from control preparations. Although these experiments with glutamate microdrops and focal uncaging of glutamate used more specific techniques than extracellular electrical stimulation to activate granule cells, and showed that the formation of new recurrent excitatory circuits and the

consequent generation and propagation of epileptiform activity in the dentate gyrus was time dependent (Wuarin and Dudek, 2001), they did not directly address the issue of monosynaptic connections between individual granule cells. Dual intracellular recordings provided direct evidence that mossy fiber sprouting in rats with pilocarpineinduced epilepsy is associated with monosynaptic recurrent excitatory connections among dentate granule cells (Scharfman et al., 2003). The most direct and quantitative ultrastructural studies also point to a high preponderance of new excitatory synapses to dentate granule cells versus interneurons (Zhang and Houser, 1999; Buckmaster et al., 2002). Although synaptic reorganization may have a unique importance in the dentate gyrus, a more likely scenario is that axon sprouting and the formation of new recurrent excitatory circuits is a widespread phenomenon after injury that occurs throughout many areas of the hippocampus and neocortex during epileptogenesis. For example, numerous laboratories have provided morphological and electrophysiological evidence that a similar mechanism of axonal sprouting and enhanced recurrent excitation occurs in the CA1 area during epileptogenesis (Meier and Dudek, 1996; Perez et al., 1996; Esclapez et al., 1999; Smith and Dudek, 2001, 2002; Shao and Dudek, 2004). As expected from previous experiments performed on recurrent excitatory circuits in the CA3 area (Traub and Wong, 1982, 1983; Miles and Wong, 1987; Christian and Dudek, 1988), several laboratories have shown that reduction of GABA-A mediated inhibition or an increase in extracellular potassium has an unmasking effect on the new local recurrent excitatory circuits in the dentate gyrus. The underlying concept is that GABA-A mediated inhibition, along with the membrane properties of the granule cells (e.g., the highly negative resting potentials, high threshold, and resistance to repetitive firing), will tend to impede the likelihood of multisynaptic interactions, which is the prime mechanism by which increased recurrent excitation would be expected to lead to seizure activity. This concept may provide a hypothetical explanation for how seizures might not occur under some circumstances, but would emerge under others; that is, seizures would be generated and/or

768

propagated when synaptic inhibition is slightly and transiently depressed and/or extracellular potassium increased in concentration, as this particular set of changes would allow multisynaptic interactions to generate a seizure.

Increased recurrent excitation versus increased inhibition as a consequence of mossy fiber sprouting Although most experiments have focused on and supported the hypothesis that mossy fiber sprouting during chronic epileptogenesis leads to increased recurrent excitation among granule cells, some data suggest that the sprouted mossy fibers preferentially synapse on interneurons (e.g., Sloviter, 1992; Kotti et al., 1997; Harvey and Sloviter, 2005). The hypothesis is that the ‘‘dormant’’ basket cells, which have presumably lost excitatory synaptic input during repetitive seizures, become re-innervated by the sprouted mossy fibers. The histopathological data include light micrographs of basket cells surrounded by Timm stain, but as pointed out previously by Ribak and Peterson (1991), similar micrographs can be obtained from the temporal pole of the hippocampus of normal animals, and the light-microscopic techniques do not show whether mossy fibers near interneurons actually synapse with them. Ultrastructural observations have reported mossy fiber synapses on interneurons (Cavazos et al., 2003; see Sutula and Dudek, this volume), but the number of synaptic contacts from mossy fibers to interneurons has not been rigorously evaluated and there is no quantitative evidence that there is an increase over normal levels. Indeed, contacts by sprouted mossy fibers on interneurons appear to be infrequent if not rare (see Sutula and Dudek, this volume). Most of the electrophysiological evidence for the ‘‘hyper-inhibition’’ hypothesis is essentially based on variations of the paired-pulse technique (Sloviter, 1992; Buckmaster and Dudek, 1997; Wilson et al., 1998; Harvey and Sloviter, 2005), which is potentially confounded by technical issues (e.g., see Waldbaum and Dudek, 2005, 2006). Thus, although mossy fiber sprouting in the dentate gyrus- and sprouting-based synaptic

reorganization in other areas of the hippocampus and temporal lobe may involve new inhibitory circuits from sprouted mossy fibers to interneurons, the preponderance of evidence from many laboratories with a wide range of techniques using several different animal models point to the conclusion that mossy fiber sprouting leads predominantly to new recurrent excitatory circuits.

Mossy fiber sprouting and increased seizure susceptibility Many studies have noted the close association between the presence of mossy fiber sprouting and decreased seizure threshold (i.e., in the kindling model) or the presence of spontaneous recurrent seizures (i.e., in the models of epileptogenesis based on SE). Nonetheless, many tissue specimens from patients with intractable TLE lack sprouting, and seizure frequency does not appear to be related to the degree of mossy fiber sprouting in animal models (e.g., Buckmaster and Dudek, 1997). This issue is complex, however, because the distribution of mossy fiber sprouting is not homogeneous (e.g., more robust sprouting generally occurs in the temporal versus septal hippocampus in animal models). Plus, axonal sprouting and the formation of new recurrent excitatory circuits likely occur in many other structures in addition to the dentate gyrus, and synaptic reorganization in these other areas is difficult to detect. Gorter et al. (2001) reported that rats subjected to amygdala-stimulated SE could be separated into two groups: (1) animals that showed a progressive increase in seizure frequency over time and that had robust mossy fiber sprouting, and (2) rats with a low seizure frequency and little or no mossy fiber sprouting. Longo and Mello (1997) have reported that pretreatment with the protein-synthesis inhibitor cycloheximide blocks mossy fiber sprouting but does not prevent the development of spontaneous recurrent seizures. Williams et al. (2002), however, found that cycloheximide pretreatment had no detectable effect on mossy fiber sprouting. In this study, some rats did not experience SE (i.e., had few if any convulsive seizures during pilocarpine treatment), but still had

769

at least a few subsequent spontaneous seizures with little or no Timm stain in the inner molecular layer, consistent with the data of Gorter et al. (2001) suggesting that mossy fiber sprouting is not necessary for the development of occasional spontaneous seizures. It should not be surprising that the dentate gyrus and mossy fiber are not required for limbic seizures (for more extensive discussion on this point, see Sutula and Dudek, this volume). More recently, Toyoda and Buckmaster (2005) also attempted to determine if cycloheximide blocked mossy fiber sprouting after pilocarpine-induced SE; they infused cycloheximide directly into the area around the hippocampus for prolonged periods from an osmotic minipump, provided evidence that the cycloheximide was present in the area around the dentate gyrus, and also observed robust Timm stain in the inner molecular layer (i.e., cycloheximide had no detectable effect on mossy fiber sprouting). It is not clear how one pretreatment injection of cycloheximide would be expected to block the weeks/months-long continuous process that is responsible for sprouting of the mossy fibers and the formation of new synaptic contacts with dentate granule cells. Presently, therefore, the proposal that cycloheximide blocks mossy fiber sprouting is problematic. The observation of spontaneous seizures before little or any mossy fiber sprouting has occurred (i.e., between 5 and 7 days after kainate-induced SE in the study of Hellier et al., 1999), and other data where animals have shown spontaneous seizures with little or no evidence of mossy fiber sprouting weeks after pilocarpine treatment (Williams et al., 2002), suggest that other mechanisms (e.g., loss of interneurons) besides mossy fiber sprouting in the dentate gyrus can lead to the development of spontaneous seizures after experimental SE. Although axonal sprouting after neuronal injury in other brain areas could account for the occurrence of spontaneous seizures, particularly in the cases of spontaneous seizures within a few weeks after SE, a more likely explanation is that the death of vulnerable interneurons during the SE (e.g., see Houser and Esclapez, 1996) decreased seizure threshold (i.e., increased seizure susceptibility) and led to occasional seizures.

Does mossy fiber sprouting inhibit seizures? After spontaneous seizures in pilocarpine-treated rats, c-fos expression has been reported to occur preferentially in interneurons versus dentate granule cells (Harvey and Sloviter, 2005), which has been interpreted to mean that mossy fiber sprouting leads to new synapses onto inhibitory interneurons rather than to granule cells (i.e., hyper-inhibition), and that mossy fiber sprouting suppresses rather than promotes seizures. Harvey and Sloviter (2005) evaluated c-fos staining 1 h after a spontaneous seizure, while in similar studies, Peng and Houser (2005) euthanized animals 15 min after a seizure and found Fos expression preferentially in granule cells. Peng and Houser (2005) found that staining was observed in interneurons when animals were euthanized after 1 h, the Fos staining in interneurons persisted longer than in granule cells, and the enhanced Fos expression in the granule cells was not observed when a second seizure occurred shortly after a previous seizure. The results of Harvey and Sloviter (2005) could have arisen if they studied animals that had had more than one seizure in close temporal proximity. In addition, progressive activation of granule cells and interneurons during spontaneous seizures would be expected to occur over hundreds of milliseconds, and seizures typically last up to 3 min; therefore, c-fos expression lacks the temporal resolution to differentiate whether granule cells or interneurons are sequentially or preferentially activated during spontaneous seizures. Harvey and Sloviter (2005) also proposed that population spikes are highly prevalent in the dentate gyrus during spontaneous seizures that occur a few days after SE, but are rare or nonexistent when spontaneous seizures occur weeks later, as might be expected if inhibition increased in the dentate gyrus over time after SE. Alternative mechanisms besides increased inhibition are possible, including depolarization inactivation, as occurs during paroxysmal depolarization shifts after blockade of GABA-A receptors. Thus, it is not clear whether these data apply to the question of whether mossy fiber sprouting leads to hyper-inhibition and is potentially compensatory.

770

Reorganization of GABAergic neurons Another hypothesis, based on increased immunohistochemical staining of GABAergic terminals in human tissue and in animal models of TLE is whether other remaining GABAergic terminals might sprout new synaptic connections to partly (or totally) restore GABA-mediated inhibition (e.g., see Davenport et al., 1990; Bausch, 2005). If this occurred, one would expect an increase in the frequency of mIPSCs with time after experimental SE, but this was not seen in rats with kainate-induced epilepsy (Shao and Dudek, 2005); however, other techniques may reveal axon sprouting, synaptic reorganization and recovery of inhibition during epileptogenesis.

The dentate gyrus as a model for studies of epileptogenesis associated with TLE Although the dentate gyrus and the hilus clearly have an important and probably have a unique role in hipocampal integration and epileptogenesis, they may also serve as a model for other cortical structures where epileptogenesis involves neuronal death and synaptic reorganization in addition to possible alterations in neurotransmitter receptors and voltage-gated channels. Although epilepsy is caused by a great variety of age-dependent etiologies and the mechanisms appear to be equally heterogeneous, the cellular and network mechanisms of (1) a loss of vulnerable interneurons, and (2) axon sprouting with progressive formation of new recurrent excitatory circuits have the most support from independent experiments in numerous laboratories and multiple animal models. Although other molecular, cellular and network mechanisms have been implicated in the dentate gyrus, these two cellular mechanisms that have been extensively documented in the dentate gyrus may be generally applicable in many other temporal and neocortical structures. Precisely how these mechanisms and others in the dentate gyrus and elsewhere lead to the epileptic seizures associated with TLE will require further studies. While the heterogeneity of TLE etiologies and the mechanisms underlying neuronal and network

synchronization demand continuing attention to a diverse range of potential mechanisms and interactions spanning molecular, cellular, and systems levels, the insights gained from study of neuronal loss and mossy fiber sprouting during the last two decades have been a substantial advance and a major accomplishment for epilepsy research.

References Ang, C.W., Carlson, G.C. and Coulter, D.A. (2006) Massive and specific dysregulation of direct cortical input to the hippocampus in temporal lobe epilepsy. J. Neurosci., 26: 11850–11856. Babb, T.L., Kupfer, W.R., Pretorius, J.K., Crandall, P.H. and Levesque, M.F. (1991) Synaptic reorganization by mossy fibers in human epileptic fascia dentata. Neuroscience, 42: 351–363. Bausch, S.B. (2005) Axonal sprouting of GABAergic interneurons in temporal lobe epilepsy. Epilepsy Behav., 7: 390–400. Behr, J., Lyson, K.J. and Mody, I. (1998) Enhanced propagation of epileptiform activity through the kindled dentate gyrus. J. Neurophysiol., 79: 1726–1732. Bekenstein, J.W. and Lothman, E.W. (1993) Dormancy of inhibitory interneurons in a model of temporal lobe epilepsy. Science, 259: 97–100. Ben-Ari, Y. (1985) Limbic seizure and brain damage produced by kainic acid: mechanisms and relevance to human temporal lobe epilepsy. Neuroscience, 14: 375–404. Bengzon, J.K.Z., Elme´r, E., Nanobashvili, A., Kokaia, M. and Lindvall, O. (1997) Apoptosis and proliferation of dentate gyrus neurons after single and intermittent limbic seizures. Proc. Natl. Acad. Sci., 94: 10432–10437. Bernard, C., Esclapez, M., Hirsch, J.C. and Bernard, C. (1998) Interneurones are not so dormant in temporal lobe epilepsy: a critical reappraisal of the dormant basket cell hypothesis. Epilepsy Res., 32: 93–103. Bertram, E.H. (1997) Functional anatomy of spontaneous seizures in a rat model of limbic epilepsy. Epilepsia, 38: 95–105. Bertram, E.H., Zhang, D.X., Mangan, P., Fountain, N. and Rempe, D. (1998) Functional anatomy of limbic epilepsy: a proposal for central synchronization of a diffusely hyperexcitable network. Epilepsy Res., 32: 194–205. Bragin, A., Mody, I., Wilson, C.L. and Engel, J. (2002) Local generation of fast ripples in epileptic brain. J. Neurosci., 22: 2012–2021. Buckmaster, P.S. and Dudek, F.E. (1997) Neuron loss, granule cell axon reorganization, and functional changes in the dentate gyrus of epileptic kainate-treated rats. J. Comp. Neurol., 385: 385–404. Buckmaster, P.S., Zhang, G.F. and Yamawaki, R. (2002) Axon sprouting in a model of temporal lobe epilepsy creates a predominantly excitatory feedback circuit. J. Neurosci., 22: 6650–6658.

