Hormones and their Actions Part I (New Comprehensive Biochemistry, Volume 18A)

HORMONES AND THEIR ACTIONS PART I New Comprehensive Biochemistry Volume 18A General Editors A. NEUBERGER London L...

Author: B. A. Cooke | R. J. B. King

65 downloads 743 Views 15MB 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

HORMONES AND THEIR ACTIONS PART I

New Comprehensive Biochemistry

Volume 18A

General Editors

A. NEUBERGER London

L.L.M. van DEENEN Utrecht

ELSEVIER Amsterdam - New York . Oxford

Hormones and their Actions Part I

Editors

B . A . COOKE Department of Biochemistry, Royal Free Hospital School of Medicine, University of London, Rowland Hill Street, London NW3 2PF, England

R.J.B. KING Hormone Biochemistry Department, Imperial Cancer Research Fund Laboratories, P.U. Box No. 123, Lincoln’s Inn Fields, London WC2A 3 P X , England

H.J . van der MOLEN Nederlandse Urganisatie voor Zuiver- Wetenschappelijk Onderzoek ( Z .W.O . ) , Postbus 93138, 2509 A C Den Haag, The Netherlands

1988 ELSEVIER Amsterdam . New York . Oxford

01988. Elsevier Science Publishers B.V. (Biomedical Division) All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the Publisher, Elsevier Science Publishers B.V. (Biomedical Division), P.O. Box 1527. 1000 BM Amsterdam, The Netherlands. No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the matcrial herein. Because of the rapid advances in the medical sciences. the Publisher recommends that independent verification of diagnoses and drug dosages should be made. Special regulafions for readers in the USA. This publication has been registered with the Copyright Clearance Center, Inc. (CCC), Salem, Massachusetts. Information can be obtained from the C C C about conditions under which the photocopying of parts of this publication may be made in the USA. All other copyright questions, including photocopying outside of the USA, should be referred to the Publisher.

ISBN 0-444-80996-1 (volume) ISBN 0-444-80303-3 (series) Published by:

Sole distributors for the USA and Canada:

Elsevier Science Publishers B.V. (Biomedical Division) P.O. Box 211 1000 A E Amsterdam The Netherlands

Elsevier Science Publishing Company, Inc. 52 Vanderbilt Avenue New York. NY 10017 USA

Library of Congress Cataloging in Publication Data Hormones and their actions / editors. B.A. Cooke, R.J.B. King, H.J. van der Molen p. cm. -- (New comprehensive biochemistry; v . 18A-) Includes bibliographies and index. ISBN 0-444-80996-1 (pt. 1) 1. Hormones--Physiological effect. I. Cooke. Brian A . 11. King, R.J.B. (Roger John Benjamin) 111. Molen, H . J . van der. 1V. Series: New comprehensive biochemistry; v. 18A, etc. [DNLM: 1. Hormones-physiology. W1 NE372 v. 18 / WK 102 H812781 QD415.NJ8 vol. 18A. etc. [ QP57 11 574.19’2 s-dc 19 [ 6 12’,4051 DN LMiDLC for Library of Congress 88-16501 CIP

Printed in The Netherlands

V

List of contributors M. Ascoli, 133 The Population Council, 1230 York Avenue, New York, N Y 10021, U.S.A. M.A. Blankenstein, 49 Department of Endocrinology, Academic Hospital Utrecht, Utrecht, The Netherlands L. Cancela, 269 Division of Biomedical Sciences, University of California, Riverside, C A 92521-0121, U.S.A. C.L. Clarke, 197 Garvan Institute of Medical Research, St. Vincent’s Hospital, Sydney, New South Wales 2010, Australia D.P. Edwards, 241 Department of Pathology, University of Colorado Health Sciences Center, 4200 East Ninth Avenue, Denver, C O 80262, U.S.A. U. Gehring, 217 Institut f u r Biologische Chemie der Universitat, Im Neuenheimer Feld 501, 6900 Heidelberg, F. R. G . D.N. Gower, 3 Division of Biochemistry, United Medical and Dental Schools (Guy’s Hospital), London SEI 9RT, England S.A. Haining, 169 Department of Biochemistry, University of Leeds, Leeds LS2 9JT, England B. Harper, 169 Department of Biochemistry, University of Leeds, Leeds LS2 9JT, England K.B. Horwitz, 241 Departments of Medicine & Pathology, University of Colorado Health Sciences Center, 4200 East Ninth Avenue, Denver, C O 80262, U.S.A. A.S. Khanna, 117 Cell Regulation Research Group, Department of Medical Biochemistry, The University of Calgary, Calgary, Alberta, Canada T2N 4N1 R.J.B. King, 29 Hormone Biochemistry Department, Imperial Cancer Research Fund, P. 0. Box 123, Lincoln’s Inn Fields, London, W C 2 A 3 P X , England N.L. Krett, 241 Department of Medicine, University of Colorado Health Sciences Center, 4200 East Ninth Avenue, Denver, C O 80262, U . S . A .

vi

W.I.P. Mainwaring, 169 Department of Biochemistry, University of Leeds, Leeds LS2 9JT, England E. Mulder, 49 Department of Biochemistry II, Erasmus University, Rotterdam, The Netherlands A.W. Norman, 269 Division of Biomedical Sciences, Univecsity of California, Riverside, C A 925214121, U.S.A. J. Nunez, 61 INSERM U 282, H6pitul Henri Mondor, 51, avenue du Martchal de Lattre de Tassigny, 94010 Crtteil, France M.G. Parker, 39 Molecular Endocrinology Laboratory, Imperial Cancer Research Fund, P. 0. Box 123, Lincoln’s Inn Fields, London WC2A 3 P X , England L.E. Reichert Jr., 105 Department of Biochemistry, Albany Medical College, Albany, N Y 12208, U.S.A. D.L. Segaloff, 133 The Population Council, 1230 York Avenue, New York, N Y 10021, U.S.A. R.L. Sutherland, 197 Garvan Institute of Medical Research, St Vincent’s Hospital, Sydney, New South Wales 2010, Australia G . Theofan, 269 Division of Biomedical Sciences, University of California, Riverside, C A 925214121, U.S.A. T.J. Visser, 81 Departments of Internal Medicine 111 and Clinical Endocrinology, Erasmus University Medical School, Rotterdam, The Netherlands D.M. Waisman, 117 Cell Regulation Research Group, Department of Medical Biochemistry, The University of Calgary, Calgary, Alberta, Canada T2N 4NI A.E. Wakeling, 151 Research Department I , Imperial Chemical Industries PLC, Pharmaceutical Division, Mereside, Alderley Park, Macclesfield, Cheshire S K I 0 4TG, England C.K.W. Watts, 197 Garvan Institute of Medical Research, St Vincent’s Hospital, Sydney, New South Wales 2010, Australia

vii

Contents List of contributors

v

. . . . . . . . . . . . . . . . . . . . . . . . .

Section I General aspects of hormones and hormone actions

Chapter 1. The biosynthesis of steroid hormones: an update. by D . B . Cower 1. Introduction . . . . . . . . . . . . . . . . . . . . . 2 . Role of lipoproteins in steroidogenesis . . . . . . . . . . . . 3 . Mitochondria1 cholesterol . . . . . . . . . . . . . . . . . 3.1, Transport of cholesterol into mitochondria . . . . . . . . . 3.2. Intramitochondrial transport of cholesterol . . . . . . . . 4 . Side-chain cleavage (SCC) of cholesterol . . . . . . . . . . . 5 . Biosynthesis of corticosteroids . . . . . . . . . . . . . . . 5.1. Enzymes involved in corticosteroid biosynthesis . . . . . . . 5.2. 1I@-and 18-hydroxylases . . . . . . . . . . . . . . . 5.3. Formation of aldosterone . . . . . . . . . . . . . . . 6 . Biosynthesis of the androgens . . . . . . . . . . . . . . . 6.1, Action and properties of 17-hydroxylase and C-17,20-lyase . . . 6.2. Conversion of S-ene-30-hydroxy- to 4-en-3-oxosteroids . . . . 6.3. Interconversion of 4-androstenedione and testosterone . . . . 6.4. Conversion of testosterone into Sa-dihydrotestosterone (Sa-DHT) 7 . Biosynthesis of oestrogens . . . . . . . . . . . . . . . . 8. Secretion of synthesized steroid hormones . . . . . . . . . . . 9 . Conclusion . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . .

3

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . .

. . . .

. . . .

. . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . . .

.

.

.

.

.

.

3 4 4 4 6 8 11 12 13 14 15 17 18 20

20 20 24 25 25 25

Chapter 2 . Overview of molecular aspects of steroid hormone actions. byR.J.B.King. . . . . . . . . . . . . . . . . . . . . . . . . . .

29

1. Introduction . . . . . . . . . . 2 . Intracellular events in steroid action . . 2.1. Intracellular location of receptors . 2.2. Receptor structure . . . . . . 2.3. D N A binding . . . . . . . . 3 . Specificity of steroid action . . . . . 3.1. Ligand availability . . . . . .

29 29 29 31 31 32 32

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

viii 3.2. Ligand specificity of receptor . . . . . . . . . 3.3. Agonismiantagonism . . . . . . . . . . . . 3.4. Availability of responsive genes . . . . . . . 3.5. Specificity of the steroid response element . . . . References . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . .

34 35 35 36 37

Chapter 3. Gene regulation by steroid hormones. by M.G. Parker . . . . .

39

1. Introduction . . . . . . . . . . . 2 . Structure and function of steroid receptors . 3 . Steroid receptor-DNA interactions . . . 3.1. Non-specific D N A binding . . . . 3.2. Specific D N A binding . . . . . . 4 . Steroid receptor-chromatin interactions . . 5 . Steroid hormone-activated gene networks . References . . . . . . . . . . . . .

. . . . . .

39 39 42 43 43 46 46 47

Chapter 4 . Characterization. assay and purification of steroid receptors. by M . A . Blankenstein and E . Mulder . . . . . . . . . . . . . . . . . .

49

1. Introduction . . . . . . . . . . . . . . . . . . . 2 . Properties of steroid receptors . . . . . . . . . . . . . 2.1. Binding properties . . . . . . . . . . . . . . . 2.2. Physico-chemical properties . . . . . . . . . . . . 3 . Assay of steroid receptors . . . . . . . . . . . . . . 3.1. General aspects and radioligand assays . . . . . . . 3.2. Separation of bound and free ligand . . . . . . . . 3.3. Immunological assays . . . . . . . . . . . . . . 3.4. Other steroid receptor assays . . . . . . . . . . . 4 . Purification of steroid receptors . . . . . . . . . . . . . 4.1. General protein purification . . . . . . . . . . . . 4.2. DNA-affinity chromatography . . . . . . . . . . . 4.3. Steroid affinity chromatography . . . . . . . . . 4.4. Immunoaffinity purification . . . . . . . . . . . . 5. Characterization of steroid receptors . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . .

49 50 50 52 53 53 54 54 55 55 55 56 56 57 57 58

. . . . . . . .

. . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . . . .

. . . . . . . . .

. . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . .

Chapter 5 . Mechanism of action of thyroid hormone. by J . Nunez . . . . .

61

1. Introduction . . . . . . . . . . . . . . . . . . . . . 2 . Thyroid hormone production. transport and uptake by the target cells 3 . Thyroid hormone nuclear receptors and cellular binding proteins . . 3.1. Nuclear receptors . . . . . . . . . . . . . . . . . 4 . Induction and repression of pituitary hormones . . . . . . . . 4.1. Growth hormone . . . . . . . . . . . . . . . . . 4.2. Thyrotropin . . . . . . . . . . . . . . . . . . . 5 . Regulation of lipogenesis in the liver . . . . . . . . . . . . . 5.1. Malic enzyme . . . . . . . . . . . . . . . . . . . 5.2. Fatty acid synthase . . . . . . . . . . . . . . . . .

61 63 64 65 66 66 68 68 68 70

. . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . .

ix 6 . Effects of thyroid hormone on the receptor-adenylate cyclase system in the adipocyte

and the hepatocyte . . . . . . . . . . . . . . . . . . . . . . . . . . 7 . The muscle cell: P-adrenergic responsiveness and the expression of myosin heavy chains . . 8 . Thyroid hormones and brain development . . . . . . . . . . . . . . . . . . 8.1, Neuronal differentiation . . . . . . . . . . . . . . . . . . . . . . 8.2. Glial cell differentiation . . . . . . . . . . . . . . . . . . . . . . 9. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

70 72 73 74 75 76 76

Chapter 6. Metabolism of thyroid hormone. by T.J. Visser . . . . . . . .

81

1. Metabolic pathways of thyroid hormone . . . . . 1.1. Introduction . . . . . . . . . . . . 1.2. Deiodination . . . . . . . . . . . . 1.3. Conjugation . . . . . . . . . . . . 2 . Type I iodothyronine deiodinase of liver and kidney 2.1. Properties and distribution . . . . . . . 2.2. Substrate specificity . . . . . . . . . . 2.3. Inhibitors and affinity labels . . . . . . . 2.4. Reaction mechanism . . . . . . . . . 2.5. Cofactor requirements . . . . . . . . . 3. Iodothyronine deiodinases of other tissues . . . . 3.1. Type I1 iodothyronine deiodinase . . . . . 3.2. Type 111 iodothyronine deiodinase . . . . 3.3. Possible other iodothyronine deiodinases . . 4 . Transport of iodothyronines into tissues . . . . . 5 . Regulation of thyroid hormone metabolism . . . References . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

81 81 82 84 85 85 86 87 89 90 93 93 95 96 97 99 100

Chapter 7. Characterization of membrane receptors: some general considerations. by L.E. Reichert. Jr . . . . . . . . . . . . . . . . . . .

105

. . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction . . . . . . . . . . . . . . . . . . . . 2 . Preparation of receptor probe . . . . . . . . . . . . . . 3 . Preparation of membrane receptors . . . . . . . . . . . . 3.1. General considerations . . . . . . . . . . . . . . 3.2. Membranes from cell cultures . . . . . . . . . . . . 3.3. Membranes from tissue homogenates . . . . . . . . 4 . Hormone binding characteristics of the membrane receptor . . . 4.1. Specificity . . . . . . . . . . . . . . . . . . . 4.2. Selection of appropriate in vitro system . . . . . . . . 4.2.1. Effects of time, temperature, buffer . . . . . . 4.2.2. Steady-state (equilibrium) conditions . . . . . . 5 . Molecular properties of the membrane receptor . . . . . . . 6 . Solubilization of the membrane receptor . . . . . . . . . 7.Summary . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . .

.

.

.

.

.

. . . . . . . . . . . . . . . . . . . . . . . .

105 106 107 . . 107 . . 108 . . . 109 . . . 111 . . 111 . . . 112 . . . 112 . . . 112 . . . 113 . . . 114 . . 115 . . 115

X

Chapter 8. Metabolism and intracellular processing of protein hormones. by A.S. Khanna and D . M . Waisman . . . . . . . . . . . . . . . . . .

117

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 . Biosynthesis of protein hormones . . . . . . . . . . . . . . . . . . . . . 2.1, Transcription and translation . . . . . . . . . . . . . . . . . . . . 2.2. Interaction of signal peptide with R E R membrane . . . . . . . . . . . . . 2.3. Cleavage of signal peptide . . . . . . . . . . . . . . . . . . . . . 3 . Processing of prohormones . . . . . . . . . . . . . . . . . . . . . . . 3.1, Structures of prohormones . . . . . . . . . . . . . . . . . . . . . 3.1.1. Pro-opiomelanocortin (POMC) peptide family . . . . . . . . . . . . 3.1.1 . 1. The POMC gene . . . . . . . . . . . . . . . . . . 3.1.1.2. Distribution and processing of POMC gene products . . . . . . 3.1.1.3. Additional modifications of POMC peptide family . . . . . . . 3.2. Significance of 'pro' sequence . . . . . . . . . . . . . . . . . . . . 3.3. Cleavage at dibasic amino acids . . . . . . . . . . . . . . . . . . . 3.4. Cleavage at monobasic amino acids . . . . . . . . . . . . . . . . . . Processingenzymes . . . . . . . . . . . . . . . . . . . . . . . . 3.5.1, Endopeptidases . . . . . . . . . . . . . . . . . . . . . . 3.5.2. Exopeptidases . . . . . . . . . . . . . . . . . . . . . . . 3.6. Post-translational modifications . . . . . . . . . . . . . . . . . . . 4 . Storage of protein hormones . . . . . . . . . . . . . . . . . . . . . . . 5 . Release of protein hormones . . . . . . . . . . . . . . . . . . . . . . 6 . Circulation in blood . . . . . . . . . . . . . . . . . . . . . . . . . . 7 . Degradation of protein hormones . . . . . . . . . . . . . . . . . . . . . 7.1. Degradation of glycoprotein hormones . . . . . . . . . . . . . . . . . 7.2. Internalization of protein hormones . . . . . . . . . . . . . . . . . . 8 . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

117 118 118 120 120 121 122 122 122 122 123 123 123 124 124 124 125 126 127 127 128 128 128 129 130 130

Chapter 9. Internalization of peptide hormones and hormone receptors. by D.L. Segaloff and M . Ascoli . . . . . . . . . . . . . . . . . . . .

133

1. Introduction . . . . . . . . . . . . . . . . . . . 2 . General features of receptor-mediated endocytosis . . . . . 3 . Methods used to assess receptor-mediated endocytosis . . . . 3.1. Morphological approaches . . . . . . . . . . . . 3.2. Biochemical approaches . . . . . . . . . . . . . 4 . Biological consequences of receptor-mediated endocytosis . . 4.1. Microaggregation . . . . . . . . . . . . . . . 4.2. Internalized and degraded hormone . . . . . . . . 4.3. Receptor down-regulation . . . . . . . . . . . . 5 . Conclusion . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . .

133 134 137 137 138 144 144 145 146 147 147

. . . . . . . . .

. . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . .

. . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

. . . . . . . . .

xi

Chapter 10. Physiological aspects of luteinizing hormone releasing factor and sex steroid actions: the interrelationship of agonist and antagonist activities. by A .E . Wakeling . . . . . . . . . . . . . . . . . . . . . .

151

1 . Introduction . . . . . . . . . . . . 2 . L H R H and L H R H analogues . . . . . . 2.1. Physiology . . . . . . . . . . . 2.2. Biological activity of L H R H analogues . 3 . Steroid antagonists . . . . . . . . . . 3.1. Physiology . . . . . . . . . . . 3.2. Antiandrogens . . . . . . . . . 3.3. Antioestrogens . . . . . . . . . References . . . . . . . . . . . . . .

151 152 152 154 156 156 160 161 104

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Section I1 Specific actions of steroid hormones

Chapter I1 . The functions of testosterone and its metabolites. by W.I.P. Mainwaring. S.A. Haining and B . Harper . . . . . . . . . . . 1 . Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 . The functions of androgens in various target organs . . . . . . . . . . 2.1. Testis . . . . . . . . . . . . . . . . . . . . . . . 2.2. Urogenital tract . . . . . . . . . . . . . . . . . . . . 2.3. Haemopoietic organs . . . . . . . . . . . . . . . . . . 2.4. Salivary glands . . . . . . . . . . . . . . . . . . . . 2.5. Kidney . . . . . . . . . . . . . . . . . . . . . . . 2.6. Muscle . . . . . . . . . . . . . . . . . . . . . . . 2.7. Liver . . . . . . . . . . . . . . . . . . . . . . . . 2.8. Central nervous system . . . . . . . . . . . . . . . . . 2.9. Anterior pituitary . . . . . . . . . . . . . . . . . . . 2.10. Breast . . . . . . . . . . . . . . . . . . . . . . . 2.11 Hair . . . . . . . . . . . . . . . . . . . . . . . . 2.12 Sebaceous glands . . . . . . . . . . . . . . . . . . . 2.13. Skin . . . . . . . . . . . . . . . . . . . . . . . . 2.14. Bone . . . . . . . . . . . . . . . . . . . . . . . . 2.15. Lymphocytic organs . . . . . . . . . . . . . . . . . . . 2.16. Accessory sexual glands . . . . . . . . . . . . . . . . . . 2.16.1. Prostate . . . . . . . . . . . . . . . . . . . . . 2.16.2. Seminal vesicle . . . . . . . . . . . . . . . . . . 2.16.3. Epididymis . . . . . . . . . . . . . . . . . . . . 2.17. Exotic systems . . . . . . . . . . . . . . . . . . . . . 3 . Concluding remarks . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

169

169 174 . . . . . 174 . . . . . 175 . . . . . 177 . . . . . 178 . . . . . 179 . . . . . 182 . . . . . 185 . . . . . 186 . . . . . 188 . . . . . 188 . . . . . 188 . . . . . 189 . . . . . 189 . . . . . 190 . . . . 190 . . . . 190 . . . . 190 . . . . 191 . . . . 191 . . . . 191 . . . . 192 . . . . 194 . . . . 194

. . . . .

xii

Chapter 12. Oestrogen actions. b y R .L . Sutherland. C . K .W . Watts and C.L. Clarke . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Introduction . . . . . . . . . 2 . Oestrogen receptors . . . . . . . 3 . Oestrogen receptor genes . . . . . 4 . Oestrogen control of gene expression . 5 . Oestrogen control of cell proliferation 6 . Antioestrogen actions . . . . . . 7 . Conclusions . . . . . . . . . Acknowledgements . . . . . . . . References . . . . . . . . . . .

197

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

197 200 203 205 . 207 . 210 . 212 . 213 . 213

Chapter 13. Glucocorticoid receptor actions. b y I/. Gehring Introduction . . . . . . . . . . . . . . . . . . . . Glucocorticoid induced lymphocytolysis . . . . . . . . . . Lymphoid cell variants with altered hormone responsiveness . . Glucocorticoid receptor defects . . . . . . . . . . . . . . Molecular weights of glucocorticoid receptor polypeptides . . . Partial proteolysis of glucocorticoid receptors . . . . . . . . Functional domains of glucocorticoid receptors . . . . . . . 7.1. The M domain . . . . . . . . . . . . . . . . . 7.2. The DNA binding domain . . . . . . . . . . . . . 7.3. The hormone binding domain . . . . . . . . . . . . 7.4. Hormone independent gene activation by truncated receptors 7.5. A chimaeric receptor . . . . . . . . . . . . . . . 8 . Glucocorticoid response elements . . . . . . . . . . . . . 9 . Higher order structures of glucocorticoid receptors . . . . . . References . . . . . . . . . . . . . . . . . . . . . .

1. 2. 3. 4. 5. 6. 7.

. . . . . . .

217

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . . . . . . . . . . . . . . . .

217

. 217 . 218 220

. 221 . 222 . 222 224 226 227 229 230 230 233 235

. . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. .

. . . . . . . . . . . . . . . . . . . . .

Chapter 14. Progesterone action and receptors. b y N . L . Krett. D .P . Edwards and K . B . Horwitz . . . . . . . . . . . . . . . . . . .

241

1. Introduction . . . . . . . . . . . . . . . . . . . 2 . Physiology and clinical uses . . . . . . . . . . . . . . 3 . Mechanisms of action . . . . . . . . . . . . . . . . 3.1. Recent technological developments . . . . . . . . 3.1.1. Receptor purification . . . . . . . . . . . 3.1.2. Affinity labeling of receptors . . . . . . . . 3.1.3. Anti-receptor antibodies . . . . . . . . . 3.1.4. Cloning of the PR cDNA . . . . . . . . . 3.2. Progesterone receptor structure . . . . . . . . . 3.2.1. The A- and B-receptor question . . . . . . . 3.2.2. Native PR structure: purification studies . . . . 3.2.3. Native PR structure: immune analyses . . . . 3.2.4. Native receptor structure: phosphorylation . . . 3.3. Intracellular localization . . . . . . . . . . . . . 3.4. Receptor function: regulation of gene expression . . .

241 241 243 243 243 244 244 245 245 245 249 251 254 255 257

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . . . . . . . . . .

.

. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . . . . . . . . . . .

xiii 3.4.1. Nuclear matrix . . . . . . . . . . . 3.4.2. Acceptor proteins . . . . . . . . . . 3.4.3. DNA hormone response elemcnts . . . 4 . Conclusions . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . 258 . . 258 . . . 259 261 . . . . 262 . . 262

Chapter 15. The pleiotropic vitamin D hormone. by L . Cancela. G. Theofan and A . W. Norman . . . . . . . . . . . . . . . . . . . . . . . . . 1. Introduction . . . . . . . . . . . . . . . . . . . . . 2 . Production and metabolism of vitamin D . . . . . . . . . . 3 . Modes of action of 1.25(OH), D, . . . . . . . . . . . . . . 3.1. Introduction . . . . . . . . . . . . . . . . . . . 3.2. Receptor-mediated genomic interactions . . . . . . . . 3.2.1. 1.25(OH),D3 receptor characteristics . . . . . . . 3.2.2. Evidence for the genomic actions of 1.25(OH),D, . . 3.3. Evidence for non-genomic actions of 1 25(OH)2D3 . . . . . 4 . Vitamin D and the maintenance of mineral homeostasis . . . . . 4.1. The kidney . . . . . . . . . . . . . . . . . . . . 4.2. The intestine . . . . . . . . . . . . . . . . . . . 4.3. Bone . . . . . . . . . . . . . . . . . . . . . . 4.4. T h e reproductive stages . . . . . . . . . . . . . . . 5 . Non-classical vitamin D responsive systems . . . . . . . . . 5.1. The pancreas . . . . . . . . . . . . . . . . . . . 5.2. Reproductive organs . . . . . . . . . . . . . . . . 5.3. Neural tissues . . . . . . . . . . . . . . . . . . . 5.4. Contractile tissues . . . . . . . . . . . . . . . . . 5.4.1. Skeletal muscle . . . . . . . . . . . . . . . 5.4.2. Cardiac muscle . . . . . . . . . . . . . . . 6 . Vitamin D and the immune system . . . . . . . . . . . . . 7 . Clinical disorders related to vitamin D . . . . . . . . . . . 8. Summary . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . .

.

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . .

. . . .

. .

. .

. . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . .

. . . .

. . . .

. . . .

. . . .

269

. . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . .

. . . . . .

269 269 271 271 271 271 272 274 . . 276 . 276 . 277 . 277 . 278 . . 280 . 280 . 280 . 281 . 281 . 281 . 282 . 282 . . 284 . 285 . 286

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

291

This Page Intentionally Left Blank

SECTION I

General aspects of hormones and hormone actions

This Page Intentionally Left Blank

B . A . Cooke. R.J.B. King and H.J. van der Molcn (eds.) Hormones and their Acrions. Purr 1 01988 Elsevier Science Publishers BV (Biomedicd Division)

3 CHAPTER 1

The biosynthesis of steroid hormones: an up-date D.B. GOWER Division of Biochemistry, United Medical and Dental Schools (Guy's Hospital), London SEI 9RT, England

1. Introduction During the past 50 years, numerous experiments have been performed in attempts to unravel the complex pathways whereby steroid hormones, that is the corticosteroids, androgens and oestrogens, are formed in mammalian and other tissues. A very large number of books and reviews have already been written on the subject and this present chapter will seek to (a) summarize the pathways concerned and how the evidence for these was obtained and (b) provide an update of advances over the past decade, particularly with regard to the properties of the steroid transforming enzymes involved and the mechanism of the reactions catalysed by such enzymes. The very early experiments which were designed to elucidate pathways of steroid hormone biosynthesis were done using large quantities of putative precursors. These were either administered to the whole animal, in which case changes in urinary output of steroids were studied, or incubated with tissue fractions, when metabolites of the added steroid were investigated. The quantities of steroids were, of course, grossly unphysiological and it was not until labelled compounds, such as acetate and cholesterol became available commercially that greater advances were made. When 'H-labelled material of high specific radioactivity became available still later, it was possible to utilize extremely small quantities of the steroid, or precursor, so that the finely balanced mechanisms of steroid hormone control were not unduly affected.

Correspondence to: Professor D.B. Gower. Division of Biochemistry, UMDS (Guy's Hospital), London SE1 9RT, England (D.B. Gower is Professor of Steroid Biochemistry, United Medical and Dental Schools, (Guy's Hospital), University of London.)

4

Having said this, no criticism of the early researchers in the field of steroid hormone metabolism is intended; they could only experiment with materials currently available to them.

2. Role of lipoproteins in steroidogenesis Although cholesterol is accepted as the major precursor of steroid hormones as a result of side-chain cleavage to pregnenolone (see below), research over the past decade or so has focused on the mechanisms by which steroidogenic tissues obtain cholesterol. It should be borne in mind that such tissues require cholesterol, not only for steroid synthesis but also for membrane synthesis, and hence require more of the precursor sterol than other tissues. Morris and Chaikoff [l]showed that the bulk of rat adrenal cholesterol was derived from circulating cholesterol, and later work revealed a similar state of affairs in humans. Through the work of many groups [2-61, there is now no doubt that steroidogenic tissues, such as adrenal, ovary, placenta and, possibly, testis of many species derive much of their cholesterol from plasma lipoproteins. These are macromolecules consisting of protein (apolipoprotein) and lipids and, depending on their hydrated densities, are classified as follows: chylomicra, very low density lipoproteins (VLDL), intermediate density lipoproteins (IDL), low density lipoproteins (LDL) and high density lipoproteins (HDL). The two last mentioned consist of a core of hydrophobic lipids, primarily cholesterol esters and triacylglycerols, surrounded by a monolayer of hydrophilic phospholipids, cholesterol and apolipoprotein. For example, plasma LDL contains apoprotein B (25%) and various lipids, of which nearly 50% is cholesterol ester. Phospholipids, cholesterol and triacylglycerols constitute the remainder of the lipids [7,8]. Cholesterol appears to be taken up from plasma lipoproteins by steroidogenic tissues by two receptor-mediated pathways - the LDL pathway and the HDL pathway. Not all tissues of all species can utilise both of these; thus, the LDL pathway appears to occur in all species, including man, whereas the HDL pathway occurs mainly in rodents. LDL lipoproteins interact specifically with cell surface-bound receptors, as shown for, e.g., adrenal [9] and ovary [lo], after which internalization occurs by endocytosis and hydrolysis of LDLs, plus their cholesterol ester complement by lysosomal action.

3. Mitochondria1 cholesterol 3.1. Transport of cholesterol into mitochondria The next events in steroidogenesis must obviously include the transport of cholesterol and cholesterol ester to the required organelles, in particular, cholesterol into

mitochondria for side-chain cleavage to occur in the first stage of steroid hormone biosynthesis. It seems likely that the cytoskeleton, including the array of microfilaments and microtubules, may play an important role in processing of lipoproteins and in intracellular cholesterol transport. Such evidence that exists has been obtained in studies using colchicine, which affects microtubules and interferes with steroid production in steroidogenic tissues, thus implying the necessity for these structures [ll].A second drug, known to alter the structure of microfilaments by causing their cross-linking and polymerization, is cytochalasin B. Treatment of adrenal and ovarian cells with this caused rapid and reversible inhibition of trophic hormone-induced steroidogenesis [ 121. Further evidence for the involvement of the cytoskeleton in steroid transport has been provided by Hall and co-workers [13], who showed that transport of cholesterol into mitochondria, and steroidogenesis, were both reduced in mouse Y-1 adrenal cells in culture in response to anti-actin antibodies. The release of cholesterol from cholesterol esters occurs extra-mitochondrially by means of a cholesterol ester hydrolase in adrenals, ovaries and testicular Leydig cells (see Ref. 6 for review). This enzyme has been studied mostly in adrenal preparations, and is known to be activated and de-activated by reversible phosphorylation [14] and that the phosphorylation was brought about by a c-AMP-dependent protein kinase [15]. Hence, ACTH stimulation of cholesterol ester activity in the adrenal occurs via the kinase and, in a similar way, trophic hormone stimulation of ovarian and testicular cholesterol ester hydrolases may occur and provide a large pool of cholesterol for steroidogenesis [ 1 ~ 8 1 . Privalle and colleagues [19] have suggested a sequence of steps culminating in the transport of cholesterol into the mitochondria of steroidogenic tissues. Stage 1 involves the binding of plasma LDL (HDL in rodents) to specific receptors, a process which is stimulated by ACTH in the adrenal due to an increase in the number of receptors [9]. Receptor-mediated endocytosis of LDLs then occurs resulting in deposition of cholesterol ester in lipid droplets. Stage 2 involves conversion of cholesterol ester to cholesterol, another process that is stimulated either by ACTH, via the c-AMP-dependent protein kinase, or by suppression of the acyl CoA-cholesterol acyltransferase (ACAT), which is needed for cholesterol ester synthesis (see Refs. 2,9 for reviews). Stage 3 involves transport of the liberated cholesterol into adrenal mitochondria; this is also stimulated by ACTH and probably depends on cell architecture since, as indicated above, anti-microtubule and anti-microfilament treatments block this process [20]. There is also evidence for the participation of sterol carrier protein(s) which seem to be present in many cells, including those of the adrenal cortex; such proteins may be involved in cholesterol transport from the cytosol [21,22]. Stage 4 involves the intra-mitochondria1 transport of cholesterol, which occurs in high concentration in the outer mitochondria1 membrane but at much lower concentrations in the inner membrane [23]. It is here that the side-chain cleavage (SCC) system resides (see below).

