Human Hypothalamus: Basic and Clinical Aspects, Part I: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

2014 Prels(1) Rev 20/11/03 9:57 am Page v CHAPTER TITLE 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 ...

Author: Dick F Swaab

726 downloads 1031 Views 14MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

2014 Prels(1) Rev

20/11/03

9:57 am

Page v

CHAPTER TITLE

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

v

Foreword

It gives us great pleasure to present the first volume of the third series of the Handbook of Clinical Neurology. The Handbook was originally conceived by Pierre Vinken and George Bruyn in the 1960s, with the first volume being published in 1968. After the series was concluded in 1982, it was realized that an update of the material was imperative. Accordingly, a revised series was planned and published over the following years, concluding with the publication of Volume 78 in 2002. Since then, George Bruyn has passed away and Pierre Vinken has retired, but the need for a new series, incorporating advances in the field, has become necessary. We are therefore pleased to take on the responsibility of supervising the preparation of a new (3rd) series. Advances have occurred in both clinical neurology and the neurosciences, and these have led to a broader understanding of neurological disorders and have had a significant impact on patient care. Such advances will be covered in the new series. In addition, new topics that were not considered in the earlier series will also be covered. We will also ensure the systematic inclusion of neurobiological aspects of the nervous system in health and disease, in order to clarify physiological and pathogenic mechanisms and provide new therapeutic strategies for neurological disorders. Furthermore, each volume in the new series will include data related to epidemiology, imaging, genetics, and therapy. The new series starts with two volumes (79, 80) on the human hypothalamus. These volumes contain the new elements that we want to develop in the third series. It covers a topic that has received insufficient attention in the past. The human hypothalamus is an extremely complex structure that consists of a large number of very different functional units (nuclei) that are not included in the standard neuropathological investigation of the human brain. In fact, on the basis of their chemical nature, many of the functional systems have only recently been distinguished by modern neurobiological techniques. The hypothalamus was traditionally considered to be a neuroendocrine structure, of limited interest to neurologists. It has now become clear, however, that this structure contributes to the memory and attention deficits in the dementias, that a disorder of the orexin/hypocretin system is the cause of narcolepsy, that hypothalamic hamartomas are responsible for gelastic epilepsy, that the subthalamic structure where depth electrodes are placed in parkinsonian patients is a hypothalamic structure, and that the source of cluster headache is situated in the hypothalamus. Moreover, the hypothalamus appears now to be the basis of many signs and symptoms of disorders situated on the border between neurology and psychiatry, such as depression, eating disorders, aggression, and mental retardation. As a consequence, the hypothalamus is a meeting point for neuroscientists, neurologists and psychiatrists, endocrinologists, and pediatricians, and as such is an important starting point for the new series. The successful preparation of the new series of the Handbook will again depend on the dedication of many persons. As in the past, each topic will be covered by one or more Volume Editors. Throughout the development and production of the Series, the editorial staff of Neurology and Neuroscience of Elsevier B.V. in Amsterdam has provided invaluable assistance. October 2003

Michael J. Aminoff François Boller Dick F. Swaab

2014 Prels(1) Rev

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 9

20/11/03

9:57 am

Page vii

Obituaries

Professor Dr. George W. Bruyn (1928–2002)

Neurologists worldwide should grieve the untimely death of George because of the enormous debt we all owe him. From 1964 until his demise, he was the inspiration and co-editor of the irreplaceable series of the Handbook of Clinical Neurology, which established itself as the world’s most comprehensive source for information about neurology and the diseases which it encompasses. It is ironic, but fitting, that George died soon after completing this gargantuan effort which is undoubtedly the last compendium to encompass the literature and lore of clinical neurology as it existed prior to the advent of the dot.com age. The impetus for it began in 1964 in the Netherlands, encouraged by Drs. Arie Biemond and Macdonald Critchley, with a difference in concept, design and make-up from its predecessor, the 18-volume Handbuch der Neurologie by Bumke and Foerster, which encompassed the knowledge pertinent to the adolescent field of neurology, some 18 years earlier. Countless hours of George’s life were spent with Pierre Vinken on this task, which began with bedside neurology verified by neuropathology and culminated with volume 78 in January 2002, during the new age of noninvasive neuroimaging. A polymath, Bruyn included neuroanatomy, neurophysiology, neurochemistry, electroencephalography, neurosurgery and neuroradiology, even though they had become more or less independent sciences.

2014 Prels(1) Rev

viii 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 9

20/11/03

9:57 am

Page viii

OBITUARIES TO GEORGE W. BRUYN

Concomitantly, he was Editor-in-Chief of the Journal of Neurological Sciences from 1983–1989, and later of Clinical Neurology and Neurosurgery. In early 2002, in one of his last contributions with Bengt Ljunggren, George told the story of the growth of the Karolinska Institute and its key role in the process of the Nobel Prize in Physiology and Medicine. How did he work so efficiently and what motivated him to contribute so much to the common good? Born in 1928 to psychiatrist Johannes Bruyn and Maria Schwencke into a family of three elder siblings and a twin sister, he trained in neurology at Utrecht before ascending to Professor and Chairman of the Department of Neurology in Leiden in 1976. He became an inspiring teacher and role model, so that 27 trainees finished their theses, covering an enormous range of topics. George relished an intellectual contest. His teaching style, like his approach to conferences, was confrontational. However, it was done with great charm, so one could not take offense. He gave unfailing advice to an array of students, conferees and friends. Gifted with a sharp wit and good sense of humor, he inspired those around him with sometimes unusual and provocative points of view. George was a fair, critical and impartial editor, who interspersed his editorial and publishing discussions with incisive wit and charm. One key aspect was his incredible knowledge of the neurological literature, which he kept at easy recall, and an unprecedented tenacity for recognizing seminal publications. Our colleague, George, is survived by his wife, Rosemary; his brother, a general practitioner; four sons, George, Richard, Terence and Geoffrey, a rheumatologist, neurologist, lawyer, and a banker, respectively; a daughter, Iris, a historian; and 16 grandchildren. We extend our sympathy to his wife, Rose, and the family. He will be greatly missed and remembered long for his many contributions to making neurology an exciting discipline. James F. Toole Winston-Salem, NC 27157-1068, USA

The international neurological community lost a distinguished member when George Willem Bruyn died suddenly on June 23, 2002 in his beloved Domaine de Caumezelles, in La Salvetat sur Agout in south-western France. Prof. Bruyn was born on December 14, 1928 in Delft, The Netherlands, the son of Johannes Willem Bruyn and Maria Schwencke. His father was the director of the psychiatric hospital Zon en Schild (Sun and Shield) in Amersfoort and had an interest in brain pathology, which he passed on to his son. George had an elder brother who became a family practitioner, two elder sisters, and a twin sister. The Bruyn family roots can be traced back to Lambert Janszoon Bruyningh, who was born in Emmelekamp, now Emlichheim, Germany, in 1638. He emigrated to Amsterdam, became a baker, and married Trijntje Jans van der Burgh. The family lived in Amsterdam until 1850, then Haarlem, and later Amersfoort. George Bruyn received his M.D. in 1951 from the State Medical School in Utrecht and was licensed in 1954. His specialty training in neurology under Prof. W.G. Sillevis Smitt was also in Utrecht, from 1954 to 1958, where he was awarded the Ph.D. for his thesis on ‘‘Pneumoencephalography in the diagnosis of cerebral atrophy’’ in 1959. He then became Chef de Clinique in the Department of Neurology of the Medical Faculty of the University of Leiden from 1959 to 1963, held the same position in the Central Military Hospital in Utrecht, 1963/1975, and was appointed as Professor and Chairman of Neurology at the University of Leiden in 1975. He became Emeritus in 1992 and retired in Bilthoven, near Utrecht, which was his home from 1962 until he died. He remained extremely active during his retirement, maintaining close ties with the University of Leiden. In 1988, he undertook a 3-month, 33-lecture tour of Indonesia, the Philippines, Thailand and Singapore, and in the 1990s served several stints as Visiting Consultant in the Armed Forces Hospital in Riyadh, Saudi Arabia. Prof. Bruyn’s distinguished career included honorary fellowships and memberships in the neurological societies of The Netherlands, Belgium, Europe, France, Peru, United Kingdom, and the United States. He was Secretary General of the XIth International Congress of Neurology in Amsterdam in 1977. He received the 1979 Lectureship Award of the National Migraine Foundation of America, and was decorated by the Government of Peru in 1984. He was a

2014 Prels(1) Rev

20/11/03

9:57 am

Page ix

OBITUARIES TO GEORGE W. BRUYN

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

ix

member of the advisory boards of numerous national and international associations and research groups on ALS, Huntington’s disease, migraine, multiple sclerosis, and on the history of medicine, one of his major interests. He had published over 450 scientific papers as well as a number of books. He was a member of the editorial board of many journals, the Editor-in-Chief of the Journal of the Neurological Sciences from 1983 to 1989, and of Clinical Neurology and Neurosurgery from 1990 until his death. One of his most memorable achievements was his editorship, with Pierre Vinken, of the prestigious 78-volume Handbook of Clinical Neurology starting in 1968. He was proof reading of the last two volumes at the time of his death. George Bruyn was truly a Renaissance man: he read and translated Greek and Latin on sight, and was fluent in French, German and English. He wrote on historical, ethical and philosophical aspects of medicine and neurology. He was knowledgeable about the arts, especially Dutch painting, literature and French wines. He maintained a lifelong interest in neuroanatomy and neuropathology, derived in part from his admiration for Ernst de Vries, the eminent neuropathologist in Utrecht. He was also an expert bridge player and a reckless driver with a cavalier disregard for speed limits. His main hobby was the collection of eponyms; he had already co-edited two books on them, and was planning a 7-volume compendium. My fondest memories of George are watching him ecstatically poring over volumes in the Rare Book Room of the Francis Countway Library of the Harvard Medical School searching for eponyms, happily jotting down new discoveries, and his pride in showing my wife and me his new plantings and other improvements in his property of Caumezelles. George Bruyn had a great sense of humor as well as a confirmed libertarian streak. His likes and dislikes were strong, the latter directed mostly against politicians of all stripes, lay and medical alike. His loyalty to family and friends was just as strong. A most remarkable man and a dear friend has passed on. He is survived by his wife Rosemary, four sons: a rheumatologist, a neurologist, a lawyer/banker and a businessman, a daughter who is a historian, and 16 grandchildren. Acknowledgements The help of Drs. Richard and Geoffrey Bruyn, Prof. Frans Jennekens, and Ms. Elly Tjoa of Elsevier, is gratefully acknowledged. Charles M. Poser Boston, MA, USA

2014 Prels(1) Rev

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

9:57 am

Page xi

Preface

It is virtually impossible to pinpoint the moment of conception of this book; all I can do is say that its direction was marked by crucial moments that have all contributed to the end result. For instance, in the summer of 1966, as a 3rd-year student of medicine, I decided I wanted to know more about medical biological research and applied for a job as student-assistant at the Netherlands Institute for Brain Research. When the then director, Professor J. Ariëns Kappers, asked me what discipline had my particular interest, I answered: “Neuroendocrinology”. That was one of those crucial moments. My father, who was very interested in hypothalamic functions and who wanted to encourage me, gave me The Pituitary Gland, three volumes, by Harris and Donovan, Neuroendocrinology, in two volumes, by Martini and Ganong, and two years later, Hypothalamic Control of the Anterior Pituitary by the Szentágothai group. In later years I came to meet the editors and authors of these books in person; which was quite an event on every occasion. Those were indeed stimulating volumes, although, to my taste, they contained too little information on the human hypothalamus. The moment when, a little later on, I laid eyes on the supraoptic nucleus in cryostat sections I had cut myself, heralded a new phase in my life: from that moment on, the borders between the study of medicine and scientific research began to blur. The next crucial moment came with my first publication, guided by my mentor, Jongkind, who taught me histochemistry and later microchemistry. It was a curious moment to see my own name in print (Jongkind and Swaab, 1967). Nineteen-seventy saw another crucial moment; in the middle of my internship in surgery, I had to ask for an afternoon off to defend my thesis. The groundwork for Chapter 8 of this book was laid then. During my internship at obstetrics and gynecology, my paper The hypothalamo-neurohypophysial system (HNS) in reproduction was discussed by the clinician Dr. J. Honnebier. In it I suggested, among other things, that the fetal HNS might play a role during labor. It turned out that Honnebier himself was also starting research involving the fetal brain and labor and for that reason had studied a few anencephalic pregnancies. His discussion of my paper resulted in a long-standing collaboration, first with him (Chapter 18.1), and later with his daughter (Chapter 4.2). For me, this meant that at last here was an opening to a field of research that had had my keen interest for a while: research of the human hypothalamus integrated with fundamental research. In 1974, I developed immunocytochemical localization of vasopressin and oxytocin, and in 1975 radioimmunoassays for these peptides were introduced in my group, all of which contributed to increasing possibilities to do research on the human brain. The early eighties saw our first attempts to apply these techniques to human material and in 1985 the research of the human hypothalamus became a main research theme. Nineteen-eighty-five was also the year of a number of key papers on the human hypothalamus in relation to sexual differentiation (Swaab and Fliers, Science, 228: 1112–1113 – see Chapters 5 and 24.5), and on aging and Alzheimer’s disease (Swaab et al., Brain Res., 342: 37–44, 1985 – see Chapters 4.3 and 29.1). Research of the human hypothalamus received an enormous boost that year, because I founded the Netherlands Brain Bank, together with Prof. F.C. Stam. Stam was professor of Neurology, Psychiatry and Neuropathology. He had come to me the year before, grumbling about my proposition during a lecture that Alzheimer’s disease was in fact nothing more than an accelerated and premature process of aging of the brain. He had been trying to convince people that it was an illness and my idea frustrated his work. My reply was that anything too rapid and too early constituted was a disease! We have learned a great deal from this pioneer of Alzheimer’s disease research. Since its foundation in 1985, the Netherlands Brain Bank has been able to provide clinically and neuropathologically well-characterized material from 2500 obductions, usually with a very brief postmortem interval, to some 340 research

2014 Prels(1) Rev

xii 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 9

20/11/03

9:57 am

Page xii

PREFACE

projects in 22 countries. Fortunately (for our own line of research) the demand for hypothalami by other groups has been modest. The core of this monograph stems from reviews from the 1970s. However, the actual set-up of this book about the human hypothalamus was shaped by a review I wrote on request in 1993 (Swaab et al., Anat. Embryol., 187: 317–330) and subsequently by a review in the Handbook of Chemical Neuroanatomy, the Primate Nervous System (eds. Bloom and Hökfelt), pp. 39–137, 1997. As difficult as it is to indicate when this book began, it was at least as hard to know when it was ready to be handed over to Elsevier and to let it go. Fortunately it is now possible to respond almost immediately to the latest scientific developments – Elsevier intends to publish the volumes of this series on the internet and to update chapters on a regular basis. It would have been impossible to complete this work without the quick, accurate, patient, and professional secretarial and editorial help of Mrs. W. Verweij, the creative bibliographical assistance by Dr. J. Kruisbrink and the artwork of G. van der Meulen and H. Stoffels. I am grateful to the late professor George Bruyn, who convinced me to publish this monograph in the Handbook of Clinical Neurology series. The colleagues and friends whose thoughtful comments helped to shape the book are thanked individually in the Acknowledgements on p. xv. All in all this has become a book I would have loved to have been given as a present by my father in 1966: hopefully it will now be a starting point for my scientifically gifted children and grandchildren. Je sais bien que le lecteur n’as pas grand besoin de savoir tout cela, mais j’ai besoin, moi, de le lui dire1 J.J. Rousseau Dick F. Swaab

1

I know very well that the reader has no great need to know all this; it is I who have a need to tell him.

Dedication

To my father from whom I got my hypothalamus and my mother who has given me the cortex to study it*

* An unacceptable simplification of the work of Keverne et al., 1996.

2014 Prels(1) Rev

20/11/03

9:57 am

Page xv

CONTENTS

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

xv

Acknowledgements

I would like to express my gratitude for the continuous help and critical remarks of all those who made this work possible, especially Afra van den Berg, Ronald Bleys, George Bruyn, Ruud Buijs, Liesbeth Dubelaar, Frank van Eerdenburg, Tini Eikelboom, Bart Fisser, Eric Fliers, Bas Gabreëls, Tony Goldstone, Louis Gooren, Valeri Goncharuk, Joop van Heerikhuize, Michel Hofman, Witte Hoogendijk, Inge Huitinga, Tatjana Ishunina, Marina Kahlmann, Dries Kalsbeek, Wouter Kamphorst, Michiel Kooreman, Berry Kremer, Frank Kruijver, Jenneke Kruisbrink, Gert Jan Lammers, Fred van Leeuwen, Rong-Yu Liu, Paul Lucassen, Gerben van der Meulen, Gerben Meynen, Jan van de Nes, Elly de Nijs, Sebastiaan Overeem, Maria Panayotacopoulou, Joris van der Post, Chris Pool, Rivka Ravid, Erik Scherder, Eus van Someren, Henk Stoffels, Elly Tjoa, Suzanne Trottier, Unga Unmehopa, Paul van der Valk, Wilma Verweij, Ronald Verwer, José Wouda, Jiang-Ning Zhou, all other participants of the Netherlands Brain Bankteam, and all staff members, students, and guest workers of the Netherlands Institute for Brain Research.

2014 Prels(1) Rev

20/11/03

9:57 am

Page xvii

xvii

LIST OF ABBREVIATIONS

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

List of abbreviations

A AADC AC AD ACTH ADHD AGRP AIDS AIP ALD ALS AM AMPA AMDLX ANP APOE AT ATD AVP BDNF BMI BST BSTdspm/ BNSTdspm BSTc BSTm CAG CAH CART CCK CDC CG ChAT CM CMV CNS CRH CSF CT DII

amygdala aromatic L-amino acid decarboxylase anterior commissure Alzheimer’s disease corticotropin attention deficit hyperactivity disorder agouti-related peptide acquired immune deficiency syndrome acute intermittent porphyria adrenoleukodystrophy amyotrophic lateral sclerosis anteromedial subnucleus of the basal nucleus -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid adhesion molecule-like X atrial natriuretic peptide apolipoprotein E angiotensin 1,4,6-androstratriene-3,17-dione (aromatase inhibitor) arginine vasopressin brain-derived neurotropic factor body mass index bed nucleus of the stria terminalis darkly staining posteromedial component of the bed nucleus of the stria terminalis central nucleus of the bed nucleus of the stria terminalis medial nucleus of the bed nucleus of the stria terminalis DNA sequence that codes for glutamine repeats. An expanded sequence is found in Huntington’s disease congenital adrenal hyperplasia cocaine- and amphetamine-regulated transcript cholecystokinin center for disease control and prevention chiasmal gray choline acetyltransferase corpora mamillaria cytomegalovirus central nervous system corticotropin-releasing hormone cerebrospinal fluid computer tomography deiodinase type II

DAX-1 DA DB/DBB DDAVP DES DHEA DHEAS DM/DMN/ DMH DMI DMV DNA DSM-III R/IV

DYN EAE ECT EEG EM ER-/ ERT FAI FO/Fx FSH GA GABA GAD GAP GFAP GHRH GnRH HCG Hcrt1-2 HD H&E HMPG 5-HIAA HIOMT HITF HIV HLA HNS HPA-axis

dosage-sensitive sex-reversal, adrenal hypoplasia, congenital, X-chromosome-1 dopamine diagonal band of Broca 1-desamine-8-D-arginine vasopressin (= desmopressin) diethylstilbestrol dehydroepiandrosterone dehydroepiandrosterone sulfate dorsomedial nucleus of the hypothalamus desmethylimipramine dorsal motor nucleus of the nervus vagus deoxyribonucleic acid diagnostic and statistical manual mental disorders (American Psychiatric Association), third revised edition/fourth edition dynorphin experimental allergic encephalomyelitis electroconvulsive therapy electroencephalogram electron microscope estrogen receptor-/ estrogen replacement therapy free androgen index fornix follicle-stimulating hormone Golgi apparatus gamma-aminobutyric acid glutamic acid decarboxylase gonadotropin hormone-releasing hormoneassociated peptide glial fibrillary acidic protein growth hormone-releasing hormone gonadotropin-releasing hormone (= LHRH) human chorionic gonadotropin hypocretin (orexin) 1-2 Huntington’s disease hematoxylin–eosin staining 3-methoxy-4-hydroxyphenylglycol 5-hydroxyindoleacetic acid hydroxyindole-O-methyltransferase human intestinal trefoil factor human immunodeficiency virus human leukocyte antigen hypothalamoneurohypophysial system hypothalamopituitary–adrenal axis

2014 Prels(1) Rev

xviii 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 9

HVA 5-HT I III ICC icv IF IFN IGF IHA IL-1 INAH1-4 INSP4 KALIG-1 LC LCA LH LHA LHRH LPH LV LVP MAO MAP(A/B) MCH MCR1-4 MDMA ME MEN MELAS MHPG MHC MPN MRI MS ()MSH (m)RNA NA NADPH NAPH NAT NBB NBM N-CAM NEI NFT NGF NKB NMDA NOS NP

20/11/03

9:57 am

Page xviii


homovanillic acid (= serotonin (5-hydroxytryptamine) infundibulum third ventricle immunocytochemistry intracerebroventricularly infundibular nucleus interferon insulin-like growth factor intermediate hypothalamic area interleukin-1 interstitial nucleus of the anterior hypothalamus 1-4 inositol-(1,3,4,5)-tetrakisphosphate Kallman’s syndrome interval gene-1 locus ceruleus leukocyte common antigen luteinizing hormone lateral hypothalamic area luteinizing hormone-releasing hormone (= gonadotropin-releasing hormone, GnRH) lipotropic hormone lateral ventricle lysine vasopressin monoamine oxidase microtubule-associated protein (A/B) melanin-concentrating hormone melanocortin1-4 receptor 3,4-methylenedioxymethamphetamine (= ecstasy) median eminence multiple endocrine neoplasia mitochondrial encephalopathy, lactic acidosis and stroke-like episode syndrome 3-methoxy-4-hydroxyphenylglycol major histocompatibility complex medial preoptic nucleus magnetic resonance imaging (fMRI = functional MRI) multiple sclerosis -melanotropin (messenger) ribonucleic acid norepinephrine nicotinamide adenine dinucleotide nicotinamide adenine dinucleotide, reduced form N-acetyl-transferase Netherlands Brain Bank nucleus basalis of Meynert neural cell adhesion molecule neuropeptide glutamic acid isoleucine neurofibrillary tangles nerve growth factor neurokinin B N-methyl-D-aspartate nitric oxide synthase neuritic plaque

NPAF NPY-IR NSM NST/NTS NT NT-3, 4/5 NTI NTL OC ORL1 OT OVLT OXT P p75 PACAP PAP PBN PC PCR PD PDD PDYN PENK PET PHM PNS POAH POMC PSP PVA PVN PWS REM RHT RIA RT-PCR SAD SCN SDN(-POA) SHBG SIADH SIDS SN SNP SNRPN SON SOREMPS SPECT SRY SSRI SWS

neuropeptide AF neuropeptide-Y-like immunoreactivity nucleus septalis medialis nucleus of the solitary tract neurotensin neurotrophin-3, 4/5 nonthyroidal illness lateral tuberal nucleus/nucleus tuberalis lateralis optic chiasm opioid receptor-like receptor optic tract organum vasculosum lamina terminalis oxytocin perikarya low-affinity neurotrophin receptor pituitary adenylcyclase-activating polypeptide peroxidase-anti-peroxidase parabrachial nucleus prohormone convertase polymerase chain reaction Parkinson’s disease pregna-4,20-diene-3,6-dione prodynorphin proenkephalin positron emission tomography peptide methionine amine peripheral nervous system preoptic anterior hypothalamic area pro-opiomelanocortin progressive supranuclear palsy periventricular area paraventricular nucleus Prader–Willi syndrome rapid eye movement retinohypothalamic tract radioimmunoassay real-time polymerase chain reaction seasonal affective disorder suprachiasmatic nucleus sexually dimorphic nucleus (of the preoptic area) = INAH-1 sex hormone-binding globulin syndrome of inappropriate secretion antidiuretic hormone sudden infant death syndrome substantia nigra single nucleotide polymorphism small nuclear riboprotein-associated polypeptide supraoptic nucleus REM sleep onset periods single-photon emission computed tomography sex-determining region Y selective serotonin reuptake inhibitor slow-wave sleep

2014 Prels(1) Rev

20/11/03

9:57 am

Page xix

xix


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

T3 T4 TBS TENS TG TH THA TH-IR TMN TR TRH

triiodothyronine thyroxine Tris-buffered saline transcutaneous electrical nerve stimulation tuberal gray tyrosine hydroxylase tetrahydroaminoacrine tyrosine hydroxylase-immunoreactive tuberomamillary nucleus thyroid hormone receptor thyrotropin-releasing hormone

Trk A, B, C TSH VR-1,2,3 VEP VIP VLPO VMN/VMH VP

tyrosine kinase neurotrophin receptor A, B or C thyrotropin vasopressin receptor 1, 2 or 3 visual evoked potential vasoactive intestinal polypeptide ventrolateral preoptic region of the hypothalamus ventromedial nucleus vasopressin

Contents of Part H Vol. 80 (3rd series vol. 2)

Part II:

Neuropathology of the Human Hypothalamus and Adjacent Brain Structures

Chapter 17. Vascular supply and vascular disorders 17.1. Blood supply to the hypothalamus and pituitary a. Stalk/median eminence region b. Pituitary c. Portal system d. Infundibular process e. Artery of the trabecula f. Vascular bed of the pars distalis g. Hypothalamus h. Optic chiasm i. Lamina terminalis 17.2. Vascular lesions of the hypothalamus a. Subarachnoidal aneurysm b. Infarction and hemorrhage c. Systemic atherosclerosis d. Cavernous malformation e. Radiation therapy 17.3. Choroid plexus of the third ventricle a. Colloid cysts b. Xanthogranuloma c. Choroid plexus papilloma

Chapter 18. Disorders of development and growth 18.1. Anencephaly a. Failures of fusion and the factors involved b. Brain pituitary remnants c. Intrauterine growth and birth d. Anencephaly, the diagnosis of death and transplantation 18.2. Transsphenoidal encephalocele and empty sella syndrome 18.3. Congenital midline defects: optic nerve hypoplasia and septo-optic dysplasia (De Morsier's syndrome) a. Optic nerve hypoplasia b. Septo-optic dysplasia 18.4. Dystopia of the neurohypophysis a. True ectopia b. Dystopia with anterior pituitary abnormalities

xxvi

C O N T E N T S O F P A R T II

18.5.

18.6.

18.7.

18.8.

The optic chiasm a. Misrouting in albinism b. Non-decussating retinal-fugal fiber syndrome c. Other optic chiasm pathology The growth hormone axis in development and aging a. Noonan syndrome b. Multiple pituitary deficiencies c. Genetic forms of GHRH deficiency d. Adult growth hormone deficiency Hydrocephalus a. Hydrocephalus and the subcommissural organ b. Hypothalamic symptoms of hydrocephalus c. Causes of hydrocephalus Septum pellucidum abnormalities

Chapter 19. Tumors 19.1.

19.2. 19.3.

19.4.

19.5.

19.6. 19.7.

19.8. 19.9. 19.10.

Symptoms due to hypothalamic tumors a. Endocrine and autonomic disturbances b. Cognitive and behavioral disorders Germinoma and teratoma Hamartoma a. Hypothalamic hamartoma b. Hamartomatous nodules c. Intrasellar gangliocytoma d. Hamartoblastomas (Pallister-Hall syndrome) Glioma a. Diencephalic syndrome: hypothalamo-optic glioma/optic pathway glioma b. Gliomas of the optic pathways c. Other gliomas Craniopharyngioma, Rathke's cleft cysts and xanthogranuloma a. Craniophalyngioma b. Rathke's cleft cysts c. Xanthogranuloma Delxnoid and epidermoid tumors Pineal region tumors a. Germ cell tumors b. Pineal parenchymal tumors c. Glial neoplasms d. Cysts, tumors of supporting elements and miscellaneous e. Clinical symptoms of pineal region tumors Tuberous sclerosis (Bourneville-Pringle syndrome) and tumors of the hypothalamus Metastases Other tumors

Chapter 20. Hypothalamic infections 20.1.

Inflammatory conditions affecting the hypothalamus a. Bacterial infections b. Acute viral meningoencephalitis


20.2. 20.3.

c. Post- and parainfectious encephalomyelitis d. Fungal infections e. Other hypothalamic infections Encephalitis lethargica (Von Economo's encephalitis) Acquired immunodeficiency syndrome (AIDS)

Chapter 21. Neuroimmunological disorders 21.1. Neurosarcoidosis of the hypothalamus a. Clinical presentation b. Pathology c. Endocrine changes d. Therapy 21.2. Multiple sclerosis (MS) and the hypothalamus a. Autonomic, behavioral and neuroendocrine symptoms b. Mood changes c. The HPA axis in relation to susceptibility and recovery d. Inflammation, demyelination and hypothalamic structures e. Differential diagnosis of optic neuritis 21.3. Langerhans' cell histiocytosis (Hand-Schtiller-Christian disease; histiocytosis-X) 21.4 Other neuroimmunological hypothalamic disorders and lesions Chapter 22. Drinking disorders 22.1. Pathology of the neurohypophysis 22.2. Diabetes insipidus a. Familial central diabetes insipidus b. Autoimmune diabetes insipidus c. Pregnancy-induced diabetes insipidus d. Other causes of central diabetes insipidus e. Nephrogenic diabetes insipidus 22.3. Primary polydipsia and adipsia a. Primary polydipsia b. Psychogenic polydipsia c. Adipsinogenic disorders 22.4. Nocturnal diuresis 22.5. Vasopressin hypersecretion in diabetes mellitus 22.6. Inappropriate secretion of vasopressin a. Syndrome of inappropriate secretion of antidiuretic hormone (Schwartz-Bartter syndrome) b. Cerebral/central salt wasting c. Other causes of hyponatremia 22.7. Wolfram's syndrome a. Clinical symptoms b. Molecular genetics, differential diagnosis and psychiatric symptoms c. The hypothalamoneurohypophysial system Chapter 23. Eating disorders a. Hypothalamic nuclei involved b. Leptin c. Neuropeptides and hormones involved

xxvii

xxviii

C O N T E N T S O R P A R T II

23.1.

23.2.

23.3.

d. Molecular genetic factors involved in obesity e. Epigenetic factors in obesity and anorexia cachexia Prader-Willi syndrome a. Symptoms and molecular genetics b. Hypothalamic abnormalities c. Behavioral disorders d. Comorbidity Anorexia nervosa and bulimia nervosa a. Symptoms b. Hypothalamic tumors mimicking anorexia nervosa c. Association with other disorders d. Therapy Other eating disorders a. Lanrence-MoordBardet-Biedl syndrome b. Biemond's syndrome c. AlstrOm's syndrome d. Night eating syndrome e. Binge eating disorder f. Miscellaneous

Chapter 24. Reproduction, olfaction and sexual behavior Disorders of gonadotropic hormone regulation 24.1. a. Hypogonadotropic hypogonadism b. Disorders of puberty c. The hypothalamopituitary gonadal axis in aging and menopause d. Polycystic ovary syndrome Olfaction, anosmia, the vomeronasal organ (Jacobson's organ) and the embryology of 24.2. LHRH neurons a. Olfaction b. Anosmia c. Neurological and psychiatric diseases d. Olfaction and sex: vomeronasal organ and the LHRH neurons of the preoptic area 24.3. Kallmann' s syndrome a. Molecular genetics and migration b. Functional deficits c. Endocrine disorders 24.4. Klinefelter's syndrome or testicular dysgenesis a. Clinical presentation b. Psychosocial problems Sexual differentiation of the brain and sexual behavior 24.5. a. Mechanism of sexual differentiation of the brain b. Sexual differentiation, the hypothalamus and amygdala c. Transsexuality and other gender identity problems d. Homosexuality e. Sexual dysfunction in hypothalamopituitary disorders Chapter 25. Hypothalamic lesions following trauma and iatrogenic disorders 25.1. Head/brain injury


25.2. 25.3.

25.4.

Chapter 26. 26.1. 26.2. 26.3. 26.4.

26.5. 26.6.

26.7.

26.8.

26.9.

Neuroleptic malignant syndrome Hypothalamic injury by radiation a. Hypothalamic symptoms following radiation of tumors b. Tumors after whose treatment these symptoms were found c. Yttrium (Y)-90 implantation in the pituitary d. Postradiation tumors e. Vascular complications f. Other complications Lesion of the pituitary stalk

Hypothalamic involvement in psychiatric disorders Psychiatric symptoms due to tumors of the third ventricle Attacks of laughter (gelastic epilepsy) Ventromedial hypothalamus syndrome and the effect of lesions on aggression Depression and mania a. Depression and neuropeptides b. Amines in the hypothalamus and depression c. Other factors and brain structures involved in the pathogenesis of depression d. CRH neurons and the symptoms of depression e. Oxytocin and vasopressin neurons and the symptoms of depression f. Biological rhythms in mood disorders g. Light therapy and the circadian system h. Other therapeutic interventions i. The thyroid axis j. Sex hormones, depression, premenstrual syndrome, antepartum depression and postpartum mood disorder k. Mania The hypothalamus in mental deficiency Obsessive-compulsive disorder a. Neuroendocrine changes b. Neuroendocrine therapies Anxiety disorders a. Panic disorder b. Social anxiety disorder Fatigue syndromes a. Chronic fatigue syndrome b. Fibromyalgic syndrome c. Postviral fatigue syndrome Aggressive behavior a. Developmental factors involved in clinical disorders associated with aggression b. Hypothalamic structures involved c. Sex hormones and aggression d. Stereotactic hypothalotomy

Chapter 27. Schizophrenia and autism 27.1. Schizophrenia a. A developmental disturbance b. Hypothalamic involvement c. Hypothalamic neurotransmitters, neuromodulators and neurohormones 27.2. Autism

xxix

XXX

CONTENTS

O F P A R T II

Chapter 28. Periodic disorders 28.1. Kleine-Levin syndrome (periodic somnolence and morbid hunger) Spontaneous periodic fever, hypothermia, Shapiro syndrome and periodic 28.2. Cushing's syndrome Acute intelTnittent porphyria 28.3. 28.4. Narcolepsy 28.5. Epileptic seizures a. Epilepsy, diurnal rhythms and sleep b. Epilepsy and hormone release c. Hypothalamic hamartomas and epilepsy d. Hypothalamic pathology in epilepsy Chapter 29. Neurodegenerative disorders 29.1. Alzheimer's disease and the hypothalamus a. Conventional neuropathology b. Sex differences and sex hormones c. Down's syndrome d. Hyperphosphorylated tan and [3-amyloid e. A[3 immunoreactivity f. Abnormally phosphorylated tau g. Relationship between Alzheimer neuropathology and decreased metabolism h. Hypothalamic changes in neuroactive substances in AD 29.2. Dementia with argyrophilic grains 29.3. Parkinson' s disease a. Autonomic symptoms b. Sleep and circadian rhythms c. Depression d. Hormones and neuropeptides in the hypothalamus e. Lewy bodies in the hypothalamus and adjacent areas Huntington's disease 29.4. 29.5. Wernicke's encephalopathy, Korsakoff's psychosis and Marchiafava-Bignami disease 29.6. Adrenomyeloneuropathy, adrenoleukodystrophy and hypothalamic-pituitary dysfunction 29.7. Other neurodegenerative disorders a. Frontotemporal dementia and parkinsonism linked to chromosome 17 b. Hippocampal sclerosis c. Progressive supranuclear palsy d. Multisystem atrophy (Shy-Drager syndrome) e. Lewy body disease f. Pick's disease g. Miscellaneous Chaper 30. 30.1. 30.2. 30.3. 30.4.

Autonomic disorders Temperature regulation Disturbed thelrnoregulation Cardiovascular regulation Cardiovascular disturbances


30.5.

30.6. 30.7.

Circumventricular organs: lamina terminalis, subfornical organ and autonomic regulation a. Organum vasculosum lamina terminalis: experimental data b. Human data c. Subfornical organ Micturition Sleep a. Hypothalamic structures involved in sleep b. Neuroendocrine changes in sleep c. Sleep and aging d. Sleep in neurological and other disorders

Chapter 31. Pain and addiction

31.1. 31.2.

31.3.

Opioid peptides and other addictive compounds Pain and the hypothalamus a. The anatomy of pain; hypothalamic structures and symptoms involved b. Placebo analgesia and other placebo effects c. Analgesia by deep brain electrostimulation, stereotactic lesions, acupuncture and TENS Headache a. Cluster headache b. Migraine c. Hypnic headache syndrome

Chapter 32. Miscellaneous hypothalamic syndromes

32.1. 32.2. 32.3. 32.4. 32.5. 32.6.

Idiopathic hypothalamic syndrome of childhood, a paraneoplastic syndrome Hypothalamic atrophy, Leigh's disease and Cornelia de Lange's syndrome Diencephalic idiopathic gliosis Mitochondrial encephalomyopathy, lactic acidosis and stroke-like episodes (MELAS) syndrome Agenesis of the diencephalon Tourette's syndrome

Chapter 33. Brain death and 'dead' neurons

a. The process of dying and brain death b. Postmortem perfusion of hypothalamic tissues and neuronal cultures: life after death References Subject index for Part I and Part II

xxxi

2014 Ch 01

20/11/03

10:02 am

Page 3

Handbook of Clinical Neurology, Vol. 79 (3rd Series Vol. 1) The Human Hypothalamus: Basic and Clinical Aspects, Part I D.F. Swaab, author © 2003 Elsevier B.V. All rights reserved

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 9

CHAPTER 1

Introduction From bench to bed and back

The human hypothalamus is a small (4 cm3; Hofman and Swaab, 1992a) but very complex structure at the base of the brain (Figs. 1.1 and 1.2). No other brain structure contains so many different small cell groups with different structural and molecular organizations that perform entirely different functions. Traditionally, hypothalamic research has focused on “lower” functions, or, as Cushing (1932) poetically phrased it:

resolution of magnetic resonance imaging (MRI) is rapidly improving. First, MRI on cadaver brains revealed the body and postcommissural part of the fornix, and the mamillothalamic tract (Miller et al., 1994). MRI now shows, in vivo, the lamina terminalis, optic nerve, optic chiasm, optic tracts, anterior commissure, corpora mamillaria, tuber cinereum, pituitary stalk, and the posterior pituitary, which is generally present as a high intensity MRI signal (Figs. 1.3 and 1.4; Chapters 16.c; 22.1). In addition, size differences of the third ventricle are measured in relation to Alzheimer’s disease (Chapter 29.1) and schizophrenia (Chapter 27.1). The advent of superconductive MRI allowed these details of the hypothalamus also to be depicted in vivo. Heavily T2-weighted MR images have a high rate of detection of the postcommissural fornix (in 100% of the cases) and even of the mamillothalamic tract (in 64% of the patients). Patients with glioblastoma multi-

“Here in this well-concealed spot, almost to be covered with a thumbnail, lies the very main spring of primitive existence – vegetative, emotional, reproductive – on which with more or less success, man has come to superimpose a cortex of inhibitions.”

However, hypothalamic involvement in “higher” functions such as memory processes (Chapters 13, 16, 29) and mood (Chapter 26.4) is becoming increasingly obvious. The

Fig. 1.1.

Medial surface of the human brain (a: overview), (b: detail with the hypothalamus): ac = anterior commissure, NII = optic nerve, lt = lamina terminalis, oc = optic chiasm, or = optic recess, III = third ventricle, cm = corpus mamillare.

3

2014 Ch 01

4 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:02 am

Page 4

D.F. SWAAB

Fig. 1.2. A block of tissue (frontal cut) containing the hypothalamus and adjacent structures; OC, optic chiasm, OVLT, organum vasculosum laminae terminalis (note that the third ventricle is shining through the thin lamina terminalis), ac = anterior commissure, on top of which the septum with the fornix at both sides is located. The lateral ventricles containing plexus choroidus are present and both sides of the septum and under the CC, corpus callosum.

forme and lacunar infarct at the hypothalamus presented with anterograde amnesia, but only when both the postcommissural fornix and the mamillothalamic tract were injured (Saeki et al., 2001; Figs. 1.3 and 1.4). In Alzheimer patients, atrophy was observed in the basal forebrain, fornix, hypothalamus, mamillary bodies and septal area, by means of MRI (Callen et al., 2001). Functional imaging, too, is beginning to yield information on hypothalamic functions. Positron emission

tomography (PET) showed an increased regional blood flow and oxygen metabolism in the hypothalamus during normal aging and a decreased blood flow and oxygen consumption in dementia (De Reuck et al., 1992). The hypothalamus of females appeared to have a higher glucose metabolism than this structure in males (Kawachi et al., 2002). Hunger was associated with increased cerebral blood flow in the hypothalamus as measured with PET (Tataranni et al., 1999), and after glucose ingestion, subjects demonstrated an inhibition of the functional (f)MRI signal in the areas corresponding to the paraventricular and ventromedial nuclei (Matsuda et al., 1999). Satiation produces a decrease in cerebral bloodflow in obese women (Gautier et al., 2001). fMRI revealed that sexual arousal in males, but not in females, is accompanied by an activation of the part of the hypothalamus located on the right (Arnow et al., 2002; Karama et al., 2002). A step forward was made by the introduction of a temporal clustering analysis technique for fMRI to demonstrate that eating-related neuronal activity peaks at two different times with distinct localization, i.e. the “upper anterior region of the hypothalamus” and the “medial hypothalamus” in relation to the plasma insulin level (Liu et al., 2000a). Interestingly, it turned out that, with PET, women who smell an androgen-like pheromone appeared to activate the hypothalamus, with the center of gravity in the preoptic and ventromedial nuclei. Men, in contrast, activate the regions of the paraventricular and ventromedial nucleus, when they smell an estrogen-like substance (Savic et al., 2001). However, the visualization of the borders of hypothalamic (sub)nuclei and the functional changes in these structures in relation to physiology or pathology is at present still only possible with the help of microscopical techniques on postmortem material, and therefore imaging (metaphorically called “intracranial voyeurism” by Charness, 1999) has so far been of only limited value for the exact localization of hypothalamic processes. The present monograph describes the functional microscopic anatomy of the human hypothalamus and adjacent structures, the stalk/median eminence, the neurohypophysis and the pineal gland as intrinsic parts of the circadian system in health and disease, mainly on the basis of postmortem tissue, clinical information, imaging, and endocrine and other functional data, in an attempt to integrate bench and bedside traditions. Part I essentially works its way from rostral to caudal through a block of brain tissue containing the hypothalamus and adjacent structures and describes the cytoarchitecture,

2014 Ch 01

20/11/03

10:02 am

Page 5

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

5

Fig. 1.3. Axial MRI images of five consecutive sections in one healthy subject shown in a–e and one section in another subject (f). a: Fornical columns (arrow) were observed in one section level above the anterior commissure. b: Postcommissural fornix (PF; white arrow) was observed as high-signal-intensity spot just behind the anterior commissure (black arrow) and exposed to the cerebrospinal fluid space of the third ventricle. c: In the next lower section, PF (arrow) was still exposed to the third ventricle. d: In the next lower section, ovoid PF (white arrow) directing posteriorly toward the mamillary body was seen. Ill-defined high-signal-intensity spot (MT; black arrow) was visible 4 mm posterior to PF. e: At the next level, PF (white arrow) directing posteriorly toward the mamillary body (black arrow) anteriorly and laterally. f: PF (white arrow) was visible but ill-defined bilaterally. Obscure PF was seen in 6% of healthy subjects. Mamillothalamic tract (MT; black arrow) was identifiable in this case. (Saeki et al., 2001; Fig. 2, with permission.)

molecular-anatomical organization, functional neuroanatomy and topographic neuropathology of hypothalamic nuclei. Part II describes the neuropathology of the hypothalamus and neurohypophysis and is basically disease- and system-oriented. The physiology and pathology of the adenohypophysis have been described in a large number of reviews and papers (e.g. Melmed, 1995; Horvath et al., 1997; Wierman, 1997; Lamberts et al., 1998) and have not been included in this monograph.

1.1. Anatomical borders of the hypothalamus Nomenclature is man-made; there is strictly speaking no correct and no incorrect way of designating nuclear groups of a region, except as certain names are sanctioned by usage. (Crosby et al., 1962)

The first to mention the hypothalamus as a distinct neuroanatomical entity was the Swiss anatomist Wilhelm His in 1893. More than one hundred years ago he proposed a subdivision of the brain on the basis of embryological

2014 Ch 01

6 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:02 am

Page 6

D.F. SWAAB

Fig. 1.4. Coronal MRI images of four consecutive sections in one healthy subject (a–d) and two sections in another subject (e and f). a: The fornical columns (large arrow) were evident at the anterior commissure (small arrow) section. b: In the next posterior section, the descending portion of the postcommissural fornix (PF; arrow) was identifiable, more clearly on the left. c: In the next posterior section, the descending portion of PF (arrow) was evident bilaterally, above the floor of the third ventricle. It was visible in 80% of the healthy subjects. d: In the next section, the mamillary body (arrow) was identifiable. At this level, the entry point of the PF was unidentifiable. e: The fornical columns (arrow) were evident. f: At the level of the mamillary body (white arrow), the origin of the mamillothalamic tract (MT) was visible. (Saeki et al., 2000; Fig 4. with permission.)

development. The point of departure was the five brainvesicles model described by Von Baer in 1828. Wilhelm His subdivided the second of these vesicles, the diencephalon, into three regions: epithalamus, thalamus and hypothalamus, which were arranged as longitudinal zones in superposition to one another. Specification of the hypothalamus in development occurs in two steps: early signals are required before or during gastrulation for forebrain induction; ventralizing and rostralizing signals provided by the axial mesendoderm are required later, to induce

cell types in the presumptive hypothalamus (Michaud, 2001). Homeobox gene Nkx2-1 expression in the hypothalamic anlage is required to maintain molecular characteristics of the developing hypothalamus and to repress molecular characteristics of dorsal alar fates (Marin et al., 2002). The diencephalon and telencephalon can be detected as early as 4 weeks postfertilization (O’Rahilly and Müller, 1999), and the fetal development of the hypothalamus from 9–10 weeks of gestation is discussed in Chapter 1.6. The exact borders of the

2014 Ch 01

20/11/03

10:02 am

Page 7

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

hypothalamus (Figs. 1.1 and 1.2) are rather arbitrary and the exact terminology has often been controversial (Le Gros Clark, 1938; Saper, 1990), but the borders are generally considered to be: rostrally, the lamina terminalis (Figs. 1.1 and 1.2) and caudally, the plane through the posterior issure and the posterior edge of the mamillary body (Fig. 1.1) or mamillo-thalamic tract or the bundle of Vicq d’Azyr (Wahren, 1959; Figs. 1.4 and 1.9). It should be noted, however, that the Ch4 region of the nucleus basalis of Meynert extends even more caudally than the mamillary bodies (Fig. 2.1). As first proposed by His in 1893 (Anderson and Haymaker, 1974), the hypothalamic sulcus (Fig. 1.9) is generally looked upon as the dorsal border. In frontal sections the hypothalamic sulcus is 1 cm more lateral, indeed, at about the level of the most ventral part of the thalamus. However, in the zone along the wall of the third ventricle, the hypothalamus continues in a dorsal direction. The paraventricular nucleus, for instance, is often found both ventrally and dorsally of the hypothalamic sulcus (Fig. 1.5). The anterior commissure (Figs. 1.1 and 1.2) has also been mentioned as a dorsal border of the hypothalamus (Wahren, 1959), but this structure might penetrate the third ventricle on different levels. Another complication is that the hypothalamus blends into the septum verum (cf. Andy and Stephan, 1968; Horváth and Palkovits, 1987). One cannot simply point to the septal nuclei as the dorsal borders of the hypothalamus (Wahren, 1959), since several chemically defined cell types seem to pass this border. For instance, LHRH neurons are scattered over the dorsal preoptic area, septum and bed nucleus of the stria terminalis (Rance et al., 1994; Chapters 7; 24.2), and the preoptic region, anterior hypothalamus and the central nucleus of the bed nucleus of the stria terminalis (BST; Chapter 7) are situated on the junction of the septum and the hypothalamus, partly dorsally and partly ventrally of the anterior commissure (Lesur et al., 1989; Walter et al., 1991; Figs. 7.1 and 7.2). The “telencephalic” (Puelles et al., 2000) BST and the septum are therefore included in this monograph. The ventral border of the hypothalamus includes the floor of the third ventricle that blends into the infundibulum of the neurohypophysis (Fig. 1.9). The exact location of the lateral boundaries, i.e. the striatum/nucleus accumbens, amygdala, the posterior limb of the internal capsule and basis pedunculi and, more caudodorsally, the border of the subthalamic nucleus (Chapter 15), is not a matter of clear-cut certainty either (Nauta and Haymaker

7

Fig. 1.5. The paraventricular nucleus (pvn), one of the major hypothalamic nuclei, is situated both ventrally and dorsally from the hypothalamic sulcus (arrow) as is shown in this corticotropin-releasing hormone staining. This illustrates that the level of the hypothalamic sulcus is not a correct dorsal boundary of the hypothalamus (v = third ventricle; bar = 300 m). (Photograph: Dr. V. Goncharuk.)

1969; Braak and Braak 1992). Cell types do not respect hypothalamic boundaries, as already declared by Malon in 1910 (Anderson and Haymaker, 1974). This monograph does, therefore, not deal with the question of which structure does or does not belong to the hypothalamus sensu stricto or sensu lato on the basis of their embryology or adult hypothalamic borders, or what the correct name of the various (sub)nuclei would be. All major areas and nuclei that were present when the “hypothalamus” was dissected en bloc were therefore included pragmatically (Figs. 1.2 and 2.1) in order to provide a basis

2014 Ch 01

8 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:02 am

Page 8

D.F. SWAAB

Fig. 1.6. Schematic representation of the nuclei of the human hypothalamus. Abbreviations: Ox: optic chiasma, NBM: nucleus basalis of Meynert, hDBB: horizontal limb of the diagonal band of Broca, SDN: sexually dimorphic nucleus of the preoptic area, SCN: suprachiasmatic nucleus, BST: bed nucleus of the stria terminalis, (c = centralis; m = medialis; l = lateralis; p = posterior); PVN: paraventricular nucleus, SON: supraoptic nucleus, DPe: periventricular nucleus dorsal zone, VPe: periventricular nucleus ventral zone, fx: fornix, 3V: third ventricle, ac: anterior commissure, VMN: ventromedial hypothalamic nucleus, INF: infundibular nucleus, OT: optic tract, MB: mamillary body i.e. MMN: medial mamillary nucleus + LMN: lateromamillary nucleus, cp: cerebral peduncle. (Adapted from Fernández-Guasti et al., 2000; Fig. 2.)

for neurobiological and neuropathological research of this brain region, including such structures as the basal cholinergic nuclei (Chapter 2), i.e. presumed telencephalic structures such as the diagonal band of Broca and the nucleus basalis of Meynert (Puelles et al., 2000), which is considered to be the lateral border of the hypothalamus by Wahren (1959), the septum pellucidum (Chapter 18.8), considered to be an archipallial splitting (Macchi, 1951), and the zona incerta (Chapter 15), considered to be a mesencephalic structure. In addition, the epithalamic pineal gland and its hormone melatonin are included (Chapter 4.5) because it is an intrinsic part of the circadian timekeeping system. For the same reason tumors of the pineal region are discussed (Chapter 19.7). Most authors distinguish three hypothalamic regions (Saper, 1990): (i) the chiasmatic or preoptic region (Figs. 1.6, 1.7 and 2.1; containing, e.g. the suprachiasmatic nucleus, the sexually dimorphic nucleus, and the supraoptic and paraventricular nucleus). It should be noted here that the paraventricular nucleus runs in a caudal

direction, all the way to the caudal border of the hypothalamus (Figs. 1.8, 1.9; Young and Stanton, 1994). In addition, the diagonal band of Broca, the nucleus basalis of Meynert, the islands of Calleja and the BST are considered in connection with the chiasmatic region; (ii) the cone-shaped tuberal region (Figs. 1.6, 1.9 and 2.1) surrounds the infundibular recess and extends to the neurohypophysis. It contains the ventromedial, dorsomedial and infundibular or arcuate nucleus. Lateral structures of this region are the lateral tuberal nucleus and the tuberomamillary1 nucleus (Fig. 1.8). The most caudal region is (iii) the posterior or mamillary region, which is dominated by the mamillary bodies that abut the midbrain tegmentum and contains the medial and lateral mamillary nucleus (Saper, 1990; Braak and Braak, 1992; Figs. 1.6 and 1.8). This region also includes the posterior hypothalamic nucleus and the incerto hypothalamic cell group (Chapters 13 and 15). Moreover, the subthalamic nucleus (Chapter 15) is included. This is a diencephalic cell group that develops in the caudal part

2014 Ch 01

20/11/03

10:02 am

Page 9

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

9

Fig. 1.7. Thionine- (left) and anti-vasopressin (right)-stained section through the chiasmatic or preoptic region of the hypothalamus. OC = optic chiasm, OVLT = organum vasculosum lamina terminalis, PVN = paraventricular nucleus, SCN = suprachiasmatic nucleus, SDN = sexually dimorphic nucleus of the preoptic area (intermediate nucleus, INAH-1), SON = supraoptic nucleus, III = third ventricle. Bar represents 1 mm.

of the hypothalamus but migrates to a position above the cerebral peduncle (Jiao et al., 2000). 1.2. Strategic research and structure–function relationships If you want to understand function, study structure.

The hypothalamus has a number of unique properties that also render it very suitable for fundamental neurobiological research of structure–function relationships. In the first place it contains, in addition to conventional neurons, neuroendocrine cells whose activity can be monitored by the measurement of plasma or urine levels of hormones secreted by these cells. Moreover, many hypothalamic nuclei can easily be delineated (Figs. 1.6 and 1.7), which makes it possible to monitor the basic

1

processes, such as cell formation, migration, maturation, sexual differentiation and cell death, quantitatively per brain area (Gahr, 1997). Although it is not necessary to delineate brain nuclei in order to determine numbers of a particular cell type (Vogels, 1997), it is impossible to determine the number of cells in a particular nucleus that are also present outside its borders, such as glial cells or small neurons, without such a delineation. The neurotransmitter, neuromodulator or neurohormonal content of many of the hypothalamic nuclei is currently becoming better known and so are their specific functions: the suprachiasmatic nucleus (Chapter 4) is the hypothalamic clock that regulates circadian and circannual rhythms; the vasopressin neurons of the supraoptic and paraventricular nuclei (Chapter 8) are involved in antidiuresis, the oxytocin neurons in reproduction, sexual behavior and eating behavior, the corticotropin-releasing hormone neurons of the paraventricular nucleus are of pivotal

The term is written as “mamillary” as it originates from mamilla and not from mamma (Lantos et al., 1995).

2014 Ch 01

10 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:02 am

Page 10

D.F. SWAAB

Fig. 1.8. (A) Three-dimensional reconstruction of the hypothalamus, displaying the cell-dense nuclei and numbered as follows: 1 = suprachiasmatic nucleus, 2 = supraoptic nucleus, 3 = interstitial nucleus of the anterior hypothalamus-1 (INAH-1 = SDN-POA), 4 = INAH-3 (note that this nucleus appears to form a medial, cell-dense border of a more diffuse circular structure), 5 = INAH-4, 6 = paraventricular nucleus, 7 = arcuate nucleus, 8 = ventromedial nucleus, 9 = dorsolateral nucleus, 10 = tuberomamillary nucleus, 11 = lateral tuberal nucleus, 12 = medial mamillary nucleus, 13 = supramamillary nucleus. The diminutive lateral mamillary nucleus was present only in a single, damaged section and is not displayed here. Magnification = 11.65. This figure illustrates the hypothalamus as viewed from its lateral aspect, using the following graphic rotation parameters: x axis rotated 20°, y axis rotated 55°, and z axis rotated 340°. (B) Medial view of the hypothalamus as seen from within the third ventricle: x axis rotated 10°, y axis rotated 310°, and z axis rotated 0°. Every fifth tissue section was traced for these views, so that the interval between sections = 320 microns. (From Young and Stanton, 1994; Fig. 2.)

2014 Ch 01

20/11/03

10:03 am

Page 11

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

11

Fig. 1.9. Schematic representation of the major hypothalamic nuclei. Lateral to the fornix and the mamillothalamic tract is the lateral hypothalamic area (in red), in which the tuberomamillary nucleus (in pink) is situated. Situated rostrally in this area is the lateral preoptic nucleus. Surrounding the fornix is the perifornical nucleus (represented as a red band), which joins the lateral hypothalamic area with the posterior hypothalamic nucleus. The medially situated nuclei (in yellow) fill much of the region between the mamillothalamic tract and the lamina terminalis. The nuclei tuberis laterales (in blue) are situated at the base of the hypothalamus, mostly in the lateral hypothalamic area. The supraoptic nucleus (in green) consists of three parts (From Nauta and Haymaker, 1969; Fig. 4.3). Note that the nucleus tuberalis lateralis (chapter 12) is depicted too small and that the suprachiasmatic nucleus is lacking.

2014 Ch 01

12 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 12

D.F. SWAAB

importance when it comes to a stress response, and the thyrotropin-releasing hormone-producing neurons of this nucleus play a vital part in thyroid regulation. Certainly not all hypothalamic neurons have yet been chemically characterized. A new releasing hormone for prolactin production was discovered some years ago on the basis of the presence of its orphan receptors in the human pituitary (Hinuma et al., 1998), and the supraoptic and paraventricular nuclei contain many neuroactive compounds whose functions are not known (Chapter 8.7). New peptides and receptors are continuously being found, but all the typical properties mentioned above make the hypothalamus an extremely suitable brain area for the study of structure–function relationships (Table 1.2). Nevertheless, relatively few neuroscientists are involved in the study of the human hypothalamus, and knowledge of its neuropathology is scarce (cf. Treip, 1992; Horvath et al., 1997; Part II). Not only is the human hypothalamus involved in a wide range of functions in the developing, adult and aging subject, it also plays a role in various diseases of different etiologies; this will necessitate strategic research for the period to come. Alterations in hypothalamic structures and functions are thought to be operative in signs and symptoms of diseases such as anorexia and bulimia nervosa (Chapter 23.2), depression (Chapter 26.4), diabetes insipidus (Chapter 22.2), Wolfram’s syndrome (Chapter 22.7), Prader–Willi syndrome (Chapter 23.1), narcolepsy (Chapter 28.4) and the malignant neuroleptic syndrome (Chapter 25.2), as well as in disturbances in cardiovascular and temperature regulation (Chapter 30). In addition, the hypothalamus is crucial for the expression of emotions (Chapter 26). A motor center for laughter has been hypothesized to be located in the caudal part of the hypothalamus (Martin, 1950), since ictal laughter is associated with hypothalamic hamartomas in that area (Chapter 26.2). The preoptic area and the posterior hypothalamus are also presumed to be involved in aggression, and the latter has even been a target for controversial stereotactic psychosurgical procedures that were claimed to prevent aggressive crises or violent behavior (Schvarcz et al., 1972). Alterations in the hypothalamus have been found in sudden-infant-death syndrome (Chapter 8.7) and in neuro-degenerative diseases, which may lead to particular symptoms in, e.g. Alzheimer’s, Parkinson’s, and Huntington’s disease (Chapter 29), and also in multiple sclerosis (Chapter 21.2). Moreover, this brain region is presumed to change as a result of endocrine effects on brain development in congenital adrenal

hyperplasia syndrome due to hormones administered during development (e.g. diethylstibestrol (DES)), as well as in transsexuality and in Turner’s, Klinefelter’s, and Kallmann’s syndrome (Chapter 24). Attention is now paid to the relationship between the structural development of the human hypothalamus, gender and sexual orientation (Swaab et al., 1992a; Swaab and Hofman, 1995; Chapters 5, 6, 24.5). More than half a century ago, Morgan (1939) investigated the hypothalamus for mental deficiency in 16 institutionalized subjects. According to Morgan, pathological involvement of the third ventricle region was evident in all but two cases. The tuberomamillary nucleus was the only cell group in the hypothalamus which did not show a marked reduction in cell density, which led Morgan to conclude “. . . that the hypothalamus plays an important role in the etiology of mental deficiency . . .”, an idea that has so far not been sufficiently followed up utilizing the assistance of modern research techniques (Chapter 26.5). 1.3. The autopsy and brain banking (Fig. 1A) Strange coincidence that all people whose heads have been opened turned out to have a brain. L. Wittgenstein, Über Gewißheit, §207

Since the imaging techniques of the living brain are now able to reveal more and more details, clinicians lose interest in the substrate of the disease and autopsy rates have dropped sharply. However, autopsies and the resulting brain material remain crucial, not only for patient care, where discrepancies are regularly found between the clinical diagnosis and the final necropsy audit, but also for fundamental research on the human brain, and for strategic research of pathogenetic mechanisms of brain diseases. The revision of laws and autopsy rules that have become more restricted in many countries during the last decades, with consent required from next-of-kin in most countries, may also have contributed to the drop in autopsy rate in hospitals in several nations (Svendsen and Hill, 1987). In addition, religious objections to autopsy may play a role (Boglioli and Taff, 1990). Because of all the important local cultural differences, the collection of brain material should, of course, take place within the framework of medico-legal and ethical guide lines, which are concomittant with local legislation (Cruz-Sánchez et al., 1997). Patients suffering from the various neurological, psychiatric or neuroendocrine disorders that are discussed

2014 Ch 01

20/11/03

10:03 am

Page 13

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

13

Fig. 1A. The central remaining fragment of Rembrandt’s (1656) painting of Dr. Deyman’s anatomical lecture. Amsterdam Historical Museum. Dr. Deyman’s anatomical lecture lasted 3 days and was open to the public for the amount of 20 cents. In the middle the praelector and doctor medicinae Jan Deyman is lifting the falx cerebri with a lancet in order to show the soul, which was thought to be localized in the pineal gland of the dissected body. This was an integral part of the punishment. The college master, Gijsbert Calkoen, waits patiently to collect the brain in the skull of the thief Joris Fonteyn (‘Black John’) who had just been executed. In 1723 a large part of this painting was destroyed by a fire in the medieval gate “De Waag”, where the teatrum anatomicum was situated. (Amsterdams Historisch Museum, with permission.)

2014 Ch 01

14 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 14

D.F. SWAAB

in this monograph generally do not die “conveniently” in a university hospital, which is why it is often very difficult to obtain well-documented brain material. This is a problem generally encountered when postmortem brain material has to be studied from patients with disorders such as dementias, psychiatric disorders such as depression, or conditions such as transsexuality. And, if these patients do die in a university hospital, the autopsy is traditionally primarily meant for diagnostic purposes and not for neurobiological research. These problems are even more serious when brain samples have to be collected from controls, i.e. patients without a brain disease. This is the reason why, in 1985, I decided to set up the Netherlands Brain Bank (NBB), in collaboration with Prof. F.C. Stam. This facility provides research groups in the Netherlands and abroad with clinically and neuropathologically well-documented, prospectively collected postmortem brain material from patients who suffered from neurological, psychiatric, or neuroendocrine disorders. The patients and/or the next of kin authorize the NBB to do a brain autopsy, with the aim of collecting brain tissue for research purposes and to have access to all the donors’ medical records after their demise. Since its establishment in 1985, the Netherlands Brain Bank has provided over 390 projects in 22 countries worldwide with material from more than 2400 autopsies. Many of our data on the human hypothalamus as they are presented in this monograph are based on findings carried out with the help of material obtained by the NBB. The NBB provides brain specimens on request, via research protocols that have been submitted in advance, and that specify the requirements of the research groups. Special emphasis is placed on rapid autopsies with a short postmortem delay – between 2 and 8 hours. Cerebrospinal fluid (CSF) is collected from a lateral brain ventricle and centrifuged in order to remove the cells, after which the pH is determined to establish the agonal state (see below, Fig. 1.13). CSF is also stored in aliquots and used to develop diagnostic tests. Once the brain is removed, it is dissected following a protocol that is different for each disorder, apart from the standard set of structures that is neuropathologically investigated in all autopsies. The fresh-dissection protocol requires highly qualified staff that have to be available around the clock. Some 80 different structures are dissected according to neuroanatomical borders: some 10 for the neuropathological diagnosis, and the rest for research purposes. On request, the brain specimens are rapidly frozen in liquid nitrogen, slowly frozen in sucrose, fixed, or used

immediately, i.e. for tissue culture or postmortem tracing (Chapter 33). (a) Clinical diagnosis Tissues obtained via the NBB are generally accompanied by a comprehensive medical history. Medical records are secured with the informed consent of the patient and/or next of kin, generally requested long in advance, together with the informed consent for performing a brain autopsy and for the use of the brain tissue and medical information for research purposes. The definitive diagnosis, established following neuropathological examination, is sent to the patient’s physician as well as to the various research groups. Although the clinical diagnosis generally has a high validity for Alzheimer’s disease (Li et al., 1997), clinical misclassification does occur, although not as frequently as in other diseases such as Pick’s. In principle, the samples are not shipped until after the neuropathological diagnosis has been completed. This is also the procedure for reasons of safety – it ensures as much as possible that no contagious tissue or fluids are sent. Good clinical information on the donors has so far prevented unexpected high-risk neuropathological autopsies such as AIDS and Creutzfeldt–Jacob disease to enter the rapid autopsy/fresh dissection protocol. The contagious brains of these patients are not freshly dissected (to prevent aerosol formation) but placed in 10% formalin (Ironside and Bell, 1996). Although the risk of sending infectious material to research groups is minimized this way, the research groups are asked to treat all specimens as risk bearers. This definitely goes for procedures such as homogenization, which causes aerosols to be formed and should be performed under a hood. Although the majority of the donors, both patients and controls, sign up for the NBB programme well in advance, so that there is plenty of time to obtain the necessary clinical information, there are always rare and interesting cases that do not follow this pattern. If brain material becomes available without sufficient clinical data, diagnosis by means of “psychological autopsy” has proved to be a promising possibility to consider, e.g. for depression (Kelly and Mann, 1996). These authors established risk factors associated with suicide. Comparison of the DSM-III R chart diagnosis generated by clinicians who had treated the subjects prior to death and the independently obtained postmortem diagnosis obtained following a structured interview with next of kin correlated very well and provided evidence for the

2014 Ch 01

20/11/03

10:03 am

Page 15

15

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

validity of the psychological autopsy as a method of determining psychiatric diagnosis. (b) Removal and documentation of the hypothalamus “Hic gaudet mors succurrere vitae” (here death joyfully helps life). Anatomic Hall of the University of Heidelberg Cruz-Sánchez et al., 1997

There are a number of specific reasons that make it difficult to perform postmortem studies on the hypothalamus. In the first place, the structure is easily damaged during brain autopsy by traction on the optic nerves and pituitary stalk, by which, e.g. the lamina terminalis, median eminence and infundibulum are often seriously damaged. Moreover, pathologists traditionally make a cut right through the optic chiasm, thereby damaging the anterior hypothalamus. For a study of the hypothalamus, the front cut should be about 2 mm in front of the optic chiasm, leaving the thin lamina terminalis intact, and the back cut just behind the mamillary bodies. The slides are then trimmed at a convenient width. Bisection of the hypothalamus often gives rise to problems in the examination of midline structures like the nucleus infundibularis. The most important reasons for the fact that the hypothalamus is so rarely studied are, however, that it can be studied adequately only by serial coronal sectioning, that its chemical neuroanatomy is very complex, and that immunocytochemistry and morphometry are often required to establish alterations. In addition, a number of ante- and postmortem factors (see below) may influence the morphology and neurochemistry of the hypothalamus. Data concerning these factors should be collected when the hypothalamus is studied (for review see Ravid et al., 1992). 1.4. Confounding factors Kein Hypothalamus sieht wie die ander aus.2 Grünthal, 1950

Observations on the hypothalamus may be confounded by a large number of factors before, during and after death. Material should be matched for such factors with appropriate controls or their effect on the measurements should be corrected for, as we did for, e.g. the effect

2

No two hypothalami are alike.

on storage time on the amount of CRH mRNA in the paraventricular nucleus in depression (Raadsheer et al., 1995). In addition to matched controls that did not die of a neurological or psychiatric disease, samples from related disorders are often useful to control for disease specificity. It is evident that the list of confounding factors will increase rapidly in the future. (a) Antemortem factors Age (Figs. 1B and 1C) And so from hour to hour, we ripe and ripe, And then, from hour to hour, we rot and rot, And thereby hangs a tale. Shakespeare, As You Like It. Act ii. Sc. 7

Age-related changes occur in many, if not all, structures in the hypothalamus. A decrease in volume and vasopressin cell number is, e.g. observed in the suprachiasmatic nucleus (SCN) in senescence (80–100 years) (Swaab et al., 1985; Chapter 4.3). The corpus mamillare decreases in size with age (Chapter 16). Another hypothalamic nucleus which shows clear age-related changes is the sexually dimorphic nucleus of the preoptic area (SDN-POA; Swaab and Fliers, 1985; Swaab and Hofman, 1988; Hofman and Swaab, 1989; Chapter 5). The SDN-POA cell number reaches a peak value at the age of 2–4 years (Fig. 5.4). Only after this age does sexual differentiation become manifest. The nucleus decreases greatly with age in a sexually dimorphic way as far as volume and cell number with age are concerned (Chapter 5; Fig. 5.6). Age-related chemical changes have also been described. Gliosis in the human hypothalamus during aging is accompanied by activation of monoamine oxidase-B activity, intensification of lipid peroxidation leading to lipofuscin, and inhibition of succinate dehydrogenase, a key enzyme in tissue respiration (Shemyakov et al., 2001). Estradiol concentrations in hypothalamic areas were significantly higher in fertile women than in postmenopausal women (Bixo et al., 1995). Age must also be taken into consideration when studying monoamines, and their metabolites and enzyme activities in the human hypothalamus (Adolfsson et al., 1979). Aging is certainly not only accompanied by degenerative signs. Activation in the course of aging is found, e.g. in the vasopressin neurons of the supraoptic nucleus of postmenopausal women (Chapter 8.3) in the

2014 Ch 01

20/11/03

10:03 am

Page 16

16 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

D.F. SWAAB

corticotropin-releasing hormone neurons of the paraventricular nucleus (Chapter 8.5) and in the subventricular nucleus in postmenopausal women (Chapter 11). Sex No one supposes that all the individuals of the same species are cast in the very same mould. These individual differences are highly important for us.

Fig. 1B. Las Edades y la Muerte (The ages of man and death). Hans Baldung Green (1484/85–1545). Museo del Prado, Madrid, Spain (No.

2220 Cat.). (With permission.)

An obvious sex difference was observed in hypothalamic glucose metabolism that was higher in women than in men, as measured by PET (Kawachi et al., 2002). There is also an increasing amount of data concerning morphological and functional sex differences on the level of the various nuclei of the hypothalamus. Morphometric analysis revealed that there is a striking sexual dimorphism in the size and cell number in the SDN-POA (Swaab and Hofman, 1984; Swaab and Fliers, 1985; Hofman and Swaab, 1989; Chapter 5). Sexual differentiation of the human SDN-POA occurs after 4 years postnatally, and only after this age does the nucleus differentiate according to sex (Fig. 5.5). This is due to a decrease in both volume and cell number in women, whereas in men it remains unaltered up to the fifth decade, after which a marked decrease in cell number is also observed (see Fig. 5.6). Sexual differentiation has also been reported for two other cell groups in the preoptic-anterior hypothalamus (INAH 2 and 3) (Allen et al., 1989a; Chapter 6). These areas (Fig. 6.1) were found to be larger in males than in females and this was later partly confirmed by LeVay (1991) and Byne et al. (2000). Other brain regions with a larger volume in males than in females are the darkly staining posteromedial part of the bed nucleus of the stria terminalis (BNST-dspm), described by Allen et al. (1990), and the central nucleus of the BST (Zhou et al., 1995c; Kruijver et al., 2000; Chung et al., 2002; Chapter 7, Fig. 7.2). The size of the anterior commissure is sexually dimorphic (Chapter 6.4). The sex difference in the number of vasoactive intestinal polypeptide-expressing neurons in the SCN is age-dependent (Zhou et al., 1995b; see Chapter 4.2; Fig. 4.25). The infundibular nucleus shows sex-dependent Alzheimer changes (Chapters 6.2; 29.1b). 5-Hydroxyindoleacetic acid (5-HIAA) levels in the hypothalamus of males were lower than those in females, indicating a higher turnover rate of serotonin in the female brain (Gottfries et al., 1974). The study that showed that females had more neurons than males in the median raphe nucleus (Cordero et al., 2000) is in line with this idea. We observed clear

2014 Ch 01

20/11/03

10:03 am

Page 17

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

Fig. 1C.

17

The five phases of life of man. Illustration for book of rhymes (ca. 1910). Jewish Historical Museum, Amsterdam (With permission.)

age-dependent sex differences in the activity of the vasopressinergic neurons of the supraoptic nucleus (Chapter 6.5), indicating that functional sex differences may be quite robust. This also holds for sex differences in sex hormone receptor distribution in the hypothalamus and adjacent areas (Chapter 6.3). Sex differences in brain and hormone levels are not only of importance for sexual behavior, they are also thought to be the structural and functional basis of the often pronounced sex differences in the prevalence of neurological and psychiatric diseases. The proportions of cases range from more than 75% women in Rett syndrome, lymphocytic hypophysitis, anorexia and bulimia nervosa and hypnic headache syndrome, to more than 75% male subjects in dyslexia, ADHD, autism, sleep apnea, Gilles de la Tourette syndrome, rabies, Kallmann syndrome and Kleine–Levin syndrome (Table 1.1). Women are more prone to anxiety disorders than men (Seeman, 1997; Piconelli and Wilkinson, 2000). Not only might the number of cases of disorders show clear sex

differences, but the signs and symptoms and the course of the disease might differ also according to sex. Male schizophrenic patients have more severe enlargement of the lateral ventricles (Nopoulos et al., 1997). Men not only suffer from schizophrenia 2.7 times more often than women, they are also prone to a more severe form of this disorder, they have a poorer premorbid functioning experience, an earlier onset, more negative symptoms and cognitive defects, and exhibit a greater number of structural brain abnormalities. Relapses are more severe, and their response to neuroleptic medication is less favorable. Women display more affective symptoms, auditory hallucinations and persecutory delusions (Castle and Murray, 1991; Leung and Chue, 2000). Moreover, an interaction with gender was observed in the second trimester of pregnancy when prenatal exposure to maternal stress was studied as a risk factor for schizophrenia (Van Os and Selten, 1998). Factors that produce normal sexual dimorphism in the brain, particularly in the cortex, may be associated with

2014 Ch 01

18 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 18

D.F. SWAAB

modulating insults producing schizophrenia (Goldstein et al., 2002). Other examples of a sex difference in a neurological disease are those following restricted left-hemisphere lesions, resulting in aphasia in 41% of the males and 11% of the women, whereas manual apraxia was found in 6% of the women and 42% of the men (Kimura and Harshman, 1984). After severe subarachnoid hemorrhage, mortality in women was lower (37%) than in men (53%), while the incidence of favorable outcome was higher in women (42%) than in men (26%). Female traumatic brain injury patients also had a better predicted outcome than male patients (Roof and Hall, 2000). The prevalence of cluster headache in men, and the fact that it is extremely rare prior to adolescence, indicates that sex hormones circulating in adulthood might be involved in the pathogenetic mechanism (Chapter 31.2; Leone and Bussoni, 1993). Sex ratios in the prevalence of disease may show changes over the years: the male/female ratio of cluster headache decreases from 6.2:1 for patients with an onset before 1960 to 2.1:1 for patients with an onset in the 1990s. Changes in life style, in particular in employment rate, and smoking habits of women are held responsible for this change (Manzoni, 1998). Other sex differences remain unaltered. Men still commit 89% of all murders and 99% of all sexual crimes (Spratt, 2000). There may also be a strong effect of age on sex differences in the prevalence of disorders. In depression up to the age of 54, the female/male ratio is 61:39, while after the age of 54 the ratio reverses to 35:65 (Bebbington et al., 1998). There is a significant variation in the male/ female ratio of episodic and chronic cluster headache with respect to age at onset, with the largest difference between 30 and 49 years of age (7.2:1 and 11.0:1) and the lowest after 50 (2.3:1 and 0.6:1, respectively) (Ekbom et al., 2002). There is an excess of men with mental retardation (Turner, 1996). However, female middle-aged Down’s syndrome patients have an earlier onset of dementia than that of male, and a more severe form of Alzheimer’s disease, which correlates with the number of neocortical neurofibrillary tangles rather than with the density of senile plaques (Raghavan et al., 1994). Women run a higher risk of developing dementia after the age of 80 than men (Seeman, 1997; Launer et al., 1999; Letenneur et al., 1999). Although Herbert et al. (2001) suggested that the excess number of women with Alzheimer’s disease is due to the longer life expectancy of women rather than to sex-

specific risk factors, Ruitenberg et al. (2001) found that the incidence of Alzheimer’s disease is higher for women than for men after the age of 90 years. The incidence of vascular dementia was found to be higher for men than for women in all age groups. However, other studies clearly indicate a prevalence of neurodegeneration in some brain structures in men. Very pronounced neurofibrillary Alzheimer changes are found in the infundibular nucleus and adjacent median eminence in 80% of males over 60 years of age and in only 6% of females (Schultz et al., 1996; Chapters 6.2, 11, 29.1). Sex differences are found in many brain structures (Chapter 24.5). With advancing age there is a loss of neurons in the pars cerebellaris loci coeruli that in women begins around the age of 40 and in men already at the age of 20 (Wree et al., 1980). Whether sex differences in the brain that arise in development (“organizing effects”; Chapter 24.5) are indeed the basis for the sex difference in neurological or psychiatric diseases has still to be established. In ADHD an association with androgen receptor haplotypes was found (Comings et al., 1999). Alternative mechanisms that are mentioned are the immediate effects of differences in circulating sex hormone levels (“activating effects”; Chapter 24), caused by sex hormone-stimulated gene transcription (Torpy et al., 1997), as presumed, for example, in sleep apnea (see Table 1.1). A number of the diseases in Table 1.1 are related to changes in catecholaminergic neurons, which are influenced, during development, by direct somatic effects of sex-specific genes (Pilgrim and Reisert, 1992). In a recent Dutch study, a higher prevalence of various psychiatric disorders was found in homosexual people as compared to heterosexual people. These differences seem to be gender-specific, with a higher prevalence of substance-use disorders in homosexual women and a higher prevalence of mood and anxiety disorders in homosexual men, both as compared to their heterosexual counterparts. It is not clear at present whether these differences result from biological or from social factors (Sandfort et al., 2001), but it does mean that sexual orientation (Chapter 24.5) should be included as a possible factor in the study of prevalence in psychiatric and neurological diseases. Nervous system birth defects also show sex differences. Some defects are more prevalent in males, such as, e.g. macrocephaly, while others, such as anencephaly, spina bifida and microcephaly, are more prevalent in females (Lary and Paulozzi, 2001).