771 Cascino, G.D. (2005) Temporal lobe epilepsy: more than hippocampal pathology. Epilepsy Curr., 5: 187–189. Cavazos, J.E., Das, I. and Sutula, T.P. (1994) Neuronal loss induced in limbic pathways by kindling: evidence for induction of hippocampal sclerosis by repeated brief seizures. J. Neurosci., 14: 3106–3121. Cavazos, J.E., Golarai, G. and Sutula, T.P. (1991) Mossy fiber synaptic reorganization induced by kindling: time course of development, progression, and permanence. J. Neurosci., 11: 2795–2803. Cavazos, J.E., Zhang, P., Qazi, R. and Sutula, T.P. (2003) Ultrastructural features of sprouted mossy fiber synapses in kindled and kainic acid-treated rats. J. Comp. Neurol., 458: 272–292. Christian, E.P. and Dudek, F.E. (1988) Characteristics of local excitatory circuits studied with glutamate microapplication in the CA3 area of rat hippocampal slices. J. Neurophysiol., 59: 90–109. Cohen, I., Navarrao, V., Clemenceau, S., Baulac, M. and Miles, R. (1998) On the origin of interictal activity in human temporal lobe epilepsy in vitro. Science, 298: 1418–1421. Cossart, R., Dinocourt, C., Hirsch, J.C., Merchan-Perez, A., De Felipe, J., Ben-Air, Y., Esclapez, M. and Bernard, C. (2001) Dendritic but not somatic GABAergic inhibition is decreased in experimental epilepsy. Nat. Neurosci., 4: 52–62. Coulter, D.A. and Carlson, G.C. (this volume) Functional regulation of the dentate gyrus by GABA-mediated inhibition. In: Scharfman, H. (Ed.), The Dentate Gyrus. Progress in Brain Research. Elsevier, Amsterdam. Crain, B.J., Westerkam, W.D., Harrison, A.H. and Nadler, J.V. (1988) Selective neuronal death after transient forebrain ischemia in the Mongolian gerbil: a silver impregnation study. Neuroscience, 27: 387–402. Cronin, J., Obenaus, A., Houser, C.R. and Dudek, F.E. (1992) Electrophysiology of dentate granule cells after kainate-induced synaptic reorganization of the mossy fibers. Brain Res., 573: 305–310. Davenport, D.J., Brown, W.J. and Babb, T.L. (1990) Sprouting of GABAergic and mossy fiber axons in dentate gyrus following intrahippocampal kainate in the rat. Exp. Neurol., 109: 180–190. Doherty, J. and Dingledine, R. (2001) Reduced excitatory drive onto interneurons in the dentate gyrus after status epilepticus. J. Neurosci., 21: 2048–2057. Engel, J. (1989) Seizures and Epilepsy (Contemporary Neurology Series). F.A. Davis Company, Philadelphia, PA. Esclapez, M., Hirsch, J.C., Ben Ari, Y. and Bernard, C. (1999) Newly formed excitatory pathways provide a substrate for hyperexcitability in experimental temporal lobe epilepsy. J. Comp. Neurol., 408: 449–460. Esclapez, M., Hirsch, J.C., Khazipov, R., Ben Ari, Y. and Bernard, C. (1997) Operative GABAergic inhibition in hippocampal CA1 pyramidal neurons in experimental epilepsy. Proc. Natl. Acad. Sci. U.S.A., 94: 12151–12156. Fricke, R.A. and Prince, D.A. (1984) Electrophysiology of dentate gyrus granule cells. J. Neurophysiol., 51: 195–209.

Gloor, P. (1992) Role of the amygdala in temporal lobe epilepsy. In: Aggleton J.P. (Ed.), The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental Dysfunction. Wiley-Liss, Inc., New York, pp. 505–538. Gorter, J.A., van Vliet, E.A., Aronica, E. and Lopes da Silva, F.H. (2001) Progression of spontaneous seizures after status epilepticus is associated with mossy fibre sprouting and extensive bilateral loss of hilar parvalbumin and somatostatinimmunoreactive neurons. Eur. J. Neurosci., 13: 657–669. Hardison, J.L., Okazaki, M.M. and Nadler, J.V. (2000) Modest increase in extracellular potassium unmasks effect of recurrent mossy fiber growth. J. Neurophysiol., 84: 2380–2389. Harvey, B.D. and Sloviter, R.S. (2005) Hippocampal granule cell activity and c-Fos expression during spontaneous seizures in awake, chronically epileptic, pilocarpine-treated rats: Implications for hippocampal epileptogenesis. J. Comp. Neurol., 488: 442–463. Heinemann, U., Beck, H., Dreier, J.P., Ficker, E., Stabel, J. and Zhang, C.L. (1992) The dentate gyrus as a regulated gate for the propagation of epileptiform activity. Epilepsy Res. Suppl., 7: 273–280. Hellier, J.L., Patrylo, P.R., Dou, P., Nett, M., Rose, G.M. and Dudek, F.E. (1999) Assessment of inhibition and epileptiform activity in the septal dentate gyrus of freely behaving rats during the first week after kainate treatment. J. Neurosci., 19: 10053–10064. Houser, C.R. (this volume) Interneurons of the dentate gyrus: An overview of cell types, terminal fields and neurochemical identity. In: Scharfman, H. (Ed.), The Dentate Gyrus. Progress in Brain Research. Elsevier, Amsterdam. Houser, C.R. and Esclapez, M. (1996) Vulnerability and plasticity of the GABA system in the pilocarpine model of spontaneous recurrent seizures. Epilepsy Res., 26: 207–218. Houser, C.R., Miyashiro, J.E., Swartz, B.E., Walsh, G.O., Rich, J.R. and Delgado-Escueta, A.V. (1990) Altered patterns of dynorphin immunoreactivity suggest mossy fiber reorganization in human hippocampal epilepsy. J. Neurosci., 10: 267–282. Hsu, M. and Buzsaki, G. (1993) Vulnerability of mossy fiber targets in the rat hippocampus to forebrain ischemia. J. Neurosci., 13: 3964–3979. Iijima, T., Witter, M.P., Ichikawa, M., Tominage, T., Kajiwara, R. and Matsumoto, G. (1996) Entorhinal-hippocampal interactions revealed by real-time imaging. Science, 272: 1176–1179. Kapur, J., Haas, K.F. and Macdonald, R.L. (1999) Physiological properties of GABAA receptors from acutely dissociated rat dentate granule cells. J. Neurophysiol., 81: 2464–2471. Kesner, R.P. (this volume) A behavioral analysis of dentate gyrus function. In: Scharfman, H. (Ed.), The Dentate Gyrus. Progress in Brain Research. Elsevier, Amsterdam. Kobayashi, M. and Buckmaster, P.S. (2003) Reduced inhibition of dentate granule cells in a model of temporal lobe epilepsy. J. Neurosci., 23: 2440–2452. Kotti, T., Riekkinen, P.J. and Miettinen, R. (1997) Characterization of target cells for aberrant mossy fiber collaterals in the dentate gyrus of epileptic rat. Exp. Neurol., 146: 323–330.

772 de Lanerolle, N.C., Kim, J.H., Robbins, R.J. and Spencer, D.D. (1989) Hippocampal interneuron loss and plasticity in human temporal lobe epilepsy. Brain Res., 495: 387–395. Longo, B.M. and Mello, L.E. (1997) Blockade of pilocarpineor kainate-induced mossy fiber sprouting by cycloheximide does not prevent subsequent epileptogenesis in rats. Neurosci. Lett., 226: 163–166. Lothman, E.W., Bertram, E.H., Kapur, J. and Stringer, J.L. (1990) Recurrent spontaneous hippocampal seizures in the rat as a chronic sequela to limbic status epilepticus. Epilepsy Res., 6: 110–118. Lothman, E.W., Stringer, J.L. and Bertram, E.H. (1992) The dentate gyrus as a control point for seizures in the hippocampus and beyond. Epilepsy Res. Suppl., 7: 301–313. Lowenstein, D.H., Thomas, M.J., Smith, D.H. and McIntosh, T.K. (1992) Selective vulnerability of dentate hilar neurons following traumatic brain injury: a potential mechanistic link between head trauma and disorders of the hippocampus. J. Neurosci., 12: 4846–4853. Lynch, M. and Sutula, T. (2000) Recurrent excitatory connectivity in the dentate gyrus of kindled and kainic acid-treated rats. J. Neurophysiol., 83: 693–704. Margerison, J.H. and Corsellis, J.A. (1966) Epilepsy and the temporal lobes. A clinical, electroencephalographic and neuropathological study of the brain in epilepsy, with particular reference to the temporal lobes. Brain, 89: 499–530. Meier, C.L. and Dudek, F.E. (1996) Spontaneous and stimulation-induced synchronized burst afterdischarges in the isolated CA1 of kainate-treated rats. J. Neurophysiol., 76: 2231–2239. Meldrum, B.S. and Bruton, C.J. (1992) Epilepsy. In: Adams J.H. and Duchen L.W. (Eds.), Greenfield’s Neuropathology. Oxford, New York, pp. 1246–1283. Miles, R. and Wong, R.K. (1987) Inhibitory control of local excitatory circuits in the guinea-pig hippocampus. J. Physiol., 388: 611–629. Molnar, P. and Nadler, J.V. (1999) Mossy fiber-granule cell synapses in the normal and epileptic rat dentate gyrus studied with minimal laser photostimulation. J. Neurophysiol., 82: 1883–1894. Nadler, J.V. (1991) Kainic acid as a tool for the study of temporal lobe epilepsy. Life Sci., 29: 2031–2042. Obenaus, A., Esclapez, M. and Houser, C.R. (1993) Loss of glutamate decarboxylase mRNA-containing neurons in the rat dentate gyrus following pilocarpine-induced seizures. J. Neurosci., 13: 4470–4485. Parent, J.M. (this volume) Adult neurogenesis in the intact and epileptic dentate gyrus. In: Scharfman, H. (Ed.), The Dentate Gyrus. Progress in Brain Research. Elsevier, Amsterdam. Patrylo, P.R. and Dudek, F.E. (1998) Physiological unmasking of new glutamatergic pathways in the dentate gyrus of hippocampal slices from kainate-induced epileptic rats. J. Neurophysiol., 79: 418–429. Peng, Z. and Houser, C.R. (2005) Temporal patterns of Fos expression in the dentate gyrus after spontaneous seizures in a mouse model of temporal lobe epilepsy. Epilepsy Curr., 6: 57–58.