6

3.2. Intrarnitochondrial transport of cholesterol In 1979, Simpson [24] postulated that the outer mitochondrial membrane was the site of action of a labile protein factor, necessary to facilitate the transport of cholesterol, and Privalle et al. [19] provided evidence to support the notion that transference of cholesterol from the outer to the inner membrane required an agent that is cycloheximide dependent. When rats were ether-stressed in vivo and cholesterol SCC was deliberately inhibited, cholesterol accumulated in the adrenal mitochondria, most (90%) of this being associated with the inner membrane cytochrome P450,,,. After administration of aminoglutethimide to rats to block SCC, there was a two-fold increase in inner membrane cholesterol, while cycloheximide abolished this increase. Thus, it appears that cholesterol accumulates in the inner mitochondrial membrane as a result of stress and that transference from outer to inner membrane requires a protein factor. Pederson and co-workers (see Ref. 25) have isolated a peptide of M , 2200, from ACTH-stimulated rat adrenals, which contained 15% of basic aminoacid residues. The polar side-chain groups were thought to alter membrane structure so that transference of cholesterol towards the cyt P-450,,, on the inner membrane would be favoured. Phospholipids are also thought to be involved in cholesterol transference. Increases in both the degree of unsaturation of fatty acyl groups and length of fatty acyl chains of mitochondrial phospholipids are known to increase the rate of cholesterol transfer [25]. Further, the concentrations of some phospholipids in the inner mitochondrial membrane of rat adrenals were shown to increase after ACTH stimulation and to be related to cholesterol SCC activity [25]. Kimura [25] has recently discussed the possibility that several factors, including phospholipids and Ca2+ions, are involved in the hexagonal phase-mediated trans-

--

-

P-450

CHOL

CHOL Bilayer (La)

Reversed Hexagonal phase (HII)

inner membrane CHOL U n s a t u r a t ion PE PS. CL

Outer membrane

CHOL

Long C Chain Ca'f

*

-

HIi HIi

HI1

Fig. 1. Factors involved in the intramitochondrial transport of cholesterol. Left, membrane fusion stimulating reversed hexagonal phase formation; right, permeation of cholesterol across membranes (from Ref. 25, with permission).

7

Mitochondrion

,

ATP

4

m

Labile Protein Synthesis

ACTH-

L

I

P-45oscc Cholesterol

Pregnenolone

---+\

L i p i d Droplet

O t h e r F a t t y Acids

1

C o r t i coi ds

Fig. 2. General scheme for transduction of ACTH signal from plasma membrane to mitochondria in the adrenal cortex. CHX, cycloheximide: T G . triglyceride; CE, cholesterol ester: Lts, leukotrienes: PGs, prostaglandins; SCP2, sterol carrier protein; PE, PC, PI, phosphatidyl ethanolamine, choline and inositol (from Ref. 25, with permission).

port of cholesterol into the inner mitochondria1 membrane, hexagonal phases being more favourable structures that lipid bilayers for effective solute permeability, without the assistance of translocating proteins [26]. Consistent with this notion are the findings that, firstly, phospholipids with highly unsaturated fatty acyl chains prefer the hexagonal phase. In this situation, the non-polar groups will be oriented towards the outside of the membrane. As cholesterol approaches this non-polar domain, free from the membrane surface, the rotatory action of the hexagonal phase clusters may transfer cholesterol from outside to inside. Secondly, if the fusion of inner and outer membranes occurs in the hexagonal phase (Fig. l ) , this would result in transference of cholesterol in the outer membrane to the matrix side. ACTH stimulation may also result in the production of factors which may cause alteration of the bilipid to the hexagonal state. In response to ACTH stimulation, increased activity of cholesterol ester hydrolase will result in more free cholesterol (see above); this will be bound by sterol-binding protein and delivered to the mitochondria. Thus, the membrane-bound cholesterol content will increase and so also will the hexagonal phase and, hence, cholesterol penetration. Fig. 2 summarizes the possible sequence of events [25],some of which have been discussed here.

8

4. Side-chain cleavage (SCC) of cholesterol Once cholesterol is transferred to the inner mitochondria1 membrane of steroidogenic tissues such as adrenals, ovaries and testes, it encounters the enzyme system known as the cholesterol SCC system. This probably comprises 20- and 22-hydroxylases and a C-20,22-lyase, all tightly bound to the inner face of the membrane and associated with a specific cytochrome P-450,,,. In addition, molecular 0, is necessary together with NADPH reductase and non-haem iron sulphur protein, which are called adrenodoxin reductase and adrenodoxin, respectively, in the adrenal [24] (Fig. 3 ) . The mechanisms whereby cholesterol is converted to pregnenolone and a C, fragment, 4-methylpentanal, have been given detailed attention by researchers during the past three decades. Earlier work in this field has been reviewed by Sulimovici and Boyd [27] and more recent developments by Mitani [28] and Gower [29]. Reaction mechanisms have been proposed which involve, as intermediates, hybrids of ionic and free-radical species [3O], hydroperoxides [31] or epoxides [32-341. Other evidence is consistent with a ‘sequential hydroxylation’ pathway, by which cholesterol is converted first by 22-hydroxylation to 22R-hydroxycholesterol, then to 20R,22R-dihydroxycholesterolby 20-hydroxylation and, finally, to pregnenolone by means of the C-20,22-1yase reaction (Fig. 4). In common with other ‘mixed function’ oxidases, the cholesterol side-chain cleavage (SCC) system requires NADPH and O,, and Shikita and Hall [35] determined stoichiometric relationships between the oxidation of NADPH, 0, consumption and pregnenolone formation. For cholesterol, 20s-hydroxycholesterol and 20R,22R-dihydroxycholesterol, these ratios were 3:3:1, 2:2:1 and l:l:l, respectively. Using a purified cytochrome P-45OsCc,Orme-Johnson et al. [36] measured the dissociation constants of

SH-Cyt P-450(Fe3+) (high spin)

Flavo protein (oxidized)

NADP+

\ ‘Cyt

A

Non-hoem iron

-Y

Flavoprotein ANon-hat (reduced) protein Fe3‘

7 SH-cyt P-450(Fe2+)-O; + H20

P-450 (Fe3+) ( l o w spin) SOH

t SH-cyt P-450(Fe3+)-0;

Fig. 3. Cyclic reduction and oxidation of cytochrome P-450 in adrenal cortical mitochondria. FP. adrenodoxin reductase; non-haem iron protein, adrenodoxin: SH, substrate: SOH. hydroxylated product (from Ref. 24. with permission).

9

Cholesterol

20, 22-6 I hydroxycholesterol

22-hydroxycholestero1

&

HO

+

H3C>CH.CH2.CH2.C, H, H3C

‘ 0

Fig. 4. A n example of a ‘sequential hydroxylation’ sequence for pregnenolone hiosynthesis. 22 indicates hydroxylation position.

0

binding at pH 7.4 of the postulated intermediates and of the product, pregnenolone. The tightest binding (4.9 nmol/l) was shown for 22R-hydroxycholesterol consistent with this being the first intermediate. 20R,22R-Dihydroxycholesterol had a binding constant of 81 and pregnenolone of 2900 nmoVl, the latter low affinity being anticipated for the product of the reaction sequence. Two studies have provided evidence for this scheme (Fig. 4). Teicher et al. [37] used a purified bovine cytochrome P -450,,, to confirm that hydroxylated cholester01s are intermediates, whereas Hume and Boyd [38] utilized an adrenal P-450,,,, with labelled cholesterol as substrate. The oxidised P-450,,-cholesterol complex was reduced chemically under anaerobic conditions and then re-oxidised with the nonhaem iron sulphur compound, adrenodoxin [24] (see Fig. 3). Further reduction-oxidation cycles were accomplished as shown in Fig. 5 . The first cycle resulted in the formation of 22-hydroxycholesterol as the major product but further oxygenation resulted in 20,22-dihydroxycholesteroland, finally, in pregnenolone. Three complete oxygenationireduction cycles were therefore suggested as being necessary for cholesterol SCC to pregnenolone. The nature of the C, side-chain fragment produced depends, to some extent, on the tissue utilised and the conditions under which the experiments are carried out, but it is generally accepted that 4-methyl pentanal is formed first and that this may be oxidised to the corresponding acid or, alternatively, reduced to 4-methyl pentanol [29]. Despite suggestions that the side-chain of cholesterol may be cleaved completely in adrenal preparations [39,40], further work indicated that little or none of the C,,

10

No reduction/oxygenation

cycles

Fig. 5 . Pattern of product formation during single turnover cycles of anaerobic reduction/oxygenation of the cytochrorne P-450,,,- cholesterol - adrenodoxin complex. The results are expressed as a percentage conversion of the total [“C]cholesterol added to the incubation. Cholesterol, 0 ; 22-hydroxycholesterol, 0 ; 20,22-dihydroxycholesterol.A ;pregnenolone, (from Ref. 38, with permission).

steroid, dehydroepiandrosterone (DHA)plus the C, fragment, 2-methyl-6-heptanone, is formed [41]. However, in testis and ovarian preparations, some 30% of the total SCC fragments was 2-methyl-6-heptanone, indicating that an alternative pathway from cholesterol to D H A may occur in these tissues (see Ref. 29 for a critical assessment). The ‘sequential hydroxylation’ proposals explained above have been criticized by Lieberman and his colleagues [42,43] in a ‘heuristic’ proposal for steroidogenic processes. In a series of papers, Lieberman’s group has presented evidence for the true intermediates in the cholesterol SCC reaction being transient, enzyme-bound complexes. Further, there may be at least two different enzyme systems that catalyse the SCC of cholesterol, cholesterol sulphate and cholesterol acetate in bovine adrenal mitochondria1 preparations [44]. Greenfield et al. [45] have purified P-450,,, (which was free of adrenodoxin and adrenodoxin reductases) and have shown monophasic binding with cholesterol, cholesterol sulphate and cholesterol acetate, with dissociation constants of 1.1, 2.6 and 1.4 pmol/l, respectively. Finally, in this section, mention must be made of the specificity of the SCC reaction. C, to C, sterols, all with saturated side-chains, underwent cleavage in rat and bovine adrenals and porcine testis, at much the same rate as cholesterol itself; however, sterols with polar side-chains, e.g. 24-, 25- or 26-hydroxycholesterol, were cleaved at higher rates [46]. The 5-cholesten-3P-ol structure seems to be a necessary requirement for the substrate for SCC [47], but the more polar sterols may enter mitochondria more readily than cholesterol itself and bind to P-450,,, which, in

11

the bovine adrenal, at least, is synthesized as a larger precursor and cleaved proteolytically either before, or on, insertion into the mitochondria [48].

5. Biosynthesis of corticosteroids As indicated in Fig. 6, pregnenolone and progesterone are precursors of the corticosteroids. In some species, such as the rat, progesterone appears to be sequentially hydroxylated at C-17, 21 and l l p , whereas in the human and rabbit adrenal cortex, pregnenolone gives rise largely (but not exclusively) to the 17-oxygenated corticosteroids, e.g., cortisol, while progesterone gives rise largely (but again not exclusively) to the 17-deoxy-corticoids such as corticosterone (for reviews, see Refs. 29,49). Since the 17- and 21-hydroxylases are microsomal enzymes, pregnenolone must pass from its site of synthesis in the mitochondria to the endoplasmic reticulum (ER), but the mitochondrial membranes do not present a significant barrier to pregnenolone efflux [50].The conversion of pregnenolone to progesterone and of 17-hydroxypregnenolone to 17-hydroxyprogesterone also occur in the smooth ER by means of the 5-ene-3~-hydroxysteroid/3-oxosteroid-4,5-isomerase (5-ene-3pHSD/4,5-isomerase) system. Thereafter, 21-hydroxylation in the smooth E R of 17hydroxyprogesterone leads to 11-deoxycortisol, while that of progesterone leads to DOC. Conversion to cortisol or corticosterone, however, can only take place in the adrenal mitochondria because the required 1I@-hydroxylaseresides there and thus necessitates the transference of the precursors back through the mitochondrial membranes. Although these various hydroxylases, including the 18-hydroxylase for corticosterone (Fig. 6), occur in all three histologically-defined zones of the cortex, the enzyme needed for aldosterone synthesis, a presumed 18-HSD, occurs in the

Cholesterol

DOC

--

4

1 0 corticosterone

I@ 1

@

--L

Pregnenolone

1

a

Progesterone

1 0 1 0

17-hydroxyprogesterone

17-hydroxypregnenolone

i 1

DHA

To androgens

11-deoxycortisol

Aldosterone Cortisol

0

Fig. 6 . Biosynthetic pathways for corticosteroids. 11 indicates position of hydroxylation; HSD, hydroxysteroid dehydrogenase.

12 zona glomerulosa (ZG). Further, both 18-hydroxylase and 18-HSD have been found associated with the inner mitochondria1 membrane [51].

5.1. Enzymes involved in corticosteroid biosynthesis A great deal of information is now available about the properties, constitution and clinical manifestations of deficiencies of the hydroxylases involved in corticosteroid synthesis [42,52,53], and only a fraction of that information can be mentioned here. All the hydroxylases are ‘mixed-function’ oxidases, requiring NADPH and 02,and some seem to be associated with cyts P-450, viz. cyt P-450,,,, cyt P-45021,cyt P450,,. The following equation represents the reaction catalysed: R-H

+ NADPH + H + + 0 2 +R-OH + NADP’ + H 2 0

Lieberman et al. [42] have criticized the kind of classical pathway diagram in Fig. 6 as giving a simplistic view of events. In this regard, evidence obtained during the past 15 years has indicated that the situation is indeed more complicated than was thought earlier [42,52]. Some of this evidence has been reviewed and suggests, for example, that there are at least two 21-hydroxylases. Kominani’s group [54-561 have purified a cyt P-450 from bovine adrenal microsomes which catalysed 21-hydroxylation of 17-hydroxyprogesterone, 21-deoxycortisol and 11Phydroxyprogesterone, the products being, respectively, 11-deoxycortisol, cortisol and corticosterone. On further purification the cyt P-450 was shown to be immunologically distinct from cyt P-450,,,, cyt P-45011, and cyt P-450,,. When a cyt P-450 with 21-hydroxylating activity for progesterone and 17-hydroxyprogesterone was mixed with cyt P-450 reductase, the 21-hydroxylating activity for pregnenolone and 17-hydroxypregnenolone was lost, although 17-hydroxylase and C-17,20-lyase activities could be reconstituted. Other data showed that progressive purification of bovine adrenal 21hydroxylase caused it to lose 21-hydroxylating ability for pregnenolone and progesterone, while retaining that for 17-hydroxyprogesterone [57]. This evidence, together with other biochemical data of Kahnt and Neher [%], is consistent with the notion that there are at least two 21-hydroxylases. In rat adrenals, experiments utilizing 21-hydroxylase inhibitors (see Ref. 52 for details) indicated that, even though cortisol was formed normally, corticosterone synthesis was inhibited. The suggestion made was that there might be one 21-hydroxylase for 17-hydroxy- and another for 17-deoxy-corticosteroids. A similar conclusion has been drawn by New and Levine [53] and may help to explain the known clinical features of the 21-hydroxylase defect of congenital adrenal hyperplasia, i.e., the existence of the simple virilizing form and the salt-losing type. It has been suggested that the 21-hydroxylase activity is impaired in the ZF for both 17-hydroxy- and 17-deoxycorticosteroid pathways, so that ll-deoxycortisol levels (and also cortisol levels) are decreased (Fig. 7), and the build-up of excess

13

Simple virilizing and salt-wasting C A H

-

Zona fasciculato ACTH

Cholesterol

1

Pregnen olone

I4,

Progesterone 21 OH

180H DOC+ll+soxycorticosterone

1

1 8 0 H B cCorticosterone

Zona glornerulosa Renin-Angiotensin

I I-

Cholesterol

1

Pregnenolone

4

17 OH - pregnenolone

1

+

17 0 H- progesterone f l M

210H

11-desoxycortisol

1

Cortisol

jimple virilizing CAH

Salt-wasting CAH

Cholesterol

Cholesterol

I Pregnenolone

Pregnenolone

1

Progesterone

1

21OH

11-desoxycorticosterone

I Corticosterone I 180HB I

t

Aldosterone

1 1

Progesterone wdzw

4

210H

11-desoxycorticosterone

1

Corticosterone

1

180H B

4

Aldosterone

Fig. 7. Adrenal steroidogenesis in the simple virilizing and salt-wasting forms of congenital adrenal hyperplasia (from Ref. 53, with permission).

17-hydroxyprogesterone results in excessive androgen production. In the ‘salt-losing’ type, however, the 21-hydroxylase defect is in the ZG so that the conversion of progesterone into DOC is blocked, resulting in aldosterone deficiency (hence the name of this defect, Fig. 7). Further, Kuhule et al. [59] have suggested that the 21hydroxylases of Z F and ZG are different, the former acting on both 17-deoxy and 17-hydroxy substrates, the latter perhaps being specific for 17-deoxy substrates, such as progesterone,

5.2.

l l p and 18-hydroxylases

There are also distinct possibilities that at least two forms of the adrenal mitochondrial llphydroxylase may exist, one in the Z G , involved in the conversion of DOC into aldosterone (Fig. 6) and another in the ZF/ZR, concerned with the conversion of D O C to cortisol and 4-androstenedione to llphydroxy-4-androstenedione. An alternative possibility is that several cyts P-450,,, may exist, which catalyse the llp hydroxylation of DOC, 11-deoxycortisol and 4-androstenedione [60,61]. To make

14 M-Enz

I

HO HO

L

0 2 M-Enz

M-Enz (step 1 1

( s t e p 21

0

1

Corticosterone

EH i 0 CH20H

18-hyd roxycorti costerone

%

Aldosterone

Fig. 8. Suggested mechanism for aldosterone biosynthesis. M-Enz represents a postulated metallo-enzyme (from Ref. 66, with permission).

the situation more complicated, it is not at all clear if there is a single cyt P-450 which possesses both 11P- and 18-hydroxylating abilities. Bjorkem and Kalmer [62] reconstituted enzyme systems from rat and bovine adrenals and studied their 18hydroxylating ability with D O C as substrate. It was found that the cyt P-450 involved was indistinguishable from that required for 11P-hydroxylation. In contrast, however, other results [63] are consistent with specific cyts P-450 from bovine mitochondria - one P-450,,,, the other P-450,,. Cheng et al. [64] found that 18-hydroxylation of corticosterone was inhibited by canrenone and other drugs to a greater extent than that of DOC, and suggested that two cyt P-4501, species might be involved. Various other pieces of evidence for the possible existence of isozymes of cyt P-450,, or for a single P-450 having multiple functions have been reviewed ~421.

5.3. Formation of aldosterone Figure 8 indicates one pathway for aldosterone biosynthesis, via corticosterone and its 18-hydroxylated derivative, which has been generally accepted as the major pathway. However, the obligatory nature of 18-hydroxycorticosterone as an intermediate has been questioned. One would expect good yields of aldosterone from this if it were the immediate precursor but this is not always the case. Further, such a pathway reIOlAquires oxidation of the -CH,OH group at C-18 by a presumed 18HSD, but there is no real evidence for this enzyme [65]. As an answer to this problem, a mechanism has been suggested [66,67] in which two successive hydroxylations at C-18 of corticosterone occur followed by spontaneous dehydration (Fig. 8). On the basis of other data, several other pathways for aldosterone biosynthesis

15

18- hydroxyprogesterone

0:

11/?- hydroxyprogesterone

0 1

t

18 - hy d roxy- DOC

@I

Progesterone

I

.-. .

DOC

Corticosterone

-.

-.

18- HSD

0 Aldosterone

Fig. 9. Biosynthetic pathways for aldosteronc. as in Fig. 6.

18 -hydroxycorticosterone

indicates postulated pathway; other abbreviations

--4

have been proposed (Fig. 9) which include 18-hydroxyprogesterone, 11phydroxyprogesterone and 18-hydroxyDOC as intermediates [29].

6. Biosynthesis of the androgens As a result of a wealth of experiments on androgen-producing tissues in numerous species [68,69], two pathways for testosterone synthesis are recognized (Fig. 10). After the formation of pregnenolone from cholesterol in the mitochondria, testosterone synthesis occurs in the endoplasmic reticulum from pregnenolone and progesterone. According to the classical view, 17-hydroxylation of these (& precursors occurs, after which SCC by a C-17,20-lyase results in dehydroepiandrosterone (DHA) from 17-hydroxypregnenolone or 4-adrostenedione from 17-hydroxyprogesterone. The action of 17P-HSD on D H A provides 5-androstene-3pJ7pdio1, which is known to be a good precursor of testosterone, through the action of 5-ene-

16

HO Cnolesterol

d-

HO

Pregnenolone

17- hydroxypregnenolone

DHA

5-androstenediol

I

Progesterone

17- h y c r s x y p r o g e ~ t e ~ o n e

4-crdroslened o l e

Testoste-one

Fig. 10. Pathways of androgen biosynthesis in rat testis. A + B + C and a + b + c are the A5 and A‘ pathways, respectively. for testosterone biosynthesis. Enzymes A.a. 17-hydroxylase; B.b. C-17,ZO-lyase; C.c, 17PHSD. Reaction c is reversible.

3p-HSD/4,5-isomerase activity [70]; the action of 17P-HSD on 4-androstenedione also provides testosterone. Inspection of Fig. 10 shows the two pathways, one involving 5-ene-3P-hydroxysteroids (sometimes called A’), the other involving 4-en3-oxosteroids (sometimes called the A4 pathway). It will be noted further that there are transitions from As to A4 at different ‘levels’ via the 5-ene-3P-HSDiisomerase enzyme system, but the reverse reactions seem to occur to only a very limited extent [52]. There is considerable species variation with respect to the predominance of one pathway or the other. For example, in the rat and mouse testis the 4-en-3-0x0 pathway seems to predominate, whereas in human testis the 5-ene-3P-pathway is quantitatively more important. One reason for the latter situation is probably the fact that 16a-hydroxyprogesterone (also produced in human testis) inhibits the SCC of 17-hydroxyprogesterone more effectively than that of 17-hydroxypregnenolone [71]. In further experiments (see Ref. 52), each substrate inhibited competitively the lyase for the other, with inhibition constant ( K , )of 19 pmolil for 17-hydroxyprogesterone and 60 pmolil for 17-hydroxypregnenolone. Testosterone inhibited the SCC of 17hydroxyprogesterone competitively but uncompetitively for that of 17-hydroxypregnenolone. These data could be taken to indicate that there are two C-17,20lyases for the two 17-hydroxysteroids or. alternatively, that a single lyase possesses

17 different active sites, but with very similar properties, for the two substrates. Whether 17-hydroxylation of the CZlsteroid precursors is a pre-requisite for C17,20-lyase action has been a point of discussion for many years. Some results suggested that 17-hydroxyprogesterone might not be an obligatory intermediate in the conversion of progesterone into testosterone and several studies in the author's laboratory have indicated that 16-unsaturated C,, steroids can be formed in boar testis from pregnenolone or progesterone without their prior 17-hydroxylation [68]. These results have recently been confirmed [72-741, although in immature porcine testis, 17-hydroxypregnenolone appears to be necessary as an intermediate in 16-androstene formation [74]. Lieberman et al. [42] favour the view that the intermediates in testosterone biosynthesis may be enzyme-bound and not, therefore, readily isolable, and these workers have reviewed the evidence that is consistent with such a notion [42]. Chasalow [75] followed up earlier work and has confirmed that 4-androstenedione is formed preferentially in rat testis from progesterone rather than from 17-hydroxyprogesterone. Other data [76] were consistent with the latter not being an intermediate. Earlier work [77], using incubations of boar testis with 17-hydroxy[I4C] progesterone plus ['Hlprogesterone, showed that the 17-hydroxy-derivative was the preferred substrate for labelled testosterone and 4-androstenedione. Progesterone, itself, gave rise to some testosterone but very largely to 'H-labelled 16-androstenes.

6.1. Action and properties of 17-hydroxylase and C-l7,20-lyase Both these enzymes are associated largely with the smooth endoplasmic reticulum, the distribution approximating that of cytochrome P-450. In addition to the P-450 which probably catalyses both enzymic functions (see below), NADPH and O2 are needed; a review has been published [52]. The mechanism of 17-hydroxylation was studied by synthesizing pregnenolone and progesterone with a tritium atom specifically at C-17, and incubating with a bovine adrenal system. The 17-hydroxypregnenolone or 17-hydroxyprogesterone formed did not contain significant quantities of 'H and indicated a direct and stereospecific substitution of the proton at C-17 by the hydroxyl group [78]: 17-['H]progesterone

+ H' + NADPH + 17-OH-progesterone + ['HIOH + NADP-

The 17-hydroxylase is found in the testis, adrenal and ovaries of many species but not in adrenal of mature rats or mice. This latter finding is consistent with the fact that the 17-deoxycorticosteroid, corticosterone, predominates, little or no cortisol being formed. At one time it was thought that the hydroxylase and lyase were separate entities and, in keeping with this, rat testis microsomes were shown to contain cyt P-450 species that were distinct for lyase and 17-hydroxylase activities [79,80]. In contrast

18

to these results, there is now evidence that both enzymic activities are linked, and that a single cyt P-450 is involved. Such a P-450 has been isolated from neonatal porcine testis [81,82]; the purified enzyme had a M , of 59000 1000 and was shown to be homogeneous, as judged by SDS-PAGE and immunochemical techniques. The purified enzyme was shown to be a glycoprotein containing haem and phospholipid, the latter being necessary for activity [82]. Both the 17-hydroxylation of progesterone and the SCC of 17-hydroxyprogesterone were catalysed when the cyt P-450 was reconstituted with an appropriate cyt P-450 reductase. However, the K,,, values differed with respect to the two substrates, progesterone (1.5 pmol/l) and 17-hydroxyprogesterone (2.4 pmol/l). Further work [82] indicated that detergent treatment of the cyt P-450 increased the 17-hydroxylase activity in relation to the C-17,20-lyase activity. If partial denaturation of the enzyme occurs through detergent treatment, then it is conceivable that the 17-hydroxylated C,, steroid intermediate may dissociate from the enzyme surface, whereas in the normal situation it would remain bound and not be readily isolable. Similar results to those of Nakajim et al. [82] have been obtained using guineapig adrenals [83], from which a cyt P-450 can catalyse both the 17-hydroxylation of progesterone and the SCC of 17-hydroxyprogesterone.

*

6.2. Conversion of 5-ene-3P-hydroxy- to 4-en-3-oxosteroids The transitions at different levels of the ‘A” to ‘A4’ pathways have been alluded to above and are illustrated in Fig. 10. Similar reactions must also occur in the formation of corticosteroids (Fig. 6). Examples of such reactions are pregnenolone -+ progesterone (and their 17-hydroxylated derivatives), DHA + 4-androstenedione and 5-androstenediol + testosterone, and these transformations are catalysed by two enzymes which probably form part of a complex associated largely with the smooth endoplasmic reticulum, namely: a 5-ene-3P-hydroxysteroid dehydrogenase/3-oxosteroid-4,5-isomerase (3pHSDlisomerase). This enzyme system occurs in the testis (leydig cells, with lesser activity in the tubules), ovary (corpus luteum), adrenal and placenta but there is species variation with regard to the adrenal, e.g., human adrenal has rather low activity [29]. The 3 p H S D requires NAD+ as cofactor (50% of activity is achieved with NADP’) and catalyses the oxidation of the 5ene-3phydroxysteroids to the corresponding 4-en-3-oxosteroids. There is marked substrate specificity [52], with DHA being oxidized with the greatest ease and cholesterol hardly at all: DHA > pregnenolone

=

5-androstenediol > 17-hydroxypregnenolone >>>> cholesterol

The intermediate 5-en-3-oxosteroids, such as 5-androstenedione (from DHA) are further converted by isomerization to the 4-en-3-oxosteroids by means of the 4 3 isomerase. This enzyme, which requires no cofactor, is associated with the smooth

19

E R of adrenal, testis, ovary, liver and placenta. It is relatively unstable, being inactivated by freezing, even when pure. A phospholipid environment appears to be an important requirement since, when bovine adrenal microsomal preparations were treated with phospholipase A, 80-85% of phospholipids were hydrolysed with a concomitant loss of 80-90% of enzymic activity [84]. Restoration of activity was achieved by adding back to the lipid-depleted membranes aqueous dispersions of microsomal total lipid mixtures [84]. A great deal of research has been undertaken to determine if the dehydrogenatiodisomerization reactions are properties of one system or of two separate enzymes (see Ref. 52); most of the evidence suggests that the former is true. Probably three, or even four, substrate-specific isomerases may occur in bovine adrenal cortex which can act on C,,, C,, and C2, steroids. Likewise, separate 5-ene-3pHSDs may exist in the adrenal cortex for C,, and C,, steroids, because the latter did not compete with C,9 steroids for active sites of the enzymes studied. As for the mechanism of the isomerase reaction, Talalay (see Ref. 52) used Pseudomonas testosteronii as enzyme source and suggested that intramolecular transfer of the 4pproton occurs to the Gpposition of the steroid molecule. The imidazole residues of histidine were suggested as playing an important role in the reaction, acting as alternate acceptors of the 4 P H and subsequent donors of the 6 P H . Smith and Brooks [85] confirmed this intramolecular 4P to 6P-H transfer (Fig. 11). Weintraub et al. [86] showed that a polar group at C-3 of the steroid substrates was necessary, binding possibly occurring through hydrogen-bonding with amino acid residues of the binding site. In contrast, ring D was thought to be involved with hydrophobic binding to the enzyme. A very precise fit of the steroid corresponding to C-11 was indicated because C-11 substituted steroids were not accepted as substrates whereas the region where the binding of ring A occurred was relatively open [521.

4 -en-3-oxosterold

5 - e n - 3 -oxosteroid

H

2-p R

R

H"+N

R

Fig. 11. Proposed mechanism for the action of 3-oxosteroid 4,5-isomrrase (from Talalay 1964, see Ref. 52).

20

6.3. Interconversion of 4-androstenedione and testosterone This interconversion is catalaysed by 17phydroxysteroid dehydrogenase (17pHSD), an enzyme generally found in the E R of numerous tissues such as adrenal, liver, testis, ovary and kidney. Like many of the enzymes described above, there appear to be different forms [52,87]. For example, rat adrenal cytosol and E R contain separate 17PHSDs, with NADH as the preferred cofactor. The rat testicular enzyme, however, prefers NADPH. Guinea-pig liver also contains two 17PHSDs, one solubilized from cytosol, the other associated with the ER [88]. These enzymes exhibit different activities towards C19steroids, the cytosolic one preferring 5Preduced 17oxosteroids and the microsomal counterpart being involved with 5a-reduced steroids, such as 5a-DHT. In this case, the product of the reaction would be 5a-androstane-3,Il-l-dione. The porcine testicular 17pHSD has been studied [89,90], and shown to be equally distributed between rough and smooth ER. The apparent K , for testosterone was 122 pmol/l (and 40 pmolil for the purified enzyme). 6.4. Conversion of testosterone into 5a-dihydrotestosterone (Sa-DHT) This conversion is catalysed by the 4-ene-5a-reductase, which has been studied in numerous tissues, including liver, testis, skin and pituitary. In androgen-target tissues, such as prostate and seminal vesicles, the reductase is associated very largely with the nucleus, but microsomal counterparts also exist, usually in androgen-sensitive tissues [52,87,91]. The enzyme has been solubilized and partially characterized from the human and rat prostate [92,93], rat epididymis [94], rat liver [95] and porcine testis [96]. The porcine testicular enzyme prefers NADPH as cofactor, only 40% activity being exhibited in the presence of NADH; the apparent K,,, was 0.6 pnolil [96]. The single most effective solubilizing agent was sodium citrate [96], as shown also for the rat epididymal enzyme [94]. Further metabolism of C,, steroids involves conversion to androstanediols, reactions which are catalysed by 3 a@)HSDs. These enzymes occur in numerous tissues and exhibit considerable heterogeneity. As reviews are available [52,91], this topic will not be discussed further here.

7. Biosynthesis of oestrogens Although it has long been known that C,, steroids, such as 4-androstenedione, give rise to oestrogens, the mechanism of this conversion has been the focus of intense study [52,97]. In pre-menopausal women the major source of oestrogen are the ovaries but, in many species, the testes make a significant contribution. The adrenals seem only to produce small quantities. However, it has been known for some years that, in post-menopausal women, most of the oestrogen formed is derived

21 mainly from plasma 4-androstenedione as a result of extra-glandular activity. Adipose tissue and muscle are important in this respect as well as liver, kidney and hypothalamus. The first step in the conversion of 4-androstenedione to oestrone is the hydroxylation at C-19, a reaction associated with the ER and which requires NADPH and 02.It was thought earlier that the 19-hydroxy derivative was then converted to the 19-aldehyde, which gave rise to oestrone or oestradiol-l7p (from 4-androstenedione or testosterone, respectively) as rupture of the bond between C-10 and the angular methyl group at C-19 occurred through C-10,19-lyase action (Fig. 12). More recent studies [42,52] have resulted in the proposal of at least three mechanisms, which all involve a second stereospecific hydroxylation at C-19 (requiring a second y

y3

3

co

Cholesteroi

FO

HO

Pregnenolone

DHA

17a -hyd roxypregnenolone

Progesterone

170-hydroxyprogesterone

I

4 -androstenedione

Testosterone

19-hydroxy-4-andr oitenedione

19-hydroxytestosterone

Oestrone

Oestrodiol,- 170

Fig. 12. 'Classical' pathways for oestrogcn biosynthesis, indicating 19-hydroxylationand 19-oxidation of 4-androstenedione and testosterone. followed by C-10.19 bond cleavage catalysed by a presumed C-10.19lyase.