2014 Ch 01

20/11/03

10:03 am

Page 19

19

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

TABLE 1.1 Ratios for women over men suffering from a selection of neurological and psychiatric diseases. Disease

Women : Men (%)

Rett syndrome Postoperative hyponatremic encephalopathy with permanent damage or death Anorexia nervosa Lymphocytic hypophysitis True (central) precocious puberty Hypnic headache syndrome Bulimia Senile dementia of the Alzheimer type Multiple sclerosis Anxiety disorder Anencephaly Posttraumatic stress disorders Dementia Unipolar depression, dysthymia Whiplash Severe learning disability Substance abuse Amyotropic lateral sclerosis Stuttering Schizophrenia REM sleep behavioral disorder Male-to-female vs. female-to-male transsexuals Dyslexia ADHD Autism Sleep apnea Kallmann syndrome Rabies REM sleep disorder Gilles de la Tourette syndrome Kleine–Levin syndrome

Use of medicines prior to death The use of corticosteroids decreases not only the amount of CRH in the paraventricular nucleus but also the amount of vasopressin in the supraoptic and paraventricular nucleus (Erkut et al., 1998; Chapters 8.4a and 8.5; Fig. 8.24). The total amount of vasopressin mRNA in the suprachiasmatic nucleus (SCN) of patients that were treated with glucocorticoids was only 50% of that in controls. There was also a 50% decrease in the total number of profiles expressing vasopressin mRNA in corticosteroid-treated people. This may be the biological basis for the circadian rhythm disturbances and sleep impairment in patients receiving glucocorticoid therapy (Liu et al., 2003, submitted; Chapter 4), and the circadian

100 : 96 93 90 90 84 75 74 67 67 67 70 64 63 60 38 34 33 29 27 24 28 23 20 20 18 17 13 13 10 0

0

: 4 : 7 : 10 : 10 : 16 : 25 : 26 : 33 : 33 : 33 : 30 : 36 : 37 : 40 : 62 : 66 : 67 : 71 : 73 : 76 : 72 : 77 : 80 : 80 : 82 : 83 : 87 : 87 : 90 : 100

(Naido, 1997; Chapter 2.5) (Ayus et al., 1992) (Whitaker et al., 1989) (Maghnie et al., 1998a) (Partsch and Sippel, 2001) (Dodick et al., 1998) (Whitaker et al., 1989) (Bachman et al., 1992) (Sadovnik and Ebers, 1993) (Seeman, 1997) (Lary and Panlozzi, 2001) (Breslau et al., 1997; Seeman, 1997) (Bachman et al., 1992) (Regier et al., 1988) (Karlsborg et al., 1997) (Castle and Murray, 1991) (Kessler et al., 1994) (Militello et al., 2002) (Castle and Murray, 1991) (Castle and Murray, 1991) (Schenk et al., 1993) (Bakker et al., 1993; Van Kesteren et al., 1996) (Castle and Murray, 1991) (Comings et al., 1999) (Skuse, 2000) (Block et al., 1979) (Rugarli and Ballabio, 1993) (Gómez-Alonso, 1998) (Schenk and Mahowald, 2002) (Caine et al., 1988) (Critchley, 1962; Chapter 28.1)

disorders in depression (Chapter 26.4f). A very high dose of metamphetamine went together with a 94% depletion of choline acetyltransferase in some autopsy brains (Kish et al., 1999), indicating that the basal forebrain nuclei (see Chapter 2) were affected. Seasonal variation Seasonal alterations have been found in the levels of hypothalamic 5-hydroxytryptamine (5-HT), with a minimum during the months of December and January and a maximum during October and November (Fig. 1.10). Using SPECT, binding to the serotonin transporter in the hypothalamus in healthy subjects was lower in winter than in summer (Neumeister et al., 2000).

2014 Ch 01

20/11/03

20 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

10:03 am

Page 20

D.F. SWAAB

Fig. 1D.

Der Anatom (The Anatomist, 1869). Gabriel von Max (1840–1915). Munich, Neue Pinakothek, Germany (With permission.)

Hypothalamic dopamine showed seasonal variation, with two peaks, i.e. during January and February, and August and September, and two nadirs, i.e. during March to June and October to December. In other brain regions the influence of biorhythms on dopamine levels was less evident (Carlsson et al., 1980a). In the pineal gland a seasonal rhythm was found in gonadotropin receptors, with higher values in winter (Luboshitzky et al., 1997). The circannual rhythms are probably based on the striking seasonal variation we observed in the SCN. The SCN contained 3 times more vasopressin-expressing neurons in October and November than in May and June, at least in young subjects (Hofman and Swaab, 1993; Hofman et al., 1993; see Chapter 4.1; Figs. 4.20 and 4.21). In addition, we found a seasonal variation in the volume of the paraventricular nucleus (PVN), with a peak during spring (Hofman and Swaab, 1992a). The month of death is

consequently a factor to consider in studies on the hypothalamus. Various factors discussed in this section might interact. An example is that the proportion of left-handed people depends on the season of birth. A higher proportion is born in the period of March to July (Martin and Jones, 1999). Circadian variation Clock time of death has been found to be a significant factor for the levels of hypothalamic monoamines. Circadian changes in the hypothalamus were observed in noradrenaline (NA), 5-HT and dopamine (DA) and their metabolites (Fig. 1.11; Carlsson et al., 1980a). The hypothalamic levels of 5-HT were found to be low between 6 a.m. and 3 p.m., and a rapid fall occurred between 5 a.m. and 8 a.m. The hypothalamic NA levels showed no seasonal variation but were found to fluctuate

2014 Ch 01

20/11/03

10:03 am

Page 21

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

21

Fig. 1.10. Level of 5-hydroxytryptamine (5-HT) in human hypothalamus determined postmortem, in relation to the month of death. Shown are the means ± S.E.M. (n) of pooled values of two consecutive months. Statistics: Student’s t-test. (With permission from Carlsson et al., 1980a; Fig. 7.)

significantly during the day in a manner similar to 5-HT. Also the level of DA in the hypothalamus showed significant circadian variation, with a nadir between 6 and 9 a.m. Circadian variation in dopamine levels in brain regions other than the hypothalamus were studied only in a few instances. These fluctuations will most probably have their basis in the circadian changes in the SCN. We observed twice as many vasopressin-expressing neurons in the SCN during the daytime than at night (Hofman and Swaab, 1994; Fig. 4.8). These data show that the clock time of death should be considered as a factor to match for when collecting hypothalamic specimens for research. Lateralization Fixing one hemisphere and freezing the other is current practice in many brain banks. If only one half of the brain is fixed for morphology, the logistic problem of satisfying the needs of both the neuropathologist who wants to see all the structures and the scientist who wants to have samples seems to be solved. Although practical approaches have been developed to solve this type of problem (Perl et al., 2000), the study of only one half of the brain prevents the recognition of possible left– right differences of various systems in the brain. Animal experiments have shown the existence of asymmetry in

neuroendocrine systems (Gerendai and Halász, 1997). So far, not a great deal of attention has been paid to the question of whether the subjects under investigation were left- or right-handed. In the human brain, several functions and transmitters are asymmetrically represented in the left or right hemisphere; lateralization of noradrenaline has been demonstrated in the human hypothalamus (Oke et al., 1978), and there is evidence of a left prominence in the distribution of thyroid-releasing hormone (TRH) in discrete nuclei of the hypothalamus, i.e. in the ventromedial dorsal and paraventricular nuclei (BorsonChazot et al., 1984), with higher concentrations on the left side (Chapter 8.6). We observed that the sex difference in size of the vasopressin neurons in the human paraventricular nucleus is pronounced on the left-hand side and absent on the right-hand side (Ishunina and Swaab, 1999). In schizophrenic patients the left (but not the right) mamillary body was found to have a larger volume (Briess et al., 1998). Consequently it is preferable to sample bilaterally. If that is impossible, it should at least be mentioned on which hemisphere the measurements were performed. Extracellular volume In patients with an antemortem hypovolemic status, increased vasopressin mRNA was found in supraoptic

2014 Ch 01

20/11/03

10:03 am

22 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

Page 22

D.F. SWAAB

Fig. 1.11. Levels of 5-hydroxytryptamine (5-HT) (O) and 5-hydroxindoleacetic acid () in human hypothalamus determined postmortem, in relation to clock time of death. Shown are the means ± S.E.M. (n) of pooled values of seven 3-h intervals. Statistics: Student’s t-test. (with permission from Carlsson et al., 1989a; Fig. 8.)

nucleus neurons (SON; Rivkees et al., 1997). In addition, increased tyrosine hydroxylase (another measure of activation of neurosecretory neurons) was observed in the SON and paraventricular nucleus (M. Panayotacopoulou et al., 2002). (b) Factors during dying Prolonged illness, gravity of illness, and agonal state Lower levels of pH were found throughout the brain in cases of death following a protracted illness such as respiratory distress, as compared to sudden death cases (Spokes et al., 1979). The changes in pH or circulating illness-associated factors like glucocorticoids may explain why, e.g. the tissue levels of neuropeptide-Y (NPY)-like immunoreactivity were found to be elevated by distress from chronic respiratory failure in the infundibular, ventromedial and paraventricular hypothalamic nucleus (Corder et al., 1990). By means of quantitative immunocytochemistry, we not only confirmed the increased amount of NPY peptide

in the infundibular nucleus with protracted illness, but also found that the amount of NPY mRNA went up. In addition, we observed that agouti-related peptide was increased during the course of illness (Fig. 1.12; Goldstone et al., 2002). Also the number of neurons staining for growth hormone-releasing hormone in the infundibular nucleus increases with the duration of premorbid illness (Goldstone et al., 2002). The gravity of the disease, too, is of great importance. During serious illness, profound changes may occur in the thyroid hormone metabolism known as non-thyroidal illness. Subjects who died a sudden death have higher TRH mRNA hybridization expression than subjects with non-thyroidal illness (Fliers et al., 1997; Alkemade, 2003, submitted). In addition, NPY immunoreactivity decreases in this condition (Fliers et al., 2001; Chapter 8.6c). Whenever possible, subjects should thus be matched for premorbid state. This is a criterion that is particularly hard to satisfy in studies of aging, since most young subjects die acutely, e.g. as a result of accidents, suicide or drug overdose, whereas older donors generally die as a result of various chronic disease states. A similar problem exists in studies

2014 Ch 01

20/11/03

10:03 am

Page 23

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

of Alzheimer’s disease patients, who frequently suffer from pneumonia and cachexia. The agonal effects associated with prolonged illness may influence the pH, and subsequently a number of chemical substances in the brain. Subjects who died

23

after a long terminal illness have a lower pH in the brain, CSF and blood. The acidosis corresponds with the increased lactic acid concentrations (Perry et al., 1982; Hardy et al., 1985). Postmortem determination of the rate of albumen transfer from plasma to CSF seems to be an indication for the duration of the agonal process, i.e. the duration of the irreversible decompensation of vital constants up to the moment of death (Mangin et al., 1983). Various enzymatic activities were found to be related to pH and lactate in postmortem brain in Alzheimer’s disease and Down’s syndrome, as well as other dementias (Yates et al., 1990). These authors found that lactate levels were higher and phosphate-activated glutaminase and glutamic acid decarboxylase (GAD) levels were lower in the hypothalamus of agonal controls than in the sudden death controls. Phosphate-activated glutaminase and GAD activities were correlated with tissue pH and lactate and were also reduced by in vitro acidification, suggesting that the low enzyme activities in agonal controls were directly due to the decreased pH. Strong positive correlations were obtained between the concentration of tryptophan, another putative agonal status marker of postmortem brain tissue, and the concentration of gammaaminobutyric acid (GABA) in all brain areas (Korpi et al., 1987a). Using an index of chemical premortem severity based upon semiquantitative estimation of terminal anoxia and hypovolemia, Montfort et al. (1985) observed a strongly positive correlation with the GABA synthesizing enzyme glutamate decarboxylase in the hypothalamic subthalamic nucleus (see Chapter 15) of Parkinson patients. Postmortem brain pH has also been reported to be a fair indication of mRNA preservation. Tissue with low pH, assumed to result from prolonged terminal hypoxia,

Fig. 1.12. Hypothalamic NPY and agouti-related peptide (AGRP) increase with duration of premorbid illness in control and obese Prader–Willi syndrome (PWS) subjects. Relationship between (a) NPY immunocytochemical (ICC) staining volumes, (b) NPY mRNA expression by in situ hybridization, and (c) AGRP ICC staining volumes in the infundibular/median eminence region (Chapter 11), and duration or premorbid illness in control adults (, slid regression line), and obese PWS adults (+) (Chapter 23.1). Note that the y-axes have log10 scales. Note also that NPY peptide and mRNA and AGRP peptide increase with illness duration in both control and obese PWS subjects. Note that correcting for illness duration, NPY ICC staining and mRNA expression, but not AGRP ICC staining, appear lower in obese PWS subjects, compared with controls. r represents Pearson correlation coefficient for controls only. (From Goldstone et al., 2002; Fig. 4.)

2014 Ch 01

24 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 24

D.F. SWAAB

contained reduced or absent mRNA (Kingsbury et al., 1995). However, the total amount of vasopressin mRNA in the paraventricular nucleus correlated negatively with CSF pH (Lucassen et al., 1997), indicating that, if these data can be confirmed, agonal state may influence the yield of different mRNAs in different brain areas in a different way. In view of the data mentioned earlier, measuring the pH in CSF makes it possible to match the samples for agonal state; it has therefore been introduced as a routine procedure in the NBB. In order to check whether the pH was affected by postmortem time, we investigated whether there is any correlation between pH in brain tissue and the postmortem interval in rat and human. The rat data made it clear that there was no significant change in the pH within a postmortem interval of 24 h (Ravid et al., 1992). This is well beyond the range of most of the postmortem intervals of autopsies in our brain bank. Comparable observations have been made on human autopsy material collected by the NBB. The pH values measured in CSF obtained by autopsy of nondemented controls and Alzheimer’s disease patients did not change significantly with postmortem delay (Fig. 1.13; Ravid et al., 1992). Consequently, measuring the pH of either the brain tissue or CSF obtained at autopsy is influenced by agonal state and not by postmortem delay, and can be used for quality control in brain banking procedures. Stress of dying In postmortem CSF, cortisol levels are some 20 times higher than in lumbar puncture CSF, probably reflecting a reaction of the hypothalamopituitary-adrenal system to the stress of dying (Swaab et al., 1994c; Erkut et al., 2003). Indeed, plasma cortisol levels rise to very high levels with impending death. Several studies have shown that severely ill patients have high corticosteroid levels in urine and plasma that are even in the range of those observed in Cushing patients, and there is an excellent correlation between plasma and CSF cortisol levels (Erkut et al., 2003, in press; Fig. 1.14). Moreover, the metabolism of corticosteroids may be decreased in moribund patients (Sandberg et al., 1956; Lamberts et al., 1997a; Klooker et al., unpublished results). As we observed even higher levels of CSF cortisol in severely demented Alzheimer patients than in mildly demented Alzheimer patients or controls, and since the administration of morphine during the agonal state did not increase the high cortisol CSF levels, the high postmortem CSF cortisol

levels seem to be due to the “physical” stress of dying that stimulates the hypothalamopituitary-adrenal axis rather than to the “psychological stress” (Erkut et al., 2003; Chapter 33a). (c) Postmortem factors It is well-known that the most exquisitely selective staining methods, like the procedures of Ehrlich and of Golgi yield good results only when they are applied to pieces of nervous tissue which are absolutely fresh, almost alive. And according to the requirements of the law, which consecrates outgrown and unfounded fears, the human cadaver does not come under the jurisdiction of the anatomist until twenty-four hours after death, when the extremely delicate and susceptible neurons and neuroglia cells have undergone serious alterations and have therefore lost their precious affinity for the reagents referred to (methylene blue and silver chromate). But in those days I was not greatly terrified by the obstacles. Determined to overcome them, I sought material for my studies in the Foundling Home and in the Maternity Hospital, domains in which, for obvious reasons, the tyranny of the law and the concern of the families are not very active. Thanks to the good offices of the staffs of these charitable institutions, and especially to the vigorous cooperation of Dr. Figueroa (an eminent physician, too soon lost to science) as well as the kindness with which I was favoured by the most worthy Sisters of Charity (who carried their amiability so far as to become autopsy assistants), my investigations went ahead as if on wheels. I am able to state that during a study of two years I had unrestricted disposal of hundreds of foetuses and children of various ages, which I dissected two or three hours after death and even while still warm. Recollections of my life Santiago Ramón Y Cajal (1852–1934)

Postmortem delay The time between death and fixation or freezing of the tissue has always received much attention, first from a morphological point of view and later also for neurochemical reasons. There is a significant negative correlation between postmortem time and hypothalamic noradrenaline levels, while a positive correlation was found for the levels of the amino acids tryptophan and tyrosine (Gottfries et al., 1980), indicating breakdown of proteins. On the other hand, many neurochemical substances, which are also present in the hypothalamus, are very stable over the period of a day or so. Neuropeptides and several receptors seem, in general, to be quite stable (Cooper et al., 1981). Even the hypothalamic content of a tripeptide like thyrotropin-releasing hormone (TRH) remains unaltered between 2.5 and 21 hours postmortem (Parker and Porter, 1982).

2014 Ch 01

20/11/03

10:03 am

Page 25

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

25

Fig. 1.13. The pH of brain tissue collected by the Netherlands Brain Bank over a period of 3 years as a function of post-mortem delay. Statistical analysis of the data was performed by applying the two-tailed Pearson correlation analysis. There was no significant correlation between pH and postmortem delay, neither in the control group (A) nor in the Alzheimer’s disease group (B). (From Ravid et al., 1992; Fig. 6.)

Fig. 1.14. A clear correlation is found between post mortem CSF and serum cortisol concentrations (controls n = 9, r = 0.75, p = 0.02 and Alzheimer patients n = 17, r = 0.80, p < 0.001). The CSF cortisol levels represent free cortisol and are approximately 20% of the total serum cortisol level. (Erkut et al., 2003, in press; Fig. 3.)

Immunocytochemical procedures have shown also that other peptides are not very sensitive to postmortem delay; an excellent staining of vasopressin, neurophysin, glycopeptide and oxytocin neurons in the supraoptic nucleus (SON) and paraventricular nucleus (PVN) was obtained on tissue which was fixed only 5–6 days after death (Gabreëls et al., 1998a, b). Bird et al. (1976) found no change in hypothalamic levels of luteinizing-hormone-releasing hormone (LHRH) in tissue maintained at 41°C for 6 days after death. However, a clear decrease was observed for LHRH mRNA in the arcuate nucleus between 4 and 24 hours postmortem (Rance and Uswandi, 1996). The postmortem stability of peptides may depend not only on the compound, but also, strongly, on the brain area involved. Somatostatin staining in the hypothalamus was found not to be affected by postmortem times up to 48 hours (Van de Nes et al., 1994). On the other hand, a rapid postmortem decomposition of somatostatin was reported in the neocortex, mainly within the first six hours after death (Sorensen, 1984). To be able to interpret the dynamics of neuropeptides, it is of great importance to investigate neuropeptide mRNA in the human brain, in addition to the neuropeptide levels themselves (e.g. Goldstone et al., 2002), since neuropeptide levels in a neuron are determined

2014 Ch 01

26 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 26

D.F. SWAAB

not only by the rate of synthesis but also by the rate of transport, release and biodegradation. Peptide and mRNA levels may, therefore, even change in opposite directions (Zhou et al., 2001). In order to test the suitability of formalin-fixed, paraffin-embedded brain tissue for vasopressin (AVP)-mRNA detection, we used symmetrical halves of 5 human hypothalami. In every case one half was formalin-fixed for 16–35 days and paraffin embedded, while the other half was frozen immediately. Following in situ hybridization on systematically obtained sections through the supraoptic and paraventricular nucleus of both halves, total amounts of AVP-mRNA were estimated, using densitometry on film autoradiographs. A nonsignificant trend with postmortem delay was found in cryostat sections but not in paraffin sections. In cryostat sections vasopressin mRNA seemed to decrease, in particular in the first 6 hours postmortem. The recovery of AVP-mRNA in formalin-fixed paraffin sections was good, i.e. about 75% as compared to cryostat sections (Lucassen et al., 1995). We studied, among other things, the effect of different post mortem times on sex hormone receptors in the rat hypothalamus, leaving the brain in the skull at room temperature for 0, 6 or 24 h after death. Following a long fixation for 20 days, the hypothalami were embedded in paraffin and sections were immunohistochemically stained for androgen receptor (AR), estrogen receptor alpha (ER), or progesterone receptor (PR). Retrieving the antigenic sites by microwave pretreatment was essential in order to achieve successful immunohistochemistry in all the groups studied. In general, immunoreactivity was restricted to the cell nuclei. The intensity of the staining appeared to be strongly dependent on the different receptor antigens and post mortem time. Both AR and ER immunoreactivity, but not PR immunoreactivity, were decreased after immersion fixation compared to the perfused sections at time point zero. In brains fixed by immersion, all three receptors decreased gradually with increasing postmortem time, and ER became hardly detectable after 24 h postmortem (Fodor et al., 2002). Most pharmacologically determined binding sites also appear to be quite stable in postmortem tissue (Hardy et al., 1983). Binding studies of imipramine and desmethylimipramine (DMI) have been proven to be stable in the post-mortem human brain, with the highest density found in the hypothalamus (Langer et al., 1981; Cortes et al., 1988; Gross-Isseroff and Biegon, 1988; Gross-Isseroff et al., 1988). Postmortem delay does not influence the distribution of high-affinity somatostatin receptors in the

tuberal nuclei of the hypothalamus obtained at routine autopsies (Reubi et al., 1986). Cooling cadavers In order to limit autolysis, cadavers are generally placed in a mortuary refrigerator. Early refrigeration was found to lead to less drastic reductions in receptor binding (Whitehouse et al., 1984). However, the cooling curve of the brain shows a very slow decrease in temperature. The cooling rate of superficial brain structures is initially faster than that of the core structures (Fig. 1.15). After only one day, temperatures of 4°C are reached (Spokes and Koch, 1978). This makes the cooling of cadavers of little use for our rapid autopsy procedure. Freezing procedures, fixation and storage time In principle, two strategies may be followed for tissue preparation in general and fixation in particular. In the first place, one may look for the optimum tissue preparation procedure for a particular compound or structure. This strategy is generally followed in animal experiments. In the second place, one may use conventionally fixed paraffin-embedded tissue and subsequently adapt the techniques such as immunocytochemistry or in situ hybridization to this tissue. We generally followed the latter strategy for our hypothalamic research because it allows the study of rare cases that have been collected by conventional fixation elsewhere, and because it is impossible to collect a new large series of brains in a different fixative for each new compound or nucleus to be studied. For many compounds it appeared to be possible to perform good, and even quantitative, immunocytochemistry on this type of tissue, as is shown by many examples in the various chapters of this book. The sensitivity of the immunocytochemical staining of formalin-fixed paraffin-embedded sections may be greatly enhanced by microwave pretreatment, especially following long fixation times (Lucassen et al., 1993) or pressure cooking (Norton et al., 1994) of the sections for antigen retrieval. Moreover, quantitative in situ hybridization can be performed on formalin-fixed, paraffin-embedded hypothalamic sections (Guldenaar et al., 1995, 1996; Lucassen et al., 1995; Fliers et al., 1997; Liu et al., 2000; Goldstone et al., 2002). When long fixation causes the in situ signal to decline, microwave pretreatment may again be of great help (U. Unmehopa, unpublished observations). Of course, sometimes it is not possible to use conventional material and one is forced to collect new tissues in a particular fixative, as was the case for

2014 Ch 01

20/11/03

10:03 am

Page 27

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

Fig. 1.15.

27

Cooling curve of the brain when the cadaver is placed in a refrigerator. Note that only after one day temperatures of 4° C are reached. (Spokes and Koch, 1998, Fig. 1.)

the immunocytochemical localization of TRH (Chapter 8.6). Alternatively, one is forced to use snap-frozen tissue. For such reasons it is essential to keep the dissection protocol for the autopsies of a brain bank flexible, so that the necessary alterations can be made quickly. Various groups have used perfusion fixation of the human head (Kalimo et al., 1974; McKenzie et al., 1991) or brain (Beach et al., 1987a) in order to obtain homogenous fixation of high quality for immunocytochemistry or electron microscopy. Perfusion of the head is performed by one or two internal carotid arteries clamping the other carotic arteries and both vertebral arteries. Both jugular veins are opened to let the 0.9% NaCl rinse pass first, later followed by the fixative drip perfusion. Another possibility is perfusion fixation of the isolated brain for which the circle of Willis should be preserved, as well as the vertebral–basilar system of arteries. The brain is suspended in phosphate-buffered saline. Solutions are introduced posteriorly into the vertebral arteries or basilar artery and anteriorly into the internal carotid arteries. The multiple (3 or 4) cannulae are fed using Y-connectors. The brain is perfused at a rate of 50–100 ml/min, first with a salt solution and then with fixative. Our brainbank dissection protocol does not allow these procedures, since different brain samples from the same patient have to be treated in a different way, i.e. frozen, fixed in different fixatives, or used fresh for tissue culture.

Fixation in formalin causes an increase in brain weight and the subsequent washing in water introduces a systematic error in brain weight, e.g. large brains gain more weight than small brains. However, the increase in brain weight during fixation is not age-dependent or sexually dimorphic (Skullerud, 1985). For the immunocytochemical localization of many compounds, the time of fixation is not crucial, certainly not when microwave pretreatment is used (Lucassen et al., 1993). In addition, proteases, formic acid, or ultrasound have been used to counter the antigenmasking effects (Shiruba et al., 1998). Conventional formaldehyde fixation for more than 600 days still results in excellent vasopressin-, neurophysin, glycopeptide, and oxytocin-staining of the hypothalamic SON and PVN neurons following microwave pretreatment (Gabreëls et al., 1998b). Some vasopressin immunoreactivity of hypothalamic neurons was still present in material that had been fixed and stored for more than 50 years, even without microwave pretreatment (Swaab, 1982). However, formalin-fixed paraffin sections are certainly not the optimum choice for studying fibers of such peptidergic neurons, for which thick cryostat or vibratome glutaraldehyde–paraformaldehyde sections are preferable. Since immersion in this fixative does not fully penetrate the intact human brain, smaller tissue blocks have to be fixed by immersion in, e.g. 2.5% glutaraldehyde and 1% paraformaldehyde for 1 week to 1 month. Subsequently the

2014 Ch 01

28 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 28

D.F. SWAAB

blocks can be frozen and stored in sealed plastic at –80°C and cryostat sections can be made for immunocytochemistry. When the sections are mounted on uncoated glass slides, they can be wiped off the glass into a buffer to be used for free-floating staining. This procedure gave good staining of VP fibers in the human brain (Fliers et al., 1986; Van Zwieten et al., 1994, 1996). Dai et al. (1997) used 4% paraformaldehyde fixation of the human hypothalamus for fiber staining of SCN effects. To facilitate the penetration of the fixative, an additional microwave treatment at low setting (80 W) was applied for 15 min. The tissue was then kept in fixative for 8 to 10 days, 25–35 m cryostat sections were cut and the distribution of the vasopressin and vasoactive intestinal polypeptide fibers of the SCN became very clear (see Chapter 4). Vibratome sections of human brain can also be stained in solution following microwave treatment, in order to unmask antigens (Shiruba et al., 1998). Storage time of paraffin sections may influence the amount of mRNA. Negative correlations were found between the storage time of paraffin sections and the total amount of corticotropin-releasing hormone mRNA in the paraventricular nucleus (Raadsheer et al., 1995) and NPY mRNA in the infundibular nucleus (Goldstone et al., 2000). On the other hand, a positive correlation was observed between storage time and AVP-mRNA in the supraoptic and paraventricular nucleus of Alzheimer patients (Lucassen et al., 1997), so that the effect of storage may depend on the species whose mRNA is studied. Human brain tissue used for biochemical studies is usually rapidly frozen and slowly thawed. However, to isolate synaptosomes which are morphologically well preserved and have retained their metabolic performance, one should use the opposite procedure, as snap-freezing generally yields metabolically and functionally inactive preparations (Hardy et al., 1983). The use of archival brain tissue in molecular genetic research One increasingly important function of brain banks is the preservation, collection and typing of DNA and RNA from clinically characterized and neuropathologically verified archival cases. In addition to major genetic defects which are responsible for familial diseases such as presenile Alzheimer’s disease, Huntington’s chorea, Parkinson’s disease, frontal lobe dementias, genetic contribution to brain diseases as risk factors, such as

ApoE-4 in Alzheimer’s disease (Chapters 2.3, 29.1) have become apparent. Brain banks can now prepare not only nucleic acid samples from freshly frozen tissues, but also from formalin-fixed and paraffin-embedded tissues. Although formalin fixation has a deleterious effect on the amount of DNA which can be extracted from archival human brain tissue, even tissue that has been in storage for more than 25 years in unbuffered formalin yields sufficient DNA for qualitative analysis. But the success rate of genotyping increased when buffered formalin was used instead of unbuffered formalin. Inhibitors of the polymerase chain reaction (PCR) process that are extracted from archival paraffin-embedded neuropathological tissues can be diluted out (Kösel and Graeber, 1994; Graeber et al., 1995). A semi-nested PCR method, suitable for providing specific high-yield PCR products from DNA that was extracted from formaldehyde-fixed specimens, which initially generate low-quality templates, has been reported for ApoE-genotyping (Ballering et al., 1997; Ghebremedhin et al., 1998). Moreover, a nested PCR was found to be effective for the determination of ApoE genotype in tissues that had been stored for 12 years in formalin (Gioia et al., 1998). Thus, both formalin-fixed and stored tissue and tissue from paraffinembedded blocks may be used for molecular-genetic gene typing of the patients studied. We should, of course, realize that genetic testing of brain material is accompanied by ethical issues for the families concerned and for society in general (Cassel, 1998). Postmortem neurons in culture conditions Survival of human brain neurons, also from the hypothalamus, with a postmortem time of up to 8 hours after death, is possible to such an extent that they still show axonal transport after 6–18 hours of incubation. For this type of study, brain samples were preincubated for 2–3 hours at –4°C. Neuronal tracers, i.e. neurobiotin or biotinylated dextran, were injected, e.g. in the optic nerve or suprachiasmatic nucleus, incubated in artificial cerebrospinal fluid at room temperature and provided with 95% O2 + 5% CO2. After 6–18 hours of incubation the tracer was transported along axons over a distance of 0.5 to 1.5 cm. Axonal transport appeared to be dependent on oxygen and glucose and thus to be an active energy-dependent process (Dai et al., 1998c). Recent developments make it possible to keep neurons obtained from postmortem tissue alive in tissue culture conditions for more than a month (Verwer et al., 2002; Chapter 33b).

2014 Ch 01

20/11/03

10:03 am

Page 29

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

1.5. Parameters of neuronal metabolic activity in postmortem tissue There are a number of measures in postmortem material that give a good impression of changes in the metabolic and synthesizing neuronal activity of specific cell groups when the subject was still alive. Some of the activity measures can be applied to various structural elements of the neuron, independent of its chemical nature, such as cell size, Golgi apparatus (GA) size and nucleolar size (Palkovits and Fisher, 1968; Hoogendijk et al., 1985; Salehi et al., 1996; Ishunina and Swaab, 1999; Ishunina et al., 1999). Nucleolar enlargement, vacuolation and multiplication, indicating neuronal metabolic activation, were observed in the infundibular nucleus of postmenopausal women and in a man with hypogonadotropic hypogonadism (Ule and Walter, 1983; Ule et al., 1983). In contrast, a decreased nucleolar size accompanies the atrophy of nucleus basalis of Meynert (NBM) neurons in Alzheimer’s disease (Tagliavini and Pilleri, 1983; Mann et al., 1984; Chapter 2.4). This change in nucleolar size seems to be quite specific for Alzheimer’s disease, since no change was observed in the NBM of multiinfarct dementia patients (Mann et al., 1986). Cell size changes appeared to be a good measure for determining the effects of estrogens in the median eminence, supraoptic and paraventricular nuclei (Rance, 1992; Ishunina and Swaab, 1999; Ishunina et al., 1999) and confirmed analogous data from measurements based on the GA size (Ishunina et al., 1999). Other parameters for neuronal activity are directed towards specific compounds in the neurons, like the Gomori staining for neurosecretory material, neurotrophin receptors, peptides (Chapters 2–16), or mRNA for specific products, such as CRH, vasopressin, NPY or AGRP (Raadsheer et al., 1995; Zhou et al., 2001; Goldstone et al., 2002). The staining of peptides (Fig. 12.3) and the expression of mRNA can be enhanced by microwave pretreatment of the sections, especially after longer fixation times. Alterations in peptide and mRNA often occur in the same direction. An example is the decreased number of vasopressin-expressing neurons in the suprachiasmatic nucleus (SCN) (Swaab et al., 1985) in Alzheimer patients, and the decreased vasopressin mRNA content in the same area in these patients (Liu et al., 2000). NPY and AGRP showed similar changes in relation to duration of illness and obesity on both the peptide and the mRNA level (Goldstone et al., 2002).

29

However, in the SCN of depressed patients, an increased number of neurons expressing vasopressin was accompanied by a decreased amount of mRNA (Zhou et al., 2001). Probably the disturbance of peptide transport and release was even stronger than the decreased production of vasopressin, as indicated by the mRNA content. Although mRNA is fragmented in formalin fixed paraffin embedded archival sections, it may still be able to serve as a template for in situ hybridization (see Van Deerlin et al., 2002). The GA size was successfully used for experimental studies in the rat and as a reliable indicator for functional changes in humans (Jongkind and Swaab, 1968; Swaab and Jongkind, 1970, 1971; Lucassen et al., 1993, 1994; Salehi et al., 1994, 1995, 1998a). Thiamine diphosphatase, often referred to as thiamine pyrophosphatase, is an enzyme that is located on the GA and appears to be a good marker for neuronal activity (Jongkind and Swaab, 1968; Swaab and Jongkind, 1970, 1971). It is, therefore, of great interest that decreased thiamine diphosphatase activities were found in the frontal and temporal cortex of Alzheimer patients (Raghavendra Rao et al., 1993). The GA has a crucial function in the processing, transport, modification and targeting of cellular proteins. We and others have put on record that the GA is atrophic and, according to some authors, also fragmented in Alzheimer’s disease (Salehi et al., 1994, 1995a, 1995c, 1998; Dal Canto, 1996; Stieber et al., 1996). Fragmentation or dispersion of the Golgi apparatus has also been observed in amyotrophic lateral sclerosis (Gonatas et al., 1998; Stieber et al., 1998). The GA size can be determined either with the help of specific antibodies or by electron microscopy. Neither method is widely used, probably because both are technically more difficult than, e.g. cell or nuclear size measurements. Neurotrophin receptors also give an impression of metabolic changes. In Alzheimer’s disease, both the highaffinity neurotrophin receptors (trkA, B, C) and the low-affinity receptor p75 are diminished in the nucleus basalis of Meynert (Salehi et al., 1996, 2000; Chapter 2.5). In contrast, in the activated supraoptic nucleus of postmenopausal women, p75 is increased and correlates significantly with another measure for neurosecretory activity, i.e. the size of the Golgi apparatus (Ishunina et al., 2000c; Chapter 8d). Another indication of a change in neuronal activity can be obtained from receptor studies. The ratio of the

4

Suprachiasmatic nucleus (SCN)

Sexually dimorphic nucleus of the preoptic area (SDN-POA) = preoptic nucleus = intermediate nucleus = interstitial nucleus of the anterior hypothalamus (INAH)-1, D14?

Gai et al., 1990; Bonnefand et al., 1990 Fliers et al., 1994 Gao and Moore, 1996a,b

Galanin

TRH GAD 65 and GAD 67

Sukhov et al., 1995

Moore, 1992; Swaab et al., 1994b Swaab et al., 1985; Moore, 1992 Moore, 1992

Goa and Moore, 1996a,b

Proenkephalin

Neurotensin

Vasopressin

Glutamic acid decarboxylase (GAD) as marker for aminobuteric acid (GABA) VIP

Meyer et al., 1989, Aldheid et al., 1990, Ikonomovic et al., 2000, Fig. 3.1b

Typical small (5–10 m) granule cells in conventional staining, vasoactive intestinal polypeptide (VIP) fibers

Sexual behavior

Biological rhythms (circadian rhythms, seasonal rhythms and possibly also monthly and circaseptan rhythms), sexual behavior, glucose homeostasis

Unknown

Memory, sleep-wake regulation, arousal, thermoregulation, baroreceptor mechanisms, aggressive behavior, feeding

Functions

Homologous to the rat SDN-POA (as described by Gorski et al., 1978)

VIP is located in the ventral and central region that receives retino-hypothalamic tract input (Dai et al., 1997), vasopressin in the dorsomedial SCN (Moore, 1992) and neurotensin throughout the SCN (Moore, 1992)

Most, if not all SCN neurons may contain GAD (Goa and Moore, 1996a,b)

Remarks

30

5

3

Islands of Calleja (insulae terminalis)

Pearson et al., 1983, McGreen et al., 1984, Chan-Palay, 1988b, Saper and Chelinsky, 1984; Gilmor et al., 1999; Blusztajn and Berse, 2000

Choline acetyltransferase, histochemistry or immunocytochemistry, vesicular acetylcholin transporter p75 neurotrophin receptor

References

10:03 am

(D13)*

2

Nucleus basalis of Meynert (NBM), diagonal band of Broca (DBB) and medial septal nucleus

Chapter

20/11/03

Basal forebrain nuclei:

Nucleus

Chemical markers of hypothalamic nuclei. Markers

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

TABLE 1.2

2014 Ch 01 Page 30

D.F. SWAAB

Vertical band of the diagonal band of Broca: cholinergic markers Division in (sub)nuclei by cytoarchitecture

7.2

8

Septum

Supraoptic nucleus (SON)

(A15)*

Horvath and Palkovits, 1987

Somatostatin neurons, enkephalin fibers, neurophysins

Lateroventral nucleus

Oxytocin

Vasopressin

See Chapter 2

VIP innervation, less dense substantia-P, enkephalin and neuropeptide-Y innervation

Medial nucleus

Dierickx and Vandesande, 1977

Dierickx and Vandesande, 1977; Van der Woude et al., 1995

Lesur et al., 1989; Walter et al., 1991

Lesur et al., 1989; Walter et al., 1991; Zhou et al., 1995c

Lesur et al., 1989; Walter et al., 1991; Zhou et al., 1995c

VIP innervation, somatostatin neurons and fibers

Central nucleus

Lesur et al., 1989 Walter et al., 1991;

Substance-P fibers

7.1

Bed nucleus of the stria terminalis (BST), D14

References

Oxytocin (adult): lactation, labor (fetal): initiation and course of labor

Vasopressin (adult): antidiuretic hormone (fetal): adaptation to stress of birth

Temperature regulation, vasodilatation, memory, sexual behavior

Sexual behavior, gender feeling

Functions

The ratio vasopressin:oxytocin is 80:20 (Fig. 8.3)

The central nucleus of the BSTc is sexually dimorphic. In male-to-female transsexuals it is female in size and somatostatin neuron number (Zhou et al., 1995c; Kruijver et al., 1999).

Remarks

10:03 am

Lateral nucleus

Chapter

20/11/03

Nucleus

Continued. Markers

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

TABLE 1.2

2014 Ch 01 Page 31

INTRODUCTION

31

(A14)*

Catecholamine-containing neurons pigmented by neuromelanin

Somatostatin

Periventricular nucleus (periventricular area)

8.7

Vasopressin

(A15)*

Spencer et al., 1985 Kitahama et al., 1998a,b Zecevic and Verney, 1995

Bouras et al., 1986, 1987 Van de Nes et al., 1994

Dierickx and Vandesande, 1977 Dierickx and Vandesande, 1977

Accessory nuclei

Oxytocin

Fliers et al., 1994 Guldenaar et al., 1996

TRH

Dierickx and Vandesande, 1977; Wierda et al., 1991

Dierickx and Vandesande, 1977; Van der Woude et al., 1995

Somatostatin: regulation of growth hormone

Vasopressin and oxytocin, see SON and PVN for putative functions

TRH: regulation of thyroid functions, autonomic functions such as temperature regulation

CRH: (adult): regulation adrenal, stress response, mood effects, anxiety (fetal): initiation of labor

Oxytocin: see SON and sedation, lowering blood pressure, antistress effects, food satiety, pair bonding, maternal behavior, parental care, sexual arousal, ejaculation and various other autonomic functions

Vasopressin: see SON and CSF production, blood pressure, temperature regulation, aggression, affiliation

Neuromelanin in pigmentation is observed in the periventricular nucleus from the fourth decade onwards (Spencer et al., 1985).

Alz-50 cross reacts with somatostatinergic neurons in healthy non-demented controls in the periventricular area (Van de Nes et al., 1994)

Colocalization of CRH and vasopressin depends on activity stage (Chapter 8.5; Raadsheer et al., 1994b).

The ratio vasopressin:oxytocin is 20:80 rostral and 60:40 in the caudal half (Fig. 8.3).

Remarks

32

Raadsheer et al., 1993, 1994a

Oxytocin

Vasopressin

Functions

10:03 am

CRH

8

8

Paraventricular nucleus (PVN)

References

20/11/03

(A15)*

Chapter

Nucleus

Continued. Markers

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

TABLE 1.2

2014 Ch 01 Page 32

D.F. SWAAB

10

Dorsomedial nucleus (DMN)

12

13

13

14a,b

Lateral tuberal nucleus (NTL)

Tuberomamillary complex

Premamillary nucleus (TMN) (D8)*

Lateral hypothalamus (D11)*

(A12)*

Infundibular nucleus (arcuate nucleus)

Orexin A, B/hypocretin1,2; dynorphin

Melanin-concentrating hormone

Dynorphin, substance-P

Histamine

Anti-somatostatin 1–12 or 1–28

-MSH, galanin, neuropeptide-Y, growth hormone-releasing hormone

The medium-sized nerve cells of the DMN are markedly richer in lipofuscin than those of the VMN

Somatostatin innervation

Pelletier et a., 1987 Abe et al., 1988; Bresson et al., 1989; Mouri et al., 1993; Sukhov et al., 1995 Peyron et al., 200o; Mignot, 2001

Sukhov et al., 1995; Chawla et al., 1997

Panula et al., 1990; Airaksinen et al., 1991a; Trottier et al., 2002

Van de Nes et al., 1994; Timmers et al., 1996

Najimi et al., 1989; Mengod et al., 1992

Désy and Pelletier, 1978; Pelletier et al., 1978; Gai et al., 1990; Goldstone et al., in prep.

Braak and Braak, 1992

Bouras et al., 1986, 1987; Najimi et al., 1989

Regulation of food intake and body weight

Goal-oriented behavior associated with hunger, thirst and reproduction

Arousal, control of vigilance, sleep and wakefulness, cerebral circulation, brain metabolism, locomotor activity, neuroendocrine functions, feeding, drinking, eating, sexual behavior, analgesia, regulation of blood pressure, temperature, influence on circadian rhythms, memory, neuronal plasticity

Feeding behavior and metabolism

Feeding behavior, growth and metabolism, reproduction

Interneurons to PVN, autonomic functions, reproduction, feeding, sexual behavior

Sexual behavior, gonadotropin secretion, feeding, aggression

Functions

Cells are distributed over the posterior dorsolateral hypothalamus and not localized in a distinct nucleus.

It is not clear whether it belongs to tuberomamillary complex

Somatostatin fibers in the VMN also stain with Alz-50 (Van de Nes et al., 1994)

Remarks

10:03 am

11

9

Ventromedial nucleus (VMN; nucleus of Cajal)

References

20/11/03

(D12)*

Chapter

Nucleus

Continued. Markers

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

TABLE 1.2

2014 Ch 01 Page 33

INTRODUCTION

33

16

14c

15.1

15.2

Corpora mamillaria Medial mamillary nucleus Lateral mamillary nucleus

Intermediate hypothalamic area

Subthalamic nucleus

Zona incerta (D10/A13)*

Tyrosine hydroxylase

Somatostatin

Björklund et al., 1975

Mengod et al., 1992

References

Nociceptive and somatosensory perception, locomotion, sociosexual behavior, feeding, drinking, arousal, attention

Motor behavior

Attack area

Episodic memory, reproduction, penile erection

Functions

A13 in the nomenclature of Björklund et al., 1975. These cells do not contain neuromelanin (Spencer et al., 1975).

Remarks

34

* D8-10: cells contain aromatic l-amino acid decarboxylase (AADC) but no tyrosine hydroxylase (TH) (Kitahama et al., 1998a,b). A11-15: cells contain AADC and TH (Kitahama et al., 1998a,b).

Chapter

Nucleus

10:03 am

Continued. Markers

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

TABLE 1.2

2014 Ch 01 Page 34

D.F. SWAAB

2014 Ch 01

20/11/03

10:03 am

Page 35

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

ER/ER-expressing cells in the supraoptic nucleus (SON), in which ER seems to mediate inhibitory signals and ER activating factors for these neurons can be such a measure (Ishunina et al., 2000). Whether this relationship holds only for a specific neuronal population such as the vasopressin neurons in the SON or whether it is a more general phenomenon is a subject for future studies. Every parameter has, of course, its advantages, disadvantages and limitations, and the application of a number of different measures of neuronal activity on the same material is the best way to come to a firm conclusion about the activity changes in a neuronal population. Gene expression profiling techniques, such as microarray technology, promise to revolutionize biology. This will soon make it possible to investigate alterations at the transcript level of the entire human genome. Good-quality postmortem tissue will be the fundamental prerequisite for such studies (Bahn et al., 2001). RNA and protein samples of human postmortem brain tissue have been found suitable for expression profiling by techniques that include RT-PCR, cDNA microassays, western blotting, immunocytochemistry and proteamics (Hynd et al., 2003). Laser-capture microdissection is a requirement in combination with these techniques when the extremely heterogeneous postmortem hypothalamus is studied. 1.6. Fetal hypothalamic development and adult markers (Table 1.2) of the human hypothalamic nuclei Early steps of hypothalamic development involve regulation both of the induction of hypothalamic identity and the migration of hypothalamic precursors. The prechordal plate and not the head endoderm provides the early signals for the establishment of the hypothalamus. A few molecular pathways involved in the specification and patterning of the hypothalamus are now known. The secreted protein Sonic hedgehog (Shh) is proposed to be such a signal, as hypothalamic tissue is absent in mice lacking Shh function, and increased Shh activity leads to ectopic expression of hypothalamic markers. Nodal signals are also required for hypothalamic development, but only for the posterior ventral region. Hedgehog (Hh) signaling inhibits the development of this region and favors the development of the anterior-dorsal hypothalamus (Mathieu et al., 2002). It has been proposed that disruption of the development of hypothalamic nuclei may be due to disruption of genes involved in neurogenesis (Otp), cell migration

35

(Otp, SF-1), cell death (Brn2, Sim1, Arnt2) and differentiation (Nkx2.1), while also genes involved in sexual differentiation of the hypothalamus emerge from experimental studies (Tobet, 2002). Brn-4 knock-out mice had a loss of the SON and paraventricular nucleus (PVN), and mice with mutations in the gene encoding for the nuclear receptor SF-1 lack the ventromedial hypothalamic nucleus (Martin and Camper, 2001). These observations raise the possibility that similar defects exist in human disease and will be revealed in years to come. In a recent, excellent, review, Koutcherov et al. (2002) have described, in detail, the development of the nuclear organization of the human hypothalamus (Fig. 1.16). As Le Gros Clark already stated in 1938, the nuclear organization in the fetal hypothalamus is in many aspects more distinct than that of the adult. Moreover, developmental data on the cytoarchitecture and chemoarchitecture of the human hypothalamus reveal crucial information on the controversial issue of homologies with the hypothalamic nuclei in rat, as is the case for the discussion on the SDN-POA /INAH-1 homology (Chapter 5.1) and NTL (Chapter 12). Weeks 9–10 of gestation (first trimester) Only minimal signs of nuclear differentiation were found in this period, but a clear subdivision into 3 longitudinal zones was found. A well-defined hypothalamic sulcus indicates the dorsal hypothalamic boundary and the lens-shaped subthalamic nucleus the lateral hypothalamic border. The tentatively designated posterior hypothalamus was separated from the lateral hypothalamus (LHA) by fiber bundle (fasciculus of Forel = fibrae hypothalamicopallidares = fasculus lenticularis) and by the mamillothalamic tract. A cell-sparse supraoptic nucleus (SON) was also found. In addition, the dorsomedial hypothalamic nucleus (DMN), ventromedial hypothalamic nucleus (VMN), the medial preoptic area, the medial mamillary body and the infundibular (= arcuate) nucleus could already be distinguished. Weeks 11–14 of gestation (first trimester). In this phase the fornix became visible and the anlage of the PVN could be distinguished. Weeks 15–17 of gestation (second trimester). The nucleus tuberalis lateralis, intermediate nucleus (= sexually dimorphic nucleus of the preoptic area = SDN-POA) differentiated. The SDN-POA was embedded in the lateral surface of the teardrop-shaped medial preoptic area. Arguments were obtained for homology between the human and rat SON-POA. The lateral mamillary body is prominent at 16 weeks of gestation. A supramamillary

2014 Ch 01

36 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 36

D.F. SWAAB

2014 Ch 01

20/11/03

10:03 am

Page 37

INTRODUCTION

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

37

Fig. 1.16. Organization of major cell groups in the developing human hypothalamus shown at landmark stages of fetal differentiation. The hypothalamus is depicted at four rostrocaudal levels (left to right) for each developmental stage. Gray scale represents hypothalamic structures revealed by cytoarchitecture of the neuroepithelial primordia and transient chemoarchitectonic labeling. Color coding indicates advanced stages of cell groups structural differentiation revealed by chemoarchitecture. Please note that these diagrams are not to scale. Abbreviations used: 3V ac Arc AVPV BST DM DMC f hs IsM LH LTu MbL MbM MEE MEI MPA MPO MPOC MPOL MPOM mt, mtt ne OT, opt ox PaD PaM PaP PaPo PeF PH SCh SChD SChC SO SUM Un VMH VMHDM VMHVL VTM

third ventricle anterior commissure arcuate nucleus anteroventral periventricular nucleus bed nucleus of the stria terminalis dorsomedial hypothalamic nucleus dorsomedial hypothalamic nucleus (compact part) fornix hypothalamic sulcus intermediate nucleus (=sexually dimorphic nucleus of the preoptic area) lateral hypothalamic area lateral tuberal hypothalamic nucleus mamillary body, lateral part mamillary body, medial part median eminence, external median eminence, internal medial preoptic area medial preoptic nucleus medial preoptic nucleus, central subnucleus medial preoptic nucleus, lateral subnucleus medial preoptic nucleus, medial subnucleus mamillothalamic tract neuronal epithelium optic tract optic chiasm paraventricular hypothalamic nucleus, dorsal subnucleus paraventricular hypothalamic nucleus, magnocellular subnucleus paraventricular hypothalamic nucleus, parvicellular subnucleus paraventricular hypothalamic nucleus, posterior subnucleus perifornical hypothalamic nucleus posterior hypothalamic area suprachiasmatic nucleus suprachiasmatic nucleus, dorsal part suprachiasmatic nucleus, central part supraoptic nucleus supramamillary nucleus uncinate nucleus ventromedial hypothalamic nucleus ventromedial hypothalamic nucleus, dorsomedial subnucleus ventromedial hypothalamic nucleus, ventrolateral subnucleus ventral tuberomamillary hypothalamic nucleus

(From Koutcherov et al., 2002; Fig. 14.)

2014 Ch 01

38 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

20/11/03

10:03 am

Page 38

D.F. SWAAB

nucleus was also visible. This nucleus cannot be seen in the adult hypothalamus. Weeks 18–23 of gestation (second trimester). By 18 weeks the posterior subnucleus of the PVN resembles the postnatal structure. The perifornical area differentiated. This area remains anchored around the fornix, whereas most LHA cells are positively displayed laterally by the successive waves of neurons of the midline and core zones that develop later. The suprachiasmatic nucleus and the retinohypothalamic tract became visible at 23 weeks of gestation. At 21 weeks of gestation the PVN evinced, for

the first time, distinct subnuclear subdivisions. Moreover, NPY-positive neurons were present in the infundibular neurons at 21 weeks of gestation. Weeks 24–33 of gestation (second trimester). By this stage the fetal hypothalamus has taken on an adult-like appearance. Week 34 to newborn (third trimester). The NTL and tuberomamillary nucleus can be distinguished. The chemical markers of the adult hypothalamic nuclei and some of their most prominent functions are given in Table 1.2.

2014 Ch 02

20/11/03

10:19 am

Page 39


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 3011 1 2 3 4 5 6 7 8 9 4011 1 2 3 4 5 6 7 8 911

CHAPTER 2

Nucleus basalis of Meynert (NBM) and diagonal band of Broca (DBB)

the extended amygdala, presenting itself as a ring of neurons encircling the internal capsule and basal ganglia (Heimer et al., 1997). Age-related memory disturbances and the loss of memory in Alzheimer’s disease have been related to cholinergic dysfunctions and degenerative changes in the nucleus basalis of Meynert complex (see below and Fig. 2.2). Large reductions in cholinergic markers were indeed found in the cerebral cortex of early-stage Alzheimer patients in biopsy material (Bowen et al., 1982). Moreover, the number of choline acetyltransferase and vesicular acetylcholine transporter neurons correlates significantly with the severity of dementia, determined by the mini-mental state examination test (Gilmor et al., 1999). Choline acetyltransferase activity in the medial, frontal and inferior parietal cortex of Alzheimer patients correlates with praxis scores, and medial frontal acetylcholinesterase activity correlates significantly with attention/registration scores. Hippocampal cholinesterase activity correlates significantly only with recent memory scores (Pappas et al., 2000). Cholinergic deficits may also contribute to behavioral disorders in Alzheimer patients (Minger et al., 2000). Moreover, neurotoxic lesions of the cholinergic system in experimental animals induce -performance deficits. The selective destruction of NBM cholinergic cells impairs the ability of the neocortex to attend to and process short, highly salient sensory stimuli (Wenk, 1997b). However, attempts to reduce memory impairments in Alzheimer’s disease clinically, by acetylcholinesterase inhibitors, have so far had only limited success. This may be because the nucleus basalis of Meynert complex is only one of the many brain systems to be affected with advanced age and in Alzheimer’s disease, and because the complex functions of a neuronal system can only partly be replaced by its neurotransmitter

The first exact description of this telencephalic nucleus, i.e. Meynert’s ganglion basale, was given in a study by A. von Kölliker (1896), who named it in honor of its discoverer (Meynert, 1872; Fig. 2A; see Wenk, 1997b). The nucleus basalis of Meynert (NBM) complex or, better, the cholinergic basal forebrain nuclei consists of the three magnocellular nuclei in the basal forebrain, i.e. the medial septal nucleus (Ch1; Chapter 7.3), the nucleus of the diagonal band of Broca (DBB) (vertical (Ch2) and horizontal (Ch3) limb) and the NBM (Ch4), which are tightly connected to each other (Mesulam et al., 1983; Ulfig et al., 1989; Fig. 2.1). The NBM is part of the substantia innominata, which also contains 39

2014 Ch 02

20/11/03

10:19 am

40 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Page 40

D.F. SWAAB

Fig. 2A.

“Am Starnberger See”, from left to right: Alzheimer, Kraepelin, Gaupp and Nissl (1906). (Source: Dr. R.J. Verhey, 1994).

(Swaab and Fliers, 1986). Moreover, in individuals with mild cognitive impairment, a condition that is often a first stage of Alzheimer’s disease, choline acetyltransferase was unchanged in the inferior parietal, superior temporal and anterior cingulate cortex, while in the superior frontal cortex this enzyme was even elevated above normal control levels. Hippocampal choline acetyltransferase was also significantly higher in these subjects. This study showed that, in contrast to the cholinergic hypothesis of Alzheimer’s disease, cognitive deficits in mild cognitive impairment and early Alzheimer’s disease are not associated with the loss of choline acetyltransferase, and the earliest cognitive deficits in Alzheimer’s disease thus involve brain changes other than cholinergic system loss (DeKosky et al., 2002). However, this system received considerable research attention because of the cholinergic hypothesis of memory dysfunction proposed a few

decades ago (Whitehouse et al., 1981; Bartus et al., 1982; Coyle et al., 1983; Collerton, 1986; Perry, 1986). In addition to their possible function in memory, the cholinergic basal forebrain nuclei are involved in a number of autonomic functions and behaviors. They are considered to be important sites of sleep–wake regulation, to mediate arousal-related functions in the cortex and, in this way, to induce EEG patterns characteristic of waking and REM sleep. Some neurons in these areas display elevated discharge rates during non-REM sleep, while electrical stimulation in the basal nuclei evokes sleep and experimental lesions cause insomnia. The basal forebrain nuclei exert their sleep-promoting functions in part via descending inhibition of caudal hypothalamic and brainstem activating systems. Cholinomimetic compounds induce REM sleep in humans, an effect that is used in the cholinergic REM induction test.

2014 Ch 02

20/11/03

10:19 am

Page 41

NUCLEUS BASALIS OF MEYNERT AND DIAGONAL BAND OF BROCA

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

Fig. 2.1. a–h: main landmarks and nuclei of the hypothalamus and adjacent structures from rostral to caudal.

41

2014 Ch 02

42 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:19 am

Page 42

D.F. SWAAB

Fig. 2.1. Continued.

2014 Ch 02

20/11/03

10:19 am

Page 43


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91


43

2014 Ch 02

44 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:19 am

Page 44

D.F. SWAAB


2014 Ch 02

20/11/03

10:19 am

Page 45


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

Fig. 2.1.

Abbreviations used:

acc al amy ap bnst ca cdt ce cgl Ch1 Ch2 Ch3 Ch4am Ch4al Ch4p ci cl cm o fx gpe gpi hip hy

nucleus accumbens ansa lenticularis amygdala ansa peduncularis bed nucleus of the stria terminalis commissura anterior caudate nucleus capsula externa corpus geniculatum laterale medial septal nucleus vertical limb nucleus of the diagonal band of Broca horizontal limb nucleus of the diagonal band of Broca anteromedial part of the nucleus basalis of Meynert anterolateral part of the nucleus basalis of Meynert posterior part of the nucleus basalis of Meynert capsula interna claustrum corpus mamillare chiasma opticum fornix globus pallidus, external segment globus pallidus, internal segment hippocampus hypothalamus

ins isC me lmi npv ns nsm nso ovlt pc put ro rub sa sn sol som spe tha tol topt vl vp vt

45

insular cortex islands of Calleja lamina medullaris externa (nuclei lentiformis) lamina medullaris interna (nuclei lentiformis) nucleus paraventricularis nucleus subthalamicus nucleus septi medialis nucleus supraopticus organum vasculosum of the lamina terminalis pedunculus cerebri putamen recessus opticus nucleus ruber subcallosal area substantia nigra stria olfactoria lateralis stria olfactoria medialis septum pellucidum thalamus tuberculum olfactorium tractus opticus ventriculus lateralis ventral pallidum ventriculus tertius

(From Vogels, 1990; Fig. 2.)

Afferents to the basal forebrain nuclei from hypothalamic and brainstem nuclei are also functionally important for sleep–wake regulation. Thermosensitive inputs from the anterior hypothalamus modulate the sleep- and arousal-related cells (Szymusiak, 1995; Baghdoyan, 1997). Thermosensitive neurons were found in experimental animals in a number of areas, including the DBB (Alan et al., 1996, 1997). Cholinergic neurons of the DBB also participate in the baroreceptor-mediated inhibition of phasic vasopressin neurons in the supraoptic nucleus (Grindstaff et al., 2000). In monkeys and cats, afferent connections to the substantia innominata/NBM complex were traced from the amygdala, hypothalamus, midline thalamus, zona incerta and fields of Forel (Irle et al., 1986). Since attack behavior can be promoted by injecting acetylcholine into the hypothalamus, and since cholinergic blockers prevent a biting attack, even in naturally aggressive cats or rats (Bear, 1991; Chapter 26.9), the cholinergic basal forebrain nuclei are thought to be also involved in aggressive behavior. Investigations in nonhuman primates have revealed feeding cells in the NBM that respond to the sight and/or taste of food if the organism is hungry. Stimulation of this region

can mimic the reward value of food (Rolls, 1984). There is a well-defined cholinergic pathway from the basal forebrain to the frontoparietal cortical microvasculature that is capable of increasing regional cerebral bloodflow in the cortical areas. 2.1. Anatomy The cholinergic basal forebrain nuclei are the major source of cholinergic innervation to the cerebral cortex, hippocampus, hypothalamus, amygdala, and olfactory bulb. The neurons of the septum and vertical limb of the DBB project mainly to the hippocampus, hypothalamus and cingulate gyrus, whereas those of the horizontal band of the DBB and the NBM project to the amygdala, hypothalamus and cerebral cortex, respectively (Parent et al., 1981; Whitehouse et al., 1981; Ribak and Kramer, 1982; Hedreen et al., 1984; Mesulam et al., 1984; German et al., 1985; Price, 1990; Heiner, 2000). Using acetylcholinesterase staining on the human fetal brain, Kostovi´c et al. (1986) found the first sign of histochemical differentiation of the basal complex at 9 weeks of gestation. At 10–15 weeks of gestation a strongly cholinesterase-

2014 Ch 02

46 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:19 am

Page 46

D.F. SWAAB

Fig. 2.2. Conventional (thianine) staining of the nucleus basalis of Meynert (NBM) of a control case (a) and a patient with Alzheimer’s disease (b). Note the large neurons in the control (a) and the presence of smaller, atrophied neurons in the Alzheimer patient. Bar indicates 20 m.

2014 Ch 02

20/11/03

10:19 am

Page 47


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

positive bundle of fibers approaches, but does not penetrate, the neocortical anlage. At 15 weeks the first acetylcholinesterase-reactive perikarya appear and the position of the future medial septal, diagonal and basal nuclei can be distinguished. Fibers from the nucleus basalis complex enter the white matter of the neocortex via the external capsule. In the next stage (18–22 weeks), the positive fibers can be followed from the NBM through the external capsule to the transient subplate zone of the neocortex. Fibers from the rostromedial part of the NBM run below the corpus callosum as part of the ventral fornix. In addition, fibers can be traced to subcortical structures such as the amygdala, putamen, and the caudate and mediodorsal thalamic nuclei. The basal nuclei act by nicotinic acetylcholine receptors – a class of ligand-gated channels composed of and subunits. They are found in, for example, the neocortex, hippocampus, entorhinal cortex, thalamus, putamen and cerebellum. The predominant subunit type is different in different brain areas, cortical layers and cell types (for review, see Court et al., 2000). In primates, the cholinergic basal forebrain nuclei can be distinguished into 4 subdivions that have been designated as Ch1–Ch4, and that are also identifiable in Nissl-stained sections in the human basal forebrain (Mesulam et al., 1983; Fig. 2.1). The Ch1 group corresponds to the cholinergic neurons in the medial septal nucleus (Saper and Chelimsky, 1984; Chapter 7.3); the Ch2 group with the vertical limb of the diagonal band of Broca; the Ch3 group corresponds with the horizontal limb nucleus of the diagonal band of Broca; the Ch4 group with the nucleus basalis of Meynert and is thus, at least, part of the poorly defined substantia innominata of Reichert. The Ch4 group contains some 210,000 cholinergic neurons (Gilmor et al., 1999). The rostral Ch4 area at the level where the anterior commissure crosses the basal forebrain can be further subdivided into an anteromedial (Ch4am), an anterolateral (Ch4a1), an intermediate (Ch4i) and a posterior part (Ch4p) (Perry et al., 1982; Mesulam et al., 1983; Hedreen et al., 1984; Saper and Chelimsky, 1984; Ulfig, 1989). In a series of sections, Vogels et al. (1990) presented a comprehensive anatomical overview of the location of the different subdivisions of the cholinergic basal forebrain nuclei (Ch1–Ch4) (Fig. 2.1). The most rostrally located part of the nucleus basalis complex (NBM), Ch1, can be found in the medial part of the septum verum (Chapter 7.3). Its boundaries are (Fig. 2.1a): dorsal, nuclei with small-sized neurons in the septum verum; lateral, nuclei with small-sized neurons in

47

the septum verum; ventral, the DBB. In monkey, the septohippocampal pathway is partly cholinergic and partly GABA-ergic (Gulyás et al., 1991). The subsequent subdivision of the NBM is the nucleus of the diagonal band of Broca (Ch2, 3). This nucleus is delineated by the following structures (Fig. 1.4a): dorsal, the medial septal nucleus; Its rostrodorsal boundary is rather arbitrarily located at the level of the anterior commissure; medial, subcallosal area and the subarachnoidal space; lateral, nucleus accumbens; the ventrocaudal boundary of the nucleus of the DBB has been disputed, as well as the division in subnuclei. Some studies subdivided the diagonal band nucleus into a vertical limb nucleus and a horizontal limb nucleus (Fig. 2.1b). The horizontal limb nucleus (Ch3) is described as a small band beneath the vertical limb nucleus, extending caudolaterally and having its greatest expansion between the preoptic region and the amygdaloid region. Mesulam et al. (1983) designated the group of cholinergic neurons in the vertical limb nucleus (without the ventral part of this vertical limb nucleus) as Ch2, and the cholinergic neurons in the horizontal limb nucleus (defined as a small band without the ventral part of the vertical limb nucleus) as Ch3 (Figs. 2.1c and 2.1d). Ch4am: according to Mesulam et al. (1983), the group of cholinergic neurons in the anteromedial part of the NBM (Ch4) extends caudally to the ansa lenticularis and is bordered by the following structures: caudal: ansa lenticularis and Ch4i region (see below); laterorostral: the olfactory tubercle; laterocaudal: the Ch4ai region (see below); dorsal: ventral pallidum (beneath the anterior commissure); medial: lateral preoptic nucleus (lateral hypothalamus); ventral: basal olfactory area and Ch3 regions. Following the rostrocaudal direction, the rostral part of the Ch4am region borders initially on the olfactory tubercle, but immediately behind the tubercle it extends laterally into the amygdaloid region. This lateral extension of the cholinergic neurons in the Ch4am region is known as the Ch4a1 region, the anterolateral part of the nucleus basalis of Meynert (Mesulam et al., 1983). The Ch4al region borders on the following structures: rostral – the olfactory tubercle and the basal olfactory area; caudal – ansa lenticularis and the Ch4i region (see below); lateral – amygdaloid region; dorsal – ventral pallidum (beneath the anterior commissure); medial – Ch4am group; ventral – basal olfactory area and the Ch3 region. The subdivision of the anterior part of the nucleus basalis of Meynert (Ch4a) into an anteromedial and

2014 Ch 02

20/11/03

10:19 am

48 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Page 48

D.F. SWAAB

anterolateral part is, rather arbitrarily, based upon the presence of a large-sized vessel or a vertically orientated small zone with lower neuron density (Fig. 2.1c). The main characteristic of Ch4a NBM is the high neuron density in the “pars compacta”, located in the centre, and the decreasing density toward the periphery of Ch4a, the “pars diffusa”. The subcommissural part of the substantia innominata (i.e. the ventral pallidum, anterior perforated substance and Ch4a; Figs. 1.4c and 1.4a) extends caudally into the sublenticular part of the substantia innominata (Figs 1.4c and 1.4f), in which a new subdivision of the NBM appears, extending in the caudolateral direction (Ch4i). Again a “pars compacta” and “pars diffusa” can be distinguished (Fig. 2.2e). In the caudal part of the subdivision, the “pars compacta” is subdivided by the presence of the ansa peduncularis into a ventromedial and a dorsolateral subnucleus (Fig. 1.4f). The most caudal part of the NBM extends caudally from behind the ansa peduncularis to the rostral part of the lateral geniculate body. Even more caudally, small cell clusters are present in the external medullary lamina (Fig. 2.1h). Mesulam et al. (1983) regarded this posterior part of the NBM as a separate subdivision and called it the Ch4p group, the group of cholinergic neurons in the posterior part of the NBM (Ch4). 2.2. Chemoarchitecture Three neuronal cell types are distinguished in the NBM and DBB: (i) 74% large multipolar neurons containing loosely packed lipofuscin granules; (ii) 9% large spindleshaped neurons with densely packed lipofuscin granules; and (iii) 18% small nerve cells (Ulfig and Braak, 1989). The cholinergic cells of the basal forebrain nuclei can be visualized by choline acetyltransferase histochemistry or immunocytochemistry (Pearson et al., 1983; McGeer et al., 1984; Saper and Chelimsky, 1984; Chan-Palay, 1988b; Gilmor et al., 1999), by staining of the vesicular acetylcholine transporter (Gilmor et al., 1999; Blusztajn and Berse, 2000), and by the presence of the p75 neurotrophin receptor (Ikonomovic et al., 2000), although the latter marker is also found in other areas, such as the SON (Ishunina et al., 2000c). Newly synthesized acetylcholine is taken up into the secretory vesicles by the vesicular acetylcholine transporter that acts as a specific carrier protein (Blusztajn and Berse, 2000). Molecular biological studies have shown that the vesicular acetylcholine transporter gene is located within the

first intron of the choline acetyltransferase gene, and the gene structure suggests that the expression of the two proteins is regulated in a coordinated fashion, mediated by cis-acting regulatory elements within the locus. For the putative regulatory elements of the rodent cholinergic gene locus, see Blusztajn and Berse (2000). In individuals under the age of 65, 72% of the choline acetyltransferase positive neurons contain the calcium binding protein calbin-D28k, a proportion that is decreasing to only 28% over the age of 65. Approximately 1.5% of the total population of magnocellular NBM and DBB neurons express tyrosine hydroxylase, a catecholamine-synthesizing enzyme and calbindin (Gouras et al., 1992; Sanghera et al., 1995). The septohippocampal pathway in monkey is partly cholinergic and partly GABA-ergic (Gulyás et al., 1991), and in rat at least half of the basal forebrain neurons, which project to the cortex, are GABA-ergic. These neurons would mediate “executive” aspects of performance (Sarter and Bruno, 2002). Moreover, quite a number of peptides is present in the neurons of the basal nuclei, colocalizing with acetylcholine, i.e. preproenkephalin (Sukhov et al., 1995), neurokinin B (Chawla et al., 1997) and LHRH (Stopa et al., 1991; Rance et al., 1994; Dudas et al., 2000; Dudas and Merchenthaler, 2002). In the DBB, LHRH cells are found that are often colocalized with delta sleep-inducing peptide (Vallet et al., 1990). Somatostatin neurons and fibers are observed in the NBM and DBB (Bennett-Clarke and Joseph, 1986). Vasopressin neurons are present in the NBM (Ulfig et al., 1990). Galanin was found to be colocalized with acetylcholine in large NBM neurons according to some authors (Chan-Palay (1998a), but not according to others (Mufson et al., 1998). Walker et al. (1991) found that only very few large NBM neurons contained galanin mRNA, while Chan-Palay (1988a) reported that galanin is present in small numbers of noncholinergic interneurons in this nucleus. We found that the DBB stained only very lightly for galanin as compared to, e.g. the sexually dimorphic nucleus (Chapter 5) (unpubl. observ.). Galanin is an inhibitory peptide that modulates cognition by regulating the cholinergic basal forebrain neurons. In humans, the basal forebrain contains a dense galaninergic fiber plexus of unknown origin. There is a high- and low-affinity galanin receptor within the basal forebrain in humans (Mufson et al., 1998). Moreover, high benzodiazepine binding is observed in the NBM and DBB (Najimi et al., 1999). Already in the human fetus and neonate, high benzodiazepine binding

2014 Ch 02

20/11/03

10:19 am

Page 49


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

was found in the DBB (Najimi et al., 2001). In addition, VIP binding sites (Sarrieau, 1994) and oxytocin binding sites are present in the NBM and DBB (Loup et al., 1991). TRH binding sites were more abundant in the infant than in the adult DBB (Najimi et al., 1991). TRH in the DBB may be involved in temperature regulation (Chapter 8.6). NADPM-diaphorase-positive neurons are also found in the NBM (Sangruchi and Kowall, 1991). NADPH diaphorase is a nitric oxide synthase. The NBM receives a dense peptidergic innervation by fibers containing somatostatin, substance-P, cholecystokinin octopeptide, VIP, metenkephalin, ACTH, MSH, and oxytocin (Candy et al., 1985), and hypocretin fibers (Moore et al., 2001), while vasopressin and oxytocin fibers, too, are present in the DBB (Fliers et al., 1986). Some cells of the DBB and of the NBM contain nestin. However, the function of this intermediate filament in mature neurons is not clear (Gu et al., 2002). The DBB is densely innervated by woolly fibers containing secretoneurin, a peptide derived from secretogranin II (Marksteiner et al., 1993). The -amino-3-hydroxy-5-methyl-4-isoaxoleproprionate (AMPA) glutamate receptor subunit GluR-1, is present in 94% of the magnocellular cholinergic elements, while the GluR2/3, which is present in young subjects, is relatively faint or non-existent in Ch1–Ch4 of aged individuals (Ikonomovic et al., 2000). A strong nuclear androgen receptor staining is present in the neurons of the horizontal band of the DBB, followed by a medium intensity in the vertical limb DBB and a weak staining in neurons of the NBM (Fig. 6.1). In males, the staining is more intense than in females in the vertical (Ch2) and horizontal band (Ch3) of the DBB (Fernández-Guasti et al., 2000). The possible effects of androgens on cognition in elderly males and Alzheimer patients are currently under investigation. In addition, NBM neurons contain estrogen receptors, especially in their rostral part (Donahue et al., 2000). This may be of clinical importance, since there is some evidence that estrogens may reduce both the risk and severity of Alzheimer’s disease, possibly by an action on the cholinergic system (Smith et al., 2001; and see Chapter 29.1b on this controversial issue). In addition, estrogen replacement therapy may enhance the response on an acetylcholinesterase inhibitor in women with Alzheimer’s disease (Schneider et al., 1996). Estrogen receptor (ER) was expressed to a higher degree in the NBM neurons than ER, and ER was mainly localized in the cell nucleus, while ER was mainly confined to the cyto-

49

plasm. A significant, positive correlation was found between the proportion of ER nuclear positive neurons and age in men, but not in women, whereas the proportion of ER cytoplasm positive neurons increased during aging in both sexes (Ishunina and Swaab, 2001). ER stained stronger in the cytoplasm of NBM and horizontal DBB neurons in women than in men (Kruijver et al., 2003). 2.3. Alzheimer’s disease In Alzheimer’s disease MRI revealed atrophy of the basal forebrain (Callen et al., 2001). The NBM is indeed severely affected in Alzheimer’s disease (Whitehouse et al., 1981, 1982, 1983b; Nakano and Hirano, 1982; Arendt et al., 1983; Candy et al., 1983; Nagai et al., 1983; Tagliavini and Pilliris, 1983; Mann et al., 1984; Coleman and Flood, 1987; Etienne et al., 1986). Fifty percent of all Alzheimer patients have markedly reduced cortical choline acetyltransferase activity, despite a preserved enzyme activity in the NBM that suggests a deficiency of axonal transport (Etienne et al., 1986). The degeneration of the cholinergic system in Alzheimer’s disease may also account, through a disorder of its cortical microvasculature innervation, for a decrease in cerebral blood flow . In addition, it has been proposed that the decrease in nerve density in the A1 segment of the anterior cerebral artery, which supplies the NBM (Bleys and Cowen, 2001), may be related to the decreased metabolic activity observed in the NBM in Alzheimer’s disease (Salehi et al., 1994). The decrease in choline acetyltransferasepositive NBM neurons in Alzheimer’s disease is dependent on the number of ApoE4 alleles. Also, dendritic length and neuronal density appear to decrease in relation to the number of ApoE4 alleles (Arendt et al., 1997; see also Chapter 2.4), but plastic neuronal remodeling also takes place in the NBM of Alzheimer patients in an inverse relationship to the ApoE4 allele copy number (Arendt et al., 1997). Many investigations have shown that aging is associated with moderate reductions in nicotinic acetylcholine receptor subunit mRNA (especially in subunit 4) and a strong loss of protein expression in Alzheimer’s disease (Court et al., 2000, 2001). But not only is a loss of neuroactive substances found in the NBM in Alzheimer’s disease; neuropeptide-Y concentrations were reported to be increased in the substantia innominata, while there was hypertrophy and