Perez, Y., Morin, F., Beaulieu, C. and Lacaille, J.C. (1996) Axonal sprouting of CA1 pyramidal cells in hyperexcitable hippocampal slices of kainate-treated rats. Eur. J. Neurosci., 8: 736–748. Ribak, C.E. and Peterson, G.M. (1991) Intragranular mossy fibers in rats and gerbils form synapses with the somata and proximal dendrites of basket cells in the dentate gyrus. Hippocampus, 1: 355–364. Sayin, U., Osting, S., Hagen, J. and Rutecki, P.S.T. (2003) Spontaneous seizures and loss of axo-axonic and axo-somatic inhibition induced by repeated brief seizures in kindled rats. J. Neurosci., 23: 2759–2768. Sayin, U., Rutecki, P. and Sutula, T. (1999) Alterations in nmda dependent currents in granule cells of the dentate gyrus contribute to the induction but not the permanence of kindling. J. Neurophysiol., 81: 564–574. Scharfman, H.E., Sollas, A.L., Berger, R.E. and Goodman, J.H. (2003) Electrophysiological evidence of monosynaptic excitatory transmission between granule cells after seizureinduced mossy fiber sprouting. J. Neurophysiol., 90: 2536–2547. Shao, L.R. and Dudek, F.E. (2004) Increased excitatory synaptic activity and local connectivity of hippocampal CA1 pyramidal cells in rats with kainate-induced epilepsy. J. Neurophysiol., 92: 1366–1373. Shao, L.-R. and Dudek, F.E. (2005) Changes in mIPSCs and sIPSCs after kainate treatment: evidence for loss of inhibitory input to dentate granule cells and possible compensatory responses. J. Neurophysiol., 94: 952–960. Sloviter, R.S. (1987) Decreased hippocampal inhibition and a selective loss of interneurons in experimental epilepsy. Science, 235: 73–76. Sloviter, R.S. (1991) Permanently altered hippocampal structure, excitability, and inhibition after experimental status epilepticus in the rat: the ‘‘dormant basket cell’’ hypothesis and its possible relevance to temporal lobe epilepsy. Hippocampus, 1: 41–66. Sloviter, R.S. (1992) Possible functional consequences of synaptic reorganization in the dentate gyrus of kainate-treated rats. Neurosci. Lett., 137: 91–96. Sloviter, R.S., Zappone, C.A., Harvey, B.D., Bumanglag, A.V., Bender, R.A. and Frotscher, M. (2003) ‘‘Dormant basket cell’’ hypothesis revisited: relative vulnerabilities of dentate gyrus mossy cells and inhibitory interneurons after hippocampal status epilepticus in the rat. J. Comp. Neurol., 459: 44–76. Smith, B.N. and Dudek, F.E. (2001) Short- and long-term changes in CA1 network excitability after kainate treatment in rats. J. Neurophysiol., 85: 1–9. Smith, B.N. and Dudek, F.E. (2002) Network interactions mediated by new excitatory connections between CA1 pyramidal cells in rats with kainate-induced epilepsy. J. Neurophysiol., 87: 1655–1658. Stables, J.P., Bertram, E., Dudek, F.E., Holmes, G., Mathern, G., Pitkanen, A. and White, H.S. (2003) Therapy discovery for pharmacoresistant epilepsy and for disease-modifying therapeutics: summary of the NIH/NINDS/AES Models II Workshop. Epilepsia, 44: 1472–1478.

773 Staley, K.J., Otis, T.S. and Mody, I. (1992) Membrane properties of dentate gyrus granule cells: comparison of sharp microelectrode and while-cell recordings. J. Neurophysiol., 67: 1346–1358. Stringer, J.L. and Lothman, E.W. (1992) Bilateral maximal dentate activation is critical for the appearance of an afterdischarge in the dentate gyrus. Neuroscience, 46: 309–314. Sutula, T., Cascino, G., Cavazos, J., Parada, I. and Ramirez, L. (1989) Mossy fiber synaptic reorganization in the epileptic human temporal lobe. Ann. Neurol., 26: 321–330. Sutula, T.P. and Dudek, F.E. (this volume) Unmasking recurrent excitation generated by mossy fiber sprouting in the epileptic dentate gyrus: an emergent property of a complex system. In: Scharfman, H. (Ed.), The Dentate Gyrus. Progress in Brain Research. Elsevier, Amsterdam. Sutula, T.P. and Steward, O. (1986) Quantitative analysis of synaptic potentiation during kindling of the perforant path. J. Neurophysiol., 56: 732–746. Sutula, T.P. and Steward, O. (1987) Facilitation of kindling by prior induction of long-term potentiation in the perforant pathway. Brain Res., 420: 109–117. Tauck, D.L. and Nadler, J.V. (1985) Evidence of functional mossy fiber sprouting in hippocampal formation of kainic acid-treated rats. J. Neurosci., 5: 1016–1022. Toth, Z., Hollrigel, G.S., Gorcs, T. and Solesz, I. (1997) Instantaneous perturbation of dentate interneuronal networks by a pressure wave-transient delivered to the neocortex. J. Neurosci., 17: 8106–8117. Toyoda, I. and Buckmaster, P.S. (2005) Prolonged infusion of cycloheximide does not block mossy fiber sprouting in a model of temporal lobe epilepsy. Epilepsia, 46: 1017–1020. Traub, R.D. and Wong, R.K. (1982) Cellular mechanism of neuronal synchronization in epilepsy. Science, 216: 745–747. Traub, R.D. and Wong, R.K. (1983) Synchronized burst discharge in disinhibited hippocampal slice. II. Model of cellular mechanism. J. Neurophysiol., 49: 459–471. Turski, W.A., Cavalheiro, E.A., Schwarz, M., Czuczwar, S.J., Kleinrok, Z. and Turski, L. (1983) Limbic seizures produced by pilocarpine in rats: behavioural, electroencephalographic and neuropathological study. Behav. Brain Res., 9: 315–335

In: H. Scharfman (Ed.), The Dentate Gyrus. Progress in Brain Research. Elsevier, Amsterdam. Waldbaum, S. and Dudek, F. (2005) Parametric assessment of the consistency of the paired-pulse protocol for studying hippocampal inhibition and hyperexcitability. Soc. Neurosci. Abstr., 31: 276.12. Waldbaum, S. and Dudek, F.E. (2006) Increased paired-pulse suppression after low-dose application of the selective GABAA receptor antagonist SR-95531, GABAzine: The paired-pulse technique can provide misleading evidence for ‘‘hyperinhibition’’ when GABAA-mediated inhibition is decreased. Soc. Neurosci. Abstr., 32: 794.18. Wierenga, C.J. and Wadman, W.J. (1999) Miniature inhibitory postsynaptic currents in CA1 pyramidal neurons after kindling epileptogenesis. J. Neurophysiol., 82: 1352–1362. Williams, P.A., Pou, P. and Dudek, F.E. (2004) Epilepsy and synaptic reorganization in a model of perinatal hypoxia-ischemia. Epilepsia, 45: 1210–1218. Williams, P.A., Wuarin, J.-P., Dou, P., Ferraro, D.J. and Dudek, F.E. (2002) Reassessment of the effects of cycloheximide on mossy fiber sprouting and epileptogenesis in the pilocarpine model of temporal lobe epilepsy. J. Neurophysiol., 88: 2075–2087. Williamson, A. and Patrylo, P.R. (this volume) Physiological studies of human dentate granule cells. In: Scharfman, H. (Ed.), The Dentate Gyrus. Progress in Brain Research. Elsevier, Amsterdam. Wilson, D.L., Kahn, S.U., Engel, J., Isokawa, M., Babb, T.L. and Behnke, E.J. (1998) Paired pulse suppression and facilitation in human epileptogenic hippocampal formation. Epilepsy Res., 31: 211–230. Wuarin, J.P. and Dudek, F.E. (1996) Electrographic seizures and new recurrent excitatory circuits in the dentate gyrus of hippocampal slices from kainate-treated epileptic rats. J. Neurosci., 16: 4438–4448. Wuarin, J.P. and Dudek, F.E. (2001) Excitatory synaptic input to granule cells increases with time after kainate treatment. J. Neurophysiol., 85: 1067–1077. Zhang, N. and Houser, C.R. (1999) Ultrastructural localization of dynorphin in the dentate gyrus in human temporal lobe epilepsy: a study of reorganized mossy fiber synapses. J. Comp. Neurol., 405: 472–490.

This page intentionally left blank

Subject Index

Seizure susceptibility 689–690 Spatial learning and memory 689 Allocentric coordinate 620 Allopregnanolone 405 Alzheimer’s disease (AD) 463–464 Dentate gyrus 723–734 Disconnected DG 732–734 Excitatory–inhibitory systems 731–732 Histopathology 724–726 Entorhinal cortex atrophy 743–745 Hippocampal atrophy 741–751 MRI 741–743 Neuropeptide Y (NPY) and its receptors in 292–293 Somatostatin (SST) levels in 268–269 b-Amyloid degradation by 268 White matter changes 748–750 Amygdalo-entorhinal cortex (AE) 46 Amyloid b deposition 725–726 Androgen receptors (AR) 399, 407 Animal models of epilepsy 139, 183–195 Antidepressants (ADs) Neurogenic effects, effects on neurogenesis 707–708 Serotonin-dependence 708–710 Aplysia 309 Apolipoprotein E receptor 2 (ApoER2) 135 Apolipoprotein E receptor 3 (ApoER3) 148 Apoptosis 355, 357, 362 Chronic stress and 364 Arachidonoylethanolamide (AEA) 321, 331 Archicortex 23 Astrocytes 513 Astrocytic b-adrenoceptors 301 Astroglial cells 404, 513–514 Autoassociative memory function 579 Autoradiograms 671 Axo-axonic cells 644–645

a-Adrenoceptors 302 a-Amino-3-hydroxy-5- methylisoxazole-4propionic acid (AMPA) 401 Aberrant neurogenesis hypothesis 536 Acetylcholine 63–78, 461 Acetylcholinesterase (AChE) Reorganization after entorhinal denervation 509–510 Activity-regulated cytoskeleton (Arc) 671 Arc-associated protein (Arc) as synapse consolidation mediator 453, 455–457 Adrenalectomy (ADX) 355, 358–361 Adrenoceptors 300–302 a-Adrenoceptors 302 b-Adrenoceptors 301 Adult neurogenesis in the intact and epileptic DG 529–537 Hippocampal neurogenesis Modulation 532 Seizure-induced 533–534 Integration and function of adult born DGCs 530–531 Neonatal and adult dentate granule cell neurogenesis 530 Regulation 531–533 Afferent and efferent connections 17, 32–34 Afferent fiber lamination formation 136–139 Cholinergic and GABAergic afferents from medial septum 138–139 Commissural/associational (C/A) fibers 137–138 Entorhinal afferents 136–137 Non-hippocampal afferent connections 33 Afterhyperpolarization (AHP) 186 Aging effects GABAergic circuitry 682–687 Glutamatergic circuitry 687–689 on dentate circuitry and function 679–690 on dentate filter function 679–681

775

776

Axon sprouting 542 Axonal degeneration techniques 44 Axonal reorganization, after entorhinal denervation 505–510 Basal forebrain inputs 18–19 Basket cells 155, 157–159, 644, 762–763, 762–764 Basolateral amygdala 622 Behavioral analysis of dentate gyrus 567–575 Conjunctive encoding 567–568 Encoding vs. retrieval of spatial information 572–574 Spatial pattern separation 568–572 Biotinylated dextran amine (BDA) 69 Blood-derived cells 513 Braak-staging 725, 733–734 Brain-derived neurotrophic factor (BDNF) 115, 373–378, 400 Arc 456 Consolidation 454–456, 460–461 Effect on neurogenesis 378 Effects on learning and memory 378 Effects on synaptic transmission 377–378 Gene regulation 375–376 in Huntington’s disease 385 in neurodegenerative diseases 385 in neuropsychiatric disease 385–386 in pain 384–385 in Rett syndrome 385 Localization, transport and release 374–375 Physiologic regulation of 377 Role in GABAergic synapses 377 Roles during development 376–377 Seizure 376 Translation control in dendrites 457–460 Brain slice methods 111 Brain-specific Na+-dependent inorganic phosphate cotransporter (BNPI) 75–76 BNPI/VGLUT1 75–76 DNPI/VGLUT2 76–77 VGLUT3 75, 77–78 Brainstem inputs 19 Brodman’s areas 28a and 28b 46 CA1 (regio superior) 4–5, 418–419 CA1 interneurons 207 CA2 4–5, 418–419 CA3 (regio inferior) 30, 123, 207, 418–419, 422

Anatomical connectivity from DG to 10, 109–111 CA3 backprojection to DG 627–636 (see also separate entry) Lesion-induced mossy fiber (MF) sprouting 99–100 Lesions 572–574 MF trajectory and termination in 86–97 Pyramidal cells 577–588 Role in spatial pattern separation 568–572 CA3 backprojection to DG 627–636 Associative networks and reverberatory circuits 635 Backprojection 629–633 Anatomical evidence 629–630 Physiological evidence 630–633 Functional implications 633–635 Hippocampal information processing 633–634 Pathological conditions 634–635 Ca2+ channels Ca2+/calmodulin-dependent protein kinase (CaMKII) 478 Ca2+/phospholipid-dependent protein kinase (PKC) 477 in human DGCs 186 in mossy fiber (MF) transmission 112 Interleukin 1-b (IL-1-b) and 348–349 Cajal-Retzius cells 137, 140, 144 Calbindin 728–729 Calcitonin generelated peptide (CGRP)immunoreactive neurons 66, 323 Calretinin 226–227, 508 Labeling comparison in mouse and rat 227 cAMP response element binding protein (CREB) 267, 427 cAMP-dependent protein kinase (PKA) 478–479 cAMP-dependent signaling pathways 267 Carbachol (CCh) 329 CART-immunoreactive mossy cells 31–32 Catecholaminergic brainstem-dentate connections 71–74 Serotonergic afferents 71–73 CB1 receptors 319–321 Endogenous ligands (arachidonoylethanolamide) 321 Expression and distribution 322, 324–326 Coexpression with other receptor subtypes 326