22 mole each of NADPH and 0,)to produce a gem diol. Mechanism (i) invokes the formation of an epoxide intermediate (19-dihydroxy-4p,5-o~ido-androstane-3,17dione) which can be aromatized subsequently [98]. Mechanisms (ii) was suggested on the basis of a 2p-hydroxylation and is consistent with 3 mol each of NADPH and 0, being required in the aromatization process [99,100]. Once the 2P-hydroxy-190x0-derivative of 4-androstenedione is formed (Fig. 13), it decomposes with loss of hydrogen at C - l p [101,102]. However, it should be noted that Caspi et al. [lo31 were unable to show that the oxygen of the 2p-hydroxyl group was transferred to formate, as required if the derivative was an obligatory intermediate. Finally, Akhtar et al. [ 1041 proposed two possible mechanisms, each worthy of further study, which assume the dihydroxylation at C-19 of 4-androstenedione as step (i), followed by oxidation to the 19-aldehyde. Thereafter, the first mechanistic proposal invokes 2p(or lp-) hydroxylation followed by reaction with the 19-carbonyl group to yield hemiacetals (Fig. 14), which then result in oestrogen formation with loss of formate and rearrangement. One oxygen atom from O2 molecule number 3 was shown to be incorporated into the formate released. Akhtar et al. [lo41 suggest that formation of a four-membered ring in XIa is unlikely, so leaving path l b of Fig. 14 as one possibility. The alternative mechanisms invoke the formation of an intermediate en-

4 - arid r o s t en e d I one

19- hydroxy-4 -androstenedlone

19,19'-d I hydroxy-4 -androstened lone (hydrated f o r m of thel9-aldehyde derlvatlve)

Fig. 13. Proposed mechanism for oestrogen biosynthesis involving double hydroxylation at C-19 and hydroxylation at C-2 (see Ref. 52).

23 OH

I

OTCH

&=! oJy{ (i)

(II1

0

(XI)

Forrnate production Rearrangement

-

(Xlaj

(V)

HO

/

Mechanism l a

(11Formate producbon Rearrangement

( I V)

(11)

\

Mechanism 1 b

HC=O

CYH O-CH

H

I

HH=j

-

5

0

(Xllo)

(XI[)

bq

0

(XIII)

Mechanism 2 NADPH

+ O,+

Enz

dZH H

HC=O

Enr-O?-O

(IV)

Enz-0-0

Patha

&o

0

Path b

Path a

Fig. 14. Alternative mechanisms proposcd for the C-10-C-19 bond cleavage step in oestrogen biosynthesis, involving hydroxylated intermediates (upper) and an enzyme-bound peroxide intermediate (lower). In the latter the arrows in structure XIV denote path b (from Ref. 104, with permission).

zyme-bound ‘peroxide’ species (Fig. 14). Aromatization is then envisaged, either through a Baeyer-Villiger type process (pathway a) or directly through a cyclic mechanism (pathway b). The former requires the intermediacy of the 10P-formyl derivative but this cannot be aromatized in human placental microsomes and thus effectively excludes this pathway as a viable mechanistic alternative. Further work by Stevenson et al. [ 1051 has shown that 16~-hydroxytestosterone can be aromatized to oestriol via the 19-dihydroxy and 19-0x0 derivatives, these changes being identical to those indicated above [ 1041, in which an enzyme-peroxide intermediate was postulated (Fig. 14). Since the ‘aromatase’ system is known to be catalysed by cytochrome P-450 [106], it is feasible that involvement of a P450-peroxide species could be envisaged, not only in the C-10,19 cleavage but also in the preceding hydroxylations (Fig. 15).

24

Fig. 1.5. Suggested dual role of a cyt P-450 - peroxide species (10) in the hydroxylation and C-10.19 bond cleavagc steps in oestrogen biosynthesis (from Ref. 105, with permission).

There is evidence [S2] that there are at least two forms of cyt P-450 involved with aromatization. Likewise, there is evidence for different aromatases in human placenta which catalyse the production of oestrone and oestriol from 4-androstenedione and 16a-hydroxytestosterone, respectively. Each enzyme system has been subfractionated into its own cyt P-450and cyt P-450 reductase [107]. This has been supported recently by Purohit and Oakey [ 1081, who measured aromatase activity for 16a-hydroxy-4-androstenedione and 4-androstenedione in the presence or absence of the other substrate. 4-Androstenedione competitively inhibited aromatization of the 16a-hydroxy derivative, with apparent K , essentially the same as its apparent K,, suggesting that both substrates bind and are aromatized independently of each other. The Iba-hydroxy derivative competitively inhibited the aromatization of 4-androstenedione, thus presumably lowering the affinity of the aromatase for the latter.

8. Secretion of synthesized steroid hormones Once the various steroids have been formed in paticular subcellular compartments, they must be released into the peripheral blood circulation. There is evidence that some steroids are released by passive diffusion, as in the case of corticosterone, but for 18-hydroxylated corticosteroids, N a f / K t -ATPase activity is necessary [6,109]. The situation is more complicated, however, because the presence of proteins in the adrenal cortex, which act as 'non-classical' receptors, may bind Czl steroids to different extents, thus reducing rates of steroid release (see Ref. 6). So far as pregnenolone is concerned, there is no barrier to its efflux from the mitochondria where it is formed from cholesterol [SO]. During incubation of rat testis [110], pregnenolone was found to travel from the mitochondria, through the E R and cytosol and then out into the medium. The release with time could be resolved into two components, one rapid and the second, much slower. More than 25% of the pregnenolone remained in the tissue after 150 min. incubation. This two-phase release may reflect the presence of two pools of steroid, the initial loss representing passive dif-

25 fusion and the slower phase being caused by pregnenolone binding to intracellular proteins [111,112]. Numerous other mechanisms, based on ultrastructural evidence, have been proposed [6] by which steroids may be secreted from their site(s) of synthesis. The steroids may be contained in secretory organelles or in lysosomes, these acting as vehicles of transport to the plasma cell membrane, where secretion occurs by exocytosis.

9. Conclusion The foregoing discussion has attempted to trace the ways in which cholesterol, derived from plasma lipoproteins, is converted into the various steroid hormones and how these are secreted back into the blood. Of necessity, many details have had to be omitted but it is hoped that this ‘up-date’ has shown the complexities of steroid biosynthetic pathways and that earlier ‘classical’ ideas have had to be modified as greater knowledge of intermediates, isoenzymes and multiple forms of cyt P-450s has become available. Perspectives for future studies are indeed exciting.

Acknowledgements Work performed in the author’s laboratory was supported by AFRC (grant nos. AG 35135 and 35144) to whom grateful thanks are expressed. Mrs. D.M. Gower kindly prepared the manuscript for publication.

References 1. Morris, M.D. and Chaikoff, 1.L. (1959) J . Biol. Chem. 234, 1095. 2. Brown, M.S., Kovanen, P.T. and Goldstein. J . L . (1979) Rec. Progr. Horm. Res. 35. 215. 3. Havcl, R . J . , Goldstein, J.L. and Brown, M.S. (1979) In: Metabolic Control and Disease (Bondy. P.K. and Roscnherg, L.E., eds.) p. 393. W.B. Saunders. Philadelphia. 4. Brown. M.S., Kovanen, P.T. and Goldstein. J.L. (1981) Science 212, 628. 5 . Gwynne, J.T. and Strauss, J.F.111. (1982) Endocr. Rev. 3, 299-329. 6. Gower. D.B. (1984a) In: Biochemistry of Steroid Hormones. 2nd Edn.. Ch. 8 (Makin , H.L.J., e d . ) pp. 293-348. Blackwell Scientific Publications, Oxford. 7. Jackson, R.L., Morisett, J . D . and Gotto. A.M. (1976) Physiol. Rev. 56. 259. 8. Kane, J.P. (1977) In: Lipid Metabolism i n Mammals, Vol.1. (Snyder, F.. ed.) p. 209. Plenum Press, New York. 9. Faust, J.R., Goldstein, J . L . and Brown. M.S. (1977) J . Riol. Chem. 252, 4861. 10. Schuler, I.A., Scavo, L.. Kirsch, T . M . . Flinkinger. G.L. and Strauss, J.F.111 (1979) J . Biol. Chem. 254. 8662. 11. Ostlund, J . E . Jr., Pfleger, B. and Schonleld, G . (1979) J . Clin. Invest. 63. 75.

12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

Miller. N.E. and Yin, J . A . (1978) Biochim. Biophys. Acta 530, 14.5. Mrotek. J.J. and Hall. P.F. (1977) Biochemistry. 16, 3177. Beckett, G.J. and Boyd, G.S. (1977) Eur. J . Biochem. 72, 223. Boyd, G.S. and Gorban, A.M.S. (1980) In: Recently Discovered Systems of Enzyme Regulation by Reversible Phosphorylation (Cohen. P., ed.) p. 95. ElsevieriNorth Holland Biomedical Press, Amsterdam. Armstrong, D.T. and Flint, A . P . F . (1973) Biochem. J . 134, 399. Moylc, W . R . . Jungas, R . L . and Greep. R . O . (1973) Biochem. J . 134. 407-414. Moyle, W.R., Jungas, R.L. and Grecp. R . O . (1973) Biochem. J . 134,415. Privalle. C.T., Crivello. J.F. and Jefcoate. C . R . (1983) Proc. Natl. Acad. Sci. U.S.A., 80. 702-706. Crivello, C . F . and Jefcoate, C.R. (1980) J. Biol. Chem. 255, 8144. Chaderblan. R.. Noland, B.J., Scallen. R.J. and Vahouny, G.V. (1982) J . Biol. Chem. 257. 8928-8934. Conneely, O.M., Headon. D.R., Olsen. C . D . , Ungar, F. and Dempsey, M.E. (1984) Proc. Natl. Acad. Sci., U.S.A . 81. 297CL2974. Cheng. B.-L. and Kimura, T. (1983) Lipids 18, 577-584. Simpson, E . R . (1979) Mol. Cell Endocrinol. 13, 218-227. Kimura, T. (1986) J . Steroid. Biochem. 25, 711-716. Gruner. S.M.. Cullis. P.R.. Hope. M.J. and Tilcock. C.P.S. (1985) Ann. Rev. Biochem. 14.211-238. Sulimovici, S.I. and Boyd, G.S. (1969) Vitamins Horm. 27, 199. Mitani. F. (1979) Mol. Cell Biochem. 24, 21. Gower, D.B. (1984) In: Biochemistry of Steroid Hormones, 2 nd. Edn. C h . 4 (Makin, H.L.J.. ed.) pp. 1 17-169, Blackwell Scientific Publications. Oxford. Luttrell, B . , Hochberg, R.B., Dixon. W.R.. McDonald, P.D. and Lieberman, S. (1972) J . Biol. Chem. 247, 1462. van Lier. J.E.. Rousseau. J . , Langlois. R . and Fisher, G.J. (1977) Biochim. Biophys. Acta 487.395. Kraaipoel, R.J., Degenhart, H . J . . Leferink. G . J r . , van Beck, V.. De Leeuw-Boon, H . and Visser, H.K.A. (1975) FEBS. Lett. 50, 204. Kraaipoel, R.J., Degenhart, H.J.. van Beck, V.. De Leeuw-Boon. H . , Abeln. G., Visser, H.K.A. and Leferink, J.G. (1975) FEBS Lett. 54. 172. Kraaipoel, R.J.. Degenhart. H . J . and Leferink. J.G. (1975) FEBS Lett. 57. 294. Shikita, M. and Hall, P.F. (1974) Proc. Natl. Acad. Sci. U.S.A. 71, 1441. Orme-Johnson, N.R., Light, D.R.. White-Stevens. R . W . and Orme-Johnson, W.H. (1979) J . Biol. Chem. 254, 2103. Teicher, B.A., Koizumi. N.. Koreeda. M.. Shikita. M . and Talalay, P. (1978) Eur. J. Biochem. 91, 11. Hume, R . and Boyd, G.S. (1978) Biochem. Soc. Trans. 6, 893. Jungmann. R . A . (1968) Biochim. Biophys. Acta 164, 110. Jungmann. R . A . (1968) Steroids 12,205. Hochberg, R.B.. Mickan, H . and Lieberman, S . (1971) Biochim. Biophys. Acta 231. 208. Lieberman, S . , Greenfield, N.J. and Wolfson, A . (1984) Endocr. Rev. 5 , 128-148. Lieberman, S . (1986) J. Endocrinol. 111, 519-529. Wolfson, A . and Lieberman, S . (1979) J . B i d . Chem. 254. 4096. Greenfield, N.J., Ponticorvo. L.. Chasalow, F. and Lieberman, S. (1980) Arch. Biochem. Biophys. 200. 232. Arthur, J . R . , Blair. H.A.F., Boyd, G.S., Mason, J.I. and Suckling, K.E. (1976) Biochem. J . 158, 47. Burnstein, S . , Nickolson. R.C.. Byon, C.Y.. Kimball, H.L. and Gut, M. (1977) Steroids 30, 439. DuBois, R.N., Simpson, E.R., Kramer, R . E . and Waterman, M.R. (1981) J . Biol. Chem. 256, 7000. Gower, D . B . and Cooke. G.M. (1983) J . Steroid Biochem. 19, 1527.

27 50. Shears. S.B. and Boyd, G.S. (1982) Eur. J. Biochem. 123, 153. 51. Aupetit, B . , Bastien, C. and Legrande, J . C (1979) Biochimie 61, 1085. 52. Gower, D.B. (1984) In: Biochemistry of Steroid Hormones, 2nd E d n . , Ch. 7 (Makin, H.L.J.. ed.) pp. 230-292. Blackwell Scientific Publications, Oxford. 53. New, M.I. and Levine, L.S. (1984) In: Biochemistry of Steroid Hormones, 2nd Edn. Ch. 16 (Makin, H.L.J., ed.) pp. 595-632, Blackwell Scientific Publications, Oxford. 54. Kominani, S . , Mori, S. and Takemori. S. (1978) FEBS Lett. 89, 215. 55. Kominani, S . , Ochi, H., Kobayashi, Y . and Takemori, S. (1980) J . Biol. Chem. 255, 3386. 56. Kominani, S., Shinzawa, K. and Takemori, S. (1982) Biochem. Biophys. Res. Commun. 109,916. 57. Mackler, B . , Haynes, B., Tattoni, D.S.. Tippit. D.F. and Kelley, V.V. (1971) Arch. Biochcm. Biophys. 145, 194. 58. Kahnt, F.W. and Neher, R . (1972) Acta Endocrinol. (Kbh) 70, 315. 59. Kuhnle, U.. Chow, D., Rapaport, R.. Pang, S., Levine, L.S. and New, M.N. (1981) J. Clin. Endocrinol. Metab. 52, 534. 60. Akhrem, A . A . , Martsev, S.P. and Chaschin, V.L. (1979) Bioorg. Khim. 5 , 786. 61. Sato, H., Ashida, N., Suhara, K., Itagaki, E . , Takemori, S. and Katagiri. M. (1978) Arch. Biochem. Biophys. 190, 307. 62. Bjorkem, I . and Kalmer, K.E. (197.5) Eur. J . Biochem. 51, 145. 63. Watanuki. M., Tilley, B.E. and Hall. P.F. (1977) Biochim. Biophys. Acta 483, 236. 64. Cheng, S.C.. Suzuki, W . , Sadee, W . and Harding. B.W. (1976) Endocrinology 99, 1097. 65. Whitehouse, B.J. and Vinson. G.P. (1981) In: Hormones in Normal and Abnormal Human Tissues, Vol. 1 (Fotherby, K. and Pal, S.B., eds.) p. 217. Walter de Gruyter, Berlin and New York. 66. Ulick, S. (1976) J . Clin. Endocrinol. Metab. 43. 92. 67. Neher, R . (1979) J. Endocrinol. 81, 25P. 68. Gower, D.B. (1984) In: Biochemistry of Steroid Hormones. 2nd. Edn., Ch.5 (Makin. H.L.J., ed.) pp. 17(&206, Blackwell Scientific Publications. Oxford. 69. Ewing, L. and Brown, B.L. (1977) In: The Testis, Vol. 4 (Johnson, A.D. and Comes, W.R., eds.) p. 239. Academic Press, New York. 70. Slaunwhitc, W.R. and Burgett, M.J. (1966) Steroids 6, 721. 71. Inano, H. and Tamaoki, B. (1978) Acta Endrocrinol. (Kbh) 88. 768. 72. Shimizu, K. (1978) J . Biol. Chem. 253, 4237. 73. Shimizu, K. (1979) Biochim. Biophys. Acta 575. 37. 74. Kwan. T . K . , Taylor, N.F., Watson, D . and Gower , D.B. (1985) Biochem. J. 227, 909-916. 75. Chasalow, F.I. (1979) J . Biol. Chem. 254. 3000. 76. Chasalow, F . I . , Marr, H . and Taylor, G . (1982) Steroids 39, 617. 77. Ahmad, N. and Gower, D.B. (1968) Biochem. J . 108, 233. 78. Kremers, P. (1976) Eur. J. Biochem. 61. 481. 79. Betz, G . , Tsai, P. and Hales, D. (1980) Endocrinology 107, 1055. 80. Betz. G . , Tsai, P. and Weakley, R. (1976) J. Biol. Chem. 251, 2839. 81. Nakajin, S., Shively, J . E . , Yuan, P . M . and Hall. P.F. (1981) Biochemistry 20, 4037. 82. Nakajin, S. and Hall, P.F. (1981) J . Biol. Chem. 256, 3871. 83. Kominani, S., Shinzawa, K. and Takemori, S. (1983) Biochim. Biophys. Acta 755, 163. 84. Geynet, P., D e Paillerets, C. and Allsen. C . (1976) Eur. J. Biochem. 71, 607. 85. Smith, A.G. and Brooks, C.J.W. (1977) Biochem. J . 167, 121. 86. Weintraub, H . , Vincent, F., Baulieu, E . E . and Alsen. A . (1977) Biochemistry 16, 5045. 87. Williamson, D.G. (1979) In: Steroid Biochemistry, Vol.1, (Hobkirk, R . , ed.) p. 83. CRC Press, Florida. 88. Kagurea, E . and Toki, S. (1977) Biochem. J . 163, 401. 89. Inano, H . and Tamaoki, B. (1974) Eur. J . Biochem. 44, 13. 90. Cooke, G.M. and Gower, D.B. (1981) J . Endocrinol. 88, 409.

91. Jeffrey, J . (1980) In: Dehydrogenases Requiring Nicotinamide Coenzymes (Jeffrey, J . . ed.) p. 85. Birkhauser-Verlag, Stuttgart. 92. Houston, B.J., Chisholm, G.D. and Hahih, F.K. (1985) J. Steroid Biochem. 22. 461-467. 93. Moore, R . J . and Wilson, J.D. (1974) Biochemistry 13, 450-456. 94. Scheer, H . and Rohaire, B. (1983) Biochem. J. 211, 65-74. 95. Graef, V. and Golf, S.W. (1975) Z. Klin. Chem. Klin. Biochem. 13, 333-339. 96. Watkins, W . J . . Wilson. L. and Gower. D.B. (1986) Biochem. Soc. Trans. 14, 979. 97. Fotherby, K. (1984) In: Biochemistry of Steroid Hormones, 2nd Edn., Ch. 6 (Makin, H.L.J., ed.) pp. 207-229. Blackwell Scientific Publications. Oxford. 98. Morand. P., Wiliamson, D.G.. Layne. D.S.. Lompa-Krzymien, L. and Salvador, J . (1975) Biochemistry 14, 635. 99. Goto, J . and Fishman, J. (1977) Science 195. 80. 100. Hosoda, H . and Fishman. J . (1974) J . A m . Chem. Soc. 96. 7325-7329. 101. Fishman, J. and Raju, M.S. (1981) J . Biol. Chem. 256. 4472-4477. 102. Hahn, E . F . and Fishman, J. (1984) J . Biol. Chem. 259, 1689. 103. Caspi, E., Wicha, J., Arunachalarn, P . , Nelson, A . and Spiteller, G . (1984) J . Am. Chem. Soc. 106. 7282. 104. Akhtar. M., Calder, M . R . . Corina, D . L . and Wright, J . N . (1982) Biochem. J . 201, 569-580. 105. Stevenson, D . E . . Wright. J . N . and Akhtar. M. (1985) J . Chem. SOC.Chem. Commun. 1078-1080. 106. Thomson, E . A . and Siiteri. P.K. (1974) J . Biol. Chem. 249, 5364. 107. Osawa. Y . and Higashyama. T. (1980) In: Microsomes, Drug Oxidations and Chemical Carcinogenesis (4th Int. Symp. Microsornes Drug Oxidation) (Coon, M.J., Conney. A.H. and Estabrook, R . W . , eds.) pp. 225-228, Academic Press. New York. 108. Purohit, A. and Oakey, R . E . (1978) J . Endocrinol. 112, Suppl., 56. 109. Sihley, C.P.. Whitehouse, B.J., Vinson, G . P . and Goddard. C. (1980) J . Steroid Biochem. 13. 1231-1239. 110. Smalley. A . D . , Taylor. N.F. and Gower. D.B. (1985) Biochem. Soc. Trans. 13, 188-189. 111. Strott. C.A. (1977) J. Biol. Chem. 252, 464470. 112. Strott, C.A. and Lyons. C . D . (1978) Biochemistry 17. 4557-4563.

B.A. Cookc. R.J.B. King and H.J. van dcr Molcn (cd\.) Hormones cind their Actions. Part I 01988 Elscvicr Science Publishers BV (Biomcdiciil Division)

29 CHAPTER 2

A n overview of molecular aspects of steroid hormone action R.J.B. KING Hormone Biochemistry Departmcvit, Itnyerial Cancer Research Fund, P. 0. Box 123, Lincoln’s Inn Fields, London W C 2 A 3 P X , Englund

1. Introduction The minor revolution in our knowledge about molecular aspects of steroid hormone action that has occurred in recent years has led to the questioning of some long-held dogmas, the confirmation of others and the introduction of several new concepts. These are discussed for individual classes of steroid hormone in separate chapters of this volume. This chapter will take a more general look at molecular aspects of hormone action with an emphasis on two topics: the generality of models of action and specificity of action.

2. Intracellular events in steroid action 2.1. Intracellular location of receptors Steroids enter responsive cells by passive diffusion, combine with an oligomeric cytoplasmic receptor, are activated, transported to the cell nucleus where they combine with DNA/chromatin. Statements such as that are legion in articles published over the past decade; some have proven to be correct, but others are being questioned. Two extreme models are presented in Fig. 1. Similarities occur with passive entry into the cell, the existence of an intracellular receptor with high ligand affinity and specificity, and the final site of action being the chromatin. Differences exist in the intracellular locus of some events and the requirements for both oligomeric structures and receptor activation. The original concept of a two-step process leading to chromatin binding remains valid but the processes involved are different in the two models. In A, a cytosolic receptor composed of dissimilar subunits is activated to the chromatin binding form, whilst model B implies that a less complex

30 A.

ACTIVATED

1

protein

\

Protein

\

/

B.

-12‘

I

Fig. 1 . Intracellular events involved in steroid hormone action. A . Model in which the receptor is cytosolic and transfers to the nucleus after binding with steroid (S). The cytosolic ‘8s’receptor consists of a ligand binding unit and other units. one of which is a 90 kDa ‘heat shock’ protein Activation is shown here as involving dimerisation of the ligand binding unit. This simplification of the true events applies to oestradiol receptor but not necessarily for other receptor classes. There is no agreement as to where activation occurs. B. Model in which unliganded receptor ) IS in the nucleus. A conformational change occurs on binding steroid (S) which may result in increased affinity for specific D N A sequences.

(a).

(0)

(0

conformational change is the sole requirement. Whether Fig. 1A or B is nearer the truth remains to be established. More detailed discussions will be found in the other chapters of this volume and in Refs. 1 and 2 but some general points can be made here. The questioning of the existence of cytosolic receptor derives mainly from histochemical experiments with antibodies to the ligand binding unit of the receptor. For oestrogen and progesterone receptors, such data are strongly in favour of a predominantly nuclear locus for unoccupied receptor but the situation is less clearcut for glucocorticoid receptor. As glucocorticoid receptors are of near universal occurrence and glucocorticoids, in contrast to oestrogens, androgens and progestins, are essential for life, it may well be that fundamental differences do exist. Much of the debate on this topic centres around potential artefacts generated either by cell fractionation or by histochemical methodologies. Good discussion of the former can be found in Refs. 3 and 4, and the latter in Refs. 5 and 6. Additional support for the nuclear model derives from experiments in which cultured cells are enucleated without homogenization [7]. Androgens and mineralocorticoids have not

31 yet entered the controversy as antibodies suitable for histochemistry have yet to be identified. As steroid receptors have homologies with thyroid receptors which are known to be nuclear in the unliganded state [8], a major nuclear location might have been anticipated and this represents the current consensus of opinion; there is no consensus as to whether the nucleus is the sole locus.

2.2. Receptor structure As most unoccupied receptor is nuclear, it has been suggested that the molecular forms found in cytosols are homogenization artefacts [l].The validity or otherwise of that view remains to be proven. What is clear is that the classical ‘8s’cytosol receptor is a heterologous structure made up of both ligand binding and non-binding subunits. Whilst the ligand binding units are different in the various steroid receptor classes, a 90 kDa heat shock protein is common to all such classes (9,101. This protein is of near universal occurrence in mammalian cells. The requirement for activation of cytosolic receptors in order to get a nuclear binding form has long been a puzzle both from the molecular point of view and because activation is not required in all cases; progesterone and 1,25-dihydroxy-vitamin D receptors are examples in which no activation is necessary. If the ‘8s’oligomeric receptor is an artefact, then much of the data on activation (Fig. 1A) may be obsolete. Whatever the role of activation might be in changing receptors to a DNA binding form, there is no doubt that a conformational change occurs in the ligand binding unit (Fig. 1B) that exposes a DNA binding domain on the receptor. The somewhat unsatisfactory state of knowledge about biologically relevant forms of receptor contrasts markedly with data on the steroid binding unit. Apart from size, the basic structure of this component is remarkably similar for oestrogen, progestin, glucocorticoid, mineralocorticoid and thyroid hormone receptors; no data are available for the androgen receptor although this omission will soon be rectified [ l l ] . At least three domains involved in ligand binding, DNA binding and regdation, respectively, have been identified, variations in size of the different receptors being primarily due to the latter domain at the N-terminal end of the protein. Furthermore, these known receptors may be members of a much larger family of proteins, the functions of which can only be guessed at. More detailed discussion of this subject will be found in subsequent chapters.

2.3. DNA binding Advances in our knowledge about molecular mechanisms involved in gene regulation have made a major contribution to ideas about nuclear events in steroid hormone action. Multiple regulatory units in the DNA upstream of the mRNA initiation site exist and complex interactions occur between these units which may contribute to the specificity of steroid hormone action (Ref. 12 and see below). One

32 type of unit has been given the general term, hormone response element (HRE) or steroid response element (SRE) if the hormone is a steroid (see Fig. 4). This element has the features of an enhancer of transcription [13] although silencer elements are also known [14]. Steroid receptor complexes bind to SREs and, frequently, multiple such sites exist within the regulatory regions of a gene. In some cases, these sites occur within the structural part of the gene (Chapter 3). The SREs are not directly adjacent to the origin of transcription and current thinking is in favour of protein:protein as well as protein:DNA interactions being involved in steroid regulation of transcription [12,15]. This may explain certain types of steroid specificity (Section 3.5). Whilst the majority of steroid-mediated actions involve transcriptional events, other mechanisms do exist which are documented in the subsequent chapters.

3. Specificity of steroid action From the chemical point of view, steroids are relatively simple molecules comprising four fused hydrocarbon rings with hydroxyl or ketone groups in certain positions. As such they have been classified as ‘low-information’ molecules but despite this nomenclature, they exhibit a very wide spectrum of biological activities. The major factor here is the combination of steroid with its receptor to form a DNA binding complex capable of regulating gene activity. It is frequently stated that the major determinant of steroid specificity is the ligand specificity of the receptor. After allowance for differing half-lives of plasma steroids, this is probably still true but several other molecular mechanisms contribute to the overall picture. Each of the more important mechanisms will be discussed with particular examples as illustrations. 3.1. Ligand availability

Even if a particular cell has multiple receptors, they will not be biologically active in the absence of ligand (Fig. 2A). Uterine epithelial cells represent such a situation, having receptors for oestrogen, androgen, glucocorticoid and progestin but, in women, progesterone is not produced in the first half of the menstrual cycle and there are therefore n o progestational responses. The uterine example also illustrates another general feature namely that oestrogens increase the efficiency of progestin effects by increasing the number of progesterone receptors. Another type of specificity determined by ligand availability occurs when ligand is present in the blood stream but is sequestered therein (Fig. 2B); it is unavailable to the target cell. An extreme form of this situation relates to hypothalamic development in the embryonicineonatal rat [16]. Rats are intrinsically female and a mechanism exists for shutting off the female behavioural and cycling centres in the

33 A.

LIGAND AVAILABILITY

CELL

B.

EXTRACELLULAR SEQUESTRATION OF LIGAND

(B)

Fig. 2. Specificity determined by availability of steroid. A . Even though glucocorticoid and progesterone receptors are present, in the absence of plasma progesterone, only glucocorticoid (G) effects are seen. B. Both oestradiol (E) and testosterone (T) may be present in the plasma but E is not available t o the hypothalamic cell due t o sequestration by nconatal oestradiol binding protein Therefore maternal oestrogen will not affcct the cell whereas, in male foetuses, testosterone can be aromatised to oestradiol within the cell.

(B)

(m).

male hypothalamus; paradoxically, this involves oestrogen. Male rats at this stage of development produce an androgen, testosterone, which, within specific hypothalamic neurones, is aromatized to oestradiol capable of permanently switching off functions attributed to female neurones. Female rats have low testosterone levels so are not susceptible to this unwanted mechanism. They do however contain potentially disastrous amounts of oestrogen of maternal origin. Fortunately for the future of the species, there is a protein in the blood, neonatal oestrogen-binding protein, which tightly binds natural oestrogens thereby protecting the hypothalamus [16,17]. Less extreme situations occur in adult life where plasma levels of cortisolbinding globulin and sex hormone-binding globulin [181 can modulate the biological activities of their ligands, cortisol/progesterone and oestradiolitestosterone, respectively. A particularly interesting recent example of the importance of this type of specificity determinant concerns the actions of mineralocorticoids [ 191. Mineralocorticoids and glucocorticoids exhibit cross-reactivity between their respective receptors, whereas only the latter will bind to cortisol-binding globulin. Hence, the high concentration of this protein in kidney but not brain may result in aldosterone

34 (mineralocorticoid) being the prime glucocorticoid in kidney whilst cortisol (glucocorticoid) assumes this function in the brain.

3.2. Ligand specificity of receptor This is the most important single factor that determines specificity of action and occurs at two levels. Presence or absence of a given receptor determines whether a cell will respond to a given class of steroids whilst ligand specificity controls which particular compound is active (Fig. 3). The oestrogen receptor has the strictest ligand requirements recognizing oestradiol about ten-times more efficiently than oestrone and about a thousand-times more than the androgen, testosterone; progesterone and cortisol are not recognized at all. These recognition efficiencies are reflections of the affinity constants of the receptor. In biological terms, this means that oestradiol is more active than oestrone whilst testosterone can have oestrogenic effects but only at pharmacologic concentrations [20,21]. Androgen, glucocorticoid, progestin and mineralocorticoid receptors have less precise ligand requirements than the oestrogen receptor (Fig. 3 ) [19,20,22]. Thus, the androgen receptor has the highest affinity for dihydrotestosterone with its metabolites being much less effective (Fig. 3). This parallels the relative biological activity of these compounds. However, many progestins, especially the synthetic ones exhibit binding to both progestin and androgen receptor. This dual specificity is reflected in the biological activities of the compounds. Thus, at one time progestins were given to pregnant women to prevent abortion, a practice that was stopped when it was noted that some offspring from such women had clinical features associated with androgen exposure [23].

RECEPTOR

ANDROGEN

ESTROGEN P,

.

A

X *

G,*' P & G inactive Fig. 3. Specificity determined by ligand specificity of receptor. Oestradiol receptor has higher affinity for oestradiol (E,) than oestrone ( E , ) ;androgens (A) such as testosterone have very low affinity whilst progestins (P) and glucocorticoids (G) are inactive. Androgen receptor has less precise specificity recognising both P and G albeit with less affinity than androgens.