2014 Ch 02

50 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:19 am

Page 50

D.F. SWAAB

hyperinnervation of remaining forebrain neurons by galanin neurons in Alzheimer’s disease (Valenti, 1996). Because galanin is supposed to inhibit the release of acetylcholine in the hippocampus, it has been suggested that the overexpression of galanin in Alzheimer’s disease may further exacerbate the cholinergic cellular dysfunction in this disorder (Mufson et al., 1998). Furthermore, galanin fibers hyperinnervate the remaining cholinergic forebrain neurons in Alzheimer’s disease, and the galanin binding sites increase (Counts et al., 2001). However, animal experimental data have shown that a loss of function mutation in the galanin gene is accompanied by a loss of a third of the cholinergic neurons in the medial septum and in the vertical limb of the DBB and memory deficits. This suggests that, at least in mice, galanin has a trophic function for cholinergic neurons (O’Meara et al., 2000; Counts et al., 2001). In late stage Alzheimer’s disease, but definitely not in the early stages, galanin binding was increased in the anterior part of Ch4. Whether the increased galanin activity in the NBM is indeed detrimental or neuroprotective is thus entirely open to discussion (Mufson et al., 2000a). When using markers for cytoskeletal alterations (e.g. the monoclonal antibody Alz-50 that is directed against hyperphosphorylated tau), the NBM of Alzheimer’s disease patients displays a pronounced staining of the perikarya and dystrophic neurites, in contrast to nondemented controls (Fig. 2.3; Swaab et al., 1992b; Van de Nes et al., 1993; see Chapter 29.1 and Fig. 29.2). A sex difference was found in the early cytoskeletal staining in the NBM of Alzheimer patients. The percentage of Alz-50-positive neurons in the NBM is significantly higher in females than in males (Salehi et al., 2000; Fig. 2.3). This may be related to the higher prevalence of Alzheimer’s disease in women as compared to men (Bachman et al., 1992). In addition, the sex difference in hyperphosphorylated tau in the NBM may be related to the observation that male Alzheimer patients have a 73% greater chance of responding to anti-acetylcholinesterase therapy than female patients (MacGowan et al., 1998). In Alzheimer’s disease (see Chapter 29.1) the proportion of neurons showing nuclear staining for both ER and and cytoplasmic staining is markedly increased, suggesting the presence of a substrate for estrogen therapy. In women, Alzheimer’s disease appears to increase the percentage of ER nuclear positive neurons, while nuclear ER increases in both sexes in the NBB (Ch4) (Ishunina and Swaab, 2001) and the vertical limb of the DBB (Ch2)

(Ishunina and Swaab, 2002). A SPECT study did not reveal an overall significant difference in the regional amount of the vesicular acetylcholine transporter between postmenopausal women treated and postmenopausal women not treated with estrogens. However, the duration of hormone treatment therapy correlated positively with this cholinergic parameter in a number of cortical areas. This suggests that hormone replacement therapy may positively affect the basal cholinergic neurons (Smith et al., 2001). In the NBM of Alzheimer patients, increased expression of -amyloid precursor protein coincides with intracellular neurofibrillary tangle (NFT) formation (Murphy et al., 1992a). Moreover, amyloid- containing senile plaques are found in the NBM (Rudelli et al., 1984), although not in large amounts (Arnold et al., 1991). On the other hand, only few /A4 (A)-staining Congo negative amorphous plaques are found in the NBM (Van de Nes et al., 1998). The evolution of Alzheimer’s disease-related cytoskeletal changes has been described by Sassin et al. (2000). The initial cytoskeletal abnormalities in the NBM are already seen in Braak stage I. Subsequently a neurofibrillary tangle is formed as a spherical somatic inclusion in this brain structure. Finally, the cell may die, leaving behind an extraneuronal “ghost tangle”. On the other hand, A-deposition in the NBM occurs late: in the third phase in the evolution of amyloidosis (Thal et al., 2002). One study showed an increase in the density of intensely staining nitric oxide-synthesizing neurons in the substantia innominata in Alzheimer’s disease, which has been interpreted as a possible source of neurotoxicity for the surrounding cholinergic neurons (Benzing and Mufson, 1995). Proteins produced by “molecular misreading”, such as ubiquitin +1, are found in the NBM of Alzheimer patients (Van Leeuwen et al., 2000). Their functional impact is currently under investigation. Down regulation of synaptophysin, synaptotagmin and a number of protein phosphatases has also been found (Mufson et al., 2002). Cholinergic deficits in Alzheimer’s disease not only correlate with cognitive impairment, but may also contribute to other behavioral disturbances in these patients. A loss of choline acetyltransferase activity in the frontal and temporal cortex correlates with increasing overactivity and aggressive behavior in Alzheimer patients (Minger et al., 2000). Whether the vulnerability of the basal forebrain neurons extends to the noncholinergic GABA-ergic neurons remains unsettled. It is of great interest, however, because of the “executive” aspects of performance in

2014 Ch 02

20/11/03

10:19 am

Page 51


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

Fig. 2.3. Immunocytochemical staining of hyperphosphorylated tau by Alz-50 in the NBM of (A) a male and (B) a female AD patient. Note the increase in the number of Alz-50 stained cells in the female AD patient. Scale bar = 80 m. (From Salehi et al., 1998c; Fig. 2.)

which these neurons are presumed to be involved (Sarter and Bruno, 2002). 2.4. Neuronal loss versus atrophy Estimations of the neuronal numbers of the NBM during normal aging vary greatly, i.e., from losses ranging from 23% to 90% (Mann et al., 1984; McGeer et al., 1984; Etienne et al., 1986; Lowes-Hommel et al., 1989; Cullen et al., 1997) to no neuronal loss at all (Whitehouse et al., 1983a; Chui et al., 1984; Bigl et al., 1987). For a mesaanalysis of the data see Lyness et al., 2003. Massive cell death in the NBM was originally presumed to be one of the major hallmarks of Alzheimer’s disease (Whitehouse et al., 1981, 1982; Arendt et al., 1983; Mann et al., 1984; Etienne et al., 1986; Whitehouse,

51

1986). In addition, a clear loss of the markers of NBM neurons, choline acetyltransferase, was reported (Pearson et al., 1983), which illustrates the importance of distinguishing a loss of cholinergic markers from a loss of neurons. The fact that cells do not show the cholinergic marker anymore does not mean that they died. They might still be there but be inactive and thus no longer express the typical marker. This is true at least for the late phases of Alzheimer’s disease, since in patients with early signs of Alzheimer’s disease, no changes were observed in neocortical areas in choline acetyltransferase or acetylcholinesterase (Davis et al., 1999a). It has been presumed that the large differences in cell loss that were reported, may, at least partly, be due to the heterogeneity of the different subdivisions of the NBM (Iraizoz et al., 1991). Indeed, Vogels et al. (1990) found an overall neuron loss in the NBM of only 10%, while neuron loss varied from 0% in the rostral to 36% in the caudal part of the NBM. However, regional heterogeneity cannot be the only explanation for the enormously varying data reported since cell loss was found to be consistent in all regions of the Ch4 system (Whitehouse, 1986). In addition, studies performed on one particular, well-defined NBM subdivision showed considerable variation. For instance, measurements performed in the Ch4a area showed differences varying from a cell loss of between 42% and 89% (Mann et al., 1984; Cullen et al., 1997) to no significant cell loss at all (Pearson et al., 1983). Gilmor et al. (1999) studied the NBM in patients without cognitive impairment, in patients with mild cognitive impairment, and in patients with early-stage Alzheimer’s disease, using choline acetyltransferase and the vesicular acetylcholine transporter as markers for the NBM neurons. No significant difference was found between the 3 groups and only a 15% non-significant reduction in the number of NBM neurons was found in the early Alzheimer cases, showing that, certainly in the early stage of the disease, these neurons are relatively preserved. The most likely explanation for the equivocal results concerning neuronal loss in the NBM in Alzheimer’s disease is the use of different criteria for the size of counted cells, which is crucial, considering the atrophy NBM neurons appear to undergo in Alzheimer’s disease. Mann et al. (1984), for instance, only counted cells with a diameter larger than 30 m and reported a 54% cell loss in the NBM, whereas Pearson et al. (1983) counted all NBM neurons regardless of their size and did not find any significant cell loss in the NBM. Indeed, while the number of large neurons decreases, the number of small

2014 Ch 02

52 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:19 am

Page 52

D.F. SWAAB

neurons increases in the NBM in Alzheimer’s disease (Whitehouse et al., 1983b; Rinne et al., 1987; Allen et al., 1988a; Vogels et al., 1990; Figs. 2.2 and 2.4), indicating that the large neurons atrophy and lose their cholinergic markers, but do not die. For this reason, the general concept of major cell loss in the NBM of Alzheimer’s disease patients has had to be abandoned and replaced by the opinion that neuronal atrophy rather than cell death is the major hallmark of Alzheimer’s disease in the NBM (Pearson et al., 1983; Rinne et al., 1987; Salehi et al., 1994; Swaab et al., 1994a; Swaab et al., 1998). Since the size of the GA originally was shown to be a sensitive parameter for neuronal activity both in animal experiments (Jongkind and Swaab, 1967; Swaab and Jongkind, 1971; Swaab et al., 1971) and in the human hypothalamus (Lucassen et al., 1993; 1994; Ishunina et al., 1999; Chapter 1.5), GA size was used to monitor activity changes in the NBM in aging and Alzheimer’s disease. The clear decrease in GA size observed in Alzheimer’s disease (49%) (Figs. 2.5 and 2.6) strongly suggests that the capacity of NBM neurons to process and target proteins decreases dramatically in Alzheimer’s disease (Salehi et al., 1994). This conclusion is consistent with studies showing a decreased volume of the nucleolus as an indication of the protein synthetic capacity of NBM neurons in Alzheimer’s disease (Tagliavini and Pilleri, 1983; Mann et al., 1984) and agrees with earlier studies providing evidence for a decrease in the activity of the enzymes choline acetyltransferase and cholinesterase in the NBM in Alzheimer’s disease (Perry et al., 1982; McGeer et al., 1984; Etienne et al., 1986, Perry, 1986; Araujo et al., 1988). There is a strong reduction in neuronal metabolic rate in the NBM of Alzheimer patients, as well as in neuronal metabolic rate in the NBM in Alzheimer patients with either one or two ApoE4 alleles (Figs. 2.7 and 29.3; Salehi et al., 1998a). This finding is in full agreement with the more severe cholinergic deficit in the temporal cortex observed in Alzheimer patients with one or two ApoE4 alleles (Poirrier et al., 1995). There are indications that ApoE genotype in the long term may affect the response to anticholinesterase therapy in Alzheimer patients. ApoE4positive women are the most likely patients to benefit (MacGowan et al., 1998). The contrast between declining neuronal activity and neurodegeneration on the one hand, like in the NBM, and increased neuronal activity and a lack of degeneration on the other, in the supraoptic and paraventricular nucleus (see Chapter 8.3), has been instrumental in the development of the hypothesis that

active neurons are protected against Alzheimer changes. We paraphrased this as “use it or lose it” (Swaab, 1991; Chapter 29.1). A key question in neurobiology of aging is, consequently, how atrophic neurons can be stimulated to regain their activity. The basal cholinergic nuclei are affected by Alzheimer’s disease. Not only the NBM (Ch4) but also the DBB and medial septal nucleus (Ch1–2) show decreased choline acetyltransferase and acetylcholinesterase activity (Henke and Lang, 1983). Although a small loss of neurons is probably present in most subdivisions of the cholinergic basal forebrain nuclei, the Ch1–2 areas seem to be an exception (stability in cell number and cell shrinkage occurs in this area) (Bigl et al., 1987). The Ch1–2 areas project mainly to the hippocampus, which is heavily affected by Alzheimer’s disease, but keep a normal level of neurotrophin receptors (Vogels et al., 1990; Salehi et al., 1998b). Nerve growth factor levels in the hippocampus were even significantly elevated in Alzheimer patients, while BDNF levels were reduced and NT-3 and NT-4/5 levels remained unchanged (Hock et al., 2000), which may be a reason for the stability of the Ch1 and Ch2 neurons. The intact levels of a number of neurotrophins in the hippocampus and the stable cell number in the DBB is support for the idea that neurotrophins may be good candidates for the reactivation of neuronal systems that are affected by Alzheimer’s disease. 2.5. Neurotrophin receptors in the NBM The basal forebrain complex contains both low- affinity nerve growth factor (NGF) receptors (p75, Hefti et al., 1986; Allen et al., 1989b) and high-affinity neurotrophin receptors (Kordower et al., 1989). All 3 family members of the high-affinity neurotrophin receptors, the tyrosine receptor kinases (trks) A, B and C are found in the NBM neurons (Muragaki et al., 1995; Shelton et al., 1995; Salehi et al., 1996). Both trk- and p75-immunoreactive neurons are already found in the NBM as early as embryonic week 14 (Chen et al., 1996). Neurotrophin receptors promote survival, neuronal differentiation, and metabolism by binding neurotrophins at the nerve terminals in the cortex and hippocampus. The trks include trkA (a receptor for NGF), trkB (a receptor for BDNF and NT4-5) and trkC (a receptor for NT3). Upon ligand-binding, trks dimerize and become catalytically active, resulting in autophosphorylation. Internalization of neurotrophins is required for activation

2014 Ch 02

20/11/03

10:19 am

Page 53


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

53

Fig. 2.4. Size-specific numerical densities (Nv) of neuronal nuclei and perikarya in non-demented controls and AD patients in the nucleus basalis of Meynert (NBM). (From Rinne et al., 1987; Fig. 3, with permission.) Note that the number of large neurons decreases, while the number of small neurons increases, which illustrates neuronal shrinkage in the NBM.

of the transcription factor following nerve terminal stimulation, but not following direct cell body stimulation. Ligand-binding stimulates receptors in nerve terminals that are internalized to form signaling endosomes. These vesicles, containing activated trk complexes, are retrogradely transported and transmit a neurotrophic signal. The signaling from the endosomes takes place within axons and in the cell bodies. The p75 receptor internalizes neurotropins and may assist trk receptors in the formation of high affinity neurotrophic binding sites. In contrast to the trks that are involved in retrograde transports of neurotrophins, the lowaffinity receptor p75 probably chaperones both the anterograde and retrograde movement of neurotrophins (Heerssen and Segal, 2002; Butowt and Von Bartheld, 2003). NGF in the NBM decreases during aging and even more so in Alzheimer’s disease (Hefti and Mash, 1989; Mufson et al., 1995), and the serum nerve growth factor concentrations diminish in preclinical Alzheimer patients. Studies from our group show that all three types of trks colocalize in the NBM neurons and decrease in Alzheimer’s disease, although trk-A decreases more than B and B decreases more than C (Salehi et al.,

1996; Figs. 2.8 and 29.5). TrkA mRNA levels decrease markedly in Alzheimer’s disease (Mufson et al., 1996). The reduction in the expression of trkA has subsequently been confirmed, not only on the messenger level, but also on the protein level (Boissiere et al., 1997; Mufson et al., 1997). Moreover, it was shown that a neuronal loss of immunoreactive trkA neurons already occurs in individuals with mild cognitive impairment without dementia, to the same degree as in early Alzheimer’s disease (Mufson et al., 2000b). By gene expression profiling trkB and -C were found to be selectively down regulated, more than trkA (Mufson et al., 2002). The reduction in trk receptors may underlie the diminished NGF levels in the NBM, leading to their decreased metabolism and function. In contrast to the decrease in trk receptors, expression of the gene encoding for the low-affinity p75 receptor was reported not to be significantly altered (Mufson et al., 1996). Also, based on a Northern blot and receptor binding, the expression of p75 in NBM neurons was reported to be unaltered in Alzheimer’s disease. These findings, however, are not without controversy, since Arendt et al. (1997) found an ApoE4-related decrease in the number of p75immunoreactive NBM neurons. We have quantified

2014 Ch 02

54 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:19 am

Page 54

D.F. SWAAB

Fig. 2.5. Immunocytochemical staining of the Golgi apparatus (GA) in the nucleus basalis of Meynert (NBM) in young (A: female, 36 years of age) and old (B: male, 85 years of age) controls and Alzheimer’s disease patients (C: female, 90 years of age; D: male, 87 years of age). Note the clear reduction in size of the GA in the NBM in Alzheimer’s disease patients when compared to the controls. Scale bar = 30 m. (From Salehi et al., 1994; Fig. 2, with permission.)

2014 Ch 02

20/11/03

10:20 am

Page 55


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

55

Fig. 2.6. Frequency distribution of the size of the Golgi apparatus (GA) in controls and Alzheimer (AD) patients. The distribution of the GA area has shifted significantly (= p < 0.001) to lower digits in AD, indicating a strong decrease in neuronal metabolism in the nucleus basalis of Meynert in AD. (From Salehi et al., 1994; Fig. 4, with permission.)

p75 immunoreactivity in the NBM of 30 Alzheimer patients and their matched controls and observed a significant decrease in p75 staining, both in cell bodies and in fibers. The fibers in the NBM contained even less p75 in younger Alzheimer patients (Salehi et al., 2000). It thus seems that both high- and low-affinity neurotrophin receptors are decreased in the NBM of Alzheimer patients. In addition, a defect in retrograde transport of NGF to the NBM of Alzheimer patients has been proposed (Mufson et al., 1995; Scott et al., 1995). This defect may be related either to the decreased amounts of trk receptors in Alzheimer’s disease (Salehi et al., 1996), to the decreased amount of p75 (Salehi et al., 2000), or to the cytoskeletal changes in the NBM (Swaab et al., 1992b) that are generally presumed to hamper axonal transport. Exactly how decreased neuronal metabolic activity (Salehi et al., 1994), cytoskeletal changes (Swaab et al., 1992b), the loss of trk and p75 receptors (Salehi et al., 1996., 2000) and the disturbed retrograde transport of NGF are related to the diminished function of the NBM neurons should be studied further. In a pilot study using a radio-controlled fully implantable pumping device delivering NGF to the lateral ventricle of a 69-year-old female Alzheimer patient who

Fig. 2.7. Immunocytochemical staining of the Golgi apparatus in Alzheimer patients with ApoE genotype 3/3 and ApoE genotype 3/4. Note the smaller Golgi apparatus in the nucleus basalis neurons of Alzheimer patients with ApoE genotype 3/4 (B) vs. the Alzheimer patients with ApoE genotype 3/3 (A). Bar = 30 m. (From Salehi et al., 1998a; Fig 1, with permission.)

had had symptoms of dementia for 8 years, increases in blood flow and nicotine binding in frontal and temporal cortex were noted, as well as a persistent increase in cortical blood flow as evaluated by positron-emission tomography. Her EEG and psychological tests also showed an improvement (Seiger et al., 1993). However, these effects were only limited and short-lasting, as may perhaps be expected from the loss of neurotrophin receptors in the cholinergic system of Alzheimer patients (see above). Moreover, the few Alzheimer cases treated with low doses of nerve growth factor experienced several serious side effects, including pain and weight loss. The pain disappeared within a couple of days after stopping the NGF infusion, and was followed by weight gain (Nordberg, 1996; Eriksdotter

2014 Ch 02

56 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:20 am

Page 56

D.F. SWAAB

for 9 months, cortical glucose metabolism was preserved ipsilaterally, temporally and parietally, while it declined elsewhere in the cortex (Turnbull et al., 1985). Controlled studies, including bilateral NBM stimulation, are probably necessary to obtain more convincing results. 2.6. Other disorders affecting the NBM and DBB

Fig. 2.8. Immunocytochemical staining of trkA in NBM neurons in a control (A) and an AD patient (B). Note the clear reduction in staining of both the large and small neurons in the AD patient. Scale bar = 35 m. (From Salehi et al., Fig 2; 1996, with permission.)

Jönhagen, 1998). A new line of research is now starting in this field. A significant and extensive decline in the number and size of cholinergic NBM neurons was found in aged rhesus monkeys. The loss of staining NBM neurons was nearly completely reversed by human nerve growth factor gene delivery (Smith et al., 1999). We have to wait and see whether a similar gene therapy in Alzheimer patients as currently performed by M.N. Tuszynski (San Diego, USA), will not lead to the same side effects as were reported earlier for NGF infusion, and to better results. We should also be aware of the possibility that neurotrophins may potentiate necrosis (Koh et al., 1995). Reactivation of nerve cells may also be obtained by electrical stimulation. Although no clinical improvements were observed following a unilateral NBM stimulation

The NBM is not only affected in Alzheimer’s disease, but also in various other neurological disorders that involve deterioration of memory and cognitive functions, such as Creutzfeldt–Jakob’s disease (Arendt et al., 1984), dementia with argyrophilic grains (Chapter 29.2; Masliah et al., 1991) and Lewy body disease, patients with hallucinations (Perry et al., 1990; Chapter 29.7). In dementia with Lewy bodies, cortical choline deficits, measured as choline-acetyltransferase. was much greater than in Alzheimer’s disease and occurred earlier, i.e. already during mid-stage disease (Tiraboschi et al., 2002). In this disorder, delusions correlate with frontotemporal cholinergic deficits (Minger et al., 2000). Moreover, the NBM is affected in fatal familial insomnia (Rossi et al., 1998a; Chapters 4b, 30.7) and, according to some authors, Pick’s disease (Uhl et al., 1983; Chapter 29.7f). Tagliavini and Pilleri (1983), who could not find a reduction in cell number in the NBM in Pick’s disease, did find a reduction in cell size, Nissl substance and nucleolar volume, which is in accordance with a decreased neuronal activity in Pick’s disease in this brain area. In progressive supranuclear palsy (PSP) and corticobasal degeneration a substantial loss of large NBM neurons was reported, with globose neurofibrillary degeneration of surviving cells, and in Parkinson dementia complex of Guam, the reduction of large neurons may amount to as much as 90% (Tagliavini et al., 1983; Kasashima and Oda, 2003). In these patients a moderate loss of choline acetyltransferase activity was found in the midfrontal and inferior parietal cortex, and a severe loss in the superior temporal cortex. The deficit was similar to that seen in Alzheimer’s disease and less severe than that observed in Lewy body disease (Masliah et al., 2001). In Wernicke’s encephalopathy, where cell death is found in the NBM (Arendt et al., 1983; Perry, 1986; Fadda and Rosetti, 1998), tau-positive granular and fibrillary inclusions are frequently observed, and increased peroxidase is found in NBM neurons (Cullen and Halliday, 1995; Chapter 29.5). In Parkinson’s disease, the NBM is affected and shows countless voluminous globular Lewy bodies and very long Lewy

2014 Ch 02

20/11/03

10:20 am

Page 57


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

neurites (Fig. 29.8). The major component of these structures are abnormally phosphorylated neurofilaments, while ubiquitin and -synuclein are present as well. The neurons perish prematurely (Arendt et al., 1983; Whitehouse et al., 1983c; Chapter 29.3; Purba et al., 1994; Braak et al., 1996, 2000). Loss of neurons in the NBM in Parkinson’s disease was first described by Lowry in 1913. In fact, pathology in this area was described before neuronal loss in the substantia nigra (Whitehouse, 1986). It is of interest to note that the NBM is the structure where Lewy originally, also in 1913, described these inclusion bodies that bear his name (for reference see Den Hartog Jager and Bethlem, 1960). The cholinergic deficit in Parkinson’s disease may lead to dementia and react on cholinesterase inhibitors (Emre, 2003). Nerve growth factor receptor immunoreactivity in the NBM neurons is reduced in Parkinson’s disease patients, either with or without dementia (Mufson et al., 1991). Microtubule-associated protein (MAP)-B is decreased in the NBM in Alzheimer’s disease and Parkinson’s disease, while in the latter condition MAP-A was also decreased (Sparks et al., 1991). It goes without saying that the problem discussed earlier (Chapter 2.4), that loss of neurons might only seem to take place when atrophic neurons lose their typical marker, also goes for the disorders discussed in this subchapter. In Huntington’s disease the NBM is not affected (Tagliavini and Pilleri, 1983). In certain neurodegenerative conditions, the degenerative changes in basal forebrain neurons are accompanied by a compensatory growth and reorganization of dendrites, as found in parameters such as dendritic length, dendritic arborization and shape of the dendritic field in Alzheimer’s disease. In aging and Wernicke’s encephalopathy, dendritic growth is largely restricted to “extensive” growth of terminal dendritic segments, resulting in an increase in the size of the dendritic tree. In Alzheimer’s disease, however, dendritic growth mainly results in an increase in the dendritic density within the dendritic field, designated as “intensive” growth. Moreover, in Alzheimer’s disease, aberrant growth processes are frequently observed in the vicinity of amyloid deposits (Arendt et al., 1995). In addition, it should be noted that a hypertrophy of the galanin network has been observed both in Alzheimer’s disease and Parkinson’s disease with dementia, indicating a plastic reaction of the NBM interneurons (Chan-Palay, 1988b). In diffuse neurofibrillary tangles with calcification or non-Alzheimer, non-Pick dementia with Fahr’s syndrome, which is a rare neurodegenerative disorder without hered-

57

itary burden, largely confined to Japanese patients, the NBM may be affected (Tsuchiya et al., 2002). Dysfunction of the NBM is thought to contribute to deficits of memory and cognition after head injury. Reduced levels of choline acetyltransferase activity have been found in the cerebral cortex of patients who died after a head injury. The majority of patients with head injury show signs of neuronal damage in the NBM due to mechanical distortion of the tissue and/or ischemic damage (Murdoch et al., 2002). Rett syndrome (1966) is a developmental disorder associated with cortical atrophy, loss of speech, sleep disorders, stereotyped hand movements mimicking hand washing, and severe mental deficiency. It almost exclusively affects females. Familial cases are an X-linked dominant disorder and the disease disorder maps to Xq28 (Shastry, 2001). Mutations encoding the X-linked methylcytosine-binding protein 2 were found in a proportion of Rett syndrome girls (Dunn and MacLeod, 2001). Reduced choline acetyltransferase and other cholinergic markers have been found in the neocortex, hippocampus, thalamus and basal ganglia, indicating that the basal cholinergic nuclei will be affected (see Chapter 26.5; Wenk, 1979a,b; Dunn and MacLeod, 2001). In addition a reduced amount of nerve growth factor is found in the frontal lobe of these patients that may also affect the functioning of the cholinergic basal forebrain nuclei. In the hypothalamus of children that died from sudden infant death syndrome (SIDS) a reduction in choline acetyltransferase was found (Sparks and Hunsaker, 2002). In autistic patients older than 18 years of age, the vertical limb of the diagonal band of Broca showed small neurons that were markedly reduced in number (Kemper and Bauman, 1998). In addition, a diminishment in nicotinic acetylcholine receptors was found in the neocortex (Court et al., 2000). No differences were found in choline acetyltransferase or acetylcholinesterase activity in the cerebral cortex and basal forebrain. Cortical muscarine 1 receptor binding in the parietal and frontal cortex was 30% lower in autistic subjects (Perry et al., 2001). Moreover, numerous swollen axon terminals (spheroids) were found in the DBB and NBM of autistic patients (Weidenheim et al., 2001). These observations implicate the cholinergic system in autism. In Down’s syndrome patients, who virtually all develop Alzheimer’s disease by the time they reach the age of 40, a cholinergic deficiency is present, as demonstrated by decreased levels of choline acetyltransferase and acetylcholinesterase, and a loss of cholinergic neurons in

2014 Ch 02

58 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:20 am

Page 58

D.F. SWAAB

the NBM (see also Chapter 26.5; 29.1). Especially in the hippocampus of older Down’s syndrome patients, significant reductions in nicotine binding were reported (Court et al., 2000). However, Down’s syndrome patients have normal levels of brain cholinergic markers during their first year of life (Casanova et al., 1985; Brooksbank and Balázs, 1988; Kish et al., 1989). The precociously impaired cholinergic activity in Down’s syndrome leads to impairment of the tuberoinfundibular cholinergic pathway, and so to somatostatinergic hyperactivity and low growth hormone responsiveness to growth hormonereleasing hormone (Beccaria et al., 1998). Despite the many neuropathological similarities between Alzheimer’s disease patients and Down’s syndrome patients with Alzheimer’s disease, Down’s syndrome patients fail to display the galanin-immunoreactive hypertrophy in the cholinergic system that is typically seen in Alzheimer patients (Mufson et al., 1998).

In neuroleptic malignant syndrome and schizophrenia, a strong loss of the large cholinergic neurons and a reduction in choline acetyltransferase in the cerebral cortex have been found (Kish et al., 1990; Caroff and Mann, 1993). Van Buttlar-Brentano already reported that many of the NBM neurons were atrophic in schizophrenic patients. In addition, he observed swelling, shrinkage, disappearing cell bodies, accumulation of lipofuscin, coarse fatty vacuolization, glassy appearance and liquefaction. Some of the earliest pharmacological treatments for schizophrenia included cholinergic agents and various authors have suggested that cholinergic activity may be reduced in schizophrenia. The cholinergic system seems to exert a damping effect on the emergence of positive symptoms and an intensification of negative symptoms, but on the basis of the pharmacological observations, an increased activity of the cholinergic system is presumed to exist in schizophrenia (Tandon et al., 1999), which

Fig. 2.9. Corpora amylacia (arrows) in the substantia innominata, ventrally of the nucleus basalis of Meynert thionine staining. Bar indicates 20 m.

2014 Ch 02

20/11/03

10:20 am

Page 59


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 91

makes the cholinergic involvement in schizophrenia a controversial topic. An increased frequency of two dinucleotide polymorphisms have been found in the 7-nicotinic receptor gene of schizophrenics and of bipolar depressive and schizoaffective patients. This not only may play a role in the pathogenesis of these psychiatric diseases, but could also be responsible for many schizophrenics’ heavy smoking (Court et al., 2000; Stassen et al., 2000). One may wonder what the scientific basis was for the operation in patients with schizophrenia and with uncontrollable aggressive states often associated with mental retardation, in which the substantia innominata was one of the areas lesioned in multitarget limbic psychosurgery. The authors reported improvements of such patients following surgery. The targets included the amygdala, substantia innominata and cingulum (Cox and Brown, 1977). These controversial operations were neither very well controlled nor very well documented. Following herpes simplex encephalitis, MRI showed involvement of the substantia innominata and of the corpora mamillaria in patients who were left with memory difficulties (Kapur et al., 1994). Exposure to very high doses of methamphetamine causes a severe (up to 94%) depletion of choline acetyltransferase in the NBM, whereas no effect seems to occur after chronic use of cocaine or heroin (Kish et al., 1999). In the sub-pial region of the substantia innominata and in the vicinity of the optic chiasm and optic tract, and in

59

the floor of the third ventricle, large numbers of corpora amylacia are often observed (Fig. 2.9). They were first described in 1827, by P.E. Purkinje, and are also called amyloid bodies, polyglucosan bodies, or starch bodies. Their sizes range from 2–20 m in diameter. While usually spherical, oval or elongated forms may occur as well, concentric laminated patterns are frequently seen, the centers staining more densely than the periphery. Corpora amylacia are present in low numbers and are smaller (5 m) in children. After the age of 30–40 years, they generally become larger both in size and number. They are said to be present in larger numbers in Alzheimer’s disease, amyotrophic lateral sclerosis and multiple sclerosis, but the quantitative data to prove this are lacking. Mostly they occur in astroglia, but they are also found in axons. They have not been reported within neuronal perikarya at any time in normal subjects. Corpora amylacia are principally composed of polysaccharides, i.e. hexoses, but also contain carbohydrates, proteins, and polyglucosans. They react with many antibodies, but one may wonder how specific the majority of these reactions are. Anyhow, components from neurons, astrocytes, macrophages, and oligodendrocytes have been reported to be present in corpora amylacia. It is presumed that potentially damaging materials and nondegradable products of the aging process form the basis for these structures (for review see Cavanagh, 1999; Hoyaux et al., 2000; Fig. 8.16).

2014 Ch 03

20/11/03

10:27 am

Page 61


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

CHAPTER 3

Islands of Calleja (insulae terminalis)

the way up to the central nucleus of the bed nucleus of the stria terminalis. Studies in rat indicate that, on the basis of morphology, connections and neurotransmitters, the islands resemble the striatopallidal systems (Fallon et al., 1983). The islands of Calleja receive terminal and vomeronasal fibers (Bossy, 1980; Schwanzel-Fukuda and Pfaff, 1994; Chapter 24.2c). Fibers in the islands contain acetylcholinesterase and choline acetyltransferase (Alheid et al., 1990). A high neuropeptide-Y fiber density has been found in the human islands of Calleja, the core of which appears to be devoid of immunoreactivity (Walter et al., 1990). In addition, substance-P fibers (Walter et al., 1991), VIP (Fig. 3.1b), and a few somatostatin, enkephalin, and tyrosine hydroxylase-positive fibers (Lesur et al., 1989) are seen. In fact, catecholaminergic fibers are already present in the islands of Calleja in human fetuses that are only 3–4 months old (Nobin and Björklund, 1973). A high density of opiate and dopamine receptors has been found (Heimer, 2000). In rat, the islands of Calleja contain receptors for estrogens and cells that produce luteinizing hormone-releasing hormone (Fallon et al., 1983), and are therefore supposed to be involved in reproductive functions. So far we have found a weak to intermediate nuclear androgen receptor staining in the islands of Calleja, but not a sex difference (FernandézGuasti et al., 2000). The nuclear staining for ER- too, was as pronounced in young men as it was in young women (Kruijver et al., 2003). However, the ER- immunoreactivity was somewhat higher in men than in women. The presence of somatostatin in the islands of Calleja (Lesur et al., 1989) explains the staining with the Alzheimer antibody Alz-50 of normal, thinly beaded fibers in nondemented young controls. This antibody

The islands of Calleja were called “insulae terminalis” by Sanides (1957), because he suggested that they might be progenitor cells arrested in development. They are also called “interface islands” and “granular islands” (Heimer, 2000), as they are characterized by a dense core of small “glia-like” granule cells that belong to the smallest neurons of the brain (5–10 m; Meyer et al., 1989; Alheid et al., 1990). Other islands, i.e. the parvicellular islands, contain somewhat larger neurons (Heimer, 2000). The islands often lie in a region with few cells and are distributed over the substantia innominata in the dorsal area of the nucleus basalis of Meynert and in the lateral area of the diagonal band of Broca (Meyer et al., 1989; Fig. 3.1a; Heimer, 2000). They are situated in a strand that runs all 61

2014 Ch 03

62 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 3.1.

20/11/03

10:27 am

Page 62

D.F. SWAAB

VIP innervation of an island of Calleja; (a) cells of an island stained by thionine; (b) VIP innervation of the same island (bar = 10 m). (J.N. Zhou, unpublished results.)

cross-reacts with somatostatin and the decreasing staining in Alzheimer patients and thus indicates that somatostatin production is affected (Van de Nes et al., 1993). In Alzheimer patients both /A4 staining of Congo-

negative amorphic plaques and Alz-50-positive dystrophic neurites and perikarya are found in the islands of Calleja, indicating amyloid and cytoskeletal Alzheimer changes in these patients (Van de Nes et al., 1993).

2014 Ch 04

22/1/04

12:16 pm

Page 63


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

CHAPTER 4

Suprachiasmatic nucleus (SCN) and pineal gland (Fig. 4A)

(Fig. 4.1). Damage to the SCN results in a disappearance of circadian rhythms (Fig. 4.2; Chapter 4.1a) Precise estimation of the periods of the endogenous circadian rhythms of melatonin, core body temperature and cortisol in healthy young and older individuals in carefully controlled lighting conditions has revealed that the intrinsic period of the human circadian pacemaker averages 24.18 ± 0.04 hours (Czeisler et al., 1999). Exposure to very dim light or the scheduled sleep–wake cycle itself entrains the near-24-h intrinsic period of the human circadian pacemaker (Wright et al., 2001). The endogenous biological rhythms enable the organism to anticipate rhythmic changes in the environment and are consequently important adaptive processes. The SCN plays a central role in the generation and regulation of biological rhythms (Buijs and Kalsbeek, 2001). From cell cultures it is apparent that the SCN contains a large population of autonomous single-cell circadian oscillators and that synapses formed in vitro are neither necessary for the operation of these oscillators, nor sufficient for their synchronization (Welsh et al., 1995). The biological basis for morning or evening patterns (“early birds” and “night owls”) is based on fundamental properties of the circadian pacemaker. The circadian pacemaker of morning types is entrained to an earlier hour with respect to both clock time and wake time (Duffy et al., 2001). A season of birth variations was found in the morningness–eveningness preference among adults. The group born during the period April to September had a lower proportion of morning types than the group born during the period October to March (Natale et al., 2002). The relationships between circadian period and morningness–eveningness, circadian phase, and waketime are lost with aging (Duffy and Czeisler, 2002). The great variability in sleep time between long sleepers (> 9 hours)

Cyclicism, which may be diurnal, lunar or seasonal, is a peculiarity of many physiological processes . . . That these processes are somehow under the control of this ancestral diencephalo-hypophysial apparatus seems most probable. Harvey Cushing, 1932, p. 38

From the moment of conception (Chapter 4.2) to the moment we die (Chapter 4.1b), biological rhythms play a prominent role in our lives. Whereas environmental periodic phenomena only entrain or synchronize biological rhythms to the environmental changes through their direct and indirect input in the suprachiasmatic nucleus (SCN), it is the SCN itself that creates these rhythms 63

2014 Ch 04

64 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:16 pm

Page 64

D.F. SWAAB

Fig. 4A. Different clocks marking the same time. The Ghetto of Prague (Reproduced with permission.)

and short sleepers (< 6 hours) persists under constant environmental conditions and seems thus to be based upon individual differences in the circadian pacemaker. Although the role of the SCN for circadian rhythms (circa = approximately, dies = day) is well established in mammals, it is less clear for rhythms with longer periods, i.e. weekly, monthly and yearly rhythms (Swaab et al., 1996; Chapter 4.1). For periods with an even longer periodicity, such as the intervals of cranial suture closing, which occur in a rhythmic pattern of 7 years (Verhulst and Onghena, 1997), a relationship to the SCN has not even been presumed so far. In addition to or as part of its function as a biological clock, the SCN may be involved in a large number of functions, such as sexual behavior (Chapter 4.4) and glucose homeostasis (Chapter 30, Nagai et al., 1996; Hall et al., 1997). Since the circa-

dian organization brings about predictable changes in the body’s tolerance and tumor responsiveness to anticancer agents, the clinical relevance of the chronotherapeutic principle is becoming more and more important. Indeed, survival largely improved, e.g. with evening rather than morning administration of maintenance chemotherapy in children with acute lymphoblastic leukemia (Lévi, 2001). (a) The circadian system The eye as the metronome of the body. Lubkin et al., 2002

The suprachiasmatic nucleus (SCN) is a small structure that is considered to be the major circadian pacemaker of the mammalian brain and to coordinate all hormonal and behavioral circadian and circannual rhythms (Rusak

2014 Ch 04

22/1/04

12:16 pm

Page 65

SUPRACHIASMATIC NUCLEUS AND PINEAL GLAND

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

65

Fig. 4.1. Schematic overview of the inputs to the suprachiasmatic nucleus and their interactions that may be relevant for the concept of SCNstimulation. For reasons of clarity, temperature input is only shown for the SCN, whereas thermosensitivity has in fact been demonstrated in the pineal, SCN, septum, raphe nuclei, locus coeruleus and somatosensory afferents. Inputs are in outlined front, structures in bold, tracts in normal font and neurotransmitters and hormones in italics. Abbreviations: 5-HT = 5-hydroxytryptamine (serotonin); DR = dorsal raphe nucleus; GABA = gamma-aminobutyric acid; GHT = geniculohypothalamic tract; Glu = glutamate; IGL = intergeniculate leaflet; LC = locus coeruleus; MR = median raphe nucleus; NA = noradrenalin; NPY = neuropeptide Y; RGC = retinal ganglion cells; RGT = retinogeniculate tract; RHT = retinohypothalamic tract; SCN = suprachiasmatic nucleus; SHT = spinohypothalamic tract. (From Van Someren et al, 1999; with permission.)

and Zucker 1979; Hofman et al., 1993). The vasopressin subnucleus of the SCN has a volume of 0.25 mm3 and contains some 10,000 vasopressin neurons on each side (Swaab et al., 1985; Figs. 1.7 and 1.8). Despite the fact that cells of many organs have the ability to retain a rhythmic function for a few days, this property disappears if it is not enforced daily by the SCN (Buijs and Kalsbeek, 2001). Animal experiments have shown that lesions restricted to the SCN make them totally arrhythmic, while transplantation of a fetal SCN may restore circadian activity rhythms in such lesioned animals (Drucker-Colin et al., 1984; Aguilar-Roblero et al., 1986; Lehman et al., 1987; De Coursey and Buggy, 1989; Saitoh et al., 1990;

Griffioen et al., 1993; Van Esseveldt et al., 2000). Ralph et al. (1990) even showed that, following transplantation of a fetal SCN into the hypothalamus of an SCN-lesioned animal, the circadian activity of the recipient adapted itself to the circadian rhythm of the donor. Interestingly, transplantation of the SCN results in restoration of activity patterns, but not of cortisone rhythms. A few clinical observations support the importance of the SCN for circadian rhythms in humans. A lesion in the suprachiasmatic region of the anterior hypothalamus, e.g. as the result of metastasis, indeed results in a decreased expression of vasopressin in the SCN and disturbed circadian rhythms in human beings (Scully

2014 Ch 04

66 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:16 pm

Page 66

D.F. SWAAB

Fig. 4.2. Metastasis affecting the suprachiasmatic nucleus function. Scully et al. (1983) and Schwartz et al. (1986) have described a 55-year-old postmenopausal woman patient with a discrete metastasis of an adenocarcinoma of the rectum in the ventral hypothalamus, optic chiasm, and neurohypophysis (a, thionine staining) who, while she was admitted into hospital for the final time, developed an abnormal daily rhythm of oral temperature. She had hypothalamic diabetes insipidus, low FSH, blurring of vision in the periphery of the right temporal field, and required more sleep at night. The metastasis was located between the infundibulum, carotid artery and optic nerve. The infundibulum was pushed into the hypothalamus. The mass also infiltrated downward along the pituitary stalk. The white granular mass extended into the supraoptic recess. The fornices were pushed laterally by the tumor. We determined 1964 vasopressin-expressing neurons in the SCN (b; in region indicated by arrowhead in (a), which was only 23% of the control values for the group of 50- to 80-year-old women (8370 ± 950 vasopressin neurons, n = 8). This observation supports the importance of the activity of vasopressin neurons for the expression of circadian rhythms in the human (Bar in (a) = 1 mm, in (b) = 100 m).

et al., 1983; Schwartz et al., 1986; Cohen and Albers 1991; Fig. 4.2). In a patient with a hypothalamic astrocytoma destroying the SCN bilaterally, reversal of the day/night rhythm of the wake/sleep pattern was also reported (Haugh and Markesbery, 1983). It should be noted, though, that in that patient not only the area of the SCN, but also a large part of the hypothalamus was affected. Moore (1992) briefly reported on a patient with an optic nerve glioma who had evidence of loss of rhythmicity in several functions and who had compression of the chiasmatic area visualized by a CT scan. In addition, a patient with septo-optic dysplasia (see Chapter 18.3b) has been described who had arrhythmicity of the type seen after a lesion of the SCN. Sleep– wake cyclicity was restored by melatonin. The scarce medical information given on the patients described by Krieger and Krieger (1966), with circadian disturbances as a result of a disease of the temporal lobe, pretectum or hypothalamus, does not allow conclusions on the possible direct involvement of the SCN in this type of disorder. Interestingly, the patient reported by Cohen and Albers (1991), who experienced disruption of circadian timing following a lesion of the SCN region resulting from

damage from a resection of a craniopharyngioma, was also tested for her temporal consistency in a later study (Cohen et al., 1997). She appeared to have a serious disruption of short-duration timing capacity. Severe impairment of time perception was also evident on duration discrimination. The SCN may thus be involved, not only in circadian regulation, but also in timing mechanisms of shorter duration. (b) Disorders of clock function and circadian rhythms in disorders (Table 4.1) Only a few reports indicate the presence of morepronounced circadian rhythms, such as the cortisol rhythm in post-traumatic stress disorders (Yehuda et al., 1994, 1995a, 1996; Chapter 8.5d). In contrast, quite a substantial number of human beings, i.e. some 33%, exhibit a desynchronization of their internal time structure. This goes, e.g. for body temperature and blood pressure (Abitbol et al., 1997). Individual differences in phase-relationship between temperature and sleep are also related to someone being a morning or an evening person. The temperature minimum for morning types occurs at 03.50 a.m., and for evening types at 06.01 a.m. (Baehr et al., 2000). The total

2014 Ch 04

22/1/04

12:16 pm

Page 67


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

67

TABLE 4.1 Disorders of clock function. Abnormalities in entrainment Advanced sleep disorder African trypanosomiasis Aging, Alzheimer’s disease Antarctic winter Blindness Delayed sleep disorder Depression Jet lag Parkinson’s disease (phase advance) Retinitis pigmentosa Rett syndrome Shift work Spaceflight Other factors outside the SCN Aging, Alzheimer’s disease Chronic liver disease Depression Glucocorticoid administration Hydrocephalus Hypothalamic tumors Oral contraceptives Surgery Disturbed pacemaker Aging, Alzheimer’s disease Depression Cushing’s syndrome, corticosteroids Familial advanced sleep-phase syndrome

heritability for morningness–eveningness is around 45% (Vink et al., 2001a). Many conditions cause disturbed circadian rhythms (Table 4.1). Several factors outside the biological clock may be involved in disorders of clock function. Circadian disturbances are present in patients with chronic liver diseases (Blei and Zee, 1998). Circadian rhythms may be disturbed by hypothalamic tumors in the region of the SCN (see Chapter 4.1a). In addition, when third ventricular tumors cause ventricular obstruction with consequent increased intracranial pressure and/or hydrocephalus, circadian temperature fluctuations disappear (Page et al., 1973). Moreover, circadian disorders occur in aging and Alzheimer’s disease (Chapter 4.3), in depression (Chapter 26.4), during surgery (Guo et al., 2002) and following glucocorticoid administration (Madjirova et al., 1995). In patients with Cushing’s syndrome, the 24-h blood

Fatal familial insomnia Hepatorenal syndrome Hypertension Irregular (non-24-hour) sleep–wake syndrome Multi-infarct dementia Neuronal ceroid lipofuscinosis (CLN5) Nocturnal diuresis Rett syndrome Septo-optic dysplasia Shy–Drager syndrome (multi-system atrophy) Stroke Smith–Magenes syndrome Circadian patterns in diseases Cancer Intracerebral hemorrhage Ischemic stroke Migraine, cluster headache, hypnic headache syndrome Non-dipping hypertensives Progressive dystonia Recurrent stupor Restless legs syndrome Seasonal mood disorders Seasonal affective aggressiveness SIDS (sudden infant death syndrome) Subarachnoid hemorrhage in hypertensives Symptoms of depression Tardive dyskinesia Tremor in Parkinson’s disease

pressure oscillation was disrupted, while that of heart rate was preserved (Piovesan et al., 1990). Glucocorticoids diminish vasopressin mRNA in the human SCN (Liu et al., 2002), which may be the basis of the disruption of circadian rhythms during glucocorticoid treatment or in Cushing’s disease. In addition, circadian rhythms are altered by oral contraceptives (Reinberg et al., 1996). Totally blind people often lack the entraining effects of light and may show free-running temperature, cortisol, and melatonin rhythms. Because of their drifting intrinsic periodicity, totally blind people may also suffer from recurrent sleep disturbances (Bodenheimer et al., 1973; Sack et al., 1992; Lewy and Sack, 1996; Skene et al., 1999). Administration of melatonin can entrain (synchronize) circadian rhythms in most blind people who have free-running rhythms (Sack et al., 2000; Lewy et al., 2001;

2014 Ch 04

68 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:16 pm

Page 68

D.F. SWAAB

Chapter 4.5). Surprisingly, some blind people maintain circadian entrainment by light and show light-induced suppression of melatonin secretion, despite the apparently total lack of pupillary light reflexes, and with no conscious perception of light (Czeisler et al., 1995; Hätönen et al., 1998). It has been proposed that in these patients the retinohypothalamic pathway that innervates the SCN (Dai et al., 1998a; Chapter 4.1e) would still be intact (Czeisler et al., 1995). It might be of practical importance to recognize these patients, since enucleation of the eyes might cause recurrent insomnia and other symptoms associated with the loss of entrained circadian rhythms. The observations in patients with a tumor in the SCN region, as well as those in blind people, emphasize the importance of the light–dark cycle for synchronization and of the SCN for circadian rhythms in the human species. A free-running rhythm has also been observed in a child with Rett syndrome (Miyamoto et al., 1999; Chapter 26.5). During the Antarctic winter melatonin and cortisol rhythms free run, while when the sun returns during spring all rhythms may again synchronize and entrain to the daylight (Kennaway and Van Dorp, 1991). Disorders of the sleep–wake cycle involving entraining effects are observed in “jet lag” syndrome, and melatonin could be used to promote adaptation to night work and jet travel (Sharkey and Eastman, 2002). The typical manifestations are insomnia during local sleep time, day fatigue, reduced concentration, irritability, and exhaustion with mild depression. Circadian rhythms such as the temperature rhythm shifts are out of phase after transmeridian flights. Psychometric evaluation showed that desynchronization affected the functioning of pilots (Ariznavarreta et al., 2002). Melatonin is used by many during jet travel, but a randomized double-blind trial to determine the effectiveness of melatonin in the alleviation of jet lag revealed a lack of response (Spitzer et al., 1999). Chronic jet lag may produce temporal lobe atrophy and deficits in learning and memory. Such deficits were more obvious in flight attendants who had a short (5-day) recovery period than in those who had a long (14 days) period between outward meridian flights. The proposed mechanism was based upon the increased cortisol levels that are significantly higher in certain crew members after repeated exposure to jet lag than after short-distance flights. Temporal lobe atrophy and spatial cognitive defects were claimed to be prevented by adequate recovery periods between consecutive jet lags (Cho, 2001; cf. Chapter 8.5e). However, this small study has quite

a number of limitations (Van Someren, 2002). In addition, performance may be affected, as it was shown that jet lag on eastward trips was disadvantageous for a U.S. baseball team (Recht et al., 1995). Sleep-onset insomnia and early morning-awakening insomnia may be caused by, respectively, delays and advances of circadian rhythms. Sleep-onset insomnias go together with body temperature delays of about 3 hours. Early morning-awakening insomnias have significant phase advances of 2–4 hours for temperature and melatonin, while the 0–4 hour advances of the sleepiness rhythm were not significant, possibly due to errors in the identification of the sleep phases. The therapeutic implication of this finding would be that early morningawakening insomnia could be treated effectively by evening bright-light therapy that would cause a phase delay in the circadian rhythms (Lack et al., 1996). Recommendations have been provided for the use of light therapy in delayed and advanced sleep-phase syndrome (Chesson et al., 1999). Phase advance of the melatonin rhythm was observed in Parkinson’s disease (Fertl et al., 1991). Bright light exposure, the most effective entraining factor, has been sucessfully employed to treat advanced and delayed sleep-phase syndrome, jet lag, shiftwork, maladaptation and non-24-hour sleep–wake syndrome (Singer and Lewy, 1999; Cole et al., 2002a). Delayed sleep-phase syndrome has been described following, e.g. traumatic brain injury, in association with a structural polymorphism in the hPer3 gene (Ebisawa et al., 2001). The genetic transmission of this syndrome is either an autosomal dominant mode of inheritance with incomplete penetration or a multifactorial mode. Both the paternal and maternal branch may contain affected individuals (Ancoli-Israel et al., 2001). As a putative mechanism for the delayed sleep-phase syndrome, hypersensitivity of melatonin suppression in the evening has been proposed, suggesting that evening light restriction may be an important measure in this syndrome. The traumatic form of the delayed-sleep phase syndrome was effectively treated with melatonin (Nagtegaal et al., 1997; Kamei et al., 2000a). A double-blind placebo controlled study has also confirmed the effectiveness of melatonin in the treatment of whiplash syndrome (Nagtegaal et al., 1998; Smits and Nagtegaal, 2000). African trypanosomiasis, caused by the tsetse fly, is accompanied by a disappearance of the 24-h rhythmicity of sleep and wakefulness. It is presumed that the serotonergic system

2014 Ch 04

22/1/04

12:16 pm

Page 69


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

is involved in the dysregulation of the SCN (Buguet, 1999). Retinitis pigmentosa, a neurodegenerative disease involving photoreceptor degeneration, is accompanied by a sleep disorder of a circadian nature (Ionescu et al., 2001). Refusal to attend school in Japanese children and adolescents was found to go together with desynchronization of their biorhythms, particularly the circadian rhythm of body temperature and sleep–wake (Tomoda et al., 1997). Sleep disturbances in astronauts occur frequently. The circadian phase of body temperature was delayed by about 2 hours compared with baseline conditions on the ground, prior to the flight. A free-running rhythm was not observed during the first 30 days in space, but sleep was shorter and subjective sleep quality was diminished. In addition, the structure of sleep was significantly altered; there was more wakefulness, less slow-wave sleep, and melatonin did not improve sleep (Gundel et al., 1997; Dijk et al., 2001; Chapter 30.7). Individual circadian desynchronization of various circadian rhythms has been documented during isolation experiments without time cues and a genetically controlled variability has been suggested (Ahskenazi et al., 1993). Irregular (non-24 hours) sleep–wake syndrome is a disorder in which the circadian pacemaker is probably disturbed (Moore, 1992; Regestein and Monk, 1995; McArthur et al., 1996; Schwartz, 1997). A patient with this syndrome responded to phototherapy. This treatment immediately changed the free-running sleep–wake and body temperature rhythm to the environmental 24-hour rhythm (Watanabe et al., 2000). Circadian disturbances are also found in Rett syndrome (Miyamoto et al., 1999), depression (Chapter 26.4f), aging and Alzheimer’s disease, where the SCN is affected (Chapter 4.3). In addition, in subcortical stroke patients, who have infarcts in the vasculature impacting the blood supply to the hypothalamus, the function of the SCN seems to be interrupted, as indicated by a greater daytime sleepiness (Bliwise et al., 2002). Tourette syndrome is associated with a circadian dysregulation of the body temperature profile (Kessler, 2002), but the SCN has not been studied in these patients. In necrotisism a weakened circadian pacemaker is presumed (Murray et al., 2002). In Cushing’s syndrome (Stewart et al., 1992; Bierwolf et al., 2000) and corticosteroid treatment the circadian rhythm is disturbed. Corticosteroids cause a decrease in vasopressin synthesis in the SCN (Liu et al., 2002). In a

69

patient with multiple system atrophy, a decrease in the number of vasopressin neurons in the SCN was found, accompanied by gliosis in the SCN and nocturnal polyuria (Ozawa et al., 1998). The decreased plasma levels of vasopressin during the night were confirmed in a larger sample of patients with this disorder (Ozawa et al., 1999). One Parkinson patient with a 48-hour sleep–wake cycle has been described (Mikami et al., 1987). In patients with primary hypertension, three major neuronal populations of the SCN, i.e. vasopressin, VIP and neurotensin, were reduced by more than 50%, which is considered to be the basis of abnormalities of diurnal rhythms of these patients (Goncharuk et al., 2001). Interestingly, a transgenic hypertensive mouse strain showed an altered light-entrainment response, accompanied by suppressed c-fos mRNA expression in the suprachiasmatic nucleus (Lemmer et al., 2000), confirming the possible involvement of the SCN in hypertension. An autosomal dominant type of familial advanced sleep phase syndrome with high penetrance has been described. The profound phase advance of the sleep–wake, melatonin and temperature rhythm was associated with a very short , i.e. a 4-hour advance of the daily sleep–wake rhythm, indicating a disorder of the circadian generating system in the biological clock (Jones et al., 1999). In one family the responsible mutation was found near the telomere of chromosome 2q, where human PER2 is also found. Affected individuals appeared to have a serine-to-glycine mutation within the casein kinase 1 binding region of hPER2, which causes hypophosphorylation of this kinase (Toh et al., 2001). Smith–Magenis syndrome is caused by interstitial deletions of chromosome 17p11-2. The children have a phase advance with a paradoxal excretion of melatonin, hyperactivity, attention deficit, self-injury, temper tantrums and a major sleep disturbance. Fatal familial insomnia is an autosomal dominant prion disease related to a point mutation at codon 128 or another point mutation of the prion protein gene on chromosome 20 (Parkes, 1999). This disease is characterized by a loss of circadian sleep–wake, autonomic and hormonal rhythms, sympathetic hyperactivity and progressive neurological motor signs and degeneration of thalamic nuclei, whereas, judging by routine pathology, the hypothalamus seemed to be spared (Avoni et al., 1991; Montagna et al., 1995; Lugaresi et al., 1998; Rossi et al., 1998a; Cortelli et al., 1999). However, investigation of the hypothalamus should certainly be repeated by the use

2014 Ch 04

70 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:16 pm

Page 70

D.F. SWAAB

of immunocytochemical procedures in order to study possible changes in the various subpopulations of the SCN, since advanced stages of the disease are characterized by a loss of circadian rhythms (Montagna et al., 1995; Cortelli et al., 1999). Interestingly, the expression of the prion protein mRNA in the rat hypothalamic nuclei shows a clear circadian rhythm (Cagampang et al., 1999). Chronic motor activity and loss of a circadian rest–activity rhythm, as found in a patient with total insomnia (Plazzi et al., 1997), suggests that the SCN is affected in this disease. Since the serotonergic system is considered to play an integral role in the regulation of sleep homeostasis and SCN function, it is of considerable interest that the number of serotonin-synthesizing neurons immunocytochemically stained for tryptophan hydroxylase in the median raphe showed a substantial increase of some 62%, indicating that a change in the serotonergic input of the SCN may be a crucial factor in this sleep disorder (Wanschitz et al., 2000). Using SPECT, a reduced availability of serotonin transporters of 53% and 73% respectively in the thalamus–hypothalamus region of two patients with fatal familial insomnia provided the first in vivo support for such an alteration in the serotonergic system in this disorder (Klöppel et al., 2002). In one patient with fatal insomnia, a microadenoma of the pituitary was found (Lugaresi et al., 1986, 1987), but it is not clear whether it has also contributed to the lack of hormonal rhythms. Circadian fluctuations of symptoms are also found in a number of diseases. Tremor, e.g. in Parkinson patients, shows strong circadian fluctuations with a clear decline during the night (Van Someren et al., 1993; Fig. 4.3). Daily rhythms are important factors in the expression of various diseases, e.g. in ischemic stroke and intracerebral hemorrhage, which show a postawakening morning peak; whereas a similar rhythm for subarachnoid hemorrhage has been recorded in hypertensive, but not in normotensive, patients (Schwartz, 1997). The ominously increased rate of cardiovascular events in the morning hours may reflect the sudden rise of sympathetic activity and the reduction of vagal tone (Furlan et al., 1990; Herlitz et al., 2002). The incidence of subarachnoid hemorrhage conforms to circadian blood pressure variation in hypertensive patients, similar to the diurnal rhythms observed with strokes and myocardial infarctions. Normotensive individuals, in contrast, have a random 24-hour distribution of subarachnoidal hemorrhage (Kleinpeter et al., 1995). In this respect it is of interest that, generally, a significant nocturnal blood pressure fall (“dippers”) is

observed. In some essential hypertensives, this nocturnal fall in blood pressure is absent (“non-dippers”) (Coca, 1994). A flattened diurnal rhythm of heart rate in uncomplicated subjects with essential hypertension is a marker of risk for subsequent all-cause mortality (Verdecchia et al., 1998). Tardive dyskinesia and progressive dystonia with diurnal variation (Segawa’s dystonia) worsen with time after awakening. Acute dystonic reactions to neuroleptics are more likely in the afternoon and evening. Migraine headaches seem to be more frequent in the morning. Headaches (Chapter 31.3) and epileptic seizures (Chapter 28.5) may be associated with sleep, while cluster headache (Chapter 31.3a) may be linked with REM sleep (Schwartz, 1997) and migraine may start during nocturnal sleep (Dexter and Riley, 1975; Chapter 31.3b). Some authors even propose that the SCN may be the site where the migraine attack is initiated (Zurak, 1997). In addition, hypnic headache syndrome of the elderly has been described, characterized by recurrent nocturnal headaches that awaken patients from sleep at a consistent time each night and respond to treatment with lithium carbonate (Newman et al., 1990; Dodick et al., 1998; Chapter 31.3c). Sudden infant death syndrome occurs with a circadian periodicity with two peaks; one at 9 a.m. and one at 9 p.m. (Bilora et al., 1997; Cornwell et al., 1998). A 50-year-old patient has been described with stupor occurring in a clear circadian pattern, i.e. almost every afternoon, for a period of 10 years. Such stuporous attacks lasted from 5 p.m. to 8 p.m. (Trenkwalder et al., 1997). Seasonal mood disorders, circadian fluctuations in the symptoms of depression, and the possible involvement of the SCN are discussed in Chapter 26.4f. A decreased amplitude of the melatonin rhythm was observed in patients with cancer, with lower levels during the night and higher levels during the day (Tarquini et al., 1999). Circadian rhythm disturbances also seem to be involved in the experience of fatigue and depression in cancer patients (Roscoe et al., 2002). However, in a preliminary study, M.A. Hofman (unpublished results) could not find a difference in the number of vasopressin-expressing neurons in the SCN of patients with solid tumors, patients with leukemia and patients with other diagnoses. The normal circadian pattern in vasopressin blood levels with higher levels during the night is absent in nocturnal diuresis (see Chapter 22.4), in hepatorenal syndrome, also known as functional renal failure of liver cirrhosis (Pasqualetti et al., 1998), and in Shy-Drager syndrome (multi-system atrophy) patients who exhibited nocturnal diuresis (Ozawa et al., 1993, 1998; Chapter 29.7d).

2014 Ch 04

22/1/04

12:16 pm

Page 71


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

71

Fig. 4.3. 24-h plots of movement classified as ‘arrhythmic activity’, ‘rest’ and ‘tremor’ on an hourly basis over, A: 10 young subjects, B: 10 aged subjects, C: 8 patients before thalamotomy, D: same 8 patients after thalamotomy. In all panels the upper shaded area represents the amount of ‘tremor’, the unshaded area the amount of ‘rest’ and the lower shaded area the amount of ‘arrhythmic activity’. Note that a considerable amount of ‘tremor’ is found only in pre-operative patients. Also note the lack of ‘tremor’ and ‘arrhythmic activity’ at night. (From Van Someren et al., 1993; Fig. 4, with permission.)

(c) Chemoarchitecture In conventionally 6- to 10-m thionine-stained paraffin sections, the human SCN cannot be recognized with certainty and therefore immunocytochemical labelling of this structure, e.g. with anti-vasopressin and antivasoactive intestinal polypeptide (VIP) or anti-neurotensin, is necessary (Swaab et al., 1985, 1990, 1994b; Moore, 1992; Dai et al., 1997; Figs. 1.7 and 4.4–4.6). The region of the SCN that receives retinohypothalamic tract input and is therefore considered to be of importance for entrainment is characterized by VIP neurons (Moore, 1992). Vasopressin is found in the remainder of the SCN (Figs. 4.4–4.6) and neurotensin is found in the entire SCN (Moore, 1992). Neurotensin may play a role in modulating circadian pacemaker function by suppression of the neuronal firing rate in the SCN. Vasopressin might

amplify the rhythm in this nucleus by its excitatory effect during the light phase, as shown by animal experiments (Ingram et al., 1996), and was found to play a role in synchronizing circadian rhythms to the light-entrainable oscillator (Murphy et al., 1998a). The observation that intranasal vasopressin markedly enhanced nocturnal slowwave sleep in humans (Perras et al., 1996) should perhaps be considered in the light of the effect of this SCN neuropeptide on other brain regions. Neurons that are immunoreactive for vasopressin, VIP, neuropeptide-Y, thyrotropin-releasing hormone (TRH), or neurotensin are present in the SCN in a particular anatomical organization (Figs. 4.4–4.6; Mai et al., 1991; Moore 1992; Fliers et al., 1994). In addition, somatostatin (Bouras et al., 1986, 1987), galanin (Gai et al., 1990), preproenkephalin (Sukhov et al., 1995), delta-sleep-inducing peptide (Najimi et al., 2001), and hypocretin fibers (Moore et al., 2001) are present in the

2014 Ch 04

72 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:16 pm

Page 72

D.F. SWAAB

Fig. 4.4. Diagram showing the organization of the human SCN. The distribution of vasopressin (VP), vasoactive intestinal polypeptide (VIP), neuropeptide-Y (NPY), neurotensin (NT), neurons (large black dots) fibers (small grey points) is shown at three levels, from rostral to caudal. (From Moore, 1992; Fig. 6, with permission.)

SCN. VIP-binding sites (Sarrieau et al., 1994), the long form of the leptin receptor (Burguera et al., 2000), ER and , the progesterone receptor (Kruijver and Swaab, 2002; Fig. 4.28) and melatonin receptors (Weaver et al., 1993) were also found in the SCN area. Since VIP is present in the SCN, it is not surprising that peptide methionine amide (PHM) is also present in the human SCN. PHM and VIP are encoded on two adjacent exons of a common prepro-VIP gene (Itoh et al., 1983; see Chapter 4.1d). Using confocal laser scanning microscopy, Romijn et al. (1999) found that a small percentage of the neurons in the human SCN colocalized vasopressin and VIP.

Following microwave treatment of sections, the staining of vasopressin and VIP becomes more sensitive. Due to this treatment, the volume of the vasopressin SCN subnucleus increased 2.4 times and that of VIP 4 times, the number of vasopressin neurons increased by 70% and the number of VIP neurons by 280%. The neurons that were visible without microwave treatment were localized mainly in the central part of the SCN, whereas the neurons that became visible only after microwave treatment could be found in the peripheral area of the subnuclei. This suggests that the vasopressin and VIP neurons in the central part of the SCN contain more peptide, possibly because they are more active than the peripheral ones (Zhou et al., 1996). One can wonder whether this is a general characteristic of hypothalamic nuclei. The shape of the human SCN as stained by antivasopressin is sexually dimorphic, i.e. more elongated in women and more spherical in men, but the vasopressin cell number and volume of this SCN subnucleus are similar in both sexes (Swaab et al., 1985). VIP-expressing neurons show strong age-dependent sex differences (Swaab et al., 1994b; Zhou et al., 1995b, Fig. 4.25). One may presume that the sex differences in the SCN in vasopressin and VIP are related to sex differences in circadian functions (Ticher et al., 1994). In addition, sex differences in the SCN may also be relevant in relation to the involvement of the SCN in sexual behavior (see also Section 4.4). Many neurons in the human SCN contain the two isoforms of glutamic acid decarboxylase (GAD), GAD65 and GAD67. GABA (-aminobutyric acid) is colocalized with one or more peptides in SCN neurons (Gao and Moore, 1996a,b). GABA is generally known as an inhibitory neurotransmitter in the brain, but its action may depend on the circadian time. Wagner et al. (1997) have shown, however, that SCN neurons can be excited by GABA through a GABAA-dependent mechanism. In rat, the excitatory response to GABA is seen only during the day, when GABA opens the chloride ion channel and the membrane potential becomes more positive and action potentials are generated. The opposite happens during the night. Then GABA acts as an inhibitory neurotransmitter. An intermediate density of benzodiazepine binding sites is present in the SCN of the human fetus and neonate (Najimi et al., 2001). From cell cultures it appeared that GABA-ergic neurotransmission is involved in synchronization of circadian rhythms in individual SCN neurons (Shirakawa et al., 2000). GABA also has time-dependent effects on the pineal gland, which is innervated by the

2014 Ch 04

22/1/04

12:16 pm

Page 73


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

73

Fig. 4.5. Consecutive sections showing distribution of vasoactive intestinal polypeptide (VIP) (A) and vasopressin (VP) (B) cell bodies in the human suprachiasmatic nucleus (SCN). Note that the VP cell bodies are located more dorsally than the VIP cell bodies. Dorsal of the SCN VP and VIP terminals are present. Asterisk, blood vessel; 3V, third ventricle. Scale bar = 200 m. (From Dai et al., 1997; Fig. 3, with permission.)

SCN. The administration of a GABA antagonist in the PVN during the day caused a rise in rat pineal melatonin levels, while it had no effect at night (Kalsbeek et al., 2000a). A dense catecholaminergic network is found in the SCN of the human fetus from as early as the 3rd and 4th months of pregnancy (Nobin and Björklund, 1973). The SCN is also called “D13”, since cell bodies contain aromatic L-amino acid decarboxylase (AADC) but no tyrosine hydroxylase (TH) (Kitahama et al., 1998a). (d) Molecular basis of circadian rhythms and the characteristics and functions of VIP The molecular basis of the circadian fluctuations is transcription of clock genes and the synthesis of the proteins they encode (Figs. 4.7 and 4.8; Van Esseveldt et al., 2000). The circadian core oscillator is thought to be composed of an autoregulatory transcription- (post)transcription-based feedback loop involving a set of clock genes (Okamura, 2003). The group of Takahashi was the first to identify a gene responsible for a clock mutation in mice. The semidominant autosomal mutation abolished rhythmicity in constant darkness. The gene was christened clock and was identified by positional cloning and functional rescue of defective mice by transgenic expression of a bacterial arti-

Fig. 4.6. Caudal level of the SCN showing distribution of VIP (A) and vasopressin (VP) (B) cell bodies and fibers (compare to the more rostral SCN in Fig. 4.5. C and D are magnifications of the anterioventral hypothalamic area of A and B. VIP and VP fibers and terminals are visible in the ventral part of PVN ((arrow)heads). 3V, third ventricle. For abbreviations see Fig. 12. (From Dai et al., 1997; Fig. 4, with permission.)

ficial chromosome clone containing the gene (Antoch et al., 1997; King et al., 1997). The human CLOCK gene was found to be 89% identical to its mouse homolog (Steeves et al., 1999). A single nucleotide polymorphism in the 3 flanking region of the human CLOCK gene has been identified. Homozygotes or heterozygotes for the 3111C allele have higher scores on a measure of evening preference for activity vs. morning preference (Desan et al., 2000). In addition, Tei et al. (1997) identified the human homologue of the period (PER) gene in Drosophila that shares a PAS domain with the human arcyl hydrocarbon nuclear translocator (Huang et al., 1993). CLOCK-BMAL1

2014 Ch 04

74 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:16 pm

Page 74

D.F. SWAAB

Fig. 4.7. Schematic representation of genes of proteins involved in the Drosophila clock and their homologs identified in mammals. Functional domains in the nucleotide sequences are indicated by shaded or black boxes. Bar lengths indicate the relative sizes of the DNA sequences. Acidic: acidic region that might act as activator domain; ATP: ATP binding site; bHLH: basic helix-loop-helix; CLD: cytoplasmic localization domain; DEDD: tetrapeptide sequence; NLS; nuclear localization signal; PAS: PAS domain (indicated are both PAS A and B domains); PB: PER binding sequence; PER-C: C-terminal region of PER; poly Q: glutamine-rich activator domain; ST: serine-threonine kinase catalytic domain; TG: threonineglycine repeat. Note that TIMELESS is not a candidate clock gene anymore (Van Esseveldt et al., 2000; Fig. 8, with permission.)

heterodimers bind to DNA via a promotor sequence termed an E-box in the promotor region of vasopressin (Okamura, 2003). This drives the positive component of PER1, PER2 and PER3 transcriptional oscillations, which are thought to underlie circadian rhythmicity (Gekakis et al., 1998). Studies in mutant mice showed that Per3 was placed outside the core circadian clockwork. Per1 influences rhythmicity, primarily through interaction with other clock proteins, while Per2 positively regulates rhythmic gene expression (Bae et al., 2001). Mice carrying a Per-1 null mutation display a shorter circadian period with reduced precision and stability, while mice that are deficient in both Per1 and -2 do not express circadian rhythms and thus have distinct and complementary roles in the mouse clock mechanism (Zheng et al., 2001). Period 3 647 Val/Gly polymorphism was associated with self-reported morningness–eveningness scores (Johansson et al., 2003). The identification of the Drosophila clock gene double time (dbt), clarified the role of phosphorylation of PER in the timing of its nuclear entry and breakdown. DBT appears to be a structural homologue of the human casein kinase I, with 86% identity in the kinase domain. DBT can bind to cytoplasmic monomers of PER in vitro and was suggested to mediate PER phosphorylation, thereby ensuring the

instability and breakdown of monomeric PER proteins (Kloss et al., 1998). In familial advanced sleep syndrome a mutation was found in the casein kinase 1 binding region of human PER2 (Toh et al., 2001). The gene cycle (cyc; Rutila et al., 1998) appeared to be the Drosophila homologue of the human gene BMAL1 (MOP3; Hogenesch et al., 1998; Ikeda and Nomura, 1997), which codes for a variety of protein products, some of which resemble cyc. cyc shares 68% homology and 55% identity with this human gene, of which the function was not known until a role was suspected in circadian rhythmicity. CLOCK and cyc encode bHLH-PAS transcription factors (Sun et al., 1997). Most clock genes have a PAS/PER-Arnt-Sim) domain that mediates protein–protein interaction, regulates circadian rhythms and is related to transcription factors that act as heterodimers (Bunney and Bunney, 2000). Using the more relevant PAS domain of Drosophila per as a probe, two independent groups isolated and cloned a mouse and human per homologue which were named m-Rigui and h-Rigui (Sun et al., 1997) or mper and hper (Tei et al., 1997). hPer and mPer share 92% homology with each other and about 50% homology with per in each of five stretches. These stretches lie within the Nls, Pas, Cld, Per-C and Per-repeat regions. A database search for

2014 Ch 04

22/1/04

12:17 pm

Page 75


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

75

Fig. 4.8. Model for the clock mechanism in mammalian SCN pacemaker cells proposed on the basis of the clock mechanism in Drosophila and reports of mammalian homologs of Drosophila Clock and clock-related genes, including the involvement of CRY and VP. Three PERIOD proteins PER1, PER2 and PER3 are active in mammalian species, which can form either homodimers or heterodimers (with an homologous or heterologous PER or with CRY1 or CRY2) that inhibit expression of the per genes in a similar way as described for the PER/TIM complex in Drosophila (negative feedback loop). These dimers enter the nucleus and interact by an as yet unknown negative mechanism with the CLK/BMAL1 complex that normally stimulates per expression through attachment to the E-box site. CRYs can also inhibit per transcription without the presence of PER. PER and PER/CRY dimers have a similar negative effect on the expression on VP and CRY and perhaps also on other clock-controlled genes (CCGs). TIM has no rhythmic expression levels in the SCN. The absence of a circadian rhythm in Tim expression in the mouse SCN and in the in vivo absence of PER/TIM complexes point to different yet unestablished role for TIM than that described in the Drosophila clock system (though PER/TIM complexes are shown to inhibit per and VP gene expression in vitro). CRYs are rhythmically expressed and the Cry/CRY loop may at the heart of the mammalian clock act as the ‘mammalian Drosophila TIM’, but the loop is blind to light as photic input does not affect Cry expression. The per/PER loop of negative feedback is sensitive to light (via an as yet unknown molecular pathway) and may thus confer photic sensitivity to the Cry/CRY loop. Vasopressin (VP) does not play a key role in the process of endogenous mechanism of the mammalian pacemaker of the clock system as its expression is clock-controlled. However, through the reported effect of the activation of the V1a receptor (also rhythmically expressed, but in reverse phase) on intracellular calcium levels, VP can influence transcription or post-transcriptional processing of clock proteins (dashed arrow lines), thereby influencing or modulating the clock mechanism. (Van Esseveldt et al., 2000; Fig. 10, with permission.) Note that TIM is not a candidate clock gene anymore.