777

Immunohistochemical studies 323 Role in GABAergic neurons 322–324 CB2 receptors 319–321 C-fos protein 604, 769 cGMP-dependent protein kinase (PKG) 479 Chandelier cells 155, 163 Chemokines 515–516 Cholecystokinin (CCK) 30, 101, 115, 201, 225 Choline acetyltransferase (ChAT) 65–66, 138–139 Cholinergic neurons 18–19, 63–78, 133–140 Chronic stress effects in dentate gyrus 362–367 Code conversion 582–585 Commissural/association (C/A) fibers 63–78 Conjunctive encoding 567–568 Cornu ammonis 594, 724 Cortical hem in DG and hippocampus development 144 Corticosteroids, in DG 355–367 Dentate function after chronic stress 362–367 Dentate function in the absence of 357–361 Cell turnover and morphology 357–359 Physiology 359–361 Functional relevance of corticosteroid effects in DG 365–367 Dose dependence 365–366 in health and disease 366–367 Physiological variations in corticosteroid levels 361–362 Cell turnover 361 Physiology 361–362 Prolonged exposure, damage due to 355 Systems activated by stress 355–357 Corticotrophin 311 Cortistatin (CST) 265 Crossed entorhinodentate fibers 507, 510 Cycloheximide 769 Cytokines 339 (see also Pro-inflammatory cytokines) Dehydroepiandrosteronesulfate (DHEAS) 401 Dendritic changes, following entorhinal denervation 514–516 (see also Transneuronal changes) Dense-core vesicles 251 Densitometry 671 Dentate gyrus (DG) (see also Hippocampal DG) Aging-related changes in 683–685

Anatomy and architecture 3–22, 418–421 as a filter or gate 601–611 Dentate gate vs. filter 605–607 Dentate filter function 607–608 Behavioral analysis CA3 backprojection to (see CA3) Cell and fiber layers development in 133–140 Cortical signals transformation in Extrinsic afferent systems to 63–78 (see also Extrinsic afferent systems to DG) Interneurons of 217–229 Modeling 639–655 Neurotrophins in 371–387 Norepinephrine and 299–312 Opioid systems in 245–258 Plastic processes in 417–442 (see also Plasticity) Pro-inflammatory cytokines and their effects in 339–350 (see also Pro-inflammatory cytokines) Depolarization-induced suppression of inhibition (DSI) 319, 327–329 Depression 464–465 NPY and its receptors in 293–294 Detonator synapses 111, 422, 433 Development of dentate gyrus 133–140 (see also Reelin) Afferent fiber 136–139 Functional considerations 139–140 Granule cell generation 134–136 Diacylglycerol lipase (DAGL) 321 Diffusion tensor imaging (DTI) 750 DiI-labeled granule neurons 170 Dopamine 63–78, 461 Dormant basket cell hypothesis of epilepsy 204 Dorsal raphe nucleus (DRN) 703 Dorsolateral (DLE) entorhinal cortex 46 Doublecortin (DCX) 157 Dual-labeling immunohistochemical techniques 323 Dynorphins 113, 245–258 Anatomical distribution of 249–250 Immunoreactivity 256 Location in hippocampal formation 248 Prodynorphin-derived opioids 246 a-Neo-endorphin 246 b-Neo-endorphin 246 Dynorphin A (1–8) 246 Dynorphin A (1–17) 246

778

Dynorphin B (1–13) 246 [Leu5]-enkephalin 246 Leumorphin 246 Electroconvulsive therapy (ECT) 293 Embryonic marginal zone 593 Emx2 gene 146 Endocannabinoids (ECs) in DG 319–332 CB1 receptors 320–321 (see also separate entry) CB2 receptor 320–321 Markers of EC system 322–327 Modulating glutamatergic transmission in DG 329–330 Non-CB receptor targets of 330–331 Physiological role in dentate 327–331 Role in neurogenesis 330 Enkephalins 113, 245–258 Anatomical distribution of 247–249 Location in hippocampal formation 248 Proenkephalin-derived opioids 246 C-terminally extended forms of [Met5]enkephalin 246 [Leu5]-enkephalin 246 [Met5]-enkephalin 246 Entorhinal afferents 136–137 Entorhinal cortex 3, 45–46 Entorhinal cortex lesions (ECL) 503–519 Projection to DG 17–18 Subdivisions 46 Entorhinal denervation, structural reorganization of DG after 501–519 Acetylcholinesterase (AChE) reorganization 509–510 Axonal reorganization 505–510 Commissural/associational (C/A) mossy cell axons 505, 508 Crossed entorhinodentate fibers, reorganization 510 Dendritic changes 514–516 (see also Transneuronal changes) Fiber systems reorganization 505–510 GABAergic C/A projection 509 Glial changes 511–514 Immune response 518 Neuroanatomical species differences 503–505 Entorhinal-dentate projection 43–58 (see also Perforant path) Epilepsy 371 (see also Epileptic DG)

Adult neurogenesis in 529–537 Epilepsy models, neurotrophins roles in 379–381 NPY and its receptors in 292–293 Epileptic DG Human dentate characteristics of 185 Mossy fiber sprouting in 541–558 (see also Mossy fiber sprouting) Epileptogenesis 650–652, 755–770 Definition 756–757 in DG 757–760 Filtering and gating properties 759–760 Propagation of seizures 760 TLE, animal models 760–762 (see also Temporal lobe epilepsy) Histopathological alterations 761–762 Kindling 761 Status epilepticus 761 Epileptic animals Pharmacology and subunit expression alterations 240 Synaptic GABAA receptor expression Changes in 240 Upregulation in 240 Tonic GABAA current 241 Zinc sensitivity 240 Estradiol 402 17b-Estradiol 401, 405 Estradiol treatment 256 Estrogen receptor (ER) 399, 402 Effect on seizures 405 ERa distribution in DG 406–408 ERb distribution in DG 408 Subcellular localization 407 Estrous cycle 402 Eukaryotic elongation factor 2 (eEF2) 458–459 Experimental autoimmune encephalomyelitis (EAE) 342 Extended preparation 44 Extracellular signal-regulated kinase (ERK) 480 Extracellular unit recording 304–305 Extrasynaptic GABAA receptors 235 Alterations in animal models of epilepsy 240 (see also Epileptic animals) Altered zinc sensitivity 240 Composition and function of 236 Expressed by dentate granule cells 236–237 Upregulation in granule cells of epileptic animals 240

779

Extrinsic afferent systems to DG 17–19, 63–78 Afferent projections 17 Basal forebrain inputs 18–19 BNPI/VGLUT1 75–76 Brainstem inputs 19 CA1–CA3 subfields 64 Catecholaminergic brainstem-dentate connections 71–74 Commissural connections 74–75 DNPI/VGLUT2 76–77 Entorhinal cortex projection 17–18 Glutamatergic innervation 75 Noradrenergic and dopaminergic afferents 73–74 Septo-hippocampal connections 65–69 Supramammillary input 19, 69–71 VGLUT3 75, 77–78 Fadrozole 405 Fatty acid amide hydrolase (FAAH) 321 Feedforward interneurons 304–305 Field excitatory postsynaptic potential (fEPSP) 359–360 Filopodia 167–180 Filtering function, dentate-hilar 607–608 Fimbria stimulation 630 Flinder’s Sensitive Line (FSL) 293 Fluoxetine 293, 708 Fornix 618 Fractional anisotropy (FA) 750 Functional cellular connectivity 202–203 Functional significance of laminated organization 139–140 Fusiform cells (spiny and aspiny) in hilus 155–164 g-EEG 299–312 g-Frequency oscillation 207 g-Oscillations 621 GABA (g-aminobutyric acid) neurons/GABAergic neurons 217, 603–604 Axonal arborizations and postsynaptic targets of 221–223 Distribution of 218–221 GABAergic afferents from medial septum 138–139 GABAergic C/A projection after entorhinal denervation 509

GABAergic commissural projection 74–75 GABAergic inhibition 548–551 in human DGCs 190 GABAergic interneurons 14–16, 63–78, 164 GABAergic neurons 322–324, 582, 587, 770 Septohippocampal cholinergic systems and 68–69 Transporters, co-localization 118–119 GABA-mediated inhibition (see also Extrasynaptic GABAA receptors) Functional regulation of DG 235–241 Gatekeeper function 237–240 Tonic and phasic inhibition in dentate granule cells 237 Genetic regulation of DG development 143–150 Cortical hem 144 Developmental signaling systems in adult DG 150 Emx2 146 Lhx5 146 Mutants with defects in development 148 Neurogenic niche 147–148 Transcription factors 144–146 Glia Glia-guided neuronal migration 134 Role of glia in synaptic remodeling 404 Glial changes, following entorhinal denervation 511–514 Astroglia 513–514 Blood-derived cells 513 Microglia 511–513 Oligodendroglia and NG2-positive cells 514 Glial fibrillary acid protein (GFAP) 35, 136, 404 Global remapping 299, 308, 312, 592 Glucocorticoid receptor (GR) 355, 357 Glutamate 113 Glutamatergic innervation of DG 63–78 Glutamatergic transmission in DG Transport, co-localization 118–119 Glutamic acid decarboxylase (GAD) 65, 77, 217–229, 641, 765 Double labeling 220 GAD expression 116–117 GAD mRNA-labeled neurons distribution 218 GAD65 219 Glycogen metabolism 299–312 Gonadal hormones 256, 399

780

Granule cells, dentate (DGCs) 8–11, 643–644, 664–666 Apical dendrites of 156 Biophysical properties of 664–665 Cell layer 5 Comparative anatomy 24–28 Co-release of glutamate and GABA 120–121 Density, regulation 148–150 Description 155–157 Glutamate as transmitter for 11 Golgi-impregnated granule cells 27 Granule cell association hypothesis 203 Granule cell generation 133–140 Compact granule cell layer formation 134–136 Sites of 134 in primates 26 Mossy cell 13–14 Mossy fibers 9–10 NPY actions 289 Pyramidal basket cell 11–13 Single unit recordings 665–666 Tonic and phasic inhibition 237 Granule neuron dendrites Development of 168–173 Maturation 173–176 Morphological development and maturation of 167–180 Primary period of 167 Temporal and spatial gradients of 168 HCN current 186–187 Hebbian plasticity 434–435 Hebb-Williams maze 571–574 Heterosynaptic LTD 437 Hilar cells with axonal projections to the perforant path (HIPP) cells 15–16, 222, 583–584, 645 Hilar commissural-association pathway-related (HICAP) cells 56, 164, 223, 645–646 Hilar interneurons 303–304 NPY effects on 289–290 Hilar mossy cells 199–213, 630 Anatomy 200–202 Basic properties 200 b/g-Oscillations 208–210 y-Oscillations 207 g-Frequency oscillation 207 Cellular properties of 203–204

Connectivity of 202–203 Firing patterns 205 Functional identification and activity in vivo 199–213 Slow oscillations of 210–213 Hilar somatotatin (SST) interneurons 269–270 Excitotoxicity 273–274 Hippocampal information processing 633–634 Hippocampal interneurons 14 Hippocampal mossy fibers and terminals Supra- and infrapyramidal bands of 87–90 Human DG 23–28 Granule cells, physiological studies of 183–195 Huntington’s disease 385 Hypothalamic-pituitary-adenocortical (HPA) system 705–707 IL-1 receptor-associated kinase (IRAK) 346 Infrapyramidal 134, 87–90, 134 Inhibitory postsynaptic potential (IPSP) 111 Interleukin 18 (IL-18) 349–350 Interleukin 1-b (IL-1-b) Action in DG 345–350 Ca2+ channels 348–349 IL-1 signalling 346 Synaptic plasticity 346–348 Interneurons of DG 14, 110, 157, 217–229 Classification 217 GABA as primary neurotransmitter 217–229 (see also GABA) GABAergic interneurons 14–16 Intracellular labeling of 217–229 Neurochemical identity 223–228 Intracellular labeling techniques 174, 217–229 k-Opioid receptors (KORs) 246–247 Immunoreactivity 254 Kainate model 405 Kainic acid 118, 761, 405 Lateral perforant path (LPP) 299–300, 307–308, 426, 462–463, 493–494 Learning and memory, BDNF effect on 378 Lef1 mutants 147 Lesioning techniques 503 Lesion-induced sprouting 85–102 Leumorphin 246 Lhx5 gene 146

781

LIM domain kinase (LIMK) gene 455 Locus coeruleus (LC) 300, 305–306 Activation 305 Glutamatergic activation of 304 Modulation of NE 306–311 Stimulation 304–307 Long-term depression (LTD) 417, 436–438, 453, 473–488 Associative LTD 437, 438 Depotentiation related to 437–438 Heterosynaptic LTD 429, 437, 493–494 Homosynaptic LTD 438 Induction, cellular mechanisms, comparison 473–494 in sparse coding 438 mGluR-induced LTD 489–491 Long-term potentiation (LTP) 245–258, 417, 423–436, 578, 590 Long-term memory (LTM) 616 Low-frequency stimulation (LFS) 111 LRP family (LRP6) 146 Lucifer yellow 733–734 Major depressive disorder (MDD) 697 (see also Depression) Genes, environment 703–707 5-HT system and hippocampal neurogenesis 703–705 Neurogenesis, stress effects on 705–707 Hippocampal dysfunction and atrophy 701–702 Mass-associated TLE (MaTLE) 184 Maximal dentate activation (MDA) 603 Medial (ME) entorhinal cortex 46, 591–592, 619 Medial perforant path (MPP) 299–301, 305, 307–308, 426 Medial septum/diagonal band of Broca (MSDB) 65 MSDB cholinergic innervation 65–67 MSDB GABAergic innervation 67–68 Medial temporal lobe sclerosis (MTS) 183–184, 189 Median raphe 63–78 Metabotropic glutamate receptors (mGluRs) 193 Metaplasticity 438–441 Methylazoxymethanolacetate (MAM) 89, 712 mGluR5 receptor 345, 349 mGluR-dependent LTD 482, 489–491