35 3.3. Agonismlantagonism Whilst ligand specificity goes a long way towards explaining which compound is active and which is not, there are two types of specificity which require additional mechanisms. The type of response to a given compound is discussed below, but a major enigma concerns the question of whether a compound, when it binds to receptor, will be an agonist or antagonist. Proposals based on rates of association and dissociation and on differing conformational states of receptor, can account for some but not all features [22,24]. For example the anti-oestrogen, tamoxifen, has a lower affinity than oestradiol for the oestrogen receptor than oestradiol, and receptor antibody studies have shown that receptor complexed with tamoxifen has a different conformation to when it binds oestradiol [25]. Taken together, this could account for the competitive interaction characteristic of antagonists producing an inactive complex. However, very high affinity antagonists (hydroxytamoxifen) are known and tamoxifen receptor complexes have some agonist activities such as induction of progesterone receptor [26]. A different type of antagonism is discussed in Section 3.5.

3.4. Availability of responsive genes Situations occur in which receptors are present in cells which are hormone insensitive. This is relatively common in tumour cells where our ignorance as to the molecular mechanisms involved are covered by the use of the term ‘post receptor defect’ [27,28]. In at least one such case, this defect lies within the steroid response element of the gene [28]. This regulatory region usually, but not always, found 5’ to the RNA initiation site, is the DNA acceptor site(s) for the steroid receptor complex (Fig. 4). It has the properties of an enhancer (see other articles for more com-

1-

STEROID RESPONSE

t mRNA INITIATION

RESPONSIVE

UNRESPONSIVE

Fig. 4. Specificity determined by availability of receptor binding regions of DNA. With an oestrogen (E) sensitive gene, the receptor complex binds to specific regions of DNA (steroid response element) which influences the efficiency of mRNA initiation. If the response element is blocked, the gene is unresponsive to the steroid.

36 plete details). Receptor attachment to the hormone response element enhances transcription from the initiation site. This inability of fully functional receptor complexes to activate a given gene also occurs in normal cells. In new-born chicks, a single injection of oestradiol has a delayed effect on vitellogenin synthesis in liver. Having been exposed once to the hormone, second injections elicit an immediate response. It has been suggested that the first injection demethylates methyl cytosine residues within the hormone response element, thereby allowing receptor attachment [29]. A different type of accessibility occurs with the rabbit uteroglobin gene. In the uterus, uteroglobin is undetectable unless stimulated by a progestin. In lung, the gene is expressed in a constitutive manner, but can be modulated by glucocorticoid [30]. In brain, the gene is not inducible at all, despite the presence of both progestin and glucocorticoid receptors. It is thought that chromatin proteins may play a part in determining this type of specificity. It is known that chromatin DNA is more accessible to added deoxyribonuclease in the steroid sensitive as compared to the insensitive state. Yamamoto [12] has pointed out that multiple enhancers and transcription factors exist which can modulate rate-limiting steps in transcriptional activity. Tissue specificity of some of these factors can have profound effects on the type and magnitude of response to a given stimulus within given cell types. 3.5. Specificity of the steroid response element This can be an important means of determining response. The uteroglobin example just mentioned indicates how response can be different from one cell type to another. A more completely analysed example is the genome of mouse mammary tumour virus which was long held to be inducible by glucocorticoids but not other steroids. This erroneous conclusion was made because the experiments were performed with cells that only contained glucocorticoid receptors; provided appropriate receptors are present, androgens, progestins and mineralocorticoids are also effective [19]. DNA binding sites for each of these receptors have been mapped to the steroid response element; only oestrogen will not stimulate the gene complex (Fig. 5A). It has been suggested that the converse situation could also be important. The limited ability of receptors for glucocorticoids, mineralocorticoids and progestins to discriminate between their respective ligands could negate some of their biological specificities. It may be that recognition specificities of the steroid response elements for different genes could counteract this effect. Thus, whether a gene responds to one hormone or another can be determined by the specificity of the steroid response element. This could be determined by a combination of DNA sequence recognition and cell specific proteins involved in mRNA initiation (see Section 2.3). Another level of sensitivity might also be determined by the steroid response element. Thus far, virtually all discussions on agonists/antagonists have centred on

37 A

RESPONSE ELEMENT

9

t mRNA INITIATION

No stimulation

B

t

mRNA INITIATION

AGONISM

ANTAGONISM

Fig. 5. Specificity of steroid response element. A . In this example (mouse mammary tumour virus), the element will bind receptors for glucocorticoids (G). androgen (A), progestin (P) and mineralocorticoid (M) so that each of these classes of steroid stimulate transcription. This type of specificity can vary from cell to cell possibly due to other protein factors (not shown). B. Although oestrogen (E) receptor will not act as an agonist for mouse mammary tumour virus transcription, it may antagonise the agonist activity of glucocorticoids (G).

competition for a given species of receptor (see above). A completely different type of antagonism could occur at the DNA level in which one type of steroid receptor complex with no agonist activity, for example oestrogen, competes for binding sites with a completely different class of receptors, glucocorticoids with agonist activity (Fig. SB). Thus far, examples of this type of competition are anecdotal but are likely to exist. The idea is appealing that, for example, oestrogens can act as antiglucocorticoids as hypothesized in Fig. SB.

References 1. King, R.J.B. (1987) J . Endocrinol. 114. 341-349. 2. Sherman, M.R. (1984) Ann. Rev. Physiol. 46. 83-105. 3. Welshons, W.V. and Gorski. J. (1986) In: The Receptors. Vol. IV, Ch. 4 (Con. P.M., ed.) pp. 97-147, Academic Press, New York. 4. Szego, C.M. and Pietras, R.J. (1985) Nature 317. 88.

5. Gustafsson, J.-A.. Carlstedt-Duke, J., Poellinger, L., Okret, S., Wikstrom, A,-C., Bronnegard, M., Gillner. M., Dong, Y . , Fuxe, K.. Cintra, A . , Harfstrand, A. and Agnati, L. (1987) Endocr. Rev, 8, 185-234. 6. Wikstrom, A.-C., Bakke, 0..Okret, S.. Bronnegard. M. and Gustafsson, J.-A. (1987) Endocrinology 120, 1232-1242. 7. Welshons, W.V., Krummel, B.M. and Gorski, J . (1985) Endocrinology 117, 214(&2147. 8. Samuels. H . H . , Perlman, A.J.. Raaka. B.M. and Stanley, F. (1982) Recent Progr. Horm. Res. 38, 557-599. 9. Okret. S., Wikstrom, A.-C. and Gustafsson, J.-A. (1985) Biochemistry 24, 6581-6586. 10. Renoir, J.-M., Buchou, T. and Baulieu, E.-E. (1986) Biochemistry 25. 6405-6413. 11. Govindan, M.V., S h a r d , J. and Labrie. F. (1987) J. Steroid Biochem. 28 (Supplement), 139s. 12. Yamamoto, K.R. (1985) Ann. Rev. Genet. 19. 109-252. 13. Chambon, P., Dierich. A., Gauh. M.-P.. Jakowley, S., Jongstra, J., Krust, A , , LePennec, J.-P., Oudet, P. and Reudelhuber. T. (1984) Recent Progr. Horm. Res. 40, 1-39. 14. Baniahmad, A , , Muller, M.. Steiner, C. and Renkawitz. R . (1987) EMBO J. 6, 2297-2303. 15. Cordingley, M.G., Richard-Foy. H . , Lichtler, A . and Hager, G.L. (1987) In: Transcriptional Control Mechanisms (Granmer, D., Rosenfeld. M.G. and Chang, s., eds.) pp. 333-342. Alan R. Liss, New York. 16. McEwen, B.S., Biegon, A., Davis, P.G., Krey, L.C., Luine, V.N.. McGinnis, M.Y., Padden, C.M., Parsons, B. and Rainbow, T.C. (1982) Recent Progr. Horm. Res. 38, 41-83. 17. McEwen, B.S., Plapinger, L.. Chaptal, C.. Gerlach, J . and Wallach, G . (1975) Brain Res. 96, 40&406. 18. Siiteri, P.K. (1986) In: Binding Proteins of Steroid Hormones (Foret, M.G. and Pugeat, M. eds.), pp. 593-609. John Libbey, London. 19. Arriza. J .L., Weinberger, C., Cerelli, G., Glaser, T.M., Handelin, B.L., Housman, D.E. and Evans, R.M. (1987) Science 237, 268-275. 20. King, R.J.B. and Mainwaring, W.I.P. (1974) Steroid-Cell Interactions. Butterworths, London. 21. Garcia. M. and Rochefort, H . (1977) Stcroids 29, 11 1-126. 22. Raynaud, J.P., Ojasoo, T. and Labrie, F. (1981) In: Mechanisms of Steroid Action, Ch. 11 (Lewis, G . P . and Ginsburg, M., eds.) pp. 145-158. Macmillan, London. 23. Aarskog, D. (1979) New Engl. J . Med. 300, 75-78. 24. Jordan, V.C., Koch, R . and Lieberman. M.E. (1986) In: EstrogeniAntiestrogen Action and Breast Cancer Therapy, Ch. 2 (Jordan, V.C., ed.) pp. 19-42. University of Wisconsin Press, Wisconsin. 25. Tate, A.C., Greene. G.L., DeSombre, E . R . , Jensen, E.V. and Jordan, V.C. (1984) Cancer Res. 44. 1012-1018. 26. Baulieu, E.-E., Robel, P., Mortel. R . and Levy, A . (1983) In: Steroids and Endometrial Cancer, Vol. 25 (Jasonni, V.M., Nenci, I. and Flarnigni, C., eds.) pp. 61-68, Raven Press, New York. 27. Gehring, U. (1986) Mol. Cell. Endocrinol. 48. 89-96. 28. Darbre, P.D. and King. R.J.B. (1987) Cell 51. 521-528. 29. Saluz, H.P., Jiricny. J. and Jost. J.P. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 7167-7171. 30. Jantzen, K., Fritton, H.P., Igo-Kemenes, T., Espel, E . , Janich, S., Cato, A.C.B., Mugele, K. and Beato, M. (1987) Nucleic Acids Res. 15. 4535-4552.

B . A . Cooke. R.J.B. King and H.J. van der Molcn (eda.) Hormones und their Actions. Purt I 01988 Elsevier Science Publishers BV (Biomedical Division)

39 CHAPTER 3

Gene regulation by steroid hormones MALCOLM G. PARKER Molecular Endocrinology Laboratory, Imperial Cancer Research Fund, P. 0. Box 123, Lincoln’s Inn Fields, London WC2A 3PX, England

1. Introduction Steroid hormones control a wide variety of physiological responses by regulating the expression of specific genes at precise stages during embryonic development and cell differentiation. This is achieved primarily by effects on rates of gene transcription but, for certain genes, post-transcriptional regulation may also be important. This chapter will concentrate on the role of steroid receptors as transcription factors and will discuss the evidence for a general model presented in Fig. 1. The main features of this model are that steroid hormones form a complex with a protein receptor which binds to specific DNA sequences, termed hormone response elements, that, upon activation, function as transcriptional enhancers. These DNA-receptor interactions are accompanied by specific changes in the structure of chromatin which may mediate the action of the steroid by stimulating the binding of transcription factors to promoter elements. The general principle of the model is not new [l]but it is only recently with the cloning of the genes for receptors and the identification of steroid response elements that it has been possible to describe mechanisms of steroid hormone action in detailed molecular terms.

2. Structure and function of steroid receptors The action of the five major classes of steroid hormone, glucocorticoid, mineralocorticoid, oestrogen, progestin, and androgen, is mediated by specific soluble receptor proteins. Following the proposal by Jensen and his co-workers [l],steroid hormone action is believed to involve a two-step process in which steroid initially interacts with a soluble receptor to form a complex which is capable of binding to specific nuclear sites to elicit a response. Recently it has been proposed that the oestrogen receptor may be a nuclear protein even in the absence of steroid [2,3], but this has not been accepted as a general feature of all classes of steroid hormone

40

6 U

m+/.m-piT

I R

Transcription

Fig. 1. A model for the mechanism whereby steroid hormones regulate rates of gcne transcription. Steroids (S) bind with receptors (R) to form a steroid-receptor complex that interacts with DNA sequences called hormone response elements (HRE). These HREs are at variable distances from the gene promoter which frequently consists of upstream promoter elements (UPE) and conserved elements such as T A T A box (TATA).

because glucocorticoid receptors appear to be localized in cytoplasm prior to steroid binding. In any event two functional domains have been defined within all steroid receptors, one is responsible for steroid binding and one is involved in nuclear binding. The two domains have been distinguished in the receptors for glucocorticoid and oestrogen biochemically and immunologically [4,5]and they have been further characterized in mutant forms of the glucocorticoid [6] and androgen receptor (71, and they appear to interact because steroid binding increases the affinity of receptor for DNA. This combination of approaches has identified a third domain in the glucocorticoid receptor, referred to as ‘immunoactive’ , which binds neither A

I

B

C I

D I

E

F

~~

100 %

PR 930

90

I

777

60

GR ER

Vit.

D

Fig. 2. Basic structure of receptors for progesterone (PR), glucocorticoid (GR), oestradiol (ER), thyroxine (T3) and vitamin D (Vit D). The receptors are divided into six domains, A-F [ 151 and the perentage homology of region C for each receptor is compared with that for the progcsterone receptor [9-141. The number of amino acids in each receptor is shown on the right-hand side.

41 DNA nor steroid but may influence DNA binding affinity in vitro and glucocorticoid action in vivo [5]. From recombinant cDNA clones it has been possible to predict the complete amino acid sequence of the receptors for glucocorticoid [8,9], oestrogen [10,11] and progesterone [12-141 and a comparison of these has led Chambon and his colleagues [15] to suggest that receptors may be divided into six domains, A-F (Fig. 2). Region C is conserved for each class of receptor, irrespective of its source and different classes of receptor share from 45 to 90% homology in this region. By alignment of region C, it appears that most variability in the sizes of receptors, results from differences in their N-terminal portions. To investigate the role of individual regions of the receptor deletion mutants of their appropriate cDNA have been expressed in cell-free systems and in intact cells. Using these approaches it has been possible to analyse the steroid and DNA binding properties of mutant forms of receptor and their ability to stimulate transcription of a responsive marker gene. In such studies of the receptors for oestrogen [16] and glucocorticoid [17,18] it has been found that the steroid binding domain resides within region E and the DNA binding domain resides within region C. Although little is known about region E, apart from its hydrophobicity, it should be possible to identify the actual amino acids which bind the lipophilic steroid ligands by a combination of fine deletion mapping and affinity ligand binding. Region C is basic and

FIRST FINGER

.Y . . .

C G S

H Y

A

G

E D

V L

.

.

Y

V

S

L T

c

c

c

c

G

*

c.

I

.

*

w

N

S

c

c

c

c

A

S

c

c

GR

E G

ER

PA

SECOND FINGEq

R R

TFlllA

I

I K N

K N C

I I

P A

c

c

.

R

D R

c GR

c PR

ER

Fig. 3. DNA-binding ‘finger’ structure proposed for ‘TFIIIA [21] and steroid reccptors. The amino acid sequence of the glucocorticoid receptor (GR) lingers are shown. Conserved amino acids in the receptors for progesterone (PR) and oestrogen ( E R ) iirc indicated by an asterisk. whereas differences are show,n.

42

DNA binding domain

Fig. 4. Derepression model proposed for receptor activation [18]. Binding of steroid (S) to receptor (R) results in exposure of a pre-existing DNA binding domain.

contains cysteine residues whose organization resembles that of cysteine and histidine residues in the transcription factors TFIIIA of the Xenopus 5 s gene [ 191 and Kruppel and Serendipity involved in Drosphilu development [20]. These amino acids appear to interact with Zn2+ and it has been proposed that DNA binding ‘fingers’ are generated (Fig. 3 ) that tend to be basic or hydrophobic and interact with the phosphate backbone of DNA in the major groove [21]. Steroid receptors may be members of this class of DNA binding proteins since they contain two potential fingers (Fig. 3 ) . Specificity of steroid hormone action would then reside, at least in part, within the amino acids in the ‘fingers’ that interact with DNA. Interestingly, in view of the overlap between certain glucocorticoid and progesterone response elements [22,23],the receptors for glucocorticoids and progestins share 90% homology in the ‘fingers’ but only 60% with those in the oestrogen receptor which do not interact with these elements. Surprisingly, mutant glucocorticoid receptors which lack the steroid-binding domain are still able to activate hormone response elements associated with the mouse mammary tumour virus (MMTV) promoter [18]. Since the steroid and DNA-binding domains are distinct, it appears that the conformational change following steroid binding, which is associated with activation or transformation of receptor, may be responsible for unmasking rather than inducing a DNA-binding domain (Fig. 4). It is presumed, from these preliminary studies, that the role of regions A and B in transcription is to interact with transcription factors or RNA polymerase 11.

3. Steroid receptor-DNA interactions It has been established that steroid binding and activation of receptor results in increased affinity of steroid-receptor complex for sites within nuclei in vivo and for DNA in vitro [4]. Such binding is low affinity, non-saturable and lacks DNA sequence specificity, yet clearly steroids regulate the transcription of specific genes. Since any protein which interacts with a specific DNA sequence must also interact with other sequences, albeit with reduced affinities, it is assumed that nonspecific

43 binding is masking the detection of binding of steroid-receptor complex to specific high affinity sites. The probability of a receptor complex occupying a specific site depends on this affinity relative to that of non-specific sites and, in consequence, as Yamamoto and Alberts [24] have emphasized, both are important.

3.1. Non-specific D N A binding Mutations in glucocorticoid receptors provide evidence for the importance of nonspecific or low affinity DNA interactions [6]. One source of mutants has been steroid-resistant lymphocytes and lymphomas which are normally killed by glucocorticoid treatment. The most frequent mutation is loss of glucocorticoid binding, referred to as r-, but two other classes with normal steroid-binding properties have been more informative. The first is nuclear transfer deficient (nt-) where receptor binds to DNA 10-100-fold less well than wild-type receptor and the second is nuclear transfer increased (nt') where receptor binds to DNA 10-100-fold more strongly than wild-type receptor. These alterations in affinity for non-specific DNA, together with possible effects on specific binding which perturb their relative affinities is probably sufficient to account for defective receptor. The molecular weights of wild-type and nt- receptor are 94000 whereas the nt' receptor is 39000. A similar relationship has been established by treating wild-type receptor with chymotrypsin in vitro since this results in a product with similar properties to the ntl receptor [ 5 ] . Both molecules have a molecular weight of 39000, lack the 'immuno-active' domain and bind with increased affinity to DNA. One possibility is that the ntl receptor could be generated in vivo by cleavage of full-length receptor but there is n o evidence for a novel chymotrypsin-like protease in mutant cells. Alternatively, since nt' mRNA is approximately 5 kb, in contrast to 7 kb for wild-type receptor mRNA, it is possible that nt' is synthesized as a molecule with a molecular weight of 39 000. 3.2. Specific D N A binding Binding of steroid to specific genomic loci was first demonstrated in Drosophila where it was found that anti-ecdysone antibodies bound to specific chromosomal puff sites at which transcription was stimulated by ecdysone [25].The exact nature of these binding sites on the polytene chromosomes could not be determined but selective binding of steroid receptor complexes directly to DNA has now been demonstrated for a variety of cloned mammalian hormone-responsive genes. The majority of specific DNA-binding sites have been detected near promoters but such sites also exist both far upstream or downstream of the initiation site of transcription (Fig. 5). Their position can be determined by a variety of DNA-receptor binding studies, of which nuclease footprints are the most precise. In this technique, receptors are bound to DNA and treated with nucleases such that binding sites are not digested and the 'footprint' produced can be mapped. It is obviously important

44

-

Transcription

RK!xL!a

lOObp

- , -

GR PR

-,

Mouse M a m m a r y Tumour V i r u s

GR GR PR GR

Human Metallothionein / I A Chicken L y s o z y m e -1

Rat Tyrosine Aminotransferase

GR

Human G r o w t h Hormone

ER

Chicken Vitellogenin

-.J

PR AR

h e

1Kb

Rabbit Uteroglobin

;-

-

R a t C3 gene

Fig. 5. Location of receptor binding sites associated with steroid responsive promoters. Cell-free binding studies were carried out to locate binding sites of receptors for glucocorticoid (GR), progesterone (PR), oestradiol (ER) and androgen (AR).

to ensure that such binding sites are of physiological significance and this can be tested by performing gene transfer experiments. In principle, the gene is introduced into cells by transfection and its expression analysed either transiently or after stable integration of the gene into the host genome. In practice, DNA containing the receptor-binding sites is fused with a promoter and marker gene and the expression of the hybrid is analysed. Many tissue-specific genes are expressed and regulated in cells resembling their natural host more readily than in heterologous cells but frequently it is not possible to culture and maintain the differentiated phenotype of such cells in vitro. It may be possible to overcome this limitation by using a constitutive promoter, which is expressed in all cells in place of the homologous promoter. For glucocorticoids, a combination of approaches has provided an indication of the DNA sequences required for receptor binding that also act as a hormone response element in vivo. Such sequences are not conserved absolutely, but a consensus sequence GGTACANNNTGTTCT has been derived, based on the hormone response elements described for mouse mammary tumour virus [26] and studies of the genes for metallothionein IIA [27], growth hormone [28], tyrosine amino transferase [29] and Moloney sarcoma virus [30,31]. This sequence comprises an imperfect palindrome of 15 nucleotides with the hexanucleotide TGTTCT being most frequently conserved. It is conceivable that the degeneracy in the consensus sequence makes for differences in DNA-receptor affinity, which in turn could lead to differences in the timing or magnitude of the hormone response. Analysis of point mutations in the hormone response elements should establish whether this is in fact the case. It has now been found that the glucocorticoid response elements of mouse mammary tumour virus and the gene for chicken lysozyme also bind progesterone re-

45 ceptors [22], and act as progestin response elements in vivo [23]. In addition, it now appears that androgens act through the same response element [32,33]. Preliminary studies indicate that, while the hormone response elements for glucocorticoids, progestins and androgens overlap, the binding sites are not identical. Fine mapping of the chicken lysozyme gene using nuclease and methylation protection studies indicate that the actual contact points for the progesterone and glucocorticoid receptor are somewhat different and, furthermore, that individual binding sites differ in their relative affinities for these receptors [34]. Therefore, in spite of the overlap in binding of different receptors in MMTV, the most likely basis for steroid specific responses is the DNA sequence of the hormone response element. Thus while some elements may bind multiple classes of steroid receptor complex, other elements may bind a single class. Finally, oestrogen response elements have recently been identified in vitellogenin genes from Xerzopus and chicken and, while they share certain homologies with each other, they appear to be slightly different from response elements for other steroids [35-371. In this respect, it is interesting that the DNA binding domain of the oestrogen receptor is appreciably more different from the receptors for glucocorticoid and progesterone than they are from one another [9-141. Response elements for steroid hormones share many of the properties of enhancer sequences [38] identified first in viruses and subsequently in cellular genes. Enhancers are cis-acting DNA sequences which stimulate transcription of RNA polymerase I1 promoters but are not promoters in their own right; cis-acting refers to the effect of DNA on adjacent DNA sequences in contrast to trans-acting where the effect is mediated from a different molecule of DNA usually via a protein. Enhancers can function independently of orientation over a range of distances from the promoter and they show varying degrees of activity in different types of cells. Such cell specificity, together with evidence from competition studies [39] suggested that trans-acting factors, presumably proteins, are involved in their function; clearly, steroid receptors are one such family of proteins which fulfil this role because they are essential for the activation of hormone response elements. Although the mechanism by which enhancers stimulate transcription rate is unknown, preliminary experiments suggest that they may function to activate promoters by recruitment of transcription factors. Evidence for this suggestion comes from work on MMTV indicating that the binding of NF-1 and factor-1 to the MMTV promoter occurs only when cells are treated with glucocorticoid [40]. An important aspect of enhancers is that they appear to consist of multiple components interacting with a variety of factors and several distinct enhancers can be associated with a single gene, for example the metallothionein gene [41]. This complexity is also apparent in many promoters since their activity may also depend on multiple elements such as the TATA box, CAAT box and GC-rich regions in the DNA. It is also likely that elements associated with promoter and enhancer function are not always discrete but may overlap [30].

46

4. Steroid receptor-chromatin interactions While the mechanism of action of enhancers is unknown there is indirect evidence that their effect is mediated by alterations in chromatin structure and that steroid hormones can induce such changes. Chromatin sturcture has been studied by analysing its sensitivity to nuclease digestion; in general inactive genes are resistant, whereas transcribed genes are sensitive to digestion. Hypersensitive sites have been detected within the sensitive regions, upstream, within and downstream of the transcribed gene and can be induced by hormones in a tissue-specific manner. The hypersensitive sites associated with genes encoding vitellogenin and ovalbumin have been mapped [42,43] and they fall into two classes. After oestrogen withdrawal when transcription diminishes one site disappears and one site persists. Therefore, there is a good correlation between the presence of certain hypersensitive sites and gene transcription, which is consistent with the notion that steroid receptor complexes bind to hormone responsive elements, induce discrete alterations in the structure of chromatin and subsequently increase transcription rate. The absence of such hypersensitive sites within other tissues containing appropriate receptors is further evidence for the involvement of additional tissue-specific factors. It has also been speculated that the second class of hypersensitive sites which persists after hormone withdrawal may play a role in the ‘memory effect’ that follows primary hormone treatment. The primary response to a hormone may proceed only after a lag period, often of several days, but after hormone withdrawal a second treatment with hormone results in a more rapid response [42]. It is conceivable that the persistent alteration in chromatin structure which initially takes place obviates the need for the lag phase in subsequent hormone treatments. DNA methylation is another aspect of chromatin structure which many workers have thought could be involved in steroid-induced gene activation. The onset of expression of certain genes correlates well with specific demethylation but upon hormone withdrawal, when transcription diminishes, the genes remain unmethylated [44]. Therefore it is likely that changes in DNA methylation are not directly involved in mediating steroid responses but may arise as a consequence of gene expression.

5. Steroid hormone-activated gene networks How can we account for the activation of different groups of genes by steroids during cell differentiation? The early models involving generalized repression of transcription by histones being overcome simply by steroid receptor complexes are clearly inadequate. Yamamoto [ 5 ] suggested that the involvement of multiple enhancers in association with promoter elements may form the basis for steroid control of gene networks. Thus, steroid receptor complexes function to activate en-

47 hancers which alone or in combination with additional enhancers may influence the interaction of transcription factors with promoters and thereby regulate rates of gene transcription. With the existence of cell-specific enhancers and promoters, it is possible to account for the wide range of transcriptional activities and cell-specific expression. Although hormone response elements are moderately well conserved, DNA sequence differences could account for differences in receptor affinity and therefore occupancy of the binding site which would change during development following the onset of receptor expression. Finally, Yamamoto [5]has speculated on the existence of multiple enhancer-activating proteins and transcription factors, some of which may be rate limiting for promoter activity. It is thus possible to envisage schemes in which a gene could be transcribed constitutively in a cell type containing one group of factors and transcribed in response to hormone in another cell type containing a different group of factors. One goal during the next few years will be to test these ideas by studying genes in appropriate cells and ultimately in tissues in whole animals, for example, by using transgenic animals.

References 1. Jensen, E.V., Suzuki, T., Kawashima. T.. Stumpf, W.E., Jungblut, P.W. and DeSombre. E.R. (1968)

Proc. Natl. Acad. Sci. U.S.A. 59, 632-638. 2. King. W.J. and Greene, G.L. (1984) Nature 309. 745-747. 3. Welshons, W.V., Lieberman, M.E. and Gorski. J . (1984) Nature 307, 747-749. 4. Greene. G.L., Sobel. N.B.. King. W.J. and Jensen. E.V. (1984) J . Steroid Biochem. 20, 51-56, 5. Yamamoto, K.R. (1985) Am. Rev. Genet. 19. 209-252. 6. Yamamoto, K.R., Gehring. U . , Stampfer. M.R. and Sibley, C. (1976) Recent Progr. Horm. Res. 32, 3-32. 7. Eil. C . (1982) J . Clin. Invest. 71. XS(L858. 8. Hollenberg, S.M., Weinberger. C., Ong. E.S., Cerelli, G . , Oro, A , , Lebo, R . , Thompson, E.B., Rosenfeld. M.G. and Evans, R.M. (1985) Nature 318, 635-641. 9. Miesfeld, R.. Rusconi, S.. Godowski. P.J., Maler. B.A., Okret, S.. Wikstrom, A.-C.. Gustafsson. J.-A. and Yamamoto, K.R. (1986) Cell 46. 389-399. 10. Green, S., Walter, P . , Kumar, V., Krust, A . , Bornet, J.H.. Argos, P. and Chambon. P. (1986) Nature 320, 134-139. 11. Greene, G.L., Gilna, P., Waterfield, M., Baker, A . , Hort, Y. and Shine, J . (1986) Science 231, 11SC-llS4. 12. Coneely, O.M.. Sullivan, W.P.. Toft. D.O., Birnbaumer, M.. Cook, R.G., Maxwell, B.L., Zarucki-Schulz, T., Greene, G.L., Schrader, W.T. and O’Malley. B.W. (1986) Science 233. 767-770. 13. Jeltsch, J.M., Krozowski, Z . , Quirin-Wicker. C . , Gronemeyer, H . , Simpson, R.J., Gamier, J.M., Krust, A., Jacob, F. and Chambon, P. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 5424-5428. 14. Loosfelt. H., Atger, M., Misrahi, M.. Guiochon-Mantel, A , , Meriel, C., Logeat, F., Benarous, R . and Milgrom, E . (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 9045-9049. 1s. Krust, A . , Green, S., Argos, P . , Kumar. V., Walter, P.. Bornert, J.-M. and Chambon, P. (1986) EMBO J . 5, 891-897. 16. Kumar, V . . Green, S., Staub, A . and Chambon. P. (1986) EMBO J . 5, 2231-2236. 17. Giguere, V., Hollenberg, S.M., Rosenfeld, M.G. and Evans, R.M. (1986) Cell 46, 645-652.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44.

Godowski, P.J., Rusconi. S., Miesfeld, R . and Yamamoto, K.R. (1987) Nature 325, 365-368. Miller, J., McLachlan, A.D. and Klug. A . (1985) EMBO J . 4. 1609-1614. Vincent, A . (1986) Nucleic Acids Res. 14, 4385-4391. Rhodes. D. and Klug, A . (1986) Cell 46, 123-132. Von der Ahe, D . , Janich, S., Scheidereit. C.. Renkawitz, R., Schutz, G . and Beato, M. (1985) Nature 313, 706709. Cato, A.C.B., Miksicek, R . , Schutz, G . , Arnemann, J . and Beato. M. (1986) EMBO J. 5,2237-2240. Yamamoto. K.R. and Alberts, B.M. (1975) Cell 4, 301-310. Gronemeyer, H. and Pongs, 0. (1980) Proc. Natl. Acad. Sci. U.S.A. 77. 2108-2112. Scheideret. C., Geisse, S., Westphal, H.M. and Beato, M. (1983) Nature 304, 749-752. Karin. M., Haslinger, A . , Holtgreve, A.. Richards, R.I., Krauter, P., Westphal. H.M. and Beato, M. (1984) Nature 308, 513-519. Slater, E . P . , Rabenan, O . , Karin. M. Baxter, J . D . and Beato, M. (1985) Mol. Cell. Biol. 5, 2984-2992. Jantzcn, H.-M., Strahle, U., Gloss, B.. Stewart. F., Schmid, W . , Boshart. M., Miksicek. R. and Schutz, G . (1987) Cell 49, 29-38. DeFranco, D. and Yamamoto, K.R. (1986) Mol. Cell. Biol. 6. 993-1001. Miksicek, R . , Heber, A . . Schmid. W., Danesch, U., Posseckert. G . , Beato. M. and Schutz, G . (1986) Cell 46. 283-290. Darbre. P . , Page, M. and King. R.J.B. (1986) Mol. Cell. Biol. 6. 2847-2854. Parker, M.G., Webb. P.. Needham. M.. White, R. and Ham, J. (1987) J. Cell. Biochem. 35,285-292. Von der Ahe. D.. Renoir, J.-M.. Buchou, T., Baulieu. E.-E. and Beato, M. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 2817-2821. Jost, J.-P.. Seldran, M. and Geiser, M. (1984) Proc. Natl. Acad. Sci. U.S.A. 81. 429-433. Klein-Hitpass. L., Schorpp, M. Wagner, U. and Ryffel. G . U . (1986) Cell 46, 1053-1061, Seiler-Tuyns. A., Walker, P.. Martiniez. E., Merillat. A . - M . , Givel. F. and Wahli, W. (1986) Nucleic Acids Res. 14, 8755-8770. Parker, M. (1983) Nature 304, 687-688. Scholer, W.R. and Gruss, P. (1984) Cell 36. 403-411. Cordingley, M.G.. Riegel, A.T. and Hager, G.L. (1987) Cell 48, 261-270. Haslinger. A. and Karin. M. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 8572-8576. Burch, J.B.E. and Weintraub, H . (1983) Cell 33, 65-76. Kaye. J.S.. Pratt-Kaye, S., Bellard, M., Dretzen. G . . Bellard, F. and Chambon, P. (1986) EMBO J . 5. 271-285. White, R. and Parker. M.G. (1983) J . Biol. Chem. 258. 8943-8948.