2014 Ch 04

76 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 76

D.F. SWAAB

human genes related to the cloned human PER gene resulted in the recognition of the human gene Kiaa347 (Nagase et al., 1997), which appeared to share 47% identity and 70% homology with human PER. Based on the sequence of this gene, another mouse homologue was identified by RT-PCR on mouse brain cDNA (Albrecht et al., 1997; Shearman et al., 1997). The PCR product showed 81% identity with the human Kiaa347 gene. The complete cDNA was sequenced and the gene, named mPer2, revealed 70% homology and 61% identity in the Pas domain and 47% overall identity with the initially cloned mouse homologue, mper1 (Albrecht et al., 1997). Subsequently, another mouse and human homologue were cloned (Zylka et al., 1998). It was shown that there is a complete loss of free-running rhythmicity in mice that lack both the plant blue-light receptor proteins CRY1 and CRY2 (Van der Horst et al., 1999). The human homologues of these proteins are not known at present. Experiments by Kume et al. (1999) indicate that CRY forms a protein complex with Per and helps to move into the nucleus, where CRY and Per turn off not only the Per genes, but the CRY genes as well. The repression in a feedback loop may be a central mechanism in the clock function of the SCN (Barinaga, 1999). Light pulses cause significant elevations of murine Tim mRNA, which may be the basis for entrainment (Bunney and Bunney, 2000). Polymorphism in the clock gene NPAS2 Leu/Ser was associated with seasonal affective disorder (Johansson et al., 2003). VIP is a member of the family of gastrointestinal hormones, which includes glucagon, secretin, gastric inhibitory peptide, growth hormone-releasing factor, and pituitary adenylate cyclase-activating peptide (PACAP) (Arimura, 1992). VIP is derived from a precursor peptide (prepro-VIP) that consists of 170 amino acid residues. The human prepro-VIP gene contains seven exons, interrupted by six introns that divide the prepro-VIP gene into a signal peptide and five additional functional domains (Obata et al., 1981; Carlquist et al., 1982; Fig. 4.9). Exon 1 encodes the 5-untranslated region, exon 2 the signal sequence (Prepro-VIP 1–21), exon 3 the N-terminal flanking peptide (Prepro-VIP 22–79), and exon 4 the sequence of a peptide that has an NH2-terminal histine and COOH-terminal methionine amide (PHM; PreproVIP 80–110), a bridging peptide (Prepro-VIP 111–122). VIP itself is encoded by exon 5, and the C-terminal flanking peptide is encoded by exon 6 (Prepro-VIP 156–170). Exon 7 consists of the 3-untranslated region of the gene (Itoh et al., 1983). It should be noted here that PHM thus comes from the same prepro-VIP gene on

chromosome 18 as VIP (Christophe, 1993). This peptide is the human analogue of the peptide with N-terminal histidine and C-terminal isoleucine amide (PHI) of the rat (Said and Mutt, 1972; Itoh et al., 1983). The amino acid sequence of rat PHI differs from the human PHM by four amino acids. The homology of the amino acid sequence of VIP itself is 89% between rat and human. VIP neurons in the basal part of the SCN are considered to be involved in entrainment (Moore, 1992). In addition, this peptide has effect on sleep and hormone levels. VIP administered as intravenous boluses causes an increased duration of both REM and non-REM sleep periods, decreased prolactin levels at low dose, and increased prolactin levels at high dose, advanced the cortisol nadir, enhanced the cortisol levels after midnight and blunted the growth hormone peak. VIP thus seems to have a phase-advancing effect on sleep cycles and cortisol secretion, possibly through actions that involve the SCN (Murck et al., 1996). (e) The retinohypothalamic tract (RHT) and other SCN afferents How the brain’s clock gets daily enlightenment. M. Baringa, 2002

The SCN itself generates biological rhythms with a period of approximately 24 hours as appears, e.g. from in vitro experiments with rat tissues in culture (Bos and Mirmiran, 1990) and from the observation that bilaterally enucleated blind human subjects show free-running rhythms of melatonin and cortisol of 24.3 to 24.5 hours (Czeisler et al., 1999; Skene et al., 1999). Because SCN oscillations do not adhere to a strict 24-hour schedule, the circadian pacemaker must be “reset” every day. The endogenous SCN rhythm is therefore synchronized to the environmental light–dark cycle for its period and phase. Bright light resets the human pacemaker (Czeisler et al., 1986). This process is called “entraining”. It is performed by a direct neuronal pathway from the retina to the SCN that also exists in humans, as was first shown by staining degenerating neurons in patients with prior optic nerve damage (Sadun et al., 1984). Initially, high light intensities of some 2500 lux were used to induce lightentrainment and to suppress nocturnal melatonin secretion. More recently, the effects of lower light intensities were demonstrated (Wright et al., 2001). In nature, the seasonal differences are most pronounced during the twilight transitions of dawn and dusk. The human circadian system

2014 Ch 04

22/1/04

12:17 pm

Page 77


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

77

Fig. 4.9. Schematic representation of the human pre-pro VIP mRNA (top line) and genomic DNA (bottom line). The 5-untranslated region (5-UT) and 3-untranslated region (3UT) are indicated by the solid line (top) and open box (bottom). The signal peptide is indicated by a dotted box. The dark boxes indicated the PHM- and VIP-coding regions. Shaded areas represent NH2-terminal, connecting and COOH-terminal peptides. VIP: vasoactive intestinal polypeptide; PHM: peptide histidine methionine. Modified from to Tsukada et al. (1985; Fig 2A.)

is also tuned to gradual increments of low-level illumination during naturalistic dawn, characterized by an accelerating time-rate-of-change that ends before sunrise (Danilenko et al., 2000). However, during the Antarctic winter, rhythms such as that of melatonin and cortisol free run. When the sun reappeared during spring all rhythms again synchronized and entrained to daylight (Kennaway and Dorp, 1991). The retinohypothalamic tract (RHT) is the principal pathway mediating the entraining effects of light on the circadian pacemaker, the SCN. In rat, the RHT was found to originate from a distinct subset of retinal neurons (type III or W cells; Moore et al., 1995). It has now become clear that PACAP cells are the origin of the RHT (see below). There was doubt about both the cones and the rods as far as origin of the RHT was concerned. One study stated that, since red light below the sensitivity of the threshold of a scotopic (i.e. rhodopsin/rod) based system, yet of sufficient strength to activate a photopic (i.e. cone) based system, was sufficient to reset the human circadian pacemaker. Cone pigments, which mediate color vision, also seem to mediate entrainment (Zeitzer et al., 1997). This may be different for the acute effect of light. Another study claimed that melatonin suppression is

4 times stronger at 505 nm than at 550 nm stimulation. Consequently, the responsible ocular photoreceptor would not be the cone system (Brainard et al., 2001). A third study concluded that melatonin suppression must be mediated by a rod-dominated system (Rea et al., 2001). However, rodless and coneless transgenic mice show circadian and pineal responses to light. These responses are driven by a single opsin/vitamin A-based photopigment with peak sensitivity around 479 nm, indicating a non-rod, non-cone photoreceptive system in the mammalian brain (Lucas et al., 2001). A novel shortwave length type of photopigment has subsequently been found to play a primary role in light-induced melatonin suppression, providing evidence for a non-rod, non-cone photoreceptive system in humans (Thapan et al., 2001). From a series of recent animal experimental papers, it has now become clear that those retinal ganglion cells that contain pituitary adenylate cyclase-activating polypeptide (PACAP) and co-store glutamate constitute the RHT. Light activates these cells that directly innervate the SCN via the photopigment melanopsin in the PACAP cells. PACAP interacts with glutamate signaling during the light-induced phase shift. The sensitivity, spectral tuning, and slow kinetics of this light response matched

2014 Ch 04

78 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 78

D.F. SWAAB

those of the photic entrainment mechanism (Gooley et al., 2001; Baringa, 2002; Berson et al., 2002; Hannibal et al., 2002; Hattar et al., 2002). The human primary visual pathway is indeed also regulated according to longterm (15–20 min) light exposure through the action of non-rod, non-cone photoreceptors via a photopigment with the characteristics of opsin: vitamin A (Hankins and Lucas, 2002). Light not only synchronizes the endogenous circadian SCN rhythm to the light–dark cycle, but also influences the SCN and its output instantaneously, as appears, e.g. from its immediate inhibition of melatonin secretion (Chapter 4.5), from the increase in resting heart rate (Scheer et al., 1999), and from the increase of salivary cortisol by early morning light (Scheer and Buijs, 1999). Luteinizing hormone excretion was also increased following light exposure in healthy young men (Yoon et al., 2003). Studies into the chemical chain of events following the RHT input into the SCN have started only relatively recently. By means of a complementary DNA subtraction method, 4 clones were isolated that were induced in the rat SCN specifically by light, including the early response genes c-fos and nur 77 (Morris et al., 1998). Recently the human RHT was studied by a newly developed postmortem tracing procedure using neurobiotin as a tracer. Remarkably, up to 6–8 hours after the death of the patients, the individual neurons are still capable of actively taking up tracer molecules and transporting them over relatively large distances (see Chapter 33b). The RHT appeared to leave the optic chiasm and enter the hypothalamus both medially and laterally of the SCN. The density of the RHT fibers decreases from rostral to caudal (Dai et al., 1998a). The RHT terminates predominantly in a zone of the SCN that contains VIP neurons (Moore, 1992) but do not only contact VIP but also neurotensin cells in the SCN. In addition some vasopressin cells are innervated by the RHT in the ventral part of the SCN. Only few projections to the dorsal part of the SCN and the ventral part of the anterior hypothalamus were found (Dai et al., 1998a; Figs. 4.10 and 4.11). Lateral RHT projections reach the ventral part of the ventromedial SON. These fibers may take part in the diurnal fluctuations of vasopressin release (cf. Nørgaard et al., 1985; Rittig et al., 1989; Forsling et al., 1998). One may hypothesize that this part of the RHT is not yet mature in enuresis nocturna, where such circadian fluctuations are lacking (Chapter 22.4). Lateral RHT projections also innervate the area lateral of the SCN. No

Fig. 4.10. Anterior level of the suprachiasmatic nucleus (SCN) showing the injection spot (asterisk, B) in the optic nerve (case no. 95053). Many labeled fibers (A) can be seen to course along the wall of the third ventricle (3V) and project to the SCN (arrows). Many fibers (B) also extend to the optic tract (arrows). A shows the high magnification of the area in B (arrowheads) and shows more clearly labeled fibers in the optic nerve and ventral part of the SCN. The morphology of labeled fibers is clearly visible. Dashed lines in A and B represent the lateral border of the SCN. Scale bar = 40 m for A, 150 m for B. (From Dai et al., 1998a; Fig. 5.)

projections to other hypothalamic areas were observed (Dai et al., 1998a). This study generally confirmed the observations of Sadun et al. (1984) with paraphenylenediamine that stains remnants of degenerated axons in patients with a lesion of the optic nerve (Sadun et al., 1983) and following DiI staining of the RHT in intact human brains (Friedman et al., 1991). Dai et al. (1998a) could, however, not confirm the existence of innervation of the paraventricular nucleus (PVN) by the human RHT that was described by the degeneration technique (Schaechter and Sadun, 1985). This technique may,

2014 Ch 04

22/1/04

12:17 pm

Page 79


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

however, have been confounded by lesions that were not restricted to the optic nerve. There are many similarities between the chemical anatomy of the SCN and its afferent connections of rat and human, but it is typical of the human SCN, as compared to monkeys and other animals, that it has: (i) a very large population of neurotensin cells; and (ii) a large population of NPY neurons that obscure a geniculohypothalamic tract that contains the same peptide, if the tract is present at all in humans (Moore, 1992). In the human brain, NPY-containing neurons are found throughout the cell groups medial of the dorsal geniculate complex that extends medially into the zona incerta. This area has been designated the pregeniculate nucleus in the primate brain, and is thus probably the homologue of the rodent intergeniculate leaflet. However, since the human SCN has a rather sparce plexus of very fine NPY axons, and itself contains a large number of NPY neurons, it is not clear whether the intergeniculate leaflet neurons indeed project to the human SCN or whether this projection is very much reduced or even absent in human beings (Moore, 1989; Moore and Speh, 1994). It is possible that the geniculohypothalamic tract is present only in nocturnal species like the rat, and absent in diurnal species, like humans (Chevassus-au-Louis and Cooper, 1998). In the rat, direct spinohypothalamic afferents were found in the anterior hypothalamus (Cliffer et al., 1991; Newman et al., 1996) that may well be involved in the motoric and sensoric effects on SCN function, as observed, e.g. in long-term fitness training in elderly people, which improves their circadian rest–activity rhythm (Van Someren et al., 1997) and in the effects of transcutaneous electrical stimulation, which improves rest–activity rhythms in Alzheimer patients (Scherder et al., 1999a). In addition, serotonin (Fig. 4.1) innervates the SCN and histamine, which is the neurotransmitter from the tuberomamillary nucleus (Chapter 13), is necessary for the circadian rhythmicity of ACTH release, food intake, drinking and the sleep–wakefulness cycle. Moreover, histamine can phase-shift circadian rhythms and some authors even consider it to be the final neurotransmitter in the entrainment of the SCN (Eaton et al., 1995; Jacobs et al., 2000; Brown et al., 2001; Tuomisto et al., 2001; Hannibal, 2002). Surprisingly, a response of the biological clock to extraocular light was also monitored after light pulses presented to the popliteal region. The entraining mechanism of extraocular “humoral phototransduction” was

79

Fig. 4.11. Distribution of retinohypothalamic tract (RHT) fibers (blue) with immunocytochemical detection of peptidergic neurons. A: Distribution of RHT fibers in relation to neurotensin (NT) cell bodies (case 95-053). NT cell bodies are distributed over the entire suprachiasmatic nucleus (SCN). Labeled fibers project to the ventral part of the SCN (arrows and arrowheads) or run in the optic tract. B: High magnification of one area of the ventral part of the SCN in A (arrowheads) showing more detailed distribution of labeled fibers in this area; some fibers seem to make contact with NT neurons (arrows). C: High magnification of another area in A. Arrows indicate extensive innervation of labeled fibers in the ventrolateral border of the SCN. D: Combining tracer detection (blue) with immunocytochemical staining for vasopressin (VP; brown; case no. 95-082). VP-positive neurons are located in the ventral part of the supraoptic nucleus (SON), and some labeled fibers leave the optic tract and project to the SON (arrows). OT = optic tract. Scale bars = 50 m in A, 25 m for B–D. (From Dai et al., 1998a; Fig. 6.)

unknown (Campbell and Murphy, 1998). Replication of Campbell and Murphy’s 1998 data was performed by the same authors, applying a 3-hour photic stimulus to the popliteal region. The proportion of rapid eye movement (REM) sleep during the 3-hour light-administration session increased by 31%. These observations seemed to confirm that extraocular light is transduced to the human

2014 Ch 04

22/1/04

80 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

12:17 pm

Page 80

D.F. SWAAB

central nervous system (Murphy and Campbell, 2001). However, since the appearance of these landmark papers, several studies have found that extraocular light does not suppress melatonin secretion in humans (Hebert et al., 1999; Skene et al., 1999; Eastman et al., 2000; Lindblom et al., 2000) and that it does not produce phase shifts. Exposure of the skin to light did not cause any significant changes in bilirubin levels in healthy adult volunteers either (Lindblom et al., 2000). In addition, the response to extraocular light therapy in patients with a seasonal affective disorder did not exceed its placebo effect. Extraocular light did not induce a phaseshift of the circadian pacemaker (Koorengevel et al., 2001). The reason for the discrepancy of the Campbell studies with the negative findings of other authors with extraocular light is most probably that the subjects’ eyes were exposed to low but biologically active light intensities during illumination of the knees, thereby confounding the experiment (Wright and Czeisler, 2002). Finally, also the rest–activity cycle and meals influence the rhythmic endocrine changes as observed during, e.g. daytime fasting, modifications in sleep schedule and psychological and social habits during Ramadan (Bogdan et al., 2001). (f) SCN efferents Immunocytochemical observations show that vasopressin and VIP fibers innervate the SCN itself and a number of other hypothalamic areas, including the contralateral SCN (Swaab et al., 1985, 1994b). Using neurobiotin, an effective and fast anterograde tracer, active, energydependent tracing up to 1–1.5 cm from the injection spot was obtained in human postmortem tissue during an incubation time of 9–12 hours (Dai et al., 1998b,c; Fig. 4.17). The tracing and immunocytochemical observations (Figs. 4.12–4.16) of SCN efferents matched each other very well. The densest projections from the SCN first reach the area between the SCN and the anteroventral part of the paraventricular nucleus (PVN), the anteroventral hypothalamic area (Dai et al., 1998b; Fig. 4.17). Immunocytochemical stainings showed vasopressin and VIP fibers in these areas (Dai et al., 1997; Figs. 4.12 and 4.13). Some of these fibers run anteriorly and enter the anteroventral part of the periventricular nucleus and of the PVN. Rat tracing data indicate that these fibers may reach the paraventricular nucleus of the thalamus and may serve to

synchronize locomotor activity with the light–dark cycle, or influence functions of the hippocampus and amygdala. Many SCN fibers continue in the posterior direction and innervate the zone below the PVN, or they reach the ventral PVN (Dai et al., 1998b; Fig. 4.17). The dense network of vasopressin and VIP-positive fibers in the sub-PVN zone (Dai et al., 1997; Figs. 4.12–4.14) is in agreement with this observation. The SCN fibers in the ventral PVN innervate vasopressin and corticotropinreleasing hormone (CRH) neurons and may thus provide an anatomical basis for the influence of the SCN on hormone secretion (Dai et al., 1998b; Fig. 4.17). This mainly concerns vasopressinergic fibers (Dai et al., 1997). Another extensive projection courses posteriorly and passes close to the third ventricle to reach the dorsomedial nucleus of the hypothalamus (DMN). Most fibers innervating the DMN are concentrated in its ventral part (Dai et al., 1998b) and VIP fibers were more abundant than vasopressin fibers (Dai et al., 1997). Also in the human brain, the DMN projects to the PVN (Dai et al., 1998d; Chapter 10). Injections in the dorsal part of the SCN showed more extensive projections to the PVN than those placed in the ventral part of the SCN. The SCN thus also influences PVN functions in an indirect way via the DMN. In the ventromedial nucleus (VMN) only a few fibers were found, either by tracing or by immunocytochemistry (Dai et al., 1997, 1998b; Figs. 4.15 and 4.16). Electrophysiological and anatomical studies in the rat have revealed a strong projection from the SCN to the supraoptic nucleus (SON) with both inhibitory (GABAergic) and excitatory (glutaminergic) components (Kalsbeek et al., 1993; Cui et al., 1997) that may also be responsible for the circadian rhythmicity in the SON. Such connections have, however, not been shown in the human brain, although SCN fibers come very close to the SON (Dai et al., 1997) and possibly even contact SON dendrites or interneurons. In addition, the lateral retinohypothalamic tract projections that innervate the ventral part of the SON (Dai et al., 1998a) may impose a diurnal rhythm on vasopressin release. Animal experiments have revealed a polysynaptic pathway between the SCN and the pineal gland, involving the autonomic subdivision of the PVN, the intermediolateral cell column in the spinal cord (Vrang et al., 1997; Teclemariam-Mesbah et al., 1999; Chapter 4.5). The SCN may be involved in, e.g. setting the sensitivity of an endocrine organ such as the adrenal cortex by sending vasopressin and VIP fibers to the paraventricular nucleus, which sends, e.g. oxytocin

2014 Ch 04

22/1/04

12:17 pm

Page 81


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

fibers to the intermedio-lateral column of the spinal cord (Buijs et al., 1999; Buijs and Kalsbeek, 2001). Retrograde transneuronal virus tracing showed, e.g. SCN control of the autonomic innervation of the thyroid gland and pancreas in the rat (Kalsbeek et al., 2000a; Buijs and Kalsbeek, 2001; Buijs et al., 2001). The innervation of endocrine organs might be at least as important for their functional fluctuations as the hypothalamic hormonal factors. Although SCN efferents are generally considered to terminate on other neurons (see earlier) and the SCN is not seen as a neuroendocrine structure releasing its products into the bloodstream, there is at least one observation that gives food for thought. Horseradish peroxidase injected intravenously in mice not only readily penetrated the median eminence and arcuate nucleus, but was also transported to the suprachiasmatic area (Youngstrom and Nunez, 1987). These observations suggest that SCN neurons also project to regions that are situated outside the blood–brain barrier, such as the median eminence or the organum vasculosum lamina terminalis (see Chapter 30.5). (g) The cerebrospinal fluid (CSF) as transport medium of the circadian message The circadian activity of the SCN is reflected in diurnal rhythms in the CSF vasopressin levels in various species. This raises the question whether circadian rhythms in brain and behavior are regulated by the SCN via vasopressin or other compounds as a hormonal messenger, and whether the CSF is a transport medium for such messengers to other brain centers. However, in the few patients from whom we could obtain CSF samples over the day and night, we found no clear circadian AVP patterns measurable in human CSF, even though normal circadian patterns of rectal temperature and plasma cortisol were present (Swaab et al., 1987a; Fig. 4.18a,b). This agrees with the absence of a circadian rhythm of CSF-vasopressin levels in patients reported by Sørensen et al. (1985, 1986). Also Burreca et al. failed to find a circadian pattern in vasopressin CSF levels in hydrocephalic patients of different etiologies. These authors did not show, however, that these patients indeed had a circadian rhythm of body temperature or cortisol. It should also be noted that others reported oxytocin levels in CSF to show a time-dependent peak (Amico et al., 1983, 1989), but this peptide is derived from the PVN

81

and not from the SCN. Our observations that indicate that circadian changes in temperature or cortisol are present – even though such an oxytocin peak in CSF was not observed (Swaab et al., 1987a) – are supported, in addition, by animal experiments. Continuously increased CSF levels of vasopressin up to 300 pg/ml were obtained by slow release (Accurel) implants into the rat cerebral ventricles. The CSF levels no longer showed any circadian fluctuation. The endogenous diurnalCSFvasopressin rhythm involving a few pg/ml must thus have been masked, while the circadian sleep/wake patterns of these animals appeared not to be disturbed. The pattern of wakefulness, quiet sleep and REM sleep over the day/night period remained fully intact (Kruisbrink et al., 1987). These observations indicate that circadian CSF-vasopressin levels in the rat are not essential for the transferral of the diurnal message from the SCN to the rest of the brain. Rather, this message will be transferred from the SCN into other brain areas by the SCN efferents (see earlier). Silver et al. (1996) transplanted isolated SCN tissue from hamsters within a semipermeable polymeric capsule, preventing neuronal signals, but allowing diffusion of humoral signals. Since such a transplant could sustain circadian activity, humoral transmission of circadian SCN activity in other species is a possibility. For the likelihood that melatonin uses the CSF as transport medium to other brain areas, see Chapter 4.5a. 4.1. Circadian, seasonal, monthly and circaseptan rhythms and the SCN For everything there is a season, and a time for every purpose under heaven; a time to be born, and a time to die. Ecclesiastes, iii, 1–2

From the moment of conception to the moment we die, biological rhythms play a prominent role in our lives. The endogenous biological rhythms enable the organism to anticipate rhythmic changes in the environment and are consequently important adaptive processes. Consistent with the role of the SCN in the temporal organization of circadian and seasonal processes in mammals, we observed clear circadian and circannual fluctuations in the number of vasopressin-expressing neurons in the human SCN. The possible role of the SCN in the menstrual cycle and weekly rhythms is, however, less clear.

2014 Ch 04

82 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 82

D.F. SWAAB

4.12. A series of line drawings arranged from rostral to caudal (A–V) to illustrate schematically the location of vasoactive intestinal polypeptide (VIP) and vasopressin (VP) cell bodies and fibers in a human hypothalamus (case no. 96-010). The dots correspond to the position and density of the cell bodies. Short lines (in J,L,N) illustrate the area through which the fibers of VP magnocellular cell bodies pass. (From Dai et al., 1997; Fig. 2.)

2014 Ch 04

22/1/04

12:17 pm

Page 83


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 4.12.

Continued.

83

2014 Ch 04

84 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 84

D.F. SWAAB


2014 Ch 04

22/1/04

12:17 pm

Page 85


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911


85

2014 Ch 04

86 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 86

D.F. SWAAB


2014 Ch 04

22/1/04

12:17 pm

Page 87


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

87

Fig. 4.12. Continued. Abbreviations: AVH BST DMH FO INF MB NBM NTL OC POA

anteroventral hypothalamic area bed nucleus of the stria terminalis dorsomedial nucleus of the hypothalamus fornix infundibular nucleus mamillary body nucleus basalis of Meynert lateral tuberal nucleus optic chiasm preoptic area

(a) Circadian rhythms “Early to bed and early to rise, makes a man healthy, wealthy and wise” Benjamin Franklin in: Poor Richard’s Almanac for the year 1757

The ubiquitous functional circadian fluctuations have their basis in circadian changes in the activity of the SCN. This holds for, e.g. such different phenomena as the rhythms in EEG (Aeschback et al., 1999), prolactin, cortisol (Touitou, 1995) and rhythms in pain rating that are highest during the night and lowest during the

PVN PH PEN SCN SDN SON sub-PVN TMN VMN VP VIP

paraventricular nucleus posterior hypothalamic nucleus periventricular nucleus suprachiasmatic nucleus sexually dimorphic nucleus of preoptic area (= INAH-1) supraoptic nucleus area below paraventricular nucleus tuberomamillary nucleus ventromedial nucleus vasopressin vasoactive intestinal polypeptide

afternoon. Totally blind subjects have similar FSH and testosterone diurnal rhythms to a comparison group, indicating that for these hormones the endogenous rhythms are more important than light as a regulatory factor (Bodenheimer et al., 1973). The biological basis of preferences for morning or evening activities are also based upon fundamental properties of the circadian pacemaker. The postawakening rise in cortisol is higher in early awakeners and shows a stronger decline than the levels of late awakeners (Duffy et al., 2001). In night workers, whatever the shift of the melatonin surge, the start of the quiescent period of cortisol

2014 Ch 04

88 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 88

D.F. SWAAB

Fig. 4.13. Illustrating the innervation of VIP (A,C) and vasopressin (VP) (B,D) fibers in the anterior hypothalamic areas. A: Many VIP fibers from the SCN project directly to the central part of the anteroventral hypothalamic area (AVH) and form an extensive plexus just lateral and ventral of the paraventricular nucleus (PVN). Some VIP neurons are visible in the ventral part of the PVN. B: VP fibers from the SCN project to the medial part of the AVH. Many fibers seem to project into the ventral part of the PVN. C and D are the magnification of the area (arrow and arrowhead) in A and B, respectively. 3V, third ventricle. For abbreviations, see Fig. 4.12. Scale bar = 200 m for A,B, 500 m for C,D. (From Dai et al., 1997; Fig. 5, with permission.)

secretion remains phase-locked to the melatonin onset with a similar time lag (1h 25±27 min). Moreover, there is a significant correlation between the timing of the melatonin onset and the timing of the start of the quiescent period. These observations show that both cortisol and melatonin are reliable markers for the assessment of circadian phase in humans (Weibel and Brandenberger,

Fig. 4.14. Distribution of dense VIP and vasopressin (VP) fibers in the sub-PVN. C is a high magnification of a part of the ventral PVN in A: VIP fibers are visible in this area (arrow). Many VIP fibers along the ventral border of the PVN are shown in B, and are presented in more detail in the high magnification D. Some of them seem to pass into the PVN. For abbreviations see Fig. 4.12. Scale bar = 200 m for A,B, 500 m for C,D. (From Dai et al., 1997; Fig. 6, with permission.)

2002). It has been proposed that the circadian temperature rhythm provides a signaling pathway for the circadian modulation of sleep and wakefulness, and that fluctuations in melatonin levels are crucial for the fluctuations in temperature (Van Someren, 1997). Aggressive offences are mainly observed in the evening and at night (Laubichler and Ruby, 1986). There is a circadian rhythm in suicides (Altamura et al., 1999) that varies with age (Preti and Miato, 2001). Both biological and sociorelational factors may contribute to the diurnal variation in suicide by age and gender (Preti

2014 Ch 04

22/1/04

12:17 pm

Page 89


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

89

Fig. 4.15. Distribution of VIP (A,C) and vasopressin (VP) (B,D) fibers in the medial and posterior level of the dorsomedial nucleus of the hypothalamus (DMH) in consecutive sections. VIP fibers are denser than VP fibers. C and D are the magnifications of A and B, respectively. 3V, third ventricle. Asterisk, blood vessel. For abbreviations see Fig. 4.12. Scale bar = 200 m for A,B, 500 m for C,D. (From Dai et al., 1997; Fig. 7, with permission.)

Fig. 4.16. Distribution of VIP (A,C) and vasopressin (VP) (B,D) fibers at the level of the posterior part of the dorsomedial nucleus of the hypothalamus (DMH) in consecutive section. VIP fibers are denser than VP fibers. C and D are the magnifications of A and B, respectively. Arrows in A and C, and arrowheads in B and D point to the same area. 3V, third ventricle. Scale bars = 200 m for A,B, 500 m for C,D. (From Dai et al., 1997; Fig. 8, with permission.)

and Mioto, 2001). Independently of time of the year, at night, train suicide rates in the Netherlands drop to about 10% of the day-time values. There are 2 daily peaks in the pattern, one shortly after sunset and the other consistently 9 to 10 hours earlier (Van Houwelingen and Beersma, 2001). The circadian fluctuations in melatonin levels (see Chapter 4.5) are influenced by seasonal factors (see below). When exposed to long nights, the duration of melatonin and prolactin secretion and the rise in cortisol are longer than the duration of secretion found during short nights. These seasonal differences are generally presumed

to be suppressed nowadays, especially in our modern urban environment (Wehr, 1998). Moreover, hospitalized patients are not exposed to a great deal of light. However, as shown in this and the following section, circadian and circannual rhythms do exist in the human SCN and in many functions. In the pineal gland, the melatonin content is significantly higher during the night than during the day, but only in the long photoperiod (April–September) and in younger individuals (30–60 years) (Luboshitzky et al., 1998). CSF production is at its peak at 2 a.m and at its lowest at 6 p.m. (Nilsson et al., 1992b). Disorders

2014 Ch 04

90 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 90

D.F. SWAAB

Fig. 4.17. A series of schematic drawings arranged from anterior to posterior (A-H) in a representative human hypothalamus (case no. 95-015) to illustrate the distribution of labeled suprachiasmatic nucleus (SCN) fibers, branching fibers with terminal boutons, and corticotropin-releasing hormone (CRH) cell bodies (dark dots in the PVN). Dark dots in the SCN represent the area in which labeled neurons can be detected following in vitro injection of the anterograde tracer neurobiotin. (From Dai et al., 1998b; Fig. 5, with permission.)

2014 Ch 04

22/1/04

12:17 pm

Page 91


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911


91

2014 Ch 04

92 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:17 pm

Page 92

D.F. SWAAB

in the circadian system are described in Chapter 4.1b. Changes during aging in the circadian fluctuations of the vasopressin-expressing neurons are discussed in Chapter 4.3. For the circadian fluctuations in seizures, see Chapter 28.5; Fig. 28.2 (Quigg, 2000).

In postmortem tissue of a group of young subjects (6–47 years of age), we observed a significant fluctuation in the number of vasopressin-expressing neurons over the 24-hour period. During the daytime, the SCN contained twice as many vasopressin-expressing neurons as during the night, with peak values in vasopressin cell number occurring in the early morning. Also the VIP neuron numbers fluctuate significantly over the 24-hour period (Hofman and Swaab, 1993, 1996; Fig. 4.19; M.A. Hofman et al., unpublished results). Since no circadian VIP mRNA fluctuations are seen in the rat under constant dark conditions, VIP neurons are presumed to be involved in entrainment rather than in pacemaker activity (Ban et al., 1997). Circadian rhythms in the human retina (Tuunainen et al., 2001) may be the consequence of circadian rhythms

Fig. 4.18a. Patient E.B., a 24-year-old woman, was admitted because of an intracerebroventricular tumor, situated immediately anterior of the pineal gland. The tumor had caused an obstruction hydrocephalus. She had not been known to suffer from any degenerative neurological condition and, prior to the onset of hydrocephalus, she had functioned normally as a primary school teacher. The tumor, a grade II astrocytoma, was extirpated by way of a left-side fronto-temporal craniotomy under dexamethasone treatment and general anesthesia. A Cordis external ventricular drain was placed in the anterior horn of the right lateral ventricle. The patient was fully conscious and without any neurological deficit from the evening of the day of operation onwards. From postoperative days one up to and including four, fresh CSF samples were obtained through the drain, while the patient remained in the Intensive Care Unit. Records of patient’s neurological status included the Glasgow Coma Scale score (Teasdale et al., 1974). Rectal temperatures were charted 4 times a day and total plasma cortisol levels were determined according to Farmer and Pierce (1974) using a Corning Immo Phase kit, twice on consecutive days in the period during which CSF samples were drawn. There was a well-defined circadian rhythm in temperature and cortisol levels. Short-term perioperative administration of dexamethasone does not appear to cause adrenal insufficiency. Methods: Patient monitoring took place in the Intensive Care Unit. Room lighting was switched off from 23.00 hrs to 06.00 hrs. Approximately 10 ml of fresh CSF was drawn from the Cordis external ventricular drains at 09.00, 12.00, 21.00 and 24.00 hrs for at least three consecutive 24-h periods. All CSF samples were obtained and kept in polyurethane tubes in ice and were immediately centrifuged for 5 min. The supernatant was kept at –20° C until measurement of vasopressin (AVP) and oxytocin (OXT) by means of radioimmunoassay according to Dogterom et al. (1977), except for the extraction before the OXT assay on CSF of patient E.B., which was performed by Seppak C 18 (La Rochelle et al., 1980) instead of Vycor. In the two lower panels of Fig. 4.18 a and b, indicates values below detection levels and high values beyond the scale, which are mentioned separately. (From Swaab et al., 1987a, Fig. 3, with permission.) Note the lack of circadian rhythmicity in the AVP and OXT levels.

2014 Ch 04

22/1/04

12:17 pm

Page 93


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

in the SCN and, via the retinohypothalamic tract, influence the rhythms in this brain structure at the same time (Chapter 4.1e). (b) Circannual rhythms In contrast to the general belief that human beings have few, if any, seasonal rhythms (Lewy and Sack, 1996), we

93

observed strong seasonal fluctuations in the SCN. The number of vasopressin- and VIP-containing neurons in the SCN was found to alter in the course of a year, with August–September values being two times higher than April–May values (Hofman and Swaab, 1992b, 2001; Hofman, 2001; Figs. 4.20 and 4.21). Photoperiod seems to be the major Zeitgeber (pacemaker) for the observed annual variations in the SCN (Hofman et al., 1993). Changes during aging in the seasonal pattern of vasopressin expression are presented in Chapter 4.3. The hypothalamic levels of serotonin and dopamine, neurotransmitters known to innervate the SCN, show diurnal rhythms and seasonal rhythms as well (Carlsson et al., 1980a; Chapter 1.3b; Figs. 1.10 and 1.11). In addition, binding to the serotonin receptor is higher in summer than it is in winter in the hypothalamus of healthy subjects (Neumeister et al., 2000). How these seasonal fluctuations relate causally to the SCN circannual rhythms has not been determined. However, the fact that both aminergic rhythms are observed in the hypothalamus indicates that at least in this respect the SCN drives the monoaminergic systems instead of the other way around. The Japanese observation, that suicide rates depend on the latitude, i.e. on the yearly total amount of sunshine (Terao et al., 2002) may be based upon this mechanism. As shown in jugular blood samples, the seratonin turnover in the brain is lowest in winter and rises rapidly with increased luminosity (Lambert et al., 2002). This clearly indicates that light via the SCN influences the activity of the serotonergic system. In addition, we observed a notable seasonal variation in the volume of the PVN in our material. This volume reached its peak during the spring (Hofman and Swaab, 1992a).

Fig. 4.18b. Patient R.H., a 67-year-old man, presented with a hematoma of the right cerebellar hemisphere following rupture of a small arteriovenous (A–Y) malformation of the right cerebello-pontine angle, as demonstrated after Seldinger angiography of the right vertebral artery. The intracerebellar hemorrhage was exacerbated by the use of oral anticoagulants, prescribed after a myocardial infarction at the age of 62 years. He was not known to suffer from any degenerative neurological condition. Prior to the rupture of the A–V anomaly he had functioned normally. Two days after the hemorrhage, he developed hydrocephalus on the basis of compression of the top of the fourth cerebral ventricle, shown on computed tomography of the posterior cranial fossa, and obtained a Cordis external ventricular drain in the anterior horn of the right lateral ventricle. From days four up to and including nine, post-bleed fresh CSF samples were obtained through this drain. For further details see Fig. 4.18a. (From Swaab et al., 1987a; Fig. 4, with permission.) Note that a rhythm is present in temperature and cortisol but not in AVP and OXT levels.

2014 Ch 04

94 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 94

D.F. SWAAB

Fig. 4.19. Circadian rhythm in the number of vasopressin-containing neurons in the human suprachiasmatic nucleus (SCN) of (A) young subjects (< 50 years of age) and (B) elderly subjects (> 50 years of age). The black bars indicate the night period (22:00–06:00 h). The general trend in the data is enhanced by using a smoothed double-plotted curve and is represented by mean ± S.E.M. values. Note the circadian rhythm in the SCN of young people with low values during the night period and peak values during the early morning. (From Hofman and Swaab, 1994; Fig. 1, with permission.)

Seasonal rhythms are also found in gonadotrophin receptors in the pineal gland, with higher values in the winter than in the summer (Luboshitzky et al., 1997). The human species is more seasonal than we have so far presumed, as appears also from, e.g., annual rhythms in FSH, LH, TSH, total T3, sex hormone-binding globulin, testosterone, plasma cortisol (Kennaway and Royles, 1986; Levine et al., 1994; Maes et al., 1997; Valero-Politi and Fuentes-Arderiu, 1998), melatonin and vasopressin plasma levels, serum osmolality (Levine et al., 1994; Asplund et al., 1998), and rectal temperature levels (Teramoto et al., 1997). The amplitude of circannual rhythms in humans may be considerable. A higher circannual than circadian amplitude was found for DHEAS in

women. In men, the amplitude of the circannual rhythm of T3 was larger than the circadian amplitude. The phasing of the melatonin rhythm changes over the seasons. The time of maximal excretion is significantly delayed in winter by 1 hour and 40 min (Kennaway and Royles, 1986). A short photoperiod appears to suppress ovarian activity, whereas melatonin secretion is increased (Ronkainen et al., 1985; Kauppila et al., 1987). The seasonal variation in testosterone in men can be explained by the variation in LH. Peak levels in both were observed in Denmark during June–July with minimum levels during winter–early spring. Air temperature rather than light exposure seems to explain these fluctuations (Anderson et al., 2003).

2014 Ch 04

22/1/04

12:18 pm

Page 95


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

95

Fig. 4.20. Normalized values of the annual cycles of the AVP- and VIP-expressing neurons in the human suprachiasmatic nucleus (SCN). Crosscorrelation analysis revealed that the two series are positively correlated at lag zero, indicating that the cycles reach their peaks and troughs in the same periods of the year. (From Hofman, 2001; Fig. 2, with permission.)

In contrast, spring is associated with increased pituitaryovarian and androgenic activity (Kaupila et al., 1987). There are also circannual rhythms in opening dates of wars (Schreiber et al., 1991a; Chapter 26.9) and in reproduction (Roenneberg and Aschoff, 1990), that seem to be influenced primarily by environmental light intensity with photo period in a secondary role. Moreover there is seasonality in pregnancies obtained with artificial insemination (Cagnacci and Volpe, 1996), birth weight (Matsuda et al., 1993) and sleep (Honma et al., 1992). Even life span depends on the months of birth. Life expectancy at age 50 appears to depend on factors that arise in utero or early in infancy and those factors increase susceptibility to disease later in life (Dobhammer and Vaupel, 2001). Also the proportion of left-handedness seems to depend on the season of birth. There are more left-handed people born in March to July (Martin and Jones, 1999). There is a preponderance of children who are born in spring and summer and who develop diabetes mellitus type 1 later in life. It has been proposed that this supports the hypothesis that viruses in the fetal or perinatal period may start the autoimmune process that leads to this disorder (Willis et al., 2002). Various psychiatric and neurological disorders show seasonal rhythms. There is a significant relationship between season and schizophrenia incidence (Battle et al., 1999), but the first episode of schizophrenic

psychosis appears to be spared this phenomenon (Strous et al., 2001). There is a high frequency of epilepsy in people born in winter, and a low frequency in September. MS, ALS and possibly Parkinson’s disease are common in those born in spring (Torrey et al., 2000) and there is seasonality in birth rates of women with drug use, indicating possible effects of environmental temperature, hormonal functions or susceptibility to viral infections during pregnancy (Goldberg and Newlin, 2000). Those who are born in the September to November period in the southern hemisphere and those born in March to May in the northern hemisphere have the highest increased frequency of suicidal and depressive symptoms. Second trimester prenatal exposure to influenza is given as an explanation (Joiner et al., 2002). Moreover, there is a season in alcoholism (Modestin et al., 1995), bulimia nervosa (Chapter 23.2), violence, (Morken and Linaker, 2000), seasonal fatigue (Meesters and Lambers, 1990), cerebral infarctions (Gallerani et al., 1993), ischemic attacks, intracerebral hemorrhage, cluster headache (Ferrari et al., 1983), cardiac arrests (Herlitz et al., 2002), childhood optic neuritis following viral infections, with the greatest number presenting in April (McDonald and Barnes, 1992), sudden death (Bilora et al., 1997), and sudden infant death syndrome (SIDS), which is more prevalent in winter (Cornwell et al., 1998). In the USA a significant annual rhythm was found in battering of women, rapes

2014 Ch 04

96 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 96

D.F. SWAAB

Fig. 4.21. Annual rhythm in the number of vasopressin-containing neurons in the human suprachiasmatic nucleus (SCN) of (A) young subjects (< 50 years of age) and (B) elderly subjects (> 50 years of age). The general trend in the data is enhanced by using a smoothed, double-plotted curve and is represented by mean ± S.E.M. Note the circannual rhythm in the SCN of young people with low values during the summer and peak values in the autumn period. (From Hofman and Swaab, 1995b; Fig. 1, with permission.)

and assault, with maximum values in summer (Michael et al., 1983, 1986). There are seasonal fluctuations in depression and suicide as observed in a number of countries (Maes et al., 1993a). A seasonal fluctuation in suicides with a dominant peak around June was found in the northern hemisphere and in December in the southern hemisphere. Various correlations indicate that sunshine may have a triggering effect particularly on violent suicide (Maes et al., 1993b; Petridou et al., 2002; Lambert et al., 2003). However, other months have also been found to show a peak in suicides, which makes such a simple relationship doubtful (Voracek and Fisher, 2002). For instance, in Sweden the monthly distribution of suicides showed a

significant peak in October/November (Brådvik, 2002). In the Netherlands, two daily peaks in train suicides were observed that shifted over the year. One peak occurred shortly after sunset and the other consistently occurred 9–10 hours earlier. Both peaks shifted with the 5.5-hour shift in sunset time (Maes et al., 1993a; Castrogiovanni et al., 1998; Altamura et al., 1999; Van Houwelingen and Beersma, 2001). In England and Wales the seasonal fluctuations in suicides were found to be diminished or even to have vanished in the period 1982–1996, possibly due to a change in lifestyle (Yip et al., 2000). There is a clear annual mortality rhythm that depends on latitude. In developed countries there is a peak of deaths in winter and a trough in summer.

2014 Ch 04

22/1/04

12:18 pm

Page 97


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

The amplitude of seasonality in mortality is greatest in mid-latitude (Israel) and low or absent near the equator (Mexico, Brazil) and subpolar regions (Finland). This cannot be explained by the weather alone: photoperiod seems to have a complex underlying role (Douglas and Rawles, 1999). It has been observed that the desertdwelling hunter-gatherers of the Kung San population of Botswana in the Kalahari desert of South Africa have a seasonal suppression of ovulation in relation to the seasonal changes in nutrition, body weight and activity (Yen, 1993). There is a seasonal clustering of sarcoidosis in spring (Wilsher, 1998). Seasonal variations have also been reported in interferon- production and in the occurrence of optic neuritis and multiple sclerosis (MS), with the highest frequencies in spring and the lowest in winter. These exacerbations of MS may be related to viral infections (Balashov et al., 1998; Jin et al., 2000). The seasonal rhythm in the SCN might be crucial in the development of seasonal depression (Parker and Walter, 1982; Chapter 26.4f) and bulimia nervosa (Blouin et al., 1992; Chapter 23.2), since the symptoms can effectively be influenced by light therapy in these disorders (Kripke, 1985; Rosenthal et al., 1988; Endo, 1993; Lingjaerde et al., 1993; Wirz-Justice et al., 1993; Lam et al., 1994). In a normal population, strong seasonal fluctuations were observed in mood, i.e. in scores for depression, hostilitis, anger, irritability, and anxiety. Females showed stronger seasonal variations than males in depression scores. In Norway, a study showed a sex difference in the binodal circannual rhythm for hospital admissions for depression. For women, the highest peak was in November and for men in April (Morken et al., 2002). Depressed patients with a seasonal pattern improved more through light therapy than patients with a non-seasonal pattern (Thalén et al., 1995), since they seem to be more responsive to external Zeitgebers (Reid and Golding, 2000). Light therapy was also effective in a patient with seasonal fatigue (Meesters and Lambers, 1990). Seasonal patterns also exist in manic episodes; these peak in early spring and have a nadir in the fall. Mixed manic admission has a peak in late summer and a nadir in November (Cassidy and Carroll, 2002). An important component in the circadian and circannual timing system is the pineal gland (see also Chapter 4.5; Penev and Zee, 1997). The nocturnal excretion of the major melatonin metabolite 6-sulfatoxymelatonin in healthy, full-term infants of 8 weeks of age, born in summer, was 3 times higher than that of those born in winter. The seasonal variations were no longer present

97

at 16 weeks of postnatal age, suggesting a prenatal influence of the photoperiod on the ontogeny of melatonin (Sivan et al., 2001). There is a relationship between diurnal and seasonal pineal rhythms, since the diurnal rhythms in pineal melatonin content of autopsy material are evident only in the long photoperiod (i.e. April to September), with melatonin concentrations being 4 times higher at night (22.00–10.00 hours) than during the day (10.00–22.00 hours) (Hofman et al., 1995; Fig. 4.22). This seasonal effect was confirmed by Luboshitzky et al. (1998). In contrast, diurnal variations in the pineal 5-methoxytryptophol contents (Fig. 4.3) are only observed in the short photoperiod (i.e. October–March) with high concentrations during the day and low concentrations at night (Hofman et al., 1995; Fig. 4.23). This shows that the synthesis of indolamines in the human pineal exhibits a diurnal rhythm that is affected by seasonal changes in day length (Hofman et al., 1995). It is generally presumed that the biological signal for photoperiodic changes is a change in the duration of melatonin secretion, reflecting the differences in duration of daylight across the year, with long nights leading to a longer duration of melatonin secretion. Some studies indicate that the seasonal variation in the endogenous circadian rhythm is relatively weak or absent in humans under constant routine conditions (Van Dongen et al., 1998). Yet there was no significant difference between blind and sighted individuals as far as the duration of melatonin secretion was concerned (Klerman et al., 2001a). The way the seasonal and circadian rhythms of the pineal gland and SCN influence each other is thus a matter for further research. (c) Monthly cycle The moon has been associated with mental disorder since antiquity, as still reflected by the word ‘lunacy’ for insanity (Luna is the Roman goddess of the moon). Old traditional Chinese medicine has also made the link between human physiological rhythms and natural rhythms. Although recent studies did not find a lunar phase effect on psychiatric hospital admission, suicide or homicide, absentism from work, traffic accident or trauma, monthly rhythms may well have been present prior to the advent of modern lighting (Raison et al., 1999). A synchronous relationship between the menstrual cycle and the lunar rhythm has been suggested by the large proportion of menstruations (28.3%) that: (i) occur around the new moon, and (ii) the zenith of 6-hydroxymelatonin levels in urine in the period of the new moon (Law, 1986).

2014 Ch 04

98 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 98

D.F. SWAAB

Fig. 4.22. The effect of photoperiod on melatonin content in the human pineal during the day (10.00–22.00 h) and night (22.00–10.00 h). Values are means ± S.E.M. Only during long photoperiods (April–September) were day- and night-time melatonin levels different (*P ≤ 0.05). (From Hofman et al., 1995; Fig. 1, with permission.)

The reproductive cycle is controlled by a cascade of signals under the direction of the pulsatile hypothalamic secretion of LHRH (Chabbert-Buffet et al., 1998). Some data indicate that the menstrual cycle is derived from the ovary (Quigley et al., 2002). However, the observation that melatonin is elevated at the time of menstrual bleeding and has its nadir at the time of ovulation (Wetterberg et al., 1976) indicates the involvement of the SCN-pineal axis in

Fig. 4.23. The effect of photoperiod on 5-methoxytryptophol (5-ml) content in the human pineal during the day (10.00–22.00 h) and night (22.00–10.00 h). Only during short photoperiods (October–March) were day- and nighttime 5-ml levels different. For further details see legend Fig. 4.22. (Reproduced from Hofman et al.., 1995; Fig. 2, with permission.)

the menstrual cycle. In the rat, SCN lesions result in a persistent estrus (Wiegand and Terasawa, 1982) and there is evidence for a monosynaptic VIP-containing pathway between the SCN and the LHRH system (Van der Beek et al., 1997). It has been presumed that a fundamental deterioration of the SCN or of the coupling to its outputs may initiate the gradual disintegration of the temporal organization of neurotransmitter rhythms that are the basis for menopause (Wise et al., 1996). Bilateral ablation of the SCN by a hypothalamic astrocytoma did indeed result in amenorrhea in one patient (Haugh and Markesbery, 1983). However, in this patient not only the SCN area, but also a large part of the remaining part of the hypothalamus was affected so that, e.g. LHRH neurons or other systems may also have been lesioned. There appeared to be no obvious difference between pre- and postmenopausal women as far as vasopressinexpressing neuron numbers in the SCN were concerned (Swaab et al., 1985), which is an argument against strong fluctuations in the activity of this subpopulation of neurons in relation to hormonal levels. However, the number of VIP-expressing neurons in the SCN of postmenopausal women is increased (J.N. Zhou, unpublished observations). Moreover, a change that might be related to the menopause is that circadian fluctuations in the number of neurons expressing vasopressin disintegrate after the age of 50, not only in males but also in females (Hofman and Swaab, 1994; Fig. 4.19). The observation that the irregularity of the menstrual cycle in stewardesses flying transmeridianly is increased as compared to controls (Preston et al., 1973) also supports a link between SCN function and the menstrual cycle. On the other hand, the menstrual cycle in one woman under social and temporal isolation did not seem to be linked to the sleep–wakefulness rhythm. In fact, in two experiments the menstrual cycle length of the same subject stayed normal, i.e. exactly 28 calender days, whereas her free-running sleep–wakefulness rhythm and rectal temperature free-running rhythm cycle length increased drastically (Chandrashekaran, 1994). These experiments still await confirmation. An interesting relationship between circadian rhythms and the menstrual cycle has been reported in a woman with premenstrual syndrome. She showed phase shifts in sleep rhythm in the menstrual cycle: progressive phase advances were found in the follicular phase and phase delays in the luteal phase. Rectal temperature also showed similar menstrual changes, but the phase advance and delay started a few days earlier than the changes in sleep–wake rhythm. It

2014 Ch 04

22/1/04

12:18 pm

Page 99


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

was suggested that the circadian rhythms in sleep and temperature were under the control of ovarian steroid hormones, the way they are in hamsters, where estrogens advance the phase of activity rhythm, whereas progesterone blocks the effect of estrogens. However, it was impossible to tell whether the circadian clock dysfunction was the cause or result of the premenstrual syndrome symptoms (Shinohara et al., 2000). Clearly more work has to be done on the possible involvement of the SCN in the menstrual cycle in relation to the possible involvement of the SCN, in both seasonal and menstrual rhythms. (d) Circaseptan rhythms The presence of circaseptan biological rhythms is still controversial and the involvement of the SCN in such rhythms is far from clear. The astronomical counterpart of this weekly rhythm may be the 6.6-day periodicity in solar radiation, resulting in a 7-day cycle in the weather (Levi and Halberg, 1982). It is possible that these circaseptan phenomena resulted in an endogenous biological rhythm that might even have led to the basic feeling of the presence of a 7-day cycle, only later expressed in the Bible and reinforced by Genesis, ii, 2 “. . . and on the seventh day God ended his work which he had made; and he rested on the seventh day from all his work which he had made.” A weekly component in physiological functions may thus have preceded the societal 7-day week. A strong argument in favor of the possible presence of such an endogenous biological weekly rhythm comes from the observation that a 7-day rhythm is present in the development of enamel on teeth. These indications of a weekly rhythm are even found in fossil hominids (Bromage and Dean, 1985) and were thus overt long before the Bible was written. Circaseptan bioperiodicities have been reported in a large number of biomedical parameters in blood and urine in the rat and humans, in myocardial infarctions and strokes (Levi and Halberg, 1982; Halberg, 1995; Pel and Heres, 1995; Swaab et al., 1996) and in the weekly distribution of suicides. The highest frequency of suicides was found to be on Mondays. Israel was said to be an exception, with the highest numbers on Sunday (Altamura et al., 1999), suggesting a social influence. However, in Sweden, too, a preponderance of suicides was later found to occur on Sundays (Brådvik, 2002). Spontaneous deliveries in humans also show a circaseptan rhythm with a trough in the weekends, the latter coinciding with higher perinatal mortality (Pel and Heres, 1995) and coronary heart disease

99

death, which occur preferentially on Mondays (Evans et al., 2000a). Moreover, it was found that subjects with strong seasonal fluctuations in mood also had stronger weekly mood cycles (Reid et al., 2000). There is also a weekly rhythm in cardiac arrests, with an increased incidence on Monday (Herlitz et al., 2002). Although most of these phenomena might be based upon masking by the calendar week, the occurrence of masking factors such as weekends and holidays does, of course, not a priori exclude the possibility of the presence of circaseptan rhythms. Masking is very unlikely in the following phenomena. The first concerns an investigator who collected his own urine daily for a period of 15 years and showed a clear circaseptan rhythm in 17-ketosteroids. During a period of 3 years he observed a desynchronization with the weekly social schedule, which suggested the presence of a free-running rhythm. A second example of a circaseptan rhythm is oviposition, in an arthropod, the insect Folsomia candida, which shows circaseptan rhythms and no within-group synchronization, also in constant darkness (Levi and Halberg, 1982). Beach beetles (Chaerodes trachyscelides White) in constant darkness show a strictly nocturnal activity period with apparently circaseptan components superimposed, which may act as adaptations to the weekly alterations between spring and neap tides. Actual 7-day periodicities have been found in a few other marine organisms (Meyer-Rochow and Brown, 1998). In addition, the presence of a circaseptan free-running rhythm was confirmed in human beings who were in isolation in a cave for over 100 days (Halberg, 1995). The fifth example that cannot be explained by masking by a calendar week is the 7-day rhythm that was observed in dentine development, from which the signs can even be seen in fossil hominids as mentioned above (Bromage and Dean, 1985), long before there was a social week. It has been hypothesized that circaseptan rhythmicities in humans are related to tidal-foraging in the history of mankind (Meyer-Rochow and Brown, 1998), but there is no evidence supporting this assumption. What role the SCN might play in a circaseptan rhythm should be investigated. 4.2. SCN development, birth and circadian rhythms (a) Circadian rhythms in the fetus and at birth The precise timing of labor is paramount to the survival of the neonate and the species. For day-active mammals,

2014 Ch 04

100 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 100

D.F. SWAAB

including human beings, the normal time of delivery is during the hours of rest, i.e. during darkness. Labor that starts at night is also shortest in duration (Ducsay, 1996) and is therefore the most efficient. Melatonin receptors (MT1 and MT2) are hypothesized to be involved in the circadian activity of the uterus (Schlabritz-Loutsevitch et al., 2003). Some circadian rhythms are present in pregnant women with a similar pattern to that in nonpregnant women, e.g. the rhythms for urinary sodium excretion, plasma cortisol and urinary prostaglandin excretion, but other circadian patterns are different as compared to nonpregnant women, e.g. for arterial pressure, for the levels of atrial natriuretic factor, aldosterone and progesterone, potassium excretion and creatinine clearance (Van der Post et al., 1997a; Lunshof et al., 1998). There are diurnal changes in maternal plasma in estrone, estradiol and estriol at 34–39 weeks of gestation. Estrogens are formed in the placenta from precursors of fetal and maternal origin. The estriol rhythm had an inverse pattern as compared to the cortisol rhythm. The estrone rhythm lags about 2 to 3 hours behind the plasma cortisol rhythm. These observations are consistent with the idea that the rhythms may reflect circadian fluctuations in the secretion of precursor steroids from the maternal adrenal (Challis et al., 1980). In both rhesus monkeys and women, the myometrium is more responsive at night (Honnebier et al., 1989b; unpublished observation; Ducsay, 1996). The rhythm in uterine activity was found to be present from 30 weeks gestation to term (SérónFerré et al., 1993) and a diurnal rhythm was observed in maternal blood for oxytocin (Lindow et al., 1996). A diurnal rhythm that is influenced by parity and season is found when human labor is in progress. A peak frequency of labor onset at 8–9 a.m. was observed in nulliparous women at 8–9 a.m. and a peak delivery at 2 p.m. (Cagnacci et al., 1998a), but a different circadian pattern of low-risk birth has also been reported (Anderka et al., 2000). One may raise the question whether the fetal or maternal SCN determines the circadian rhythm in delivery. A 24-hour rhythm of fetal adrenal cortisol is present and the data suggest a fetal circadian pacemaker (Serón-Ferré et al., 2001). However, various diurnal rhythms of the fetus disappear immediately after birth and re-emerge later in the neonate and continue to develop, postnatally, over a period of several weeks to 3 months, which suggests that fetal rhythms are predominantly driven by the mother (Honnebier et al., 1989a; SérónFerré et al., 1993). Moreover, a loss of circadian rhythms

was found in plasma cortisol, ACTH, 17 estradiol, estriol, and fetal heart rate following administration of the corticosteroid triamcinolone to 5 healthy pregnant women at 35 weeks of gestation, supporting the idea that the maternal adrenal gland entrains fetal rhythms. The possibility that fetal rhythms are driven by the mother is reinforced by the observation that postnatal development of various overt rhythms, for example in N-acetyltransferase and sleep/wakefulness patterns, is paralleled by the maturation of the SCN, as evidenced by the strong increase in the number of vasopressin-expressing neurons in this nucleus (Swaab et al., 1990, see below). On the other hand, the fetal SCN itself already shows metabolic circadian changes in the squirrel monkey, and melatonin receptors manifest themselves in the human SCN area as early as the 18th week of gestation (Reppert, 1992). Moreover, temperature rhythms are reported to be already present in some 50% of premature human babies (Mirmiran et al., 1990; Mirmiran and Kok, 1991), although later research indicates that such rhythms may, at least partly, be due to masking effects. Although circadian rhythms of insulated skin temperature and heartrate were sometimes present in premature babies of 24–29 weeks, they were not found very often (Tenreire et al., 1991). Another argument in favor of the idea that the mother may drive circadian rhythms is the observation by Lunshof et al. (1998b) of a circadian rhythm in fetal heart rate variability in only 46%–38% of the fetuses of 26–28 weeks of gestation. However, in 73% of these fetuses she observed a circadian rhythm in basal heart rate, while longitudinal evaluation of salivary cortisol levels in neonates revealed a distinct endogenous rhythm only in 2 out of the 10 healthy full-term neonates and in 3 out of the 10 healthy preterm neonates, with a periodicity of 12–30 hours (Bettendorf et al., 1998). In a study on the development of circadian rhythms in human infants, temperature rhythms became significant 1 week after birth, and the wake rhythm on day 45, at the same time as increased melatonin secretion began to occur at sunset. The sleep circadian rhythm appeared last, after day 56 (McGraw et al., 1999). This indicates that, while the fetal SCN neurons may already have a circadian rhythm, the SCN efferents or the brain areas where they terminate do not mature until much later, after birth. We can, however, at present not exclude that, while most fetal rhythms are driven by the mother, some overt circadian rhythms (for example temperature rhythms) may be present as early as the premature period. A preliminary observation in this connection came from a study on a

2014 Ch 04

22/1/04

12:18 pm

Page 101


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

discordant anencephalic twin that showed that the fetal brain may contribute to the generation of diurnal rhythm in fetal heart rate and synchronization of maternal–fetal brains. In normal twins the diurnal pattern in fetal heart rate was found to follow the maternal heart rate by 1–2 hours. However, neither the anencephalic fetus nor its normal twin had a diurnal rhythm, which was proposed by the authors to be due to tactile influences of the anencephalic fetus. The anencephalic fetus may, by its forceful jerky movements, have disturbed the diurnal rhythm of the normal twin (Lunshof et al., 1997). It has not yet been established which mechanism of the mother drives the fetal circadian rhythms. It has been suggested that maternal cortisol or melatonin may act as Zeitgeber for fetal diurnal rhythms, although this has not been proven (Lunshof et al., 1998, 2000b). Fetal heart rate, fetal movement, cortisol, estriol, and 17-estradiol rhythms were absent in a patient following total adrenalectomy and radiation of the sella turcica for Cushing’s syndrome. The patient was on a regimen of 50 mg of cortisone-acetate daily at 8 a.m. and 8 p.m. This case history suggested that maternal cortisol rhythms may modulate fetal behavior (Arduini et al., 1987). On the other hand, Lunshof et al. (2000b) found that the correlation between maternal and fetal heart rate rhythms did not change after betamethasone administration, which does not support the presumed crucial role of maternal corticosteroids for the generation of fetal rhythms. Postnatal transfer of melatonin via the milk has also been proposed as a possible maternal entraining signal. Although the evidence is only circumstantial, a clear rhythm of melatonin with peak levels at night has been observed in human milk (Illnerova et al., 1993). (b) The SCN during development The observations on the presence of some overt rhythms in the fetus mentioned above raise the possibility of an involvement of the fetal SCN in early circadian rhythms, induced by a cell type that is already mature well before birth. May et al. (1998) found neurophysin-positive neurons in the SCN from 18 weeks of gestation onwards. However, vasopressin neurons do not seem to mature early; at birth the SCN contains only some 13% of the number of vasopressin-expressing neurons found in adulthood (Fig. 4.24a). There appears to be a relationship between the increasing number of SCN neurons expressing vasopressin in the first postnatal months and the occurrence of the circadian activity of N-acetyltrans-

101

ferase in the pineal gland at around 3 months of postnatal life and sleep–wakefulness rhythms that begin to emerge around 2 months postnatally (Swaab et al., 1990; Antonini et al., 2000). Measurements of a urinary metabolite of melatonin showed a gradual appearance of a diurnal rhythm between 8 and 12 weeks postnatally (Sérón-Ferré et al., 1993), and in a recent study day–night rhythms in the melatonin metabolite were found from 27–41 days postnatally (Ardura et al., 2003). Cortisol circadian rhythms emerge mostly between 2 and 3 months postnatally (Antonini et al., 2000). The vasopressin cell numbers rise to maximum values around 1–2 years postnatally, after which they gradually decrease to some 50% of these numbers in adulthood (Swaab et al., 1990). Experiments in rats have shown that VIP neurons in the SCN develop well before the vasopressin neurons do (Laemle, 1988). In order to assess the course of maturation of the VIP neurons in the human SCN, the number of VIP-expressing neurons was determined by immunocytochemistry and morphometry in 43 subjects ranging in age from mid-gestation up to 30 years (Swaab et al., 1994b; Fig. 4.24b). Both VIP and vasopressin neurons were first observed at 31 weeks of gestation in the ventrolateral part of the SCN. From the postnatal age of 3 months onwards, VIP-positive neurons were found in some subjects in the centromedial part of the SCN, but a majority of the individuals did not yet show VIP positive neurons. The centromedial VIP staining became a constant finding only from about 20 years of age. Postnatally, the number of VIP neurons increased gradually until adult values were reached around the age of 3 years (Swaab et al., 1994b; Fig. 4.24b). After the age of 10, a clear sex difference was found in the number of VIP neurons, with the male SCNs displaying, on average, twice as many VIP neurons as those of the female. However, after the age of 40 this sex difference reverses (Fig. 4.25; Zhou et al., 1995b). In adults the number of VIP cells in the SCN is clearly lower than the number of those containing vasopressin (Fig. 4.24). The ratio of VIP to vasopressin-expressing neurons varies between 12% in middle-aged men and 40–65% in older women. With respect to the sex differences in age-related changes in VIP neurons, it is of interest to mention that there are differences between healthy elderly women and men as far as entrained circadian temperature rhythms are concerned that suggest that aging may affect the circadian timing system in a sexually dimorphic way. The acrophase of body temperature was phase-advanced by an average

2014 Ch 04

102 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 102

D.F. SWAAB

Fig. 4.25. Lifespan changes in the number of vasoactive intestinal polypeptide (VIP)-immunoreactive neurons of the human SCN in control subjects. The blank bar indicates the males and the hatched bar indicates the females. The SCN of young males (10 to 40 years) contains twice as many neurons as that of young females (**p < 0.02). This sex difference reverses in middle-aged subjects (*p < 0.04). Note that the decrease in the number of VIP cells started already in middleaged males and the significant reduction in the elderly males compared with young males (#p < 0.02). (From Zhou et al., 1995b; Fig. 2, with permisson.)

Fig. 4.24. a and b: development of the human suprachiasmatic nucleus (SCN) of the hypothalamus. Log-log scale. The period at term (38–42 weeks of gestation) is indicated by the vertical bar. (a) Note that vasopressin (AVP)-expressing cell number is low at the moment of birth (21% of the cell number found in adulthood). There is no difference in the developmental course of the SCN in boys and girls. Cell numbers around 1 to 1.5 years postnatally are more than twice the amount of adult cell numbers. After these high levels a decrease to adult vasopressin cell number is found (From Swaab et al., 1990, Fig. 2, with permission). (b) Until the end of term the VIP-expressing cell numbers are low, whereas the majority of subjects do not show any VIP expression at all. After term there is a gradual increase in VIP neuron numbers and after the age of 10 years the values for males are distinctly higher than those for females. (From Swaab et al., 1994b; Fig, 3, with permission.)

of 1.25 hours in older women compared to age-matched men. Women woke up earlier and slept for shorter periods of time (Moe et al., 1991). In addition, the sex differences in the human SCN reinforce the ideas on the possible involvement of this nucleus in sexual behavior or reproduction (see Chapter 4.4). Our data on the ontogeny of the SCN did not point to a particular role for VIP neurons in those rhythms that may already be present in early fetal development, for example, the temperature rhythm observed in some prematures around 30 weeks gestational age. However, recently, by improving the sensitivity of the VIP staining by microwave treatment of the sections, it was found that VIP was already present from 24 weeks of gestation. The most surprising finding was, moreover, that neurotensin neurons were present in the SCN from the youngest stages studied onwards, i.e. 20 weeks of gestation. This means that neurotensin neurons are at present the best candidates for generating rhythms in the fetal SCN at an early stage (Xu et al., 2003). The developmental pattern of the vasopressin neurons seems to be related to the postnatal pattern of the sleep and wakefulness rhythm development. Around the 7th week of postnatal development, circadian

2014 Ch 04

22/1/04

12:18 pm

Page 103


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

components emerge in this pattern, while after 12 weeks of postnatal development a prominent circadian positive peak is found and a negative peak at 12 hours periodicity, which represents a very scarce amount of daytime sleep (Fukuda and Ishihara, 1997). The differences between men and women and earlier (animal) studies suggest an additional role for VIP neurons in the SCN in sexually dimorphic functions such as reproduction and sexual behavior (Swaab et al., 1994b; Zhou et al., 1995b; see Section 4.4). Entrainment of circadian rhythms during pregnancy may have important physiological consequences. The optic chiasm is complete and the optic tract traceable from the optic chiasm over the lateral aspect of the diencephalon in a dorsal direction by as early as 7 weeks of gestation (Cooper, 1945). Although myelin sheets are already present for the first time around the fibers of the optic nerve at 32 weeks of gestation, significant increases are still seen in myelination during the first 2 postnatal years (Magoon and Robb, 1981). The retinohypothalamic tract is visible from 23 weeks of gestation (Koutcherov et al., 2002). The biological clock of very premature baboon infants is already responsive to light. Moreover, since there is no difference between the circadian sleepwake rhythms of preterm infants that were entrained during a similar time of exposure to an environment with daily time cues and those of term infants, it seems to be the period of exposure to environmental time cues rather than the neurological maturity that determines the entrainment of the circadian rhythm of sleep and wakefulness in the human infant (McMillen et al., 1991). Preterm birth and a subsequent stay in hospital have no influence on the development of a circadian sleeping pattern. Parental care-giving behavior and nursing are major determinants of time-of-night sleeping or “settling” of infants (Lunshof et al., 2000) and the mother–infant synchronization is probably the first factor in the entrainment of the infant’s circadian sleep–wake rhythm (Nishihara et al., 2002). In full-term infants, a similar conclusion was reached (Nishihara et al., 2000). Interestingly, the development of premature children exposed to a nursery environment with diurnal cycles is better. They gain weight faster, can be fed orally sooner, spend fewer days on the ventilator and on phototherapy and display enhanced motor coordination. They develop sleep–wake cycles sooner after discharge than infants cared for in constant lighting (Mann et al., 1986; Fajardo et al., 1990; Miller et al., 1995; Brandon et al., 2002), which is an indication of the important role of the circadian system

103

in the neurological development of the child (for review see Mirmiran et al., 1992a). This is in contrast to the assumption that a dark environment would mimic the in utero condition and be beneficial for the child. Reducing light conditions in the neonatal care unit does not improve medical outcomes in very low birth weight infants (Kennedy et al., 2001). 4.3. Circadian and circannual rhythms in aging and Alzheimer’s disease Ageing seems to be the only available way to live a long life. Daniel Francois, Esprit Auber

(a) Disruption of rhythms in aging and Alzheimer’s disease From temporal isolation experiments it appeared that there is a negative relationship between the period of the clock and the age of the individual. In addition, 80% of subjects in the 50- to 80-year range show a spontaneous internal desynchronization of rhythms that may affect sleep patterns and other aspects of biological aging (Weitzmann et al., 1982; Mirmiran et al., 1992b). Habitual bedtime and wake time are earlier in people in their forties and fifties than in young subjects. In addition, the middleaged have a greater orientation toward morningness and have an earlier phase of temperature rhythm (Carrier et al., 2002). Moreover, older subjects are sleeping and waking earlier relative to their nightly melatonin secretory episode. Consequently, older subjects wake up at a time when their relative melatonin levels are higher (Duffy et al., 2002). However, it was found that the intrinsic circadian period () in totally blind people of 40 to 50 years of age appears to lengthen slightly but significantly. Age-related shortening of thus does not seem to be the explanation for early morning awakenings in older people (Kendall et al., 2001). In a study measuring sleep and circadian rhythm of activity under natural environmental conditions, weakened and fragmented circadian sleep and rest–activity rhythms were found during aging, while no gender-related difference was found. A strong decline in “actual sleep time” and “sleep efficiency”, as well as increased “sleep latency” was observed in the old and oldest volunteers (Huang et al., 2002). Changes in circadian rhythms are frequently associated with a reduction in night-time sleep quality, a decrease in daytime alertness, and an attenuation in cognitive performance (Myers and Badia, 1995). In the elderly,

2014 Ch 04

104 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 104

D.F. SWAAB

a disordered sleep architecture is found with reduced time spent in slow wave sleep (SWS) and in REM sleep. Aging is frequently associated with complaints about earlier bed and wake times. These changes in sleep timing are associated with an earlier timing of multiple endogenous rhythms, including core body temperatures and plasma melatonin. There is, however, a maintained responsiveness of the circadian pacemaker to light, which implies that scheduled bright light exposure can be used to treat circadian phase disturbances in older people (Klerman et al., 2001b). In fact, elderly persons, especially the insomniacs, are exposed to significantly less environmental light. Supplementary exposure to midday bright light significantly increased melatonin secretion to levels similar to those in young adults (Mishima et al., 2001). The observation that elderly reporting visual impairment are also likely to report sleep complaints (Zizi et al., 2002) also indicates the importance of light input. Also other input mechanisms to the SCN seem to be intact and can be used to improve sleep in the elderly. An increased level of physical activity improves circadian rhythmicity in healthy elderly people, as was found following a 3-month fitness training period (Van Someren et al., 1997b). Melatonin treatment for elderly insomniacs improved sleep (Zhdanova et al., 2001). Age-related changes have been found in, e.g., rhythmic levels of cortisol, vasopressin, blood pressure, pulsatile LH, testosterone secretion, -endorphine levels and many other endocrine circadian rhythms in humans (Tenover et al., 1988; Asplund and Åberg, 1991; Touitou, 1995; Magri et al., 1997; Forsling et al., 1998), but the entrained body temperature rhythms appear to be only slightly affected (Monk et al., 1995; Touitou et al., 1997; Ferrari et al., 2001). Not only the testos-terone levels decrease in aging men, a process that is accompanied by reduced virility and libido (see Chapter 24), but also the circadian testosterone rhythm was markedly attenuated or absent in healthy elderly men (Bremner et al., 1983; Schill, 2001). The amplitude of the DHEAS rhythm (Guagnano et al., 2001) was also affected. A placebo-controlled, double-blind randomized study showed that intranasal vasopressin treatment (20 IU before bed time and after awakening) increased the total sleep time, time spent in SWS and duration of REM sleep, althoughthe scores of subjective sleep quality did not change. Although this effect was explained as a compensation for an age-related decrease in vasopressin content of the SCN (Perras et al., 1999a; see

below), a central mechanism of action of the beneficial effect of vasopressin on sleep in the elderly still has to be proved. The fragmented sleep–wake pattern which occurs in senescence is even more pronounced in Alzheimer’s disease (Witting et al., 1990; Mirmiran et al., 1992b; Prinz and Vitiello, 1993; Bliwise et al., 1995; AncoliIsrael, 1997). Continuous measurement of the circadian rest–activity cycle for 589 days in a demented patient with probable Alzheimer’s disease revealed slow progressive changes in temporal organization until death. There was a gradual insertion of wakefulness into rest, and of rest into wakefulness. Pacing acted as a “side-effect”, a non-photic Zeitgeber, improving the synchronization of the rest–activity cycle (Werth et al., 2002). In Alzheimer’s disease, disruptions of the circadian rhythms are often so severe that they are even thought to contribute to mental decline (Moe et al., 1995). The circadian fluctuations of salivary cortisol are also less marked in Alzheimer patients than in controls (Giubilei et al., 2001). In Alzheimer patients with disturbed sleep–wake rhythms there is a higher degree of irregularities in melatonin secretion (Mishima et al., 1999). An impairment of melatonin secretion is present that is related to both age and severity of mental impairment. The nocturnal growth hormone secretion and both the mean levels and nadir values of plasma cortisol are also related to mental impairments (Magri et al., 1997). Demented patients frequently (about 45%) suffer from sundowning, characterized by an exacerbation of symptoms indicating increased arousal and agitation in the late afternoon, evening or night (Cardinali et al., 2002). Sundowning is considered to be a chronobiological disturbance (Lebert et al., 1996) related to a phase delay of body temperature caused by Alzheimer’s disease (Volicer et al., 2001). Disruption of the sleep of the caregiver due to nocturnal problems of the patient is a more important reason for placement of a dementing patient in a nursing home than cognitive impairment (Pollak and Perlick, 1991). Mishima et al. (1997) claim that Alzheimer patients had an intact body temperature rhythm, in agreement with Touitou et al. (1997) and an earlier study of Prinz et al. (1984), while multi-infarct dementia patients had a low amplitude and disorganized pattern of body temperature. This was in marked contrast to the disturbed rest–activity rhythm in both groups. Rest–activity rhythm disturbances and temperature rhythm disturbances may thus have a different pathological basis, but it should be noted that

2014 Ch 04

22/1/04

12:18 pm

Page 105


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

no neuropathological confirmation of the different types of dementia was available in that study. A general point for the studies claiming that a greater disruption of circadian rhythms is present in multi-infarct dementia or dementia accompanied by subcortical, hypodense white matter regions (see, e.g. Bliwise et al., 2002), is that: (i) circadian rhythms are generally not systematically studied, (ii) neuropathology of the diagnosis is lacking, and (iii) the SCN is not studied either. In addition, a substantial proportion of both nursing home residents with night time incontinence and frail geriatric patients experience a reversal of the normal diurnal pattern of urine excretion (Ouslander et al., 1998). The circadian rhythm in blood pressure is preserved in the early stages of Alzheimer’s disease, but is disrupted in advanced or institutionalized patients (Cugini et al., 1999). (b) Peptide changes in the SCN in aging and Alzheimer’s disease The disruption of circadian and circannual rhythms and the increased incidence of disturbed sleep during aging (Van Someren, 2000a) are paralleled by age-related alterations in the circadian timing system, a decreased input to the SCN, and in Alzheimer’s disease also with the presence of pretangles (Swaab et al., 1992; Van de Nes et al., 1994, 1998) and tangles (Stopa et al., 1999) in the SCN. Diffuse amyloid plaques are only seldom noted in this nucleus (Van de Nes et al., 1998; Stopa et al., 1999). The circadian and circannual fluctuations in vasopressin-expressing neuron numbers in the SCN decrease during aging. The marked diurnal oscillation in the number of vasopressin-expressing neurons in the SCN of young subjects, i.e. low vasopressin neuron numbers during the night and peak values during the early morning, disappears in subjects over the age of 50 (Hofman and Swaab, 1994; Fig. 4.19). Whereas in young subjects low vasopressin and VIP neuron numbers are found during the spring and summer, and peak values in autumn and winter, the SCN of people over 50 years of age showed a disruption of the annual cycle with a reduced amplitude (Hofman and Swaab, 1995; Figs. 4.20 and 4.21). A marked decrease was found in the number of vasopressin-expressing neurons in the SCN only in subjects of 80 to 100 years of age, while in Alzheimer’s disease these changes occurred earlier and were more dramatic (Swaab et al., 1985, 1987a; Fig. 4.26). Stopa

105

et al. (1999) found a decrease in the density of vasopressin and neurotensin neurons as well as a corresponding increase in the GFAP stained astrocytes in Alzheimer patients. The immunocytochemical data indicating decreased activity of the SCN in Alzheimer’s disease have been confirmed by in situ hybridization. The total amount of vasopressin-mRNA was 3 times lower in Alzheimer patients than in age and sex-matched controls. In addition, the SCN vasopressin mRNA-expressing neurons showed only a marked day–night difference in controls under 80 years of age. The amount of vasopressin mRNA was more than 3 times higher during the day than at night in controls, whereas no clear diurnal rhythm of vasopressin mRNA was observed in AD patients (Liu et al., 2000; Chapter 29.1g). The data mentioned above support the idea that damage to the SCN is the underlying anatomical substrate for the clinically often-observed disturbances in circadian rhythmicity in Alzheimer’s disease.

Fig. 4.26. Number of vasopressin (VP) expressing neurons in the suprachiasmatic nucleus. Note the low values in the 81–100-year-old group and the very low numbers in the AD patients (DEM) that were 78 ± 5 years of age. The decreased number of cells expressing VP is considered to be an indication for low metabolic activity of the SCN in old people and AD patients and the changes in the SCN in AD are held responsible for sleep disturbances and nightly restlessness. (Based upon Swaab et al., 1987a; Fig. 1, with permission.) The variability is largely due to circadian and circannual changes (Figs. 4.19 and 4.21).