Microglia 511–513 MicroRNAs (miRNAs) 459–460 Mild cognitive impairment (MCI) 742, 748–749 Mineralocorticoid receptor (MR) 355–356 Mini mental state examination (MMSE) 732–733, 744 Modeling DG 639–655 Molecular layer perforant path-associated cells (MOPP) 15, 56, 223 Monoglyceride lipase (MGL) 321 Mossy cells 13–14, 159–161, 644 (see also Hilar mossy cells) Mossy fiber sprouting in epileptic DG 541–558 (see also Seizure-induced mossy fiber sprouting) Mossy fibers (MFs), dentate 9–10, 58–102, 568 Anatomy 110–111 Development and cognitive function Development 97–98 Genetics and breeding experiments 99 in CA3 (regio inferior) Intra- and supragranular mossy fiber collaterals 94–95 Mossy fiber LTP 432 Mossy fiber sprouting 183–195 Plasticity of 431–436 Projections to CA1 (regio superior) 91–92 Projections to CA3 10 Projections to the hilus 9–10, 92, 95–96 Sprouting 99–100, 765–769 Structural plasticity 85–102 Supra- and infrapyramidal 87–90 Synaptic transmission 109–123 Terminals 287–289, 581–582 Tetanic stimulation 590 Timm staining 93–94 Trajectory and termination 86–97 Transverse and longitudinal trajectories of 90–91 Visualization 86 Zinc in 33 a-Neo-endorphin 246 b-Neo-endorphin 246 Na+ currents 186 Naloxone 568 NE (see Norepinephrine) Neonates 167–180

782

Neprilysin 268 Nerve growth factor (NGF) 115, 371–373 Neuroanatomical organization, DG 3–22 Cell types and their connectivity 8–19 (see also Granule cells; Mossy cells; Pyramidal basket cells) Comparative neuroanatomy 6–8 among rat, monkey, humans 6–8 Extrinsic inputs 17–19 Layers of, 5 Granule cell layer 5 (see also granule cells, dentate) Molecular layer 5 Polymorphic layer (hilus) 5, 9–10 Long-spined cell 16 Nissl and Timm’s staining 6 Septotemporal axis 5 Neurochemical identity, of interneurons 223–228 Calretinin 226–227 Cholecystokinin (CCK) 225 Neuropeptide Y (NPY) 227–228 Parvalbumin (PV)-labeled neurons 224–225 Somatostatin 225–226 NeuroD mutant 149–150 Neurodegenerative diseases, BDNF role in 385 Neurogenesis 143–144, 355, 464, 662, 697–715 and antidepressants 707 and MDD (see Major depressive disorder) and sex steroids 402–403 Antidepressant treatments, neurogenic effects of 707–708 BDNF effect on 378 Endocannabinoid (EC) role in 330 NPY effects on 291–292 Opioids and 256–257 Preclinical studies 710–713 Antidepressant (AD) drugs, behavioral effects of 712–713 Learning, hippocampal neurogenesis in 711–712 Neurogenesis blockade and depression symptoms 711 Stress based depression paradigms 710–711 Serotonin-dependent ADs and hippocampal neurogenesis 708–710 Neurokinin B 115

Neuronal cell death, TNF-a and 341 Neuronal migration 134 Neuronal network, connectivity matrix for 642 Neurons of DG 14–17 Axo-axonic cell 15 Long-spined cell 16 of the molecular layer 15 Neurons of polymorphic cell layer 15–17 Neuropeptide Y (NPY) in DG 115, 227–228, 285–294, 686 (see also Y1 receptors; Y2 receptors; Y4 receptors; Y5 receptors) Effects on hilar interneurons 289–290 Effects on LTP 290 Effects on neurogenesis 291–292 Effects on synaptosomes 290–291 Electrophysiological effects of 287–294 in the fascia dentata 286–287 in disease 292–294 Alzheimer’s disease (AD) 292–293 Depression 293–294 Epilepsy 292–293 Neurons containing NPY and their synaptic contacts 286 Neuropil 702 Neuroprotection 399 Neuropsychiatric disease, BDNF role in 385–386 Neurotransmitter transport of human dentate granule cells 191–193 Neurotrophins (NT) in DG 371–387 Brain-derived neurotrophic factor (BDNF) (see also separate entry) Nerve growth factor (NGF) 372–373 Neurotrophin-3 (NT-3) 371, 378–379 Neurotrophin-4/5 371, 379 Role in epilepsy models 379–381 (see also Epilepsy) Seizure regulation of 379 Signal transduction 372 Structure 371–372 NF-kB 340, 342 Nissl staining 762 Nissl and Timm’s staining 6 NMDAR/mGluR-dependent LTD 474–481 Ca2+/calmodulin-dependent protein kinase (CaMKII) 478 Ca2+/phospholipid-dependent protein kinase (PKC) 477

783

cAMP-dependent protein kinase (PKA) 478–479 Cascades involved with 475 cGMP-dependent protein kinase (PKG) 479 Expression 476 Extracellular signal-regulated kinase (ERK) 480 Induction 475 Kinase pathways 477 Phosphatase pathways 476 Phospholipase A2 480 Presynaptic vesicular release 479–480 NMDAR-dependent LTD 481–482 mGluR-dependent LTD 482 (see also separate entry) N-methyl-D-aspartate receptors (NMDAR) 119, 175, 401, 474–488, 567 (see also NMDARdependent LTD; NMDAR/mGluRdependent LTD) Non-hippocampal afferent connections 33 Non-human primates, 23–38 Non-neuronal cells 511 Non-self/self duality 619–620 Norepinephrine (NE) 299–312, 461 Adrenoceptor distribution 300–302 (see also Adrenoceptors) Hypotheses 299 Locus coeruleus(LC)-NE modulation and environmental events 310–312 NE innervation 300 NE-induced plasticity 306–310 LC firing and NE release patterns 309–310 Pathway selectivity 307–308 Physiology 302–306 EEG recording 305 Evoked potential recording 305–306 Extracellular unit recording 304–305 Glycogen metabolism 302–303 Hilar interneurons 303–304 Intracellular recording 303 Release modulation by glutamate and vice-versa 310 Normal aging, hippocampal granule cells in 661–674 Aged granule cells Biophysical properties of 664–665 Cholinergic responses in vitro in 668 Single unit recordings in granule cells 665–666 DG, vulnerability of 672–673 Granule cell dendrites 664

Granule cell neurogenesis 662–663 Granule cell numbers 662 Granule cell synaptic contacts, changes in 663–664 Immediate early genes 670–672 In situ hybridization 672 Perforant path-granule cell synapse 666–670 Novel environment 311 Nuclear translocation 134 d-Opioid receptors (DORs) 246–247, 251–252 Opioid physiology 254–255 Subcellular locations of 249 m-Opioid receptors (MORs) 246–247, 252–253 Offline behavioral states 579 Oligodendrocytes 158, 514 Opioid systems in DG 245–258 b-Endorphin 246 Anatomical distribution of 247–251 Endomorphin 246 Gonadal steroids and 256 Neurogenesis and 256–257 Opioid peptides 246 (see also Dynorphins; Enkephalins) Release 251 Role in mossy fiber LTP 432–433 Opioid physiology 254–255 Opioid receptors 246–247 ORL1 247 Prenatal morphine and 257 Seizures and 255–256 Subcellular locations of 249 p55 TNFR 340, 342 Pain, BDNF role in 384–385 Parachlorophenylalanine (PCPA) 703–704 Paradoxical TLE (PTLE) 184 Parahippocampal region 43–58 Parvalbumin (PV)-labeled neurons 224–225, 514–515 Parvalbumin-immunostained basket cells in human DG 37 Pathway selectivity, in NE-induced plasticity 307–308 Pediatric epilepsy 756 Perforant path 17, 43–58 Contralateral entorhinal-dentate projection 43–58

784

Longitudinal organization 43, 54–58 Nomenclature 45–46 Radial organization in 47–54 Synaptic organization of 56–57 Transection 503–519 in mice 503 Transverse organization in 51, 53 Phaseolus vulgaris-leucoagglutinin (PHAL) tracing 508, 510 Phenylchloromethamphetamine (PCA) 704 Phosphatase pathways 476 Phospholipase A2 480 Physiological studies of human dentate granule cells 183–195 GABAergic inhibition 190 Membrane properties 184–187 Neuromodulation 193–195 (see also Neuromodulation) Synaptic properties 187–193 Picrotoxin 238 Pilocarpine 761 Plaques 727 Plasticity 555–556 at mossy fiber (MF) CA3 synapse 121–122 Excitatory and inhibitory circuits 555–556 in somatostain (SST) receptors expression 272–273 in the DG 417–442 Metaplasticity 438–441 of MFs dentate mossy fibers 85–102 Neural grafting, testing MF growth and 100–101 Sex steroids and 402–405 Synaptic plasticity 423–436 Polymorphic layer (hilus) 5, 9–10 Mossy fibers 9–10 Neurons of 15–17 Postmitotic neurons 530 Postsynaptic targets of GABA neurons 221–223 Pregnenolone 405 Prenatal morphine 257 Presynaptic mechanisms in synaptic consolidation 462 Presynaptic vesicular release 479–480 Progenitor cells 529–537 Progesterone (PR) 399, 402 3a,5a-tetrahyroprogesterone 405–406 Distribution in DG 409

Subcellular localization of 407 Pro-inflammatory cytokines, and their effects in DG 339–350 (see also Interleukin 1-b; Interleukin 18; Tumour necrosis factor-a) in CNS, distribution 340 Projection cells Ultrastructure and synaptic connectivity of 155–164 Pyramidal basket cell, dentate 11–13, 158, 160 Radial glial cells in the adult DG 157 Radial glial scaffold 135 Radial organization in perforant path 47–52, 52–54 Rate remapping 592 Receptor tyrosine kinases (RTKs) 372 Recurrent excitation, in DG 551–558 Circuit formation 554 Emergent property 557 Functional effect 558 Inhibition state 551–553 Reelin 133–140, 148, 536 Effect on granule cell migration 135 Effect on granule cell orientation 135 Effect on radial glial scaffold 135 Reinnervation 507 REM (rapid eye movement) 602 Rett syndrome, BDNF role in 385 Schaffer collateral axons 473–494, 628 Seizure disorders 689–690 Hippocampal-dependent seizures 690 Seizure regulation of BDNF expression 376 of NGF expression 373 of trk receptor expression 374 trk receptor activation following 381–384 Seizures and hippocampal neurogenesis 533–534 Caudal subventricular zone (SVZ) 535–536 Functional significance 534–536 Mechanisms 536–537 Temporal lobe epilepsy (TLE) models 533 Seizure-induced mossy fiber sprouting 541–558 Selective serotonin reuptake inhibitor (SSRI) 708–709 Septal fibers 18–19 Septo-hippocampal (SH) connections 65–69

785

GABAergic systems and, interactions between 68–69 Medial septum/diagonal band (MSDB) cholinergic innervation of DG 65–67 MSDB GABAergic innervation of DG 67–68 Septotemporal axis 5 Serotonin 63–78, 403, 461, 532 Serotonin-dependent antidepressants 708 Sex steroids and dentate physiology 400–402 and dentate plasticity 402–405 (see also Plasticity) and neurogenesis 402–403 and synaptic remodeling 403–405 and the DG 399–410 Effect, mediation of 406–410 ERa distribution in DG 406–408 ERb distribution in DG 408 Local steroid synthesis, role of 409–410 PR distribution in DG 409 Electrophysiological studies 401 in adulthood 401 Neuroprotective effects 405–406 Subcortical mediation of sex steroid effects 410 Sharp waves-ripple complexes (SPW) 207 Slow oscillations, of hilar mossy cells 210–213 Somatostatin (SST) in DG 225–226, 265–278, 686 Activity-dependent expression of 270–272 as a subset of hilar interneurons 269–270 Cortistatin (CST) and 265 Distribution 266 Electrophysiological effects of 274–277 in hippocampus 270 Antiepileptic properties in 271 Inhibiting LTP 275 K+ channels activated by 266 Levels in Alzheimer’s disease 268–269 SST receptor expression in DG 272 Plasticity in 272–273 Synaptic potentiation reduction by 265 Spatial information encoding vs. retrieval 572–574 Spatial learning 689 Spikes, dentate 586 Potentiation and pairing requirements 308

in anesthetized rats 306–307 In vitro 307 Spinophilin 403–404 Spoiled gradient recalled (SPGR) pulse 742 Sprouting 507, 534, 541–558 (see also Mossy fiber sprouting) Cholinergic 509 Commissural/associational fibers 507–508 Fiber systems 507 Reinnervation process 507 Mossy fibers (MFs), lesion-induced 99–100 Species difference in 517–518 Status epilepticus (SE) 533 Stem cell 529–530 Stimulus/response couplet 616–617 Stimulus-induced mGluR-dependent LTD 488 Stress Exposure of rat to 356 Systems activated by 355–357 Subcortical and commissural afferents 63–78 Subiculum 43–58 Supragranular mossy fibers 94–95, 156 Supramamillo-dentate connections (SUM) 69–71 Supramammillary area (SUM) 63–78 Suprapyramidal bands of DG 134 of hippocampal mossy fibers and terminals 87–90 Synapse 505 Synaptic consolidation 453–465 (see also Brain-derived neurotrophic factor) Extrinsic modulation of 461–462 Functions and clinical implications of 463–465 Alzheimer’s disease 463–464 Depression 464–465 Neurogenesis 464 Lateral perforant path (LPP) 462–463 Mechanisms, functions, and therapeutic implications 453–465 Presynaptic mechanisms 462 Synaptic consolidation 453 Translocation 460 Synaptic plasticity in DG 299, 311–312, 423–436 in hippocampus, forms LTD 344 LTP 344 IL-1b and 346–348 TNF-a and 343–345