B . A . Cooke, R.J.B. King and H.J. van der Molcn (eds ) Hormones und their Actions, Purr 1 0 198s Elsevier Science Publishers BV (Biomcdical Division)

49 CHAPTER 4

Characterization, assay and purification of steroid receptors M.A. BLANKENSTEIN” a n d E . MULDERb “Department of Endocrinology, Academic Hospital Utrecht and ’Department of Biochemistry 11, Erasmus University Rotterdam, The Netherlands

1. Introduction Steroid hormones achieve their effects on target tissues through intracellular receptor proteins. According to recent views, oestrogen and progestin receptors are localized in the nuclear compartment of the cells, whereas glucocorticoid receptors may reside in both the cytoplasm and the nucleus. Determination of the intracellular localization of androgen receptors awaits the development of (monoclonal) antibodies which will enable immunohistochemical studies. The molecular aspects of the mechanism of action of steroid hormones will be covered in other chapters [ 1-31 in this volume. The present chapter deals with the characterization, assay and purification of steroid receptors. According to an early definition [4],steroid receptors should have: 1) a high affinity towards the ligand; 2) a limited binding capacity; 3) a high ligand specificity; and 4) a high degree of tissue specificity. Application of these criteria is useful for the identification of steroid-binding macromolecules in tissues previously not identified as steroid target tissues. In general, the first three criteria are relatively easy to check, but the fourth criterion. i.e., the demonstration that steroids exert effects in tissues not belonging to the classical steroid target tissues like prostate, uterus or mammary gland, may present some difficulty. Presence of a receptor in a tissue does not always mean that the receptor also serves a purpose in that tissue. This is illustrated by the liver of the cockerel, which contains large amounts of oestrogen receptors, but only synthesizes egg protein (vitellogenin) after administration of oestradiol [ 5 ] .

Correspondence to: M . A . Blankenstein, Department of Endocrinology, Academic Hospital Utrecht, P.O. Box 16250, 3500 CG Utrecht, The Netherlands.

50

2. Properties of steroid receptors 2.1. Binding properties The first two properties of steroid receptors, i.e., the binding affinity and capacity are derived from binding curves. The interaction between a steroid and its receptor obeys the law of mass action and can, accordingly, be represented by: (1)

S+R$SR

in which S represents the unbound steroid, R the unoccupied receptor and SR the steroid-receptor complex. The equilibrium dissociation constant of this reaction can be written as:

Substituting [B] for the concentration of bound steroid, [F] for the concentration of free steroid, and [R,,,] = [R] + [SR] for the total number of receptor sites, this relationship can be rewritten as:

Bound 3H-ORG 2058 (nrnolll)

(A)

0.~1

1

2

3

4

5

6

Total 3 H - O ~2058 (nmolll)

Bound 'H-ORG 2058 (nmolll)

Fig. 1. Assay for progestin receptors in human breast cancer cytosol. A. Saturation analysis. Total ( 0 - 0 ) and aspecific (0-0) binding are measured; Specific binding (solid line) is calculated by subtraction. B. Scatchard plot derived from the binding data by the method of Chamness and McGuire [8]. Numerical data: protein concentration of cytosol. 1.8 mgiml: assay volume, 100 pl; cytosol volume, 50 pi. Calculated values: K d = 0.76 nmolil; progestin receptor content, 510 fmolimg protein.

51 A plot of [B]/[F] vs. [B] will therefore be a straight line with a slope of - l/Kd, the intercept on the [B]-axis being equal to the total concentration of receptors in the preparation. Such a plot is known as a Scatchard plot [6] and it is one of the most popular ways to describe the interaction between a steroid and its receptor. In practice, interference may occur because of the binding of the ligand to other, non-specific, binding sites. Since aspecific binding is relatively unsaturable, the aspecific binding can be corrected for relatively easily. An example of a saturation curve and a Scatchard plot is given in Fig. 1. Both the binding capacity and the affinity of the binding can be read from a Scatchard plot. Occurrence of multiple binding sites and correction for non-specific binding are treated extensively in several reviews [7-9]. The third property of steroid receptors, i.e., their ligand specificity, is inferred from competition studies. In such studies, receptors are labelled with radioactive ligand and competitors are added at different concentrations. After reaching equilibrium, the amount of radioactive label bound to the receptor is a measure of the affinity of the receptor for the competitor. The ligand specificity may depend on the temperature and the duration of the incubation [lo]. This should be taken into account when attempts are being made to extrapolate data obtained in vitro to the living cell. An idealized example of the possible result of a steroid specificity experiment is shown in Fig. 2. The criterion of tissue specificity implies that steroid receptors occur only in target cells and are absent from non-target cells. The syn"1.8

/ B,

loo

1

10-1

100

101

102

103

104

molar excess competitor

Fig. 2. Simulated experiment on the ligand specificity of a steroid receptor. The receptor is labelled with a near-saturating concentration of radioactive ligand. Different concentrations of ligand ( 0 ) and competitors are added. After equilibrium has been reached, the amount of radioactive ligand bound to the receptor is determined. The relative binding affinity of the added competitors is calculated to be 10% (m) and 0.5% (A)respectively.

52 thesis of radioactive ligands with very high specific activities has made it possible to detect minute amounts of binding. The physiological meaning of such observations at present is unclear. We are inclined to accept that the criterion of tissue specificity is fulfilled when the steroid can be shown to have an effect on the cells of the putative target tissue.

2.2. Physico-chemical properties Steroid receptors are thermolabile proteins. Part of the receptors can be extracted from homogenized tissue with low ionic strength. This part is generally referred to as ‘cytosolic’ receptors, i.e., the fraction of receptors that is no longer particle bound after homogenization. With buffers of high ionic strength, i.e., 0.4-0.6 M, receptor sites resistant to extraction at low ionic strength can be extracted. These receptor sites are considered to represent the receptors closely associated to the nuclear material. It has been shown that there are nuclear receptor sites which resist extraction with buffers of high and low ionic strength and treatment with nucleases. These receptor sites are bound to the nuclear matrix of the target cells [ll].The precise role which these receptors play in the mechanism of action of steroids is presently unknown. Due to their protein nature, steroid receptors are prone to proteolytic degradation. Thus several fragments can be made, many of which retain the ability to bind the steroid. The smallest fragment still capable of binding the ligand is called the meroreceptor [12]. Due to the many different forms in which steroid receptors can occur, different values have been reported for the sedimentation coefficient, Stokes’ radius, and other parameters reflecting molecular size. In addition, the cytosolic or non-activated (non DNA-binding) forms of the receptor have a strong tendency to form complexes, e.g., with a 90 kDa heat-shock protein, or with RNA [13]. Some physical properties of different forms of steroid receptors are given in Table I.

TABLE 1 Selected phvsical properties of different forms of steroid receptors

Macromolecular forms’ Multimeric forms Monomeric forms Intermediate forms Mero-receptors

Molecular mass (kDa)

R, (nm)

SZ,l u

25C400 15c-220 7C-100 30-70 17-25

X.(k9.0 6.C7.0 5.2-6.0 2.5-4.0 1.9-2.4

9-12 6-8 4-5 3-4 2-3

.‘May contain non-steroid binding proteins.

53

3. Assay of steroid receptors 3.1. General aspects and radioligand assays For the assay of steroid receptors, one or more of their properties is employed. Classically, receptor assays are based on binding of a radioactive form of the ligand. Such assays are relatively simple to perform. A suitable receptor preparation is incubated with either a single saturating concentration or a number of different nonsaturating concentrations of the radioactive ligand. After equilibrium has been reached, the excess ligand is separated from the receptor-bound ligand and the latter is quantified. When multiple ligand concentrations are used, a Scatchard plot can be constructed from which the number of receptors as well as the binding affinity can be read as explained in the previous section of this chapter. When a single saturating dose assay is used, which does not allow extrapolation to the true number of receptors, the number of receptors is slightly underestimated. Moreover, no information on the affinity of the binding observed is obtained. In those cases where one deals with a well characterized preparation and relatively high receptor levels these drawbacks can be overcome. The amount of non-specific binding is generally estimated by performing parallel incubations in the presence of a large (10U-1000-fold) molar excess of radioinert ligand. When the native hormone is prone to metabolic degradation during the assay procedure, a synthetic ligand which is not metabolized is often used. Thus progestin receptors are preferably assayed with [3H]ORG 2058 (16a-ethyl-21-hydroxy19-nor-pregn-4-ene-3,20-dione). When binding of the ligand can also occur to binders which are likely to contaminate the receptor preparation such as sex hormone binding globulin (SHBG) in human breast cancer cytosols, care must be taken to correctly estimate the amount of aspecific binding. Thus radioinert diethylstilboestrol (DES), which does not bind to SHBG, is used instead of radioinert oestradiol for the estimation of oestrogen receptors in human breast cancer tissues [ 141. Serious underestimation of the number of receptor sites may occur when the receptors are occupied with endogenous ligand. To circumvent this problem, assays have been developed, in which the endogenous ligand is exchanged for the radioactive ligand during prolonged incubation at temperatures of 2CL37"C. Such conditions must be chosen carefully, in order not to destroy the receptors as a result of increased proteolytic activity due to the elevation of temperature. Different tissues may require different conditions in this respect. Because of the inherent risk of underestimation of the number of receptor sites, exchange assays in general are not particularly well suited for absolute measurements. By contrast, these procedures can be very useful for comparative measurements.

3.2. Separation of bound and free ligand For the separation of receptor-bound and free ligand, many procedures have been devised. Among these are: absorption of unbound steroid to dextran-coated charcoal (DCC); absorption of steroid-receptor complexes to hydroxyl-apatite; precipitation of steroid-receptor complexes by protamine sulphate; agar gel electrophoresis; gel filtration on Sephadex G25 or LH-20. More sophisticated methods include high-performance liquid chromatography (HPLC) sucrose gradient analysis and isoelectric focusing. Absorption to DCC and precipitation by protamine sulphate are rapid and simple methods for the separation of bound and free steroid. They have the advantage that many incubations can be performed simultaneously, but the disadvantage is that no other information but the amount of radioactivity bound becomes available. Because of its simplicity, the DCC method is very popular when Scatchard plot analysis is performed. In addition to the concentration of steroid receptors, many of the other methods listed provide a physico-chemical property which may be helpful in the identification of the binding agent under investigation. Thus sucrose gradient analysis provides the S-value of the binding protein, isoelectric focusing the isoelectric point, gel electrophoresis the electrophoretic mobility, and the various chromatographic procedures may reveal information on the molecular size.

3.3. Immunological assays Receptor assays discussed thus far are all based on binding of the ligand and require homogenization of tissue. This causes several disadvantages. First of all, the assays critically depend on the presence of the steroid binding site. It is not possible to study synthesis and processing of the non-steroid-binding part of the receptor molecule unless the steroid binding moiety is attached to it and the latter is labelled with the radioactive hormone. The second disadvantage is that due to the homogenization procedure information on the localization of receptor-containing cells in a tissue may be lost. T o overcome these problems many attempts have been made to prepare fluorescent steroid derivatives, which would visualize the receptors histochemically. Unfortunately, however, these attempts have not resulted in reliable methods for the histochemical detection of steroid receptors, mainly because the fluorescent derivatives did not have sufficient affinity for the receptors (15,161. The development of monoclonal antibodies against steroid receptors has made it possible to design immunohistochemical assays [ 17,181 as well as enzyme immunoassays [19,20] for steroid receptors. Results of enzyme immunoassays and radioligand binding assays for oestrogen and progestin receptors correlate very well [21]. These assays have the modest advantage that no radioactivity is being used, but the cost of reagents has prevented general introduction thus far. Nevertheless, some remarkable successes have been achieved. In fact, our current view on the intracellular localization of steroid receptors [22,23] is based to a large extent on the re-

55 sults obtained with immunohistochemical assays. One of the potentials of these assays is to provide information on the heterogeneity of tissues, e.g., breast cancer, and to perform receptor assays on tissue samples too small to be processed for radioligand assays, such as fine needle aspirates. Thus, monoclonal antibody-based immunohistochemical receptor assays may help to improve the success of breast cancer treatment.

3.4. Other steroid receptor assays When only limited amounts of cells are available (e.g., of human leukocytes for estimation of corticoid receptors or of human skin fibroblasts for estimation of androgen receptors) an estimation of receptors in separated cell fractions might not be feasible. Total numbers of binding sites per cell may be obtainend if different concentrations of suitable highly specific ligands are used and incubation conditions of intact cells are carefully calibrated. In addition a good separation between medium with free steroid or steroids loosely adsorbed to the cell surface and steroid bound to the receptor inside the cell is essential. In one procedure [24] cells are rapidly centrifuged through an oil layer with a density in between the density of cells and medium. In this way cells and medium are not only separated within seconds, preventing redistribution of bound steroid, but also steroid adhering to the cell surface is removed by the lipophilic separation layer. Information on the rate of synthesis or turnover of receptors can be obtained from studies with dense amino acids, e.g., in studies with cultured cells. The isolated steroid receptor complexes are separated by density gradient centrifugation and the presence of a faster sedimenting form of the receptor, containing the dense aminoacids is monitored [25].

4. Purification of steroid receptors Purification of steroid receptors has been of crucial importance for the progress of research on the mechanism of action of steroid hormones. Four approaches for purification of the receptors will be summarized. 4.1. General protein purification

In early studies partial purification of oestrogen and progestin receptors was achieved with ammonium sulphate precipitation and DEAE-Sephadex chromatography. Thus, partial purification of oestrogen, progestin and glucocorticoid receptors [26] has been reported. In more recent work HPLC has been used for separation on basis of ion exchange properties of intact steroid-receptor complexes [27]. The con-

56 centration of steroid receptors in tissues is extremely low (between 0.02% of the proteins for glucocorticoid receptors and 0.001 % for androgen receptors) and, consequently, extensive purification is necessary.

4.2. DNA-afinity chromatography Steroid receptors possess both a steroid-binding site and a DNA-binding site. Therefore advantage has been taken of the properties of steroid receptors to interact with DNA and DNA-like structures and, in addition, of the property of receptors that this affinity is not always present but can be acquired after an activation or transformation process. In a differential chromatography procedure, receptors in the inactivated form, with a low affinity for polyanions (cytoplasmic form, described in Section 2.2), are passed through DNA- or phosphocellulose. Other DNAbinding proteins are then retained by the column and are eluted and discarded. After activation to the DNA-binding state the receptor is selectively bound to these matrices and can be eluted free from contaminating DNA-binding proteins. With this approach considerable success has been obtained in the purification of glucocorticoid [28] and progestin receptors. An example of one of the first successful isolations of glucocorticoid receptors based only on a difference in the affinity for DNAand phosphocellulose is given in Table 11.

4.3. Steroid affinity chromatography Purification of some steroid receptors to near homogeneity has been accomplished with steroid affinity chromatography. Advantage has been taken of the availability of an unoccupied steroid-binding site on the receptor in tissues obtained from sources with low endogenous hormone concentration (e.g., oestrogen receptors not occupied by oestradiol in immature calf ulterus). In these studies matrices with immobilized steroids were used. Considerable attention was paid to the length of the sidearm connecting the steroid with the carrier matrix, to permit optimal interaction

TABLE I1 Purification of glucocorticoid receptors by differential chromatography [24] mg protein

Cytosol Phosphocellulose flowthrough DNA-cellulose flowthrough activation (30 min, 25°C) DNA-eluate after binding Recovery of receptor: 38% Receptor purification: 6700-times

4450 3500 3500 0.2

57

HO

JXY

Sepharose-66

Fig. 3. Structure of an oestradiol derivative. used for affinity chromatography [29].

with the receptor. In addition, the high degree of receptor purification needed, and the large excess of steroid used, on these affinity-matrices made it necessary to choose very stable linkages between steroid and carrier to prevent even a limited loss of steroid due to chemical or enzymatic attack. An example of a ligand used for steroid affinity purification of oestrogen receptors [29] is shown in Fig. 3.

4.4. Immunoaffinity purification Partially purified receptor preparations have been used for production of polyclonal and monoclonal antibodies to receptors. If these antibodies are available they might be used for immunoaffinity chromatography of receptors. This procedure permits rapid isolation of highly purified receptors and has been used, e.g., for separation of different phosphorylated forms of progestin receptors [30].The right choice of a suitable reagent for elution of receptors from the immunoaffinity column in a form still able to bind the steroid was crucial in these studies.

5. Characterization of steroid receptors Classical characterization procedures of non-purified receptors were based on behavior of the radioactive label of the steroid bound to the receptor. Sucrose gradient centrifugation and gel filtration have been used for estimation of size and molecular mass (Table I, Section 2.2). Iso-electric focussing under non-denaturing conditions revealed iso-electric points between 4.5 and 6.0 for the monomeric forms. With respect to the estimation of molecular mass of the receptors SDS-PAGE electrophoresis has recently also been successfully applied. A prerequisite for the application of this technique is the covalent attachment of the ligand to the steroid binding domain of the receptor molecule (affinity labelling). Several procedures have been used for covalent attachment of the steroid to proteins, e.g., via photoactivation of highly conjugated synthetic ligands [31].. An example of the result of a photoaffinity labelling experiment in which R1881, a synthetic androgen was used for androgen receptors [32] is shown in Fig. 4. Chemical linkages with protein might be obtained through irreversible covalent coupling with reactive groups (e.g., by reaction with thiol groups of methanesulfonate esters of dexamethasone, a synthetic glucocorticoid or with an aziridine derivative of tamoxifen, an anti-estrogen)

dpm loo

MW

1

93K

69K I

L6K I

98K

I

10

20

30 LO gel slice number

Fig. 4. SDS-PAGE profiles of photolysed ['H]R1881-labelled androgen receptors. belled receptor only; 0 ,control experiment without labelled receptor [32].

0,

irradiation of la-

[33,34]. Affinity labelling not only permits the identification of proteins at very low abundance in cytosolic preparations of high complexity, but also the unequivocal identification of steroid binding proteins at any stage of a purification procedure. The availability of purified receptor preparations and monoclonal antibodies against almost all steroid hormone receptors has recently permitted the use of the standard procedures of molecular biology for isolation of complementary DNA of different receptors and deduction of the primary amino acid composition [35,36]. DNA-binding domains with considerable homology were found in all receptors both for steroid and thyroid hormones. The domains have the potential to form two DNAbinding fingers ('Zn-fingers', each with two pairs of cysteines). The exact position of the steroid molecule in the receptor molecule and the amino acids involved in the binding of the steroid are not yet known, but the tools for these studies are now available (transfection of receptor cDNA and expression in suitable cells, affinity labelling and estimation of amino-acid sequence).

References 1. 2. 3. 4.

King, R.J.B. (1988) This volume. Sutherland, R. (1988) This volume. Mainwaring, W.I.P. (1988) This volume. Baulieu, E.-E., Alberga, A., Jung, I.. Lebeau, M.-C., Mercier-Bodard, C.. Milgrom, E., Raynaud,

59 J.P., Raynaud-Jammet, C., Rochefort, H.. Truong, H. and Robel, P. (1971) Recent Progr. Horm. Res. 27, 351-412. 5 . Tata, J.R. and Smith, D.F. (1979) Recent Progr. Horm. Res. 35,47-90. 6. Scatchard, G. (1949) Ann. N.Y. Acad. Sci. 51, 66C-672. 7. Munck, A. (1976) In: Receptors and Mechanism of Action of Steroid Hormones, Part I, Ch. 1 (Pasqualini, J.R., ed.) pp. 1-40. Marcel Dekker, New York. 8. Chamness, G.C. and McGuire W.L. (1975) Steroids 26, 538-542. 9. Clark, J.H. and Peck, E.J. (1977) In: Receptors and Hormone Actions, Vol. I , Ch. 11 (O'Malley, B.W. and Birnbaumer, L., eds.) pp. 383-410. Academic Press, New York. 10. Raynaud, J.-P., Bouton, M.M., Moguilewsky. M., Ojasoo, T., Philibert, D., Beck, G . , Labrie, F. and Mornon, J.P. (1980) J. Steroid Biochem. 12, 143-157. 11. Barrack, E.R. and Coffey, D.S. (1982) Recent Progr. Horm. Res. 38, 133-189. 12. Sherman, M.R., Pickering, L.A., Rollwagen, F.M. and Miller, L.K. (1978) Fed. Proc. 37, 167-173. 13. Sullivan, W.P., Vroman, B.T., Bauer, V., Puri, R.K., Riehl, R.M., Pearson, G.R. and Toft, D.O. (1985) Biochemistry 24, 4214-4222. 14. EORTC Breast Cancer Cooperative Group (1980) Eur. J . Cancer 16, 1513-1515. 15. Berns, P.M.J.J., Mulder, E., Rommerts, F.F.G., Blankenstein, M.A., de Graaf, E. and van der Molen, H.J. (1984) Breast Cancer Res. Treatm. 4, 195-204. 16. Lamml, A,, Krieg, M. and Klotzl (1983) Prostate 4. 271-282. 17. Press, M.F. and Greene, G.L. (1984) Lab. Invest. 50,48&486. 18. Perrat-Applanat, M., Logeat, F., Groyer-Picard. M.T. and Milgrom, E. (1985) Endocrinology 116, 1473-1 484. 19. Leclercq, G., Bojar, H., Goussard, J.. Nicholson. R.I., Pichon. M.-F., Piffanelli, A , , Pousette. A , , Thorpe, S. and Lonsdorfer, M. (1986) Cancer Res. 46,4233s-4236s. 20. Jordan, V.C., Jacobson, H.I. and Keenan, E.J. (1986) Cancer Res. 46, 4237s-4236s. 21. Blankenstein, M.A., van der Meulen-Dijk, C. and Thijssen, J.H.H. (1987) Clin. Chim. Acta 165, 189-195. 22. King, W.L. and Greene, G.L. (1984) Nature 307. 745-747. 23. Welshons, W.V., Lieberman, M.E. and Gorski, J . (1984) Nature 307, 747-749. 24. McLaughlin, W.H., Milius, R.A.. Gill, L.M., Adelstein, S.J. and Bloomer, W.D. (1984) J. Steroid Biochem. 20, 1129-1 133. 25. Scholl, S. and Lippman, M.E. (1984) Endocrinology 115, 1295-1301. 26. King, R.J.B. and Mainwaring, W.I.P. (1974) Steroid-Cell Interactions. Butterworths, London. 27. Hyder, S.M., Wiehle, R.D., Brandt, D.W. and Wittliff, J.L. (1985) J . Chromatogr. 327, 237-246. 28. Wrange. O., Carlstedt-Duke, J. and Gustafsson, J.-A. (1979) J. Biol. Chem. 254, 9284-9290. 29. Greene, G.L., Nolan, C., Engler, J.P. and Jensen, E.V. (1980) Proc. Natl. Acad. Sci. 77, 5115-5119. 30. Logeat, F., Le Cunff, M., Pamphile, R . and Milgrom, E. (1985) Biochem. Biophys. Res. Commun. 131, 421-427. 31. Dure, L.S., Schrader, W.T. and O'Malley. B.W. (1980) Nature 283,784-786. 32. Brinkmann, A.O., Kuiper, G.G.J.M., de Boer, W., Mulder, E., Bolt, J., van Steenbrugge, G.J. and van der Molen, H.J. (1986) J. Steroid Biochem. 24, 245-249. 33. Katzenellenbogen, J., (1984) Vitamins Hormones 41, 213. 34. Gronemeyer, H. and Govindan, M.V. (1986) Mol. Cell. Endocrinol. 46, 1-19. 35. Green, S. and Chambon, P. (1986) Nature 324, 615-617. 36. Weinberger, C., Hollenberg, S.M., Rosenfeid, M.G. and Evans, R.M. (1985) Nature 318,67@672.

This Page Intentionally Left Blank

B . A Cooke. R.J.B. King and H . J . van der Molen (eds.) Hormones und their Action>. Purr 1 @ 1988 Elsevier Science Publishers BV (Biomedical Division)

61 CHAPTER 5

Mechanism of action of thyroid hormone JACQUES NUNEZ INSERM U 282, Hbpital Henri Mondor, 51, avenue du Markchal de Lattre de Tassigny, 94010 Creteil, France

1. Introduction The dominant view during the last two decades [l] is that thyroid hormones (Fig. 1) stimulate the synthesis of specific proteins in a variety of tissues of vertebrates both during development and in adulthood [ 2 ] .It is also widely accepted that their action is mediated by the interaction of the active form of the hormone, 3,5,3’-triiodothyronine (TJ, with specific nuclear receptors [3,4] which are present in all types of target cells. The hormone-receptor complex then reacts with specific genes thus inducing the corresponding m-RNAs. Depending on the cell type and/or the developmental stage different proteins are induced or repressed. In other words the response to hormone stimulation is specific to the cell type and to the developmental stage. This probably explains why thyroid hormone produces so many different effects in vivo, i.e., stimulation of carbohydrate, lipid , cholesterol, bone and skin metabolism, changes in muscle and heart activity, developmental regulation of maturation of brain, muscle, bone, etc. The well known calorigenic effect of thyroid hormone, which has been documented in a variety of tissues, probably results both from an action on the expression of membrane Na+-K+-ATPase [ 5 ] and of several mitochondria1 enzymes. Important features of thyroid hormone action are that some of the induced proteins are either 1) other hormones or growth factors that regulate other cell types or 2) receptors for regulatory signals; this can change the sensitivity of the target cell to other hormones, growth factors or neurotransmitters or 3) regulatory enzymes catalysing the production, degradation or release of intracellular second

Abbreviarions: 3,5,3‘,5’-tetraiodothyronineor thyroxine, T,; 3,5,3-triiodothyronine, T,; 3,3’,5’-triioacid, TETRAC; 3,5,3’-triiodothyroacetic dothyronine, reverse Tzor r-T3;3,5,3’.S’-tetraiodothyroacetic acid, TRIAC; 3,5-diiodotyrosine, DIT; 3-monoiodotyrosine, MIT; thyrotropic hormone, TSH; thyreoliberin, TRH; growth hormone, GH; microtubule associated proteins. MAPs; TAU protein, one of the brain MAPs.

62

yf OH

0

Thyroxine CT,)

3,3’-DI iodothyronine

$‘ I 0

3.5,3’-Tr110dothyron1ne(T3)

3,3,5’ - T r I I o d o t hyron I n e

( r -T3 )

Fig. 1. Major thyroid hormones (3.5.3’,5’-tetraiodothyronineor thyroxine, T,) and (3,5,3’-triiodothyronine, T,) and other important iodothyronines (3,3S’-triiodothyronine or reverse-T,, r-T,, and 3.3‘diiodothyronine 3,3’-T2). T, and T, are active, r-T, and 3.3’-T2 are inactive. The acetic derivatives of T, (TETRAC) and T, (TRIAC) are produced by oxidative decarboxylation o f the alanine side-chain and have thyrornimetic activities.

messengers or 4) key unidirectional enzymes belonging to major metabolic pathways which are subject to multihormonal regulation or 5 ) ionic pumps or cytoskeletal and contractile proteins etc. This means that thyroid hormone action triggers in vivo both a large number of direct effects, which differ depending on the cell type, and a cascade of secondary effects. Yet their primary mechanism of action, i.e., the interaction with a specific and almost ubiquitous nuclear receptor seems to be identical whatever the cell type and the developmental stage. What probably changes from cell to cell (and depending on the developmental stage) are the genes which are able to bind the hormone-receptor complex and which are activated or unactivated as a result of this interaction. In this chapter we will first briefly describe the major parameters controlling the level of thyroid hormone production and its concentration by the target cell (see Ref. 6), i.e., biosynthesis and output from the thyroid gland, transport in the blood, conversion of the prohormone thyroxine (T4), to the active form, 3,5,3’-triiodo-

63 yronine (T,) in the target cells. The regulatory mechanisms which control thyroid hormone biosynthesis and secretion in the thyroid gland (i.e., thyrotropin (TSH) and thyreoliberin (TRH)) actions are described in another chapter of this volume. In a second section of this chapter we will review the available data on the different cellular thyroid hormone-binding proteins so far described and on the nuclear receptor. Finally, we will analyse in some detail few examples of specific thyroid hormone effects. We have selected them from many others to illustrate the major types of response elicited by these hormones depending either on the cell type or on the developmental stage, namely 1) induction of growth hormone (GH) and repression of thyrotrophic hormone (TSH) in the pituitary; 2) regulation of lipogenesis in the liver (induction of malic enzyme and of fatty acid synthase); 3) changes in the expression of fetal and adult myosin heavy chains in the heart; 4) changes in P-adrenergic responsiveness in the heart and of the activity of the adenylate cyclase system in the adipocyte; 5 ) stimulation of neuronal and glial differentiation during brain development. No attempt has been made in this chapter to refer to all articles published on the topics listed above.

2. Thyroid hormone production, transport and uptake by the target cells The concentration of thyroid hormones in a given responsive cell depends on a number of factors (Fig. 2) including I) the production of the two major hormones in the thyroid gland and their secretion in the blood (see Ref. 6) (both the synthesis and the secretion processess are themselves regulated by TSH); 2) the transport of T, and T, in the blood and then their degradation in the liver and in the kidney; 3) the uptake of thyroid hormones by the different target cells; 4) the peripheral conversion of the prohormone, T,, to the active (T,) and the inactive (‘reverse’ T, or r-T,) derivatives in the responsive cells. The biosyntheszs of thyroid hormone (see Ref. 7) occurs in specialized epithelial cells of the thyroid gland. An enzyme, thyroid peroxidase, catalyses the iodination of several tyrosine residues of a large glycoprotein, thyroglobulin (660000 Da). Some of the monoiodotyrosine (MIT) and diiodotyrosine (DIT) residues produced during the iodination reaction are ‘coupled’ to hormone residues by the same enzyme. Thyroxine is thus formed by the coupling of two DIT residues and T, by the coupling of one MIT and one DIT. Small amounts of r-T, and 3,3’-diiodothyronine are also produced. The amino acid sequence of the thyroglobulin molecule, which has been established recently [8], have confirmed that this protein contains a limited number of ‘coupling’ sites, i.e., three for the formation of T, and one for T,. Since the level of iodination of thyroglobulin varies in vivo depending on the iodide sup-

64 Vypothalamus

Pi t u i t o r y

Iodide

-

A

bi osyn thesis

Entero-hepatic circulation

C o n ] u go t I on Deiodinotion

f E l i m i n o t ion

-

I T a r a e t Cells

II T4

\

T3

r- Tg

conversion

Fig. 2. Different parameters contributing to the hormone concentration in blood and in the target cells: 1) hypothalamic (TRH)-pituitary (TSH) control of 2) thyroid hormone biosynthesis by the thyroid gland, 3) enterohepatic circulation of T, and T, and 4) conversion of T, to either T, or r-T, by the target cells.

ply this protein may contain a varying number of MIT and DIT residues and therefore different T, and T, contents. Once iodinated the molecule of thyroglobulin is stored within the thyroid follicle: this provides a large store of hormones readily available upon TSH stimulation. A complex pathway involving proteolytic breakdown of thyroglobulin results in the release of free T, and T, that are secreted in the blood where they are transported by specific serum proteins. They are then taken up by the different target cells. The translocation of thyroid hormone across the plasma membrane seems to depend on a carrier-mediated, energy-dependent transport [9,10]. In these cells T4 is converted to T3 or r-T, by partial deiodination which is catalysed by different enzymes (see Chapter 6 in this volume). Both T, and T3can also be converted to their acetic derivatives (TRIAC and TETRAC) which have thyromimetic activities and to other inactive metabolites (diiodothyronines and monoiodothyronines). Most of the cellular T, and r-T, is thus formed peripherally.

3. Thyroid hormones, nuclear receptors and cellular binding protein, Several proteins, with binding activities for thyroid hormones, have been detected in a variety of cell types and with different subcellular localizations, i.e., in the plasma membrane, the sarcoplasmic reticulum, the cytosol, the mitochondria and

65 the muscles. However little is known about the structure and the role of the nonnuclear-binding sites. We will see below that a large body of evidence shows that the nuclear-binding sites are the major thyroid hormone receptors if not the only ones. A mitochondrial T3-binding protein has been detected [ l l ] but its function and physiological relevance remain largely unknown. Recently, it has been shown that a mitochondrial enzyme, adenine nucleotide transferase, exhibits high affinity binding of T, [12]. Membrane-containing fractions displaying T3-binding activities have been detected in a variety of cell types [13-161. Rat liver (151 and erythrocyte [16] plasma membranes, for instance, contain T,- and T,-binding sites with affinities ranging from 1 to 10 x lop"' M for T,. It is not clear whether the function of these sites is related to the transport of thyroid hormones from the blood to the cell or if they represent receptors responsible for non-nuclear effects of thyroid hormones [17,18]. Cytosol-binding proteins have been detected in almost all responsive tissues including muscle, brain liver, kidney and blood cells [19-301. The molecular mass of these proteins varies from 45 000 to more than 100000 Da. The binding affinity also varies depending on the tissue examined. In most cases the affinities of the cytosolic receptors for T4 have been found to be greater than those for T,, but it is not established whether there exist separated binding proteins for T, and T,. Since the capacity of the cytosol to bind T, and T3 has been found to be much larger than that of the nucleus, the assumption has been made that the soluble sites are a reservoir for thyroid hormones. Other unsolved possibilities are that the cytosolicbinding sites play a role either in the conversion of T, to T, or in their transport into the cell.