2014 Ch 04

22/1/04

12:18 pm

Page 106

106 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

D.F. SWAAB

(c) Alzheimer’s disease: circadian disorders and light/melatonin therapy Old men with progressive ischemia By daylight get dreamier and dreamier. But when it is night They get up and fight From hypothalamic anaemia. T.G. Howell, 1943, from Jan Godderis

Disturbances in the sleep–activity rhythm are prominent and disabling symptoms of Alzheimer’s disease. Nighttime sleep is less efficient and lighter, and exhibits a high frequency of arousals and awakenings. Daytime activity is disrupted by multiple short daytime napping episodes, and in many patients afternoon or evening delirium (sundowning) is found, with agitation, wandering, irritability and confusion. A significant phase delay is found in the core body temperature rhythm. Night-time insomnia and nocturnal wandering in Alzheimer patients often poses unbearable problems for caregivers. Hypnotic or antipsychotic medication is only slightly effective (Witting et al., 1990; Van Someren et al., 1993; Dowling, 1996; Harper et al., 2001). Benzodiazepines have negligible effects on sundowning (Burney-Puckett, 1996), while sleep–wake cycle disturbances may even be aggravated by a classic neuroleptic-like haloperidol (Wirtz-Justice et al., 2000). Circadian rhythms of sleep–wake activity are more disturbed and affected in nursing home patients that are more seriously demented (Ancoli-Israel et al., 1997). Delirium in Alzheimer patients may be related to urinary tract infections, stressful events, surgery, medical illness and medication (Lerner et al., 1997). The SCN is affected by Alzheimer’s disease, since the typical cytoskeletal alterations have also been found in the SCN of these patients (Swaab et al., 1992b; Van de Nes et al., 1993; see Chapter 4.3b). With respect to the occurrence of degenerative changes of the SCN in Alzheimer’s disease, it is important to note that several factors attenuate the input of environmental light to the circadian timing system during aging and in Alzheimer’s disease. In the first place, Alzheimer patients were found to be exposed to less environmental light than their agematched controls (Campbell et al., 1988). The median light exposure of institutionalized demented patients was 54 lux and a median time spend of only 10.5 min was spent over 1000 lux. Higher light levels predicted fewer night-time awakenings and severe dementia predicted more daytime sleep and lower mean activity (Shochat et al., 2000). In addition, light adsorption of the lens changes

with aging (Sample et al., 1988). Especially light with short wavelengths is often not transmitted. Not only the retina but also the optic nerve, which provides direct and indirect light input to the SCN, shows degenerative changes in Alzheimer’s disease (Hinton et al., 1986; Katz et al., 1989; Trick et al., 1989; Blanks et al., 1996a,b). Moreover, cataract and maculopathy are more common in the elderly (Meisami, 1988) and age-related maculopathy is associated with Alzheimer’s disease (Klaver et al., 1999). In the macula of Alzheimer patients, retinal cell degeneration has been observed without neurofibrillary tangles, neuritic plaques or amyloid angiopathy being present in the retina or optic nerves (Blanks et al., 1989, 1996a,b). An extensive neuronal loss of some 36% was reported throughout the entire retina of Alzheimer patients, but was the most pronounced in the superior and inferior quadrants. The ratio of astrocytes to neurons is significantly higher in Alzheimer’s disease (Blanks et al., 1996b), indicating that a process of neurodegeneration takes place. On the other hand, there are also a number of observations suggesting that visual deficits in Alzheimer’s disease do not stem from neuroretinal dysfunction. Although a higher proportion of abnormalities was found in the retinal nerve fiber layer of Alzheimer patients than in controls (Hedges et al., 1996), and nerve fiber layer thickness in vivo was diminished in each quadrant in Alzheimer patients as measured by optical coherence tomography (Parisi et al., 2001), others found (by means of scanning laser polarimetry) that the retinal nerve fiber layer thickness was not altered in the earlier stages of Alzheimer’s disease (Kergoat et al., 2001). It should be added here that, in contrast to the observed degenerative changes mentioned above, in one study the reduced density of retinal ganglion cells in Alzheimer patients was found to be similar to that of aged controls in another study (Curcio and Drucker, 1993), and myelinated axon number in the optic nerve of Alzheimer patients was reported to be unaffected in another report (Davies et al., 1995). Moreover, the nerve cells located more distally in the retina give rise to electrical signals in early and moderate Alzheimer patients that are not different from those in controls in one study (Justino et al., 2001), but pattern electroretinogram responses were found to be abnormal in Alzheimer patients (Parisi et al., 2001). These discrepancies in the literature on retinal and optic nerve degeneration in Alzheimer’s disease should be resolved in future studies in order to make clear whether both the input of the visual system to the SCN and the SCN itself are indeed affected

2014 Ch 04

22/1/04

12:18 pm

Page 107


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

in Alzheimer’s disease. The finding that patients with Alzheimer’s disease may have an increased occurrence rate of glaucoma may also have implications on the input to the circadian system. Visual field defects and/or optic disc cupping compatible with the diagnosis “glaucoma” were found 5 times more frequently in Alzheimer patients than in controls (Bayer et al., 2002a,b). The more demented the Alzheimer patients, the more fragmented their sleep. Increased wandering at night and more frequent aggressive behavior during the day are associated with the use of sedative-hypnotics and with going to bed early (Ancoli-Israel et al., 1994, 1997). A subgroup of Alzheimer patients have a diminished capacity to synchronize the rhythm of core body temperature with the circadian cycle of rest activity (Satlin et al., 1995). Regression analysis showed that rest–activity rhythm disturbances are influenced by daytime activity and light (Van Someren et al., 1996), and that sleep–wake variables were highly correlated with, and explained a significant part of, the variance in cognitive and functional measures (Moe et al., 1995). Indeed, following exposure to extra amounts of bright light, behavioral disorders such as sundowning wandering, agitation or delirium almost disappeared, and sleep–wake rhythm disorders improved in Alzheimer patients (Hozumi et al., 1990; Okawa et al., 1991; Satlin et al., 1992; Mishima et al., 1994; Van Someren et al., 1997a; Fig. 4.27; Yamadera et al., 2000). As the compliance of demented patients in front of a light source is minimal, continuous attendance by the nursing staff is necessary to keep the patients in front of the source. This means that, at least in the Netherlands, it is barely feasible to get attendance from nursing staff for such a protocol and that, if they have the time to collaborate, a placebo effect is introduced by the simultaneous increase in attention of this staff. This is an important point, since social interaction with nurses is also effective for improving circadian rhythms (Okawa et al., 1991). Therefore, we have investigated whether the effect is also present if unattended exposure of patients to increased levels of indirect ceiling-mounted light during the daytime is maintained. This appeared indeed to be the case (Fig. 4.27), but only in demented patients with relatively unimpaired vision. It did not affect circadian rhythms in patients with severe visual deficits such as macular degeneration. The latter observation indicates strongly against a placebo effect of such a light treatment (Van Someren et al., 1997a). In a controlled trial with demented nursing home patients, Ancoli-Israel et al. (2002) showed that increasing exposure to morning bright light delays the

107

acrophase of the activity rhythm, which makes it easier to provide nursing care to patients and makes the circadian rhythm more robust. Improved circadian rhythmicity was also observed following transcutaneous nerve stimulation (Scherder et al., 1999a). A first study indicated that brightlight therapy in Alzheimer patients not only improved circadian rhythms but also the cognitive state of these patients, especially in the early stages of the disease (Yamadera et al., 2000). The favorable effect of a bright light source on mini-mental state score was confirmed by Graf et al. (2001). These observations indicate that stimulation of the circadian system by nonpharmacological means may have important therapeutic consequences for Alzheimer patients. It also shows that there is still plasticity in neuronal systems of aged individuals, even if they suffer from Alzheimer’s disease. Whether, in addition to light, the administration of melatonin is also effective in cases of sundowning (Cohen-Mansfield et al., 2000; Cardinali et al., 2002) is currently investigated by our group. In contrast to the literature mentioned above, Mishima et al. (1998) found that bright daytime light treatment induced a significant reduction in nighttime activity, but only in patients with vascular dementia and not in patients with dementia of the Alzheimer type. It should be noted, however, that the proportion of vascular dementias has always been overestimated when no neuropathological confirmation is performed, since in many demented patients with vascular lesions Alzheimer changes are found as well. Especially in Japan the proportion of vascular dementias has in the past always been estimated to be much too high for this reason. 4.4. The SCN in relation to sex, reproduction and sexual orientation (cf. Fig. 20) (a) Sex differences in sleep In the rat SCN there is a sex-specific diurnal pattern in VIP mRNA but not in vasopressin mRNA (Krajnak et al., 1998). The sex difference in shape in the vasopressinergic subnucleus of the SCN and the sex difference in VIP-expressing cell numbers (Swaab et al., 1985, 1994b; Zhou et al., 1995b) suggest the possibility that sex differences in circadian patterns are also present in humans. Moreover, in the human SCN, nuclear androgen receptor staining was more apparent in men than in women (Fernández-Guasti et al., 2000; Figs. 6.2 and 6.3;

2014 Ch 04

108 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 108

D.F. SWAAB

Fig. 4.27. Raw activity data (left panels) of a patient with Alzheimer’s disease assessed three times for five days; before (upper left panel), during (middle left panel) and after (lower left panel) light treatment. The right panels show double plots of the average 2- to 4-hour activity level (solid line) and one standard deviation above this level (dashed line). Note the decreased variability, the smoother average and the clearer difference between the day and the night during light treatment. (Van Someren et al., 1997a; Fig. 2, with permission.)

Table 6.1), while nuclear estrogen receptor and staining were stronger in women than in men and no sexual dimorphism was observed for nuclear progesterone receptors (Kruijver et al., 2002a,b, 2003; Table 6.2; Kruijver and Swaab, 2002; Fig. 4.28). This suggests that sex hormone receptor-dependent mechanisms may be instrumental in the functional sex differences in this nucleus, and that sex hormones affect the SCN neurons directly. Indeed, sex differences have been reported in sleep patterns that may be related to SCN sex differences. Women have higher percentages of slow-wave sleep and lower percentages of stage 1 sleep than men (Van Hilten et al., 1994), and about twice as many sleep spindles as males. Females tend to spend more time sleeping than men in a free-running environment. Also, middle-aged women display more slow-wave sleep than middle-aged men. The

period of free-running circadian rhythm is shorter and the fraction of sleep is significantly larger in women than in men (Wever, 1984). In addition, testosterone has relatively specific and discrete effects on sleep and hormonal rhythms in men (Leibenluft et al., 1997). Moreover, there are sex differences in morningness–eveningness preference (Adan and Natale, 2002). In healthy elderly women and men there are differences in entrained circadian temperature rhythms and sleep patterns that indicate that aging may affect the circadian timing system in a sexually dimorphic way (Campbell et al., 1989; Moe et al., 1991). Animal experiments indicate that only some of the sex differences in paradoxical sleep are dependent on circulating hormones (Fang and Fishbein, 1996) so that also organizing effects of sex hormones may be involved (see Chapters 4.4b and 24.5).

2014 Ch 04

22/1/04

12:18 pm

Page 109


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

109

Fig. 4.28. (A) Estrogen receptor (ER), (B) ER and (C) progesteron receptor immunoreactivity in SCN neurons. The asterisk (*) points to nuclear ER-immunoreactivity in smooth muscles and endothelial cells of a small blood vessel. Note the positive and negative nuclear ER, ER and PR staining in adjacent SCN neurons, as indicated by the arrows in A, B and C. Scale bar represents 8 m. (From Kruijver and Swaab, 2002; Fig. 3.)

2014 Ch 04

22/1/04

12:18 pm

Page 110

110 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

D.F. SWAAB

(b) The SCN in relation to sexual orientation, sleep and reproduction In addition to its possible involvement in reproduction (see below), the SCN might also play a role in sexual orientation. In fact, the first difference in the human brain in relation to sexual orientation was observed in the SCN. Morphometric analysis of the SCN of 10 homosexual men revealed that the volume of this nucleus was 1.7 times larger than that of a reference group of 18 presumed heterosexual male subjects, and that it contained 2.1 times as many cells (Fig. 4.29; Swaab and Hofman, 1990). In fact, the same high number of SCN vasopressin neurons as observed in 1- to 2-year-old children (Swaab et al., 1990; Fig. 4.24a) were also found in homosexual men. It seems as if the programmed postnatal cell death, which seems to begin in the SCN between 13 to 16 months after birth (Fig. 4.24a) does not occur to the same extent in homosexual men. The increased number of vasopressinexpressing neurons in the SCN of homosexual men appeared to be quite specific for this subgroup of neurons, since the number of VIP-expressing neurons was not changed. However, in both the vasopressin and VIP neurons in the SCN, a reduced nuclear diameter was observed in homosexual men, suggesting metabolic alterations in the SCN in relation to sexual orientation (Zhou et al., 1995a). Interestingly, homosexual orientation seems to be accompanied by changes in the sleep–wake cycle. Compared to heterosexuals, homosexual men and women wake up earlier, while only homosexual men go to sleep later. A different setting of the circadian pacemaker in homosexual subjects has been proposed on the basis of this study (Rahman and Silber, 2000). There are also a number of experimental data and observations on human material that indicate that the SCN is involved in aspects of sexual behavior and reproduction. Already in the early seventies, post-coital ultrastructural changes indicating neuronal activation were reported in the SCN of the female rabbit (Clattenburg et al., 1972). Also important is that the activity of SCN neurons increases suddenly around puberty (Anderson, 1981), indicating the addition of a reproductive function to the already mature circadian functions of the SCN. In addition, efferents of the SCN innervate the preoptic area, which is involved in reproductive behaviors. Extensive lesioning of the SCN area results in failure of ovulation in the female rat (Brown-Grant and Raisman, 1977). Rat studies indicate that the ovarian reproductive cycle is controlled by the SCN. For this function a direct

monosynaptic innervation of luteinizing hormonereleasing hormone (LHRH) neurons by VIP fibers is important (Van der Beek et al., 1993, 1997). Moreover, vasopressin fibers from the SCN that innervate the preoptic region may act as a circadian signal during a specific time window to induce a luteinizing hormone surge (Palm et al., 2001). Several morphological sex differences have been reported that support putative reproductive functions of the SCN. The SCN of male rats contains a larger amount of axospinal synapses, postsynaptic density material, asymmetrical synapses, and their neurons contain more nucleoli than those of female rats (Güldner, 1982, 1983). The sex difference in synaptic number in the rat SCN depends on androgens in development (Le Blond et al., 1982). In gerbils the volume of the SCN is sexually dimorphic (Holman and Hutchison, 1991) and so is the organization of astroglia in the SCN (Collado et al., 1995). A sex difference was found in the shape of the vasopressin subdivision of the human SCN (Swaab et al., 1985) as well as in the number of VIP-containing neurons in the human SCN. The number of VIP-expressing neurons in the SCN is larger in men of 10–40 years and larger in women of 41–65 years of age (Swaab et al., 1994b; Zhou et al., 1995b; Fig. 4.25). These observations are also consistent with sexually dimorphic functions of the SCN that are, however, still in need of a better definition. It is interesting to note, moreover, that the pineal hormone 5-methoxytryptophol shows significant sex differences: plasma concentrations increase in boys and decrease in girls after the age of 8 (Molina-Carballo et al., 1996). In seasonal breeders, VIP immunoreactivity in the SCN changes in relation to seasonal fluctuations in sexual activity (Lakhdar-Ghazal et al., 1992). The activation of c-Fos in the SCN by sexual stimulation (Pfaus et al., 1993) also points to a role of the SCN in reproduction. Bakker et al. (1993) found that male rats treated neonatally with the aromatase inhibitor ATD showed a clear sexual partner preference for females when tested in the late dark phase. When tested in the early dark phase, however, they showed a diminished preference for the female, or no preference at all. This is the first indication of the involvement of the clock, i.e. the SCN, in sexual orientation. The number of vasopressinexpressing neurons in the SCN of these ATD-treated bisexual animals was increased (Swaab et al., 1995b), something that was also found in homosexual men (Swaab and Hofman, 1990). This observation supports the possibility that the increased number of vasopressin-expressing

2014 Ch 04

22/1/04

12:18 pm

Page 111


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

111

Fig. 4.29. (a) Volume of the human suprachiasmatic nucleus (SCN) and sexually dimorphic nucleus of the preoptic area (SDN) as measured in three groups of adult subjects: (1) a male reference group (n = 18); (2) male homosexuals who died of AIDS (n = 10); (3) heterosexuals who died of AIDS (n = 6; 4 males and 2 females). The values indicate medians and the standard deviation of the median. The differences in the volume of the SCN between homosexuals and the subjects from both other groups are statistically significant (Kruskal–Wallis multiple comparison test, p* < 0.05; **p < 0.01; ***p < 0.001). Note that none of the parameters measured in the SDN (A,B) showed significant differences among the three groups (p always > 0.4). (b) Total number of cells in the human SCN and SDN. The SCN in homosexual men contains 2.1 times as many cells as the SCN in the reference group of male subjects and 2.4 times as many cells as the SCN in heterosexual AIDS patients. (c) The number of vasopressin neurons in the human SCN (the SDN does not contain vasopressin-producing cells). The SCN in homosexual men contains, on average, 1.9 times as many vasopressin neurons as the SCN in heterosexual AIDS patients. Notice that the SCN of heterosexual individuals who died of AIDS contains fewer vasopressin cells than the SCN of the subjects from the reference group. (From Swaab and Hofman, 1990; Fig. 2, with permission.)

2014 Ch 04

22/1/04

12:18 pm

Page 112

112 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

D.F. SWAAB

neurons in the SCN of adult homosexual men reflects a difference in the interaction between sex hormones and developing SCN neurons in the early stages of development in the brain. 4.5. Melatonin and its receptors I chose “mela” because the hormone lightens the frog’s melanophores, and “tonin” because the hormone is derived from serotonin. A. Lerner, 1958

(a) The pineal gland and other structures in the pineal region The pineal gland or epiphysis cerebri is a structure of the epithalamus of the diencephalon. It is 5 mm long, 1–4 mm thick and weighs about 90 mg, both in men and in women (Hasegawa et al., 1987). It has an ovoid shape, like a pine cone (pina in Latin). The arterial supply of the pineal gland comes through various groups of pineal arteries stemming mainly from the medial posterior choroidal arteries. The stalk lines the pineal recess, whose superior lip links the pineal gland to the habenular commissure and habenular nuclei and inferior lip to the posterior commissure (Duvernoy et al., 2000). Below the rostral part of the posterior commissure, the subcommissural organ (Chapter 18.7) is situated. Descartes (1596–1650) believed that the pineal gland was a sphincter, the point at which the soul preeminently controls the body. In 1958 Lerner and colleagues succeeded in isolating melatonin from bovine pineal glands. The compound was termed melatonin because of its blending effects on melanophores (Karasek, 1999). The pineal gland is a key structure of the circadian system and is connected to the SCN. The SCN is the clock of the brain but innervates only a small number of hypothalamic nuclei directly (Dai et al., 1997, 1998b). However, the SCN imposes circadian fluctuations indirectly on many more brain structures by means of melatonin from the pineal gland (Fig. 4.30). The pineal gland contains pinealocytes, which produce melatonin, and astrocytes. The pinealocytes are arranged in cords or lobules embedded in a matrix of neuroglia surrounded by septa. The pinealocytes have club-like endings. In the pineal gland, calcium and magnesium concrements are present, also known as acervuli, corpora arenacea, brain sand, or psammoma bodies. They contain hydroxyapatite and calcium phosphate. The incidence of concretions in

Fig. 4.30. Diagram of the human brain (mid-sagittal section) showing the neural pathways (dashed line) by which photoperiodic information reaches the pineal. Abbreviations: SCN, suprachiasmatic nucleus; PVN, paraventricular nucleus; SCG, superior cervical ganglion.

the pineal increases with age (McKinley and Oldfield, 1990), at least up to the age of 30 years, after which the degree of calcification did not seem to increase anymore (Hasegawa et al., 1987). Calcification of the pineal gland has been related to disturbed circadian rhythmicity in the sleep–wake cycle (Kunz et al., 1998) and the decline in melatonin production with age (Kunz et al., 1999; see Chapter 4.5d). The human pineal gland contains small, fusiform dopamine -hydroxylase positive neurons whose function is not known (Jengeleski et al., 1989). Serum melatonin levels are high at night and low during the day (Weaver et al., 1993). In the pineal itself the melatonin content is also higher during the night than during the day, but this difference is only significant in the long photoperiod (April–September) and in young subjects (30–60 years of age) (Hofman et al., 1995; Figs. 4.22 and 4.23; Luboshitzky et al., 1998; for circadian and seasonal fluctuations see Chapter 4.1). The synthesis of melatonin is presented in Fig. 4.31. L-Tryptophan is taken up from the circulation into the

2014 Ch 04

22/1/04

12:18 pm

Page 113


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 4.31.

Biosynthesis and metabolism of melatonin. (From Karasek, 1999; Fig. 1, with permission.)

113

2014 Ch 04

114 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 114

D.F. SWAAB

pineal gland and catalized to 5-hydroxytryptophan, decarboxylated to serotonin, and by N-acetyl transferase turned into N-acetylserotonin. The final step is the O-methylation of this compound to melatonin. Melatonin is metabolized in the liver, and secondarily in the kidney, to 6-hydroxymelatonin, followed by sulfate or glucuronate conjugation (Karasek, 1999). The main environmental stimulus for the entrainment of rhythmic melatonin fluctuations is light intensity. Descartes already intuitively proposed a pathway linking the eye and pineal gland (Descartes, 1662). The light information travels from the retina to the SCN via the retinohypothalamic tract and from the SCN via a multisynaptic pathway to the pineal gland (Fig. 4.30). However, since persistent 24-h variations in urinary 6-hydroxymelatonin sulphate and cortisol have been found in males in Antarctica, where a strong light–dark cycle of melatonin production is absent, it may also be entrained by other factors (Griffiths et al., 1986). Although rods or a rod-dominated mechanism were implicated in the suppression of melatonin by light (Rea et al., 2001), the light information is probably not mediated by the rods or classic three-cone photopic visual system (Brainard et al., 2001). The existence of a novel short-wavelength photo-pigment, melanopsin, in light-induced melatonin suppression has been proposed that would be a non-rod, non-cone photoreceptive system (Thapan et al., 2001; Barinaga, 2002; Berson et al., 2002; Hannibal et al., 2002) as discussed in Chapter 4.2e. It is remarkable that in totally blind people the duration of the melatonin secretion was not significantly different from that in healthy, sighted individuals (Klerman et al., 2001a). This was also found for other circadian rhythms (Bodenheimer et al., 1973). The human pineal gland has a dense noradrenergic plexus (Jengeleski et al., 1989). During darkness, noradrenaline is released from the sympathic nerve endings in the pineal gland to activate N-acetyl-transferase, the enzyme which catalyses the rate-limiting step of the synthesis from serotonin to melatonin (Mayeda et al., 1998; Fig. 4.31). In addition, it was found that the density of adrenoceptors in the pineal gland becomes higher between 18.00 and 20.00 hours. The upregulation of receptors coincided with an increase in the concentration of serotonin and N-acetylserotonin (Oxenkrug et al., 1990). The rapid adrenergic c-AMP regulation of N-acetyltransferase activity is mediated by rapid reversible control of selective proteasomal proteolysis (Gastel et al., 1998). Propanolol, a -receptor antagonist, causes a dosedependent decrease in melatonin levels under both light

and dark conditions, or even totally abolished the nighttime surge (Brown, 1992; Mayeda et al., 1998). In addition, somatostatin and its receptors are found in the human pineal gland. The somatostatin receptors sst 1, 2, 3 and 5 are present, but not 4. Moreover, synaptophysin, neurofilaments and chromagronin A were detected in the human pineal (Champiere et al., 2003). The pineal gland is innervated by a multisynaptic pathway from the SCN to the paraventricular nucleus (PVN), which is inhibited in the rat, by the biological clock through GABAergic transmission (Kalsbeek et al., 2000a). The sympathetic pathway subsequently goes to the intermediolateral column of the upper thoracic spinal cord and the superior cervical ganglion that sends noradrenergic fibers to the pineal gland (Fig. 4.29; AriënsKappers, 1965; Jengeleski et al., 1989; Larsen, 1999; Teclemariam-Mesbah et al., 1999; Duvernoy et al., 2000). A very large goiter may compress the superior cervical ganglia, thus altering the melatonin synthesis (Karasek et al., 2000). Under the influence of the noradrenergic innervation, melatonin is produced and released causing circadian fluctuations in many brain functions. In a patient with hyperhidrosis a prominent melatonin rhythm was observed preoperatively in the CSF and plasma. After bilateral T1–T2 ganglionectomy, melatonin levels were markedly reduced and the diurnal rhythm was abolished, providing evidence for the importance of the sympathetic innervation of the pineal gland (Bruce et al., 1991). The sympathetic fibers enter the posterior tip of the pineal gland. The paired, unmyelinated superior cervical ganglion afferent is called the nervi conarii. The afferent courses to the pineal within the inferior surface of the cistern connecting the posterior pineal gland and the falx (Sparks, 1998). This innervation is thought to induce the fluctuations in melatonin. Six additional pineal afferents enter the gland at its anterior aspect: (i) a single midline unmyelinated tract that courses from the rostral tectal area to the apex of the gland is only found in the fetus. It may be a vestigial tract similar to the nervus pinealis of lower animals; (ii, iii) two other inputs are made through the commissural peduncles. Commissural fibers from the habenular commissure and the posterior commissure entering the pineal gland appear to be confined to the anterior region; moreover (iv) a paired ventrolateral pineal tract has been described that is associated with the anterior inferio-ventro-lateral pineal gland. They form a structural component of the wall of the pineal recess and are myelinated. It has been suggested that these tracts were seen by Voltaire, when he asserted that “if

2014 Ch 04

22/1/04

12:18 pm

Page 115


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

the pineal gland is the seat of the soul, as Descartes suggested, then these must be the reins of the soul”. It is at present not known whether the last tracts are pinealofugal, pinealopetal, or both. Their function is equally obscure (Sparks, 1998); it seems that (v) hypothalamic information may reach the pineal gland more directly through the stria medullaris thalami, the habenular nuclei and the pineal stalk; moreover (vi) parasympathetic, cholinergic supply may reach the pineal gland from peripheral origin, e.g. from the superior salivatory nucleus via pterygopalatin and sphenopalatin ganglia and trigeminal ganglia (Larsen, 1999; Duvernoy et al., 2000). In its turn, melatonin elicits two distinct, separable, effects on the SCN, i.e. acute neuronal inhibition and phase shifting (Liu et al., 1997a). In the rat, melatonin was shown to inhibit the vasopressin release of SCN neurons and the electrical activity of SCN neurons (Watanabe et al., 1998). A recommended technique for removal of the human pineal gland at autopsy so as to retain its anatomical integrity is given in Sparks et al. (1997). It is generally thought that melatonin reaches its targets in the brain via the peripheral circulation. Melatonin would be secreted into large sinusoid capillaries in the central part of the gland. The venous drainage takes place via the lateral pineal veins that, in most cases, flow into the Galen vein, which drains into the sagittal sinus before entering the jugular vein. Venous melatonin is then transported back to the brain via the carotid arteries (Duvernoy et al., 2000). An alternative hypothesis is, however, that melatonin reaches its targets in the brain through diffusion to the cerebrospinal fluid (CSF) of the third ventricle. Experiments in ewes that show that melatonin levels in the third ventricle CSF were 7-fold higher than those in the lateral ventricle, and observations in various species that show that melatonin levels in the lateral ventricle CSF exceed those in the jugular plasma, support this possibility (Skinner and Malpaux, 1999). Moreover, melatonin concentrations are much higher in the pineal recess than in the third ventricle of the sheep. Since surgical sealing off of the pineal recess decreased melatonin levels in the third ventricle and because in the superior part of the pineal recess pinealocytes are in direct contact with the CSF, melatonin seems to enter the CSF, at least in this species, through the pineal recess (Tricoire et al., 2002, 2003). The observation that in humans the level of melatonin was significantly higher in CSF than in simultaneously sampled serum, is evidence against venous drainage of melatonin from the pineal into the peripheral circulation

115

and that it is compatible with the hypothesis of a direct pineal secretion of melatonin into the ventricular system. Also interestingly, a significant correlation was found between CSF and serum melatonin levels, which indicates that serum levels may be used to predict the concentrations in CSF (Rousseau et al., 1999). (b) Melatonin, the “vampire hormone” that acts like darkness Melatonin circulates, bound to plasma proteins, while saliva melatonin, present in concentrations up to 70% lower than those in plasma, reflects the circulating free hormone (Kennaway and Vaultsios, 1998). Melatonin is secreted from the pineal gland in a pulsatile pattern with a rate of 3–6 pulses/12 hours in normal adults and 9 pulses/hour in pre- and postpubertal children (Luboshitzky and Lavie, 1999). Pineal function shows circadian, monthly and circannual fluctuations. The onset of the nocturnal rise occurs around 21.00 and 22.00 hours and the offset between 07.00 and 09.00 hours. Peak levels occur between 02.00 and 04.00 hours (Luboshitzky and Lavie, 1999). The synthesis of indolamines in the human pineal exhibits a diurnal rhythm which is affected by seasonal changes in day length. A diurnal rhythm in pineal melatonin was evident only in the long photoperiod (April-September). In contrast, diurnal variations in the pineal 5-methoxythryptophol content were only observed in the short photoperiod (October-March). In general, night-time concentrations of melatonin and 5-methoxytryptophol were higher in the pineal gland in summer than in winter (Hofman et al., 1995; Figs. 4.22 and 4.23; Luboshitzky et al., 1998). A dark season is characterized by increased melatonin levels and decreased ovarian and androgen activities (Ronkainen et al., 1985; Kauppila et al., 1987). Since totally blind males have normal FSH and testosterone rhythms, the effect of light on pineal function is not considered to be crucial in the regulation of gonadal function in man (Bodenheimer et al., 1973). Since significant circadian and circannual ovarian follicular fluid concentrations were found, melatonin may interfere with the regulation of reproductive function, also at the follicular level (Rönnberg et al., 1990). The nocturnal rise in plasma melatonin is highest in autumn and lowest in winter (Asplund et al., 1998). The compressed melatonin waveform in humans experiencing a long natural summer photoperiod from sunrise until sunset may adapt rapidly to a shortening of the photoperiod (Vondrasˇová-Jelinková et al., 1999).

2014 Ch 04

116 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 116

D.F. SWAAB

Melatonin is involved in the mechanism of light therapy in depressed patients (Thalén et al., 1995). It was generally supposed that human beings required light of considerably higher intensity for melatonin suppression (1500–2500 lux for an intermediate and good response, and no response at 500 lux) than other mammals (e.g. 10 lux in the rat; Lewy et al, 1980). However, a more recent study showed that three cycles of a weak photic stimulus of about 180 lux of 5 hours centered 1.5 hour after the endogenous temperature nadir, significantly phase-advanced the plasma melatonin rhythm in healthy young men. Ordinary indoor room light can thus shift the pacemaker (Boivin and Czeisler, 1998). Humans are highly responsive to the phase-delaying effects of light during the early biological night, while both the phase resetting response to light and the acute suppressive effects of light on plasma melatonin follow a logistic dose-response curve. About half the maximum phasedelaying response and melatonin level suppression achieved in response to a single short episode of 9000 lux during the evening was obtained with 1% of this light (100 lux). Even small changes in ordinary light exposure during the late evening hours can thus significantly affect plasma melatonin concentrations and the entrained phase of the human circadian pacemaker (Zeitzer et al., 2000). Light-induced melatonin suppression is accompanied by phase changes of sleep propensity and core body temperature rhythms (Kubota et al., 2002). The importance of endogenous melatonin for circadian rhythms is supported by observations in totally blind people who have drifting, active and quiescent phases of melatonin production, each of about 12 hours duration (Lewy and Sack, 1996; Palm et al., 1997). A striking relationship has been observed between the timing of daytime production of melatonin and the timing of daytime naps in blind subjects (Lockley et al., 1997). All blind subjects who were bilaterally enucleated showed free-running melatonin rhythms (Skene et al., 1999). Moreover, melatonin administration can entrain free-running circadian rhythms in some (4 out of 7) blind subjects (Lockley et al., 2000). However, the integrity of the circadian system, e.g. presence or absence of the retinohypothalamic tract in blind people, should be better documented in order to be able to interpret these observations. It should be noted that melatonin is not only produced in the pineal gland, but also, e.g. in the Harderian gland, thymus, thyroid, pancreas, carotid body, placenta, endometrium, blood platelets, gut mucosa, kidney, adrenal, liver, cerebellum, airway epithelium, human

ovary, and retina (Kvetnoy et al., 1997, 1999; Itoh et al., 1999; Karasek, 1999; Savaskan et al., 2002a). Melatonin synthesis in the retina of mammals is under the control of a circadian oscillator located within the retina (Tosini, 2000; Savaskan et al., 2002a). It has therefore been questioned whether the various observations on plasma levels of melatonin, as mentioned above, may have been influenced by the extrapineal production of melatonin. However, in patients with a pineal germinoma, regardless of treatment option, melatonin plasma levels were nearly absent. In contrast, melatonin secretion and its circadian rhythms were not affected in patients with a hypothalamoneurohypophysial germinoma (Sawamura et al., 1998). These observations point to the pineal gland as the major source for plasma melatonin, also in the human. Retinal melatonin is thought to be involved in local cellular modulation (Savaskan et al., 2002a). (c) Hypothermic, cardiovascular, hypnotic and phaseshift effects Melatonin not only influences circadian rhythms and seasonal responses but has hypnotic and hypothermic effects as well (Nave et al., 1996; Hughes and Badia, 1997; Penev and Zee, 1997). Exogenous melatonin induces hypothermia in a dose-dependent manner (Satoh and Mishima, 2001). Light-induced melatonin suppression may be directly reflected in core body temperature rhythms rather than in melatonin rhythm (Kubota et al., 2002). In relation to the effects of melatonin on sleep, it is of interest that melatonin and body temperature rhythms are inversely coupled and that the hypothermic properties of melatonin are accountable for the generation of at least 40% of the amplitude of the circadian body temperature rhythm (Cagnacci et al., 1992, 1996). It is well known that nocturnal bright light exposure suppresses melatonin secretion and inhibits the fall in core temperature. In contrast, bright light exposure during the day reduces tympanic temperature and increases melatonin levels (Aizawa and Tokura, 1999). Other studies, too, support the notion that the therapeutic effect of melatonin on circadian rhythm sleep disorder is produced by an acute, transient hypothermic action (Mishima et al., 1997). Melatonin may thus execute its hypnotic effect by lowering body temperature, resulting from an increase in peripheral vasodilation and a concurrent heat loss (Van Someren, 2000a). Melatonin not only is presumed to be the neuroendocrine mediator of the circadian rhythm of sleep

2014 Ch 04

22/1/04

12:18 pm

Page 117


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

(Shochat et al., 1997; Van Someren, 1997), but also greatly influences arterial blood flow, decreases blood pressure in pharmacological amounts (Kitajima et al., 2001), and blunts noradrenergic activation. The nocturnal rise in melatonin may protect against cardiovasculatory accidents and seems to be beneficial in the treatment of essential hypertension. Vasoconstrictive effects of melatonin have been postulated to be mediated by the melatonin 1a receptor. This receptor is also present in the adventia of hippocampal arteries and is upregulated in Alzheimer’s disease (Savaskan et al., 2001), so that central effects of melatonin may also be mediated by its action on blood vessels. The effect of melatonin may depend on sex hormones, since the cardiovascular response to melatonin, causing a reduction in systolic and diastolic blood pressure, is only maintained in postmenopausal women if they receive hormone replacement therapy (Cagnacci et al., 1998a, 2001a). The onset of nocturnal melatonin secretion initiates the chain of events that 2 hours later leads to the opening of the sleep gate. Once the secretion into the bloodstream has begun, melatonin inhibits the SCN wakefulnessgenerating mechanism (Lavie and Luboshitzky, 1997). In fact, bright light prior to sleep not only suppresses melatonin levels and raises rectal temperature but also enhances low-frequency power of the EEG and delays REM sleep, indicating a phase-shift (Bunnell et al., 1992). Indeed, ingestion of melatonin affects sleep propensity (the speed with which one falls asleep) as well as the duration and the quality of sleep (Cagnacci, 1996; Brzezinski, 1997). The notion that melatonin possesses direct hypnotic effects (Mishima et al., 1997) is supported by a number of case histories. A child with a germ cell tumor involving the pineal region had markedly suppressed melatonin secretion associated with severe insomnia. Melatonin administration restored sleep continuity (Etzioni et al., 1996). In addition, in a patient whose pineal gland was removed 5 years earlier, in the course of treatment for a pineal astrocytoma, melatonin administration was associated with improvements in selfreported sleep and mood ratings (Petterborg et al., 1991). Another report supporting the involvement of endogenous melatonin in sleep regulation is that of a child with a pineal tumor and severe chronic sleep disorder. Oral melatonin greatly improved her sleep (Jan et al., 2001). In a blind 7-year-old child with a longstanding sleep– wake disorder, sleep onset and sleep offset improved significantly with melatonin administration (Cavallo et al., 2002).

117

Altered circadian melatonin secretion patterns were found in patients with various forms of chronic sleep– wake rhythm disorders (Rodenbeck et al., 1998). Nocturnal melatonin secretion was significantly diminished in patients with primary insomnia (Riemann et al., 2002). Melatonin has been used effectively in the treatment of sleep disorders in all phases of life, and in a multitude of disorders, such as Rett syndrome (McArthur and Budden, 1998; Miyamoto et al., 1999; Yamashita et al., 1999), autism (Hayashi, 2000), children with severe learning disorders (Gordon, 2000) and with other developmental disabilities (Dodge and Wilson, 2001; Ross et al., 2002), in children with chronic sleep onset insomnia (Smits et al., 2001), delayed sleep phase syndrome following traumatic brain injury and whiplash syndrome (Nagtegaal et al., 1998; Smits and Nagtegaal, 2000), in sleep–wake disturbances in visually handicapped children, and in young adults (Palm et al., 1997; Gordon, 2000; Cavallo et al., 2002) and elderly people with sleep–wake disturbances (Jean-Louis et al., 1998; Zhdanova et al., 2001). In REM sleep behavioral disorder, characterized by vigorous sleep behaviors accompanying vivid, striking dreams and REM sleep without muscle atonia, a dramatic clinical improvement was observed following melatonin treatment, possibly by restoring REM-atonia (Kunz and Bes, 1999; Schenck and Mahowald, 2002). In septo-optic dysplasia, where the SCN seems to be absent, arrhythmicity could be changed into normal sleep–wake cyclicity by melatonin administration (Chapter 18.3b; Rivkees, 2001). In Angelman syndrome, melatonin promotes sleep and reduces motor activity during the sleep period (Zhdanova et al., 1999), while melatonin was also used to improve the quality of sleep and mood of patients with major depressive disorder (De Vries and Peeters, 1997; Dolberg et al., 1998). In addition, melatonin enhanced the rest–activity rhythm in elderly people and improved sleep quality, although total sleep time was not significantly increased. Interestingly, memory and mood were significantly improved by melatonin. However, another study reported that 5 mg of fast release melatonin taken at bedtime does not improve the quality of sleep in older people with age-related sleep maintenance problems (Baskett et al., 2003). In medically ill patients with initial insomnia, melatonin hastened sleep onset without producing drowsiness, improved the quality and depth of sleep, increased the duration of sleep, and increased morning freshness on awakening (Andrade et al., 2001). In periodic limb movement disorder, melatonin seems to have a chronobiologic effect, in that the output amplitude

2014 Ch 04

118 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 118

D.F. SWAAB

of the circadian rhythmicity of locomotor activity is enhanced, with a reduction of sleep motor activity (Kunz and Bes, 2001). Moreover, melatonin was found to be effective in the treatment of tardive dyskinesia (Shamir et al., 2001). On the basis of observations in patients with acute intermittent porphyria, it was concluded that melatonin may have a protective effect on seizures (Bylesjö et al., 2000). The anti-epileptic activity of melatonin, which may be based upon its antioxidant activity as a free radical scavenger, was confirmed in children with severe intractable seizures. In addition, melatonin reduces glucose tolerance and insulin sensitivity (Cagnacci et al., 2001b). Melatonin was inactive as a sleeping pill in patients with neuronal ceroid lipofuscinosis with fragmented or normal motor activity rhythms recorded by wrist actigraphy (Hätönen et al., 1999). Melatonin is capable of entraining (synchronizing) a free-running circadian rhythm in most blind people (Lockley et al., 2000; Sack et al., 2000). The ability of melatonin to phase-shift the circadian system has been extensively investigated in humans. The time of melatonin administration rather than the pharmacological dose seems to be the crucial factor as far as its phase-shift potency is concerned (Luboshitzky and Lavie, 1999). In order to cause a phase advance, light should be scheduled in the morning and melatonin administered in the afternoon. In order to cause a phase delay, light should be scheduled in the evening and melatonin should be administered in the morning (Lewy and Sack, 1996). Administration of melatonin has phase-shifting therapeutic actions in circadian sleep disorders, including disorders associated with jet lag, blindness, shift work, delayed phase sleep disorder, non-24 hour sleep–wake syndrome, REM sleep behavioral disorder (a parasomnia characterized by a loss of REM-associated atonia, usually accompanied by vivid dreams, punching, kicking, yelling and leaping out of bed in sleep, which often result in injuries) (Takeuchi et al., 2001), periodic sleep disorder in blindness, and sleep and behavioral disorders in children with or without severe mental retardation, autism, adolescents, multiple brain damage or Rett syndrome (McArthur and Budden, 1998; Yamashita et al., 1999; Akaboshi et al., 2000; Hayashi et al., 2000; Kamei et al., 2000a; Kayumov et al., 2001), tuberous sclerosis (O’Callaghan et al., 1999; Ross, 1999), winter depression (Lewy et al., 1998) and major depression (see Chapter 26.4; Palm et al., 1991, 1997; Brown, 1995; Lewy and

Sack, 1996; Smits et al., 1996; Arendt et al., 1997; Kunz and Bes, 1997; Nagtegaal et al., 1997; Jan et al., 1998, 1999). Melatonin could be used to promote adaptation to night work and jet travel (Sharkey and Eastman, 2002). A double-blind trial could, however, not confirm the effectiveness of melatonin in the alleviation of jet lag (Spitzer et al., 1999). An improvement of the symptoms of sundowning was seen in demented patients following melatonin administration (Cohen-Mansfield et al., 2000). Moreover, a man with a non-24-hour sleep–wake syndrome with a period of 25.1 hours and a subsensitivity to bright light was entrained to the 24-hour day/night cycle by melatonin (McArthur et al., 1996). The endogenous melatonin profile is not affected by melatonin treatment in human beings, although it can shift the phase. Consequently there is no indication for a feedback inhibition of pineal melatonin by such a therapy (Matsumoto et al., 1997). The phase delay of the circadian system by evening light appeared to be independent of an immediate hyperthermic effect and is not mediated by melatonin (Kräuchi et al., 1997). (d) Age and sex From 15–17 weeks of fetal life onwards the human hypothalamus contains melatonin (MT2) receptors in the suprachiasmatic, ventromedial, arcuate and mamillary nucleus (Thomas et al., 2002). The fetus is probably exposed to melatonin from the maternal pineal gland, as evidenced by the presence of melatonin in umbilical cord blood. The fetal pineal gland synthesizes melatonin as early as 26 weeks of gestation. Melatonin in humans is transferred from the maternal to the fetal circulation easily and rapidly. In addition, the neonate may get melatonin from the mother via maternal milk. The postnatal development of melatonin rhythm does not occur until 49–55 weeks postconception or 2–3 months postnatally (Thomas et al., 1998; Luboshitzky and Lavie, 1999), when the vasopressin-expressing neurons in the SCN increased (Swaab et al., 1990). In a recent study significant day–night differences in urinary 6-sulfatoxymelatonin levels were observed already between 27–41 days postnatally (Ardura et al., 2003). Low nocturnal 6-sulfoxy melatonin excretion in the first weeks of life correlates with delayed psychomotor achievements at 3 and 6 months of life. Whether this is a causal or predictive link should be investigated (Tauman et al., 2002). There seems to be an association between melatonin secretion patterns and the evolving sleep–wake organization in the second

2014 Ch 04

22/1/04

12:18 pm

Page 119


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

half of the first year of life during development. A delayed peak of melatonin was associated with poorer sleep quality (Sadeh, 1997). Plasma melatonin levels show maximum levels around the ages of 3–7 years and decline with age, with a major decline occurring before puberty (Penev and Zee, 1997; Touitou, 1997; Touitou et al., 1997; Waldhauser et al., 1988; Luboshitzky and Lavie, 1999). The pineal hormone 5-methoxytryptophol plasma levels show age and sex differences. Plasma levels increase in boys and decrease in girls from the age of 8 onwards (Molina-Carballo et al., 1996), a pattern that is the reverse of the one we found for the number of VIP-expressing neurons in the SCN (Zhou et al., 1995b). The pattern of changes in 5-methoxytryptophol in girls may have a permissive effect on puberty (Molina-Carballo et al., 1996). The idea that the pineal gland may affect puberty dates back to 1898 when Otto Heubner described a 4.5-year-old boy with precocious puberty and a nonparenchymal tumor that had destroyed the pineal gland (Brzezinski, 1997). Although effects of pineal region tumors on puberty may also be due to local pressure of such tumors on the hypothalamus (Chapters 19.1, 19.7), the case of the 21year-old male patient described by Puig-Domingo et al. (1992) supports the idea that melatonin might play a crucial part in the development of reproductive activity in human beings. When the patient’s melatonin levels were 15–20 times the normal value, the patient’s pituitarygonadal function, including sperm production, was disturbed, as observed in many seasonally breeding animals. Full sexual capability was restored when the melatonin secretion gradually decreased. The patient’s hypogonadotrophic hypogonadism, in fact an extreme form of delayed puberty, was thus probably caused by hypermelatoninemia. In pharmacological doses, melatonin indeed induces decreased serum LHRH levels and increased prolactin levels (Brzezinski, 1997). On the other hand, decreased levels of sex hormones as found in hypothalamic amenorrhea (see below) and hypogonadal men increase melatonin levels. Following testosterone replacement therapy in hypogonadal men, the high melatonin levels decreased both during the day and at night (Rajmil et al., 1997). A study in Klinefelter patients (see Chapter 24.4) showed higher melatonin levels and lower levels of metabolites in urine than in controls. After testosterone treatment, melatonin levels fell, whereas the urine metabolites increased. The effect of testosterone on melatonin levels thus seems to be mediated by an enhanced metabolism of melatonin and not by any effect on sympathetic outflow (Caglayon et al., 2001). Clearance

119

studies showed that melatonin secretion occurs at a similar rate in men and women (Fourtillan et al., 2001). Although stimulatory effects of melatonin on gonadotropin secretion have been reported in the follicular phase of the menstrual cycle (Cagnacci, 1996), melatonindecreased LH levels in women between 43 and 49 years old (Bellipanni et al., 2001). Melatonin regulates reproduction in animals and may thus be involved in the development of puberty and have an effect on gonadotropin levels in humans. Yet, attempts to utilize melatonin for contraception have so far failed (Rohr and Herold, 2002). It is interesting to note that in the human pineal gland, LH and FSH receptors show a significant seasonal variation with 20-fold higher values in winter than in summer. For FSH receptors day/night differences were present only during the summer. Androgen and estrogen receptors in the pineal do not reveal any seasonal changes. These receptors are already present in the first 2–4 months of life (Luboshitzky et al., 1997). Elderly subjects (61–84 years of age) had lower pineal melatonin contents than younger (30–60 years of age) subjects. Although the age-related difference was not statistically significant in some studies (Luboshitzky et al., 1998), the nocturnal decline of melatonin levels was significant at the age of 60 in another study and further declined in the 70’s and 80’s (Zhao et al., 2002). In addition, the hypothermic response to melatonin is markedly blunted and inconsistent in aged individuals. In postmenopausal women, who have higher nocturnal melatonin levels than men (Zhao et al., 2002), the effect of melatonin on cerebral blood flow is reduced or absent (Cagnacci et al., 1997). Although many reports indicate that melatonin levels decline with age, especially the nocturnal melatonin peak (Ferrari et al., 2000), some recent studies did not support such a diminishment (Zeitzer et al., 1999; Fourtillan et al., 2001), although they were again challenged by others (Cornelissen et al., 2000). We found that the age-related decline in saliva nocturnal peak levels of melatonin begins during middle age (Zhao et al., 2003). The daily variation in melatonin content in the pineal with higher melatonin levels in the younger age group (18–54 years) was not maintained in the older age group (55–92 years; Skene et al., 1990). The peak of circadian melatonin occurs later within the sleep of elderly subjects (Duffy et al., 2002). The pinal calcification increases with age. The degree of calcification correlated positively with the incidence of chronic daytime tiredness as well as with the subjective

2014 Ch 04

120 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 120

D.F. SWAAB

perception of sleep disturbances (Kunz et al., 1998). However, even in very old subjects, the pineal parenchyma is histologically still very active (Arieti, 1954). The size of the pineal calcification is not associated with melatonin secretion. However, an approximation of the size of the uncalcified pineal tissue, presumably representing active pinealocytes, is positively associated with the total amount of 24-hour 6-sulphatoxymelatonin excretion in the urine. Calcification of pinealocytes is considered by some to result from death or degeneration of the cell itself, thus leading to an overall decrease in pineal activity (Kunz et al., 1999). However, the case described by Puig-Domingo et al. (1992) of the 21-year-old patient with very high melatonin levels and hypogonadotropic hypogonadism, also supports the view that pineal calcification may be the result of hyperactivity of the pineal and not a sign of inactivity or atrophy. There thus seems to be little if any support for the hypothesis that the pineal gland would be a centralized clock for aging, i.e. that the calcification process in the pineal gland would provide the bioinorganic timing mechanism for the aging process and that the secreted melatonin would carry the signal to all cells in the organism (Kloeden et al., 1990). A review of the literature on the excretion of the melatonin metabolite, 6-sulfatoxymelatonin, shows that the melatonin production is lower in older people, but that the most important change occurs very early in life, around 20–30 years of age (Kennaway et al., 1999). Moreover, there is a huge (20-fold) genetically determined interindividual variability in humans in the amount of melatonin secretion and in the size of the pineal gland (Kunz et al., 1999). Elderly persons, especially the insomniacs among them, have diminished nocturnal melatonin secretion and are often exposed to lower amounts of environmental light (Mishima et al., 2001). Supplementary exposure to midday bright light significantly increased melatonin secretion in these subjects. The amount of environmental light is thus a crucial factor in this type of study. On the other hand, older people of over 65 years of age with age-related sleep maintenance problems do not have lower melatonin levels than older people reporting normal sleep (Baskett et al., 2001). Animal experiments have shown that dark-cycle night administration of melatonin in drinking water or transplantation of pineal glands from young to old mice may prolong survival and preserve cellular functions, despite age (Pierpaoli and Regelson, 1994). Whether melatonin may also have such antiaging effects in humans should be studied.

(e) Alterations in melatonin levels in various disorders Intrauterine growth retardation or fetal distress in human infants is associated with a pronounced reduction in melatonin secretion during the first months of life. Remarkably, urinary 6-sulfadoxymelatonin excretion was impaired in adults who were growth-restricted after 40 weeks of gestation (Kennaway et al., 2001). The circadian and circannual pattern of occurrence of sudden infant death syndrome (SIDS), i.e. death occurring preferentially during the winter months and the night, the occurrence of sleep and temperature regulation disturbances and pineal dysfunction, has implicated melatonin in the etiology of SIDS (Thomas et al., 1998). CSF melatonin levels are lower in children with SIDS, indicating the presence of circadian disturbances or altered pineal function in this disorder (Sturner et al., 1990). The pineal is indeed reported to be smaller in children with SIDS (Sparks and Hunsaker, 1988, 2002). It is not clear whether the SCN itself is affected in SIDS. Children with a life-threatening event also had lower melatonin production (Sivan et al., 2000), raising doubt about the specificity of the melatonin changes found in SIDS. In Smith–Magenis syndrome, caused by a deletion on chromosome 17p11-2, a completely inverted melatonin secretion cycle was found. Tantrums occur when melatonin increases and sleep attacks take place when it peaks (McBride, 1999). A -adrenergic agonist and melatonin administration may help to manage hyperactivity, enhances cognitive performance and reduces sleep disorder in children with this syndrome. In boys with fragile-X syndrome, both nocturnal and daytime melatonin levels were higher, possibly due to overactivity of the sympathetic nervous system. These children have greater variability in total sleep time and difficulty in sleep maintenance (Gould et al., 2000). In children, irregular ultradian rhythms of melatonin are found during the night, in contrast to the regular ultradian rhythms of the adults. In children with growth disorders, abnormal melatonin levels are found (Muñoz-Hoyos et al., 2001). A remarkable finding is that prolonged use of cellular phones may lead to reduced melatonin production and elevated 60Hz magnetic field may potentiate the effect (Burch et al., 2002). Human serum melatonin levels change during the menstrual cycle. Elevated levels are found at the time of the menstrual bleeding and a nadir is observed at the time of ovulation (Wetterberg et al., 1976). Nocturnal

2014 Ch 04

22/1/04

12:18 pm

Page 121


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

melatonin levels are elevated in women with hypothalamic amenorrhea (Berga et al., 1988; Brzezinski et al., 1988). The finding that the levels of melatonin and its metabolites are elevated both in LHRH deficient women with Kallmann’s syndrome and in women with idiopathic hypogonadotropic- hypogonadism suggests that nocturnal melatonin is elevated as a consequence of LHRH deficiency, irrespective of its etiology (Kadva et al., 1998). Indeed, male patients with gonadotropin-releasing hormone deficiency have increased nocturnal melatonin secretion that decreases to normal levels during testosterone treatment (Luboshitzky et al., 1995, 1997a,b). This idea also fits in with the observation that, in hypergonadotropic-hypogonadal males, melatonin secretion is decreased (Luboshitzky et al., 1997b). Other studies found day and night elevations in melatonin levels in anovulatory states. Disturbed melatonin patterns were observed in women with exercise-induced amenorrhea, and in anorexic women, although these observations were not always confirmed (Cagnacci and Volpe, 1996). However, these reports in general support the idea that increased levels of melatonin may induce anovulation, but the reverse may also be true (see earlier). Melatonin levels were significantly higher in premenopausal female patients with hyperprolactinemia and hyperandrogenemia, while obese women showed lower melatonin production (Blaicher et al., 1999a,b). In postmenopausal women with visceral obesity, day-night fluctuations in melatonin levels were accompanied by fluctuations in markers for bone mass loss. The circadian rhythm of melatonin is disturbed by oral contraceptives (Reinberg et al., 1996), and melatonin secretion is augmented in women on oral contraceptives (Kostoglou-Athanassiou et al., 1998b; Wright et al., 2000). In the past, melatonin has been hailed as a potential therapy for aging and cancer because it seems to be a potent radical scavenger, but most of these claims have little credible scientific support (Rivkees, 1997). However, there have not been sufficient well-controlled clinical trials. In addition, pinealectomy enhances tumor growth and metastatic spread in experimental animals. The effect is only partly due to melatonin, since melatonin-free pineal extracts containing as yet unidentified pineal substances have also shown tumorinhibiting activity. Disturbances of melatonin secretion were found in the case of duodenal ulcer (Malinovska et al., 2001). A relationship between melatonin and breast cancer has been suggested but not proved (Penev and Zee, 1997). Epidemiological studies showing that blind people have half the rate of breast cancers support this

121

hypothesis (Rohr and Herold, 2002). In cancer patients, a decreased amplitude of the melatonin rhythm has been reported with lower levels at night and higher levels during the day (Tarquini et al., 1999). In women with clinical stage I or II breast cancer with estrogen receptor positivity, the nocturnal increase in melatonin was much lower than observed in control subjects. Women with the lowest peak concentration of melatonin had tumors with the highest concentrations of estrogen receptors. Since low nocturnal melatonin concentrations might thus indicate the presence of estrogen positive breast cancer, melatonin was presumed to be involved in the pathogenesis of this condition (Tamarkin et al., 1982). Melatonin was found to inhibit growth in one out of six mamillary carcinomas in a dose-dependent manner. Cell cultures of ovarian tumors were either inhibited or stimulated. For alterations in melatonin secretion in tumors of the pineal region, see Chapter 19.7. Depressed nocturnal concentrations of melatonin have also been found in other malignancies such as prostate cancer, colorectal carcinoma and adenocarcinoma of the corpus uteri (Karasek, 1999). Some preliminary data suggest antiproliferative properties of melatonin in neoplastic patients (Karasek and Pawlikowski, 1999). In addition, women who work on rotating night shifts with at least 3 nights per month on duty appear to have a moderately increased risk of breast cancer after extended periods of working in these shifts (Schernhammer et al., 2001). Melatonin is presumed to be involved in this effect. Impaired nocturnal secretion of melatonin is associated with acute coronary heart disease (Brugger et al., 1995; Dominguez-Rodriguez et al., 2002), is found in chronic renal failure (Karasek et al., 2002), and postoperative patients who underwent cardiac surgery lost the rhythmicity of melatonin and body temperature. Postoperative sleep disturbances may be related to the decreased plasma melatonin levels during the first day after the operation (Cronin et al., 2000). After coronary surgery, melatonin secretion is disrupted, but circadian rhythms have generally returned on postoperative day 2 (Guo et al., 2002). Anesthesia in conjunction with minor orthopedic surgery was found to be accompanied by lower saliva melatonin levels on the first postoperative day (Kärkelä et al., 2002), while another study found increased melatonin levels on postoperative days 2 and 3 that are thought to be related to stress, as stress, either acute or chronic, of any kind increases melatonin secretion. It has been found that postoperative delirium is associated either

2014 Ch 04

122 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 122

D.F. SWAAB

with an increase in melatonin (Uchida et al., 1999) or with decreased melatonin levels. It was proposed that unless serious postoperative complications occur, which would markedly increase melatonin levels through the noradrenergic pathway, decreased melatonin secretion after surgical operation triggers sleep disturbances in elderly patients, which in turn cause delirium (Shigeta et al., 2001). Melatonin indeed has been used successfully to treat or prevent postoperative delirium in a few cases (Hanania and Kitain, 2002). Well-controlled trials are needed to confirm such effects. Melatonin also has anticonvulsive properties. Patients with intractable epilepsy have low melatonin levels that increase following seizures (Bazil et al., 2000). During cluster headache periods the acrophase of melatonin is moved forward and the night-time peak is blunted and significantly reduced. The daytime levels of the melatonin metabolite 6-sulfatoxymelatonin did not differ from those at night, nor did they differ during cluster headache or remission, indicating the involvement of hypothalamic rhythm regulating centers in this disorder (Leone et al., 1998; see Chapter 31.2). Melatonin may alleviate cluster headache attacks according to one publication (Peres and Rozen, 2001), but not according to another (Pringsheim et al., 2002). In migraine, melatonin levels are low, and administration of melatonin to these patients would normalize the circadian cycle and relieve migraines (Gagnier, 2001). In multiple sclerosis (MS) pineal failure may be present. During exacerbation of MS symptoms, nocturnal melatonin levels below daytime values are frequently found (Sandyk and Awerbuch, 1992). Melatonin enhances immune function (Nelson et al., 1995; Haimov et al., 1997). In this respect it is important that CD4+ T-lymphocytes have specific highaffinity binding sites for melatonin. This suggests the possibility of a direct effect of melatonin on immune functions (Karasek, 1999). In critically ill patients with sepsis, circadian melatonin fluctuations were impaired. In contrast, in severely ill nonseptic patients this is not the case, indicating it is the sepsis that causes the impaired rhythms (Mundigler et al., 2002). A small study indicated that melatonin treatment of septic newborns may reduce oxidative stress and improve clinical outcome (Gitto et al., 2001), but a large randomized study is needed to come to a firm conclusion. In patients with liver cirrhosis, a delayed timing of the nocturnal rise in plasma melatonin and increased daytime levels of the hormone have been described (Penev and Zee, 1997). Familial pineal hyperplasia, in association

with insulin-resistant diabetes mellitus, is part of a rare familial condition known as Rabson–Mendenhall’s syndrome (Penev and Zee, 1997). In a number of psychiatric disorders, abnormal melatonin levels have been reported. A reduction of nocturnal levels of melatonin has been reported in the majority of depressed patients (Brown et al., 1995; Chapter 26.4) in seasonal affective disorder, bipolar disorder, and unipolar depression (Pacchierotti et al., 2001), but this was not confirmed in some other studies (Penev and Zee, 1997; Voderholzer et al., 1997; Kripke et al., 2003). Melancholic depressed patients had lower evening levels of melatonin in a Greek study (Fountoulaki et al., 2001). Melatonin secretion abnormalities were found in a subgroup of patients with a bipolar disorder (Nurnberger et al., 2000). In postmenopausal women with a positive family history of depression (Tuunainen et al., 2002), melatonin treatment significantly decreased depression ratings compared to treatment with a placebo in a pilot study (Lewy et al., 1998a), improved mood in two case studies (De Vries and Peeters, 1997; Petterborg et al., 1991) and in elderly people with sleep-wake disorders (Jean-Louis et al., 1998). However, Dolberg et al. (1998) could not find an effect on the rate of improvement in symptoms of major depression disorder. Whereas higher melatonin levels have been found during the manic phase in patients with bipolar depression, patients with schizophrenia would have decreased night-time levels of melatonin (Penev and Zee, 1997; Pacchieratti et al., 2001; Vigano et al., 2001). Moreover, MRI measurements showed that the pineal volume was smaller in schizophrenic patients than in controls (Bersanim et al., 2002). In drug-free paranoid schizophrenics, no circadian rhythm of plasma melatonin was found, whereas the circadian rhythm of plasma cortisol was preserved (Monteleone et al., 1992). Intravenous melatonin administration to schizophrenic patients in remission caused a worsening of psychotic symptoms which persists after the treatment is stopped (Pacchierotti et al., 2001), suggesting that melatonin may be involved in the pathogenesis of schizophrenia. In addition, melatonin appeared to exaggerate the first-night effect, i.e. the tendency to have a bad first night’s sleep in a clinic, in schizophrenic patients (Shamir et al., 2000). One study reported increased nocturnal melatonin levels in panic disorder (Brown, 1996). Moreover, an altered melatonin secretion pattern was found in bulimia and anorexia nervosa, and obsessive compulsive disorder (Pacchierotti

2014 Ch 04

22/1/04

12:18 pm

Page 123


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

et al., 2001). According to some studies, chronic fatigue syndrome and fibromyalgia (Chapter 26.8) are associated with low melatonin secretion, while other studies reported higher night-time melatonin levels in these patients (Korszun et al., 1999; Knook et al., 2000). The reason for this discrepancy is not clear, while the presence of decreased quality of sleep is obvious in these syndromes. Hypothalamic lesions due to craniopharyngeoma or hypothalamic pilocytic astrocytoma resulted in increased daytime sleepiness and decreased nocturnal melatonin levels (Müller et al., 2002a). Absence of melatonin rhythms has been reported more frequently in demented patients than in controls. In patients who lack serum melatonin rhythms, clinical symptoms of rhythm disturbances such as delirium and sleep–wake disturbance were frequently but not always observed (Uchida et al., 1996). A significant decrease in agitated behaviors and confusion in the evening hours, or ‘sundowning’ was seen following the administration of melatonin (Cohen-Mansfield et al., 2000). The finding that the daily variations in pineal melatonin and 5-methoxy-tryptophol content disappeared in Alzheimer patients is consistent with the clinical observations of sleep disorders and sundowning in these patients (Skene et al., 1990; Fig. 23; Chapter 4.3). Others have reported that there is a selective impairment of the nocturnal melatonin peak in dementia (Ferrari et al., 2000) or that the melatonin levels are increased in Alzheimer patients during daytime and that these patients do not react to bright light (Ohashi et al., 1999), indicating that the neurodegenerative process has affected the circadianpineal system. Melatonin deficiency may contribute to the pathogenesis of Alzheimer’s disease, since it was found in in vitro experiments that melatonin functions as an antioxidant and neuroprotector in rat and primate brain tissue (Papolla et al., 2000; Reiter et al., 2000; Tan et al., 2000), and that it inhibits the progressive formation of sheets and amyloid fibrils and the secretion of soluble A (Pappola et al., 1998, 2000). Melatonin might also act via apolipoprotein (ApoE4). ApoE4 binds to the Alzheimer amyloid protein A and, under experimental conditions, promotes the formation of -sheet structures and amyloid fibrils. In vitro, melatonin inhibits this fibril formation. Although patients with Alzheimer’s disease have diminished functioning of the pineal gland, no evidence was observed in this structure of neurofibrillary tangles, the accumulation of neurofilaments, tau, hyperphosphorylated tau (stained by Alz-50) or /A4 amyloid deposition in pinealocytes. So, although the decreased

123

melatonin levels indicate decreased functioning of pinealocytes in Alzheimer’s disease, this does not involve typical neuropathological Alzheimer changes of intrinsic cells or afferent fibers (Pardo et al., 1990). The decline in pineal function may be due to alterations in the afferent pathway, i.e. in the superior cervical ganglia, the noradrenergic innervation of the pineal gland that shows swollen axons in aged individuals and Alzheimer patients, or in the SCN of aged individuals and Alzheimer patients (Swaab et al., 1985; Jengeleski et al., 1989; Pardo et al., 1990; Chapter 4.3; Fig. 4.26). In Alzheimer patients with disturbed sleep–wake patterns, melatonin secretion patterns are irregular (Mishima et al., 1999). We have observed strongly decreased postmortem CSFmelatonin levels in Alzheimer patients. The melatonin levels in CSF of Alzheimer patients were only onefifth of those in control subjects. The melatonin levels of patients with ApoE3/4 type were significantly higher than those expressing ApoE4/4 (Chapter 29.1). The impairment of nocturnal melatonin secretion is related to mental impairment (Magri et al., 1997), and the suppletion of melatonin in Alzheimer patients may improve circadian rhythmicity and suppress sundowning (Cardinali et al., 2002). Beneficial effects of melatonin on memory, sleep-disturbances and reduction of sundowning were reported in a case report on a monozygotic twin and in a small retrospective study on Alzheimer patients (Brusco et al., 1998, 2000) while a randomized placebo controlled trial of melatonin administration to demented patients did not improve sleep (Serfaty et al., 2002). However, because in elderly people with sleep disorders melatonin improved the ability to remember previously learned things (Jean-Louis et al., 1998), positive cognitive effects of melatonin in Alzheimer patients seem to be a possibility, but this has to be confirmed in well-controlled studies. On the other hand, exposing Alzheimer patients to bright light did not lower their serum melatonin levels (Ohashi et al., 1999) so that the plasticity of this part of the system is not apparent at present. (f) Melatonin receptors, additional effects and side effects of melatonin For hypnotic, hypothermic, cardiovascular and phase-shift effects, see Chapter 4.5c. Specific high-affinity melatonin (MT1) binding sites have been observed consistently in the human SCN area, ventromedial nucleus, arcuate nucleus and mamillary

2014 Ch 04

124 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

22/1/04

12:18 pm

Page 124

D.F. SWAAB

nucleus as early as 18–19 weeks of gestation (Reppert et al., 1988; Reppert, 1992; Thomas et al., 1998, 2002). In contrast, such binding is generally not detectable in the pars tuberalis of the pituitary of fetal human subjects and only rarely in adults. Melatonin binding was also detected in the pars distalis of several subjects, but with an inconsistent distribution (Weaver et al., 1993; Thomas et al., 2002). A family of three subtypes of melatonin receptors has been revealed: Mel1a (MT1), Mel1b (MT2) and MT3 (Reppert et al., 1996; Savaskan et al., 2002b). Two sub-types have been identified in the human with the Mel1a receptor mainly expressed in the SCN and the pars tuberalis of the pituitary and the Mel1b receptor expressed in the retina (Thomas et al., 1998). The Mel1a receptor has a size of 60 kDa and belongs to the superfamily of guanine nucleotide-binding regulatory protein (G-protein)-coupled receptors. Mel1a receptors couple functionally to both pertussus toxin-sensitive and insensitive G-proteins (Brydon et al., 1999). In the human melatonin 1a receptor gene, 7 mutations have been found, two of which were presumed to be related to non-24hour sleep-wake syndrome (Ebisawa et al., 2000). Two genetic polymorphisms have been detected in the human melatonin 1b receptor gene. However, neither of these missense mutations was likely to be associated with sleep disorders (Ebisawa et al., 2000). In the adventitia of the arteries in the hippocampus of Alzheimer patients, the vascular Mel1a receptor, which has been postulated to mediate the vasoconstrictive effects of melatonin, is increased (Savaskan et al., 2001, 2002b). This may be a response to the decreased melatonin levels in Alzheimer’s disease (see above). In vitro, melatonin acts directly as a free radical scavenger and neutralizes reactive oxygen and nitrogen species. It also stimulates antioxidative enzymes such as superoxide dismutase, glutathione peroxidase and glutathione reductase. Animal experiments indicate that melatonin may protect the eye lens from the damaging effects of ultraviolet exposure (Anwar and Moustafa, 2001). In cultured human retinal neurons, melatonin appeared to have antioxidant effects in a dose-dependent manner and to be able to rescue retinal neurons from injury caused by reactive oxygen species (Lee et al., 2001a). Free radicals have also been implicated in the pathogenesis of neonatal sepsis and its complications. Indeed, melatonin induced a decrease in serum levels of lipid peroxidation products in septic newborns and improved clinical outcome (Gitto et al., 2001). In addition, melatonin has anti-amyloidogenic properties (Papolla

et al., 2000; Reiter et al., 2000; Tan et al., 2000) and stimulates electron transport and ATP production in the inner mitochondrial membrane (Reiter et al., 2002). However, whether such effects may indeed protect the organism against the process of aging or Alzheimer’s disease remains to be proved. In Chinese traditional medicine plants are used that contain high levels of melatonin. Melatonin is thus a highly conserved molecule that is not only present in animals, but also in bacteria, unicellular organisms and plants (Chen et al., 2003). The MT1, but not the MT2, receptor is expressed in human breast tumor cell lines, and melatonin-induced growth suppression can be mimicked by MT1 and MT2 agonist and blocked by an antagonist (Ram et al., 2002). Epidemiological data show that blind people have half the rate of breast cancers. Yet the ‘melatonin hypothesis’ of cancer is still controversial (Rohr and Herold, 2002). The human myometrium expresses the MT1 and MT2 melatonin receptor isoforms that may modulate circadian myometrical function (Schlabritz-Loutzevitch et al., 2003). In humans, melatonin has been shown to inhibit vasopressin and oxytocin secretion, but there is controversy over these effects (Chiodera et al., 1998c). In fact, melatonin inhibits vasopressin at high doses and enhances the response at low doses. The administration of melatonin increases cortisol levels in postmenopausal women, but this effect completely disappeared during estrogen administration (Cagnacci et al., 1997). In mice, melatonin possesses powerful immune augmenting properties (Maestroni et al., 1988) and in human controls and asthma patients melatonin is proinflammatory (Sutherland et al., 2002). Melatonin is also effective in the treatment of seizures in children and adults (Rohr and Herold, 2002; Chapter 30.7). Melatonin is currently available in the USA as a “natural food supplement”. If melatonin is used, e.g. against jet lag or circadian rhythm disturbances in blind people (Palm et al., 1997) or in Alzheimer patients (Chapter 4.3), one should be aware of the fact that the timing of melatonin administration is important and that adverse drug reactions may occur, such as: (i) fever on the first day of melatonin treatment, which is possibly a reaction to the thermoregulatory function of melatonin, (ii) hyperkinesia or complaints of restless legs, (iii) menorrhagia, which may be explained by a decrease in plasma FSH and LH, (iv) pigmentations on arms and legs, (v) headache and abdominal reactions, such as nausea, dyspepsia and abdominal pain, (vi) thrombosis, and (vii) drowsiness

2014 Ch 04

22/1/04

12:18 pm

Page 125


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

(Nagtegaal et al., 1996; Avery et al., 1998). A pharmacological dose of melatonin (3.0 mg) to elderly people not only induced sleep but also induced hypothermia and caused melatonin plasma levels to remain elevated into the daylight hours. Moreover, intravenous administration of melatonin to schizophrenic patients in remission causes a worsening of psychotic symptoms which persists even after the treatment is interrupted (Pacchierotti et al., 2001). Since tumor growth was promoted in animal experiments, patients with non-hormone-dependent tumors like leukemia should avoid melatonin (Rohr

125

and Herold, 2002). The implications of the observation that the in vitro addition of melatonin to normal semen caused time- and dose-dependent inhibition of sperm motility (Luboshitzky and Lavie, 1999) should be further investigated in vivo. Pregnant women should avoid melatonin, since its (functional) teratological effects are not known. Apart from the concern about reproductive functions, exacerbation of epilepsy and withdrawn problems in psychiatric patients have been mentioned as possible side effects of melatonin that need attention.

2014 Ch 05

20/11/03

10:45 am

Page 127


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

CHAPTER 5

Sexually dimorphic nucleus of the preoptic area (SDN-POA) = intermediate nucleus = interstitial nucleus of the anterior hypothalamus (INAH-1) = preoptic nucleus

found in hypothalamic pathologies such as the Alzheimer changes in the infundibular nucleus (Chapters 11g; 29.1b), the commissura anterior and intrathalamic adhesion (Chapter 6.2), the functional activity of the hypothalamoneurohypophysial system (Chapter 8.d) and sex hormone receptor distribution (Chapter 6.5). The human homologue of the rat anteroventral periventricular nucleus (AVPV), which is larger in the female rat than in the male rat (Gu and Simerly, 1997), has not yet been identified. Due to differences in perinatal steroid levels, the SDNPOA in the male rat is 3–8 times larger than in the female rat (Jacobson et al., 1980). There are indications that the SDN-POA area is involved in aspects of sexual behavior. Electrical stimulation of the median preoptic area in the rat induces highly exaggerated, stimulation-bound sexual behavior (Merari and Ginton, 1975). In squirrel monkeys, electrical stimulation of the preoptic area elicited penile erection (MacLean and Ploog, 1962). Although mounting, intromission and ejaculation are eliminated after lesion of the medial preoptic, the animals do not lose the ability to achieve an erection (McKenna, 1998). While the median preoptic area does not seem to organize copulatory behavior, it is crucial for the recognition of sensory stimuli as appropriate sexual targets, and for the integration of this recognition with sexual motivation and copulatory motor programs. The paraventricular nucleus (PVN) receives extensive input from the medial preoptic area. The parvocellular oxytocinergic neurons of the PVN project to the spinal cord and synapse on neurons that innervate the penis (McKenna, 1998). It is, however, not clear what the exact role of the SDN-POA is in these functions of the preoptic area. The electrophysiological experiments in rat that showed increased multiple unit activity in the preoptic area during mounting, both in intact males and females (Kartha and

The hypothalamus as our sexiest part

Functional magnetic resonance imaging (MRI) revealed a significant activation of the right hypothalamus in males who were sexually aroused (Arnow et al., 2002). The hypothalamus contains a number of structurally sexually dimorphic structures (Fig. 5.1) that are presumed to be involved in sexual behavior, such as the suprachiasmatic nucleus (SCN; see Chapter 4.4), the sexually dimorphic nucleus of the preoptic area (SDN-POA), which was first described in the rat brain by Gorski et al. (1978), the interstitial nucleus of the anterior hypothalamus (INAH-2,3; Chapter 6), and the bed nucleus of the stria terminalis (BST) (Chapter 7). Sexual dimorphism was, in addition, 127

2014 Ch 05

128 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:45 am

Page 128

D.F. SWAAB

Fig. 5.1. Topography of the sexually dimorphic structures in the human hypothalamus. A is a more rostral view than B. Abbreviations: III, third ventricle; AC, anterior commissure; BNST-DSPM, darkly staining posteromedial component of the bed nucleus of the stria terminalis; Fx, fornix; I, infundibulum; INAH1-4, interstitial nucleus of the anterior hypothalamus 1-4; LV, lateral ventricle; OC, optic chiasm; OT, optic tract; PVN, paraventricular nucleus; SCN, suprachiasmatic nucleus; SDN, sexually dimorphic nucleus of the preoptic area = INAH-1; SON, supraoptic nucleus. Scale bar = 5 mm. The AC, BSTc, BNST-DSPM, INAH2,3,4, SCN and SDN vary according to sex. The SCN, INAH3 and AC are different in relation to sexual orientation.