786

Synaptic potentiation and spike potentiation in vitro 307 Synaptic remodeling 399 Glia in 404 Sex steroids and 403–405 Synaptic transmission and TNF-a 343 BDNF effects on 377–378 Synaptic transmission, MFs CA3-interneuron synapses 123 Communication from DG to CA3 area 109–123 ‘Detonator synapse’ 111 GABA transporters, co-localization 118–119 Glutamate transporters, co-localization 118–119 Long-term plasticity at 121–122 Multiple neurotransmitters/modulators in 113–117 Pre- and post-synaptic constituents 114, 119–120 Presynaptic modulation 117–118 Transmission at MF-CA3 pyramidal neuron synapse, quantal nature of 113 Transmitter release, properties of 112–113 Voltage clamp methods to the study 111 Synaptodendrosome (SD) 458 Synaptophysin immunohistochemistry 687 y-EEG 303, 311 y-Oscillations 621 Tau 724 Temporal lobe epilepsy (TLE) 183, 204, 240–241, 533–537, 755 Animal models of 183–195 Dormant basket cell hypothesis 762–764 GABAergic neurons, reorganization of 770 Hilar damage 762 Interneurons, loss of 764–765 Mass-associated TLE (MaTLE) 184 Mossy fiber sprouting 765–769 Axon sprouting 765–768 Recurrent excitation vs inhibition 768 Seizure susceptibility 768–769 Paradoxical TLE (PTLE) 184 Pilocarpine model 534–536 Tetrodotoxin (TTX) 454–455 Thorny excrescences 28

Tianeptine 708 Timm staining 18, 760, 765–766 and species differences, in mossy fibers 93–94 of DG 25 Nissl and Timm’s staining 6 Timm stained hippocampal sections of cat 88 of dog 88 of European hedgehog 88 of guinea pig 88 of man 88 of rat 88 T-maze 622 TNF receptors (TNFR) 340 kB (NF-kB) 340 p55 TNFR 340 Topographical organization 43–58 Tracing techniques 508 Transcription factors, patterning the cortex and hippocampus 144–146 Transneuronal changes Cellular mechanisms in 515 Chemokines role in 515–516 Degeneration 512–516 in granule cells 514–515 in parvalbumin-positive neurons 514–515 Transplants 85–102 Tri-synaptic circuit/pathway 63, 109, 417 Tropomyosinrelated kinase (trk) receptors 372, 381–384, 460 Tumour necrosis factor (TNF) 339–350 (see also TNF receptors) Vasopressin 311 Very low-density lipoprotein receptor (VLDLR) 135, 148 Vesicular glutamate transporters (VGLUT), 75–76 BNPI/VGLUT1 75–76 DNPI/VGLUT2 76–77 VGLUT1, 75–76, 325 VGLUT2, 75 VGLUT3, 75, 77–78 Voltage sensitive dye imaging techniques 238 Wnt signaling pathway 146–148

787

Y1 Y2 Y4 Y5

receptors receptors receptors receptors

285–294 285–294 286 286

Z scores 749 Zinc (Zn2+) in mossy fibers 33 Zinc-sensitive synaptic GABAA receptors 240–241

This page intentionally left blank

Plate 1.1. The rat hippocampal formation. (A) Nissl-stained horizontal section through the hippocampal formation of the rat. The major fields are indicated. Projections (1) originate from layer II of the entorhinal cortex (EC) and terminate in the molecular layer of the dentate gyrus (DG) and in the stratum lacunosum-moleculare of the CA3 field of the hippocampus. An additional component of the perforant path originates in layer III and terminates in the CA1 field of the hippocampus and the subiculum. Granule cells of the DG give rise to the mossy fibers (2) that terminate both within the polymorphic layer of the DG and within stratum lucidum of the CA3 field of the hippocampus. The CA3 field, in turn, gives rise to the Schaffer collaterals (3) that innervate the CA1 field of the hippocampus. Pyramidal cells in CA1 project to the subiculum and to the deep layers of the EC. The subiculum also gives rise to projections to the deep layers of the EC. (B) Line drawing to illustrate the major regional and laminar organization of the DG. The DG is divided into a molecular layer (ml) a granule cell layer (gcl) and a polymorphic layer (pl). The molecular layer is divided into three sublayers based on the laminar organization of inputs. The hippocampus is divided into CA3, CA2 and CA1 subfields. Within CA3, a number of layers are defined. The main cell layer is the pyramidal cell layer (pcl). Deep to the pyramidal cell layer is the stratum oriens (so); deep to this is the white matter of the alveus (al). Superficial to the pyramidal cell layer is stratum lucidum (sl), stratum radiatum (sr) and stratum lacunosum-moleculare (sl-m). Fields CA2 and CA1 have the same layers as CA3 except for stratum lucidum. The remainder of the hippocampal formation is made up of the subiculum (Sub), presubiculum (Pre), parasubiculum (Para) and EC. Layers of these latter structures are indicated with roman numerals. Additional abbreviations: ab, angular bundle; fi, fimbria; hf, hippocampal fissure. (C) Schematic illustration of DG and hippocampus to illustrate position of suprapyramidal blade, infrapyramidal blade and crest of the DG. This model is used in subsequent illustrations to demonstrate the major cell types and connections of the DG. (For B/W version, see page 4 in the volume.)

Plate 1.3. Horizontal sections through the rat hippocampal formation. This figure illustrates a more dorsally situated (A) and a more ventrally situated (B) horizontal section through the rat hippocampal formation. The approximate level of the section is illustrated on a 3D reconstruction of a magnetic resonance image series of the rat brain. Subtle differences in the cytoarchitectonic organization are seen throughout the hippocampal formation. The dentate gyrus, takes on a more ‘‘V’’ shape dorsally and a more ‘‘U’’ shape ventrally. Calibration bar ¼ 250 mm. (For B/W version, see page 7 in the volume.)

Plate 1.4. The dentate granule cell. The characteristic features of the dentate granule cell are illustrated, including its axonal arbor. A collateral plexus gives rise to numerous (
200) typical synapses on cells located within the polymorphic layer. Most of these synapses are onto the dendrites of inhibitory interneurons. Some of the large mossy fiber expansions are also distributed in the polymorphic layer. Many of these terminate on the proximal dendrites of mossy cells. The mossy fiber axons ultimately enter the CA3 field where they travel through the full transverse extent of the field. On their course, they terminate with mossy fiber expansions on a small number (15–20) of CA3 pyramidal cells. Additional abbreviations: gc, granule cell; pc, pyramidal cell. (For B/W version, see page 8 in the volume.)

Plate 1.6. The pyramidal basket cell. The cell body of the pyramidal basket cell is located at the interface between the granule cell layer and the polymorphic cell layer. The axon (arrow) emerges from the apical dendrite. Collaterals of this axon form a curtain of terminals that synapse with the granule cell bodies. Additional abbreviations: pbc, pyramidal basket cell. (For B/W version, see page 12 in the volume.)

Plate 1.8. The mossy cell. A line drawing of a mossy cell (mc) in the polymorphic layer. The axon (arrow) develops a plexus within the polymorphic layer, and also an ipsilateral projection to the inner molecular layer, known as the associational pathway. The main axon also projects contralaterally to the inner molecular layer, forming the commissural pathway. The ipsilateral projection increases in density with distance from the cell body of origin. (For B/W version, see page 13 in the volume.)

Plate 2.1. Photomicrographs of Timm-stained sections of the dentate gyrus (A and B) and Ammon’s horn (C) in adult rat. Timmstained fibers are evenly distributed throughout the hilus (A), whereas they form a bundle in stratum lucidum of the CA3 area (A and C). Timm-stained puncta outline dendrites inside the granule cell layer (g) in B. Timm-stained mossy fibers terminate at the border of CA3/CA2 areas (arrow), but a few Timm-labeled terminals (curved open arrows) are in the pyramidal layer (p) of CA2 area. There are no Timm-labeled terminals in the CA1 area. Double arrows mark the border between the CA2 and CA1 areas (C). Scale bar ¼ 100 mm for A, 20 mm for B and C. (For B/W version, see page 25 in the volume.)

Plate 2.2. Photomicrographs of Timm-stained sections of the dentate gyrus (A and B) and Ammon’s horn (C and D) in adult rhesus monkey. The hilus is outlined by the Timm-stained fibers (A). Timm-stained puncta outline dendrites (arrows) in the granule cell layer (g), and a few puncta (small arrows) are always found in the molecular layer (B). Timm-stained fibers occupy not only stratum lucidum (lm) but also the entire width of the pyramidal layer (p) of the CA3 area (C). In the pyramidal layer (p) and in stratum lucidum (lm) Timm-stained puncta clearly outline dendrites (C). In stratum oriens (o) some puncta form clusters (arrows) (D). Scale bar ¼ 200 mm for A, 20 mm for B, 100 mm for C, 20 mm for D. (For B/W version, see page 26 in the volume.)

Plate 2.3. Photomicrographs of Golgi-impregnated granule cells without basal dendrites (A and D) and with basal dendrites (B, C and D) in the dentate gyrus of adult rhesus monkey. Some basal dendrites (arrows) branch close to the cell body (B), and others extend into the hilus (arrows on C and white arrows on D). Morphology of dendritic tree of conventional granule cells differs depending on the location of the soma within the granule cell layer. Those cells at the hilar border (h) have one main dendrite and narrower dendritic tree (open curved arrow on D), whereas those at the border with the molecular layer have multiple dendrites that arise from the soma (open arrowhead on D). Basal dendrites are similarly covered by spines as the apical dendrites (C). Axons usually originate from the basal pole of the soma (A and D). Two Golgi-impregnated pyramidal cells in the pyramidal layer (p) of the CA3 area display thorny excrescences (arrows) on both the apical and basal dendrites (E). Scale bar ¼ 20 mm. (For B/W version, see page 27 in the volume.)

Plate 2.4. Photomicrographs of Golgi-impregnated mossy cells in the hilus of the dentate gyrus of adult rhesus monkey (A) and humans (B). Thorny excrescences are considerably larger on the human mossy cells than in rats and monkey. In primates, dendrites of many mossy cells penetrate the molecular layer (m) (A). The dendrite (arrows) is covered by spines in the molecular layer (m), but the segment inside the granule cell layer (g) is practically spine free (C). The proximal dendrites and soma (arrow) of the hilar mossy cell is covered by thorny excrescences (D). Scale bar ¼ 100 mm for A, 50 mm for B, 25 mm for C and D. (For B/W version, see page 29 in the volume.)

Plate 2.5. CART-immunoreactive mossy cells and their axons in the dentate gyrus of adult rhesus monkey. Somata of mossy cells are located in the hilus (h), whereas their axons form the associational pathway and project mainly to inner third of the molecular layer (m) of the dentate gyrus (A). Somata of mossy cells at higher magnification (B). CART-positive puncta inside the granule cell layer (g) shows the ascending axons from the hilus to the molecular layer (B). Large magnification of CART-positive mossy cell displays thorny excrescences (arrowheads) in the monkey dentate gyrus (C). Scale bar ¼ 250 mm for A, 50 mm for B, 10 mm for C. (For B/W version, see page 31 in the volume.)

Plate 2.6. High magnification photomicrographs of a CART-immunostained mossy cell displaying large and complex thorny excrescences in the hilus of the human dentate gyrus (A and B). Scale bar ¼ 10 mm for A and B. (For B/W version, see page 32 in the volume.)

Plate 2.7. Photomicrographs of Golgi-impregnated basket cells in the granule cell layer (g) of the dentate gyrus in newborn (A and B) and 1month-old (C) rhesus monkey. Soma of a fusiform type basket cell located in the granule cell layer with an axon (arrow) originating from the main apical dendrite (A). High magnification photomicrograph of the basket cell shows the point of origin of the axon (arrows), as well as the well-developed axonal branches among the granule cells (B). A multipolar basket cell at the hilar border (h) has dendrites that extend into the molecular layer, and an axon (arrow) with many branches that ramify within the granule cell layer. Scale bar ¼ 20 mm for A, B, C. (For B/W version, see page 36 in the volume.)