3.1, Nuclear receptors Limited capacity and high affinity binding sites for thyroid hormones are present in the nucleus [31] of thyroid hormone responsive cells [32], where they are associated to the chromatin [33]. These sites bind thyroid hormones at physiological concentrations. The estimated equilibrium dissociation constants (Kd) for L-T3is 0.029 nM and for T, 0.26 nM indicating that the affinity for T, is ten-times higher than that for T, [34]. In most cells D-T, is also bound to isolated nuclei with an affinity higher than that of L-T, [35]. The deaminated derivative of T,, L-TRIAC, exhibits an affinity similar or even higher than T, for the nuclear receptor. However, in general thyroid hormone analogues bind to the nuclear sites in direct proportion of the analogues thyromimetic potency when the in vivo half life of the analogues and their ability to enter the cell have been taken in consideration. The nuclear receptor is an intrinsic chromosomal acidic non-histone protein whose localization in the nucleus is not dependent on the presence of the hormone. It can be extracted by salts [36] yielding a 3.5-3.8 S [37] protein having a molecular mass

66 of approximately 50 000 Da [38]. Treatment of chromatin by micrococcal nuclease releases a predominant 6.5 S form and a 12.5 S less-abundant species (which represents mononucleosome particles) from GH, cell nuclei [39]. The 6.5 S entity seems to be composed of the 3.8 S receptor and additional proteins which are associated with a 35-40 base fragment of DNA [38]. After complete micrococcal nuclease treatment no binding sites are present in the residual nuclear matrix fraction. Affinity labelling of thyroid hormones to nuclear receptors indicated the presence of an abundant 47000 Da component and a less abundant 57000 Da species [40]. Micrococcal nuclease also excises the two receptor forms. It is yet not clear whether these two forms are products of different genes or if the 57000 Da form is converted or processed to the 47000 Da species. Recently it has been shown independently by two groups [41,42] that the cellular counterpart (c-erb-A) of the viral oncogene V-erb-A encodes the thyroid hormone receptor. The protein encoded by this oncogene binds T,, T, and other thyroid hormone analogues similarly to the nuclear receptor. Moreover, the amino acid sequences of the receptor molecules of oestradiol, corticosteroids and thyroid hormones show high levels of homology suggesting that they all belong to a super family of regulatory proteins that have evolved from a primordial receptor gene. The Cerb-A thyroid hormone receptor contains a domain which is responsible for its binding to DNA; this domain has a sequence very similar to those of the corresponding regions of the oestrogen and corticosteroids receptors (52 and 47% homology, respectively). The sequences of the carboxy terminal region of the three receptors, which contain the binding sites for the hormone, also show homology (17%). Another important finding is that there exists three chromosomally linked c-erb-A genes implying the existence of one or more other molecules closely related to the thyroid hormone receptor. In contrast the v-erb-A protein does not bind T, and seems therefore to be a constitutively active form of the thyroid hormone receptor.

4. Induction and repression of pituitary hormones 4.1. Growth hormone

Several important features of thyroid hormone action have been obtained by using pituitary cell lines (GH,, GHI. GC) derived from the somatotrophs, i.e., the cells which synthesize and secrete growth hormone [43,46]. When cultured in thyroid hormone depleted media such cells are responsive to thyroid hormone. Tsai and Samuels [47] first demonstrated that GH, cells synthesize 3-10-times more GH when cultured in the presence of physiological concentration of T3 (0.1-1 nM). A detectable increase in G H synthesis occurs 45-60 min after significant binding to nuclear sites. Such an increase is paralleled by changes in total cytoplasmic G H mRNA

67 levels [44-46,4&56] and is synergistically potentiated by glucocorticoids. T, and glucocorticoid induce or repress synergistically or independently a small number of other proteins in these cells [54]. The free hormone concentration required for half maximal stimulation of G H ~ 0.19 nM for its acetic derivative TRIAC [51]. For synthesis is 0.17 nM for L - Tand D-T, and L - T the ~ half maximal responses were obtained at much higher concentrations (4 and 9 nM, respectively). The same range of relative affinities for the nuclear receptor were obtained with the same thyroid hormone analogues suggesting a good correlation between hormone binding to the nuclear receptor and the G H response. With the exception of D-T,, which probably enters the cell less easily than the other analogues, the relative receptor affinities of the different iodothyronines was identical using intact cells or isolated nuclei. Several groups have demonstrated both by cell free translation of poly A' RNA extracted from induced and non-induced cells and by molecular hybridization techniques that thyroid hormones increase the level of G H mRNA. Nyborg et al. [56] have, for instance, measured the rate of G H gene transcription by elongation of in vivo initiated RNA chains in nuclei isolated from G C cells. They have found that 1) the rate of G H gene transcription is proportional to T, receptor occupancy during the initial phase of hormone induction, 2) the gene response to receptor occupancy is very fast suggesting that the unoccupied receptor resides close to the gene regulatory site, and 3) dexamethasone potentiates T, action on both G H gene transcription and G H mRNA level. Glucocorticoid alone has little or no effect on G H synthesis and mRNA acccumulation, whereas in the presence of T, two peaks of enhanced transcription were produced. This suggests a direct interaction of the occupied receptors at their regulatory sites. Finally comparison of the rates of decay of G H mRNA suggest that G H mRNA is much more stable when T, is present. Other data also suggest that glucocorticoids enhance G H mRNA accumulation both at the transcriptional and post transcriptional levels. Recently published results [57,58] led to the identification of the sequences of the GH promotor region which are required for responsiveness to T,. Chimeric genes containing different GH promotor sequences were linked to bacterial genes and then transfected to G C cells. Such studies allowed the identification of a fragment of the G H promotor gene (-235 + l l ) which was able to direct T, responsive expression of the bacterial gene by the transfected cell. Such regulation seems to be mediated by factors present only in pituitary cells since other cell types, when transfected by the same promotor containing constructions, failed to respond to T, although they contained the nuclear receptor. Other recent results also suggest that the interaction of the receptor-T, complex with the chromatin alters the local chromatin structure [59]. This was shown by treating the chromatin prepared from cells exposed to T, with DNAse I: three hypersensitive sites were identified in the region spanning the transcriptional initiation site from -200 to +150. The first site was in the first intron of the gene. The

68 second site was located in a region of 5' flanking DNA which promotes T-regulated transcriptional initiation. The third site was centered between the two other sites and was located at the position of the TATA sequence. It is possible that such changes in local chromatin structure produced by the occupied T, receptor are related to the mechanism by which T, induces G H gene activation. In contrast the same authors have shown that dexamethasone had no discernible effects on the chromatin structure or flanking DNA. It might be interesting to know whether the potentiation by dexamethasone of the T, effects on G H expression depends or not on such changes in chromatin structure induced by T,.

4.2. Thyrotropin Circulating levels of thyroid hormones regulate both the pituitary and blood levels of TSH. This provides a tight control of the hormogenic activity of the thyroid gland. Thyroid hormone excess in blood has been shown to decrease the levels of TSH both in the pituitary and in the blood [60,61]. TSH consists of two subunits ( a and p) encoded by separate genes on different chromosomes [62]. T, decreases the level of the mRNAs encoding both subunits [63,64]. Recently it has been shown [65] that isolated nuclei prepared from hypothyroid mice bearing TtT97 thyrotropic tumors previously injected with T, synthesize a and p mRNA sequences. These data suggested that T, regulates TSH and gene transcription and that the transcription of the p gene is affected to a greater extent than that of the a subunit gene. It is interesting in the these respects that the a subunit is common to TSH, FSH (folliclestimulating hormone) and LH (luteinizing hormones), whereas the p subunits are unique for each of these hormones and confer biological specificity.

5. Regulation of lipogenesis in the liver The synthesis of several lipogenic enzymes is stimulated in the liver by thyroid hormones. For instance the concentration of acetyl CoA carboxylase, fatty acid synthetase and malic enzyme is increased in vivo after T, injection [66-691.

5.1. Malic e n z y m e Malic enzyme catalyses the NADP-dependent oxidative decarboxylation of malate to pyruvate and C 0 2 with the production of NADP which is utilized for the synthesis of long chain saturated fatty acids from malonyl CoA. Upon thyroid hormone stimulation the activity of malic enzyme is stimulated over 20-fold and this stimulation is paralleled by an increase in the mRNA level [70-721. The rate of malic enzyme mRNA accumulation is stimulated by T, both in the rat liver and the heart (ll-16-fold in the liver, 3-4-fold in the heart) [73]. Nuclear

.

69

MaIic Enzyme

T3

-

I

M a l i c E n z y m e m-RNA

Reg u I a t o r y

Transcription

Peptide

M a l i c Enzyme m-RNA Level

Glucagon M a l i c Enzyme

Fig. 3 . Regulation of malic enzyme mRNA levels in the liver by thyroid hormone. T, both increases malic enzyme gene transcription and malic enzyme mRNA levels (stabilization). Malic enzyme mRNA levels are also reduced by glucagon and increased by insulin at a post transcriptional level.

run off experiments demonstrated that the rate of transcription of the malic enzyme gene is stimulated to similar extent in liver and heart (3-4-fold). The additional increase in cellular malic mRNA (Fig. 3 ) produced by T, in the liver seems to be due to a tissue-specific change either in the rate of degradation of cytoplasmic mRNA or of its processing in the nucleus [73]. A diet high in carbohydrates also stimulates malic enzyme activity and synthesis in rodent and avian liver while starvation or low carbohydrates has the opposite effect [67,74]. Recent data demonstrated [75]that a high-carbohydrate diet increases the cytoplasmic malic enzyme mRNA at a post-transcriptional level probably by retarding its degradation. Such a control is liver specific since no response was observed in brain, heart, kidney and other non-hepatic tissues. The amplitude of the response to the high carbohydrate diet is increased several fold by T,. Goodridge and co-workers (see Ref. 76) have developed a system of chick embryo hepatocytes which, when cultured in defined medium, respond to T,, insulin and glucagon. Low concentrations of T, (Ks0 4 ~ 1 0 ~M)" increase by 15-fold the malic enzyme level and 7-fold the concentration of its mRNA [77]. Insulin alone had no effect both on the enzyme and the mRNA levels, whereas in combination with T, it caused an 11-fold increase in malic enzyme mRNA levels. Glucagon almost completely abolished the stimulatory effect caused by insulin + T,. Experiments performed with puromycin showed that this inhibitor of protein translation blocks the accumulation of malic enzyme mRNA stimulated by T, suggesting that most of the T, effect on malic enzyme takes place at a post-transcriptional step. Glucagon had no effect on transcription but caused malic enzyme mRNA to decay

70 with a much shorter half-life (1 .S h) than in the presence of amanitin or actinomycin (8-1 1 h). The effect of glucagon is therefore entirely post-transcriptional. From these data the authors’ proposals are that T, primarily regulates production of a peptide that stabilizes malic enzyme mRNA and that glucagon inhibits the activity of the same peptide (Fig. 4).

5.2. Fatty acid synthase Another major lipogenic enzyme, fatty acid synthase, is also regulated in the liver by nutritional status, insulin, glucagon and T,. Wilson et al. [78] have found that stimulation of fatty acid synthase requires both thyroid hormones and insulin (40fold stimulation), whereas T, or insulin alone had much smaller effects (2.S.-fold). Experiments performed in the presence or the absence of puromycin suggest that a common T,-induced peptide intermediate regulates the level of both fatty acid synthase and malic enzyme mRNAs.

6. Effects of thyroid hormone o n the receptor-adenylate cyclase system in the adipocyte and the hepatocyte The effect of altered thyroid state on the activity of the p-adrenergic receptor-adenylate cyclase complex have been documented in a number of cell types and tissues such as turkey erytrhrocytes, marrow cells, salivary glands, pancreatic cells, isolated hepatocytes, cardiac membranes and adipocytes. However changes in padrenergic responsiveness do not occur in all cell types. In fat cells epinephrine stimulation of cyclic AMP accumulation and lipolysis is markedly reduced in hypothyroidism but enhanced in hyperthyroidism (see Ref. 79). Similar effects of altered thyroid status on the response to two other lipolytic hormones, ACTH and glucagon, have been reported suggesting that thyroid hormones regulate similarly either the different receptors of the various lipolytic hormones and/or a common step of the lipolytic pathway [80]. Several hypotheses have been tested to account for the altered adipocyte sensitivity of the adipocyte in hypothyroidism: 1) decrease in the p-adrenergic receptors [81]; 2) altered coupling of agonist-receptor binding to activation of adenylate cyclase [82]; 3) increased phosphodiesterase activity [80,82,83]; 4) enhanced sensitivity to adenosine [84]. Whatever the site of action of thyroid hormones it is clear however that cyclic AMP does not accumulate in fat cells from hypothyroid young animals [80,83], both because its production is markedly reduced and because its degradation is increased (Fig. 4). The effect on cyclic AMP production has been actually described whatever the stage [80,81,83] of post natal rat development, whereas the enhancement in phosphodiesterase activity was observed only when the adipocytes were prepared from young animals [go].

71 L I pol y t i c Hormones

Receptor- Adenylate Cyclase Complex

it I

?DE 5'AM?

*

C- AM?

T3

Fig. 4. Action of thyroid hormones on cyclic AMP production and degradation in the adipocyte. The response of the adipocyte to different lipolytic hormones ( p catecholamines, ACTH and glucagon) is under thyroid hormone control both at the level of the receptor-adenylate cyclase complex and at the level of the phosphodiesterase. T, also regulates the expression of several key lipogenic enzymes.

These effects of T, on the hormone-dependent lipolysis in the adipocyte seem to be direct. They have been reproduced in a culture system of 3T3-L preadipocytes [85]. When maintained in thyroid hormone-depleted media these cells exhibit lower sensitivity to isoproterenol, which can be explained by alterations in both production and degradation of CAMP. In vitro, T, did not affect receptor number or affinity suggesting that the decrease in cAMP production depends on some unknown alteration of the coupling system. It is worth noting that hypothyroidism also leads to an increased lipogenesis in the adipocyte which is independent of cyclic AMP [86,87]. In other words, the activity of several enzymes of the lipogenic pathway are stimulated in hypothyroidism independently from the effects of T, on the cyclic AMP system. Little is known about the mechanisms by which T, exerts these cyclic AMP-dependent and independent effects (Fig. 4). One may assume that T, increases or decreases the synthesis of some of the enzymes responsible for these pathways by acting on the transcription and/or the stability of their specific mRNAs. In contrast to the situation in the adipocyte, hypothyroidism potentiates @-adrenergic receptor-mediated cAMP and glycogen phosphorylase response in rat hepatocytes [88]. Thyroid hormones suppress @-adrenergic-stimulated phosphorylase b kinase and phosphorylase a activities, while enhancing phosphoprotein phosphatase activity in the same cells [89,90]. In other words, thyroid hormones seem to

72 exert opposite effects on the adenylate cyclase system present in the adipocyte and the hepatocyte, respectively.

7. The muscle cell: P-adrenergic responsiveness and the expression of myosin heavy chains The changes in cardiac activity (see Ref. 91) occurring during hyperthyroidism are similar to those induced by sympathetic nerve stimulation and infusion of epinephrine, and can be ameliorated by adrenergic blockade. Conversely, in hypothyroidism, the alteration in cardiac activity suggests diminished adrenergic responsiveness. Williams and Lefkovitz first reported [92] that cardiac membranes from rats made hyperthyroid demonstrate an increase in Padrenergic receptor number. These observations might explain the increased sensitivity to adrenergic activation of cardiac glycogen phosphorylase which is seen in hyperthyroidism. This would imply that the cyclic AMP level in the heart would be increased in experimental hyperthyroidism and decreased in hypothyroidism but contradictory results have been reported in this respect (see Ref. 91). Recently Exton and co-workers [93] have proposed that adrenergic responsiveness in skeletal muscle is regulated by thyroid hormones at two levels, i.e., 1) stimulation of p-adrenergic receptors and adenylate cyclase activity; and 2) increased activity of phosphoprotein phosphatases. Such results would explain the effect of thyroid hormones on glycogen metabolism in muscle although the primary mechanism of these actions remains unknown. Another effect of thyroid hormones in the heart takes place at the level of the expression of myosin heavy chain (Fig. 5 ) . Myosin is composed of two heavy chains, (200000 Da), two phosphorylable light chains (18 000-20000 Da) and two nonphosphorylable light chains (16000-27 000 Da). In some animal species, three isomyosins (V,, V2, V,) have been resolved which display decreasing ATPase activity [94]. These isomyosins have identical light chains but differ in heavy chains (aor p HMC). Multiple genes encode these heavy chains [95,96]. The expression of the HMC p gene is predominant during late fetal stages and the isomyosin is a pp hornodimer (V,). After birth the synthesis of p HMC is decreased whereas that of a HMC is increased. Consequently the aa homodimer (V,) progressively replaces the pp form (V,). V2 is a heterodimer containing both the a and p isoforms [97,98]. Such changes in the expression of the heavy chain isoforms is subject to complex regulation, one major factor being the thyroid hormone status of the animal [99,100]. For instance: 1) the increase in endogenous serum level of T, and T, which is seen during postnatal development in the rat correlates with the enhanced expression of the a HMC isoform; 2) when postnatal hypothyroidism is established soon after birth p HMC is the only expressed isoform (thyroid hormone administration restores the expression of the adult isoform, a HMC [99,100]); 3) a close correlation has been

HMC genes T,

P

I

-

action

H M C isozyrne

/3

m-RNA

p

HMC

+I (2

I

c rn - R N A

a

HMC

"1

ATPase

low

high

Contracti I ity

I ow

high

Ca"

73

Fig. 5. Regulation of myosin heavy chains by thyroid hormones during development and in adulthood. T, represses the synthesis of the fetal p myosin heavy chain mRNA while inducing the expression of the adult (Y isomyosin mRNA.

shown between the expression of the a and p HMC isoforms and the relative abundance of the corresponding (Y and p mRNAs suggesting that their genes are regulated in an antithetic fashion [loo]; 4) hypothyroidism, when established in adulthood, is associated with a shift from a to p HMC, which can be reversed by administration of thyroid hormones. Such a shift probably has a physiological significance since the actomyosin p/3 complex has a lower Ca2+ATPase activity than the aa complex. Recently Izumo et al. [loo] have reported that the myosin multichain family is composed of six different myosin heavy chains that are all responsive to thyroid hormones. The same myosin heavy chain gene can be regulated differently by thyroid hormones, even in opposite directions, depending on the tissue in which it is expressed. Differential expression and regulation by thyroid hormones have been demonstrated not only in heart muscle cells (atrium and ventricle) but also in several skeletal muscles (soleus, diaphragm, masseter, etc.).

8. Thyroid hormones and brain development It has long been known that thyroid hormone deficiency, when established prior to the critical period of brain development, produces severe and permanent mental retardation both in humans ('cretinism') and in experimental animals (see Refs. 6 and 101). Most of the early studies aimed at describing the behavioural, physiolog-

74 ical and morphological brain defects induced by neonatal hypo- or hyperthyroidism. The conclusions [lo21 were that: 1) all the developmental processes contributing to normal brain maturation seem to be under thyroid hormone control, i.e., cell acquisition, cell migration, proliferation and branching of neuronal processes, synaptogenesis and myelin formation; 2) most of the brain abnormalities observed as a consequence of early hypothyroidism are permanent if early replacement therapy with adequate amounts of thyroid hormones is not performed; 3) hyperthyroidism seems to be as detrimental as hypothyroidism suggesting that a precise hormonal concentration is required to synchronize the different brain developmental events listed above; 4) most of the brain regions (forebrain, cerebellum, hippocampus, etc.) and cell types (neurons, astroglial and oligodendroglial cells) seem to be under thyroid hormone control. Such a heterogeneity in cell types and cell populations which are segregated into several regions undergoing asynchronous development have precluded, until recently, detailed studies on the mechanism of action of thyroid hormones, i.e., on the expression of specific genes involved in the differentiation process of a given type of neural cell. The recent availability of relatively homogenous culture systems of neurons, astrocytes and oligodendrocytes which are able to differentiate in vitro have led to the conclusion that all these cell types are targets for thyroid hormones. 8.I . Neuronal differentiation

The most significant of the abnormalities observed in a hypothyroid brain is a hypoplastic neuropile, i.e., a marked reduction in the number of connections between neurons [102]. This has been observed both in the cerebrum and the cerebellum. For instance a permanent and dramatic reduction in the arborization of the dendritic tree of the Purkinje cell is observed in the hypothyroid cerebellum [103]. The length of the primary dendritic trunk is increased and a deficit in the number, density and branching of the dendritic spines is noticed. In contrast neonatal hyperthyroidism accelerates development of spines. Similar findings have been reported for the cerebrum, i.e., reduction in length and branching of pyramidal neurons, of the density of axonal terminals and of the number of spines [102]. Most of the culture systems of neurons are also responsive to thyroid hormones. For instance cultures of fetal neurons [104,105] develop progressively a dense array of branched neurites and a similar morphological differentiation is stimulated by thyroid hormones [104,105]. Little is known on the mechanism(s) responsible for neurite outgrowth and branching. One of the possible mechanism implies that the marked and specific change in neuronal shape occurring during neurite outgrowth depends on underlying modification in the rate and extent of microtubule assembly. Indirect proofs of these assumptions have been provided by the observation that antimitotic drugs which inhibit microtubule assembly also prevent neurite outgrowth [10&108]. Several observations also favour the hypothesis that microtubule

75 assembly is 1) a key regulator of neurite outgrowth and 2) is under thyroid hormone control. Microtubules are composed with a major protein, tubulin and several microtubule associated proteins (MAPs). MAPs are promotors of microtubule assembly and seem also to be required to stabilize the microtubule lattice (see Ref. 109). Differential expression of tubulin isoforms and MAPs occur during brain development (see Ref. 110). For instance, one of the major MAPs, ‘immature’ T AU protein (48 kDa), is progressively replaced by several adult TAU isoforms (52-70 kDa) [111,112]. Such immature and mature T A U entities differ in their polymerization activity when tested with purified tubulin. This may explain why the microtubules present in immature brain are less stable than those assembled at adulthood [113]. Similarly, microtubule assembly is impaired in hypothyroid rat brain [114]. This was shown by assaying the rate of assembly in crude brain supernatants prepared from young hypothyroid rats. Such a decrease in polymerization activity could be corrected in vivo by early replacement therapy with thyroid hormones and in vitro by adding TAU protein [115]. Other experiments showed that thyroid hormone changes both the number of TAU isoforms and their concentration [loll. These observations are consistent with the finding that MAPs concentration increases during in vitro differentiation both of normal primary neuronal cultures [116] and of PC,, cells cultured in the presence of NGF [117]. In other words MAPs seem to be key regulators of microtubule assembly and of neurite outgrowth. Direct counting of the microtubules present both in the dendrites of the Purkinje cells [118] and in the sciatic nerve [119] have recently confirmed that hypothyroidism markedly decreases microtubule assembly and/or stability. Little is known about the mechanism of action of thyroid hormones on microtubule assembly. It might be that, as in other cell types, thyroid hormones regulate the expression of the specific TAU mRNAs and/or their accumulation. One has also to take into account that a large number of other molecules are required to build up neurites during neuronal differentiation (membranes, components of the growth cone, etc.).

8.2. Glial cell differentiation Another major target of thyroid hormones seems to be the oligodendrocyte. This type of glial cell is responsible for myelination in the central nervous system. In vivo thyroid hormone deficiency both decreases the myelin content and retards its formation [120,121]. These changes in myelination might be secondary to neurite outgrowth. However a direct effect of thyroid hormone on the synthesis of sulfatides (a specific marker of myelin) has been demonstrated in cultures of oligodendrocytes [122]. Little is known about the mechanism of action of thyroid hormones on oligodendrocyte differentiation and myelination but it can be assumed that specific enzymes involved in myelin synthesis are induced by T,. Finally it has also been

76 reported recently that astrocytes in culture contain nuclear receptors for T, but their concentration in this type of glial cell is lower than that present in neurones; little is known about the effects of T, on astrocyte differentiation [123].

9. Conclusions Increasing evidence suggests that the nuclear-receptor mechanism is, at least partly, able to explain a number of effects of thyroid hormones both during development and in adulthood. According to the schema proposed for G H induction and TSH repression, receptor occupancy seems to be linearly related to gene expression. The sequences of the G H gene responsible both for receptor binding and for hormone responsiveness are in the process of being identified although several questions related to the mechanism of gene activation remain largely unsolved. The discovery that the product of the oncogene c-erb-A is a T, receptor will probably help to identify the binding domains of the T,-responsive genes and to understand the mechanism of gene activation by the occupied receptor. A second type of mechanism is illustrated by the malic enzyme and fatty synthetase systems. In such systems the effect of T, on the accumulation of the specific mRNAs does not seem to depend only on direct gene activation by the T,-nuclear receptor. An amplification post-transcriptional mechanism seems to contribute to the accumulation of the specific mRNAs. Stabilization of the mRNA level is also differently modulated by other hormones and by the diet. It is still not clear whether all of the effects of thyroid hormones on specific mRNA accumulation in other target cells follow the linear or amplified mechanisms described for G H and for the malic enzyme. Few examples even suggest the existence of non-nuclear effects of thyroid hormones [17,18], but their precise significance remains largely unknown. What is suggested at the present stage of the research is that 1) more than one mechanism probably leads to the accumulation of specific mRNAs and 2) the response to thyroid hormone is highly specific to the cell type, i.e., probably depends, for each cell type, on the organization of its chromatin and/or on the presence of other factors present in the nucleus or the cytoplasm.

References 1. Tata, J . R . and Widnell, C.C. (1966) Biochem. J. 98, 604-620. 2. Wolff, E.C. and Wolff. J . (1964) In: The Thyroid Gland Vol. I, (Pitt-Rivers, R. and Trotter. W.R. eds.) p. 237. Butterworth, London. 3. Oppenheimer. J . H . (1983) In: Molecular Basis of Thyroid Hormone Action. (Oppenheimer J . H . and Samuels H . H . , eds.) pp. 1-34. Academic Press, New York. 4. Samuels, H.H. (1983) In: Molecular Basis of Thyroid Hormone Action (Oppenheimer J . H . and Samuels H . H . . eds.) pp. 35-65. Academic Press, New York.

77 5 . Ismael-Beigi, F. and Edelman, I.S. (1970) Proc. Natl. Acad. Sci. U.S.A. 67, 1071-1078. Guerney, D.L. and Edelman I.S. (1983) In: Molecular Basis of Thyroid Hormone Action. pp. 293-324. 6 . Nunez, J . (1985) In: Handbook of Neurochemistry. Vol. 8 (Lajtha A , . ed.) pp. 1-28. Plenum Publishing Co., New York. 7. Nunez, J . (1984) In: Methods in Enzymology, Vol. 107 (Colowick, S. and Kaplan, O., eds.) pp. 47G488. Academic Press, New York. 8. Mercken, L., Simons, M.J., Swillens, S., Massaer, M. and Vassart, G . (1985) Nature 316, 647-651. 9. Eckel, J . , Rao, G.S., Rao, M.L. and Breuer, H . (1979) Biochem. J . 182, 473-491. 10. Mol, J . A . , Krenning, E.P., Docter, R.. Rozing, J . and Henneman. G. (1986) J . Biol. Chem. 261, 764G-7643. 11. Sterling, K., Milch, P.O., Lazarus. J . H . , Sakurada, T. and Brenner, M.A. (1978) Science 201, 1126-1129. 12. Sterling, K.(1986) Endocrinology 119, 292-295. 13. Gharbi, J . and Torresani, J . (1979) Biochem. Biophys. Res. Commun. 88, 17G177. 14. Cheng, S.Y., Hasumura, S . , Willingham. M.C. and Pastan, I. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 947-95 1. 15. Pliam, N.B. and Goldfine, 1 . 0 . (1977) Biochem. Biophys. Res. Commun. 79, 166-172. 16. Botta, J.A. and Farfas, R.N. (1985) Biochem. Biophys. Res. Commun. 133, 442-448. 17. Goldfine, I.D., Simons, G.G. and Ingbar, S.H. (1975) Endocrinology 96, 802-805. 18. Segal, J . and Gordon, A. (1977) Endocrinology 101, 150-156. 19. Segal. J. and Ingbar, S.H. (1979) J . Clin. Inv. 63. 507-517. 20. Hamada, S . , Torizuka, K., Mikyake, T. and Fusaze, M. (1970) Biochim. Biophys. Acta 201,479-492. 21. Spaulding, S.W. and Davis, P.J. (1971) Biochim. Biophys. Acta 229, 279-283. 22. Refetoff, S . , Matalon, R. and Bigazzi. M. (1972) Endocrinology 91, 934-947. 23. Davis, P.J., Handwerberg, B.S. and Glaser, F. (1974) J . Biol. Chem. 249, 6208-6217. 24. Dillman, W., Surks, M.I. and Oppenheimer. J.H. (1974) Endocrinology 95, 492-498. 25. Samuels, H.H., Tsai, J.S., Casanova, J . and Stanley, F. (1974) J . Clin. Inv. 54, 853-865. 26. Geel, S.E. (1977) Nature 269, 428-430. 27. Yoshida, K. and Davis, P.J. (1977) Biochem. Biophys. Res. Commun. 78, 697-705. 28. Dozin-Van Roye. B . and De Nayer, P. (1978) FEBS Lett. 96, 152-154. 29. Michelot, J . , Dastugue, B., Defer, N . and Meyniel, G. (1979) Biochem. Biophys. Res. Commun. 88. 1368-1374. 30. Lennon, A.M., Osty, J. and Nunez, J . (1980) Mol. Cell Endocrinol. 18, 201-214. 31. Oppenheimer, J . H . , Koerner. D., Schwartz. H.L. and Surks, M.I. (1972) J . Clin. Endocrinol. Metab. 35, 33C-333. 32. Oppenheimer, J . H . , Schwartz, H.L. and Surks. M.I. (1974) Endocrinology 95, 897-903. 33. Thomopoulos, P.. Dastugue, B. and Defer. N . (1974) Biochem. Biophys. Res. Commun. 58,499-506. 34. Samuels, H . H . , Stanley, F. and Casanova. J . (1979) J . Clin. Invest. 63, 1229-3492. 35. Samuels, H . H . , Stanley, F. and Casanova. J . (1979) J. Clin. Invest. 63, 1229-1240. 36. Samuels, H.H., Tsai, J.S. and Casanova, J . (1974) J . Clin. Invest. 53, 656659. 37. Latham, K.R., Ring, J.C. and Baxter, J.D. (1976) J. Biol. Chem. 251, 7388-7397. 38. Perlman, A.J., Stanley, F. and Samuels, H.H. (1982) J. Biol. Chem. 257, 93G938. 39. Samuels, H.H., Stanley, F., Casanova, J. and Shao, T.C. (1980) J . Biol. Chem. 255, 2499-2508. 40. Casanova, J., Horowitz, Z.D., Copp, R.P., McIntyre. W.R., Pascual, A. and Samuels, H.H. (1984) J. Biol. Chem. 259, 12084-12091. 41. Sap, J., Munoz, A., Damm, K . , Goldberg, Y , Ghysdael, J . , Leutz, A . , Beug, H . and Vennstriim, B. (1986) Nature 324, 635-640. 42. Weinberger, C., Thompson, C.C., Ong, E.S., Leho, R., Gruol, D.J. and Evans, R.M. (1986) Nature 324, 641-646. 43. Tashjian, A . H . , Bancroft, F.C. and Levine. (1970) J. Cell Biol. 47, 61-70.