Ramakrishna, 1996), do not clarify the exact contribution of the SDN-POA part to these activity changes. Although lesion experiments in rats indicated that the SDN-POA may be involved in aspects of male sexual behavior, i.e. mounting, intromission and ejaculation (Turkenburg et al., 1988; De Jonge et al., 1989), the effects of lesions on sexual behavior were only slight. Gorski (2002) reported that electrical stimulation of the male SDN-POA markedly enhanced sexual behavior and possibly also aggressive behavior. Penile erection following medial preoptic area stimulation in monkey and rat (MacLean

and Ploog, 1962; Giuliano et al., 1996) is also considered to be one of the putative functions by some authors. However, such a function does not tally with the observation that, following lesions of the medial preoptic area, rats do not lose the ability to achieve an erection (McKenna, 1998). The SDN-POA is a galanin-containing area and it is therefore presumed to be homologous to the ventrolateral preoptic nucleus in the rat and involved in sleep regulation (Gaus et al., 2002). This is not a very likely homologue, however, because the SDN-POA is localized more mediodorsal than the rat ventrolateral

2014 Ch 05

20/11/03

10:45 am

Page 129

SDN-POA

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

= INTERMEDIATE

NUCLEUS

preoptic nucleus. On the basis of some animal experiments, the SDN-POA area is presumed to be involved in sexual orientation. Lesions of the area of the SDN-POA in the ferret caused a significant shift in the males’ preference from estrous females to stud males, i.e., from a male-typical pattern of sexual behavior to a more female-typical pattern (Paredes and Baum, 1995; Kindon et al., 1996). Intact female rats show a preference for interaction with males, and males show a tendency to interact with females. After lesion of the medial preoptic area of the anterior hypothalamus (mPOA), the females’ preference was not modified. However, mPOA-lesioned male rats changed their partner preference and the coital behavior of these males was significantly reduced (Paredes et al., 1998). The human preoptic area also seems to be involved in responses to pheromones in a sexually dimorphic way (Savic et al., 2001). Although in our studies of the human SDN-POA we did not find differences in size or cell numbers of this nucleus in relation to sexual orientation (Swaab and Hofman, 1990; Fig. 4.29 and see below), this does not, of course, exclude a functional involvement of this structure in sexual orientation. 5.1. Nomenclature and homology to the rat SDN-POA The SDN-POA is located between the dorsolateral supraoptic nucleus (SON) and the mediorostral pole of the paraventricular nucleus (PVN), at the same rostrocaudal level as the suprachiasmatic nucleus (Figs. 5.1 and 5.2). The SDN-POA in the young adult human brain is twice as large in males (0.20 mm3) as in females (0.10 mm3 on one side) and contains twice as many cells (Swaab and Fliers, 1985). The original observations on 13 males and 18 females were extended and confirmed in a group of 103 subjects containing a reference group of 42 males and 38 females (Swaab and Hofman, 1988). The sexual dimorphism of the SDN-POA in the human has been confirmed by H. Braak, although findings have not been formally published (Braak and Braak, 1992, p. 14). The SDN-POA is also present in rhesus monkey (Braak and Braak, 1992). However, the fact that Byne (1998), who called this nucleus “the lateroanterior nucleus of the rhesus monkey” did not find a sex difference in its volume, makes this a complex comparison. Daniel and Prichard (1975) used the term “Preoptic Nucleus” for the human SDN-POA, but this name has not been used in the literature since. The SDN-POA is also identical to

= INAH-1 = PREOPTIC

NUCLEUS

129

the “Intermediate Nucleus” described by Brockhaus (1942) and Braak and Braak (1987a), and to the Interstitial Nucleus of the Anterior Hypothalamus-1 (INAH-1) of Allen et al. (1989a). The term “intermediate nucleus” is, however, controversial since Feremutsch (1948) called the scattered vasopressin or oxytocin cells and islands between the SON and PVN “intermediate nucleus” as well. How confusing the term “intermediate nucleus” is appears, e.g. from the paper of Morton (1969), who used this name, by mistake, also for the clusters of accessory SON cells, but now referring to the 1942 Brockhaus paper on the SDN-POA. Judging by the sex difference in the human SDN-POA in size and cell number, rostrocaudal position, cytoarchitecture, peptide and GABA content (see below), this nucleus is most probably homologous to the SDN-POA in the rat (Gorski et al., 1978), in spite of the fact that the rat SDN-POA is located in a more medial position than its human counterpart (Koutcherov et al., 2002). The SDN-POA contains galanin (Gai et al., 1990; Fig. 5.3) – named after its N-terminal residue glycine and its C-terminal alanine (Tatemoto et al., 1983) – galaninmRNA (Bonnefond et al., 1990; Gaus et al., 2002), thyrotropin-releasing hormone (TRH) neurons (Fliers et al., 1994; Fig. 5.4) and glutamic acid decarboxylase (GAD) 65 and 67 (Gao and Moore, 1996a,b). This supports the possible homology with the SDN in the rat, in which these peptide- and GABA-containing neurons have also been described (Bloch et al., 1993; Gao and Moore, 1996a,b). In addition, some scattered substanceP neurons are present in the human SDN-POA (Chawla et al., 1997) and moderate substance-P cell numbers were found in this area in the rat (Simerley et al., 1986). Although we think that, at present, all the data favor the homology between the human and the rat SDN-POA, and are supported in this by the study by Koutcherov et al. (2002), it should be noted that others have claimed, on the basis of the presence of a sex difference, a possible homology between the rat SDN-POA and INAH-3 (Allen et al., 1989; LeVay, 1991; Byne et al., 2000, 2001; Fig. 6.1). However, this claim did not take into account a homology in neuropeptide and GABA content of those nuclei (see below), nor the presence of benzodiazepine binding sites in this area, as found in the area containing the SDN-POA (Najimi et al., 2001). One of the main arguments against homology between the human and rat SDN-POA has been the location of the human SDN-POA in the lateral POA, in contrast to the more medial position of the rat SDN. In a very detailed developmental study, Koutcherov et al. (2002) have shown that, during fetal

2014 Ch 05

130 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:45 am

Page 130

D.F. SWAAB

Fig. 5.2. Thionine-stained frontal section (6 m) of the hypothalamus of (A) a 28-year-old man and (B) a 10-year-old girl. Arrows show the extent of the SDN-POA. Note the large blood vessel penetrating the SDN-POA and note that the male SDN is larger than that of the female. Bar represents 1 mm. (From Swaab and Fliers, 1985; Fig. 1, with permission.)

development, the human SDN-POA abuts the POA laterally at 16 weeks of pregnancy. This original of the human SDN from the POA supports the homology with the rat SDN-POA. In men, we recently observed a more intense staining for the androgen receptor and for the estrogen receptor in the SDN-POA than in women, further supporting the presence of a sex difference in this nucleus (Fernández-Guasti et al., 2000; Kruijver et al., 2002; Fig. 6.2; Tables 6.1 and 6.2). Concerning the presence of estrogen receptors in the SDN-POA, it is relevant to note that an estrogen response element has been found within the human galanin gene (Howard et al., 1997). It should also be noted that microinjection of galanin in the medial preoptic nucleus facilitates female-typical and male-typical sexual behaviors in the female rat (Bloch et al., 1996). The human SDN-POA has been reported to contain the second-step catecholamine synthesizing enzyme aromatic l-amino acid decarboxylase (AADC), but not tyrosine hydroxylase (TH) (D14; Kitahama et al., 1998a). In addition, hypocretin fibers are present in the medial preoptic area (Moore et al., 2001). 5.2. Development, sexual differentiation, aging and Alzheimer’s disease Sexual dimorphism does not seem to be present in the human SDN-POA at the time of birth. At that moment, total cell numbers are still similar in boys and girls, and the SDN-POA contains no more than some 20% of the

total cell number found between 2 to 4 years of age. From birth up to this age, cell numbers increase equally rapidly in both sexes (Fig. 5.5). The postnatal increase in cell numbers in the SDN-POA raises the question as to where they originate. The exhaustion of the matrix layer around the third ventricle – the site where hypothalamic cells are generally thought to be born (also paraphrased as “brain marrow”) – was reported to be complete in the fetus around 2 to 3 weeks of gestation (Staudt and Stüber, 1977). However, an alternative source for the cells may be the recently described subventricular zone in the ventral part of the third ventricle that may represent a zone even of adult neurogenesis in the human brain. The cells express a polysialylated embryonic form of neural cell adhesion molecule and -tubulin III, which is an early marker of neuronal determination (Bernier et al., 2000). A sex difference in the SDN-POA does not occur until about the 4th year postnatally, when cell numbers start to decrease in girls, whereas in boys the cell numbers in the SDN-POA remain stable until their rapid decrease at approximately 50 years of age. In females, a second phase of marked cell loss sets in after the age of 70 (Fig. 5.6; Swaab and Hofman 1988; Hofman and Swaab 1989). The sharp decrease in cell numbers in the SDN-POA later in life might be related to the hormonal changes that accompany both male and female senescence (Hofman and Swaab 1989; Chapter 24.1c), and to the decrease in male sexual activity around 50 years of age (Vermeulen, 1990).

2014 Ch 05

20/11/03

10:45 am

Page 131

SDN-POA

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

= INTERMEDIATE

NUCLEUS

= INAH-1 = PREOPTIC

NUCLEUS

131

Fig. 5.3. Galanin immunoreactivity in the human SDN-POA of a 17-year-old man (no. 97-152). F.P.M. Kruijver and E.A.M. Ligtenberg, unpublished data.

Fig. 5.4. Detail from the SDN-POA of a male subject, 72 years of age. TRH staining counterstained with hematoxylin-eosin. Note darkly stained TRH-positive cell (arrowhead) and moderate TRH fiber density. Bar = 50 m. (From Fliers et al., 1994; Fig. 4, with permission.)

However, it is not clear whether the hormonal changes are directly related to these changes in various functions, either as cause or as effect of the observed cell loss in this nucleus. The sex difference in the pattern of aging, and the fact that sexual differentiation in the human SDN-POA only occurs after the 4th year of age (Swaab and Hofman, 1988; Fig. 5.5) might explain why Allen et al. (1989a), who had a sample of human adults biased for aged individuals, did not find a significant sex difference in the size of the SDN-POA, which they called INAH-1. In the study of Allen et al., 40% of the adult subjects came from the age group in which the SDN-POA sex difference is

minimal compared with the 29% found in our study (Hofman and Swaab, 1989). Moreover, the age group of elderly subjects (over 70 years of age) was underrepresented in Allen’s study: 20% compared with the 37.5% that would be a proportional distribution of all ages. In our study, 32% of the subjects belonged to this old-age group. It therefore seems likely that Allen et al. (1989a) were unable to establish a sex difference in the INAH-1 (= SDN-POA) because they used a non-representative sample. A further argument for this assumption is that, if we, in our material, had studied only subjects of the age distribution of Allen’s study, the sex difference in SDN-POA volume would have been reduced from

2014 Ch 05

132 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:45 am

Page 132

D.F. SWAAB

Fig. 5.5. Developmental and sexual differentiation of the human sexually dimorphic nucleus of the preoptic area (SDN-POA) of the preoptic area of the hypothalamus in 99 subjects, log-log scale. Note that at the moment of birth the SDN-POA is equally small in boys (▲) and girls (❍) and contains about 20% of the cell number found at 2–4 years of age. Cell numbers reach a peak value around 2–4 years postnatally, after which a sexual differentiation occurs in the SDN due to a decrease in cell number in the SDN of women, whereas the cell number in men remains approximately unchanged up to the age of 50. The SDN-POA cell number in homosexual men (■) does not differ from that in the male reference group (▲). The curves are quintic polynomial functions fitted to the original data for males (full line) and females (dashed line). (Adapted from Swaab and Hofman, 1988; Fig. 1, with permission.)

a factor 2 (Hofman and Swaab, 1989) to only 1.4 times, and this difference would no longer have been statistically significant. Moreover, the sex difference in the SDN-POA merges only between the ages of 4 and puberty (Swaab and Hofman, 1988; Fig. 5.5); therefore the brain of the 5-year-old boy and 4-year-old girl (she indeed had by far the largest volume of the entire series of female INAH-1) also produced a bias in the Allen et al. (1989a) material. The age distribution, however, does not explain why LeVay (1991) and Byne et al. (2000) could not find a sex difference in the volume or neuron number (Byne et al., 2001) of INAH-1. Although the numbers of subjects they studied were much smaller than those in our study (Swaab and Hofman, 1988), technical differences such as section thickness may be a possible explanation for the controversy. Be that as it may, the finding that nuclear androgen and estrogen receptor staining in the SDNPOA was more intense in males than in females (Fernandez-Guasti et al., 2000; Kruijver et al., 2001, 2002;

Tables 6.1 and 6.2) supports the presence of a sex difference in this nucleus. A prominent theory is that sexual orientation develops as a result of an interaction between the developing brain and sex hormones (Gladue et al., 1984; Dörner,1988; Chapter 24.5). According to Dörner’s hypothesis, male homosexuals would have a female differentiation of the hypothalamus. Although LeVay’s (1991) data on the small, female-sized INAH-3 in homosexual men are in agreement with this theory (Chapter 6.1), this idea was not supported by our data on the SDN-POA in homosexual men. Neither the SDN-POA volume nor the cell number of homosexual men who died of AIDS differed from that of the male reference groups in the same age range, nor from that of heterosexuals also suffering from AIDS (Swaab and Hofman 1988, 1990; Figs. 4.29). The fact that no difference in SDN-POA cell number was observed between homo- and heterosexual men, along with the large SCN found in homosexual men (Swaab

2014 Ch 05

20/11/03

10:45 am

Page 133

SDN-POA

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

= INTERMEDIATE

NUCLEUS

= INAH-1 = PREOPTIC

NUCLEUS

133

Fig. 5.6. Age-related changes in the total cell number of the sexually dimorphic nucleus of the preoptic area (SDN-POA) in the human hypothalamus. The general trend in the data is enhanced by using smoothed growth curves. Note that in males SDN cell number steeply declines between the ages of 50 to 70 years, whereas in females a more gradual cell loss is observed around the age of 80 years. These curves demonstrate that the reduction in cell number in the human SDN in the course of aging is a nonlinear, sex-dependent process. (From Hofman and Swaab, 1989; Fig. 5, with permission.)

and Hofman, 1990), refutes the general formulation of Dörner’s (1988) hypothesis that male homosexuals would have “a female hypothalamus” and rather favors the idea that homosexual men are a “third sex”, i.e. different from heterosexual men and women. In Alzheimer’s disease – not in controls – SDN-POA neurons and dystrophic neurites are stained with cytoskeletal markers such as Alz-50, anti-tau, anti-paired helical filaments and anti-ubiquitin (Swaab et al., 1992b; Van de Nes et al., 1993). In spite of these pretangle Alzheimer changes and of the /A4-staining Congo-

negative amorphic plaques that were present in this nucleus, no difference was found between Alzheimer patients and controls as far as SDN-POA cell numbers were concerned (Swaab and Hofman, 1988), indicating that there is no direct relationship between the occurrence of these hallmarks of Alzheimer’s disease and cell death (see Chapter 29.1). AIDS has a small effect on SDN-POA volume. HIVpositive heterosexual men and women had an 8% increase in the volume of this nucleus compared to HIV-negative individuals (Byne et al., 2001).

2014 Ch 06

20/11/03

10:48 am

Page 135


1 2 3 4 5 6 7 8 9 111 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

CHAPTER 6

Other sexual dimorphisms

Moreover, LeVay (1991) found that INAH-3 was twice as large in heterosexual men than in homosexual men. The observation that INAH-3 was sexually dimorphic in size and neuron number was confirmed by Byne et al. (2000, 2001). However, in Byne’s studies the size of the INAH-3 in homosexual males showed only a trend to be smaller than in heterosexual males, and the number of neurons in INAH-3 from homosexual males did not differ from those of heterosexual males (Byne et al., 2001). INAH-2 in the human hypothalamus is said to correspond to the anterocentral nucleus in the rhesus monkey and INAH-3 to the dorsocentral portion of the anterior hypothalamic nucleus in rhesus monkey (Byne, 1998). Since nothing is known about their neurotransmitter content, it is at present not clear which nuclei in the rat (Fig. 6.1) or rhesus monkey are homologous to the human INAH-2 and -3. Recently we found galanin-containing cells and fibers, not only in INAH-1(=SDN-POA) but also in INAH-2. On the basis of serial sections, one could even raise the question whether INAH-1 and INAH-2 are indeed two separate nuclei or whether they are both part of a continuous horseshoe-shaped structure. As long as no chemical marker is known for INAH-3 it is not clear either whether this nucleus has to be considered as, for example, a perifornical cell group, an island of the paraventricular nucleus (PVN) (Koutcherov et al., 2002), part of the BST or as a separate anatomical entity. Although the nuclear organization of the human hypothalamus is more distinct in fetal development than in the adult, INAH-3 and -4 could not be distinguished during development in a recent study (Koutcherov et al., 2002). There is a discrepancy in the literature concerning the sex difference in the size of INAH-2 as described by Allen et al. (1989a) that could not be confirmed by LeVay (1991), nor by Byne et al. (2000, 2001). The fact that

6.1. Interstitial nucleus of anterior hypothalamus (INAH)-2 and -3 (Figs. 5.1 and 6.1) In addition to the sex differences observed in the suprachiasmatic nucleus (SCN) (Chapter 4.4), in the sexually dimorphic nucleus of the preoptic area (SDN-POA) (see Chapter 5) and in the bed nucleus of stria terminalis (BST) (see Chapter 7), and the functional sex differences in the supraoptic nucleus (SON) and paraventricular nucleus (PVN) (Chapter 8.d), Allen et al. (1989a) described two other cell groups (the interstitial nucleus of the anterior hypothalamus (INAH)-2 and -3) that were larger in the male brain than in the female brain (Figs. 5.1 and 6.1). 135

2014 Ch 06

136 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:48 am

Page 136

D.F. SWAAB

Fig. 6.1. Three-dimensional reconstructions of the medial preoptic-anterior hypothalamic continuum of the human (A) and rat (B). According to Byne, INAH3 in the human, like the SDN-POA of the rat, is a component of the MP-AHN. In contrast, the other INAH are situated outside the MP-AHN. For a different view (i.e. that the SDN-POA of the rat is homologous to INAH-1 in human), see Fig. 1.16 and Koutcherov et al., 2002. In the rat, an expansion of the ventricle (V) is seen behind the anterior commissure (ac). In the human, the region of the reconstruction did not extend through ac posteriorly. Reconstructions were prepared from thionine-stained serial sections with the assistance of Application Visualization System software (Advanced Visual Systems, Inc., Waltham, MA). Abbreviations: ac, anterior commissure; INAH, interstitial nucleus of the anterior hypothalamus; MP-AHN, medial preoptic-anterior hypothalamic nucleus; oc, optic chiasm; SDN-POA, sexually dimorphic nucleus of the preoptic area; SON, supraoptic nucleus. (From Byne et al., 2001; Fig. 2, with permission.)

LeVay did not observe a smaller INAH-2 in women was proposed to be explained by an age-related sex difference in this nucleus. INAH-2 shows this sex difference only after the child-bearing age, with one exception: a 44-yearold woman who had a hysterectomy with ovarian removal 3 years prior to her death and who had a small INAH-2 (Allen et al., 1989a). The sex difference in INAH-2 thus seems to come to expression only after menopause, when circulating estrogens are absent. This would also explain why LeVay (1991) could not confirm the difference in INAH-2 in his group of young patients. The sex difference in INAH-2 was considered to be the first human example of a sex difference depending on circulating levels of sex hormones, i.e. a difference based upon a lack of activating effects of sex hormones in menopause rather than an organizing effect of sex hormones in development. However, Byne et al. (2000, 2001) could not confirm the relationship between INAH-2 and reproductive status that was suggested by the data of Allen et al. (1989a). A second example of a functional sex difference can be found in the supraoptic and paraventricular nucleus (see

Chapter 8.d and 6.3). INAH-4 is not sexually dimorphic (Allen et al., 1989a). 6.2. Anterior commissure, the interthalamic adhesion, corpora mamillaria and the third ventricle The anterior commissure appears as early as gestational day 47 (Hori, 1997) and contains about 7 million fibers. It anteriorly connects two-thirds of the right and left temporal neocortices and the posterior part of the orbital aspect of the frontal lobes. Anatomical variants of the anterior commissure have been observed, such as the accidental finding of an asymmetrical commissure whose left wing had shifted in a posterior direction and passed behind the left columna fornicis (Hori, 1997). The anterior commissure was investigated in humans who had circumscribed hemispheric lesions. The largest contingent of commissural axons appeared from this study to originate in the inferior part of the temporal lobe. In addition, axons originating

2014 Ch 06

20/11/03

10:48 am

Page 137

OTHER SEXUAL DIMORPHISMS

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

from the inferior part of the occipital lobe, occipital convexity, central fissure and prefrontal convexity were found to cross the anterior commissure. The anterior commissure mediates interhemispheric transfer of visual information, including the visual recall of dreams, and auditory and olfactory information (Risse et al., 1978; Martin, 1985; Allen and Gorski, 1991; BotezMarquard and Botez, 1992). There is also a small olfactory part rostrally (Sylvester, 1986). This may at least partly explain the loss of axons, reactive astrogliosis and dystrophic calcifications found in the anterior commissure of patients with Wolfram’s syndrome (Dean et al., unpublished results), since these patients have an olfactory bulb and tract atrophy (Chapter 22.7). Degeneration accompanied by a destruction of myelinated fibers occurs in the anterior commissure in Marchiafava– Bignami disease, a rare complication of alcoholism (see Chapter 29.5; Victor, 1994; Moreau et al., 1996). In a case of septo-optic dysplasia in Cornelia de Lange syndrome, the anterior commissure was rudimentarily present (Hayashi et al., 1996; Chapter 32.2), and a hypoplastic anterior commissure was found in a mentally retarded patient (Shaw, 1987). In a variant of Shapiro’s syndrome, abnormalities of the corpus callosum, hypoplasia of the anterior commissure and absence of the septum pellucidum and columns of the fornix were found. It concerned a woman with episodic sweating and shivering with reduced core temperature (Klein et al., 2001). A novel brain malformation is characterized by an absence of the anterior commissure without callosal agenesis, but associated with gross unilateral panhemispheric malformations, incorporating subependymal heterotopia, and gyral abnormalities, including temporal malformation and polymicrogyria. The cause is currently not known. In addition, absence of the anterior commissure without callosal agenesis was observed in subjects who were heterozygous for mutations in the PAX6 gene. These patients did not have gross neocortical abnormalities (Mitchell et al., 2002). The anterior commissure was found to be 12% larger in women than in men (Allen and Gorski, 1991). In an earlier study only a trend towards such a sex difference was noted. The anterior commissure tended to be larger in men in that study. Recently Byne et al. (2001, 2002) could not replicate a sex difference in the cross-sectional area of the anterior commissure. However, Highley et al. (1999) found a greater number of fibers in women than in men. Their study suggested

137

that the cross-sectional area of the anterior commissure is a poor predictor of the number of fibers it conveys. This may be an explanation for at least some of the controversies in the literature. A sex difference in the size of the commissure was also observed in Down’s syndrome subjects (Sylvester, 1986). Allen and Gorski (1992) found that the anterior commissure was larger in homosexual men than in heterosexual men, but this finding could not be confirmed by Byne et al. (2002). Sexual dimorphism in the anterior commissure is presumed to underlie sex differences in cognitive skills, developmental language disorders and functional asymmetries. In rats, the anterior commissure is also larger in females than in males. Prenatal stress – known to disrupt both sexual differentiation and sexual behavior – leads to an absence of the sex difference in the size of the anterior commissure (Jones et al., 1997), which indicates that organizing effects of sex hormones during development are instrumental in the sex difference of this structure. The anterior commissure has been reported to be smaller than normal in trisomy 21 (Sylvester, 1986) and larger than normal in patients with agenesis of the corpus callosum. The anterior commissure can be absent in patients with callosal agenesis, trisomy 18, occipital encephalocele or holoprosencephaly (Golden, 1998). In schizophrenia, where there appears to be a reduction in brain asymmetry, a reduction in density of the fibers in the anterior commissure was found in female but not in male patients (Highley et al., 1999). Allen and Gorski (1991) found that the interthalamic adhesion or massa intermedia, a gray structure that crosses the third ventricle between the two thalami, was present in more females (78%) than males (68%), confirming the old study of Morel (1947). In patients with schizophrenia, there is a sex-by-diagnosis interaction in the absence of the massa intermedia. Female schizophrenic patients had a significantly higher incidence of absent massa intermedia (33%) than healthy controls (14%), whereas in male patients no difference in absence of this structure was found (Nopoulos et al., 2001). The latter lack of a difference in male patients was confirmed later (Meisenzahl et al., 2002). Males have larger corpora mamillaria than females (Sheedy et al., 1999). In a comparison between 10 control women and 8 men, all with a mean age of 79 years, MRI measurements showed that the third ventricle volume was 67% larger in men than in women (Wahlund et al., 1993).

2014 Ch 06

138 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:48 am

Page 138

D.F. SWAAB

Fig. 6.2. Schematic representation of the sex differences in the intensity of androgen-receptor immunoreactivity in the human hypothalamus. Abbreviations: Ox: optic chiasma, NBM: nucleus basalis of Meynert, hDBB: horizontal limb of the diagonal band of Broca, SDN: sexually dimorphic nucleus of the preoptic area, SCN: suprachiasmatic nucleus, BST: bed nucleus of the stria terminalis, PVN: paraventricular nucleus, SON: supraoptic nucleus, DPe: periventricular nucleus dorsal zone, VPe: periventricular nucleus ventral zone, fx: fornix, 3V: third ventricle, ac: anterior commissure, VMN: ventromedial hypothalamic nucleus, INF: infundibular nucleus, OT: optic tract, MB: mamillary body, i.e. MMN: medial mamillary nucleus + LMN: lateromamillary nucleus, cp: cerebral peduncle. (From Fernández-Guasti et al., 2000; Fig. 2.)

2014 Ch 06

20/11/03

10:48 am

Page 139


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

139

Fig. 6.3. Photomicrographs showing androgen-receptor immunoreactivity in various areas in the anterior part of human hypothalamus (a–e). Note the clear sex difference in the SCN (a,b) and the hDBB/Ch3 (d,e). The SDN revealed sexually dimorphic labeling with a medium to weak intensity, while the area surrounding this nucleus was only weakly stained. (c) SDN of a female hypothalamus. Bar (a,b,c) = 1400 m and bar (d,e) = 600 m. (From Fernández-Guasti et al., 2000; Fig. 3, with permission.)

2014 Ch 06

20/11/03

10:48 am

Page 140

140 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

D.F. SWAAB

TABLE 6.1 Median intensity of label for AR-ir in various hypothalamic brain structures. Hypothalamic area

Men n

DBB (Ch2) NBM Ch4) hDBB (Ch3) BST SDN-mPOA mPOA DPe VPe SCN PVN SON VMN INF LMN MMN Calleja islands

Women c

++ (44) + (5/5) +++ (5/5) + (5/5) ++ (5/5) + (5/5) + (5/5) + (5/5) ++ (5/5) ++ (4/5) + (3/5) ++ (5/5) ++ (5/5) +++ (5/5) +++ (5/5) + (5/5)

+ (3/4) + (5/5) + (4/5) – (4/5) + (3/5) + (3/5) – (5/5) – (5/5) + (4/5) – (2/5) + (3/5) – (2/5) + (3/5) + (4/5) + (4/5) – (5/5)

n

c

+ (3.5) + (5/5) ++ (5/5) + (5/5) + (3/5) – (2/5) – (5/5) – (5/5) + (3/5) – (2/5) – (1/5) + (5/5) + (3/5) ++ (5/5) + (5/5) + (5/5)

– (2/5) + (4/5) + (4/5) – (2/5) – (2/5) – (2/5) + (5/5) + (5/5) – (2/5) – (2/5) – (1/5) + (3/5) – (2/5) ++ (3/5) + (4/5) – (5/5)

The category assigned to a given brain region corresponds to the predominant cell type according to the following scale: – = no staining; + = staining diffuse and transparent, ++ = staining nontransparent but individual granules of the reaction product still distinguishable, and +++ = intense opaque stain. Proportions in parentheses indicate number of patients stained/total number. AR-ir: androgen receptor immunoreactivity. n: nuclear staining; c: cytoplasmic staining. For other abbreviations see Table 6.2.

6.3. Sex hormone receptor distribution The brain is our biggest sexual organ. A pity it is hidden in the skull.

In most hypothalamic areas that contain androgenreceptors, staining, nuclear staining in particular, is less intense in women than in men (Figs. 6.2, 6.3 and Table 6.1). The strongest sex difference was found in the lateral and the medial mamillary nucleus (Fernández-Guasti et al., 2000; Fig. 6.5). The mamillary body complex is known to be involved in several aspects of sexual behavior (see Chapter 16). In addition, a sex difference in androgen-receptor staining was present in the horizontal diagonal band of Broca (Fig 6.3), SDN-POA, medial preoptic area, the dorsal and ventral zone of the periventricular nucleus, PVN, SON, ventromedial hypothalamic nucleus (VMN) (Fig. 6.4) and the infundibular nucleus. No sex differences were observed in androgen-receptor staining in the BST, the nucleus basalis of Meynert (NBM) and the island of Calleja (Fernández-Guasti et al., 2000). Nuclear androgen-receptor activity in the mamillary complex of heterosexual men did not differ from that

of homosexual men, but it was significantly stronger than in women (Fig. 6.6). A female-like pattern was found in men with low testosterone levels, e.g. in a 26-year-old and a 53-year-old castrated man, and in intact old men. These data indicate that the amount of nuclear-receptor staining in the mamillary complex is dependent on the circulating levels of androgens, rather than on gender identity or sexual orientation. This idea is supported by the finding that a male-like pattern of androgen-receptor staining was found in a 36-year-old bisexual noncastrated male-to-female transsexual, and in a heterosexual virilized woman of 46 years of age (Kruijver et al., 2001). Various sex differences were observed for estrogen receptor- (ER) staining in the hypothalamus and adjacent areas of young human subjects (Fig. 6.7). More intense nuclear ER immunoreactivity (-ir) was found in young men as compared to young women in neurons of the medial part of the bed nucleus of the stria terminalis (BSTm), the SDN-POA, the SON, the PVN, the dorsal periventricular zone (Dpe) and the lateral hypothalamic area (LHA). Women revealed a stronger nucleus ER-ir in the diagonal band of Broca (DBB/Ch2), SCN, VMN,

2014 Ch 06

20/11/03

10:48 am

Page 141


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 6.4.

141

Photomicrographs showing androgen-receptor immunoreactivity in the ventromedial hypothalamic (VMN) and infundibular nucleus (INF). Bar (a) = 600 m and bar (b) = 1400 m. (From Fernández-Guasti et al., 2000; Fig. 4., with permission.)

2014 Ch 06

142 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:49 am

Page 142

D.F. SWAAB

Fig. 6.5. Photomicrographs showing androgen-receptor immunoreactivity in the mamillary body (MB) and lateromamillary nucleus (LMN). Notice the conspicuous sexual dimorphism: (a) and (c) represent strong AR-immunoreactivity in the male MB and LMN compared with weak staining in the female (b,d). Example of AR-immunopositive neurons are shown at higher magnification (*). Note the strong immunoreactivity within the nucleus with no staining in the nucleolus as well as the presence of weak cytoplasmic labeling. Bar (a–d) = 1400 m. (From Fernández-Guasti et al., 2000.)

2014 Ch 06

20/11/03

10:49 am

Page 143


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 6.6. Photomicrographs showing AR-ir in neurons of the MMN of the mamillary body of a heterosexual man (A), a heterosexual woman (B), a homosexual man (C), and a woman with high levels of androgens (D). Note that in the mamillary body there is a clear sex difference in AR-ir (see A and B), whereas there is no difference in the intensity of AR staining between the representative heterosexual man (A), the homosexual man (C) and the virilized (androgenized) woman. (D) Scale bar, 150 m. (From Kruijver et al., 2001; Fig. 1, with permission.)

and medial mamillary nucleus (MMN). No sex differences in nuclear ER staining were found in, e.g. the lateral septum (LS), the central part of the BST (BSTc), the islands of Calleja or in the infundibular nucleus (INF). Sex differences in cytoplasmic staining with a stronger staining in men were found in the BST, the SCN, the NBM, the INF, the tuberomamillary complex (TM) and the lateromamillary nucleus (LMN). An ovariectomized 46-year-old female subject, a castrated and estrogen-treated 50-year-old male-to-female

143

transsexual, and a 31-year-old male subject with high estrogen levels due to an estrogen-producing tumor, revealed ER distribution patterns according to their level of circulating estrogens in most areas, suggesting that the majority of the reported sex differences in ER-ir are “activating” rather than “organizing” in nature (Kruijver et al., 2002). In general, ER-immunoreactivity (ER-ir) was observed more frequently in the cytoplasm than in the nucleus, with a stronger staining in women in the NBM, hDBB and TM and in men in the medial preoptic area (MPO) (Fig. 6.8 and Table 6.2). A more intense nuclear ER staining of a low to intermediate level was found, in men, in neurons of the BSTc, the BSTm, the islands of Calleja, the SDNPOA, the DBB/Ch2, and the VMN, as well as the paratenial nucleus (PT) and the paraventricular nucleus of the thalamus. Women revealed more nuclear ER of a low to intermediate level in the SCN, the SON, the PVN, the INF, the nucleus tuberalis lateralis and the MMN. ER-ir was not only observed in neurons but also in endothelial cells and perivascular smooth muscle cells. Interestingly, a striking ER-ir was observed in fibers of the internal capsule and in the BSTc, while in the latter structure also a “basket-like” neuronal staining pattern suggestive of nerve-terminal appositions was observed. An ovariectomized 46-year-old female subject, a castrated and estrogen-treated 50-year-old male-to-female transsexual and a 31-year-old male subject with high estrogen levels due to an estrogen-producing tumor revealed, in most areas, ER-ir distribution patterns according to their level of circulating estrogens, suggesting that the majority of the reported sex differences in ER-ir are “activational” rather than “organizational” in nature. The presence of ER and - in the hypothalamus on the mRNA level has been reported in the VMN, INF, SON and PVN. In contrast to ER mRNA, expression of the subtype was generally very low in these areas (Österlund and Hurd, 2001). Sex differences in receptor distribution may be agerelated. The dorsolateral SON is the main production site of plasma vasopressin. Plasma vasopressin levels and the activity of vasopressin neurons are higher in men than in premenopausal women. On the other hand, an increased activity of vasopressin neurons becomes prominent in postmenopausal women with strongly decreased estrogen levels (Chapter 8.d). As estrogens are presumed to inhibit vasopressin production in a receptor-mediated way,

2014 Ch 06

144 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:49 am

Page 144

D.F. SWAAB

Fig. 6.7. Schematic representation of the intensity of nuclear estrogen receptor (ER) staining in the hypothalamus of men and women between 20 and 40 years of age. Note the presence of region-dependent sex differences. (Kruijver et al., 2002.) For abbreviations see Table 6.2.

2014 Ch 06

20/11/03

10:49 am

Page 145


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 6.8.

145

Schematic representation of the intensity of nuclear estrogen receptor (ER) staining in the hypothalamus of men and women between 20 and 40 years of age. Note the presence of region-dependent sex differences. For abbreviations see Table 6.2.

2014 Ch 06

20/11/03

10:49 am

Page 146

146 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

D.F. SWAAB

TABLE 6.2 Distribution and sex differences in nuclear (n) ER-ir versus nER-ir. (Peri)-hypothalamic-area

Men ER n

Women ER n

ER n

ER n

Preoptic region DBB (CH2) NBM (CH4) hDBB (CH3) Cal LS MS BSTm BSTc BSTl CdM ic EGP SDN-mPOA mPOA DPe VPe SCN PVN SON

+ (5/5) – (5/5) – (5/5) + (4/5) ++ (5/5) n.d. ++ (5/5) + (5/5) – (5/5) – (5/5) – (4/5) – (5/5) + (5/5) – (5/5) + (5/5) + (5/5) + (5/5) – (5/5) – (5/5)

– (5/5) + (5/5) ++ (5/5) ++ (4/5) ++ (5/5) n.d. ++ (5/5) + (5/5) + (5/5) + (5/5) + (4/5) – (5/5) ++ (5/5) + (5/5) ++ (5/5) ++ (5/5) ++ (5/5) ++ (5/5) ++ (5/5)

– (4/5) – (5/5) – (5/5) - (4/5) ++ (4/5) n.d. – (5/5) – (5/5) – (5/5) – (5/5) – (5/5) – (4/5) – (5/5) – (5/5) + (5/5) ++ (5/5) ++ (5/5) + (5/5) + (5/5)

++ (4/5) + (5/5) ++ (5/5) ++ (4/5) ++ (4/5) n.d. + (5/5) + (5/5) + (5/5) + (5/5) + (5/5) – (4/5) + (5/5) – (5/5) + (5/5) ++ (5/5) +++ (5/5) + (5/5) + (5/5)

++ (5/5) ++ (5/5) + (5/5) – (5/5) + (5/5) – (5/5) – (5/5) – (5/5) – (5/5)

++ (5/5) ++ (5/5) ++ (5/5) + (5/5) + (5/5) ++ (5/5) + (5/5) ++ (5/5) ++ (5/5)

– (4/5) – (4/5) + (5/5) – (5/5) – (5/5) + (5/5) + (5/5) – (4/5) – (5/5)

++ (4/5) ++ (4/5) ++ (5/5) + (5/5) ++ (5/5) ++ (5/5) + (5/5) ++ (4/5) ++ (5/5)

– (4/5) + (4/5)

++ (4/5) +++ (4/5)

Tuberal region PT PV BSTP DMN VMN INF NTL TM LHA

Mamillary region LMN MMN

– (5/5) – (5/5)

++ (5/5) + (5/5)

Legend. Median intensity of nuclear label for ER-ir and ER-ir in various hypothalamic and adjacent brain structures. The category assigned to a given brain region corresponds to the predominant cell type according to the following scale: – = no staining, + = staining diffuse and transparent, ++ = staining non-transparent but individual granules of the reaction product still distinguishable, and +++ = intense opaque staining. n = nuclear staining. N.d: not determined. Proportions in parentheses indicate number of patients stained/total number. The areas in bold point to reversed sex differences for ER-ir and ER-ir. The hypothalamus is subdivided into its three main regions, which are the preoptic, tuberal and mamillary region according to most authors. (Saper, 1990.)

2014 Ch 06

20/11/03

10:49 am

Page 147


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

147

TABLE 6.2 Continued. Abbreviations 3V ac Ac AR-ir Astr BLv BST BSTc BSTI BSTm BSTp CAL Cdm chp cp DBB DMN DPe EGP ER fx hDBB INF INAH-1 ic ithp LHA LMN

third ventricle anterior commissure nucleus accumbens androgen receptor immunoreactivity astrocytes blood vessel bed nucleus of the stria terminalis central BST lateral BST medial BST posterior BST Calleja islands medial caudate nucleus choroid plexus cerebral peduncle diagonal band of Broca dorsomedial hypothalamic nucleus periventricular nucleus, dorsal zone external globus pallidus estrogen receptor fornix horizontal limb of the diagonal band of Broca (Ch3) infundibular (arcuate) nucleus interstitial nucleus of the anterior hypothalamus-1 internal capsule inferior thalamic peduncle lateral hypothalamic area lateromamillary nucleus or nucleus intercalatus

we studied ER and - immunoreactivity in the dl-SON. The vasopressin part of the dl-SON of young women contained 50 times more neurons with ER nuclear staining than that of young men, and 250 times more than that of elderly women (Fig. 8.12). In addition, young women also showed more ER cytoplasmic staining than young men and elderly women. In contrast to the ER-ir, no differences were found in the number of ERpositive neurons in the 4 groups, but the age and sex pattern of ER staining was basically the opposite of that of ER. Significant correlations between the percentage of ER and --positive and negative vasopressin neurons

LS LV MB MBC ME MMN mPOA NBM NBMC NTL ot OVLT ox PMN PT PV PVN SCN SDN-POA SON ST TM TMvp/E3 and TM/E3 VMN VPe

lateral septum lateral ventricle mamillary body mamillary body complex median eminence medial nucleus of the mamillary body medial preoptic area nucleus basalis of Meynert (Ch4) nucleus basalis of Meynert complex (containing DBB, hDBB, and NBM) nucleus tuberalis lateralis optic tract organum vasculosum lamina terminalis optic chiasm premamillary nucleus paratenial nucleus of the thalamus paraventricular nucleus of the thalamus paraventricular nucleus suprachiasmatic nucleus sexually dimorphic nucleus of the preoptic part supraoptic nucleus stria terminalis tuberomamillary nucleus/complex tuberomamillary complex, ventral perimamillary part ventromedial hypothalamic nucleus periventricular nucleus, ventral zone

and age were found in women, but not in men. These data demonstrate a strong decrease in ER and an increase in ER-ir in vasopressin neurons of the dl-SON of postmenopausal women. Both receptor changes are proposed to participate in the activation of the vasopressin neurons in postmenopausal women (Ishunina et al., 2000b). The ER gene is located on chromosome 6q25-1. Putative associations have been reported between ER polymorphisms and personality traits such as anxiety, conduct disorder, non-conformity (including “indirect aggression and irrationality”) and psychotism (including “suspicion”) (Westberg et al., 2003).

2014 Ch 07

20/11/03

10:55 am

Page 149


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

CHAPTER 7

Bed nucleus of the stria terminalis (BST) and the Septum

estrogens. The BST in the rat receives projections from the amygdala, paraventricular and periventricular nuclei, the parabrachial nucleus, the nucleus of the solitary tract dorsal vagal nucleus and from mesencephalic structures (see below) (Kozicz et al., 1998), and provides a strong input to the preoptic-hypothalamic region. Reciprocal connections between the hypothalamus, BST and amygdala are well documented in experimental animals (Zhou et al., 1995c; Liu et al., 1997b). There is a strong innervation of galanin fibers in the BST and galanin receptors have also been shown in this structure (Mufson et al., 1998). The BST and centromedial amygdala have common cyto- and chemoarchitectonic characteristics, and these regions are considered to be two components of one distinct neuronal complex. Neurons in the substantia innominata form cellular bridges between the BST and amygdala (Lesur et al., 1989; Martin et al., 1991; Walter et al., 1991; Heimer, 2000; Figs. 7.1 and 7.5). In most mammals, including human, the extended amygdala presents itself as a ring of neurons encircling the internal capsule and the basal ganglia (Heimer et al., 1997). The BST–amygdala continuum contains, e.g. luteinizing hormone-releasing hormone (LHRH) neurons (Rance et al., 1994). May et al. (1998) found neurophysin containing neurons from 20 weeks of gestation onwards in the BST, and in the adult human BST vasopressin is found (Fliers et al., 1986). The BST also contains aromatic l-amino acid decarboxylase (AADC), but, according to some, no tyrosine hydroxylase (TH) (D14; Kitahama et al., 1998). However, TH mRNA has been found by others in this structure (Gouras et al., 1992). Five principal sectors have been identified in the BST (Figs. 7.1 and 7.2): (i) a lateral nucleus (Walter et al., 1991) or lateral sector (Lesur et al., 1991) with neuropeptide-Y cells and fibers and substance-P fibers;

7.1. The BST The BST is situated at the junction of hypothalamus, septum and amygdala (Lesur et al., 1989; Walter et al., 1991; Figs. 7.1 and 7.2). It plays an essential part in rodent sexual behavior (Liu et al., 1997b) and participates in certain types of anxiety and stress responses (Walker et al., 2003). Androgen and estrogen receptors have been found in the human BST (Fernández-Guasti et al., 2000; Kruijver et al., 2002, 2003; Tables 6.1 and 6.2) and it is a major aromatization center in the developing rat brain, i.e. converting androgens into 149

2014 Ch 07

150 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:55 am

Page 150

D.F. SWAAB

(ii) a central nucleus (Zhou et al., 1995c; Fig. 7.3) also known as the supracommissural part of the central nucleus of the BST (Walter et al., 1991) or central sector (Lesur et al., 1989) or BSTLD (BST dorsal part) (Martin et al., 1991). The central nucleus is sheathed in myelinated fibers and characterized by a high density of somatostatin neurons and fibers (Kruijver et al., 2000; Fig. 7.4) and VIP innervation (Zhou et al., 1995c; Fig. 7.3). This innervation was presumed to originate from the amygdala. However, in the rat it was established that the VIP innervation of the BST comes from the mesencephalic periaqueductal gray and the dorsal and linear raphe nuclei. In relation to the sex differences in the BSTc (Chapter 7.2), it is therefore of interest that the median raphe nucleus contains more neurons in women than in men (Cordero et al., 2000). The VIP fibers are involved in the modulation of the stress response and reach the BST from the mesencephalon via the ventral amygdalofugal pathway and the stria terminalis passing by the amygdala (Kozics et al., 1998). In addition, enkephalin cells and fibers and neurotensin cells are found;(iii) a medial nucleus (Walter et al., 1991) or medial sector (Lesur et al., 1991) with VIP innervation (Zhou et al., 1995c; Fig. 7.3), and a less dense aminergic and peptidergic (i.e. substance-P) and scarce enkephalin and neuropeptide-Y innervation; and (iv) a lateroventral nucleus (Walter et al., 1991) or lateroventral sector (Lesur et al., 1989) where somatostatin (Fig. 9.2) and enkephalin plexuses are prominent and where neurophysins are present (Lesur et al., 1989; Walter et al., 1991; Zhou et al., 1995c); in addition, (v) a “darkly staining posteromedial component (dspm) of the BST” was distinguished by Allen et al. (1990). This part of the BST is situated in the zone that lies dorsolaterally the fornix (Fig. 5.1) and is sexually dimorphic (Chapter 7.2). This sex difference does not seem to occur before adulthood. Its chemical composition and relationship to the other 4 principal BST sectors (see above) is unknown. Moreover, the BST contains numerous neurokinin-B neurons, neuropeptide-Y binding sites and diazepine binding sites and only a few substance-P neurons, but the subnuclei where these peptides were present were not specified (Chawla et al., 1997; Dumont et al., 2000; Najimi et al., 2001). The BST is densely innervated by CART (cocaine and amphetamine-regulated transcript)containing nerve fibers. All subdivisions of the BST display a prominent staining of secretoneurin, a 33-amino acid neuropeptide produced by endoproteolytic processing

from secretogranin II (Kaufmann et al., 1997). Although the BST contains nuclear androgen receptors, no sex difference was observed in staining of this receptor (Fernández-Guasti et al., 2000; Fig. 6.2). Estrogen receptor (ER) is present in the BSTc but no sex difference was observed (Kruijver et al., 2002; Fig. 6.7); while more nuclear ER was found in men than in women in the medial and central BST (Kruijver et al., 2002; Fig. 6.8). Interestingly, a striking ER-ir was observed in fibers of the internal capsule (IC), stria terminalis and BSTc, which went together with a selective ‘basket-like’ neuronal staining pattern suggestive of nerve terminal appositions (Kruijver et al., 2002a,b; Chapter 6.3). The stria terminalis makes a dorsally convex detour behind and above the thalamus (Fig. 7.5). In the flow of the lateral ventricle, it accompanies the thalamostriate vein. Cell groups along the arch of the stria terminalis, the supracapsula bed nucleus of the stria terminalis, provide important evidence for Heimer’s “extended amygdaloid” concept (Heimer et al., 1999). These cell groups are said to form more prominent continuities between the lateral and medial bed nuclei and the central and medial amygdaloid nuclei in early development. In Alzheimer’s disease, /A4-staining Congo-negative amorphic plaques (Van de Nes et al., 1998) and Alz-50positive dystrophic neurites and cell bodies are found (Van de Nes et al., 1993; Figs. 7.6 and 9.2), indicating its involvement in Alzheimer pathology. In dementia with argyrophilic grains, a modest amount of grains are found in the BST (Braak and Braak, 1987a; 1989) and in Parkinson’s disease the BST is also affected (Braak and Braak, 2000). In schizophrenia, increased levels of norepinephrine were found in the BST (Farley et al., 1978). In the subependymal region in the vicinity of the BST, large numbers of corpora amylacia are often observed (Cavanagh, 1999), just like in the subpial region of the substantia innominata (see Chapter 2.6). 7.2. Reversed sex differences in the BST in transsexuals The woman shall not wear that which pertaineth unto a man, neither shall a man put on a woman’s garment: for all that do so are abomination unto the Lord thy God. Deuteronomy 22:5

The volume of the BST-dspm is 2.5 times larger in males than in females (Allen et al., 1990). We have found that the central nucleus of the BST (the BSTc; Fig. 5.1), which

2014 Ch 07

20/11/03

10:55 am

Page 151

BED NUCLEUS OF THE STRIA TERMINALIS AND THE SEPTUM

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

151

Fig. 7.1. Schematic drawings of four different section levels in the frontal plane (a–d) based on brain A 58 of the Vogt collection. The subdivisions of the bed nucleus of the stria terminalis (BNST) are displayed in black. In (c) and (d) ventrolateral extensions are symbolized by black spots in the substantia innominata: thin lines should indicate their connection to the BNST, which can only be realized by viewing a sequence of consecutive sections (Walter et al., 1991; Fig. 1, with permission). CC cdm ci CL CO DB Fpt Fsvm

corpus callosum nc. caudatus medialis capsula interna claustrum chiasma opticum diagonal band fundus putaminis fundus subventricularis medialis

GPe ITm ITs NS prp ptm VL

globus pallidus externus insula terminalis magna insula terminalis substriatales nuclei septi prepiriform cortex putamen mediale ventricularis lateralis

2014 Ch 07

152 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:55 am

Page 152

D.F. SWAAB

Fig. 7.2. Diagrams of the BST on four representative verticofrontal levels spaced by 500 m, from rostral (I) to caudal (IV). Column (a) indicates the anatomic parcellation proposed; column (b) represents the distribution of somatostatin-like immunoreactivity (SST) LIR; column (c) the distribution of Met-enkephalin (MET-E)-like immunoreactivity. Full triangles are for labeled perikarya; open circles indicate basket-like pericellular fibers; curved thick lines indicate the peridendritic plexuses of the granular type; immunolabeled terminal-like networks are depicted by small dots; stippling indicates the ‘pipe-like’ ENK-IR peridendritic plexuses fibers, characteristic of the globus pallidus. Abbreviations: Ac: anterior commissure; Acc: nucleus accumbens; Cd: caudate nucleus; Fx: fornix; GP: globus pallidus; GPv: ventral globus pallidus; GPe: external GP; GPi: internal GP; IC: internal capsule; Lat V: lateral ventricle: LS: lateral septum; mfb: medial forebrain bundle; MS: medial septum; NDB: nucleus of the diagonal band of Broca; NBM: nucleus basalis of Meynert; NSTa: rostral extension of the bed nucleus of the stria terminalis; pr: preoptic area; st: stria terminalis (Lesur et al., 1989; Figs. 3 and 4).

2014 Ch 07

20/11/03

10:55 am

Page 153


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 7.2.

Continued.

153

2014 Ch 07

154 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:55 am

Page 154

D.F. SWAAB

Fig. 7.3. Representative sections of the BSTc innervated by vasoactive intestinal polypeptide (VIP). A. heterosexual man; B. heterosexual woman; C. homosexual man; D. male-to-female transsexual. Scale bar, 0.5 mm. LV, lateral ventricle. Note there are two parts of the BST in A and B: small medial subdivision (BSTm) and large oval-sized central subdivision (BSTc) Note also the sex difference (A vs. B) and the fact that the male-to-female transsexual (D) has a female BSTc in size and type of innervation (from Zhou et al., 1995c; Fig. 2, with permission).

2014 Ch 07

20/11/03

10:55 am

Page 155


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

155

Fig. 7.4. a–d: representative immunocytochemical stainings of the somatostatin neurons and fibers in the BSTc of a reference man (a), reference woman (b), homosexual man (c), male-to-female transsexual (d). There is a clear sex difference with male-to-female transsexuals having a BSTc in the female range. * = blood vessel. Bar represents 0.35 mm (from Kruijver et al., 2000; Fig. 2, with permission).

2014 Ch 07

156 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:55 am

Page 156

D.F. SWAAB

Fig. 7.5. The extended amygdala (in color) shown in isolation from the rest of the brain, with the extensions of the central (Ce) and medial (Me) amygdaloid nuclei alongside the stria terminalis (st) and through the sublenticular region to the bed nucleus of stria terminalis (BST). The central division of extended amygdala is color-coded in yellow and the medial division in green. The supracapsular part of the bed nucleus of stria terminalis is depicted as a continuum, although the neuronal cell bodies of especially the medial division (green) do not form a continuous column (see text). Associated dendrites and neuropil, however, are likely to form a continuous columnar structure within the stria terminalis. Note that the laterobasal complex of the amygdala (lateral, basolateral, basomedial and paralaminar amygdaloid nuclei) and cortical amygdaloid nuclei are not included as part of the extended amygdala (Art by Medical and Scientific Illustration, Crozet, Virginia). From Heimer et al., 1999; Fig. 1, with permission.

was defined by its dense VIP innervation (Fig. 7.3) or by its somatostatin fibers and neurons (Fig. 7.7), is sexually dimorphic. The BSTc is 40% smaller in women than in men (Fig. 7.7) and contains some 40% fewer somatostatin neurons (Fig. 7.8). No relationship was observed between BSTc volume or somatostatin cell number and sexual orientation: in the heterosexual reference group and a group of homosexual males a similar BSTc volume and somatostatin cell number were observed. The size and somatostatin cell number of the BSTc were, moreover, not influenced by abnormal hormone levels in adulthood. However, a remarkably small BSTc (40% of the male reference volume and somatostatin neuron number) was observed in a group of 6 male-to-female transsexuals (Figs. 7.7 and 7.8). These data suggest that the female size of this nucleus in male-to-female transsexuals was established during development and that the BSTc is part of a network that might be involved in gender, i.e. the feeling of being either male or female (Zhou et al., 1995c; Kruijver et al., 2000; Chapter 24.5c). In order to determine at what age the BSTc becomes sexually dimorphic, the BSTc volume in males and females was studied from mid-gestation into adulthood. Using vasoactive intestinal polypeptide and somatostatin immunocytochemical staining as markers, we confirmed that the BSTc was larger and contained more neurons in men than in women. Unexpectedly, this difference became significant only in adulthood, showing that sexual differentiation of the human brain may extend into adulthood (Fig. 7.9). There are several possible explanations for the lack of a sex difference in the BSTc shortly after fetal or neonatal sex differences in testosterone levels emerge. Organizational effects of testosterone may become overt much later in life, as has been shown for the sexually dimorphic anteroventral periventricular nucleus in the rat brain. Alternatively, peripubertal sex hormone levels may have an effect, or sexual differentiation may depend on a late sexually dimorphic innervation of the BSTc. Sex steroid-independent mechanisms cannot be excluded either. The discrepancy between the late structural sexual differentiation of the BSTc and the early occurrence of gender problems in transsexualism raises the question whether these two are indeed directly related and if they are, how? It is of course possible that functional sex differences in the BSTc, e.g. in synaptic density, neuronal activity or neurochemical content precede the structural sex differences in the course of development (Chung et al., 2002).

2014 Ch 07

20/11/03

10:55 am

Page 157


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

157

Fig. 7.6. (a) Alzheimer patient, 40 years of age, SP28 (somatostatin)/Alz-50 (hyperphosphorylated tau) staining. Anti-A staining in the tuberal gray (TG) shows a low degree of mainly large-sized, intensely stained amorphous plaques (ap). The amount of Alz-50-stained perikarya (p) and dystrophic neurites (dn) is low. There is no clear intimate relationship between amorphous plaques and cytoskeletal changes. (b) Alzheimer patient, 56 years of age, anti-A (#1G102) staining. Staining of amorphous plaques in the tuberal gray (TG) is characterized by large amounts of often fairly granular and small-sized deposits (continued on next page). (c) Alzheimer patient, 45 years of age, SP28 (somatostatin)/Alz-50 (hyperphosphorylated tau) staining. The central part of the bed nucleus of the stria terminalis (BSTc) can be easily delineated from the lateral sector of the bed nucleus (indicated with an asterisk). Arrows indicate faint-granular /A4-reactive guirland-like /A4 deposits in the rim of the nucleus. Amorphous plaques in the BSTc of AD patients are often more spheroid and more intensively stained with anti-A. Cytoskeletal changes are not observed. (d) Alzheimer patient, 70 years of age, anti-A- staining (#1G102). Again, the central sector of the bed nucleus of the stria terminalis (BSTc) is obviously distinguishable from the surroundings, as indicated by arrows. The BSTc shows some /A4-reactive deposits only in the periphery of the nucleus. The surrounding nuclei are as follows: 1 refers to the caudate nucleus, 2 to the lateral and 3 to the medial part of the bed nucleus of the stria terminalis. CG: chiasmatic gray (from Van de Nes et al., 1998; Fig. 2, with permission).

2014 Ch 07

158 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:55 am

Page 158

D.F. SWAAB

Fig. 7.7. Volume of the BSTc innervated by VIP fibers in presumed heterosexual males (M), homosexual males (HM), presumed heterosexual females (F) and male-to-female transsexuals (TM). The six transsexuals are numbered T1–T6. The patients with abnormal sex hormone levels are numbered S1–S4. M1 and M2, postmenopausal women. Bars indicate mean ± S.E.M. Open symbols: individuals who died of AIDS. Note the sex difference in BSTc volume, that the volume is not affected by abnormal sex hormone levels in adulthood, and the female volume found in male-to-female transsexuals (from Zhou et al., 1995c; Fig. 3, with permission).

7.3. The Septum Verum The human septum is a telencephalic structure bordered by the corpus callosum, the lateral ventricles, the subcallosal gyrus, the nucleus accumbens, the interstitial nucleus of the stria terminalis, the preoptic region of the anterior hypothalamus and the anterior commissure. The midline region of the septum, the septal recess, displays a particularly intense staining for subventricular zone markers, suggesting a zone of adult neurogenesis in the human brain (Bernier et al., 2000). The lateral septal nucleus borders the BST. The septum consists of two main parts, (i) the, phylogenetically new, dorsally located plate-like septum pellucidum that is situated below the corpus callosum and between the lateral ventricles; it contains myelinated fibers, glial cells and in its caudal part some neurons; and (ii) the ventral area or septum verum (Andy and Stephan, 1968). The septal area is innervated by the processes of the terminal and vomeronasal fibers (Bossy, 1980; Schwanzel-Fukuda and Pfaff, 1994; Chapter 24.2c). The septum verum is 14–16 mm long and its largest dorsoventral diameter is 12 mm and its thickest part is 8.2 mm

2014 Ch 07

20/11/03

10:55 am

Page 159


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

159

Fig. 7.8. BSTc neuron numbers. Distribution of the BSTc neuron numbers among the different groups according to sex, sexual orientation and gender identity. M (heterosexual male reference group), HM (homosexual male group), F (female group), TM (male-to-female transsexuals). The sex hormone disorder patients S1,2,3,5,6 and M2 indicate that changes in sex hormone levels in adulthood do not change the neuron numbers of the BSTc. The difference between the M and the TM group (< 0.04) becomes also statistically significant according to the sequential Bonferonni method if S2, S3 and S5 are included in the M group or if S7 is included in the TM group (p ≤ 0.01). Note that the number of neurons of the female-to-male transsexual (FMT) is fully in the male range. A = AIDS patient. The BSTc number of neurons in the heterosexual man and woman with AIDS remained well within the corresponding reference group, so AIDS did not seem to affect the somatostatin neuron numbers in the BSTc. P = Postmenopausal woman. S1 ( 25 years of age): Turner syndrome (45,X0; ovarian hypoplasia). M2 ( 73 years of age): postmenopausal status (from Kruijver et al., 2000; Fig. 1).

wide (Horváth and Palkovits, 1987). The 4 septal nuclei and their subdivisions and the vertical band of Broca (Chapter 2), which extends into the septum, are summarized in Table 7.1 and in Fig. 7.10. The typical giant cells of the vertical band of the diagonal band of Broca (DBB) are generally not considered to be part of the septal nuclei, although the DBB extends into the medial septal nucleus (Ulfig, 1989). The diagonal band of Broca in experimental animals contains temperature-sensitive neurons. The DBB Ch1 and Ch2 neurons project through the fornix to the hippocampus (Mesulam et al., 1983; Chapter 2)

and this system is involved in memory processes. In the monkey it was shown that the medial septal nucleus projects to the hippocampus and that this septohippocampal pathway contains GABA-ergic axons (Gulyás et al., 1991). The septal area, and more specifically its LHRH neurons present in this region (Dudas et al., 2000), are considered to be involved in temperature regulation and the vasodilatative effects of LHRH may be related to the etiology of menopausal hot flushes (Hosomo et al., 1997; Chapter 11f). LHRH neurons often colocalize with delta sleep-inducing peptide (Vallet et al., 1990), which may

2014 Ch 07

160 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:55 am

Page 160

D.F. SWAAB

Fig. 7.9. BSTc development in males and females. (A, C) BSTc volume as delineated by its VIP innervation. (B, D) BSTc volume as delineated by its somatostatin innervation. (E) Total number of BSTc neurons in males and females in adulthood. Note that the sex difference in volumes only develops after puberty (from Chung et al., 2002; Fig. 2, with permission).

also have central effects. Retrograde tracing studies in the rat have shown that at least some LHRH neurons of the septal nuclei project to the median eminence and may thus be involved in reproductive functions (Silverman et al., 1987). Intense sexual disinhibition was observed following the placement of the tip of a ventriculoperitoneal shunt into the septum (Miller et al., 1986; Gorman and Cummings, 1992; Frohman et al., 2002). The observation that electrical stimulation of the medial septal region in squirrel monkeys elicited penile erections

(MacLean and Ploog, 1962) supports the involvement of this area in primate reproduction. The LHRH neurons of the septopreoptic area originate in development from the olfactory pit (Chapter 24.2c). LHRH immunoreactivity is found in the developing olfactory pit at 5 weeks gestation. By 10 weeks of gestation, the bulk of the cells have travelled into the nasal cavity and by 20 weeks migration of LHRH neurons into the septopreoptic area is complete (MacColl et al., 2002).

2014 Ch 07

20/11/03

10:55 am

Page 161


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Table 7.1 Nuclei and subdivisions of the human septum. Nuclei and subdivisions of the medial area Medial septal nucleus pars dorsalis pars ventralis pars fimbrialis pars intermedia pars posterior Triangular nucleus Nucleus of the diagonal band (of Broca), vertical part Nuclei and subdivisions of the lateral area Dorsal septal nucleus Lateral septal nucleus pars anterior pars dorsalis pars ventralis From Horváth and Palkovits, 1987.

Fig. 7.10. BD C CI D DE DI DM F

= = = = = = = =

The dorsal part of the lateral septal nucleus and the vertical limb of the DBB have vasopressin binding sites and the ventral part of the lateral septal nucleus has oxytocin binding sites (Loup et al., 1991). Some of the medial and lateral septal nuclei and the DBB contain nestin. The exact function of these intermediate filaments in the mature neurons is not clear (Gu et al., 2002). Moreover, benzodiazepine binding sites are found in the septal area (Najimi et al., 2001). The medial-dorsal part of the septum contains some tyrosine hydroxylase-containing neurons (Dudás and Merchenthaler, 2001). Some neuritic plaques containing amyloid may be found in the diagonal band of Broca and in the septal nuclei in Alzheimer’s disease (Rudelli et al., 1984). In the septal area of Alzheimer patients, the diagonal band of Broca stained much less intensely for choline acetyltransferase and acetylcholinesterase (Henke and Lang, 1983). Although the entire nucleus basalis complex seems

Frontal views of the human septum, brain AC, 7.2, fiber stain.

Nucleus of the diagonal band of Broca Bed nucleus of the anterior commissure Capsula interna Nucleus septalis dorsalis Nucleus septalis dorsalis pars externa Nucleus septalis dorsalis pars interna Nucleus septalis dorsalis pars intermedius Nucleus septalis fimbrialis

From Andy and Stephan, 1968, plate 3, with permission.

161

Fx GS L MP NA PM SW T

= = = = = = = =

Fornix Gyrus subcallosus Nucleus septalis lateralis Nucleus septalis medialis pars posterior Nucleus accumbens Nucleus praeopticus medianus Septal warts Bed nucleus of the stria terminalis

2014 Ch 07

162 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

10:55 am

Page 162

D.F. SWAAB

to be heavily affected in Alzheimer’s disease (Chapter 2.3), and the septal area in Alzheimer patients is atrophied as determined by MRI (Callen et al., 2001), the Ch1-2 area, which includes the medial septal nucleus, seems to be spared. The total number of neurons is not changed in the Ch1-2 area in Alzheimer’s disease; only the cell size is diminished. One may wonder whether the Ch1-2 area is relatively spared, because the neurotrophin receptors, nerve growth factor and NT3, 4/5 levels in its main projection area, the hippocampus, are not decreased (Vogels et al., 1990; Salehi et al., 1998b; Hock et al., 2000). In autism, neuropathology was reported in the medial septal nucleus (Kemper and Bauman, 1998). The ventral part of the septum contains increased norepinephrine levels in schizophrenia. Unusually large septal nuclei were found in New Guinea cases of Kuru (Beck and Gajdusek, 1966); however,

microscopic investigation did not reveal any pathological changes. Later it appeared that the size of the septal nuclei of New Guinea patients who did not die of Kuru were no different from those of European brains, supporting the possible relationship between Kuru and large septal nuclei. Tumors in the septal area may be associated with outbursts of temper and violence (Albert et al., 1993; Chapter 26.9). In addition, vasopressin is an antipyretic peptide acting on the ventral septal area in experimental animals (Kasting et al., 1989). Pathologies of the septum pellucidum are described in Chapter 18.8, including the cavum septum pellucidum, which may be related to neurobiological and psychiatric disorders such as schizophrenia (Chapter 27.1) or boxing injuries, and the cavum vergae. Absence of the septum pellucidum is observed in, e.g. septo-optic dysplasia (Chapter 18.3), and tumors may be present in the septum pellucidum.

2014 Ch 08

20/11/03

11:03 am

Page 163


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

CHAPTER 8

Supraoptic and paraventricular nucleus (SON, PVN)

system (HNS), which represents the classic example of a neuroendocrine system. A second type of neuroendocrine cells are the PVN cells, which release their peptides into the portal capillaries that transport them to the anterior pituitary (Chapter 8.1c). A third type of PVN cells projects on other neurons, where the peptides act as neurotransmitters/neuromodulators (Chapter 8.1f). Due to its shape and localization along the optic tract, the SON was formally called the “tangential nucleus” (Cajal, 1911; Gurdijan, 1927). The SON and PVN are supplied with unusually rich capillary beds. The density of this capillary bed was reported to decrease with age in the PVN, but not in the SON (Abernethy et al., 1993). The decrease in the PVN is, however, rather unexpected and needs confirmation, as the number of oxytocin neurons is not changed in aging (Wierda et al., 1991), and corticotropin-releasing hormone (CRH) and vasopressincontaining PVN neurons were even seen to be activated (Raadsheer et al., 1994a,b; Van der Woude et al., 1995; Ishunina et al., 1999; Chapters 8.3 and 8.5). An agerelated increase in capillary bed in the PVN would thus rather have been expected. In order to establish the proportion of SON and PVN cells that project to the neurohypophysis, Morton (1969) determined neuronal numbers in these nuclei for a period of 12–45 months following hypophysectomy, an operation performed in those days as a palliative measure in the treatment of hormone-dependent metastatic mammary carcinoma. After hypophysectomy there was an average loss of neurons from both the SON and PVN of over 80%. Following hypophysectomy or transsection of the stalk, it took until about a year after the operation for the stump of the stalk to be innervated again throughout (Daniel and Prichard, 1972; Chapter 25.4). From these observations it was concluded that most neurons of the human SON and

One of my anatomist friends who has made notable contributions to our knowledge of the anterior lobe hormones has challenged me to produce any corresponding clinical or experimental evidence of posterior lobe activity. To this challenge, this and the succeeding papers are a partial answer. H. Cushing, 1932, p. 60

(a) The hypothalamoneurohypophysial system The supraoptic and paraventricular nucleus (SON and PVN) (Fig. 1.7) and their axons running to the neurohypophysis form the hypothalamoneurohypophysial 163

2014 Ch 08

164 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 164

D.F. SWAAB

PVN project to the neurohypophysis, where they would release vasopressin and oxytocin as neurohormones (see 8.1b). In addition, Golgi studies have revealed very small neurons in the human SON that are most probably interneurons (Al-Hussain and Al-Jomard, 1996). In view of the apparent lack of retrograde changes in the accessory SON cells following hypophysectomy, it seems likely that the axons of these islands between the SON and PVN project more proximally to the stalk of the pituitary (Morton, 1969; Chapter 25.4), to the median eminence region, or elsewhere in the brain. In addition, vasopressin from the PVN is released into the portal capillaries and triggers corticotropin (ACTH) release from the anterior lobe of the pituitary (René et al., 2000). As part of this stressregulating system, vasopressin is colocalized with CRH in an activity-dependent way (Raadsheer et al., 1993, 1994b; Chapter 8.5). This pathway is hyperactive in depression (Chapter 26.4d). The SON is subdivided into three parts, i.e. the dorsolateral, dorsomedial and ventromedial (Fig. 8.1 and Table 8.1). The largest part, the dorsolateral SON, has a volume of 3 mm3 (Goudsmit et al., 1990) and contains 53,000 neurons, 90% of which contain vasopressin and 10% oxytocin (Fliers et al., 1985). The oxytocin neurons are mainly localized as a cap on top of the dorsolateral SON (Fig. 8.2). Indeed, J. Purba (unpublished results) counted 49,240 vasopressin and 5,460 oxytocin neurons in this part of the SON. The dorsomedial and ventromedial SON together contain some 23,000 neurons (Morton, 1969, 1970). The ventromedial part is also called the postchiasmatic SON (Brockhaus, 1942). According to Dierickx and Vandesande (1977), 85% of the neurons of the medial part of the SON contain vasopressin and 15% oxytocin. The entire SON thus contains some 78,000 neurons on one side (Morton, 1969). The PVN (Fig. 1.5) has a volume of 6 mm3 (Goudsmit et al., 1990) and was estimated to consist of about 56,000 neurons (Morton, 1969) of which some 25,000 contain oxytocin and 21,000 express vasopressin (Wierda et al., 1991; Purba et al., 1993; Van der Woude et al., 1995; Table 8.1). A rostrocaudal gradient in the ratio between oxytocin and vasopressin in cells is present in the PVN (Fig. 8.7). The estimate of the exact neuron numbers in the SON and PVN depends strongly, however, on the methods used (Harding et al., 1995). For instance, the high prevalence of multinucleated neurons in the SON of young patients with pulmonary pathology (Ishunina et al., 2000a), may have led to an overestimation of total cell number in studies in which a deconvolution method was used to

Fig. 8.1. Coronal section of the human hypothalamus, showing the three parts of the supraoptic nucleus: ventromedial (VM), dorsomedial (DM), and dorsolateral (DL). The former two are joined by a band of cells. OT, optic tract; P, paraventricular nucleus; 3V, third ventricle (courtesy of Dr. Walter Freeman, Sunnyvale, California). (Nauta and Haymaker, 1969; Fig. 4.6, p. 144.)

measure the size of nuclear profiles (Goudsmit et al., 1990). Using nucleoli as unique marker will give a similar problem in such multinucleated cells. Of course the history of the patients included may also explain part of the variability in neuron numbers (see, e.g. Harding et al., 1996). The 6-m sections of the human PVN we generally use do not show the clear topographic division in subnuclei as is found in the rat. The absence of a clearcut arrangement of the PVN into subnuclei has also been observed in the cow, cat and guinea pig (Raadsheer et al., 1993; Swaab et al., 1995), and the populations of magno- and parvicellular neurons are not clearly separated. Quite a number of intermediate-sized cells are present in the PVN (Fig. 8.25). However, recently the cyto- and chemoarchitecture of the human PVN was studied on serial 50-m sections with the aid of three-dimensional computer reconstruction (Koutcherov et al., 2000). Chemoarchitecture revealed the following five subnuclei in the human PVN (Fig. 8.3). The most prominent one is the paraventricular (Pa) magnocellular subnucleus (PaM), which occupies the ventrolateral quadrant of the PVN, which comprised a concentration of large arginine vasopressin (AVP)- and acetylcholinesterase (AChE)-positive cells, and smaller calbindin (Cb)-positive neurons. Rostrally, the PaM is succeeded by the small anterior parvicellular subnucleus (PaAP), which contains small AChE-, AVP- and tyrosine

2014 Ch 08

20/11/03

11:03 am

Page 165

165

SUPRAOPTIC AND PARAVENTRICULAR NUCLEUS

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

TABLE 8.1 Neuron numbera

Proportion vasopressin : oxytocin

Supraoptic nucleus Total subnuclei:

77,700

88% : 12%

Dorsolateral part

54,700

90% : 10%

J. Purba, unpubl. results

Dorsomedial part plus ventromedial part

23,000

85% : 15%

Morton, 1969, 1970

Paraventricular nucleus

56,000

54% : 46%

Morton, 1969; Wierda et al., 1991; Purba et al., 1993; Van der Woude et al., 1995.

a

Source

One sided.

hydroxylase (TH)-positive cells. Dorsal to the PaM, the dorsal subnucleus (PaD) is found, containing large, spindle-shaped TH-, oxytocin (OXY)-, and AChE-positive cells, as well as a population of small Cb-positive neurons. Abutting the wall of the third ventricle and medial to PaM and PaD, is the parvicellular subnucleus (PaP). The PaP contains small cells immunoreactive (IR) for CRH, neuromedin K receptor (NK3), and nonphosphorylated

neurofilament protein (SMI32). The posterior subnucleus (PaPo) is situated posterior to the descending column of the fornix; it replaces all above-mentioned subdivisions caudally and is a chemoarchitectonic amalgam that includes dispersed large AChE-, OXY-, AVP- and THpositive cells, as well as small NK3-, CRF-, SMI32- and Cb-IR neurons. This description suggests that the human PaM and PaPo correspond to the rat medial parvicellular and posterior subnuclei, respectively. (b) Neurosecretion In the 1940s practically everybody vigorously, or even viciously, rejected the concept of neurosecretion. (Bertha Scharrer, letter to DFS, 1984)

Fig. 8.2. Consecutive sections of a 49-year-old female control stained for vasopressin and oxytocin. (a) dorsolateral supraoptic nucleus (SON) stained with an antiglycopeptide (Boris Y-2) against the vasopressin precursor and (b) oxytocin (0-1-V, purified). Note that the relatively small oxytocin cell population is clearly separated from the vasopressin cell population. Asterisk indicates a blood vessel that is present in both in consecutive sections: OC = optic chiasm. Bar = 100 m. (From Evans et al., 1996; Fig. 1, with permission.)

For a long time it was believed that the neurohypophysis made a product and that this function was regulated by the innervating nerves. In 1908, P.T. Herring first described the “peculiar hyaline bodies” seen in sections of the neurohypophysis and expressed the belief that what we now call “Herring bodies” represented the secretory product of the epithelial investment of the posterior lobe known as the pars intermedia. When Collin (1928) found stained droplets in the hypothalamus of the guinea pig, droplets similar to the secretory material found in the neural lobe, he suggested that this material was transported to the hypothalamus. Indeed, the “Herring bodies” are accumulations of neurosecretory material. However, Cushing’s description was that the globules appeared to find their way toward the tuber cinereum and “in favourable histological preparations could be seen passing between the bodies of the ependymal cells to enter

166

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03 11:03 am

Fig. 8.3. A model of the human paraventricular hypothalamic nucleus (Pa) showing the subnuclei delineations based on the Magellan 3.1-generated plots of six histochemical markers (AChE, SM132, Cb, TH, NK3, CRF) and verified with the distribution of AVP, OXY and NPH. Individual subnuclei are illustrated by different colors. The stereotaxic coordinates were borrowed from Mai et al. (1997a). The third ventricle wall lies to the right. (Koutcherov et al., 2000; Fig. 1.) (For explanation of the abbreviations see text, Chapter 8.1a.)

2014 Ch 08 Page 166

D.F. SWAAB

2014 Ch 08

20/11/03

11:03 am

Page 167


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

the infundibular cavity” (Cushing, 1932, p. 59). A famous case report tells of a man who, in 1910, received a gunshot wound, resulting in a strong polyuria and “sexual dystrophy”. At autopsy the bullet was found to be lodged in the sella-turcica and to have destroyed the infundibular process and posterior lobe (Vonderahe, 1940; Brooks, 1988). In 1921, Starling and Verney showed that a posterior pituitary extract given to patients with diabetes insipidus caused a strong antidiuretic effect on an isolated kidney and concluded that the hypotonic urine of the isolated kidney preparation was almost certainly due to the absence of an antidiuretic substance. Although such observations revealed the antidiuretic function of the neurohypophysis, the concept that “neurosecretion” in vertebrates was found in the large neurons of the human supraoptic and paraventricular nucleus was not proposed until 1939, by the Scharrers (Scharrer and Scharrer, 1940; Brooks, 1988; Meites, 1992). According to the critics of those days, this concept was based on “nothing more than signs of pathological processes, postmortem changes or fixation artifacts”. In the 1940s “practically everybody vigorously or even viciously” rejected the concept that a neuron could have a glandular function (B. Scharrer, personal communication, see also citation above, Chapter 8.1c). The initially highly charged rejection of the neurosecretion concept was followed by acceptance only when Bargmann (1949) demonstrated the same Gomori-positive material in the neurohypophysis and in the neurons of the SON and PVN and concluded that the axons from the SON and PVN transport material to the neurohypophysis. He called the aggregate of fibers “the neurosecretory pathway”. Seeking a new method for revealing neurosecretory material, he (Bargmann) placed sections from a dog’s brain into acid-permanganate-chrome-alum hematoxylin, according to Gomori’s method . . . and was astonished at what he found (instead of shouting Eureka, Bargmann, waving the cigar that was always in his hand exclaimed (magna voice): “Donnerwetter!”), the cells of the supraoptic and paraventricular nuclei and the fibers extending into the infundibulum and reaching the posterior lobe had selectively taken on a blue hue! . . . When, everafter, in papers dealing with hypothalamic neurosecretion “Gomori-positive” and “Gomori-negative” results were cited, Gomori would comment in conversation that he found the terms distasteful but amusing. “Right now” he once said before lunch to a fellow-Hungarian, Jacob Furth: “I feel Gomori-negative”. (Anderson and Haymaker, 1974)

167

In 1955, Dr. Vigeaud received the Nobel Prize for the elucidation of the chemical nature of vasopressin and oxytocin and the synthesis of vasopressin. The nonapeptides vasopressin and oxytocin are synthesized in the hypothalamus as part of a large precursor that includes a neurophysin for both peptides and a C-terminal glycoprotein for the vasopressin precursor only (Bahnsen et al., 1992; Chapter 22.2). The vasopressin and oxytocin precursor genes are only separated by 12 kb in the human genome, located on the distal short arm of chromosome 20, and are transcribed towards each other (Schmale et al., 1993; Evans, 1997). When the neurophysins were discovered by Asher and co-workers, they were recognized as the inactive fragments of the precursor, with a higher molecular weight, and were proposed to act as “carriers” for vasopressin and oxytocin. The role of neurophysins is at present considered in the light of the knowledge on mutations in the neurophysin part that cause diabetes insipidus (see Chapter 22.2). A disruption of the structure and thus of the threedimensional conformation of neurophysins by mutations (Fig. 22.3) may cause a decline in the binding and activity of endopeptidases responsible for the cleavage of vasopressin. Mutations in the neurophysins may also produce a change in their polymerization and salt bridges and thus in their intracellular trafficking, resulting in an accelerated, aspecific enzymatic degradation of the hormone accumulation in an organelle or degeneration of the neuron, revealing the clinical symptomatology. So, rather than being a mere inactive part of the precursor, neurophysins are considered an essential system for carrying and protecting the nonapeptides (Legros and Geenen, 1996; Chapter 22.2). The importance of neurophysins for intracellular trafficking is supported by observations on, e.g. a Dutch family with hereditary hypothalamic diabetes insipidus, based upon a single G to T transversion in the neurophysin-encoding exon B (Bahnsen et al., 1992). When this mutant DNA was stably expressed in a mouse pituitary cell line, the mutant precursor was synthesized, but processing and secretion were dramatically reduced and the protein did not seem to reach the trans-Golgi network (Olias et al., 1996), which is localized in the perinuclear region of these neurons (e.g. Fig. 8.9). Other studies on reversal mutant AVP genes showed accumulation of mutant AVP precursors in the endoplasmic reticulum, which is located in the peripheral part of the cytoplasm in these neurosecretory neurons (e.g.