Plate 2.8. Photomicrographs of parvalbumin-immunostained basket cells and their axons in the human dentate gyrus: (A) 1 monthold; (B) 8 years old; (C) 10 year old. In the dentate gyrus of a 1 month-old child, there are no parvalbumin-stained somata, but a few axonal branches were visible (A). At older ages, axons arborize among the granule cells (B) in a manner characteristic of adults (C). Scale bar ¼ 20 mm for A, B, C. (For B/W version, see page 37 in the volume.)

Plate 4.3. Light micrograph (kindly provided by Dr. Attila Gulyas) shows the result of a combined anterograde tracing and calretinin immunostaining study, in the dentate gyrus. The anterograde tracer, biotinylated dextran amine (BDA) was injected into the medial septum diagonal band. Large boutons of BDA-containing axons form multiple, basket-like, putative synaptic contacts with the soma of calretinin-immunoreactive (labeled with a brown diaminobenzidine reaction product) neurons. One of these cells (arrows) located in the supragranular layer (SgL) the other is at the border between the granule cell layer (GcL) and dentate hilar area (H). Bar scale ¼ 50 mm. (For B/W version, see page 69 in the volume.)

Plate 4.5. Light micrograph taken from a double immunostained vibratome section of the monkey dentate gyrus. Immunoreactivity for substance P was labeled with a dark-blue Ni-diaminobenzidine reaction, while immunostaining for parvalbumin was visualized by a brown diaminobenzidine reaction. The soma and dendrites of the parvalbumin-containing cell embedded into the granule cell layer (GcL) is contacted by several substance P-immunoreactive axon terminals (arrows). Bar scale ¼ 10 mm. (For B/W version, see page 71 in the volume.)

Plate 5.2. Timm stained hippocampal sections from rat (A), guinea pig (B), cat (C), European hedgehog (D), dog (E) and man (F), illustrating the distribution of intensely stained, black MF terminals in the different species, as well as other terminal fields and general organization. Note species specific characteristics, like the ‘‘end bulb’’ at the terminal part of the MF layer in guinea pig (black asterisk, B) and the layering of the dentate hilus with MF terminal-free, intermediate or plexiforme sublayer in the same species (white asterisk, B), and MF projections into CA1 in some cats (arrows, C) and European hedgehog (D). Except for the black MF terminal staining, the neuropil of the human hippocampus is virtually unstained, due to the use of special Timm staining procedures for non-perfused tissue (Danscher and Zimmer, 1978). Abbreviations: CA1, hippocampal regio superior or subfield CA1; CA3, hippocampal regio inferior or subfield CA3; FD, fascia dentata (or denate gyrus); g, dentate granule cell layer; h, dentate hilus or CA4; m, dentate molecular layer; mf, MF layer; Sub, subiculum. Scale bars: 500 mm (For B/W version, see page 88 in the volume.)

Plate 5.3. Timm stained mouse hippocampal sections, illustrating strain differences in MFs terminal projections between BALB7C (A) and C57B mice (B) and Reeler mutants (C). BALB/c mice have no infrapyramidal MFs, but a short intrapyramidal bundle (black asterisk, A), compared to long infrapyramidal bundles in C57B mice (black asterisk, B). In Reeler mutant mice, granule and pyramidal cell layers are disorganized, resulting in a corresponding distribution of MF terminals in the confluent granule cell layer/hilar area and CA3. However, MFs still avoid CA1, similar to normal mice. Abbreviations: g, dentate granule cell layer; h, dentate hilus (CA4); m, dentate molecular layer. Scale bar: 500 mm (For B/W version, see page 89 in the volume.)

Plate 5.5. Timm stained section from dorsoposterior level of young adult, rat hippocampus. When newborn, the rat received a mechanical lesion of the CA3–CA1 transition (les) as well a transection of the entorhinal perforant path to the fascia dentata. The CA3–CA1 lesion induced an aberrant ingrowth of MFs into CA1, illustrated by presence of small, Timm stained MF terminals (asterisk) along the CA1 pyramidal cell layer. The removal of the perforant path projection to the outer molecular layer (m) induced spread of associational-commissural projections from inner part of the layer, and induction of supragranular MF terminals (arrow) (see Zimmer, 1973, 1974). Abbreviations: g, granule cell layer. Scale bar: 500 mm (For B/W version, see page 92 in the volume.)

Plate 5.6. A: Timm stained MF (collateral) terminals, found in normal adult rats to accompany dendrites (arrows), arising from neurons in the limiting subzone just below the dentate granule cell layer or in the dentate hilus, and extending into and sometimes through the dentate granule cell layer. (g) into the, commissural-associational terminal zone (c/a). B: Timm stained, dense supragranular MF terminal projection (sgr) found normally at temporal levels of the cat fascia dentata as well as in other species. h, dentate hilus. Scale bar: 50 mm (A); 100 mm (B). (For B/W version, see page 94 in the volume.)

Plate 5.7. (A) Timm stained section from midposterior level of 5-day-old rat, demonstrating a distinct Timm stained MF layer in CA3 and the hilus at this age (arrows). (B, C) Parallel sections of hippocampal slice culture, derived from 7-day-old rat and grown for 3 weeks. Timm stain reveals the distribution of aberrant supragranular and normal CA3 MFs (arrow and mf, respectively, in B) and thionine cell staining to visualize general cellular organization (C). Abbreviations: g, granule cell layer. Scale bars: 500 mm (For B/W version, see page 95 in the volume.)

Plate 5.8. (A) Timm stained sections of hippocampus from adult rat, subjected to hippocampal x-irradiation as newborn, stop odentate granule cell formation and lead to few MF terminals. (B) Timm stained section from the contralateral hemisphere of the rat shown in A, depicting a well-integrated dentate transplant (tpl), derived from a small block of fascia dentata, grafted just after the xirradiation. From the graft, an apparently normal, laminar-specific MF projection (mf) developed, connecting dentate granule cells of the graft (g) with the host CA3. The molecular layer of the graft (m) received a comparably laminar-specific projection of host entorhinal perforant path projections, normalizing the Timm stained laminar appearance of the layer. (C) Slightly displaced dentate transplant (tpl), grafted to x-irradiated newborn hippocampus just after irradiation as part of same experiment (Sunde et al., 1984). Encroaching on the recipient CA3, MFs from the granule cells of this graft has entered adjacent parts of CA3 and project ‘‘downstream’’, but appear to stop at the border with CA1. Surprisingly, no MFs from the graft projected in the ‘‘upstream’’ direction, i.e., toward the host fascia dentata. Scale bar: 500 mm (For B/W version, see page 100 in the volume.)

Plate 6.1. Interneurons are the primary target of DG granule cells. Diagram representing the multiple types of excitatory contacts made by granule cells. Within the dentate/hilar region collaterals of the mossy fibers (MFs) innervate both mossy cells and inhibitory basket cells. In area CA3, the MFs contact pyramidal neurons via large expansions, but also excite interneurons through either filopodial extensions from the large boutons or smaller en passant expansions along MF axons. (For B/W version, see page 110 in the volume.)

Plate 6.2. Summary of the pre- and post-synaptic constituents of the different MF terminals and of their plasticity. (A) Schematic representation of the giant MF boutons, which contact CA3 pyramidal cells. These terminals are characterized by low probability of basal release and multiple release sites. (B) Schematic representation of filopodial extensions and en passant contacts, which synapse on to interneurons. These have a high probability of basal release and single release sites. Both types of MF terminals contain several neuromodulators and the neurotransmitters glutamate and GABA. Note that both the presynaptic and postsynaptic sites contain several ionotropic GluRs and metabotropic GluRs, which confer to the MF a high degree of plasticity and the capacity for synaptic integration. (C) Relative expression of the different releasable contents and of the receptors and transporters during development, in the adult (D), and after epileptic activity (E). Arrows and font size indicate relative differences between the three states. (For B/W version, see page 114 in the volume.)

Plate 7.1. Original drawing of the dentate gyrus and hippocampus proper by Camillo Golgi (Golgi, 1886). While Golgi was aware of the laminated, bipolar arrangement of the granule cells, he did not correctly draw the course of the mossy fibers. As is known from numerous more recent tracer studies, granule cell axons mainly run in stratum lucidum of CA3 and impinge on proximal dendritic segments of pyramidal neurons. (For B/W version, see page 135 in the volume.)

Plate 7.2. Rescue of granule cell lamination in a slice culture of reeler dentate gyrus is achieved by coculturing with a wildtype culture providing a reelin-containing marginal zone. Two reeler cultures (rl/1 and rl/2) are cocultured with a rat hippocampal slice. A compact cell layer (arrow) has only formed in rl/1, which was cocultured next to the outer molecular layer of the rat dentate gyrus containing reelin-synthesizing Cajal-Retzius cells. In rl/2, which was cultured next to the stratum oriens of CA1, the reeler-specific loose distribution of neurons in the dentate gyrus is retained (arrowhead). Dashed lines represent borders between cultures. CA1, CA3, hippocampal regions CA1 and CA3; DG, dentate gyrus; g, granule cell layer. Scale bar: 200 mm (from Zhao et al., 2004, with permission). (For B/W version, see page 136 in the volume.)

Plate 7.3. Cocultures of two hippocampal slices to allow for the formation of ‘‘commissural’’ projections, traced by anterogradely transported biocytin injected into the hilar region of one of the two hippocampal slices (sites of biocytin injections labeled by asterisks). (A) Coculturing of two wildtype slices results in the formation of a compact ‘‘commissural’’ projection in the inner molecular layer (black arrow). Open arrow labels ipsilateral mossy fiber projection in stratum lucidum of CA3. CA1, CA3, hippocampal regions CA1 and CA3; g, granule cell layer. (B) Coculture of wildtype (wt) hippocampus and reeler hippocampus. Biocytin-labeled ‘‘commissural’’ fibers from the wildtype culture terminate profusely in the reeler culture, thus reflecting the scattered distribution of their target neurons, the granule cells. Open arrow labels the mossy fiber projection in the wildtype culture. DG, dentate gyrus of the reeler culture with scattered granule cells; P, pyramidal layer in the wildtype culture; P1, P2, double pyramidal layer in the reeler culture. (C) ‘‘Commissural’’ fibers from a reeler culture form a compact projection in the inner molecular layer of the wildtype culture (black arrow), indicating that reeler commissural fibers project to the inner molecular layer as is normal, provided that their target cells, the granule cells, form a tightly packed layer. g, granule cell layer. Scale bar: 100 mm (applies to A–C) (modified from Zhao et al., 2003, with permission; copyright by the Society for Neuro-science). (For B/W version, see page 138 in the volume.)

Plate 8.2. NeuroD mutants have radial glial scaffolding defects in the postnatal dentate. We examined the organization of the GFAP+ radially oriented dentate precursors usually residing in the subgranular zone in NeuroD mutant mice. In control mice, Prox1 clearly marks the dentate granule cell layer with the adjacent subgranular zone that serves as the focus of radially oriented GFAP+ fibers. In contrast, in NeuroD mutant mice there is only a small cap of Prox1+ granule cells in the dentate of P21 mutants and there are essentially no GFAP+ processes present in the mutant dentate. (For B/W version, see page 149 in the volume.)

Plate 11.2. Granule cell hyperexcitability is associated with mossy fiber sprouting. A shows representative traces comparing the response to molecular layer (orthodromic) and hilar (antidromic) stimulation in a patient with MTS. Note that there is a greater response to maximal hilar stimulation than to stimulation of outer molecular layer. Moreover, note that hilar stimulation produced a synaptic response prior to an antidromic spike. These data support the hypothesis that mossy fibers are primarily excitatory in resected tissue. B shows Timm staining from the region of the dentate where the cell shown in A was recorded. Note the robust supergranular labeling indicating the presence of sprouted fibers. (C1) Stimulation protocols have the problem of activating multiple fiber systems. Therefore, in the presence of 30 mM bicuculline, we applied a bolus of glutamate to the granule cell layer at a distance from the recorded cell and were able to evoke a long lasting barrage of EPSPs. The glutamate was applied at a distance from the recorded cell and thus we were not observing direct glutamate effects. C2 shows that a similar response was not observed when the glutamate was applied to the hilus. (For B/W version, see page 188 in the volume.)

Spontaneous

Evoked 20 mV 0.5 mS

IC EC

50 microV 0.5ms

40 microV 0.5ms

Plate 12.4. Extracellular waveforms of mossy cells during spontaneously and evoked spiking. A recording was obtained where the intracellularly recorded mossy cell was also observed on the extracellular electrode. The extended shape of the mossy cell dendrites allowed the extracellular signal to be observed on three shanks of the extracellular silicon probe (150 mm between shanks). The extracellular electrode track is indicated by the white arrows. The soma of the mossy cell is indicated by the yellow arrow. There was a difference in the shape of the action potentials, both intracellular and extracellular, suggesting differences in the site of spike initiation for spontaneous and evoked spikes. (For B/W version, see page 206 in the volume.)