78 44. Shapiro, L.E., Samuels, H.H. and Yaffe, B.M. (1978) Proc. Natl. Acad. Sci. U.S.A. 75.45-49. 45. Martial, J .A,, Baxter, J.D., Goodman. H.M. and Seehurg, P.H. (1977) Proc. Natl. Acad. Sci. U.S.A. 74, 1816-1820. 46. Seo, H., Vassart, G., Brocas, H . and Refetoff, S. (1977) Proc. Natl. Acad. Sci. U.S.A. 74,2054-2058. 47. Tsai, J.S. and Samuels, H.H. (1974) Biochem. Biophys. Res. Commun. 59, 42C-428. 48. Yaffe, B.M. and Samuels, H.H. (1984) J . Biol. Chem. 259, 6284-6291. 49. Samuels, H . H . and Shapiro, L.E. (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 3369-3373. 50. Samuels, H . H . , Stanley, F. and Shapiro, L.E. (1976) Proc. Natl. Acad. Sci. U.S.A. 73,3877-3881. 51. Samuels, H.H., Stanley, F. and Shapiro. L.E. (1979) Biochemistry 18, 715-721. 52. Spindler, S.R., Mellon, S.H. and Baxter, J.D. (1982) J. Biol. Chem. 257, 11627-11632. 53. Dobner, P.R., Kawasaki, E.S., Yu, L.Y. and Bancroft, F.C. (1981) Proc. Natl. Acad. Sci. U.S.A. 78,2230-2234. 54. Ivaric, R.D., Morris, J.A. and Ebcrhardt, N.L. (1980) Recent Progr. Horm. Rcs. 36, 195-235. 55. Samuels, H . H . , Horwitz, Z.D.. Stanley, F., Casanova, J . and Shapiro, L.E. (1977) Nature 268. 254257. 56. Nyborg, J . K . , Nguyen, A.P. and Spindler, S.R. (1984) J. B i d . Chem. 259, 12377-12381. 57. Larsen, P.R., Harney, J.W. and Moore, D . (1986) J . Biol. Chem. 261, 14373-14376. 58. Crew, M.D. and Spindler, S.R. (1986) J . Biol. Chem. 261, 5018-5022. 59. Nyborg, J.K. and Spindler, S.R. (1986) J . Biol. Chem. 261, 5685-5688. 60. Kourides, I.A., Weintraub, B.D., Ridgway, E.C. and Maloof, F. (1975) J . Clin. Endocrinol. Metab. 40, 872-885. 61. Surks, M.I. and Lifschitz. B.M. (1977) Endocrinology 101, 769-775. 62. Kourides, I.A., Barker, P.E.. Gurr, J.A., Pravtchcva. D.D. and Ruddle, F.H. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 517-519. 63. Gurr, J.A., Cattcrall. J.F. and Kouridcs, T.A. (1983) Proc. Natl. Acad. Sci. U.S.A. 80,2122-2126. 64. Shupnik, M.A., Chin, W.W.. Ross, D.S., Downing, M.F.. Habemer. J.F. and Ridgway, E.C. (1983) J., Biol. Chcm. 258, 1512(k15124. 65. Shupnik, M . A . , Chin, W.W., Habcner, J.F. and Ridgway, E.C. (1985) J . Biol. Chcm. 260, 2900-2903. 66. Tcpperman, H.M. and Teppcrman. J . (1964) Am. J . Physiol. 206, 357-361. 67. Wise, E.M. and Ball, E.G. (1964) Proc. Natl. Acad. Sci. U.S.A. 52, 1255-1263. 68. Goodridge, A.G. and Adelman, T.G. (1976) J. Biol. Chem. 251, 3027-3032. 69. Mariash, C.N.. McSwigan. C.R., Towlc, H.C., Schwartz, H.L. and Oppenhcimer, J.H. (1981) J . Clin. Invest. 68, 1485-1490. 70. Towlc, H . C . , Mariash, C.N., Schwartz, H.L. and Oppenhcimer, J . H . (1981) Biochemistry 19, 579-585: Biochemistry 20, 3486-3492. 71. Magnuson, M.A. and Nikodem, V.M. (1983) J . Biol. Chem. 258, 12712-12717. 72. Magnuson, M.A., Dozin, B. and Nikodem, V.M. (1985) J. Biol. Chem. 260, 5906-5912. 73. Dozin, B., Magnuson, M.A. and Nikodem, V. M. (1986b) J. Biol. Chem. 261, 10290-10292. 74. Goodridge. A.G. (1968) Biochem. J . 108. 667-673. 75. Dozin, B., Rall, J.E. and Nikodem, V.M. (1986a) Proc. Natl. Acad. Sci. U.S.A. 83, 4705-4709. 76. Goodridge. A.G. (1983) In: Molecular Mechanism of Thyroid Hormone Action (Oppenheimer, J.H. and Samuels, H . H . , eds.) pp. 368-383. Academic Press, New York. 77. Back, D.W., Wilson, S.B., Morris, S.M. and Goodridge, A.G. (1986) J . Biol. Chem. 261, 12555-12561. 78. Wilson, S.B., Back, D.W., Morris, S.M., Jr., Swicrczynski, J . and Goodridge, A.G. (1986) J . Biol. Chcm. 261, 15179-15182. 79. Nunez, J. and Correze, C. (1981) In: Advances in Cyclic Nucleotide Research, Vol. 14 (Dumont, J., Greengard, P. and Robinson, G . , eds.) pp. 539-553. Raven Press, New York. 80. Correze, C., Laudat. M.H., Laudat, P. and Nunez, J . (1974) Mol. Cell. Endocrinol. 1, 309-327.

79 81. Malbon, C.C., Moreno. F.J., Cabelli, R.J. and Fain, J.N. (1978) J . Biol. Chem. 253, 671-678. 82. Amstrong, K.J., Stouffer, J.E., Van Inwegen. R.G., Thompson, W.J. and Robison, G . A . (1974) J. Biol. Chem. 249, 42264231. 83. Van Inwegen, R.G., Robison, G.A. Thompson, W.J., Amstrong, K.J. and Stouffer, J . E . (1975) J. Biol. Chem. 250, 2452-2456. 84. Ohisalo, J.J. and Stouffer, J.E. (1979) Biochem. J . 178, 249-251. 85. Elk, M.L. and Manganiello, V.C. (1985) Endocrinology 117, 947-953. 86. Correze, C., Nunez. J. and Gordon. A. (1977) Mol. Cell. Endocrinol. 9, 133-144. 87. Correze, C., Berriche, S . , Tamayo, L. and Nunez, J. (1982) Eur. J. Biochem. 122, 387-392. 88. Malbon, C.C. (1980) J. Biol. Chem. 255, 8692-8699. 89. Malbon, C.C. and Lo Presti, J.J. ( I Y 8 1 ) J . Biol. Chem. 256, 12199-12204. 90. Malbon, C.C. and Campbell, R. (1982) Endocrinology 111, 1791-1796. 91. Williams, R.S. and Lefkowitz, R.J. (19x3) In: Molecular Basis of Thyroid Hormone Action (Oppenheimer, J.H. and Samuels, H . H . , eds.) pp. 325-349. Academic Press, New York. 92. Williams, L.T. and Lefkowitz, R.J. (1977) J . Biol. Chem. 252, 2787-2789. 93. Chu, D.T.W., Skikama. H., Khatra, R.S. and Exton, J . H . (1985) J . Biol. Chem. 260,999410000. 94. Lompre, A.M., Mercadier, J.J., Wisnewsky. C . , Bouveret, P., Pantaloni, C., D’Albis, A. and Schwartz, K. (1981) Dev. Biol. 84, 286290. 95. Mahdavi, V . , Periasamy, M. and Nadal-Ginard. B. (1982) Nature 297, 659-664. 96. Sinha, A.M., Umeda, P.K., Kavinsky, C.J., Rajamanickam, C . , Hsu, H.J., Jacovcic, S. and Rabinovitz, M. (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 5847-5851. 97. Whalen, R.G., Sell, S.M., Butler-Brownc, G.S., Schwartz, K., Bouveret, P. and Pinset-H2rstrom., I. (1981) Nature 292. 805-809. 98. Chizzonite, R.A. and Zak, R. (1984) J . Biol. Chem. 259, 12628-12632. 99. LomprC. A.M., Nadal-Ginard, B. and Mahdavi. V. (1984) J. Biol. Chem. 259, 6437-6446. 100. Izumo, S., Nadal-Ginard, B. and Mahdnvi. V. (1986) Science 231. 597-600. 101. Nunez, J. (1985) Neurochem. Int. 7, 959-968. 102. Eayrs, J.T. (1960) Br. Med. Bull. 16, 122-127. 103. Legrand, J . (1967) Arch. Anat. Microsc. Morphol. Exp. 56, 205-244. 104. Romijn, A.M., Habets, M.C., Mud, M.T. and Wolters, P.S. (1982) Brain Res. 2, 583-589. 105. Puymirat, J., Loudes, C., Faivre-Bauman. A.. Tixier-Vidal, A. and Bourre, J.M. (1982) Cold Spring Harbor Conferences on Cell Proliferation 9, 1033-1051. 106. Seeds, N. W., Gilman, A.G., Ammo. 7‘. and Niremberg, M.N. (1970) Proc. Natl. Acad. Sci. U S A . 66, 160-167. 107. Yamada. K.M., Spooner, B.S. and Wessels, M.K. (1970) Proc. Natl. Acad. Sci. U.S.A. 66, 1 2 0 6 1212. 108. Daniels, M.P. (1972) J . Cell Biol. 53. 164176. 109. Dustin. P. (1984) In: Microtubules. 2nd Edn.. Springer-Verlag. Berlin. 110. Nunez, J . (1986) Dev. Neurosci. 8. 125-141. 111. Mareck, A., Fellous, A., Francon. J . and Nunez. J. (1980) Nature 284, 353-355. 112. Francon, J . , Lennon, A.M., Fellous. A . . Mareck, A., Pierre, M. and Nunez, J. (1982) Eur. J . Biochem. 129, 465-471. 113. Fellous, A . , Lennon, A . M . , Francon. J . and Nunez, J. (1979) Eur. J. Biochem. 101, 365-376. 114. Francon, J . , Fellous, A., Lennon, A.M. and Nunez, J. (1977) Nature 266. 188-190. 115. Lennon, A.M., Francon, J . , Fellous. A . and Nunez, J. (1980) J . Neurochem. 35, 804813, 116. Couchie, D . , Faivre-Bauman, A . , Puimyrat, J . , Guilleminot, J.. Tixier-Vidal, A . and Nunez, J . (1986) J . Neurochem. 47, 1255-1261. 117. Drubin, D . G . , Feinstein, S.C., Shooter, E.M. and Kirshchner, M.W. (1985) J . Cell Biol. 101, 1799-1807. 118. Faivre, C . , Legrand, C.H. and Rabie, A. (1983) Dev. Brain. Res. 8, 21-30.

80 119. Marc, C. and Rabie, A. (1985) Intl. J . Dev. Neurosci. 3, 353-358. 120. Hamburgh. M . , Mendoza, L.A., Burkart. J.F. and Weil. F. (1971) In: Cellular Aspects of Neuronal Growth and Differentiation (D.C. Please, ed.) pp. 321-328. University of California Press, Berkeley. 121. Clos, J.. Rebiere. A . and Legrand. J . (1973) Brain Res. 63, 445-449. 122. Bhat. N.R., Rao, G.S. and Pieringer, R . A . (1981) J. Biol. Chem. 256, 1167-1171. 123. Pascual, A . , Aranda, A , , Ferret-Sena, V . . Gabellec, M.M., Rebel, G . and Sarlieve, L.L. (1986) Dev. Neurosci. 8. 89-101.

B.A. Cooke, R.J.B. King and H.J. van der Mole11 (cds.) Hormones und their Actions, Port I 01988 Elsevier Science Publishers BV (Biomedical Division)

81 CHAPTER 6

Metabolism of thyroid hormone THE0 J . VISSER Department of Internal Medicine I l l and Clinical Endocrinology, Erasmus University Medical School, Rotterdam, The Netherlands

1. Metabolic pathways of thyroid hormone 1.1. Introduction

The follicular cells of the thyroid gland produce predominantly thyroxine (3,3’ ,5 3’tetraiodothyronine, T4). This is accomplished by iodination of tyrosyl residues of thyroglobulin (to monoiodotyrosine, MIT, and diiodotyrosine, DIT) and subsequent coupling of two DIT residues [l].In thyroglobulin of normal iodine content MIT is more abundant than DIT. Nevertheless, production of 3,3’ ,5-triiodothyronine (T,) and even more so of 3,3’,5‘-triiodothyronine (reverse T,, rT,) by coupling of MIT and DIT is quantitatively less important. In healthy humans, thyroidal secretion comprises on average 115 nmolid T,, approximately 9 nmolid T, and only about 2 nmoliday rT, [2,3].Formation of 3,3’-diiodothyronine (3,3’-T2) by coupling of two MITs also appears negligible. Thyroid hormone is essential for the regulation of various metabolic processes and the energy consumption of the organism [4].This hormonal action is exerted primarily through interaction with nuclear receptors controlling the transcription of thyroid hormone-responsive genes. T, demonstrates a 10- to 100-fold higher affinity for these receptors than T, while binding of rT, is negligible. T, is considered to have little intrinsic bioactivity and the biological effects elicited in vivo are largely Correspondence to: Prof. Theo J . Visser, Department of Internal Medicine 111, Erasmus University Medical School, P.O. Box 1738. 3000 D R Rotterdam, The Netherlands. Abbreviations: BAT, brown adipose tissuc; BrAc, N-bromoacetyl-; BSA, bovine serum albumin; CNS, central nervous system; D E P , diethylpyrocarbonatc; DIT. diiodotyrosine; DTT. dithiothreitol; G , glucuronide; Grx, glutaredoxin; GSH, reduced glutathione; GSSG. oxidized glutathione; IRD. inner ring deiodination; MIT, monoiodotyrosine; O R D , outer ring deiodination; PTU, propylthiouracil; S , sul(reverse T?);Trx, thioredoxin; T,, diiodothyronine; T,. 3,3’,5-triiofate; rT,, 3,3’,5’-triiodothyronine dothyronine ; T,, 3,3‘,5,5’-tetraiodothyronine (thyroxine).

82 deiodination

HO 0

conjugation

0

ether bond cleavage

0 CH2-CH-COOH

oxidative deamination

Fig. 1. Metabolism of thyroxine

owing to its conversion to T,. The potential activity of T, is lost by its conversion to rT,. Thyroid hormone undergoes in principle four different metabolic reactions (Fig. 1). Deiodination is the most important pathway not only in quantitative terms but also because it can both activate and inactivate the hormone [5-81. Conjugation of the phenolic hydroxyl group serves a dual purpose: sulfation facilitates the deiodinative breakdown especially of T, in the liver [7,8], and glucuronidation is the first step in the enterohepatic cycle of T, and T, [9-111. The latter compounds are also metabolized to a minor extent by oxidative deamination of the alanine side-chain, leading to the formation of the corresponding iodothyroacetic acid derivatives [ll]. Finally, iodothyronines may be degraded by ether bond cleavage. Although normally this is the least significant pathway, it may be induced in conditions of increased peroxidative activity of the tissues, especially in stimulated macrophages [ll].This is a possible explanation for the increased T, turnover in bacterial infections [ll]. Thyroid hormone activity is determined predominantly by intracellular concentrations of (free) T, in the tissues. This T, bioavailability depends on 1) the secretion of T, and T, by the thyroid, 2) the conversion of T, to T, outside the thyroid, 3) the metabolic clearance of T, and 4) exchange of T, and T3 between plasma and tissues. It is the purpose of this chapter to review recent advancements in the study of the transport and metabolism of thyroid hormone. Especially the role of deiodination, conjugation and tissue uptake mechanisms will be emphasized in the regulation of thyroid hormone action. 1.2. Deiodination

The plasma appearance rate of both T, and rT, in healthy humans amounts to approximately 40-45 nmol/day, which is roughly 5 (T,) and 20 (rT,) times the thyroidal secretion rates of these compounds [2,3]. Consequently, 80% of circulating

83 I

I

I

I

I

I

I

I

3, 3’-T

Fig. 2. Stepwise deiodination of thyroxine

T, and 95% of rT, are derived from conversion of T, in peripheral tissues (Fig. 2). T, is produced by monodeiodination of the phenolic ring of T,, a process which is also called 5’-deiodination or outer ring deiodination (ORD). Reverse T, is generated by monodeiodination of the tyrosyl ring of T,, also termed 5-deiodination or inner ring deiodination (IRD). A further metabolite in the stepwise deiodination of thyroid hormone is 3,3’-T2, which is produced by IRD of T, as well as by O R D of rT,. Although conversion to 3,3’-T2 is only one of the possible pathways for the metabolism of T,, it is the main route for the clearance of rT, [5-81. Little is known about the possible occurrence of O R D of T, to 3,5-T2 and IRD of rT, to 3’,5’-T2 but these are probably minor reactions. An important problem associated with in vivo studies of the metabolism of iodothyronines is that products that are generated in the tissues may be further degraded before release into the circulation and, therefore, escape detection. In rats equilibrated with radioiodine-labelled T, or T, roughly half of the radioactivity appears as I- in the urine and the other half as free iodothyronines in the feces [12]. Treatment of the rats with 6-propyl-2-thiouracil (PTU) results in a marked decrease in urinary radioactivity and a reciprocal increase in fecal clearance [ 121. Also, in humans, PTU has been shown to inhibit peripheral iodothyronine deiodination besides its well-known effect on thyroid hormone biosynthesis [13]. Compared with the rat, deiodination is an even more important pathway for the clearance of thyroid hormone in man as evidenced by the greater proportion undergoing urinary clearance [2]. Furthermore, estimation of iodothyronine turnover kinetics in humans has demonstrated that a major fraction of T, disposal is accounted for by plasma production rates of T, and rT, [2,3].

84 TABLE I Three types of iodothyroninc deiodinaaes I

Location Substrate preference Thiols PTU Hypothyroidism Hyperthyroidism

liver, kidney rT, >> T, > T, stimulation inhibition decrease increase

brain, BAT. pituitary T, > rT, stimulation no effect increase decrease

brain, skin, placenta

T, > T, stimulation no effect decrease increase

Both chemically and physiologically O R D and IRD are distinct processes. Because of the vicinal hydroxyl group, the I substituents in the outer ring are readily removed by electrophilic substitution reactions in contrast to the stable bonds of the I atoms with the inner ring. As to the biological implications, O R D is regarded as an activation step that converts the T, to bioactive T3, whereas IRD is a catabolic reaction that not only inactivates T, but also prevents its formation from T,. Recent investigations of the metabolism of iodothyronines in different tissues especially of the rat have led to the recognition of at least three different iodothyronine-deiodinating enzymes [5-81 (Table I). These deiodinases have in common that they are located in the membrane fractions of the tissues and that they are stimulated by sulfhydryl (SH) compounds, especially dithiols [5-81. However, important differences exist between the specificities and catalytic mechanisms of these enzymes, their tissue distribution, sensitivity to PTU and other inhibitors, and regulation by thyroid hormone [5-81. The characteristics of the different deiodinases will be discussed in more detail in Sections 2 and 3 . 1.3. Conjugatiori

Conjugation is a phase I1 reaction which transforms lipophilic endogenous and foreign compounds into water-soluble derivatives to facilitate their excretion in bile and urine [ 141. Glucuronidation is performed by a group of related enzymes located in the endoplasmic reticulum of the liver but also in other tissues. These enzymes use UDP-glucuronic acid as cofactor for the conjugation of hydroxyl and other functional groups in different classes of compounds [ 141. The identity of the UDPglucuronyl transferase for thyroid hormone has not been established but the enzyme is probably shared with simple aromatic substrates such as p-nitrophenol [15]. The preference of this enzyme for the glucuronidation of different iodothyronines is also unknown.

85 Different phenol sulfotransferases have been identified in the soluble fraction of especially the liver that use 3’-phosphoadenosine-5’-phosphosulfate as sulfate donor [16]. Examination of the substrate preferences of partially purified phenol sulfotransferases has indicated higher rates of sulfation of 3,3’-T2 than of T3 and negligible sulfation of T, and rT, [17]. T, and T, are excreted in bile of normal rats predominantly as the glucuronide conjugates [9,18]. Biliary excretion of sulfate conjugates, however, is greatly enhanced if hepatic deiodinase activity is inhibited [9,18]. Also, in vitro studies with isolated rat hepatocytes indicate that T, is metabolized to similar extents by glucuronidation and sulfation [19]. However, while T, glucuronide (T,G) is a stable conjugate that accumulates in the culture medium, the sulfate conjugates of T, (T,S) and 3,3’-T2 (T,S) are rapidly degraded by enzymatic deiodination [19,20] (see also Section 2.1). Little is known about the conjugation of thyroid hormone in man other than the identification of T, glucuronide (TAG)in human bile [21]. A greater proportion of administered T, is excreted in bile of dogs as T3S and T,S than is the case in rats [9]. That conjugation of thyroid hormone is not restricted to the liver is evidenced by the excretion of large amounts of iodothyronine conjugates, especially sulfates, in the urine of hepatectomized dogs [Y]. Strains of obligately anaerobic bacteria in rat (and human) intestinal microflora have been shown to produce iodothyronine glucuronidase and sulfatase activities [22,23]. This explains why T, and T, appear in feces of rats as free iodothyronines [24] although they are excreted in bile mainly as glucuronide conjugates. Recent studies indicate that a large fraction of T,G introduced in the rat intestinal tract is resorbed as free T, liberated by bacterial hydrolysis [25,26]. In rats, therefore, glucuronidation does not seem an irreversible pathway for the elimination of T, and T, but an essential step in the enterohepatic cycle of these compounds.

2. Type I iodothyronine deiodinase of liver and kidney 2.1. Properties and distribution Type I iodothyronine deiodinase is defined as the enzyme which catalyses the (m0no)deiodination of the inner or the outer ring of different iodothyronines and which is inhibited by p M concentrations of PTU [5-81. In rats and humans, such enzyme activities are present at high levels in liver, kidneys and interestingly also in thyroid, and at low levels in many other tissues [5-81. The deiodinase is associated with the microsomal fractions of these tissues but is only active in the presence of a cytoplasmic cofactor [5-81. Also, in the absence of cytosol, deiodinase activity is stimulated by simple thiols such as dithiothreitol (DTT) but the physiological cofactor has not yet been identified with certainty. The effects of synthetic and natural thiols will be discussed in Section 2.5.

86 Type I iodothyronine deiodinase is an integral membrane protein that in liver appears to be associated with the endoplasmic reticulum while in kidney it is located in the plasma membranes [S-8]. The enzyme has an apparent molecular mass of about SO000 Da [7,8] and probably consists of two dissimilar subunits [27,28]. The delipidated enzyme is a basic protein with a p l value of 9.3 [29]. Active enzyme may be solubilized using different ionic and nonionic detergents [30].Soluble deiodinase activity has been purified from rat liver microsomes over 2000-fold through a series of chromatographic steps [311. Although purification was still incomplete, the equal enrichment of inner ring and outer ring deiodinase activities was remarkable. 2.2. Substrate specificity The kinetic parameters of the deiodination of different iodothyronines and their sulfate conjugates by rat liver microsomes as determined in this laboratory are summarized in Table 11. Although these reactions have been carried out in the presence of DTT, the V,,,,,/K, ratio is thought to be a cofactor-independent indicator of the efficiency of the catalytic process [8] (see also Section 2.4). Reverse T, is the preferred sustrate for the type I deiodinase; the efficacy of its O R D to 3,3’-T2 is at least 500-fold higher than the deiodination of T, and T, [32]. Reverse T, is converted quantitatively to 3,3’-T,, and no evidence exists for the hepatic production of 3’,S’-T2 [32,33]. T, undergoes either O R D to T, or IRD to rT, but little of the latter is recovered as it is rapidly further degraded to 3,3’-T, [32]. The kinetics of T, I R D have, therefore, been estimated by summation of the rT, and 3,3’-T2productions [32]. Production of 3,3’-T, by IRD of T, is a relatively slow process while O R D of T, to 3,5-T, has not been observed in liver. A recently recognized property of the type I deiodinase is its particular activity towards sulfated iodothyronine substrates [ 19,20,34]. This was first discovered in

TABLE I1 Substrate specificity of rat liver type I iodothyronine deiodinase

T, T, sulfate

Outer ring

T4

Inner ring

T, sulfate

2.3 ND 1.9 0.3

30 ND

13 ND

18 526

9 2020

rT, rT, sulfate

Outer ring

0.06 0.06

559 516

8730 8600

T.3 T, Sulfate

Inner ring

6.2 4.6

36 1050

6 230

Kinetic parameters were determined using liver microsomes from euthyroid rats in 0.1 M phosphate (pH 7.2). 2 mM E D T A and 3-5 mM DTT. K,,, is expressed in WM and V,,,, in pmolirnin per mg protein. N D , not detectable. Data are taken from Refs. 20. 32 and 34.

87 the metabolism of 3,3'-T2 and subsequently that of T, by cultured rat hepatocytes [19]. Little T2S and T,S formation was seen in these incubations unless type I deiodinase activity was inhibited with PTU. Tests with synthetic T,S and T2S demonstrated that the IRD of T, and subsequent O R D of 3,3'-T, by microsomes are greatly accelerated by sulfation without changing the precise order of these reactions [ 19,201. However, the specificity of the deiodination pattern of T4 is changed dramatically after sulfate conjugation [34]. In contrast to free T,, T,S is deiodinated rapidly and selectively in the inner ring to rT,S, whereas deiodination of the outer ring of T4S does not occur. Deiodination of rT, is not affected by sulfation perhaps because it is already an optimal substrate in the non-conjugated form [34]. The deiodinative degradation of T4 and T, induced by sulfation may represent a mechanism for the irreversible inactivation of thyroid hormone that allows re-utilization of the iodine. Sulfation appears to be an important pathway for the metabolism of T, in rats [9,18,35]. Although T,S has also been identified in bile of PTU-treated rats (M. Rutgers and T.J. Visser, unpublished work), the extent to which T4 is sulfated in vivo remains to be elucidated.

2.3. Inhibitors and affinity labels Type I deiodination of iodothyronines is inhibited competitively by a wide variety of aromatic substances, especially those with halogen substituents in the ortho position of hydroxyl and amino groups. The following classes of compounds are distinguished. 1) Iodothyronines acting as competitive substrates, where it has been demonstrated that a) inhibition is independent of whether the analogue undergoes the same or the alternative reaction (IRD or O R D ) and b ) the K, value of the analogue is identical to its K, as a substrate [7,8]. 2) Simple iodinated phenol and aniline derivatives, among which are the highly potent 2,4,6-triiodophenol ( K , 0.03 pM) and different X-ray contrast agents such as iopanoic acid [36]. 3) Various halogenated derivatives of phenolphthalein and fluorescein such as bromophenol blue ( K , 0.04 p M ) [37], erythrosin [38] and rose bengal ( K , 0.06 pM) [39], of which the latter not only act as true competitive inhibitors but, when irradiated with invisible light, also induce the photochemical inactivation of enzyme [39] (see also below). 4) Substances isolated from plants including flavones, aurones and chalcones [40], and also coumarin derivatives [41] with K , values in the pM range. The type I deiodinase of liver and kidney is inactivated by different SH-selective reagents. In particular, it shows an extremely high susceptibility to carboxymethylation by iodoacetate and a somewhat lesser sensitivity for iodoacetamide and bromoacetate [42]. In comparison, N-alkylmaleimides are only inhibitory at high concentrations (> 0.1 mM) [42]. Enzyme inactivation by iodoacetate follows pseudo

88 first-order rate kinetics and is prevented in the presence of substrate, especially rT,, suggesting the location of an essential SH group in or near the enzyme active center [421. N-Bromoacetyl-T, (BrAcT,) is the most potent inhibitor of the type I deiodinase presently known [27]. It is remarkable that, although BrAcT, is a derivative of a pure IRD substrate, it inhibits the O R D of other iodothyronines as has been shown before for T, itself [43]. BrAcT, is a competitive inhibitor of the O R D of rT, by rat liver microsomes with a K , value of 0.1 nM [27] compared with 10 pM for T, [43]. Pretreatment of microsomes with sub-nM concentrations of BrAcT, induces the rapid and irreversible loss of deiodinase activity. Initial enzyme inactivation obeys pseudo first-order reaction kinetics reaching maximal rates at increasing BrAcT, concentrations. This reaction is characterized by a limiting rate constant (k3)of 0.35 min-', a K iof 0.2 nM and, consequently. a bimolecular rate constant ( k , / K i )of 2 ~ 1 M-'.min-' 0 ~ [27]. In comparison, for iodoacetate these values amount to 1.56 min-', 5 pM and 3x10' M-'.min-' [42]. Substrate provides protection against inactivation of deiodinase by BrAcT,. Analysis of the reaction of microsomal protein with ['"I]BrAcT, indicates the labelling of a subunit of the deiodinase with approximate molecular mass of 25000 Da [27]. Similar findings have also been obtained with ['251]BrAcT, [28]. It is possible that BrAcT, reacts with the same Cys that is also modified by iodoacetate but bromoacetylated compounds have been shown to react with Lys and His residues as well. The presence of an essential His at the catalytic center of the deiodinase has been indicated by studies utilizing diethylpyrocarbonate (DEP) and rose bengal WI. Modification of the type I deiodinase with D E P impairs enzyme activity which is a) characterized by pseudo first-order reaction kinetics, b) prevented in the presence of rT, or iopanoic acid and c) partially reversed by subsequent treatment of the modified enzyme with hydroxyl amine [39]. The latter observation especially, strongly implies the involvement of a His residue. This is further supported by the finding that rose bengal induces the photo-inactivation of the enzyme [39]. It is tempting to speculate that the imidazole group functions in the formation of a hydrogen bond with the essential SH group to increase the nucleophilic character of the latter. The relatively high pK value of 7.5 for the His residue modified by DEP [39] would be in agreement with this view. Thiourea derivatives are known for their anti-thyroid effects due to inhibition of thyroid peroxidase [l]. Two thiourea compounds especially, have found wide application in the treatment of patients with hyperthyroidism, i.e., PTU and 2-mercapto-l-methylimidazole (methimazole). It was soon recognized, however, that while methimazole only blocks thyroid hormone synthesis PTU has an additional effect on thyroid hormone metabolism [ 131. These clinical findings have been confirmed in vitro showing that PTU, but not methimazole, is a potent inhibitor of the type I deiodinase [5-81. Structure-activity studies of thiourea analogues [44,45] have

-

89 imputed the lack of deiodinase inhibition by methimazole to alkylation of N'.The inhibitory effect of thiouracil is strongly augmented by iodination of C5, indicating the use of 5-['2sI]iodothiouracil as a specific affinity-label for the deiodinase [46]. Inhibition of the type I deiodinase by PTU is uncompetitive with substrate and competitive with cofactor. This is the case for the O R D of T4 and rT, as well as for the IRD of T, and T,S [7,8]. Persistent inactivation of enzyme by PTU and covalent labelling with radioactive inhibitor requires the presence of substrate and is only reversed with high DTT [42,47]. All available evidence indicates that PTU reacts with a substrate-induced enzyme intermediate. As thiourea derivatives are particularly reactive towards sulfenyl iodide (SI) groups, generation of an enzyme-SI intermediate is thought to precede thiouracil inhibition through mixed disulfide formation [7,8].

2.4. Reaction mechanism As mentioned above, rat liver cytosol contains one or more factors that stimulate microsomal iodothyronine deiodinase activity. It has been realized for more than a decade now that the enzymatic deiodination of iodothyronines is a reductive process which is supported by different synthetic and natural SH compounds [48]. Most investigations of the catalytic mechanism of the deiodinase have utilized artificial cofactors such as the dithiol DTT. The results have demonstrated that both O R D and IRD follow ping-pong type reaction kinetics, indicating that the enzyme exists in two alternating forms induced by the reactions with substrate and cofactor [7,8]. The current concept of the catalytic mechanism of the type I iodothyronine deiodinase is presented in Fig. 3 . The iodine is removed from the substrate in the form of the iodonium (It) ion and transferred to an enzyme SH group (E-SH). The resultant enzyme SI (E-SI) intermediate represents an oxidized form of the deiodinase from which native enzyme is regenerated by reduction with cofactor. The latter reaction is inhibited by PTU which reacts with E-SI under formation of a stable enzyme-PTU mixed disulfide. Type I deiodination of iodothyronines is not related to the enzymatic deiodina-

HNt

E-S-SN '

R

+I-

Fig. 3. Mechanism of type I deiodination of T, to T,

90 tion of iodotyrosines by microsomal fractions of thyroid and liver [49]. The latter is catalysed by a flavoprotein which, through an unidentified reductase, uses NADPH as a cofactor and which is not inhibited by PTU. The catalytic properties of the type I deiodinase, however, bears some resemblance with those reported for thymidylate synthetase in the dehalogenation of 5-bromo- and S-iodo-2'-deoxyuridylate [SO]. This reaction also appears to involve the active participation of an enzyme SH group and is also stimulated by DTT. Electrophilic displacement of 1' with a proton as has been suggested for the non-enzymatic deiodination of DIT (with Cys as I + acceptor) [Sl] seems a possible mechanism for O R D but not for IRD. Therefore, the exact molecular mechanism of type I deiodination remains to be elucidated.