2014 Ch 08

168 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 168

D.F. SWAAB

Fig. 8.6) and which reduced viability of the cell lines (Ito et al., 1997; Chapter 22.2). The oxytocin and vasopressin receptors form a subfamily within the much larger superfamily of G-protein-coupled receptors. The vasopressin (V)-1 receptor, formally known as V1a, is expressed in the liver, blood vessels, smooth muscle cells and most other peripheral tissues that express vasopressin receptors. The V2 receptor is expressed in the kidney and the brain of newborn rats, while the V3 receptor (formally known as V1b) is expressed in the majority of the anterior pituitary corticotropin cells, in multiple brain regions and in a number of peripheral tissues, including kidney, thymus, heart, lung, spleen, uterus and breast (Burbach et al., 2001). The V2 receptor gene has been assigned to chromosome Xq28 and the V3 receptor gene to the chromosome band 1q32. (c) Vasopressin and oxytocin: production and release The evidence that such cells secrete colloid and are to be considered a ‘diencephalic gland’ is morphological evidence and does not deserve acceptance at this time. H.B. Van Dyke, 1939; Scharrer, 1975

In 1957, in a paper of 130 pages (summarizing 14 years of work), Verney concluded that osmo-receptors regulating the release of antidiuretic hormone (= vasopressin) are located in the hypothalamus, probably in the anterior or preoptic areas. Connections with the SON and PVN need to be intact to enable hormone release (for references see Sawin, 2000). Likely candidates for the function of osmosensor in the hypothalamus are the SON neurons themselves and the astrocytes that are in intimate contact with the SON cells. It has been proposed that the water channel aquaporin 4, located in the perivascular glial endfoot processes, may play a central role in this function (Venero et al., 2001). In addition, osmoreceptors are present in the organum vasculosum lamina terminalis and subfornical organ (Chapter 30.5a). Vasopressin is synthesized in the SON, the PVN and some accessory cells (Fig. 8.4) and released after dehydration (Husain et al., 1973) or other types of osmotic stimulation (Valloton et al., 1983; Pederson et al., 2001). The function of vasopressin is to reduce the rate of urine flow by increasing the readsorption of solute free water in the distal and collecting tubules of the kidneys (Robertson, 2001). In humans, vasopressin secretion in response to dehydration is under the stimulatory influence

Fig. 8.4. Photomicrograph depicting AVP-mRNA signal on film in the SON and PVN of (A) a control, age 85, and (B) an Alzheimer patient, age 87. Bar represents 1 mm. S: SON, P: PVN. (From Lucassen et al., 1997; Fig. 1.)

of histamine from the tuberomamillary nucleus (Chapter 13), mediated by H2 receptors (Kjaer et al., 2000). In addition to the SON and PVN, vasopressin is synthesized in the suprachiasmatic nucleus (SCN) (Swaab et al., 1985; Chapter 4), while some vasopressin cells are found in the diagonal band of Broca (DBB), nucleus basalis of Meynert (NBM) (Ulfig et al., 1990; Chapter 2), and BST (Fliers et al., 1986; Mai et al., 1993; Chapter 7). Oxytocin is produced in the PVN (Fig. 8.5), the accessory nuclei and the dorsal part of the SON (Dierickx and Vandesande, 1979; Evans et al., 1996; Fig. 8.2). Mai et al. (1993)

2014 Ch 08

20/11/03

11:03 am

Page 169


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

pointed to an additional oxytocin-containing cell group dorsolateral to the fornix, which they refer to as the parafornical cell group. It is not clear why this cell group is considered to be separate from the PVN. Clusters of magnocellular neurosecretory neurons containing oxytocin or vasopressin are found throughout the hypothalamic gray in between the PVN and SON (Figs. 8.4 and 8.5). These ectopic clusters, which tend to be arranged around blood vessels, are generally referred to as “accessory nuclei”. They contain more oxytocinergic neurons than vasopressinergic neurons (Dierickx and Vandesande, 1977). However, in the older literature, Feremutch (1948) called the scattered cells and islands of neurosecretory cells between the SON and PVN the “intermediate nucleus”. This is a confusing term, since Brockhaus (1942) originally used the same name for the sexually dimorphic nucleus of the preoptic area (SDN-POA) (Braak and Braak, 1992; Chapter 5). In the rat, a particular group of accessory cells, located laterally between the PVN and SON, the nucleus circularis, was proposed to play a role as an osmoreceptor. Electrical stimulation of this nucleus produced longlasting and substantial antidiuresis and water deprivation, signs of cellular activation (Hatson, 1976; Tweedle and Hatton, 1976). In addition, neurons of the nucleus circularis in the hamster were activated in offensive aggression (Delville et al., 2000). A homologous nucleus in the human anterior hypothalamus has so far not been described. Vasopressin and oxytocin are produced in different neurons (Dierickx and Vandesande, 1979; Hoogendijk et al., 1985; Fig. 8.6). In the SON, the vasopressin neurons are obviously larger than the oxytocin neurons (Fig. 8.2). Although in the paraventricular nucleus the size of the two cell types shows considerable variation, the vasopressinergic neurons are generally larger than the oxytocinergic neurons (Dierickx and Vandesande, 1977). This is also true for the accessory nuclei. The observation in the rat that, under extreme forms of stimulation, the neurons may produce both peptides (Mezey and Kiss, 1991) has not yet been followed up in humans. A rostrocaudal gradient in the ratio between vasopressin and oxytocin neurons is present in the PVN. Whereas the ratio of vasopressin to oxytocin cells remained 80% from rostral to caudal over a distance of 1.5 mm in the dorsolateral SON, in the PVN this ratio starts below 20% rostrally, goes up to 60% in the caudal half, after which the ratio decreased again (Swaab et al., 1987b; Fig. 8.7).

169

Fig. 8.5. Adjacent sections through the hypothalamus of an AIDS patient (no. 89793) demonstrate the hybridization of the oxytocin probe in the human paraventricular nucleus (pvn) (a). The specificity of the hybridization of the oxytocin probe was supported by the absence of signal when a sense probe was used on an adjacent section, (b) and by the hybridization in the dorsal cap of the supraoptic nucleus (son) (a), which is in agreement with the established, preferential location of oxytocin cells in this nucleus. III, third ventricle; oc, optic chiasm. Bar = 5 mm. (From Guldenaar and Swaab, 1995; Fig. 1.)

Based on work in experimental animals, knowledge about the way in which the rate of hormone production is controlled (Fig. 8.8), about the physiological integration of the HNS in the control of peripheral functions, and about how this integration is accomplished on the level of the cell and the molecules involved, is reviewed by Burbach et al. (2001). On a minutes-to-hours scale, physiological activation of the HNS in the rat induces a coordinated astrocyte withdrawal from between the magnocellular somata and the parallel-projecting dendrites of the SON. These changes are accompanied by increased direct apposition of both somatic and dendritic membranes and the appearance of novel multiple synapses in both the dendritic and somatic zone. Additionally, activation results in increased interneuronal coupling. These changes play an important role in the coordinated release of oxytocin and vasopressin during, e.g. lactation and dehydration (Hatton, 1997). The SON and PVN send their axons to the neurohypophysis to release the neuropeptides into the circulation. The axons also make synaptoid contacts with the pituicytes, also in the human neurohypophysis (Okado and Yokota, 1982). Animal research has revealed that the pituicyte

2014 Ch 08

170 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 170

D.F. SWAAB

Fig. 8.6. Alternating sections from the central part of the PVN of a 72-year-old woman, stained for vasopressin (A) and oxytocin (B) and counterstained with gallocyanin. Note the absence of cross-reaction, in cells 1–5 stained in (A) and not in (B), and vice versa for cells 6–8. Also note the sharply bordered nucleoli (arrows). The bars represent 50 m and 10 m, respectively, in the low-magnification photomicrograph and in the inset. III, third ventricle; b, blood vessels. The peripheral localization of the Nissl material (endoplasmic reticulum in the neurosecretory neurons is clear in, e.g. B cells 3 and 4). (From Hoogendijk et al., 1985; Fig. 1, with permission.)

2014 Ch 08

20/11/03

11:03 am

Page 171


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 8.7. A 70-year-old woman. Vasopressin/oxytocin cell ratios were followed from rostral to caudal in six sections, for a total distance of 1.5 mm. While in the dorsolateral SON, this ratio was relatively constant for this distance, the relative number of oxytocin cells in the PVN was larger in the rostral and caudal areas as compared with the central region (cf. Fliers et al., 1985a), similar to earlier observations in the rat (Swaab et al., 1975b). The percentage of vasopressin cells was calculated on the basis of numerical density of vasopressin neurons/numerical density of vasopressin plus oxytocin neurons 100% using a discrete deconvolution procedure of nuclear measurements of immunocytochemically identified cells. (From Swaab et al., 1987; Fig. 2.)

receives a very diverse input, not only from the SON and PVN, but from other hypothalamic brain regions as well. As pituicytes are electrically coupled, activation of their receptors by innervating nerve fibers may result in a coordinated retraction from their usual position along the basal lamina that allows increased neurohypophysial hormone release (Boersma and Van Leeuwen, 1994; Hatton, 1997). Similar mechanisms in the human neurohypophysis still have to be shown. Moreover, data in the rat indicate that a subset of pituicytes in the neurohypophysis may be able to synthesize vasopressin mRNA and vasopressin themselves in response to osmotic stimulation (Pu et al., 1995). This observation revives the old discussion on the possible role of pituicytes in the production of neuropeptides (Chapter 8.1b). Vasopressin from the PVN is released not only as a neurohormone in the neurohypophysis, but also in the portal system of the anterior lobe of the pituitary, at least partly with CRH, whose action on ACTH is potentiated by vasopressin (see Chapter 8.5). In this way vasopressin is involved in the regulation of ACTH and, hence, corticosteroid secretion. The pituitary actions of vasopressin are mediated by plasma membrane receptors of the V3

171

(= V1b) subtype, coupled to calcium-phospholipid signaling systems. This system is critical in the stress response, as indicated by the preferential increase in expression of vasopressin over CRH in the PVN and the upregulation of pituitary vasopressin receptors during stress in animal experiments. The number of CRH and vasopressin-colocalizing neurons is increased during activation, e.g. in multiple sclerosis (Chapter 21.2) and depression (Chapter 26.4). V3 receptor mRNA levels and coupling of the receptor to phospholipase C are stimulated by glucocorticoids. Consequently, vasopressin upregulation may be critical for sustaining corticotroph responsiveness in the presence of high circulating glucocorticoid levels during chronic stress or depression (Aguilera and Rabadon-Diehl, 2000). Interestingly, while vasopressin is an ACTH-stimulating hormone, oxytocin inhibits ACTH release (Legros, 2001). (d) Sex differences in vasopressin neurons Although we did not find a sex difference in vasopressin neuron number, a sex difference was reported in vasopressin plasma levels. Men have higher vasopressin levels than women (Asplund and Aberg, 1991; Van London et al., 1997). In addition, the posterior lobe of the pituitary is larger in boys than in girls (Takano et al., 1999). These sex differences are explained by the higher metabolic activity we found in vasopressin neurons in the SON in young men as compared to women, as determined by the size of the Golgi apparatus (Fig. 8.9). The Golgi apparatus is located in the perinuclear region of the SON and PVN neurons and is a sensitive measure for changes in neuronal metabolism (Chapter 1.5). While estrogens in young women inhibit vasopressin synthesis, neuronal activity, whilst remaining stable in men, gradually increases in women in the course of aging, a process that is probably triggered by the decrease in estrogen levels. The sex difference in neuronal activity in the SON thus disappears after the age of 50 (Ishunina et al., 1999). This is a clear example of a hypothalamic system that shows a functional sex difference instead of a structural sex difference. It is also an example of a sex difference based on the “activating” (or in this case “inhibiting”) effect of sex hormones in adulthood. The activation of neurosecretory vasopressin neurons in postmenopausal women was confirmed by in situ hybridization (Fig. 8.10; Ishunina et al., 2000a; Ishunina and Swaab, 2002) and by measurement of the cell size as a parameter for neuronal activity (Chapter 1.5). The

2014 Ch 08

172 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 172

D.F. SWAAB

Fig. 8.8. Peptidergic neuron. Cellular and molecular properties of a peptidergic neuron (neurosecretory cell) are shown schematically. The structure of the neurosecretory cell is depicted with notations of the various cell biological processes that occur in each topographic domain. Gene expression, protein biosynthesis, and packaging of the protein into large dense core vesicles (LDCVs) in the cell body, where the nucleus, rough endoplasmic reticulum (RER) and Golgi apparatus are located. Enzymatic processing of the precursor proteins into the biologically active peptides occurs primarily in the LDCVs (see inset), often during the process of anterograde axonal transport of the LDCVs to the nerve terminals on microtubule tracks in the axon. Upon reaching the nerve terminal, the LDCVs are usually stored in preparation for secretion. Conduction of a nerve impulse (action potential) down the axon and its arrival in the nerve terminal causes an influx of calcium ion through calcium channels. The increased calcium ion concentration causes a cascade of molecular events (see inset) that leads to neurosecretion (exocytosis). Recovery of the excess LDCV membrane after exocytosis is performed by endocytosis, but this membrane is not recycled locally, and instead is retrogradely transported to the cell body for reuse or degradation in lysosomes. TGN, trans-Golgi network; SSV, small secretory vesicles; PC1 or PC2, prohormone convertase 1 or 2, respectively; CP-H, carboxypeptidase H; PAM, peptiylglycine -amidating monooxygenase. (From Burbach et al., 2001; Fig. 2, with permission.)

2014 Ch 08

20/11/03

11:03 am

Page 173


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

173

Fig. 8.9. Immunocytochemical staining of the Golgi apparatus in dorsolateral SON neurons in a young woman (A,B) and a young man (C,D). Note the clear difference at both low (B,D)and high (A,C) magnification between the male and female patients. One subject with an ovariectomy following an ovarian carcinoma (no. 80002) shows a very intense and large Golgi apparatus (E,F). Scale bar: A,C,E = 64 m; B,D,F = 300 m. (From Ishunina et al., 1999; Fig. 1.)

minimum and maximum diameters were determined in order to estimate the volumes of cell somata and cell nuclei in vasopressin neurons stained with an antibody against human glycoprotein, a part of the vasopressin precursor, and a monoclonal anti-oxytocin antibody in men and women ranging in age from 29 to 94 years. The vasopressin neurons in the SON and PVN appeared to be larger in young men than in young women. The vasopressin cell size of elderly women (>50 years old) considerably exceeded that of young women. In addition, vasopressin cell size correlated positively with age in women, but not in men. Sex differences in the size of the PVN vasopressin neurons were pronounced on the left side (p = 0.048) and absent at the right side (p = 0.368), indicating the presence of functional later-

alization of this nucleus (Fig. 8.11). No difference was found in any morphometric parameter of oxytocin neurons in the PVN among the 4 groups studied. These data demonstrate sex differences in the size of the vasopressin neurons, and thus presumably in their neurosecretory activity, that are age- and probably also side-dependent, and the absence of such changes in oxytocin neurons in the PVN (Ishunina and Swaab, 1999). The activation of vasopressin neurons in postmenopausal women is probably mediated by a decrease in the expression of estrogen receptor- in these neurons and an increase in estrogen receptor- nuclear staining (Fig. 8.12; Ishunina et al., 2000b), by which the original sex difference of these receptors in young individuals (see Chapter 6.3) disappears.

2014 Ch 08

174 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 174

D.F. SWAAB

Fig. 8.11. Graph depicting differences in the AVP cell somata volumes between the left and right parts of the SON and PVN. The Mann–Whitney test showed that there is a trend for left-right asymmetry in the PVN in men (p = 0.057). As a result, sex differences in the PVN were pronounced at the left side (p = 0.048*) and absent at the right side (p = 0.368). (From Ishunina and Swaab, 1999; Fig. 5, with permission.)

Fig. 8.10. Vasopressin mRNA in the dorsolateral part of the human supraoptic nucleus (dl-SON) of a young man (A), a young woman (B), an elderly man (C) and an elderly woman (D). Note that AVP mRNA production is significantly higher in the dl-SON of a young man (A) than of a young woman (B) and is markedly increased in an elderly woman (D) compared with a young woman (A). Bar = 42 m. (From Ishunina and Swaab, 2002; Fig. 2.)

The low-affinity neurotrophin receptor p75 (p75NTR) may also be involved in the mechanism of activation of vasopressin neurons in postmenopausal women. We investigated whether p75NTR immunoreactivity in SON neurons was age- and sex-dependent in postmortem brains of control patients ranging in age from 29 to 94 years. To study whether the p75NTR might also participate in the activation of SON neurons, we related Golgi apparatus size to the amount of p75NTR in the same patients. p75NTR

immunoreactivity indeed correlated significantly with age (Fig. 8.13) and with Golgi-apparatus size as a measure for neurosecretory activity, but only in women. These observations suggest that p75NTR participates in the activation of the SON following the reduction of estrogen levels in postmenopausal women (Ishunina et al., 2000c). The sex differences in activity of the SON are consistent with other observations showing oestrogen and progesteron interference with renal actions of vasopressin. Women have a higher turnover than men, and the greatest difference is present during the lateral phase of the menstrual cycle (Claybough et al., 2000). (e) Neuroendocrine functions, afferent fibers and other factors affecting vasopressin and oxytocin release The pituitary is a rudimentary organ without any functional meaning. (G. Van Rijnberk, 1901, cited by Prof. C. Winkler, 1994)

In spite of the suggestions of H. Cushing (1932) that, in those days of posterior lobe function, there was no clinical

2014 Ch 08

20/11/03

11:03 am

Page 175


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

175

Fig. 8.12. Graph depicting differential expression of nuclear ER and ER in vasopressin neurons in the dl-SON in relation to age and sex. In young women (39.5 ± 3.47 yr old; 6 subjects), the percentage of nuclear ER-positive neurons is 50 times higher than that in young men (36.63 ± 3.6 yr old; 8 subjects) and 250 times higher than that in elderly women (70.9 ± 4.46 yr old; 10 subjects), whereas the proportion of ER-positive cells in elderly women and in young and elderly (68.88 ± 5.2 yr old; 8 subjects) men exceeds that in young women 4.5 and 3 times, respectively. (From Ishunina et al., 2000b; Fig. 1, with permission.)

or experimental evidence (see citation, start of Chapter 8), endocrine effects on the kidney and uterus had already been known for years. In 1913 Von den Velden and Farini had already described the antidiuretic effect of posterior pituitary extracts in patients suffering from diabetes insipidus, and in 1909 Blair-Bell had reported the oxytocic effects of posterior pituitary extracts in labor. Vasopressin and oxytocin released into the blood circulation act as neurohormones. In humans, 90% of the circulating vasopressin is bound to platelets (Preibisz, 1983; Bichet et al., 1986; Nussey et al., 1986; Van der Post et al., 1994), and a patient with autoimmune thrombocytopenia (D.F.S.) indeed had low total vasopressin levels in the platelet fraction (unpublished observation). In addition, a small amount of vasopressin is present in the cytoplasmic and nuclear extracts of human peripheral blood lymphocytes that also contain a vasopressin receptor (Ekman et al., 2001). Vasopressin is released from the neurohypophysis during osmotic stimulation, hypotension or hypovolemia (Husain et al., 1973; Kakiya et al., 2000; Pedersen et al., 2001;

Fig. 8.13. Microphotographs demonstrating p75NTR immunoreactivity in the supraoptic nucleus (SON) of a young woman (A,B), an elderly woman (C,D). Note the difference in the intensity of the staining which is more prominent in an elderly woman (C,D) when compared with a young woman (A,B). Bar represents 50 m. (From Ishunina et al., 2000b; Fig. 1, with permission.)

Robertson, 2001). Vasopressin acts as an antidiuretic hormone on the kidney and regulates free water clearance by V2-type vasopressin receptors and the subsequent formation of water channels of aquaporins (Knepper, 1994; Mayinger and Hensen, 1999; Fig. 8.14; Saito et al., 2000a; Pederson et al., 2001). The G-proteincoupled vasopressin receptors signal their binding through phosphorylation of intracellular segments. Phosphorylation facilitates binding of arrestins and internalization of the receptor. Arrestins behave as adaptor proteins, facilitating the recruitment of receptors to the plasma membrane domains where the calthrin-coated pits develop. The dephosphorylated receptor returns to the

2014 Ch 08

176 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 176

D.F. SWAAB

cell surface ready to be activated and internalized once more (Bowen-Pidgeon et al., 2001). The water channel aquaporin-2, which is expressed in the renal collecting duct, is redistributed to the apical membrane in response to an intracellular signaling cascade initiated by binding of the antidiuretic hormone vasopressin to its receptor (Deen et al., 2000). Osmotic stimuli thus cause a rise in plasma vasopressin levels, induce the formation of aquaporin-2, and so water retention. Approximately 3% of the aquaporin-2 in collecting duct cells is excreted into urine. Urinary excretion of aquaporin-2 thus reflects the action of vasopressin on the collecting ducts (Pederson et al., 2001). Aquaporin-2 excretion is even considered to be a more sensitive measure of the vasopressin effect on the collecting ducts than vasopressin levels themselves (Ishikawa et al., 2001a). In central diabetes insipidus (Chapter 22.2a), aquaporin-2 excretion is diminished, and in inappropriate secretion of vasopressin, e.g. due to decreased cardiac output in cardiac failure or liver cirrhosis (Chapter 22.6), excretion of the water channels is enhanced (Ishikawa, 2000; Schrier et al., 2001). Seven renal aquaporins (1–4, 6–8) have been identified that are involved in antidiuresis and acid–base balance (King and Yasui, 2002). A lack of vasopressin may cause diabetes insipidus (Chapter 22.2) or nocturia (Chapter 22.4). The hypothalamus integrates signals from osmoreceptors that are probably located in the SON (Leng et al., 1982), from the organum vasculosum lamina terminalis, the subformical organ (Chapter 30.5) and from stretch receptors of the vascular tree. Astrocytes expressing aquaporin-4 in the SON, PVN and the subfornical organ are presumed to play a key role in osmoreception (Badaut et al., 2000). The cholinergic neurons of the DBB (Chapter 2) participate in the baroreceptor-mediated inhibition of phasic vasopressin neurons in the SON (Grindstaff et al., 2000). Predominant stimuli for thirst and vasopressin release in human are osmolality of the extracellular fluid and hypovolemia (Wells, 1998). An increased hybridization density was observed for vasopressin mRNA in the SON in subjects who had an antemortem hypovolemic status (Rivkees et al., 1989) and in our own observations (M.T. Panayotacopoulou et al., 2002). It has been proposed that Verney’s hypothalamic “vesicular osmometers” are in fact the aquaporin-4 molecules expressed in the astrocyte membranes in the SON and the ependymal cell membranes in the subfornical organ (Wells, 1998). Suppression of thirst sensation and decreased vasopressin secretion are probably mediated by stimulation of oropha-

ryngeal receptors and/or distention of the stomach. In this way, plasma osmolality is regulated within narrow boundaries of 2–3 mosmol/kg (Jenkins, 1991). In humans, osmotic stimuli affect vasopressin neurons in a selective way. Osmotic stimuli such as dehydration or high sodium intake cause a release of vasopressin (Helderman et al., 1978; Robertson and Rowe, 1980; Kirkland et al., 1984; Phillips et al., 1984; Williams et al., 1986) but not of oxytocin (Williams et al., 1986), in contrast to the rat, where both are released. Electrophysiological studies in the rat have shown that there is also sensory nervous information originating in renal and liver receptors that alter the activity of vasopressin and oxytocin neurons in the paraventricular nucleus. These receptors may also contribute to the hypothalamic control of vasopressin and oxytocin release into the circulation (Ciriello, 1998). Moreover, orthostasis is accompanied by a release of vasopressin (Rowe et al., 1982) and orthostatic hypotension may be treated by the vasopressin analogue desmopressin (Kallas et al., 1999). In addition, smoking and nicotine cause vasopressin release (Rowe et al., 1982), and alcohol inhibits vasopressin release in young subjects (Helderman et al., 1978). Moderate hypotension does not increase vasopressin levels in humans. Tonic reflex restraint of vasopressin secretion is less important in humans than in experimental animals (Goldsmith, 1989). Morphine was reported to cause a release of vasopressin in one study (Van Wimersma Greidanus and Grossman, 1991) and to have no effect in another (Philbin and Coggins, 1978), whereas naloxone is without effect on basal vasopressin levels in man (Van Wimersma Greidanus and Grossman, 1991). The tuberomamillary nucleus (Chapter 13) innervates the SON and PVN. Histamine activates both vasopressin and oxytocin neurons, which may play a role in pregnancy, parturition, lactation and novelty stress (Brown et al., 2001; Burbach et al., 2001). Following ecstasy administration, vasopressin levels are elevated for some 8 hours (Henry et al., 1998; Forsling et al., 2001). In addition, vasopressin release is inhibited by glucocorticoids and stimulated in cases of adrenal insufficiency (Ahmed et al., 1967; Erkut et al., 1998; Fig. 8.24). Vasopressin plasma and urine levels normally show a circadian rhythm, with higher levels during the night (Forsling et al., 1998). This circadian pattern is absent in nocturnal diuresis (Rittig et al., 1989; Chapter 22.4), in a significant proportion of nursing home residents with night-time incontinence (Ouslander et al., 1998; Chapter 8.3) and in the hepatorenal syndrome, also known as

2014 Ch 08

20/11/03

11:03 am

Page 177


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

177

Fig. 8.14. Schematic view of the mechanism of water regulation induced by vasopressin: Vasopressin (AVP) regulates water permeability by stimulation of vasopressin 2-receptors, which are located in the basolateral membrane of the renal tubular cell. Binding of AVP to the V2-receptors activates a guanine nucleotide (GTP) binding protein of the Gs subtype, which stimulates adenylate cyclase activity. This leads to an increase in intracellular concentration of cyclic-3,5-adenosine monophosphate (cAMP) and activation of protein kinase A (PKA). The aquaporin-2 (AQP2) water channels are the target for the action of vasopressin. Vasopressin triggers fusion of aquaporin-2-bearing vesicles with the luminal plasma membrane, and water enters the cell. Water leaves the cell passively through the basolateral membrane (AQP3 and AQP4), which results in water retention. (From Mayinger and Hensen, 1999.)

2014 Ch 08

178 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 178

D.F. SWAAB

functional renal failure or liver cirrhosis (Pasqualetti et al., 1998; Chapter 22.6). A direct projection of the SCN to the rat SON has been identified by electrophysiological means and was presumed to be the basis for these circadian rhythms (Cui et al., 1997). Such direct connections have not been found so far in the human hypothalamus, although SCN fibers come close to the SON (Dai et al., 1997, 1998a). Such fibers may terminate on SON dendrites or interneurons and in this way induce the circadian rhythm in vasopressin release. In addition, the optic nerve sends lateral retinohypothalamic tract projections to the ventromedial SON (Dai et al., 1998b), which may also induce day/night rhythms in vasopressin levels. There are, moreover, circannual fluctuations in vasopressin blood levels. In both men and women the plasma vasopressin levels were higher in winter than in any other season (Asplund et al., 1998; Chapter 4.1b). In women, oxytocin is involved in labor and lactation (Evans, 1997; Lindow et al., 1999). Pregnancies complicated by diabetes insipidus, uterine atony and very prolonged labor have been reported, although in other patients with diabetes insipidus the release of oxytocin may be undisturbed (Marañón, 1947; Chau et al., 1969), depending on the nature and level of the lesion. Only a failure of lactation (not of labor) was found in oxytocin knock-out mice (Russell and Leng, 1998), which can probably be explained by back-up systems for the other functions of oxytocin. During normal delivery, stretching of the lower birth canal triggers the neurohormonal “Ferguson” reflex, leading to rapid secretion of oxytocin by the pituitary gland, which results in strong expulsive contractions. In the rat, it was shown that at-term pregnancy uterine contractions activate both oxytocin and vasopressin neurons in the SON and that this activation involves a noradrenergic pathway (Douglas et al., 2001). A significant increase in oxytocin blood levels was found at the onset of full cervical dilation and crowning of the fetal head, while this increase in oxytocin levels did not occur following a lumbar epidural block, probably because the Ferguson reflex is blocked by epidural analgesia. This would explain why, following epidurals, more arrests of the descent during the second stage of labor and more forceps deliveries are required. Indeed, oxytocin treatment during the second stage of labor with epidural analgesia reduces the need for forceps (Goodfellow et al., 1983). Morphine suppresses oxytocin secretion in the first stage of human labor, while an effect of naloxone was not demonstrated (Lindow et al., 1992; Feinstein et al., 2002).

It is of considerable interest that oxytocin plasma levels rise during the night while -endorphin levels decrease in pregnancy (Lindow et al., 1996) while labor takes place preferentially during the night. Oxytocin might be involved in the initiation of preterm labor. Competitive antagonists of oxytocin inhibit contractions in uncomplicated preterm labor (Åkerlund et al., 1987; Turnbull, 1987) and can prolong uterine quiescence after successful treatment of an acute episode of preterm labor (Valenzuela et al., 2000). The inhibitory effect of alcohol on oxytocin release (Thornton et al., 1992) has been used when premature labor threatened (Fuchs et al., 1967), however, without even considering the frightening possibility of detrimental effects of this therapy on the developing fetal brain. Infants born within 12 hours of administration of ethanol had significantly lower 1-min Apgar scores and a higher incidence of respiratory stress syndrome (Zervoudakis et al., 1980), which supports this possibility of detrimental effects. It should also be noted though that the role of oxytocin in the initiation of normal labor is controversial (Chapter 8.1h) and that more recent research shows that a marked increase in oxytocin blood levels occurs only in a minority of the patients with a normal delivery. Progress of labor was not found to be related to an increase in oxytocin blood levels in that study either, which does not support a crucial role for an increase in circulating oxytocin during spontaneous at-term labor (Thornton et al., 1992). It is of great importance that changes in oxytocin and vasopressin (V1) receptors could modify uterine activity, even in the absence of a change in blood levels of these peptides. Such receptors are present both in pregnant human myometrium and decidua (Ivanisevic et al., 1989; Åkerlund et al., 1999). While the vasopressin V2 receptor does not seem to be involved to any significant degree in the activation of the pregnant human uterus (Åkerlund et al., 1999), oxytocin receptors increase during pregnancy in the human uterus (Tence et al., 1990). In fact, uterine oxytocin receptor expression peaks in early labor (Maggi et al., 1990), suggesting a role in the process of parturition. Since oxytocin receptor mRNA expression in myometrium increases in late pregnancy, whereas decidual expression was much lower and did not go up at term (Wathes et al., 1999), an increased sensitivity to oxytocin of the myometrium may be crucial for the onset and/or course of labor. As shown by the presence of oxytocin mRNA in human amnion, chorion and decidua, the peptide oxytocin is also produced in these very same tissues, while oxytocin gene expression

2014 Ch 08

20/11/03

11:03 am

Page 179


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

increased three- to four-fold around the time of onset of labor. The possible role of this extrahypothalamic oxytocin in the onset and course of labor would not be reflected in increased blood levels (Chibbar et al., 1993). The putative role of fetal vasopressin and oxytocin in parturition is discussed below (see Chapter 8.1). Oxytocin levels in CSF increase during labor (Takeda et al., 1985), a change that is presumed to be associated with the induction of maternal behavior by its central effects. Concluding, at present the literature provides strong support for a role of the oxytocin/oxytocin receptor system in the regulation of human parturition, in particular in the expulsion phase. However, maternal paracrine rather than endocrine mechanisms might be more important, and the regulation of the receptor seems to have more profound effects than regulation of the ligand. The system is regulated by a wide variety of chemical and physical factors, including sex steroids, orphan receptors and uterine stretch (Mitchell and Schmid, 2001). In women, oxytocin plasma levels increase following estrogen administration (Kostoglou-Athanassiou et al., 1998b). In breastfeeding women, oxytocin secretion is inhibited by morphine (Lindow et al., 1999). In addition, estrogens reduce the set point for osmoregulation of vasopressin (Stackenfeld et al., 1999). Pharmacological studies in humans and animals suggest the existence of vascular endothelial vasopressin and oxytocin receptors that mediate vasodilatory effects (Thibonnier et al., 1999a). In addition, abundant amounts of oxytocin mRNA were found in the rat vena cava and pulmonary vein. The amount of mRNA was enhanced in some blood vessels by estrogens (Jankowski et al., 2000). Vasculature thus also seems to contain an intrinsic oxytocin system. A possible relationship was found between oxytocin for induction of labor and the risk of sudden infant death syndrome (Einspieler and Kenner, 1985) and for neonatal jaundice (Chalmers et al., 1975). (f) Neurotransmitters and neuromodulators: the central pathways involved . . . and evidence will be presented to show that posterior lobe extracts are far more potent when injected into the cerebral ventricles than by any other means of administration. (Harvey Cushing, 1932, p. 21)

Commercially available posterior lobe extract (“obstetrical pituitrin”) which Harvey Cushing (1932, p. 59) injected into the ventricle in 38 instances in 24 subjects convalescing (!) from operations for pituitary adenoma,

179

appeared to have a pronounced stimulatory effect, essentially parasympathetic in character, and, apparently, central in origin (Fig. 30.1). Since intramuscularly injected pituitrin had a reverse effect, and no central effects were noted when the hypothalamus was affected by hydrocephaly or by a tumor, Cushing indeed had solid evidence to propose a central action (i.e. on the hypothalamus) of neurohypophyseal extracts. These impressive observations passed into oblivion, perhaps because the scientific community chose to fight the “neurosecretion” concept proposed by the Scharrers (see earlier). And when the neurosecretion concept was accepted because Bargmann (1949) demonstrated the same Gomori-positive material in the neurohypophysis and in the neurons of the SON and PVN, this did not explain central effects of neurohypophysial hormones. In the concept of neurosecretion, Barry’s proposition (1954) concerning the existence of Gomori-positive endings outside the hypothalamus (“synapses neurosécrétoires”) could not be properly appreciated and was, therefore, eventually forgotten, a process likely to have been expedited by the fact that his articles were in French. When De Wied (1965) showed that posterior lobectomy in the rat resulted in an accelerated extinction of conditioned shuttle-box avoidance response, and, later on, that this behavioral deficiency could be alleviated by peripherally administered vasopressin (De Wied and Bohus, 1966), these effects on memory processes were explained in terms of an endocrine concept, i.e. on the basis of the release of neurohypophysial hormones into the bloodstream, which in turn would transport the peptides back to the brain. However, a number of observations have made it extremely unlikely that the vascular route or the cerebrospinal fluid (CSF, see Chapter 4.1g), would be major channels for physiological central actions of neurohypophysial hormones (Swaab, 1982). In the meantime, moreover, immunocytochemical observations had revealed yet another site of production for vasopressin – the suprachiasmatic nucleus (SCN) (Swaab et al., 1975). Immunocytochemistry for vasopressin and neurophysin subsequently led to the rediscovery of Barry’s (1954) extensive extrahypothalamic pathways terminating in various structures ranging from the olfactory bulb to the spinal cord (Swanson, 1977; Buijs et al., 1979; Sofroniew and Weindl, 1978). The main sources of thesevasopressinergic and oxytocinergic pathways were thought to be the SCN and the PVN, although the SON could not be excluded altogether as an additional site of origin, since, in the lateral rat hypothalamus, morphological and

2014 Ch 08

180 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 180

D.F. SWAAB

electrophysiological evidence has been obtained for the presence of axon collaterals from the SON (Mason et al., 1984). Moreover, there is one autoradiographic study claiming the presence of extrahypothalamic fibers originating in the rat SON (Alonso et al., 1986). Also, Golgi studies of the human hypothalamus demonstrated the presence of axon collaterals that may stay within the SON and/or project to other hypothalamic nuclei (Al-Hussain and Al-Jomard, 1996). However, this study has never been confirmed and, following lesions in the rat PVN, nearly all brainstem innervation disappears (Lang et al., 1983), which indicates that such SON fibers are not extensive. Bilateral lesions of the rat PVN did not eliminate the entire vasopressin content of the brainstem and the spinal cord but caused a reduction of 50% and 80%, respectively, while oxytocin almost completely disappeared. This suggested that the PVN was not the only site of origin of extrahypothalamic vasopressin (Lang et al., 1983). Colchicine administration has subsequently revealed vasopressin-containing neurons in the rat BST, medial amygdala, dorsomedial hypothalamic nucleus and locus coeruleus (Caffé and Van Leeuwen, 1983; Van Leeuwen and Caffé, 1983) as possible sources for extrahypothalamic pathways. Previously, the putative source of the fiber projections had been established only by tracing immunocytochemically stained fibers in serial sections. However, this method proved to be unreliable because, when vasopressin projections from the SCN were checked by lesioning this nucleus, it turned out not to be the source of the lateral septum VP innervation. Other proposed target regions of the SCN, viz. the dorsomedial nucleus of the hypothalamus and the organum vasculosum of the lamina terminalis, however, were confirmed (Hoorneman and Buijs, 1982). Additional cell bodies containing vasopressin or oxytocin have been reported in the human brain BST (Fliers et al., 1986; Mai et al., 1993), DBB and NBM (Ulfig et al., 1990), in a parafornical cell group (Mai et al., 1993), the lateral septal nucleus, globus pallidus and the anterior amygdaloid nucleus (Mai et al., 1993). Ultrastructural immunocytochemical observations in the rat have shown that extrahypothalamic peptidergic fibers terminate synaptically on other neurons (Buijs and Swaab, 1979). In the human brain, vasopressin and oxytocin fibers, probably originating from the PVN, project to the NBM, DBB, septum, BST, locus coeruleus, the parabrachial nucleus, the nucleus of the solitary tract, the dorsal motor nucleus of the nervus vagus, substantia nigra, dorsal raphe nucleus and spinal cord (e.g.

Sofroniew, 1980; Fliers et al., 1986; Mazurek et al., 1987; Unger and Lange, 1991; Fodor et al., 1992; Van Zwieten et al., 1994, 1996; Figs. 8.15 and 8.16). The PVN is a central structure in the autonomic regulation of endocrine glands and even of adipous tissue by these peptidergic nerve fibers (Buijs and Kalsbeek, 2001). An extension to the extrahypothalamic neurophysin-containing fiber distribution in the human brain was given by Mai et al. (1993), though without distinguishing between vasopressinergic and oxytocinergic fibers. Vasopressin-containing fibers innervate the fissures of the choroid plexus in rat (Brownfield and Kozlowski, 1977), while the choroid plexus of the human brain was found to contain vasopressin binding sites. In Alzheimer’s disease a twofold increase in vasopressin binding sites was found in this structure (Korting et al., 1996). Vasopressin is thought to play a role in the choroid plexus with respect to ion and water transport and to reduce CSF production (Nilsson et al., 1992a). Using tritiated vasopressin, oxytocin and agonists, high-affinity binding sites were found in a number of brain areas, which were different for the two peptides in the forebrain, while there was overlap in the brainstem. Vasopressin binding sites were detected in the dorsal part of the lateral septal nucleus, in midline nuclei and adjacent intralaminar nuclei of the thalamus, in the hilus of the hippocampal dentate gyrus, the dorsolateral part of the basal amygdaloid nucleus and the brainstem. Oxytocin binding sites were observed in the NBM, the vertical limb of the DB, the ventral part of the lateral septal nucleus, the preoptic/anterior hypothalamic area, the posterior hypothalamic area and variably in the globus pallidus and ventral pallidum (Loup et al., 1991). In the brainstem and spinal cord, oxytocin binding sites were the most intense in the substantia nigra pars compacta, substantia gelatinosa of the caudal spinal trigeminal nucleus and of the dorsal horn of the upper spinal cord, as well as in the mediodorsal region of the nucleus of the solitary tract. Binding was also present in the rest of this tract and in the spinal trigeminal nucleus (Loup et al., 1989). The central vasopressinergic fibers may be involved in blood pressure and temperature regulation, regulation of osmolality and corticosteroid secretion and may thus influence cognitive functions, aggression, paternal behavior and social attachment (Legros et al., 1980; De Wied and Van Ree, 1982; Buijs et al., 1983; Fliers et al., 1986; Holsboer et al., 1992; Legros and Anseau, 1992;

2014 Ch 08

20/11/03

11:03 am

Page 181


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

181

Fig. 8.15. Vasopressin pathways in the human brain. Question marks indicate that at present no site of origin or termination is known. A: amygdala; AM: anteromedial subnucleus of the basal nucleus; BST: bed nucleus of the stria terminalis; DBB: diagonal band of Broca; DMV: dorsal motor nucleus of the nervus vagus; LC: locus coeruleus; NSM: nucleus septalis medialis; NTS: nucleus of the solitary tract; PBN: parabrachial nucleus; PVN: paraventricular nucleus; SCN: suprachiasmatic nucleus; SN: substantia nigra; SON: supraoptic nucleus. (Scheme from E.J. van Zwieten, 1995.)

Insel, 1997). In addition, the PVN area is activated in men by smelling an estrogen-like pheromone (Savic et al., 2001; Chapter 24.2). Oxytocinergic central pathways are involved in reproduction (see 8.1g), cognition, tolerance, adaptation, and in the regulation of cardiovascular and respiratory functions (Gutkowska et al., 2000; Mack et al., 2002). A large dose of oxytocin impairs recall in human subjects (Ferrier et al., 1980; Kennett et al., 1982). Oxytocin

neurons from the PVN that innervate the brainstem nuclei are involved in blood pressure and heartrate regulation (Maier et al., 1998). The role of oxytocin in cardiorenal functions is mediated by the release of atrial natriuretic peptide from the heart. There are oxytocin receptors present in all heart compartments and in the vasculature. In addition, oxytocin is synthesized in the heart, and in large vessels like the aorta and vena cava (Gutkowska et al., 2000). Centrally released oxytocin would also give

2014 Ch 08

182 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 182

D.F. SWAAB

expressing neurons in the PVN was found in this disorder (Swaab et al., 1995a; Chapter 23.1). The increased activity of oxytocinergic neurons in depression may contribute to the eating disturbances in this disorder (Purba et al., 1996; Chapter 26.4). Oxytocin has been administered to patients with obsessive-compulsive disorder (Evans, 1997; Chapter 26.6). Alterations in PVN function have been proposed in schizophrenia, too, on the basis of the highly decreased numbers of neurophysin-containing neurons that were found in the PVN of schizophrenic patients. No differences were found in the SON of schizophrenic patients, however (Mai et al., 1993), but it is not known whether and how the PVN changes may contribute to the symptomatology of the disorder (see Chapter 27.1). Vasopressin and oxytocin: Yin-yang hormones. (Legros, 2001)

Fig. 8.16. Photomicrographs of vasopressin and oxytocin-immunoreactive fibers in the parabrachial nucleus of the human brain. Vasopressin (A) and oxytocin (B) innervation in the dorsal part of the lateral PBN. The arrowheads (B) indicate corpora amylacea. (C) Vasopressin and (D) oxytocin innervation of the ventral part of the lateral. The vasopressin innervation in the dorsal (E) and ventral (F) part of the medial PBN. Note the much denser vasopressin innervation than the oxytocin innervation. Scale bar = 80 m. (Van Zwieten et al., 1996; Fig. 2, with permission.)

rise to sedation. Oxytocin levels rise after non-noxious stimulation such as touch, and after exposure to light and high temperatures. Oxytocin is held responsible for the antistress effects that occur during lactation (UvnäsMoberg, 1997) and elicits yawning (Argiolas et al., 1986). Oxytocin induces a decline in cortisol (Chiodera et al., 1991), which may be essential for the formation of social bonds (see 8.1g). Brain oxytocin modulates a range of social behaviors, from parental care to mate guarding. Social amnesia was found in mice lacking the oxytocin gene. Treatment with oxytocin, but not with vasopressin, rescued social memory in these mice (Ferguson et al., 2000). Oxytocin administration decreased fatigue, anger and anxiety (Evans et al., 1997). Oxytocin also has central effects on food intake, and oxytocin neurons are considered to be the putative satiety neurons for eating behavior. Our observations in Prader– Willi patients, who have an insatiable hunger and extreme obesity, support this idea, as a 42% decrease in oxytocin-

Vasopressin and oxytocin have opposite functions. Vasopressin potentiates the effect of CRH on ACTH, while oxytocin inhibits ACTH release. These two closely related peptides also have opposite actions on cognition. Vasopressin effects are all directed towards protecting homeostasis of the individual (water retention, blood pressure regulation, increased arousal and memory), whereas oxytocin actions are all directed towards the social group (fetal expulsion, milk let-down, social behavior and interaction (see 8.1g). Vasopressin can therefore be seen as a ‘selfish’ peptide and oxytocin as an ‘altruistic’ peptide (Legros, 2001). (g) Oxytocin, vasopressin and reproductive behavior Oxytocin and vasopressin are thought to be involved in affiliation, including pair bonding, parental care, and territorial aggression in monogamous animals (Insel, 1997; Young et al., 1998), maternal behavior and other aspects of reproductive behavior (Carter, 1992; Insel, 1992; Anderson-Hunt and Dennerstein, 1995; McKenna, 1998). Human sexually dimorphic reactions to pheromones (Savic et al., 2001) may be involved in such processes in the PVN. Lesions in the male rat PVN also indicate that this nucleus is involved in erection and that the magnocellular and parvicellular elements play different parts in this function (Liu et al., 1997b). Electrical stimulation of the paraventricular nucleus in squirrel monkeys elicited penile erection (MacLean and Ploog, 1962). Electrical

2014 Ch 08

20/11/03

11:03 am

Page 183


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

stimulation of the dorsal penile nerve or of the glans penis excited 60% of the oxytocin cells in the contralateral and ipsilateral SON of the rat. In another experiment, too, oxytocin cells of the PVN were found to be activated by sensory information from the penis. In the male rat, oxytocin is an extremely potent inducer of penile erection (Argiolas et al., 1986; McKenna, 1998). Since intracerebroventricular injection of an oxytocin antagonist reduced non-contact penile erections in rat dose-dependently, the involvement of central oxytocin in the expression of penile erections does not only seem to be a pharmacological effect but to have a physiological function as well (Melis et al., 1999). The direct contractile effects of vasopressin on penile blood vessels, together with its amplifying effects on adrenergic-mediated constriction support the idea that circulating oxytocin may also be involved in penile erection (Segarra et al., 1998). Tactile stimulation of the penis during male copulatory behavior further activates oxytocin cells, both in the PVN and in the SON, and this seems to induce both central and peripheral oxytocin release. Excitatory amino acid transmission increases in the PVN during noncontact erections. This may contribute to the nitric oxide production in the PVN and activates oxytocin neurons, thus mediating this sexual response (Melis et al., 2000). In the erectogenic effects of oxytocin, MSH and its MC4 receptor are also involved (Martin et al., 2002). Dopamine neurotransmission to the PVN is also supposed to be involved in penile erection (Chen, 2000). In patients with psychogenic erectile dysfunction apomorphin sublingual caused an extra activation of the hypothalamus during erotic video stimulation. Apomorphin acts on the oxytocinergic neurons in the PVN. In addition, oxytocin may reduce maternal aggression in a period shortly after the birth when lactating females show naturally high levels of this behavior (Giovenardi et al., 1998). In women, basal levels of oxytocin during lactation are associated with a desire to please, give and interact socially (Uvnäs-Moberg et al., 1998). Oxytocin released during labor and lactation may influence human maternal responsiveness and perhaps attachment (Carter, 1998). The increased oxytocin levels in CSF during labor in humans are presumed to be associated with the induction of maternal behavior (Takeda et al., 1985). In a number of cases, milk let-down, indicating oxytocin release, has been reported during the sexual act in women who were in their lactating period (Campbell and Petersen, 1953). In men, oxytocin may also be involved in sexual arousal and ejaculation (Carmichael et al., 1987;

183

Murphy et al., 1987). In women, too, oxytocin secretion seems to be related to smooth muscle contractions during orgasm (Carmichael et al., 1987). One minute after orgasm, oxytocin levels are increased in women. In agreement with this case report, oxytocin levels are known to rise during sexual arousal and peak during orgasm in both women and men. In multiorgasmic subjects, oxytocin peaks immediately prior to and during terminative orgasm, i.e. the oxytocin peak coincides with sexual saturation. The intensity of orgasmic contractions, but not their duration, correlated positively with increases in oxytocin levels. Naloxon decreases the level of pleasure at orgasm and blocks the periorgasmic rise of oxytocin levels (Murphy et al., 1990; Carmichael et al., 1994). A recent study showed an increase in plasma oxytocin levels immediately following orgasm in men, after which a rapid decline occurred to basal levels within 10 min (Krüger et al., 2003). Administration of vasopressin inhibits copulatory behavior in female rats, while a vasopressin antagonist facilitated the lordosis response (Meyerson et al., 1988). Vasopressin plasma levels remained, however, unaltered during sexual arousal and orgasm according to a recent study in humans (Krüger et al., 2003). Both in men and women, oxytocin induces contractions of smooth muscle cells and may thus facilitate transport of eggs and sperm (Carmichael et al., 1994). Animal experiments in the rat confirm the possibility of such roles (Ackerman et al., 1998). In female animals, oxytocin was found to facilitate estrus, sexual arousal, receptivity and other mating behaviors, including lordosis. An as-yet unconfirmed case report has described a woman who began to take a contraceptive pill containing progesteron only. She experienced accentuated physiological and psychological sexual arousal after she had coincidentally used a prescribed synthetic oxytocin spray for let-down of breast milk (Anderson-Hunt, 1994; Anderson-Hunt and Dennerstein, 1995). Animal experiments show a mechanism of interaction between sex hormones and oxytocin by initiating the production of receptors for this peptide (Anderson-Hunt and Dennerstein, 1995). The human oxytocin gene promotor has an imperfect palindrome with sequence similarities to other estrogen response elements (Richard and Zingg, 1990). Oxytocin is also produced by the male reproductive tract and modulates not only its contractility but also steroidogenesis. The finding that the oxytocin receptor is present in the interstitial tissue and in Sertoli cells in the testes supports the presence of such biological actions of oxytocin (Frayne and Nicholson, 1998). Indeed,

2014 Ch 08

20/11/03

11:03 am

Page 184

184 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

D.F. SWAAB

oxytocin levels are significantly higher in women on oral contraceptives, and oxytocin levels recorded during the menstrual week are significantly lower than at other times (Uvnäs-Moberg et al., 1989). Oxytocin plasma levels increase on estrogen administration (KostoglouAthanassiou et al., 1998b). The plasma concentration of vasopressin during the menstrual cycle is doubled on day 16–18 as compared to day 1 (Forsling et al., 1981), while others found a tendency of vasopressin to increase on day 11–13, when peak concentrations of estradiol occur (Punnonen et al., 1983). Not only were the basal vasopressin levels higher in the follicular phase of the natural menstrual cycle, their nocturnal peaks were also higher (Kostoglou-Athanassiou et al., 1998b). Increased oxytocin plasma levels were reported in a few women around the time of ovulation (Mitchell et al., 1980). In the nonpregnant human uterus, oxytocin and vasopressin receptors of the V1 subtype are present (Guillon et al., 1987) and vasopressin may thus have a role in stimulating contractions of the nonpregnant uterus. In this respect it is of interest that women with premenstrual pain or primary dysmenorrhea, who have an increased uterine activity and a decreased blood flow in the uterus, have increased plasma levels of vasopressin (Åkerlund et al., 1979; Strömberg et al., 1984). A V1a vasopressin antagonist was effective in the prevention of dysmenorrhea in one study (Brouard et al., 2000) but not in another (Valentin et al., 2000). Moreover, vasopressin has central antinociceptive effects (Chapter 31.2a). (h) Vasopressin, oxytocin and osmotic regulation in pregnancy Normal pregnancy is associated with a 30–50% increase in total plasma and extracellular fluid volumes, and a substantial rise in cardiac output. Hyponatremia is also the rule. Mean arterial pressure decreases by 6 weeks’ gestation, in association with an increase in cardiac output and plasma volume, and a decrease in systemic vascular resistance (Schrier et al., 2001). Concurrent with an increase in body circulatory fluid volume, there is a significant decrease in the volume of the maternal brain size, and an increase in ventricular size, which reverses by 6 months after delivery (Oatridge et al., 2002). During early pregnancy, the thirst threshold, i.e. the level of plasma osmolality above which thirst is sensed and which therefore regulates drinking behavior, as well as the osmotic threshold, i.e. the level of plasma osmolality above which vasopressin is released, are generally

reported to decrease. The thirst threshold decreases from ±300 to 290 and the osmotic threshold from ±290 to 280 mosmol/kg (Davison, 1984). It should be noted, though, that some authors have found no difference for the thirst threshold or the osmotic threshold as compared to the thresholds after pregnancy (Thompson et al., 1991). During pregnancy, in the Brattleboro rat (which has an autosomal, recessive hereditary hypothalamic vasopressin deficiency), plasma osmolality is also lowered about 10 mosmol/kg (Barron et al., 1985). This means that an increase in vasopressin plasma levels is not necessary to induce the changes in plasma osmolality as seen in pregnancy. Indeed, no significant increase in plasma vasopressin levels was found in early (5–8 weeks gestational age) (Davison et al., 1988), nor in late human pregnancy (Buemi et al., 2001). In nonpregnant patients with hypothalamic diabetes insipidus, oxytocin was found to have only 5% of the antidiuretic effects of vasopressin (Kelley et al., 1992). However, during pregnancy, oxytocin is reported to act antidiuretically (Liggins, 1962; Whalley and Pritchard, 1963; Abdul-Karim and Rizk, 1970; Gupta and Cohen, 1972), yet it is considered to be unlikely that oxytocin causes the pregnancy-associated decrease in plasma osmolality, since this phenomenon already occurs very early in pregnancy and oxytocin levels around that time are invariably found to be low (Kumaresan et al., 1974; Leake et al., 1981a; Padayachi et al., 1988; Thornton et al., 1992). On the other hand, a more recent study showed a gradual increase in plasma oxytocin levels during the course of pregnancy (Buemi et al., 2001). The reported plasma levels of vasopressin in pregnant women show a large variation. Average basal plasma levels of vasopressin during pregnancy vary between 0.7 and 5.1 pg/ml (Rosenbloom et al., 1975; Leung et al., 1980; Davison et al., 1984; Barron et al., 1985; DeVane, 1985; Lindheimer et al., 1985; Pedersen et al., 1985; Brown et al., 1988; Davison et al., 1988; Davison et al., 1989; Baylis and Munger, 1990) and are not different from the nonpregnant situation. Only one author (DeVane, 1985) described a small reduction in plasma vasopressin levels in the first and second trimester of pregnancy of approximately 1 pg/ml (compared to 4 pg/ml in the third trimester and in nonpregnant women). Others found no differences compared to nonpregnant levels or during labor. Differences in assay conditions during blood sampling, vasopressin platelet binding and pulsatile hormonal release in the case of oxytocin and intersubject variation might account for the differences in reported

2014 Ch 08

20/11/03

11:03 am

Page 185


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

plasma levels (Van der Post et al., 1994). Probably the most important factor for variation is in vivo activity of cystyl-aminopeptidase (CAP). The human placenta produces CAP, an enzyme also called vasopressinase or oxytocinase. It degrades nmol/l amounts of vasopressin or oxytocin (Vmax = 10–6 mol/l/min) very rapidly, while the peptides circulate in pmol/l amounts (Tuppy, 1968; Lauson, 1974; Van Oudheusden, 1986; Burd et al., 1987). In vitro degradation of vasopressin and oxytocin in human pregnancy plasma is, therefore, massive, and the determination of these neuropeptides during pregnancy is thus seriously hampered (Van der Post et al., 1994). Moreover, inhibitors of CAP were found seriously to disturb our assays. We therefore developed a radioimmunoassay for vasopressin in which o-phenanthroline effectively inhibits CAP activity in pregnant women and in which the enzyme is removed during the extraction of vasopressin to prevent disturbance of the assay (Van der Post et al., 1994). The metabolic clearance rate (MCR) of vasopressin increases 3- to 4-fold during the course of human pregnancy (Davison et al., 1989). Plasma CAP levels increase during gestation and are positively correlated with the increase in MCR of vasopressin. Both free-circulating CAP as well as placental CAP contribute to the elevated MCR in mid- and late pregnancy (Davison et al., 1989). Consequently, the slope that describes the linear relationship between plasma vasopressin levels and plasma osmolality in individuals is less steep in the third trimester of pregnancy compared with the first trimester and to nonpregnant values (Davison et al., 1984, 1988; Brown et al., 1988). In nonpregnant patients this slope correlates positively with the platelet-bound vasopressin fraction (Bichet et al., 1987). Since plasma levels of unbound vasopressin only diminish slightly, if at all, during pregnancy, vasopressin production and release must be increased in view of the increased MCR. The increased neurophysin plasma levels during pregnancy (see below) and the increased aquaporin-2 urinary excretion during pregnancy (Buemi et al., 2001) support this view. That urinary excretion of immunoreactive vasopressin was not found to be altered in third trimester pregnancy (Davison et al., 1981) may thus be explained by CAP degradation of vasopressin, so that an elevation could not be detected. This possibility has not been studied systematically. Reported basal plasma levels of oxytocin vary considerably (see also Chapter 8.1e). Levels between belowdetection limit to over 400 pg/ml have been described throughout pregnancy and labor (Kumaresan et al., 1974; Leake et al., 1981a; De Geest et al., 1985; Takada et al.,

185

1985; Kuwabara et al., 1987; Padayachi et al., 1988; Thornton et al., 1990; Fuchs et al., 1992; Thornton et al., 1992). Differences in immunological detection of oxytocin following degradation by CAP or differences in assay sensitivity for inhibitors, among other factors, might explain this variability in analogy with the factors mentioned before, for vasopressin measurement (Fuchs et al., 1991). Only in more recent studies, in which measurement problems are dealt with adequately, are oxytocin plasma levels described to be low in late gestation: 0.24–0.34 pg/ml (Fuchs et al., 1992), < 1.5 pg/ml in 75% of the measurements (Thornton et al., 1992) and < 0.42 pg/ml in all patients (Fuchs et al., 1991). Oxytocin plasma levels are thought to rise only during delivery by an increase in pulse frequency and duration (Thornton et al., 1992). However, more recently Buemi et al. (2001) observed a gradual increase in plasma oxytocin levels during the course of pregnancy; also, data on the MCR of oxytocin are equivocal. Some find a 3- to 4-fold increase in late gestation (Thornton et al., 1990), others do not find differences compared with nonpregnant individuals (Ryden and Sjvöholm, 1971; Amico et al., 1987; Takeda et al., 1989). We measured neurohypophysial hormones in healthy nulliparous women in the course of pregnancy. Although elevated vasopressin-neurophysin and oxytocin-neurophysin levels during pregnancy seem to indicate increased release of neurohypophysial hormones, until 36 weeks of gestation pregnancy is accompanied by low circulating vasopressin and oxytocin levels. This discrepancy may be explained by increased CAP breakdown of these neuropeptides during pregnancy. Reduced circulating platelet-bound vasopressin levels were found during pregnancy. Randomized sodium restriction only diminished 24-h urinary vasopressin excretion in nonsmoking pregnant women without changing circulating levels of vasopressin or oxytocin (Van der Post et al., 1997b). The recently observed increased urinary excretion of aquaporin-2 and the gradual increase in oxytocin plasma levels support the idea that the HNS is activated during human pregnancy (Buemi et al., 2001). (i) Pre-eclampsia and hypertension in pregnancy Classically, pre-eclampsia is defined as hypertension in combination with proteinuria and/or edema. Preeclampsia is an important cause of fetal and maternal morbidity and mortality. Hemodynamics of women with severe hypertension and proteinuria are characterized by a

2014 Ch 08

186 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 186

D.F. SWAAB

reduced cardiac output combined with an increased systemic vascular resistance and low plasma volume in comparison with normal pregnancy. The increase in vascular resistance leads to reduced organ perfusion and subsequent organ dysfunction. Cerebral vasoconstriction, too, can be severe, as shown by arteriography (Lewis et al., 1988), and may lead to ischemia, resulting, e.g. in cortical blindness or convulsions (eclampsia). In pre-eclamptic patients the reduction in brain size during pregnancy was stronger than in healthy pregnant women (Oatridge et al., 2002). As early as the beginning of the 20th century, Zweifel called pre-eclampsia “the disease of theories”. Neurohypophysial hormones have been included in these theories. Vasopressin is a very potent vasoconstrictor (Cowley and Liard, 1987) and has therefore also been presumed to be involved in the pathogenesis of pre-eclampsia. Vascular reactivity for infused vasopressin is increased in pre-eclampsia and precedes the onset of the development of hypertension (Steegers and Van der Post, 1998). There has, however, been no report of an increase in circulating vasopressin in pre-eclamptics (Pedersen et al., 1985) or in hypertensive women (Brown et al., 1986). On the other hand, these reports showed methodological shortcomings concerning the vasopressin assay, as discussed before. We have therefore investigated this possibility again, using the radioimmunoassay that was specially developed for use during pregnancy (Van der Post et al., 1994). However, there was no difference in vasopressin levels of platelet-poor plasma or in the amount of platelet-bound vasopressin or the vasopressin platelet receptor density and affinity between hypertensive/pre-eclamptic women and nondiseased pregnant women (Van der Post et al., 1993). Oxytocin administration during pregnancy is known to cause water retention (Storch, 1971), and even water intoxication (Whalley and Pritchard, 1963; Lilien, 1968) due to the antidiuretic effect of oxytocin on the kidney during pregnancy (Douglas, 1965). In addition, oxytocin may elevate cardiac output during pregnancy (Weis et al., 1975). In spite of these effects on kidney and circulation, we found that the oxytocin plasma levels of pre-eclamptic levels were not elevated (Van der Post et al., 1993), so no support was obtained for the idea that increased peripheral circulating neurohypophysial hormones might play a role in the pathogenesis of pre-eclampsia. The possibility of a role of central vasopressin in the development of preeclampsia has so far not been studied. Other factors than vasopressin may certainly play a role in pre-eclampsia. Neurokinin-B, a tachykinin that causes hypertension,

is produced by the placenta. Plasma concentrations of this peptide are grossly elevated in pregnancy-induced hypertension and pre-eclampsia (Page et al., 2000). Moreover, a function mutation of the mineral corticoid receptor was found that causes early-onset hypertension, probably because the receptor specificity has been altered in such a way that progesterone and other steroids become potent agonists for this receptor (Geller et al., 2000). 8.1. The fetal SON, PVN in birth and development . . . since the abnormal process of birth frequently produces no effect, difficult birth in itself in certain cases is merely a symptom of deeper effects that influenced the development of the fetus. Sigmund Freud, 1897

Not only maternal but also fetal neurohypophysial hormones play a role in the birth process. Fetal oxytocin has been proposed to initiate parturition (Schriefer et al., 1982) or accelerate the course of labor (Swaab et al., 1977; Boer et al., 1980). However, other data show that labor is not associated with an increase in fetal oxytocin levels (Patient et al., 1999) and that umbilical plasma levels of oxytocin in anencephalics are no different from that of controls (Oosterbaan and Swaab, 1987), arguing against an active release of oxytocin by the fetus during normal labor. However, fetal vasopressin levels in umbilical cord blood are much higher following normal delivery than at any other stage of life (Chard et al., 1971; Oosterbaan and Swaab, 1989). The umbilical arterial vasopressin concentrations are up to 200-fold greater than those caused by water deprivation (Thornton et al., 2002). In human anencephalics, such a rise in fetal vasopressin levels does not occur (Oosterbaan and Swaab, 1987), confirming its origin from the fetal brain. Fetal vasopressin is one of the hormones that plays a role in the adaptation of the fetus to the stress of labor, for example by redistribution of the fetal blood flow, with a marked reduction in the flow to gastrointestinal and peripheral circulations and an increase in the flow to essential organs such as the brain, the pituitary, the heart and the adrenals (Iwamoto et al., 1979; Pohjavuori and Fyhrquist, 1980). This adaptive vasopressin response has been said to be induced by the stress of birth (Chard et al., 1971), by hypoxemia, acidemia (Parboosingh et al., 1982; Daniel et al., 1983) or by a rise in intracranial pressure associated with delivery. Perinatal hypoxia stimulates the vasopressin neurons in particular, as suggested by the increased coexpression of tyrosine hydroxylase in

2014 Ch 08

20/11/03

11:03 am

Page 187


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

these neurons (Panayotacopoulou et al., 1994; Chapter 8.2). It has, moreover, been shown in vitro that the human myometrial contractile response to vasopressin was much stronger than that to oxytocin, after, but not before, labor. Encouraging results have been obtained from clinical trials with the oxytocin antagonist atosiban in the treatment of threatening preterm labor. As atosiban is more selective for the vasopressin V1a than the oxytocin receptor, there is an ongoing debate about the possible role of vasopressin in term and preterm labor. Since the V1a receptor expression is lower in term-pregnant than in nonpregnant uterus, high expression of this receptor is not required for a vasopressin response. Also, a high umbilical vasopressin production by the fetus is not mandatory for normal labor, since, in many individuals in whom labor progressed in a normal fashion, a marked arteriovenous difference in umbilical cord vasopressin levels was observed. A role of fetal vasopressin in the onset or course of labor is thus far from settled (Thornton et al., 2002). Since parturition is prolonged in the Brattleboro rat, which lacks vasopressin (Boer et al., 1981), vasopressin may be involved in the course of labor. The neurons producing the neurohypophysial peptides are already present early in fetal life, and, in the 8.5-week-old human fetus, fenestration of capillaries, granular vesicles and clear vesicles is already found in the neural lobe (Okado and Yokata, 1980), indicating hormone release. Most cells of the magnocellular system seem to be derived from the region of the hypothalamic sulcus and migrate laterally and ventrally before expressing neurophysin at their site of settling. The cells possibly migrate along epidermal growth factor receptorpositive radial glial cells, which extend from the hypothalamic sulcus into the lateral hypothalamus. Using neurophysin staining, which did not distinguish between oxytocin or vasopressin neurons, Mai et al. (1997) detected staining of SON and PVN neurons from as early as 10 and 14 weeks of gestation onwards, respectively, roughly coinciding with the arrival of the optic tract fibers. Vasopressin and oxytocin have been found as early as 11 and 14 weeks of gestational age, respectively (Paulin and Dubois, 1978; Fellmann et al., 1979; Burford and Robinson, 1982). Neurophysin was demonstrable at 13 weeks of gestation in the accessory nuclei and at 14 weeks of gestation in the PVN. Vasopressin-neurophysin was detected from 18 weeks of gestation onwards, and vasopressin-mRNA from 21 weeks of gestation in the SON, PVN and accessory nuclei (Murayama et al., 1993). A dense catecholaminergic network of fibers is already

187

present in the PVN of a 3- to 4-month-old fetus (Nobin and Björklund, 1973). An increase in vasopressin and oxytocin levels in the pituitary and the brain during the development of the fetus has been described by a number of researchers (Skowsky and Fisher, 1977; Schubert et al., 1981; Burford and Robinson, 1982; Khan-Dawood and Dawood, 1984). It has been proposed that chronic fetal secretion of vasopressin may produce oligohydramnios by decreasing lung fluid and urinary contribution to amniotic fluid volume decrease (Leake et al., 1985). Vasopressin secretion is strongly increased during labor (see earlier) and levels in the urine already decrease gradually during the first 24–36 hours after birth. However, elevated vasopressin excretion remains present in children with insults such as intracranial hemorrhage, hypoxic encephalopathy and pneumothorax (Wiriyathian et al., 1986). In the perinatal period, those children with a mean gestational age of 27 weeks that developed a chronic lung disease were more immature, of lower birthweight and had higher plasma vasopressin levels and higher urine osmolality. The vasopressin levels were significantly correlated with the duration of oxygen dependency (Kavvadia et al., 2000) and are thus a measure of perinatal stress. Aquaporin-2 encodes the vasopressinregulated water channels of the renal collecting duct (Fig. 8.14) and is excreted in human urine. Aquaporin-2 is present in the urine of term and preterm children but in concentrations that are several times lower than in adulthood. Its excretion correlates positively with urine osmolality and its levels decrease postnatally (Tsukahara et al., 1998). Plasma oxytocin levels also decrease rapidly after delivery, but they remain elevated over adult basal levels. It should be noted in this connection that breast milk contains some 10 l oxytocin/ml (Leake et al., 1981b), which may explain the higher levels in newborns. Vasopressin excretion in urine decreases gradually between 2 and 11 years of age when expressed in relation to creatin excretion. However, when vasopressin excretion is expressed in relation to body surface area, no difference with age or with adult excretion was found (AllevardBurguburu et al., 1981). In boys between the ages of 5 and 19 years, the posterior pituitary is larger than in girls (Takano et al., 1999), which is probably a reflection of a sex difference in neurohypophysial activity (see Chapter 8.1d). The vigorous vasopressin response to hypotension in the fetus is partially mediated by arterial baroreceptors. In contrast, vasopressin responses to hypoxia are relatively small and are mediated by the generation of adenosine (Wood and Tong, 1999).

2014 Ch 08

188 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 188

D.F. SWAAB

. . . I pushed out of the womb against my mother’s strength – I felt free. Soledad Brother The prison letters of George Jackson, 1970

Because premature children are more sensitive to the stress of birth, and the SON and PVN neuropeptides seem to be involved in adaptation to the stress of birth, we determined the number of oxytocin and vasopressin-expressing neurons in the human fetal (HNS) in premature and at-term children. From the youngest fetus of our study onwards, so from a gestational age of 26 weeks, adult vasopressin and oxytocin cell numbers were found in both the SON and the PVN (Fig. 8.17 cf. Wierda et al., 1991; Van der Woude et al., 1995). This is in agreement with Dörner and Staudt’s (1972) qualitative estimation that the hypothalamic nuclei are already completely formed around 25 weeks of gestation. This conclusion was based on the disappearance of the matrix layer around the third ventricle at that age (Staudt and Stüber, 1977). The fetal HNS, however, is far from mature at term, in spite of the adult cell numbers. This is apparent, for example, from the neuronal densities, which are still decreasing after this period. Rinne et al. (1962) found a gradual increase in nuclear volume of the neurons of the SON and PVN during fetal development, but did not distinguish between oxytocin and vasopressin neurons. Judging by the strongly increasing nuclear size of the oxytocin neurons in the fetus during the last part of gestation (our unpublished observation), these neurons seem to become gradually strongly activated towards term. This should, however, be confirmed by better measures of neuronal activity, such as in-situ hybridization for oxytocin-mRNA. It therefore seems quite possible that less-mature oxytocin neurons in premature children would be, at least partly, responsible for the increased incidence of obstetrical problems. The idea of an active fetal role of oxytocin neurons in delivery is reinforced by a number of clinical observations. Firstly, human anencephalics do not have a neurohypophysis and have impaired neurohypophysial hormone release (Visser and Swaab, 1979; Oosterbaan and Swaab, 1987). In anencephalics expulsion takes twice as long and the birth of the placenta even takes three times longer, suggesting a role of fetal brain and possible of neuroendocrine mechanisms in speeding up the course of labor. In addition, the observation that about half of the anencephalics die during the course of labor is a strong indication of the importance of an intact fetal brain to withstand the stress of birth (Honnebier and Swaab, 1973; Swaab et al., 1977; Chapter 18.1). The second observation

Fig. 8.17. Vasopressin (light bars) and oxytocin (dark bars) cell numbers in the PVN of premature (26–37 weeks) and mature (37–42 weeks) fetuses and adults. Adult numbers were already present around 26 weeks gestation. (From Goudsmit et al., 1992; Fig. 6, with permission.)

is derived from children suffering from Prader–Willi syndrome. These children frequently suffer from considerable obstetrical problems (Wharton and Bresman, 1989), and we found that Prader–Willi patients have only 58% of the normal number of oxytocin neurons in adulthood, but a normal number of vasopressin neurons (Swaab et al., 1995a; see Section 23.1). The third argument is based on the frequent perinatal problems found in septo-optic dysplasia (De Morsier syndrome), in which the fetal HNS is often damaged (Chapter 18.3b). Moreover, prolonged labor and breech delivery have been documented in 50–60% of the idiopathic growth hormone-deficient children, a disorder that seems to be based on congenital hypothalamopituitary abnormalities (Maghnie et al., 1991; Chapters 18.4 and 18.6). Importantly, the causality of the relationship between obstetric complications and neurological or psychiatric diseases such as schizophrenia or autism (Geddes and Lawrie, 1995; Verdoux and Sutter, 2002) might thus be quite the reverse of what is generally thought. A disturbed labor might thus be the first symptom of a brain disorder, probably even more often than that disturbed labor is the cause of the brain disorder. How the fetal hypothalamus might play a role in the disorders of fetal presentation, e.g. in Prader–Willi syndrome (Wharton and Bresman, 1989) is not yet known.