Plate 13.2. Double labeling of GAD65 mRNA and either NeuN (A) or NADPH-diaphorase (B) in the rat dentate gyrus. (A) GAD65 mRNA-labeled neurons (black) are numerous throughout the dentate hilus (H) where they are intermingled with NeuN single-labeled (non-GABA) neurons (red; examples at arrows). Some GAD65 mRNA-labeled neurons are present in the granule cell layer (G) but are most abundant at the base of this layer. Most neurons in the molecular layer (M) are labeled for GAD65 mRNA. Granule cells and CA3 pyramidal cells are single-labeled for NeuN. (B) NADPH-diaphorase (dark blue) is present in only subgroups of the GAD mRNA-labeled neurons (red) in the hilus (H), granule cell layer (G) and molecular layer (M) (examples at arrows). (For B/W version, see page 220 in the volume.)

Plate 14.1. ‘Gatekeeper’ function of the dentate gyrus is maintained by GABAergic inhibition. Simultaneous voltage sensitive dye (A snapshot taken at the peak of the response, B trace illustrating the VSD response over time), patch clamp (C), and field potential (C) recording of dentate gyrus response to perforant path activation in control ACSF. Note robust activation of dentate gyrus molecular layer (red color in A, corresponding to a 10–15 mV EPSP in B), which does not result in activation of downstream structures (note lack of response in area CA3 in A and B). This lack of CA3 activation is because dentate granule cells do not fire action potentials in response to perforant path activation under these conditions. This is evident in both the patch (C, top trace, the neuron depolarized to Vm of 50 mV) and field potential recording (C, bottom trace), due to powerful feedforward inhibition activated by perforant path stimulation (C, note large IPSP in patch recording). The importance of inhibition in mediating this ‘gatekeeper’ function is illustrated in responses in panels D, E, and F, following perfusion with 5 mM picrotoxin, a non-competitive GABAA receptor antagonist. This concentration blocks 20–25% of inhibition (see inset [located above panel E] depicting an averaged spontaneous IPSC [sIPSC] before and after perfusion with 5 mM picrotoxin). During 25% GABAergic blockade, perforant path activation resulted in powerful activation of both the dentate gyrus and downstream structures (CA3 and hilus; D, E). It also triggered action potential firing in dentate granule cells (see patch and field potential recordings in F, both of which exhibit action potential firing). A grayscale image of the slice, with patch and field potential recording electrode location is depicted in the inset above A. From Carlson and Coulter (unpublished). (For B/W version, see page 239 in the volume.)

SO

CA1 p

SP

SUB

SR SLM

TAT PP p

SLu

HIL

m g

CA3

DG GCL ML

ENK DYN

Plate 15.1. Location of enkephalins and dynorphins in the hippocampal formation. Opioid peptides are found in all regions of the hippocampal formation: the hippocampus proper (fields CA1 and CA3), the dentate gyrus (DG), and the subiculum (SUB). The glutamatergic neurons [pyramidal cells (p), granule cells (g), mossy cells (m)] and their projections are shown. GABAergic interneurons are scattered in all laminae. The perforant path (PP) from entorhinal cortex provides the major excitatory innervation. Subcortical afferents and contralateral hippocampal projections enter via the fimbria/fornix (not shown) adjacent to CA3. Enkephalins (green) are in some granule cells, particularly their mossy fiber terminals in hilus and stratum lucidum (SLu). Mossy fibers extensively innervate hilar mossy cells and CA3 pyramidal cells. Enkephalins are in a portion of the projection from entorhinal cortex; the lateral portion of the perforant path (PP) in the DG and the temporal-ammonic tract (TAT) in CA1. Enkephalins are in a few scattered interneurons, which are relatively abundant on the border of stratum radiatum (SR) and stratum lacunosum-moleculare (SLM). Dynorphins (yellow) are in the granule cells. Dynorphin peptides are most striking in the mossy fiber terminals, and are also in the mossy fiber axons, granule cell somata, and granule cell dendrites. Abbreviations: SO, stratum oriens; SP, stratum pyramidale; GCL, granule cell layer. (For B/W version, see page 248 in the volume.)

+ ENK

Molecular Layer

_

PP

GABA

? ?

+ ACh

SubP Granule Cell Layer

Granule Cell

?

GABA PARV

Hilus

+ ENK/DYN

+ ACh

GABA NPY SOM

MOR

to CA3

+

DOR KOR

Plate 15.2. Subcellular locations of opioid peptides and receptors in the dentate gyrus. Enkephalins (green shading) are in: (1) somatodendritic, axon, and terminal portions of granule cells; (2) lateral perforant path axons and terminals in the outer molecular layer; and (3) occasional GABAergic interneurons (not shown). Mu-opioid receptors (MORs; blue) are most frequently in all portions of parvalbumin (PARV)-containing basket cells that innervate granule cell somata and proximal dendrites. MORs are also in cholinergic and GABAergic afferents from the lateral septum/diagonal band. Delta-opioid receptors (DORs; pink) are in many hilar interneurons that contain somatostatin or neuropeptide Y (SOM/NPY) and inhibit the distal dendrites of granule cells. PARVcontaining neurons express DOR mRNA. DORs, and to a lesser extent MORs, are occasionally present in granule cell dendrites. Kappa-opioid receptors (KORs) in guinea pigs are in substance P containing afferents to granule cells, in granule cell mossy fibers, and most likely on perforant path terminals in the outer molecular layer. (For B/W version, see page 249 in the volume.)

Plate 19.2. DSI in dentate granule cells and hilar mossy cells. (AA) Dentate gyrus of the hippocampal slice (scale, 50 mm). (AB) Differential interference contrast image of dentate granule cells (scale, 10 mm). (AC) Single dentate granule cell filled, and visualized with fluorescence microscopy during whole cell recording (scale, 10 mm). (AD) DSI of eIPSCs (left, top) and sIPSCs (right, top) is observed in dentate granule cells. In both cases, magnitude of DSI is reduced by bath application of thapsigargin (TG, 2 mM, bottom). (AE) Thapsigargin also reduced the calcium transients produced by DSI-inducing depolarizing voltage steps. Adapted with permission from Isokawa and Alger (2005). (B) Depolarization induced release of endogenous cannabinoids from hilar mossy cells preferentially inhibits calcium-dependent exocytosis. DSI in of eIPSCs is apparent in an individual mossy cell. DSI of miniature IPSCs is absent in the same cell. Following bath application of KCl and CaCl2, DSI of mIPSCs becomes apparent. Summary data are presented in BD, where numbers on the bars are n values. Adapted with permission from Hofmann et al. (2006). (For B/W version, see page 328 in the volume.)

Plate 23.2. Subcellular localization of estrogen (ER), androgen (AR), and progestin (PR) receptors in the dentate gyrus. A subset of GABAergic interneurons contains nuclear ERa (dark pink). Granule cells, newly born cells (identified by DCX) and some GABAergic interneurons contain cytosolic and plasma membrane-associated ERb (blue). Dendritic spines, many originating from granule cells contain ERa, ERb, AR (dark green), and PR (purple). A few dendritic spines in the hilus, likely originating from mossy cells, contain ERa and ERb. ERa, ERb, AR, and PR are found in axons and axon terminals. Some ERa-containing terminals are cholinergic (acetylcholine, orange); some ERb-containing terminals resemble monoaminergic boutons. Lot of astrocytes (stars), mostly in the molecular layer, also contain ERa, ERb, AR, and PR. (For B/W version, see page 407 in the volume.)

Plate 28.2. Increased cell proliferation and altered dentate gyrus neuroblast migration after SE. (A, B) BrdU incorporation (arrows) in adult rat dentate gyrus 7 days after saline (A) or pilocarpine (B) treatment followed by a 2 day post-BrdU survival. Subgranular zone BrdU labeling increases markedly after SE (B) compared to the control (A). (C, D) Doublecortin (DCX) immunoreactivity in the dentate gyrus of a control (Con; C) and an adult rat 14 days after pilocarpine-induced SE (14d; D). Note the increased hilar DCX expression and chains of DCX-positive cells extending into the hilus (arrows in D) after SE. E, PSA-NCAM+ neuroblast chains (red, arrows) alongside GFAP+ hilar astrocytes (green). Dashed lines in C–E denote the granule cell layer (gcl) — hilar (h) border. (For B/W version, see page 534 in the volume.)

Plate 28.3. Schematic showing the effects of prolonged seizures on caudal subventricular zone (SVZ) and dentate gyrus progenitors. In the intact adult rat (top), progenitors located in the infracallosal SVZ (purple) give rise to white matter oligodendrocytes (orange), while those in the dentate gyrus (green) give rise to DGCs in the granule cell layer (gcl). After seizure-induced hilar and pyramidal cell layer (pcl) injury (bottom; jagged arrows), progenitors migrate aberrantly from the caudal SVZ to form glia (a; purple) in the hippocampus proper or from the dentate subgranular zone to form hilar-ectopic DGCs (b). (For B/W version, see page 535 in the volume.)

Plate 33.1. Wiring diagram of the hippocampus (interneurons excluded) showing dual inputs from the lateral and medial entorhinal cortex. The layer 2 cortical inputs to the dentate and CA3 diverge fan out (f) widely over these networks and then provide convergent input to individual dentate granule cells. In contrast, the layer 3 cortical inputs are specialized for individual subregions of CA1. This is an example of a point-to-point (p-p) connection. Within the hippocampus, granule cells provide input, via mossy fibers to the mossy cells of the dentate and to CA3 cells. CA3 cells make feedback connections to themselves and to the mossy cells of the dentate. These, in turn, provide excitatory input to granule cells in the inner third of the granule cell dendritic tree. (For B/W version, see page 616 in the volume.)

Plate 36.1. (A) Two representative granule cells filled with 5,6 carboxyfluorescein from the dentate gyrus of a 24-month-old rat. (B) Histograms showing the numbers of carboxyfluorescein injections that resulted in single, double, triple, or greater numbers of granule cells filled with dye. Aged rats showed significantly increased electrotonic coupling compared to young animals. This may account for the increased excitability of old granule cells. Adapted with modifications from Barnes et al. (1987). (For B/W version, see page 665 in the volume.)

Plate 36.5. Image of granule cells (blue) following in situ hybridization using a c-RNA probe for the IEG Arc. (A) Arc positive cells (shown in red) in the granule cell layer from the upper blade of a young rat following two 5 min (separated by a 20 min rest period in the home cage) episodes of spatial exploratory behavior. (B) Granule cells from a caged control animal of the same age, that did not perform the exploratory behavior, are shown, illustrating baseline or negligible expression of Arc mRNA. (For B/W version, see page 672 in the volume.)

Plate 36.6. (A) Examples of gadolinium-induced change in MRI signal, (a measure of cerebral blood flow volume, CBV, that is a correlate of brain metabolism) from a young and old rhesus monkey. Warm colors indicate higher CBV values. The white circle, identified in precontrast images, overlies the dentate gyrus. (B) CBV measurements plotted vs. age in the dentate gyrus indicates a significant decline with advancing age. (C) CBV measurements in the dentate gyrus of old monkeys plotted against memory performance on the delayed nonmatching-to-sample test. A significant relationship between CBV and memory performance is observed, selectively in the old animals. Adapted with modifications from Small et al. (2004). (For B/W version, see page 673 in the volume.)

Plate 36.7. Behavior-induced Arc mRNA expression pattern in young and old rat granule cells of the dentate gyrus. (A) Individual examples of behaviorally induced Arc RNA. Fluorescence-tagged CY3 Arc RNA containing cells are green, and granule cell nuclei are shown in red (counterstained with propydium iodide, a nucleic acid stain). (B) The average proportion of Arc-positive neurons (% of total cells) measured from the three different age groups (9, 15, and 24 months) is shown for the dentate gyrus. Note that the dentate gyrus of old rats contains a significantly lower proportion of behaviorally induced Arc-positive granule cells compared to young or middle-aged animals. (For B/W version, see page 674 in the volume.)

Plate 38.1. A schematic of the dentate gyrus granule cell layer (GCL) illustrating the different ways by which neurogenesis can influence its structure and function. Boxed panel reveals a cross section of the dentate GCL with the different populations that reside within it: mature granule cells born during development (light blue), adult-generated mature granule cells (dark blue), adult-born immature neurons (red) and interneurons (green). Over the lifespan, the GCL may increase in size due to a net addition of new neurons (A) or may remain unchanged due to a net replacement of developmentally generated granule cells (B). Changes in neurogenesis can result in increased representation of interneurons (C), a larger pool of adult generated immature neurons (D) or the generation of mature neurons with distinct physiological and biochemical properties (E). Conceivably, neurogenesis may be altered in any one of these ways in MDD. Conversely, AD drugs may influence DG function in more than one way to exert their behavioral effects. (For B/W version, see page 700 in the volume.)

Plate 40.4. Color map showing significant (p ¼ 0.001) voxels of decreased white matter density in participants with amnestic MCI compared with controls, superimposed on coronal and sagittal slices of a template based on data for all subjects in the study. The image is masked to include white matter regions. The colors correspond to the t values shown on the color bar. Note the bilaterality of significant differences in the white matter of the parahippocampal gyrus (adapted with permission from Stoub et al., 2006). (For B/W version, see page 750 in the volume.)