2014 Ch 08

20/11/03

11:03 am

Page 189


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Immaturity at term may also affect the function of vasopressin neurons, though to a lesser degree than oxytocin neurons, since the vasopressin neurons are already further advanced in the developmental process, something that has also been reported for other species. This is in agreement with the vasopressin levels in fetal cord blood, which are extremely high after delivery (Oosterbaan and Swaab, 1989). Hereditary hypothalamic diabetes insipidus children from similar mothers without vasopressin did not have a history of difficult labor. At first sight this seemed to suggest that the absence of vasopressin alone does not prevent an adequate neuroendocrine adaptive response of the fetus during labor (Swaab et al., 1982). However, later it became clear that vasopressin production in these children may be quite normal up to the age of 9 years or so (Bahnsen et al., 1992; Chapter 22.2), so that they are not a useful source of information on the putative involvement of fetal vasopressin in the process of labor. For the possible involvement of the fetal brain – and in particular the fetal CRH neurons – in timing the moment of birth, see Chapters 18.1 and 8.5. Male infants are more likely to have arrest of descent during the second stage of labor, to require oxytocin augmentation, instrumental vaginal delivery or caesarean section. Female infants are more likely to have meconium stained liquor (Feinstein et al., 2002; Eogan et al., 2003). The possibly neuroendocrine sex differences behind these different risks for male and female children need clarification. 8.2. Colocalization of tyrosine hydroxylase (TH) with oxytocin and vasopressin Immunohistochemical studies have indicated that, in the adult human PVN and SON, a large proportion of neurons contains the catecholamine-synthesizing enzyme tyrosine hydroxylase (TH) (Spencer et al., 1985; Li et al., 1988; Panayotacopoulou et al., 1991). The SON and PVN have therefore also been considered to be part of the catecholaminergic system and designated “A15”, according to the nomenclature of Dahlström and Fuxe (1964). However, whether or not dopamine is indeed produced in the human SON and PVN still remains a controversial topic, as aromatic L-amino acid decarboxylase (AADC) has not been found in these nuclei (Kitahama et al., 1998a). A difference with the classical catecholaminergic neurons is, moreover, that melanin pigment, considered to be a

189

by-product of L-DOPA synthesis, is not observed in the PVN and SON, in contrast to their presence in, for instance, a number of periventricular and arcuate nucleus neurons (Spencer et al., 1985). On the other hand, SON and PVN neurons contain calbindin, just like the midbrain dopaminergic neurons (Sanghera et al., 1995). In adulthood, TH staining in the PVN varies strongly between individuals. TH is certainly not present in lesser amounts in Parkinson patients (Purba et al., 1994), which supports the notion that in Parkinson’s disease dopaminergic neurons of the mesencephalon, but not of the hypothalamus, are affected. In the fetal human, PVN and SON some TH-staining perikarya are present from 4.5 to 6 gestational weeks onwards (Zecevic and Verney, 1995. A large number of TH neurons was found in fullterm neonates who had died of perinatal hypoxia, while only a few were evident in the premature ones (Panayotacopoulou and Swaab, 1993). A clear difference between the neonate and adult cases of our sample was observed in the proportion of TH neurons that colocalize oxytocin or vasopressin (Fig. 8.18). In neonates, the majority of the TH-IR perikarya stained for vasopressin, while only few TH neurons were also positive for oxytocin. The reverse was observed in adults, where the majority of the double-stained TH neurons colocalized oxytocin while only few TH-IR perikarya appeared to contain vasopressin. In the rat, an increase in TH-IR in vasopressin producing neurons was observed after experimental manipulations that activate vasopressin synthesis. (Kiss and Mezey, 1986). This indicates that the expression of TH is a sign of hyperactivity of SON and PVN neurons. This can also be concluded from animal experiments showing increased TH in the SON and PVN during lactation, following osmotic stress (Meister et al., 1990), and in the Brattleboro rat (Kiss and Mezey, 1986), which lacks vasopressin and, therefore, has osmotically activated neurosecretory neurons. Increased fetal vasopressin secretion has been found in human neonates following fetal stress, asphyxia or rises in intracranial pressure (see Chapter 8.1). The colocalization of TH with vasopressin in the neonatal PVN and SON may therefore indicate that antemortem stress factors such as perinatal hypoxia have increased TH production in the vasopressin neurons during the process of birth (Panayotacopoulou et al., 1994; Fig. 8.18). This possibility has diagnostic consequences and therefore needs further investigation. In adulthood, increased TH staining was found in the SON and PVN in patients following osmotic activation due to dehydration and

2014 Ch 08

190 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:03 am

Page 190

D.F. SWAAB

Fig. 8.18. (1) PVN of a 74-year-old male (no. 81.032) double-stained for tyrosine hydroxylase (TH) and oxytocin (OXT). TH-IR perikarya are stained in red (open arrow), OXT-IR neurons in blue and doublestained for TH + OXT neurons in purple (arrow). (2) SON of a 3-month-old infant (no. 85.002.3) double-stained for TH + OXT. Note neurons containing both TH and OXT stained in purple (arrow) but also single-stained TH-IR perikarya in red (arrow). (3) PVN of a 37year-old male (no. 84.248) double-stained for TH + vasopressin (VP). TH-IR neurons are revealed in red, VP-IR neurons in blue and doublestained TH + VP positive perikarya in violet (arrow). (4) SON of a 37-year-old male (no. 84.248) double-stained for TH + VP. Violet neurons containing both TH and VP are evident (arrow) among the blue VP-IR perikarya. (5) PVN of a neonate of 40 weeks of gestation (no. 88.353.2) double-stained for TH + VP. Note many double-stained TH + VP positive neurons revealed in purple (arrows). (6) SON of a neonate of 37 weeks of gestation (no. 89.153.3) double-stained for TH + VP. Note some perikarya containing both TH + VP revealed in purple (arrow) among the blue VP-IR perikarya. The bar represents 25 m. (From Panayotacopoulou et al., 1994; Figs. 1–6.)

nonosmotic activation of the neurosecretory neurons in case of pulmonary portal hypertension or liver cirrhosis that lead to a decrease in “effective” blood volume (Panayotacopoulou et al., 2002). These observations confirm the idea that TH expression in the SON and PVN is a sign of hyperactivity of these neurons. Observations in the rat indicate that, although TH is expressed in response to hyperosmotic stimulation and

coexists with vasopressin in magnocellular neurons, there is a lack of L-DOPA and of the second-step catecholamine synthesizing enzyme, i.e. AADC. Zucker obese (fa/fa) rats, which have an activated SON and PVN, while heterozygous lean (Fa/fa) rats spontaneously express TH in magnocellular vasopressin neurons and a few oxytocin neurons, independent of an osmotic challenge. The lack of L-DOPA and AADC in these neurons agrees with the presumed absence of mechanisms necessary for catecholamine synthesis in these cells (Fetissov et al., 1997). A similar conclusion may be derived from the observation that in the human and monkey PVN, the neurons that are IR to TH were not immunopositive for AADC (Kitahama et al., 1998a). Of course, there is always the possibility that the sensitivity of the methods by which AADC has been detected so far may not have been sufficient. Alternatively, catecholamines may be produced at the level of the hypophysis, using the AADC that is present in the capillaries. A similar situation exists in the ventrolateral part of the arcuate nucleus (Meister et al., 1988; Meister and Elde, 1993). So, although the expression of TH in SON and PVN neurons may be a valuable measure of hyperactivity of these neurons, it remains controversial whether the production of dopamine takes place in the vasopressin and oxytocin neurons. Also, the functional consequence of the colocalization of vasopressin and TH is not clear at present. However, in case dopamine is indeed produced in the SON, PVN or neurohypophysis, it is interesting to note that it is capable of both inhibiting the release of vasopressin (Lightman and Forsling, 1980) and stimulating its release (Spigset and Hedenmalm, 1995), and that it facilitates drinking in animals (Naitoh and Burrell, 1998). In addition, rat studies show that the D1, D2 dopamine receptor, as well as the DARP-32 protein (the third messenger for dopaminergic neurotransmission) have been located in the rat neurointermediate lobe, indicating that dopamine, but also neuroleptics, can act at the level of the pituitary (Holzbauer et al., 1983). Moreover, L-DOPA in vivo has several vasopressin-related functions. It not only increases myocardial contractile force, but also renal plasma flow, glomerular filtration rate, and potassium and sodium excretion. The latter effects were observed in hypokalemic but not in normokalemic patients, and a correlation between aldosteron production and this renal effect of L-DOPA was suggested (Finlay et al., 1971; Granerus et al., 1977). A different potential function of the THcontaining neurons in the PVN is the regulation of LHRH neurons in the infundibular and medial preoptic areas,

2014 Ch 08

20/11/03

11:03 am

Page 191


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

since catecholamines can modulate LHRH production (Dudás and Merchenthaler, 2001). 8.3. The SON and PVN in aging and Alzheimer’s disease It’s a fortunate person whose brain Is trained early, again and again, And who continues to use it To be sure not to lose it, So the brain, in old age, may not wane. Rosenzweig and Bennett, 1996

Various observations have provided evidence in support of our hypothesis that activation of neurons interferes in a positive way with the process of aging and with Alzheimer’s diseases. We proposed that a high neuronal metabolic activity may prolong the life span of neurons and that activation may restore their function. This hypothesis is paraphrased as “use it or lose it” (Swaab, 1991). The neurons of the SON and PVN have been instrumental in the formulation of this hypothesis, since they form a population of extremely stable cells in normal aging and in Alzheimer’s disease. The classical Alzheimer changes are generally absent in the SON and PVN (Ishii, 1966; Saper and German, 1987). Despite the use of several antibodies, neither cytoskeletal alterations nor /A4 plaques are found in the neurons of the SON of most Alzheimer patients (Standaert et al., 1991; Swaab et al., 1992b). Only in 5–8% of the old subjects studied were Alzheimer changes observed in the PVN and SON (Schultz et al., 1997). Although in the PVN of some Alzheimer patients some neuronal and dystrophic neurite staining with cytoskeletal antibodies can be observed (Swaab et al., 1992b), no /A4 plaques are present (Standaert et al., 1991) and the total cell number in the PVN remains unaltered (Goudsmit et al., 1990). In addition, the low-affinity p75 neurotrophin receptor is still present in the SON and PVN of aged subjects (Moga and Duong, 1997), which is a sign of high activity. In women, even an age-related increase in the p75 neurotrophin receptor was found, which correlated with the increase in neurosecretory activity of the vasopressin-containing SON neurons as measured by increased in situ vasopressin mRNA (Ishunina et al., 2000a; Fig. 8.10) and the increased size of the Golgi apparatus (Ishunina et al., 2000c; Fig. 8.13). These observations were in accordance with the hypothesis that metabolic activation protects against aging

191

and Alzheimer’s disease. SON and PVN neurons are thus not only metabolically highly active throughout life, but are even extra-activated in senescence, as can be judged from the increase in the size of the vasopressin-producing perikarya (Fliers et al., 1985), nucleoli (Hoogendijk et al., 1985) and Golgi apparatus (Lucassen et al., 1993, 1994), and from the enhanced plasma levels of vasopressin (Frolkis et al., 1982; Rondeau et al., 1982; Johnson et al., 1994; Frolkis et al., 1999) and neurophysins (Legros et al., 1980). It should be noted, though, that there are also reports that do not confirm an increased basal vasopressin plasma level in aging, either in the basal levels (Helderman et al., 1978; Rowe et al., 1982; Naitoh and Burrell, 1998) or after an overnight fast (Duggan et al., 1993). Also, in contrast to our data, Mann et al. (1985a) found a decrease in nucleolar volume in Alzheimer’s disease in the PVN and SON. The fact that we later found that the age-related increase in SON activity only takes place in women and not in men (Ishunina et al., 1998; Figs. 8.9, 8.10 and 8.19) may explain at least a number of the discrepancies in literature. The number of neurons expressing vasopressin in the PVN increases during the course of aging in controls – a change that is also interpreted as hyperactivation – and remains stable in Alzheimer patients. In the SON the number of vasopressin neurons did not correlate with age and was similar for Alzheimer patients and nondemented controls (Van der Woude et al., 1995; Fig. 8.20a,b). The observation that the plasma vasopressin response to osmotic stimulation by hypertonic saline infusion is intact in Alzheimer patients (Peskind et al., 1995) supports the idea that this system remains largely unaffected in this disease. This observation also questions the importance of the cholinergic system in the osmotic response of vasopressin neurons, since this system was shown in many studies to be seriously affected in Alzheimer’s disease (see Chapter 2). The increased activity of the SON and PVN during aging may also alter the balance between storage and release of neurohypophysial hormones and thus be the basis of the age-related decline in the frequency of the posterior pituitary bright spot found by MRI, a frequency estimated to decline with a rate of approximately 1% per year (Brooks et al., 1989). Elderly people are less inclined to notice that they are thirsty. After 24 hours of water deprivation, there appeared to be a deficit in thirst in healthy elderly men. However, osmoreceptor sensitivity as estimated by the vasopressin response to hypertonic saline increases with

2014 Ch 08

192 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 8.19.

20/11/03

11:04 am

Page 192

D.F. SWAAB

Graph depicting mean Golgi apparatus (GA) and cell profile area per neuron in males and females as measured in the maximal vasopressinergic area of the dorsolateral supraoptic nucleus (dl-SON). (From Ishunina et al., 1999; Fig. 2, with permission.)

age, which is in agreement with the increased neurosecretory activity we observed. The enhanced vasopressin response to osmotic stimulation may be a response to reduced renal function (see below). However, elderly people also frequently fail to release vasopressin in response to orthostasis and, after a short decrease in vasopressin levels, have a paradoxical increase following ethanol infusion (Helderman et al., 1978; Robertson and Rowe, 1980; Rowe et al., 1982; Kirkland et al., 1984; Phillips et al., 1984; Stout et al., 1999). The alterations in water and sodium balance in the elderly is readily influenced by many disease states and medications and may be expressed as either hyponatremia or hypernatremia (Miller, 1997). Drinking does not seem to activate the oropharyngeal inhibition of vasopressin secretion in elderly patients, which predisposes for hyponatremia (McKenna and Thompson, 1998). Alzheimer patients are

even more at risk of dehydration due to a loss of the protective “thirst” response. Plasma vasopressin levels were not different from controls after overnight dehydration, but these levels can be considered as inappropriately low for the level of serum osmolality (Albert et al., 1994). Endocrine studies suggest that elderly subjects are more frequently in a state similar to partial central diabetes insipidus; they complain of polyuria, especially at night. Although vasopressin deficiency has been proposed as a cause (Faull et al., 1993), circadian disturbances (Chapter 4.3), renal resistance to vasopressin or a mild heart failure are more probable explanations. In normal elderly people, thirst is significantly reduced. Since the relationship between vasopressin and osmolality is unchanged, the increased plasma osmolality seen in elderly people may be due to the kidney’s reduced response to vasopressin

2014 Ch 08

20/11/03

11:04 am

Page 193


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

193

Fig. 8.20. (a) Linear regression between vasopressin (AVP) cell number in the PVN and age. Data of male () and female () control subjects did not differ and were pooled. A significant correlation between age and cell number was found in control subjects (r = 0.583, p < 0.01; n = 20). Old control subjects had a significantly higher cell number compared with young controls. Values of male (▲) and female (•) Alzheimer’s disease patients are delineated by a minimum convex polygon and were reduced compared to old controls. Note that the rise in AVP cell member with age in controls does not occur in Alzheimer patients, but neither was a decrease in cell number found. (From Van der Woude et al., 1995; Fig. 3, with permission.) (b) Linear regression between vasopressin (AVP) cell number in the dorsolateral supraoptic nucleus (SON) and age. Data of male () and female () control subjects did not differ and were pooled. No statistically significant correlations with age were observed in either young or old subjects. Values of male (▲) and female (•) AD patients were within the range of the controls (from Van der Woude et al., 1995, with permission). These data show that the SON cells that are very active during the process of aging, both in controls and AD patients, are not lost in AD. (Van der Woude et al., 1995; Fig. 4, with permission.)

(Ledingham et al., 1987; Naitoh and Burrell, 1998). In this respect it is of clinical interest that plasma osmolality may be a predictor of outcome in acutely ill elderly patients (O’Neill et al., 1990). Vasopressin levels are significantly increased in the temporal lobe of Down’s syndrome patients and Alzheimer patients and significantly reduced in the cerebellum of Down’s syndrome patients (Labudova et al., 1998). The functional significance of these observations is not clear. Animal experiments suggest that the increased activity of SON and PVN neurons during the course of aging might be a compensatory activation due to an age-related disorder of the vasopressin receptor systems in the kidney (Fliers and Swaab, 1983; Ravid et al,. 1987; Goudsmit et al., 1988; Herzberg et al., 1989). It has been known for a long time now that the concentrating ability of kidneys in humans declines during aging (Lewis and Alving, 1938). A more recent study reported that, in a 30-month-old WAG/Rij rat, the age-related polyuria was not related to changes in vasopressin V2 receptor mRNA, that the cAMP content of the papilla was unchanged

and that the vasopressin binding sites were reduced by only 30%. However, aquaporin-2 and 3 expression were downregulated by 80% and 50%, respectively, in the medullary collecting duct, which seem to be the main cause of the age-related polyuria (Preisser et al., 2000). The number of oxytocin-expressing neurons in the PVN remains unaltered in aging and Alzheimer’s disease (Wierda et al., 1991; Fig. 8.21). In some projection areas, i.e. the hippocampus and temporal cortex of Alzheimer patients, the oxytocin concentration even increased, whereas it remained unaltered in other areas (Mazurek et al., 1987). Both observations are in agreement with a stable oxytocin neuron population in aging. Interestingly, chronic administration of a crude extract of the posterior pituitary gland or synthetic oxytocin extended the life span of old male rats (Friedman and Friedman, 1963; Bodansky and Engel, 1966). To my knowledge this effect has not been the focus of further studies on aging. Perhaps this effect of oxytocin is related to the antiproliferative effect of this peptide, as was demonstrated in vitro in various neoplastic cells (Cassoni et al., 2001).

2014 Ch 08

194 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:04 am

Page 194

D.F. SWAAB

Fig. 8.21. Linear regression between oxytocin (OXT) cell number in the PVN and age. Data of male and female control patients did not differ and were pooled. No statistically significant correlations were observed in either young or old control subjects. Values of male and female Alzheimer disease patients (closed symbols) are delineated by a minimum convex polygon and were within the range of the controls. (From Wierda et al., 1991; Fig. 3, with permission.)

Fig. 8.22. Representative immunocytochemical staining of solitary neurons of the human hypothalamus expressing mutant vasopressin (VP + 1, a–c) and oxytocin (OT + 1, d) precursors displaying the +1 reading frame. a–c: solitary neurons of the supraoptic nucleus stained with huva + 1 antiserum directed against VP + 1 precursors. The arrowheads in b and c point out to fibers immunoreactive for VP + 1 precursors; d: solitary neuron of the paraventricular nucleus stained with humox + 1 antiserum directed against the OT + 1 precursor; Note that immunoreactivity in all positive cells is exclusively present in the cytoplasm. Bar = 25 m. (From Evans et al., 1996; Fig. 2, with permission.)

8.4. (a) Vasopressin secretion in various disorders

It is intriguing that the occurrence of a high-frequency +1 frameshift mutation has been described in vasopressin transcripts in the rat. Dinucleotide deletions (GA) occur in the rat, predominantly at GAGAG motifs, and the number of vasopressin neurons expressing the +1 protein increases with age (Evans et al., 1994). Indications for similar frameshift mutations have now been found in the human SON and PVN vasopressin, and to a lesser degree in oxytocin precursors (Fig. 8.22). So far no age-related increase in these mutations has been found in the human hypothalamus (Evans et al., 1996). Since the number of mutated cells is about 3 per 10,000 neurosecretory neurons, this mutation rate will probably not affect the function of the HNS and not give rise to diabetes insipidus. It may well be, however, that in other neuronal systems or in other molecules the frequency of such frameshift mutations is much higher. In the cerebral cortex of Alzheimer and Down syndrome patients, +1 mutations have been observed in various disease-related proteins, such as in the neurofibrillary tangles of Alzheimer’s disease. They are present in mRNA but not in DNA and are due to a recently discovered process designated “molecular misreading”.

Increased release of vasopressin during osmotic stimulation (Husain et al., 1973; Chapter 8.1c) and during birth and aging have been dealt with earlier in Chapters 8.1, 8.2 and 8.3. Chronic alcohol consumption affects the vasopressinexpressing neurons mainly in the SON, but also in the PVN. In 10 chronic male alcoholics who consumed over 80 g of ethanol per day, the volume of the SON and PVN and the number of vasopressin-expressing neurons correlated negatively with the alcohol intake. With a consumption level of over 100 g of ethanol per day, a loss of vasopressin neurons was the result. In addition, neuronophagia, pyknosis and neuronal loss were noted in the SON and PVN. Alcoholics respond inappropriately with suppressed vasopressin levels under osmotic challenge. Concluding, chronic alcohol consumption is toxic to hypothalamic vasopressin neurons in a concentration- and time-dependent manner, not only in patients with Wernicke’s encephalopathy (see Chapter 29.5), where the SON and PVN are also affected (Harding et al., 1996). Although alcohol is generally recognized as an inhibitor for vasopressin and oxytocin blood levels, acute alcohol intoxication and hangover stimulate vasopressin secretion (Taivainen et al., 1995). Observations

2014 Ch 08

20/11/03

11:04 am

Page 195


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

in alcohol-preferring rats and high alcohol-drinking rats have shown increased plasma vasopressin levels and increased vasopressin mRNA in the PVN, but not in the SON (Hwang et al., 1998). In chronic alcoholic patients, neurophysin II (oxytocin neurophysin) serum levels were elevated and normalized during alcohol withdrawal. Neurophysin I (vasopressin-neurophysin) levels were normal on admission and did not change during hospitalization (Legros et al., 1983). Since V1 receptor antagonists do not alter blood pressure in healthy human volunteers, vasopressin does not seem to be required for blood pressure maintenance in resting conditions (Thibonnier et al., 1999b). However, in acute hypotension during ganglion blockade, there is a delayed and sustained pressor response and increased vasopressin release (Jordan et al., 2000). In an early qualitative study in patients with essential hypertension, the SON and PVN were reported to be unchanged in number and Gomori staining (Wehrle, 1950). In a later paper the SON and PVN were reported to be hypertrophied in hypertensive patients. The neuronal nuclei in the SON and PVN were increased in diameter, the total volume of the SON was increased and the capillary networks of the SON and PVN were dilated. These alterations have been interpreted as hyperactivity of the neurosecretory neurons (Postnov et al., 1974). Indeed, recent observations confirmed that the PVN is hyperactive in hypertensive patients. Quantitative analysis showed an increase in the total number of CRH-expressing neurons that was approximately twofold, and a more than fivefold increase in the amount of CRH-mRNA in hypertensive patients (Goncharuk et al., 2002). The decreased activity of the SCN in those patients (Goncharuk et al., 2001; see below) may contribute to the hyperactivity of the PVN, since animal experiments have shown that the vasopressinergic efferents of the SCN normally inhibit the CRH neurons in the PVN (Kalsbeek et al., 1992). The vasopressin cells in the PVN of hypertensive patients have so far not been studied, but elevated vasopressin levels have been found in hypertensive subjects (Cowley et al., 1981; Zheng et al., 1995; Frolkis et al., 1999; Zhang et al., 1999). More attention has to be paid to the observation that high vasopressin levels in essential hypertension were confined almost exclusively to males (Cowley et al., 1987). Morbidity and mortality caused by cardiovascular diseases increase in postmenopausal women, possibly as a consequence of the lack of estrogens (Cicconetti et al.,

195

2000). Monitoring of 24-hour ambulatory blood pressure revealed that blood pressure is higher in men than in premenopausal women at similar ages. After menopause, however, blood pressure increases in women to levels that are even higher than those in men. In most cases, hormone replacement therapy with estrogens does not significantly reduce blood pressure in postmenopausal women, suggesting that the loss of estrogens may not be the only component involved in the higher blood pressure in women after menopause (Reckelhoff, 2001). Although we have indeed found, in postmenopausal women, that vasopressin neurons are stimulated, as indicated by an increase in vasopressin mRNA (Ishunina et al., 2000a), accompanied by a decrease in estrogen receptor , an increase in estrogen receptor (Ishunina et al., 2000b) and an increase in the neurotrophin receptor p75 (Ishunina et al., 2000c), this may not be the full explanation of the rise in blood pressure in postmenopausal women. There are also sex differences in the response to vasopressin. Pressor responsiveness to vasopressin was greater and baroreflex sensitivity was attenuated to a lesser extent in hypertensive males than in hypertensive females (Share et al., 1988), which argues in favor of a relationship between hypertension and vasopressin. However, others state that the higher blood pressure levels, mortality and morbidity observed in postmenopausal women are simply attributable to their older age and not to the change in hormone levels in menopause (Casiglia et al., 1996). A particularly strong relationship between high blood pressure and vasopressin levels seems to be present in subjects with low levels of renin (Zhang et al., 1999). In patients with severe salt-induced hypertension of end-stage renal disease where plasma vasopressin was found to be increased, and in patients with a malignant hypertension, a V1 receptor antagonist produced a modest but consistent fall in supine blood pressure (Thibonnier et al., 1999b). We have to await the results of ongoing studies to see whether nonpeptide vasopressin receptor antagonists are effective in hypertension (Paramjape and Thibonnier, 2000). Yet the possible involvement of vasopressin is a controversial topic. There was no change in blood pressure when levels of vasopressin up to five times the size of those found in malignant hypertensive patients were infused into normal subjects (Padfield et al., 1976), nor was there a change in patients with vasopressin excess, as in inappropriate vasopressin secretion (Zhang et al., 1999). In patients with inappropriate vasopressin syndrome, plasma levels as high as 700 pg/ml were observed (control

2014 Ch 08

196 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:04 am

Page 196

D.F. SWAAB

levels are about 1 pg/ml) without hypertension (Padfield et al., 1976). One could, however, argue that in many cases the vasopressin elevation is in fact highly appropriate, i.e. to compensate for cardiac failure or backward failure (Chapter 22.6). The same can be said of the observation of the small increases of plasma vasopressin that are associated with moderate dehydration. They do not seem to play a role in the maintenance of arterial pressure (Huch et al., 1998). The high vasopressin plasma levels might be caused by hypertension, e.g. be secondary to the stimulus of a contracted plasma-volume following salt and water loss due to renal injury, and not play a causal role in the pathogenesis of hypertension. On the other hand, hypertensive patients showed a tendency towards higher rates of vasopressin responses when plasma osmolality was raised (Henneberry et al., 1992). So far, molecular variants of the V1 receptor gene have not been found to be involved in essential hypertension (Thibonnier et al., 2000). Although it is thus by no means settled that vasopressin acts as a vasoconstrictor in human hypertension, there are indications that it is indirectly involved through its volume-retaining properties or its central actions on the cardiovascular medullary centers, the baroreflex, the autonomous nervous system or catecholamine metabolism (Johnston, 1985). A positive correlation was found between increased plasma levels of oxytocin and increased systolic blood levels with orgasm. Since blood pressure was also increased following intracisternal administration of oxytocin in dogs, this peptide may be involved in the central control of blood pressure (Carmichael et al., 1994). In this respect it is of interest that, in the spontaneously hypertensive rat, a decrease was found in PVN concentration of vasopressin and oxytocin, while SON levels remained unchanged. It is the PVN that is thought to be involved in the control of autonomic functions (Morris and Keller, 1982; Chapter 30). In genetically hypertensive rats, circadian rhythm disturbances in blood pressure were found (Ikonomov et al., 1998), while also the SCN itself is affected, since the light-entrainment response is disturbed and accompanied by a suppressed c-Fos mRNA expression in the SCN (Lemmer et al., 2000). The occurrence of non24-hour rhythms was indeed more frequent in hypertensive individuals than in normotensive subjects. The period of circadian rhythms is disturbed in about 30% of the hypertensive subjects (Abitbul et al., 1997). Generally a significant nocturnal fall in blood pressure takes place (“dippers”). In some essential hypertensives this nocturnal

fall in blood pressure is not found (“nondippers”). Nondipper essential hypertensive patients are subject to central sympathetic hyperactivity, responsible for quantitative and qualitative alterations of sleep (Pedulla et al., 1995). In relation to the circadian disturbances in at least some of the hypertensive patients, it is interesting to note that melatonin was found to be beneficial in the treatment of essential hypertension (Cagnacci et al., 1998b; Chapter 4.5). Moreover, in essential hypertension, morphological signs of hypertrophy were observed in the nucleus habenularis epithalamis (Postnov et al., 1974), which is a field of termination of suprachiasmatic nucleus efferents. The abnormalities in diurnal rhythms in hypertensive patients suggested deteriorations in the functioning of the SCN. Indeed, the staining for three main neuronal populations, i.e. vasopressin, VIP and neurotensin, appeared to be reduced by more than 50% in hypertensives as compared to controls, indicating a serious dysregulation of the biological clock in hypertensive patients. The difficulty in adjusting from inactivity to activity, based upon the observed SCN alterations, may be involved in the morning clustering of cardiovascular events (Goncharuk et al., 2001). Various systems other than vasopressin may be involved in essential hypertension. Animal experiments have shown that programming of the hypothalamopituitary–adrenal axis by events in fetal life may be one of the mechanisms linking reduced size at birth to raised blood pressure in later life (Clark, 1998). In this connection, it is of interest that cortisol may increase blood pressure in a dose-dependent fashion (Kelly et al., 1998), and that Goncharuk et al. (2002) observed strongly increased CRH levels in hypertensive individuals (Chapter 8.5d). Hypertension is also observed in hereditary glucocorticoid resistance (Lamberts, 2001; Chapter 8.5d). Moreover, an active mineralocorticoid receptor mutation has been found that causes early-onset hypertension that is markedly exacerbated in pregnancy. This mutation alters receptor specificity with progesterone and other steroids that become potent agonists (Geller et al., 2000). Other potential factors involved in essential hypertension are leptin, melanocortin-4 receptors, NPY, angiotensin II, the sympathetic nervous system and obesity, which may account for 65–75% of human essential hypertension (Hall et al., 2001). Hepatic osmoreceptors are sensitive to changes in portal blood osmolality and cause variations in plasma vasopressin and water diuresis, which prevent major systematic osmotic changes. Insensitivity of hepatic

2014 Ch 08

20/11/03

11:04 am

Page 197


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

osmoreceptors could contribute to maintaining increased plasma levels of vasopressin and negative free water clearance in some cirrhotic patients (Castellano et al., 1994). In the hepatorenal syndrome, also known as functional renal failure of liver cirrhosis, which evolves in patients with advanced liver disease, there is increasing difficulty in handling free water and the syndrome of inappropriate antidiuretic hormone secretion may occur (Chapter 22.6). The vasopressin levels are strongly increased and the circadian pattterns have disappeared. It is a nonosmotic type of vasopressin release that occurs, despite a reduced plasma osmolality (Pasqualetti et al., 1998), which can be considered as a compensatory reaction to the backward failure of the heart (Chapter 22.6). The syndrome is characterized by the production of urine that is almost totally devoid of sodium, by progressive oliguria, leading to anuria, and almost invariably has a poor prognosis. Current theories favor intense renal sympathetic and renin-angiotensin vasoconstriction as a reaction to the accumulation of vasodilatator mediators. Blockade of V2-renal vasopressin receptors by orally active nonpeptide vasopressin receptor antagonists appears effective in correcting abnormal water handling (Paranjape and Thibonnier, 2001; Wong et al., 2003). However, low-dose vasopressin may also have a beneficial effect, due to its vasoconstrictor features (Inoue et al., 1998; Schrier et al., 1998; Eisenman et al., 1999). In patients with chronic renal failure, whether or not they were undergoing hemodialysis and in the nephrotic syndrome, plasma vasopressin and neurophysin levels are significantly higher (Legros and Franchiment, 1972; Pyo et al., 1995). Plasma vasopressin levels are increased, partly due to decreased clearances. The polyuria associated with renal failure is, at least partly, the result of decreased expression of collecting duct aquaporins (Cadnapaphornchai and Schrier, 2000). Plasma vasopressin increases in patients with congestive cardiac failure, stimulated by a decrease in cardiac output or peripheral arterial vasodilation (Szantalowicz et al., 1981; Cadnapaphornchai and Schier, 2000). The increased vasopressin levels may contribute to the vasoconstriction in that syndrome (Szatalowicz et al., 1981; Rondeau et al., 1982; Goldsmith, 1987), while some authors described patients with congestive heart failure who showed increased plasma vasopressin levels. Others showed suppressed levels (Bichet et al., 1986). In conscious dogs, it was shown that a small fall in left arterial pressure produces a rise in plasma vasopressin levels and antidiuresis (Yaron and Bennett, 1978). It is proposed that

197

cardiac failure is associated with arterial underfilling, which stimulates the nonosmotic release of vasopressin with upregulation of aquaporin-2 water channels (Chapter 22.6). The resultant water retention may be reversed by orally active V2 renal receptor antagonists (Schrier et al., 1998; Paranjape and Thibonnier, 2001). The plasma and CSF levels of vasopressin and oxytocin are elevated in meningitis (Unger et al., 1971; Garcia et al., 1981). Either low, high, or normal vasopressin levels have earlier been reported in septic or cardiogenic shock (Landry et al., 1997; Goldsmith, 1998; Chen et al., 1999a; Landry and Oliver, 2001; Patel et al., 2002a). The observation that in patients with septic shock, short-term vasopressin infusion spared conventional (i.e. norepinephrine) use, while maintaining blood pressure and cardiac output. Increased urine output and creatinine clearance were observed during vasopressin infusion (see 8.4b) and fits best with the idea that septic shock is associated with vasopressin deficiency and a hypersensitivity to its exogenous administration (Patel et al., 2002a). In patients with cerebrovascular accidents such as subarachnoid hemorrhage, especially if they had bled from the anterior communicating artery, in intracranial hemorrhage and in ischemic stroke, increased concentrations of vasopressin were found in CSF and sometimes also in plasma (Mather et al., 1981; Sørensen, 1986). The increase in plasma vasopressin levels in stroke correlates with the size of the lesion and the severity of the neurological deficits. The possibility that vasopressin may play a role in neuronal damage following cerebral ischemia has been put forward (Barreca et al., 2001). In addition, increased intracranial pressure was proposed to be associated with elevated levels of vasopressin in the CSF (Sørensen et al., 1984; Sørensen, 1986). In contrast to this report, the concentration of vasopressin in plasma was found not to be related to intracranial pressure in a single patient, reported by Bohnen et al. (1992), who did not find any changes in CSF vasopressin levels with increasing intracranial pressure. However, recent experimental evidence also indicates that centrally released vasopressin induces capillary water permeability and may play a causal role in vasogenic and ischemic brain edema. A dose-dependent decrease in brain edema was found after treatment of rats with a selective vasopressin V1 receptor antagonist in cold-brain-injured rats (Bemana et al., 1999). Glucocorticoids have a suppressive effect on the expression of processed vasopressin, while the precursor of vasopressin is not decreased (Erkut et al., 1998, 2002;

2014 Ch 08

198 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:04 am

Page 198

D.F. SWAAB

Fig. 8.24), indicating a processing disturbance. In glucocorticoid-deficient patients, vasopressin levels are increased (Ahmed et al., 1967; for references see Erkut et al., 2002). The vasopressin gene promotor contains a glucocorticoid response element. In this way, glucocorticoids inhibit vasopressin promotor activity, explaining part of the nonosmotic increase in vasopressin secretion with glucocorticoid deficiency. Other disorders that are accompanied by elevated plasma vasopressin levels are pain (Chapter 31.2) and smoking, which causes a several hundred-fold increase in vasopressin levels (Husain et al., 1973; Kendler et al., 1978). Surgery increases vasopressin levels (Philbin and Coggins, 1978). The increased vasopressin release during laparoscopic donor nephrectomy is presumed to be related to increased abdominal pressure (Hazebroek et al., 2002). After surgery, probably due to vasopressin release, hyponatremia and hyponatremic encephalopathy may develop. Menstruant women are about 25 times more likely to die or have permanent brain damage than men or postmenopausal women when this disorder develops (Ayus et al., 1992). After induction of anesthesia, there is a marked increase in the plasma and urinary aquaporin-2, although no changes in plasma osmolality or serum sodium concentrations have been observed (Otsuka et al., 1999). In acute illness of status asthmaticus, elevated plasma vasopressin levels are found (Baker et al., 1976). In various pulmonary pathologies, such as bronchopneumonia, lung edema, asthmatic bronchitis, hypoxia, lung carcinoma, lung tumors, sepsis, and lung emphysema in the vasopressin neurons of the SON, we observed a high percentage of multinucleated neurons, especially in young subjects. In those subjects with the highest proportion of multinucleated neurons, vasopressin mRNA expression was lowest. Multinucleated neurons thus appear to be a hallmark of pulmonary pathology, especially in young subjects (Ishunina et al., 2000a), although the mechanism causing this phenomenon cannot be explained at present. Angiotensin II infusion induced increased levels of both vasopressin and oxytocin in human beings (Vallotton et al., 1983; Chiodera et al., 1998a,b). Vasopressin plasma levels also rise during migraine attacks (Hasselblatt et al., 1999; Chapter 31.3b). For the abnormal secretion of vasopressin and oxytocin in schizophrenia, see Chapter 27.1. In autism, swollen axon terminals (spheroids) are found in the paraventricular nucleus (Weidenheim et al., 2001). Increased vasopressin plasma levels have been reported in patients with inner ear disorders caused by endo-

lymphatic hydrops, including Ménière’s disease. Indeed, animal experiments have shown that chronic administration of vasopressin may induce endolymphatic hydrops (Takeda et al., 1995, 2000). Since the frequency of gastroduodenal ulceration is lower in humans suffering from hypothalamic diabetes insipidus, endogenous vasopressin could have a harmful effect on the gastroduodenal mucosa (Pávó et al., 2000). Bed rest and supine position suppress vasopressin plasma levels while +3G acceleration provokes a strong increase in vasopressin (Keil and Ellis, 1976; Pump et al., 1999). In astronauts, a deficit in plasma volume occurs; body fluid regulation in space appears to depend partly on activation of the vasopressin system (Drummer et al., 2002). Water immersion causes a suppression of vasopressin release (Hammerum et al., 1998) and a graded increase in arterial blood pressure, central blood volume and cardiac output (Gabrielse et al., 2000a,b). Hypersecretion of vasopressin was observed in patients with pituitary tumors (Pawlikowski and Lesnik, 1971). Blockade of V3 pituitary vasopressin receptor by orally active nonpeptide vasopressin receptor antagonists was proposed as a potential therapeutic strategy in the case of ACTH-secreting pituitary tumors (Paranjape and Thibonnier, 2001). Abuse of 3,4-methylene dioxymethamphetamine (MDMA, commonly called ecstasy) has been associated with acute hyponatremia due to increased vasopressin secretion (Henry et al., 1998; Fallon et al., 2002). 8.4. (b) Vasopressin administration in various disorders DDAVP (1-desamine-8 D-arginine vasopressin = desmopressin) is used to treat central diabetes insipidus (see Chapter 22.2a). Pitressin has been used for the emergency control of gastrointestinal bleeding from esophagal varices, also known as the Mallory–Weiss syndrome. When administered intravenously, 20 pressure units produce a significant decrease in portal blood flow and pressure by vasoconstriction of portal arterioles in the liver. The rationale for this use of vasopressin is based on its ability to constrict splanchnic arterioles (Dill et al., 1971; Kraft et al., 1991). However, the vasoconstrictor activities do not spare the coronary circulation; massive myocardial injury has been reported in an old paper as a side effect (Beller et al.,

2014 Ch 08

20/11/03

11:04 am

Page 199


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

1971). Because of the coronary constriction caused by vasopressin, it is of considerable interest that there is a vasopressin system in the heart as well. Both de novo synthesis in the heart and release of this peptide by cardiac efferents have been shown (Hupf et al., 1999). In addition, paradoxical hypotension and bradycardia have been reported following vasopressin infusion; these side effects might also reflect cardiac ischemia (Kraft et al., 1998). Massive pulmonary haemorrhage due to leptospirosis has also been treated successfully with desmopressin infusions (Pea et al., 2003). Pericervical injection of vasopressin into the uterus in case of abdominal hysterectomy significantly reduces blood loss without increasing morbidity, infection or blood pressure (Kammerer-Doak et al., 2001; Okin et al., 2001). DDAVP has been used in children for prophylaxis of bleeding, e.g. in adenotonsillectomy, and to stop bleeding in mild hemophilia, type I von Willebrand’s disease, in hemophilia B and in patients with various platelet function defects. It reduces the bleeding diathesis of children with uremia and drug-induced bleeding complications. It may act on platelets (Sutor, 1998) and bleeding time (Fuse et al., 2003) and is used prophylactically for surgery (Ehl and Sutor, 2000; Fuse et al., 2003). Desmopressin raises endogenous factor VIII and von Willebrand factor 3–5 times, and thereby corrects, in von Willebrand type I disease, the intrinsic coagulation and the primary hemostatic defects. In type III or II disease, desmopressin is ineffective (Mannucci, 2001). However, Allen et al. (1999) cautions that even a single dose of desmopressin may lead to substantial hyponatremia if accompanied by aggressive, intravenous hydration and poor oral intake. For children with von Willebrand disease who require adenotonsillar surgery, they therefore recommend a protocol for fluid and electrolyte management. As the function of DDAVP is dependent on requirements such as a resting level of factor VIII, a von Willebrand factor of at least 10% and a platelet number of at least 50 109/l, a test dose has to be given to the child to predict its hemostatic effect. Side effects such as facial flushing, headache, increase in pulse rate and drop in systolic blood pressure are mild and transient. Children under the age of 18 months should be under close surveillance in order to prevent water intoxication and electrolytic imbalance (Sutor, 2000). According to some studies, vasopressin injection followed by defibrillation in cardiopulmonary reanimation seems to give better results than epinephrine injections (Wenzel et al., 1998), but another study failed to detect

199

any survival advantage for vasopressin over epinephrine (Stell et al., 2001). The baroreflex-mediated secretion of vasopressin is defective in a variety of vasodilatory shock states such as postcardiotomy shock. Administration of vasopressin markedly improves vasomotor tone and blood pressure in catecholamine-resistent septic and postcardiotomy shock (Morales et al., 2000; Dünser et al., 2001). Vasopressin has been used for the treatment of refractory hypotension after cardiopulmonary bypass and may sometimes provide a dramatic improvement of hemodynamic conditions (Overand and Teply, 1998). Vasopressin infusion of 0.01–0.04 U/min in patients with septic shock and vasodilatory shock due to systemic inflammatory response syndrome may cause an increase in urinary output, and pulmonary vascular resistance may decrease. However, infusions of more than 0.04 U/min may lead to adverse effects (Holmes et al., 2001; Patel et al., 2002). In advanced vasodilatory shock the combined infusion of vasopressin and noradrenaline proved to be superior to noradrenaline alone (Dünser et al., 2003). Terlipressin or long acting vasopressin analogue induced a significant rise in blood pressure in patients with a noradrenaline resistent septic shock (O’Brien et al., 2002). Vasopressin administration suppresses leptin levels, an effect from which the physiological meaning is not yet clear (Rubin et al., 2003). 8.5. Corticotropin-releasing hormone (CRH) neurons in the PVN . . . the central nervous system plays a key role in the response to stressful physical and psychological events in the environment, and does so by means of factors that are channeled through the hypothalamo-hypophysial portal vascular system to control the secretion of ACTH by the pituitary gland. G.W. Harris. 1955

Corticotropin-releasing hormone (CRH) is not only a crucial neuropeptide in the regulation of the HPA axis, the final common pathway in the stress response, as indicated in the quote above, but also has central effects, including cardiovascular regulation, respiration (Chapter 30), appetite control (Chapter 23), stress-related behavior and mood (Chapter 26.4), cerebral blood flow regulation (Lehnert et al., 1998) and stress-induced analgesia (Lariviere and Melzack, 2000). In addition, the hormonal end-product of the HPA axis, cortisol, is one of the most

2014 Ch 08

200 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:04 am

Page 200

D.F. SWAAB

powerful endogenous feedback compounds on the proinflammatory signal transduction machinery (Rivest, 2001; Chapter 21.2). The stress response consists of alterations in levels of anxiety (Chapter 26.4d), a loss of cognitive and affective flexibility, activation of the HPA axis and autonomic nervous system (Chapter 30) and inhibition of vegetative processes that are likely to impede survival during a lifethreatening situation, such as sleep (Chapter 30.7), sexual activity (Chapter 24) and endocrine programs for growth and reproduction (Gold and Chrousos, 2002). CRH is a 41 amino acid peptide that was isolated from ovine hypothalamus in 1981 by Vale and co-workers. It is produced by parvicellular neurons of the PVN (Raadsheer et al., 1993; Fig. 8.23; Koutcherov et al., 2000), but a considerable amount of CRH neurons was also found in the periventricular and infundibular nuclei. The perifornical area and the dorsomedial nucleus contain only scattered CRH neurons (Mihaly et al., 2002). CRH plays a key role in the response of the HPA axis to stress, by stimulating the release of ACTH from the anterior pituitary gland (Vale et al., 1981, 1983). ACTH stimulates the adrenal to produce cortisol, the main corticosteroid in human beings and corticosteroids inhibit CRH production (Erkut et al., 1998, 2003, in press; Fig. 8.24). Corticosteroids act on many organs and various brain areas by two types of receptor, i.e. mineralocorticoid receptors and glucocorticoid receptors (Hollenberg et al., 1985; Reul and De Kloet, 1985; Arriza et al., 1987). The ACTH-releasing activity of CRH is strongly potentiated by vasopressin (AVP) (Gillies et al., 1982; Rivier and Vale, 1983; Rubin et al., 2003), when released into the portal capillaries. CRH and vasopressin are colocalized in the PVN and increased activity of CRH neurons is accompanied by a higher proportion of CRH neurons that express vasopressin (Raadsleer et al., 1993, 1994; Erkut et al., 1995; Fig. 8.23). It would be interesting to know whether a similar potentiating effect exists for central effects. In the pituitary, vasopressin triggers ACTH release through a specific receptor subtype termed V3 or V1b, which is almost exclusively expressed by pituitary corticotrophs and some corticotroph tumors (René et al., 2000). Although in rat also, a synergistic action of oxytocin on CRH-induced ACTH release was reported, oxytocin did not affect basal vasopressin levels or CRH-induced ACTH release in humans, and even appeared to inhibit the potentiating effect of vasopressin on CRH-induced ACTH release (Suh et al., 1986). The inhibiting effect of oxytocin on ACTH release has now

Fig. 8.23. Immunocytochemical double staining on frontal paraffin sections (6 m) through the human hypothalamus. CRH cells stained blue, AVP cells red, and neurons containing both CRH and AVP stain purple. (A) section through the PVN of a young patient (male, 37 years of age), not showing colocalization of AVP and CRH; (B) section through the↓ PVN of an old patient (male, 74 years of age) showing red (*), blue ( ) and purple cells (▲). (C) section through the SON of the same patient as in B, showing only red AVP cells. (D) staining by solid-phase AVP-preadsorbed anti-AVP and with 10–5 M preincubated CRH anti-CRH of a section through the PVN of the same patient as in B, showing no immunoreactive cells. Bar = 25 m. (From Raadsheer et al., 1993; Fig. 3, with permission.)

been confirmed in various species (Legros, 2001), providing yet another example of opposite actions of vasopressin and oxytocin. In women, both suckling and breast stimulation induces a significant increase in oxytocin plasma levels and a decrease in plasma ACTH, which agrees with an inhibitory influence of oxytocin on ACTH and cortisol secretion in humans (Chiodera et al., 1991). CRH and vasopressin mediate ACTH release via different second-messenger systems. CRH also activates G protein-linked adenylate cyclase, leading to cAMP formation and protein kinase-A activation. Vasopressin

2014 Ch 08

20/11/03

11:04 am

Page 201


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

201

Fig. 8.24. Estimated number of CRH-immunoreactive cells in the hypothalamic PVN (A), the total integrated immunoreactivity for AVP (B), the mean staining intensity of AVP-immunoreactive cells in the PVN and SON (C), and the mean staining intensity for OXT in the PVN of the controls and the corticosteroid-exposed subjects (D; CST). The numbers of the plotted data refer to the numbers of subjects in the paper. The bars and error lines represent the mean and SEM, and the p values are according to the Mann–Whitney U test. Note that corticosteroids do not only decrease the number of CRH neurons in the PVN, but also the amount of vasopressin staining in the SON and PVN, while OXT stays unaffected. (From Erkut et al., 1998; Fig. 2.)

2014 Ch 08

202 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:04 am

Page 202

D.F. SWAAB

activates phospholipase C; prostaglandin may have specific interactions with both pathways. In this respect it is interesting that aspirin, an inhibitor of prostaglandin synthesis, significantly reduces the cortisol response to vasopressin in humans (Nye et al., 1997). Atrial natriuretic factor, which is elevated in patients with panic disorder, inhibits the CRH-stimulated release of ACTH in humans (Kellner et al., 1992; Dieterich et al., 1997), even though the HPA axis is activated in this disorder (Chapter 26.7). NPY-containing axonal varicosities are juxtaposed to both dendrites and perikarya of the majority of CRH neurons residing in the PVN, and in periventricular and infundibular nuclei. The NPY axons either form baskets around their perikarya or they completely ensheath the emanating CRH dendrites. Since only a small proportion of CRH neurons was contacted by agoutirelated peptide axons (a peptide-colocalizing with NPY in the infundibular nucleus), the NPY innervation is presumed to be derived mainly from regions outside the infundibular nucleus (Mihaly et al., 2002). In the fetus, stimulation of CRH neurons by NPY may be involved in the initiation of labor (Chapter 18.5a). In postmortem tissue we found that, following different types of corticosteroid treatment in different disorders or during the presence of high levels of endogenous corticosteroids produced by a tumor, not only CRH-expressing neurons are hardly detectable any longer, but also vasopressin expression in the SON and PVN is strongly decreased. Oxytocin neurons, however, were not affected (Erkut et al., 1998, 2002; Fig. 8.24). This illustrates that in the human brain, negative cortisol feedback is present, both in the CRH cells that coexpress vasopressin, and in those that do not. Secondly, it makes it clear how important information on the use of medicines may be for a study on postmortem brain tissues. It should be noted also that inhaled corticosteroids may inhibit the HPA axis (Levine and Boston, 2000). The human equivalent of urocortin, a CRH-related peptide, has been cloned and characterized. It has 45% homology with CRH. Based upon its location in the rat brain, it was proposed to be a natural ligand for the type-2 CRH receptor (Donaldson et al., 1996; Dieterich et al., 1997; see below). In animal experiments urocortin has anxiogenic-like properties (Moreau et al., 1997) and delays gastric emptying (Taylor, 1999). Neither urocortin immunoreactivity nor mRNA hybridization signals were, however, detected in human hypothalamus or pituitary stalk, whereas this peptide was found in extrahypothalamic sites such as Purkinje cells in the cerebellum and

anterior horn cells in the human spinal cord (Lino et al., 2000). More recently, human stress-coping peptide and stress-coping-related peptide were identified as specific ligands for the CRH-2 receptor (Hsu and Hsueh, 2001; see below). Also a hypothalamic corticotropin-release inhibiting factor is present, which inhibits ACTH synthesis. It has been proposed that this would be identical to the 22-amino acid peptide prepro TRH (178–199), which points to an integrated regulation of the HPA and thyroid axes (Redei et al., 1995). CRH immunoreactivity is present in the human hypothalamus, in parvicellular neurons of the PVN (Fig. 8.23). CRH-vasopressin double-staining neurons are also parvicellular (Fig. 8.25; Raadsheer et al., 1993; Raadsheer, 1994). CRH fibers innervate the median eminence, where CRH is released into the portal vessels, and other CRH fibers run into the brain (Raadsheer et al., 1993). CRH fibers are found to innervate LHRH neurons in the infundibular nucleus, which may be a substrate for CRH controlled LHRH secretion (Dudás and Merchenthaler, 2002b). CRH-positive cells and fibers are present in the human brain from fetal week 12–16 onwards (Bresson et al., 1987). Originally only few CRH-expressing neurons of the PVN were found in the rostral PVN (Pelletier et al., 1983; Raadsheer et al., 1993). Later, a more sensitive technique managed to locate CRH neurons in the most rostral part of the PVN and in the medially situated parvicellular nucleus and posterior subnucleus of the PVN as well (Koutcherov et al., 2000). Since interleukin-1 (IL1) mediates the acute phase reaction, it is interesting that a dense innervation of this cytokine is present in the PVN (Breder et al., 1988). Recent studies show, moreover, that IL-1 is present not only in glia cells in the hypothalamus, but also in oxytocin-producing neurons of the PVN and in oxytocin neurons in the dorsal cap of the SON and in the islands of neurosecretory neurons in between these nuclei. Changes occur in these neurons in MS (Chapter 21.2), but the functional meaning of the neuronal colocalization is not yet clear. CSF-CRH reflects the activity of non-HPA axis sources of CRH. Although both yohimbine and naloxone stimulate the HPA axis, only yohimbine appeared to have stimulatory effects on CSF-CRH (Vythiligam et al., 2000). Two types of G-protein-coupled brain CRH receptors, CRH-1 and CRH-2, have been extensively characterized and localized in the rat. CRH-1 receptor expression is very high in cerebral cortex, septal region, amygdala and pituitary. The receptor mediates the “fight or flight” response, characterized by the activation of the

2014 Ch 08

20/11/03

11:04 am

Page 203


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

203

receptors, suggesting that CRH-binding protein might limit CRH actions at CRH receptors. It has also been presumed that CRH-binding protein may stop placental CRH from overstimulating the pituitary, and that, when pregnancy progresses, CRH-binding protein levels fall, which causes a diminished blockade of CRH effects (Thomson, 1998). A large proportion of the CRH in human brain thus seems to be complexed to CRH-binding protein, and the highest levels of both are found in the hypothalamus (Behan et al., 1997). (a) Fetal CRH and birth Fig. 8.25. Distribution patterns of only CRH-expressing neurons (open triangle), both CRH and AVP-expressing neurons (closed triangle), and only AVP-expressing neurons (open circle) in frontal paraffin sections (6 m) of hypothalami of 13 control subjects immunocytochemically double labeled for AVP and CRH and of all AVP-expressing neurons (closed circle) in adjacent sections immunocytochemically stained for AVP only. All neurons contained a cell nucleus. Note that in the double-labeled sections, all CRH-expressing neurons were parvocellular, whereas almost all exclusively AVP-expressing neurons (open circle) were magnocellular. In the sections which were only stained for AVP (closed circle), both magno- and parvocellular neurons were found. (From Raadsheer, 1994; Fig. 1, p. 133.)

CRH-ACTH-cortisol axis (Hsu and Hsueh, 2001). CRH2 receptor expression is confined to subcortical structures such as the lateral septum, ventromedial hypothalamic nucleus, choroid plexus, supraoptic and paraventricular nucleus. The type-2 CRH receptor mediates the stress-coping response during the recovery phase of stress. Human stress coping and stress-related peptide are specific ligands for the CRH-2 receptor (Hsu and Hsueh, 2001). Two isoforms of the CRH-2 receptors are distinguished, CRH-2 and 2, which differ in their N-terminal part only. The human CRH-2 is found mainly in the heart or skeletal tissues, while the CRH-2 form is present mainly in the brain. The human CRH-2 receptor has a functional response to urocortin and also binds CRH (Valdenaire et al., 1997). In addition, CRH-binding protein is present in plasma and also in the brain. CRHbinding protein is thought to modify the actions of CRH by intra- and extracellular mechanisms. CRH-binding protein also binds human urocortin with high affinity (Donaldson et al., 1996; Dieterich et al., 1997). On a molar base, CRH-binding protein is present in at least 10-fold higher amounts in brain regions than CRH

Labor is too important to leave it entirely to our mother.

Normal development of the fetal HPA axis leads to a late gestational cortisol surge that is essential for the regulation of intrauterine homeostasis and the timely differentiation and maturation of vital organ systems, including the lung and central nervous system, and is necessary for immediate neonatal survival after birth (Bolt et al., 2002). Stressful events experienced in fetal and early neonatal life can produce enduring changes in programming of the HPA axis function and predispose to psychopathology, such as depression in later life (Chapter 26.4). Small size at birth is also associated with an alteration in the set point of the HPA axis, resulting in increased cortisol responsiveness and increased risk of depression in adulthood (Phillips, 2001; Thompson et al., 2001). In addition, acting together with the placenta, the HPA axis may control the normal timing of parturition (Ng, 2000). Before midterm, placental ACTH, placental CRH, and estrogens may regulate fetal adrenal steroidogenesis. After this period ACTH produced by the fetal pituitary gland seems to be important (Bolt et al., 2002). In a series of observations in sheep, G.C. Liggins has introduced the concept that fetal hypothalamic CRH may be the start of a cascade that is primarily responsible for the initiation of labor (for references see Swaab et al., 1978; McMillen, 1995; Liggins, 2000; Chapter 18.1c). The observations that CRH mRNA levels increase in the fetal sheep PVN during late gestation, and that the infusion of a CRH antagonist on the fetal side delays parturition with a week (Chan et al., 1998) agree with this intriguing possibility. A crucial experiment was performed by McDonald and Nathanielsz (1991), who showed that, after bilateral selective stereotactic destruction of the fetal PVN, gestation is prolonged

2014 Ch 08

204 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:04 am

Page 204

D.F. SWAAB

in sheep. In the human fetus, CRH-containing cell bodies are present from 19 weeks of pregnancy onwards and CRH-containing nerve fibers are found in the median eminence from the 16th week of fetal life onwards (Bresson et al., 1985). During the last weeks of gestation, a surge of cortisol occurs. However, both the adrenal and the pituitary are immature in their reaction to CRH in human preterm infants. Preterm infants that are small for their gestational age have a lower adrenal response to a stimulation by ACTH. This may be of importance in neonatal morbidity and possibly in the development of disease in later life (Bolt et al., 2002a,b,c). Both hyperand hypocortisolism may arise as a consequence of fetal programming of the HPA axis during intrauterine life. In adulthood such changes may result in coronary heart disease, insulin-resistance syndrome (Kajantie et al., 2002) and depression (Chapter 26.4). The human fetal brain does not seem to determine the mean gestational length in the same way it does in sheep, but rather to be responsible for the strict timing of the moment of birth around 40 weeks of pregnancy, i.e. preventing pre- and postmaturity. In human anencephalics, who lack a hypothalamus, this mechanism is absent and the timing of birth seems to be entirely lost. One-third of these children are born prematurely and one-third postmaturely (Honnebier and Swaab, 1973; Chapter 18.1c; Fig. 18.5). The observation that ACTH administration to the human fetus initiates labor in postmature pregnancies and not at term, while a tendency towards prolongation of labor has been reported following corticosteroid administration to pregnant women, suggests that the fetal hypothalamopituitary axis in humans might indeed be important for timing the moment of birth within a narrow period around 40 weeks of pregnancy (Swaab et al., 1978). Interestingly, pregnant women with high stress and anxiety levels are at risk for premature birth (Mulder et al., 2002), pointing to a role of the maternal HPA axis in the timing of labor. Glucocorticoids may act by stimulating prostoglandin synthesis in fetal membranes and by increasing intrauterine CRH expression (Whittle et al., 2001). Administration of corticosteroids to pregnant women at risk for early preterm birth is an established intervention with proven reduction in the rates of mortality, decreasing the incidence of neonatal respiratory distress syndrome and intraventricular hemorrhage. Antenatal steroids also reduce the incidence of periventricular leukomalacia and ventriculomegaly and improve long-term health (Finer et al., 2000). However, many obstetricians now prescribe repeated courses of

corticosteroids for those women who continue to be at risk for preterm delivery but remain undelivered 7 days after the previous course. In a cohort study, repeated antenatal corticosteroid courses were found to be accompanied by a significant decrease in birthweight and head circumference of the child (French et al., 1999). However, determination of a causal relationship would require a randomized trial of repeated corticosteroid use. On the basis of the short-term positive results of prenatal corticosteroids, currently over 50% of the extremely low birthweight infants are likely to receive postnatal steroids during their stay in the neonatal unit. However, there are significant short-term and longterm adverse effects of this treatment, such as a decrease in somatic and head growth, an increased incidence of neurological abnormalities such as cerebral palsy, and mental and psychomotor development disturbances. There is, of course, extensive experimental literature on the inhibitory effects of corticosteroids on brain development in rodents. This led Finer et al. (2000) to summarize the literature by “short-term gain, long-term pain?” and to plead for good clinical trials that include long-term neurodevelopmental outcome and somatic growth. Intrauterine infection is associated with activation of the fetal HPA axis, increased androgen, cortisol and estrogen synthesis and preterm labor within 7 days. This may be the result of proinflammatory cytokines. It is not known, however, what the nature or source is of the stimulus that results in activation of the fetal HPA axis during the last weeks of gestation (Bolt et al., 2002a,b,c). It has been hypothesized that hypoglycemia may be one of the candidates acting as a physiological stressor in late gestation, which is in full agreement with the citation of Hippocrates quoted at the beginning of Chapter 18.1c. There is indeed some evidence that parturition is preceded by an increased sensitivity of the fetal HPA axis to prevailing blood glucose concentrations (McMillen et al., 1995). It has also been proposed that NPY production in neurons of the fetal infundibular nucleus would be activated by fetal undernutrition and by glucocorticoids that may be released by preterm stress or by HPA axis maturation at term. NPY innervation of the fetal PVN would thus activate the fetal HPA axis. An intriguing possibility for the initiation of labor is thus a positive feedback of cortisol, stimulating NPY, which increases CRH activity and thus cortisol production. This positive feedback may be initiated by the decreased fetal glucose concentrations that occur during late gestation. Moreover, a decrease

2014 Ch 08

20/11/03

11:04 am

Page 205


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

in leptin may be a signal for NPY expression (Warnes et al., 1998). In addition, the placenta seems to produce factors that contribute to the sustained elevated cortisol concentrations that are needed to initiate parturition (Thorburn et al., 1991), and, at the time of labor, CRH receptor subtype 1 is upregulated in the human myometrium (Stevens et al., 1998a). The placenta can be viewed from a neuroendocrine perspective as a “third brain”, linking maternal and fetal neuroendocrine functions (Yen, 1994). Indeed, during pregnancy, CRH is also secreted by the human placenta, contributing to high plasma levels (McLean and Smith, 1999). An exponential rise in maternal plasma CRH is found in advancing human pregnancy, concomitant with a fall in concentrations of the specific CRH-binding protein. In women who are destined to experience preterm delivery, CRH levels are higher, whereas in women destined to have post-term delivery, they are lower, suggesting that CRH acts as a trigger for parturition (McLean et al., 1995; Korebrits et al., 1998; Smith, 1998; Majzoub et al., 1999; McLean and Smith, 1999). CRH thus seems to be coupled to a placental clock which determines the length of gestation. Umbilical cord blood CRH levels are extremely elevated in growthretarded fetuses (Goland et al., 1993). Cortisol is a trigger for not only increased production of fetal hypothalamic CRH, but also placental CRH (Clifton and Challis, 1997; Majzoub et al., 1999; Whittle et al., 2001), linking these systems together. CRH production of the placenta is, in addition to glucocorticoids, also stimulated by oxytocin and vasopressin and a number of other factors (McLean and Smith, 1999). Stress causing an elevation of CRH either in the mother, fetus or placenta may thus result in premature onset of labor (Majzoub et al., 1999). CRH-receptor antagonists are therefore proposed to be therapeutic compounds in the case of premature labor (Grammatopoulos and Chrousos, 2002). (b) CRH in relation to sex, aging, Alzheimer’s disease, depression and multiple sclerosis When joy is at its highest Sad thoughts run rife Youth and strength, how short they last How hopelessly we age! Emperor Han Wudi, Western Han Dynasty

Sex profoundly affects the dexamethasone-CRH test outcome: females, regardless of age, have an increased

205

hormonal secretion in comparison to males (Heuser et al., 1994a). Cortisol levels are 19% higher in women than in men (Laughlin and Barrett-Connor, 2000), while the timing of cortisol secretion alters with the stage of the menstrual cycle (Parry et al., 2000). In postmortem CSF we found higher cortisol levels in women than in men, both in Alzheimer’s disease patients and in controls (Erkut et al., 2002). In women, the CRH neurons are presumed to be more active than in men (Antonijevic et al., 1999), while cortisol production rate in men is clearly higher than that in women (Vierhapper et al., 1998; Shamim et al., 2000). In a small sample of controls we did not, however, find an indication of the presence of more CRH-expressing neurons in women (Raadsheer et al., 1994a). Although sexual dimorphism in cortisol metabolism is not dependent on estrogens (Toogood et al., 2000), ovarian steroids do increase HPA axis activity, enhance the HPA axis response to psychological stress, and sensitize the hypothalamopituitary-ovarian axis to stress-induced inhibition (Roy et al., 1999). The exact way in which the HPA axis activity is modulated by sex hormones has yet to be investigated. Estradiol may enhance HPA axis activity, e.g. by reducing glucocorticoid receptor function or by stimulation of CRH gene transcription, as the human CRH gene contains 5 perfect, half-palidromic estrogen-responsive elements (Torpy, 1997). In addition, sex differences in free cortisol levels may, at least partly, be explained by estradiolinduced changes in cortisol-binding protein levels (Kirsbaum et al., 1999). In premenopausal women a significant reduction of ACTH and cortisol is found after ovariectomy, while the response of ACTH, but not of cortisol, to CRH is reduced (De Leo et al., 1998). How the sex differences in the HPA axis may contribute to the sex difference in depression (see Table 8.5.I and Chapter 26.4) should be further investigated. We observed clear signs of activation of CRH neurons in both sexes during aging. The total number of CRHproducing neurons (Fig. 8.26) and the proportion of vasopressin coexpressing CRH neurons (Figs. 8.23 and 8.27) went strongly up from the age of 40 onwards (Raadsheer et al., 1994a,b). There are also sex-dependent effects of aging in the HPA axis hormone levels. Higher mean hypothalamic CRH levels were found in women than in men (Frederiksen et al., 1991), and the adrenal response to CRH is elevated in elderly women and in subjects with a chronic disease (Greenspan et al.,

2014 Ch 08

206 1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03

11:04 am

Page 206

D.F. SWAAB

1993). In response to a cognitive challenge, older women exhibited greater increases of salivary free cortisol than older men, whereas in young adults, men exhibited greater increases (Seeman et al., 2001). Healthy senior women who were exposed to mild stress and had elevated urinary free cortisol appeared to suffer more from impaired sleep and earlier awakenings. The sex differences in the basal levels of hormones of the HPA axis of elderly people are relatively small. In both men and women cortisol CSF and plasma levels increase progressively between 20 and 80 years of age (Guazzo et al., 1996; Swaab et al., 1996; Van Cauter et al., 1996; Laughlin and Barrett-Connor, 2000). In addition, an age-related decline in the plasma and CSF levels of the most abundant adrenal steroid dehydroepiandrosterone sulphate (DHEAS) has been found in both men and women. A specific function of DHEAS, however, remains to be found (Orentreich et al., 1992; Guazzo et al., 1996; Casson et al., 1998). DHEAS levels in plasma already decline from the first postnatal month onward to the age of 5 years. They subsequently rise rapidly from age 9 onwards in boys and age 7 in girls until peak concentrations are reached between ages 20 and 40. After that the levels decline and 20–30% of the peak concentrations are found by the ages of 70–80. The time course of DHEAS levels is similar. In a small study we did not find differences in cortisol or DHEAS levels between males and females, either in controls or in Alzheimer patients or patients with other possible forms of dementias (Tables 8.2.I and 8.2.II). However, in a large study, DHEA and DHEAS levels were 40% lower and cortisol 10% higher in women (Laughlin and BarretConnor, 2000). Moreover, a sex-dependent decrease in DHEA and DHEAS was observed with age (Laughlin and Barrett-Connor, 2000). It is interesting that cortisol and DHEAS have opposite effects on the brain, in particular on the hippocampal region. DHEA and DHEAS have been found to have antiglucocorticoid actions and are claimed to be effective as antidepressant drugs (Reus, 1997). In a prospective study among healthy elderly subjects, basal-free cortisol levels were positively related to cognitive impairment and cortisol levels after dexamethasone treatment. The age-related decline of DHEAS was confirmed by others and, in addition, an inverted but nonsignificant association between DHEAS and cognitive impairment and decline was observed (Kalmijn et al., 1998; Ferrari et al., 2001). In addition, the amplitude of the DHEAS circadian rhythm declined with age (Guagnano et al., 2001). DHEA and DHEAS belong to

Fig. 8.26. Linear regression between age and corticotropin-releasing hormone (CRH) cell number in the PVN estimated by the disector method. Filled circles and solid lines indicate control subjects; open circles and dashed lines indicate Alzheimer’s disease patients. A significant correlation was found between age and absolute CRH cell number for control subjects (rho = 0.66, p = 0.02). In Alzheimer’s disease patients, the age effect was almost significant (rho = 0.53, p = 0.06). (From Raadsheer et al., 1994a; Fig. 3, with permission.)

the “neurosteroids” because they can be synthesized de novo in the brain. Their concentrations are considerably higher in the brain than in other organs (Kroboth et al., 1999). In Alzheimer’s disease the DHEAS levels are reported to be decreased (Hillen et al., 2000). However, we could not confirm the presence of decreased DHEA levels in Alzheimer patients as reported earlier (Sunderland et al., 1989; Tables 8.2.I and 8.2.II). DHEAS treatment of age-advanced men but not of women decreased body fat and increased muscle strength, while an increase in insulin-like growth factor-1 was observed (Lamberts et al., 1997b; Morales et al., 1998). However, the many claims as to its usefulness during aging, including being “a fountain of youth” (e.g. Valenti, 1997; Casson et al., 1998), and inducing improved concentration in the elderly (Achermann and Silverman, 2001) should be further investigated. A double-blind study could not confirm the sense of well-being or improved sexual functions reported by others (Flynn et al., 1999). The level of the nocturnal plasma cortisol nadir increases progressively with aging in both sexes. An age-

2014 Ch 08

20/11/03

11:04 am

Page 207


1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

Fig. 8.27. Age-related increase in the ratios of the number of CRH neurons showing colocalization with AVP over those of CRH neurons not showing colocalization with AVP. Control subjects (), Alzheimer patients (). Correlations between age and ratio are highly significant (p = 0.02) in controls and Alzheimer’s disease patients and show Spearman’s correlation coefficients of 0.71 and 0.72, respectively. Note that the majority of CRH neurons colocalize AVP above the age of approximately 60 years. (From Raadsheer et al., 1994b; Fig. 1, with permission.)

related elevation in the morning acrophase occurred in women but not in men (Van Cauter et al., 1996). These data may be related to the age-related changes in SCN function (Chapter 4.3). Mild hypercortisolism is a frequent concomitant of Alzheimer’s disease and an increased glucocorticoid production seems to be an early feature of it (Rasmussen et al., 2001, 2002; Chapter 29.1). It appears to be related to the severity of dementia and the clinical progression of the disease, but not to the age or to the length of survival (Weiner et al., 1997; Swanwick et al., 1998). However, in a small study, we did not find a difference in basal plasma cortisol levels between elderly controls and probable Alzheimer patients (Tables 8.2.I and 8.2.II). Others pointed to the reduced amplitude of cortisol rhythm and higher evening and night-time levels (Ferrari et al., 2001). In the hypothalamus, the total number of CRH-expressing neurons in the human PVN increases with age in controls and Alzheimer’s disease brains to the same degree (Raadsheer et al., 1994a; Fig. 8.27), which is in agreement with the age-related increase in hypothalamic CRH content reported by Frederiksen et al.

207

(1991). Endocrine studies suggest an increased basal activity, a flattened diurnal amplitude of the HPA axis in the elderly (Deutschle et al., 1997) and increased post-dexamethasone cortisol levels with increasing age (O’Brien et al., 1966). In addition to increased levels of serum cortisol, dehydroepiandrosterone and androstenedione levels are significantly increased in AD patients (Rasmussen et al., 2002). We did not find an obvious sex difference in this pattern, in contrast with the higher hypothalamic CRH content in females, compared with males reported by Frederiksen et al. (1991), but the number of subjects we studied was quite small. The age-dependent increase in the absolute number of neurons expressing CRH (Fig. 8.23) in the PVN of both control and Alzheimer’s disease patients (Figs. 8.26 and 8.27) is interpreted as a sign that CRH neurons become increasingly active with age. Also, parvicellular neurons containing both CRH and vasopressin were found in increasing numbers of control subjects and Alzheimer patients (Fig. 8.27) ranging between 43 and 91 years of age, whereas the CRH neurons in the PVN of younger subjects (23–27 years of age) did not contain vasopressin (Figs. 8.23). The colocalization of vasopressin in CRH neurons is a measure of the activity of CRH neurons (De Goeij et al., 1991, 1992a,b,c; Bartanusz et al., 1993; Whitnall et al., 1993), which was much the same in controls and Alzheimer’s disease patients. In both groups a similar increase with age was present in the number of CRH neurons that colocalize AVP (Raadsheer et al., 1994b; Fig. 8.27). The third parameter for activity of CRH neurons measured in this material was the total amount of CRH-mRNA as determined by quantitative in situ hybridization. In contrast to the two parameters mentioned earlier, CRH-mRNA was found to be higher in Alzheimer’s disease patients than in age-matched controls (Raadsheer et al., 1995; Fig. 26.2; Chapter 26.4). In conclusion, CRH neurons in Alzheimer’s disease patients were moderately activated as compared to normal controls, as appeared from the difference in CRH-mRNA only, confirming the endocrine parameters indicating a moderate hypercortisolism. We could thus not confirm the increased CRH immunoreactivity in Alzheimer patients as reported by Powers et al. (1987), a qualitative study based on only three controls and two Alzheimer patients. However, our data do agree with those of Bissette et al. (1985), who did not find a difference in the hypothalamic CRH content in Alzheimer’s disease by means of radioimmunoassay. Why Behan et al. (1997) could not find a significant change in hypothalamic values

2014 Ch 08

208

20/11/03 11:04 am

TABLE 8.2.I

Group

Age (n)

Cortisol (mol/l)

DHEAS (mol/l)

Testosterone (nmol/l)

SHBG (mg/l)

FAI (nmol/mg)100

1700 h

0900 h

1700 h

0900 h

1700 h

0900 h

1700 h

0900 h

1700 h

0.51 ± 0.07

0.27 ± 0.02**

2.8 ± 0.7

2.7 ± 0.7

11.8 ± 1.8

10.9 ± 1.5

44 ± 8

44 ± 10

33 ± 6

34 ± 8

Alzheimer’s disease 70 ± 3 (6)

0.48 ± 0.09

0.20 ± 0.04*

2.5 ± 0.6

2.4 ± 0.6

15.6 ± 1.5

12.8 ± 1.1

42 ± 6

41 ± 6

39 ± 4

32 ± 4

Other dementia cases 72 ± 2 (6)

0.45 ± 0.03

0.26 ± 0.05*

2.3 ± 0.6

2.3 ± 0.5

15.7 ± 2.5

14.6 ± 2.1

42 ± 4

42 ± 5

37 ± 2

36 ± 4

76 ± 3 (8)1

1 Afternoon samples (17.00 h) were only taken in 7 control cases. * Different from values at 09.00 h; p < 0.05 (Mann–Whitney U test). ** Idem; p < 0.01.

Plasma levels of hormones were determined in 16 patients with the diagnosis of probable Alzheimer’s disease, 9 other dementia cases and 17 controls matched for age and sex at 9 a.m. and 5 p.m., with a view to circadian fluctuations. No differences between probable Alzheimer patients, other dementia cases and controls were observed for any of the parameters (Kruskal–Wallis one-way ANOVA, p > 0.32 for males and > 0.25 for females). It is concluded that baseline levels of steroid hormones do not seem to play an important role in the etiology of Alzheimer’s disease. E. Goudsmit, Ph. Scheltens, E. Endert, E. Fliers and D.F. Swaab, unpublished results.

D.F. SWAAB

0900 h Controls

Page 208

Plasma levels of cortisol, dehydroepiandrosterone sulphate (DHEAS), testosterone, sex hormone-binding globulin (SHBG) and free androgen index (FAI) in male control subjects, probable AD cases and patients suffering from “other dementias”.

2014 Ch 08

1 2 3 4 5 6 7 8 9 101 1 2 3 4 5 6 7 8 9 201 1 2 3 4 5 6 7 8 9 301 1 2 3 4 5 6 7 8 9 401 1 2 3 4 5 6 7 8 911

20/11/03 11:04 am

Group

Age (n)

Cortisol (mol/l)

DHEAS (mol/l)

0900 h

1700 h

0900 h

1700 h

Controls

76 ± 2 (9)

0.50 ± 0.0

0.29 ± 0.07** 1.9 ± 0.3

2.0 ± 0.3

Human Hypothalamus: Basic and Clinical Aspects, Part I: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Human Hypothalamus: Basic and Clinical Aspects, Part 2: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Sleep Disorders Part I: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Sleep Disorders Part II: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

History of Neurology: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Myopathies and Muscle Diseases: Handbook of Clinical Neurology Vol 86 (Series Editors: Aminoff, Boller and Swaab)

Neuromuscular Junction Disorders: Handbook of Clinical Neurology Vol 91 (Series Editors: Aminoff, Boller and Swaab)

Stroke Part II: Clinical manifestations and pathogenesis: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab) (Handbook of Clinical Neurology)

Malformations of the Nervous System: Handbook of Clinical Neurology Vol 87 (Series Editors: Aminoff, Boller and Swaab)

Neurology and Clinical Neuroscience

Handbook of Biomineralization - Medical and Clinical Aspects

Cancer: Basic Science and Clinical Aspects

Clinical Neurology

Clinical Neurology

Neuropsychology and Behavioral Neurology: Handbook of Clinical Neurology Vol 88

Clinical Neurology

Clinical Neurology

Clinical Neurology

Handbook of Basic Clinical Manipulation (Clinical Handbook of Psychotropic Drugs)

Basic and Clinical Endocrinology

Basic and Clinical Pharmacology

Clinical and Basic Immunodermatology

Clinical biochemistry : metabolic and clinical aspects

textbook of clinical neurology

Disorders of Consciousness: Handbook of Clinical Neurology

Textbook of Clinical Neurology

Handbook of Stroke Prevention in Clinical Practice (Current Clinical Neurology)

Malformations of the Nervous System Handbook of Clinical Neurology Series

Handbook of Neurocritical Care (Current Clinical Neurology)

Parkinson's Disease and Related Disorders Part I: Handbook of Clinical Neurology

Clinical Evaluation and Management of Spasticity (Current Clinical Neurology)

Human Hypothalamus: Basic and Clinical Aspects, Part I: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Human Hypothalamus: Basic and Clinical Aspects, Part 2: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Sleep Disorders Part I: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Sleep Disorders Part II: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

History of Neurology: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab)

Myopathies and Muscle Diseases: Handbook of Clinical Neurology Vol 86 (Series Editors: Aminoff, Boller and Swaab)

Neuromuscular Junction Disorders: Handbook of Clinical Neurology Vol 91 (Series Editors: Aminoff, Boller and Swaab)

Stroke Part II: Clinical manifestations and pathogenesis: Handbook of Clinical Neurology (Series Editors: Aminoff, Boller and Swaab) (Handbook of Clinical Neurology)

Malformations of the Nervous System: Handbook of Clinical Neurology Vol 87 (Series Editors: Aminoff, Boller and Swaab)

Neurology and Clinical Neuroscience

Handbook of Biomineralization - Medical and Clinical Aspects

Cancer: Basic Science and Clinical Aspects

Clinical Neurology

Clinical Neurology

Neuropsychology and Behavioral Neurology: Handbook of Clinical Neurology Vol 88

Clinical Neurology

Clinical Neurology

Clinical Neurology

Handbook of Basic Clinical Manipulation (Clinical Handbook of Psychotropic Drugs)

Basic and Clinical Endocrinology

Basic and Clinical Pharmacology

Clinical and Basic Immunodermatology

Clinical biochemistry : metabolic and clinical aspects

textbook of clinical neurology

Disorders of Consciousness: Handbook of Clinical Neurology

Textbook of Clinical Neurology

Handbook of Stroke Prevention in Clinical Practice (Current Clinical Neurology)

Malformations of the Nervous System Handbook of Clinical Neurology Series

Handbook of Neurocritical Care (Current Clinical Neurology)

Parkinson's Disease and Related Disorders Part I: Handbook of Clinical Neurology

Clinical Evaluation and Management of Spasticity (Current Clinical Neurology)

Recommend Documents