2.5. Cofactor requirements The dithiols, DTT and dithioerythritol, are particularly potent cofactors for the type I deiodinase, much more so than monothiols such as 2-mercaptoethanol [48,52]. This is explained by the low oxidation-reduction potential of the 1,4-dithiols owing to the close proximity of the SH groups which are oxidized to stable cyclic disulfides. However, the low activity of 1,4-dimercaptobutane [S2] suggests that other factors are involved as well. Interesting is the behavior of dihydrolipoamide which is as strong a reductant as DTT but gives the same stimulation of deiodinase at 10-fold lower concentrations [S2]. The minute concentrations of dihydrolipoamide in the cytosol, however, excludes its role as a physiological cofactor. Reduced glutathione (GSH) is the most abundant thiol in liver and other tissues with intracellular concentrations usually exceeding 5 mM. Combined with findings of parallel changes in GSH and deiodinase-stimulating activity in liver cytosol of fasted rats [53,S4], this has led to the proposal that GSH is the physiological cofactor of the type I enzyme. Furthermore, the fasting-induced decrease in deiodinasesupporting activity was found to be restored with exogenous thiols or by addition of NADPH [53,54]. Presumably, the latter is mediated through an increased GSH generation by glutathione reductase. However, the role of GSH has become questionable in the light of more recent findings that show a dissociation between GSH levels and deiodinase activities in liver homogenates from rats fed with different diets [55]. Such discrepancies have also been observed in studies with fasted-refed [56] and diabetic rats [57]which emphasized the importance of the insulin/glucagon ratio for the regulation of hepatic iodothyronine deiodinase activity. Studies by Sat0 and co-workers [58,59] utilizing normal or tumour liver cells in culture have produced evidence that it is not the level of GSH itself but rather the redox state of glutathione which determines the activity of the deiodinase. Deprivation of the cultures of Met and Cys results in a depletion of total glutathione to less than 10% of control without affecting the IRD or OR D of iodothyronines incubated with these cells [%]. If a similar decrease in GSH is induced by oxidation to GSSG with diamide or t-butylhydroperoxide, both IRD and O R D are strongly

diminished [5Y]. Incubation in the absence of glucose potentiates the effects of the oxidative challenge on the GSH/GSSG ratio and deiodination rates. This is thought to be due to a decreased glucose oxidation in the hexose monophosphate shunt, leading to a diminished supply of NADPH for reduction of GSSG by glutathione reductase [59]. Relative to the dithiol DTT but also to other monothiols such as 2-mercaptoethanol, GSH is a poor stimulator of microsomal deiodinase activity even when tested in the presence of NADPH and glutathione reductase [52,60,61]. Deiodinase activity of isolated microsomes is supported to a limited extent by GSH if tested with low (nM) but not high (pM)rT, concentrations or with T, as the substrate. This low potency of GSH has led investigators to explore other physiological cofactors. As mentioned above, the paucity of cytoplasmic dihydrolipoamide makes it an unlikely candidate despite its unsurpassed potency [52]. This is supported by the finding that addition of NADH, the cofactor €or lipoamide hydrogenase, does not stimulate deiodinase activity of kidney homogenates unless supplemented with lipoamide [521. While GSH itself is unable to support deiodination rates greater than -1 pmol/min per mg rat liver microsomal protein, this is greatly enhanced in the presence of glutaredoxin (Grx) [60,61]. This is a polypeptide found in the soluble fraction of many tissues and organisms, that has a molecular mass of 11000 Da and is similar to, if not identical with, thioltransferase and ‘soluble protein factor’ [8,6&62]. It contains a characteristic -Cys-Pro-Tyr-Cys- sequence that forms a disulfide when glutaredoxin is oxidized (Grx-S,) [62]. The latter is reduced by two GSH molecules to the active dithiol (Grx-(SH),) that presumably acts in the same way as DTT to reduce the E-SI intermediate of the deiodinase. Although the stimulation of the type I deiodinase by GSH is greatly enhanced by glutaredoxin, at physiological concentrations of these factors the enzyme operates at only a fraction of the rates achieved with DTT [60,61]. Studies by Sawada et al. [63] and Goswami and Rosenberg [64] have also suggested the possible involvement of thioredoxin (Trx) as a physiological cofactor of the type I deiodinase. This factor resembles glutaredoxin with respect to its ubiquitous distribution, molecular size and the cyclic interconversion of a dithiol-disulfide as the basis of its catalytic activity [62]. The active site of thioredoxin contains, in the dithiol form (Trx-(SH),), the sequence -Cys-Gly-Pro-Cys- which is produced by reduction of the disulfide of oxidized thioredoxin (Trx-S,) by NADPH and the flavoenzyme thioredoxin reductase [62]. In contrast to glutaredoxin, the thioredoxin system only appeared to support the deiodination of low concentrations of rT, and not of T, [64]. On the basis of the different characteristics of the deiodination of nM iodothyronines by liver or kidney microsomes in the presence of D’IT, GSH, glutaredoxin or thioredoxin, Goswami and Rosenberg suggested the existence of multiple ‘low K,’ deiodinases in addition to the type I enzyme [60,64,65]. However, the evidence to support this conclusion should be regarded as inconclusive (see also Section 3.3).

-

92

A

I

glutathione reductase

NADP

I

2GSH

NADPH

B

x xT4

ASSG

NADPH

E-SH E-SI

-

T3

x; x x >: Grx-(SH)*

\/

E-SI

GSSG

w reductase

NADP

NADiH

C

E-SH

PGSH

Grx-S2

Tr;1-S2

E-SH

T,

E-SI

T3

T4

thioredoxin reductase

NADP

Trx-(SH)2

Fig. 4. Possible roles of glutathione, glutaredoxin (Grx) and thioredoxin (Trx) in type I iodothyronine deiodination.

As mentioned above, addition of NADPH stimulates deiodinase activity in liver homogenates from fasted rats [54]. It is now possible to envisage different cascades of enzymatic reactions that transfer the reductive equivalents from NADPH to the deiodinase (Fig. 4). Glutathione reductase catalyses the reduction of GSSG by NADPH to 2-GSH which could serve as a cofactor for the deiodinase either directly (A) or via glutaredoxin (B). NADPH may also generate a third possible cofactor for the deiodinase by reduction of thioredoxin through thioredoxin reductase (C). To what extent these different pathways are involved with the hepatic deiodination of thyroid hormone in vivo remains uncertain, although the glutaredoxin system appears to have the greatest capacity in vitro [60,61]. However, in view of the low rates at which the deiodinase operates in vivo (- 0.01 fmolimin per mg microsomal protein [S]), the contribution of other cofactors is not excluded. Significant stimulation of type I deiodinase activity is observed with GSH only if the rate of E-SI generation is limited, i.e., at low rT, concentrations or with T4 as the substrate which is slowly deiodinated. The low reactivity of GSH in the absence of glutaredoxin may be due to the formation of a stable enzyme-glutathione mixed

93 disulfide (E-SSG) by reaction of GSH with E-SI (Fig. 4). Reduction of E-SSG with a second GSH may be impeded by steric hindrance in the enzyme active site. GSSG is an inhibitor of the type I deiodinase [66] which may also be due to E-SSG formation with the catalytic SH group or another Cys residue. This is a well-known mechanism for the regulation of enzyme activities by GSH/GSSG [67,68]. It has been suggested that the fasting-induced defect in hepatic deiodinase activity is due to a diminished glucose metabolism in the hexose monophosphate shunt with a resultant decrease in the NADPHiNADP and, therefore, GSH/GSSG ratios [53,54]. This leads to an increase in the formation of protein-glutathione mixed disulfides that not only inhibit the deiodinase directly but also the activity of glutaredoxin [62]. Indeed, an increase in protein-glutathione mixed disulfides in livers of fasted rats has been reported [67,68]. However, the above hypothesis of how this is brought about does not appear to be correct as hepatic NADPH levels are not lowered in starvation [69]. An alternative explanation implicates an increased production of peroxide (accompanying glucagon-induced fatty acid oxidation) which is neutralized at the expense of GSH [67,68]. Besides a decreased deiodinase activity, a defect in tissue iodothyronine uptake may be an even more important factor contributing to the changes in thyroid hormone metabolism induced by fasting [70].

3. Iodothyronine deiodinases of other tissues 3.1. T y p e I I iodothyronine deiodinase Type 11 iodothyronine deiodinase is defined as the enzyme which selectively deiodinates the outer ring of iodothyronines and which is not inhibited by p M concentrations of PTU [ M I . Thus, the type I1 deiodinase converts T, to T, but not to rT, and it catalyses the production of 3,3’-T, from rT, but not from T, (Table I). The enzyme has been found in the CNS [71], in the pituitary [72], in brown adipose tissue (BAT) [73] and in placenta [74]. In the CNS, type I1 deiodinase activity has been localized especially in cerebral cortex and cerebellum [75]. Studies with fetal rat brain cell cultures have suggested that the enzyme is associated with neurons [76]which is supported by findings with neuroblastoma cell cultures [77]. However, type I1 deiodinase activity has also been correlated with the number of astrocytes in glial cell cultures from neonatal rat brain [78,79]. In the pituitary, type I1 enzyme activity is higher in the anterior than in the posterior lobe [72].Fractionatior, of isolated pituitary cells have indicated higher enzyme levels in somatotrophs and lactotrophs than in thyrotrophs and gonadotrophs [80]. Active T, to T, conversion has also been observed in growth hormone and prolactin-secreting pituitary tumor cells [811. When assayed in the presence of DTT, deiodinase activity has been found in the membrane fractions of the above-mentioned tissues, specific enzyme activity being

94 TABLE 111 Kinetics of type I1 iodothyronine deiodinase from rat cerebral cortex

1.1 2.8

0.64 0.34

Kinetic parameters were determined using cerebrocortical microsornes from rats thyroidectornized 12 days previously in 0.1 M phosphate (pH 7), 1 mM E D T A and 20 rnM DTT. K , , is expressed in nM and V,,,, in prnolih per mg protein. Data are taken from Ref. 82.

greatest in the microsomes [72-74,82-851. Analysis by discontinuous sucrose density centrifugation has suggested that the enzyme is associated with the rough endoplasmic reticulum of bovine anterior pituitary [84] and with nerve terminal plasma membranes of rat cerebral cortex [85]. In contrast to the type I deiodinase which shows a high preference for rT, over T, as the substrate (Table II), the type I1 enzyme is somewhat more effective in the deiodination of T4 than of rT, (Table 111). Under the conditions tested, the K , value of T, for the type I1 enzyme is three orders of magnitude lower than the K , of T, for the type I deiodinase. The K , of rT, for the type I1 deiodinase is somewhat greater than that of T, and differs less from the K , of rT, for the type I enzyme. The V,,,,, of the conversion of T, to T, by the type I1 enzyme depends on the tissue and the thyroid status of the animal (see below). In cerebral cortex of hypothyroid rats [82] it is roughly one-thousandth of the maximum T, production by the hepatic type I deiodinase of euthyroid animals determined under similar conditions [32]. The V,,,/K, ratio of this reaction is, therefore, similar for the type I1 deiodinase of hypothyroid rat brain and the type I deiodinase of euthyroid rat liver and much greater than that for the hepatic enzyme of hypothyroid rats [86]. In view of the reaction kinetics of the type I1 deiodinase (see below), it is questionable if the V,,,IK, ratios estimated in vitro also apply to physiological conditions with unknown cofactor availability. That the type I1 deiodinase represents a common enzyme for the ORD of T, and rT, is supported by their mutual competitive inhibition with corresponding K , and Ki values [72-74,82,83]. T,, which is not a substrate for the type I1 deiodinase, also does not inhibit the deiodination of T, and rT, in vitro. In addition to competitive substrate inhibition, other mechanisms exist for the regulation of type I1 enzyme activity by thyroid hormone in vivo. Experimental hypothyroidism in rats induces a large increase in type I1 activity in the CNS [71,82], pituitary [72,83,87] and BAT [73] at least in part by prolongation of the half-life of the enzyme [88]. Treatment of hypothyroid rats with T, produces a rapid fall in type I1 deiodinase in CNS and pituitary which appears to be due to an accelerated inactivation of the enzyme [88]. T, seems to act through a post-transcriptional mechanism that perhaps does not involve the classical thyroid hormone receptor [4]. This is supported by the finding that T4 and rT, are even more potent regulators of the type I1 deiodinase [81,8%91],

9.5 although they possess low to negligible affinity for the nuclear receptor [4]. These effects of T, and rT, have been observed in vivo [90,91] and in cell cultures [81,89]. They are not explained by competitive substrate inhibition but may represent inactivation of enzyme during catalysis. However, substrate inhibition is not observed in vitro, and also the mechanism of enzyme regulation by T, remains unexplained. In BAT, T, effects on the type I1 deiodinase are mediated by growth hormone probably via IGF-1 [91]. The enzyme in BAT is primarily under positive control of norepinephrine and is also stimulated by insulin and glucagon [92]. Type I1 deiodinase activity is low in unsupplemented tissue homogenates but is stimulated by DTT [71-74,82,83] and to a lesser extent also by GSH [72]. The DTT concentrations required for maximal enzyme stimulation in the CNS and pituitary seem higher than in BAT and also than those necessary for the type I deiodinase in liver and kidney. Kinetic analysis of the deiodination of varying substrate (T4, rT,) concentrations at different cofactor (DTT) levels have indicated a sequential reaction mechanism for the type I1 deiodinase [73,82,83]. This is very suggestive of the formation of a ternary enzyme-substrate-cofactor complex in the catalytic process [82]. The physiological cofactor of the type I1 deiodinase has not been identified but it has been observed that GSH depletion with diamide or diethylmaleate impairs T, to T, conversion in GH3 pituitary tumor cells [93]. The insensitivity of the type I1 enzyme to PTU seems to exclude the generation of an enzyme SI intermediate as is the case with the type I deiodinase (see Section 2.4). The lack of involvement of a catalytic enzyme SH group in type I1 deiodination is also suggested by the weak effects of iodoacetate [82], a potent inhibitor of the type I deiodinase. It may be speculated that the type I1 enzyme catalyses the transfer of 1' from the substrate directly to the SH group of the cofactor [82]. I n contrast to PTU, iopanoic acid has similar inhibitory effects on the type I and I1 deiodinases [71-73,84,89].

3.2. Type 111 iodothyronine deiodinuse Evidence has accumulated for the existence of a specific deiodinase for the inner ring of iodothyronines which is further distinguished from the type I enzyme because of its insensitivity to sub-mM PTU concentrations. Thus, type 111 iodothyronine deiodinase converts T, to rT, but not to T, and produces 3,3'-T, from T, but not from rT, (Table I). It has been detected in chick embryo heart [94] and liver [95] cells, monkey hepatocarcinoma cells [96], rat CNS [71,75,97], human [98], rat [98] and guinea pig [99] placenta, and rat skin [loo]. With higher enzyme activities in cerebral cortex than in cerebellum, the distribution of the type 111 deiodinase is different from that of the type I1 enzyme [7.5]. In brain cell cultures type 111 deiodination appears associated with the presence of glial cells [76,78,79]. In common with the other deiodinases, the type 111 enzyme is located in the microsomal fractions of the tissues [97,100] and is activated by thiols. Kinetic analysis

96 of the type 111 deiodinase has provided evidence for a sequential reaction mechanism [97] as has been found for the type I1 enzyme (Section 3.1) but different from the ping-pong kinetics for type I (Section 2.4). The type 111 deiodinase of brain requires high DTT concentrations (- 0.1 M) for maximal activity [97]; its physiological cofactor is unknown. Differences in substrate specificity between the type I11 deiodinase and the inner ring deiodinase activity of the type I enzyme are 1) the lower K , values of T4and T, (- 0.1 pM)for the type 111 deiodinase [97] and 2) the apparent inability of this enzyme to deiodinate sulfated iodothyronines [94,96]. The activities of both enzymes is decreased in hypothyroid and increased in hyperthyroid rats [71].

3.3. Possible other iodothyronine deiodinases The distinction of three types of iodothyronine deiodinases is based on circumstantial evidence concerning the catalytic properties of the different tissue activities. Until the enzymes are characterized at the molecular level with specific antibodies or by determination of their primary structure, this classification must remain putative. It is, therefore, no surprise that the criteria to distinguish the type I and I1 enzymes have been questioned [loll, and the existence of additional enzymes capable of ORD has been proposed [60,64,65]. Thus, it has been shown that sensitivity to PTU is not an absolute indicator for the type I deiodinase. At very high (mM) concentrations it also inhibits the type I1 enzyme especially with limited DTT [101,102]. Moreover, since PTU is an uncompetitive inhibitor of the type I deiodinase (Section 2 . 3 ) , it is relatively less potent at low substrate levels. It is, therefore, important to realize that under assay conditions for the low-K, type I1 enzyme PTU may not always completely inhibit the type I deiodinase. This is also a potential pitfall of the use PTU to investigate the origin of T, in plasma and tissues of rats, especially at the low iodothyronine levels in hypothyroid animals. Additional information is then required to assess the contribution of the different enzymes. Previous findings [ 1031 suggesting that peripheral production of T, in hypothyroid rats is provided primarily by PTU-insensitive (type 11) deiodination of T, have been confirmed in subsequent experiments [90]. Goswami and Rosenberg have suggested that liver and kidney microsomes contain in addition to the type I deiodinase multiple low-K, enzymes for the O R D of T, and rT, that differ from the type I1 enzyme [60.64,65]. This was mainly based on different susceptibilities to iopanoic acid and PTU if reactions were carried out at low substrate concentrations in the presence of various cofactors, i.e., DTT, GSH, glutaredoxin and thioredoxin [60,64,65]. It was even reported that the deiodinase activity stimulated by the thioredoxin system accepted rT, but not T, as substrate [64]. The uncertainty in the estimation of the low conversion rates in the nM substrate range which are not accounted for by residual activity of the type I deiodinase, however, questions the validity of the above conclusions. The possible exist-

97 ence of a family of low-K, isoenzymes for the ORD of T4and rT,, therefore, remains to be established.

4. Transport of iodothyronines into tissues In view of the hydrophobic character of iodothyronines it has been generally assumed that they are transported into tissues by simple diffusion. However, increasing evidence has been obtained in recent years indicating that the penetration of thyroid hormone into tissue cells is an active process mediated by specific carrier systems located in the plasma membranes [ 1041. Krenning et al. have found that the uptake of T,, T, and rT, into rat liver cells occurs at least in part through high-affinity, limited-capacity transport systems (reviewed in Ref. 104). The general properties of these systems are given in Table IV. The saturable component of thyroid hormone entry into hepatocytes is characterized by apparent K , values of about 20 and 1 nM for free T, and T,, respectively. Although T, is a competitive inhibitor of the transport of T, and vice versa, the discrepancy between K , and K , values makes it unlikely that these compounds are taken up by the same carrier. The translocation of thyroid hormone across the liver cell membrane is an energy-dependent process as evidenced by the impediment of T, and T4 entry by lowering of the incubation temperature or by diminution of the cellular ATP. Varying ATP depletion was induced a) by pretreatment of hepatocytes with T,, the mechanism of which effect is unknown, b) by addition of metabolic inhibitors such as KCN and oligomycin or c ) by a progressive decrease in medium glucose and further replacement with fructose. Under these conditions a strong correlation was found between the decrease in ATP and the loss of transport activity, although uptake of T, and rT, was more readily affected than uptake of Ti. This finding supports the view that the transport of T, and rT1 is mediated by a different system than uptake of Ti. Both systems appear to require an intact TABLE IV Plasma membrane transport of thyroid hormone into rat hepatocytes Saturation: Stimulation:

K,,, T, 1 nM K,,, T, 20 nM temperature

ATP dlhumin Inhibition:

T,, T, ouahain amiodarone radiographic agents monoclonal antibody ER-22 NTI-serum factors

Na' gradient across the cell membrane, as evidenced by the effects of the Naf,K+ATPase inhibitor ouabain [ 1041. The possibility that plasma iodothyronine-binding proteins play a more than passive role in the cellular uptake of thyroid hormone has been recognized recently. Saturation of T, entry into hepatocytes in monolayer culture is not observed when a protein-free medium is used, perhaps due to rate-limiting diffusion of the hormone through the unstirred water layer around the cells. Addition of bovine serum albumin (BSA) to the medium obviates this effect by providing a buffer of proteinbound T, at the cell surface which is not readily depleted by the cellular uptake. Although the effect of BSA on diffusion is maximal at 0.1%, further increases in BSA also stimulate carrier-mediated uptake of T, provided that the free hormone concentration is kept constant. The latter finding may reflect an increased dissociation of T, from BSA induced by interaction with the cell surface as has also been proposed for the facilitatory role of albumin in the hepatic uptake of fatty acids and certain drugs [105]. Also, in vivo studies have suggested a greater availability of albumin-bound T, and T, and perhaps also prealbumin-bound T, for tissue uptake than predicted from kinetic parameters determined in vitro [106]. The concept that plasma membrane transport plays a key role in the regulation of intracellular thyroid hormone levels is supported by studies with a monoclonal antibody against an antigen exposed on rat liver cells [107]. This antibody inhibited the uptake of different iodothyronines by rat hepatocytes under initial rate conditions as well as the metabolism of these compounds during prolonged incubations [107]. Uptake and metabolism of T,, T, and rT, were affected to the same extent, suggesting that a single system operates in the transport of different iodothyronines, which is opposite to the view advanced above. However, it is not excluded that the antibody interacts with a component of the plasma membrane and thereby affects multiple transport systems. Different diagnostic and therapeutic agents have been shown to interefere with the transport of T, and T, into hepatocytes [104]. Among these are iodinated substances such as radiocontrast agents and amiodarone which are structural analogues of thyroid hormone and probably compete for binding to the transporter. Inhibitory effects have also been observed with propranolol but this is thought to be due to a decrease in the ATP content of the cultured hepatocytes which probably does not occur in vivo [104]. Of special interest is the observation that plasma of patients with severe non-thyroidal illness (NTI) contains a factor that inhibits the binding of T, to plasma proteins as well as its uptake by human hepatocarcinoma (Hep G2) cells [ 1081. Energy-dependent, carrier-mediated uptake of thyroid hormone has been demonstrated in a variety of other cell types, including fibroblasts [109,1lo], pituitary tumor cells [ l l l ] and muscle cells [112]. In addition to the uptake processes in the plasma membrane, indirect evidence has been reported recently for active transport of T, between subcellular compartments, i.e., from the cytoplasm to the nu-

99 cleus [113]. The existence of such gradients would also have major implications for the regulation of thyroid hormone bioactivity.

5. Regulation of thyroid hormone metabolism As discussed in previous sections, the stepwise deiodination of T, is mediated by at least three different enzymes. Deiodination of the outer ring of T, and reverse T, is mediated by the type I and I1 enzymes while deiodination of the inner ring of T, and T, is catalysed by the type I and I11 enzymes. The contribution of the different enzymes to the peripheral production and clearance of T, and rT, can be estimated using PTU as a specific inhibitor of the type I deiodinase (for potential pitfalls of this approach, see Section 3.3). Thyroid hormone has a positive effect on the type I and type I11 enzymes but down-regulates the type I1 deiodinase. It has been demonstrated [lo31 that in euthyroid rats at least 70% of peripheral T, to T, conversion is derived from a PTU-sensitive mechanism which, therefore, must involve the type I deiodinase. PTU also inhibits the production of T, from T, as well as the clearance of plasma rT, in humans [13]. The important role of the type I deiodinase in the latter process is not surprising considering that rT, is the preferred substrate for the enzyme (Section 2.2). For this reason it is also logical to assume that any rT, generated by type I inner ring deiodination of T, is rapidly further degraded before being released into the circulation. The hypothesis that the type I enzyme of liver is a major site for the clearance but not for the production of plasma rT, is supported by direct estimates of arteriovenous gradients of rT, across the liver in patients with mild liver failure [114]. Although in principle the metabolic clearance of plasma T, may occur via several pathways, direct deiodination of the inner ring of T, by the type I enzyme seems of minor importance (Section 2.2). Also, glucuronidation in the liver does not represent an irreversible pathway of T, elimination since enzymatic hydrolysis of the conjugate in the intestine allows for the reabsorption of free T, (enterohepatic cycle). Further, the finding that plasma T, clearance is not affected in patients with liver cirrhosis [llS] suggests that hepatic metabolism of T, by sulfation and subsequent deiodination is less important than in rats. It appears, therefore, that the type I11 deiodinase of extrahepatic tissues is a major site for the clearance of plasma T, as it is also for the production of plasma rT,. Figure S is a model of the peripheral metabolism of thyroid hormone in normal humans which places the production of plasma T, and the clearance of plasma rT, predominantly in tissues with PTU-sensitive, type I deiodinase activity. Although the role of the liver is emphasized, contribution of the kidneys is not excluded. Clearance of plasma T, and production of plasma rT, is located mainly in tissues such as brain and perhaps skin with PTU-insensitive, type I11 deiodinase activity. It should be pointed out that the model suggests that type I1 deiodination of T,

T35

"

11 /

PT U - sensi t ive t i s s u e

PTU-insensitive t i s s u e

Fig. 5. Peripheral metabolism of thyroid hormone

does not contribute significantly to the production of plasma T, in euthyroid subjects. However, this does not negate the important function of the type I1 enzyme as a local source of intracellular T, in, for instance, brain and pituitary but also as a major producer of plasma T, in hypothyroidism [103]. The latter is understandable in light of the regulation of the different deiodinases by thyroid hormone. The 'low T, syndrome' is induced by a decrease in the production of plasma T, as well as the clearance of plasma rT, and is observed in several clinical situations such as starvation, systemic illness and the use of certain drugs [115]. In fasting [70] and illness [lo81 the abnormal thyroid hormone metabolism appears to result from a defective liver uptake and, therefore, a decreased supply of T4 and rT, for intracellular deiodination. In other conditions such as treatment with PTU or propran0101 [116], the defect appears localized in the type I deiodinase itself leading to a decline in T, formation and rT, breakdown.

References 1. Taurog, A. (1986) In: The Thyroid (Ingbar, S.H. and Braverman, L . E . , eds.) pp. 53-97. Lippin-

cott, Philadelphia. 2. Engler, D. and Burger, A.G. (1984) Endocr. Rev. 5 , 151-184. 3 . Hennemann, G. (1986) In: Thyroid Hormone Metabolism (Hennmann, G.. ed.) pp. 277-295. Marcel Dekker, New York. 4. Oppenheimer, J.H. and Samuels, H.H. eds. (1983) Molecular Basis of Thyroid Hormone Action. Academic Press, New York.

101 5 . Kaplan, M.M. (1984) Neuroendocrinology 38, 254260. 6. Hesch. R . D . and Koehrle, J. (1986) In: The Thyroid (Ingbar. S.H. and Braverman, L.E., eds.) pp. 154-200. Lippincott, Philadelphia. 7. Leonard, J.L. and Visser. T.J. (1986) In: Thyroid Hormone Metabolism (Henneman, G., ed.) pp. 189-229. Marcel Dekker, New York. 8. Visser, T . J . (1988) In: Coenzymes and Cofactors. Vol. 111, Glutathione: Chemical, Biochemical and Medical Aspects (Dolphin, D., Poulson. R. and Avramovic, 0..eds.) in press. Wiley, New York. 9. Bollman, J.L. and Flock, E.V. (1965) In: The Biliary System (Taylor, W., ed.) pp. 345-365. Blackwell, Oxford. 10. Miller. J.L.. Gorman, C.A. and Go. V.L.W. (1978) Gastroenterology 75, 901-911. 11. Burger, A.G. (1986) In: Thyroid Hormone Metabolism (Henneman, G., ed.), pp. 255-276. Marcel Dckker, New York. 12. Morreale de Escobar, G. and Escobar el Rey. F. (1967) Recent Progr. Horm. Res. 23, 87-137. 13. Cooper, D.S. (1984) N. Engl. J . Med. 311, 1353-1362. 14. Dutton. G.J., ed. (1980) Glucuronidation of Drugs and Other Compounds. CRC Press, Boca Raton. 15. Chowdhury. J.R., Chowdhury, N.R., Moscioni. A.D., Tukey. R., Tephly, T. and Arias. I.M. (1983) Biochim. Biophys. Acta 761, 58-65. 16. Mulder, G.J., ed. (1981) Sulfation of Drugs and Related Compounds. CRC Press, Boca Raton. 17. Sekura, R.D.. Sato. K., Cahnmann, H.J.. Robbins, J. and Jakoby, W.B. (1981) Endocrinology 108, 454-456. 18. De Herder, W.W., Bonthuis, F., Rutgers. M.. Otten, M.H., Hazenberg, M.P. and Visser, T.J. (1988) Endocrinology 122, 153-157. 19. Otten, M.H.. Mol, J.A. and Visser. T.J. (1983) Science 221, 81-83. 20. Visser, T . J . , Mol, J.A. and Otten, M.H. (1983) Endocrinology 112, 1547-1549. 21. Myant, N.B. (1956) Clin. Sci. 15. 551-555. 22. De Herder, W.W., Hazenberg. M.P., Pennock-Schroder, A.M., Henneman, G. and Visser, T.J. (1986) FEMS Microbiol. Lett. 35, 249-253. 23. De Herder, W.W., Hazenberg. M.P.. Pennock-Schroder, A.M., Henneman, G. and Visser, T.J. (1986) Med. Biol. 64, 31-35. 24. DiStefano, J.J. and Sapin, V. (1987) Endocrinology 121, 1742-1750. 25. De Herder. W.W., Bonthuis, F., Hazenberg, M.P.. Otten, M.H. and Visser, T.J. (1986) Ann. Endocrinol. 47. 23. 26. Rutgers, M., Hazenberg, M.P., Heusdens, F.A., Bonthuis, F. and Visser. T.J. (1987) Ann. Endocrinol. 48, 134. 27. Mol, J.A., Docter, R., Kaptein. E.. Jansen. G.. Hennemann, G . and Visser, T.J. (1984) Biochem. Biophys. Res. Commun. 124, 475-483. 28. Kohrle, J . , Kaiser, C., Rokos, H . , Hesch, R.D. and Leonard, J.L. (1987) Ann. Endocrinol. 48, 132. 29. Fekkes, D . , Hennemann, G. and Visser. T.J. (1983) Biochim. Biophys. Acta 742, 324-333. 30. Mol, J . A . , Van den Berg, T.P. and Visser, T.J. (1988) Mol. Cell. Endocrinol. 55, 149-157. 31. Mol, J.A., Van den Berg, T.P. and Visser, T.J. (1988) Mol. Cell. Endocrinol. 55, 159-166. 32. Visser, T.J.. Fekkes, D., Docter, R . and Hennemann. G. (1979) Biochem. J. 179, 489-495. 33. Eelkman Rooda, S.J., Van Loon, M.A.C. and Visser, T.J. (1987) J . Clin. Invest. 79, 1740-1748. 34. Mol, J.A. and Visser, T.J. (1985) Endocrinology 117. 8-12. 35. Rutgers, M., Bonthuis, F., De Herder. W.W. and Visser, T.J. (1987) J . Clin. Invest. 80, 758-762. 36. Fekkes, D., Hennemann, G. and Visser. T.J. (1982) Biochem. Pharmacol. 31, 1705-1709. 37. Fekkes, D.. Hennemann, G. and Visser. T.J. (1982) FEBS Lett. 137, 40-44. 38. Ruiz, M. and Ingbar, S.H. (1982) Endocrinology 110. 1613-1617. 39. Mol. J.A., Docter, R . , Hennemann. G . and Visser. T.J. (1984) Biochem. Biophys. Res. Commun. 120, 28-36,

102 40. Auf'mkolk, M., Koehrle, J., Hesch, R.D. and Cody, V. (1986) J. Biol. Chem. 261, 11623-11630. 41. Goswami, A., Leonard, J.L. and Rosenberg. I.N. (1982) Biochem. Biophys. Res. Commun. 104, 1231-1238. 42. Leonard, J.L. and Visser. T.J. (1984) Biochim. Biophys. Acta 787, 122-130. 43. Fekkes, D. , Hennemann, G. and Visser, T.J. (1982) Biochem. J. 201, 673-676. 44. Visser, T.J. and Van Overmeeren, E . , Fekkes, D., Docter, R . and Hennemann, G. (1979) FEBS Lett. 103, 314-318. 45. Chopra, I.J., Chua Teco. G.N., Eisenberg, J.B., Wiersinga. W.M. and Solomon, D.H. (1982) Endocrinology 110, 163-168. 46. Visser, T.J. and Van Overmeeren, E. (1979) Biochem. J . 183, 167-169. 47. Leonard, J.L. and Rosenberg, I.N. (1980) Endocrinology 106, 444-451. 48. Visser, T.J., Van der Does-Tobe, Docter, R. and Hennemann, G. (1976) Biochem. J. 157,479-482. 49. Goswami, A . and Rosenberg, I . N . (1981) J. Bid. Chem. 256, 893-899. 50. Garrett, C., Wataya, Y. and Santi, D.V. (1979) Biochemistry 18, 2798-2804. 51. Hartmann, K., Hartmann, N. and Bulka, E. (1971) Z . Chem. 9, 344-345. 52. Goswami, A. and Rosenberg, I.N. (1983) Endocrinology 112, 1180-1187. 53. Harris, A.R.C., Fang, S.L., Hinerveld. L., Braverrnan, L.E. and Vagenakis, A.G. (1979) J . Clin. Invest. 63, 516524. 54. Balsam, A. and Ingbar, S.H. (1979) J. Clin. Invest. 63, 1145-1156. 55. Gavin, L.A., McMahon, F.A. and Moeller, M. (1981) Endocrinology 109, 530-536. 56. Gavin, L.A. and Moeller, M. (1983) Metabolism 32, 543-551. 57. Gavin, L.A., McMahon, F.A. and Moeller, M. (1983) Diabetes 30, 694-699. 58. Sato, K. and Robbins, J. (1981) Endocrinology 109, 844-852. 59. Sato, K., Mimura, H . , Wakai, K., Tomori, T., Tsushima, T. and Shizume, K. (1983) Endocrinology 113, 878-886. 60. Goswami, A . and Rosenberg, I.N. (1985) J . Biol.Chem. 260, 6012-6029. 61. Visser. T.J., Mol, J.A. and Holmgren. A. (1986) In: Thioredoxin and Glutaredoxin Systems: Structure and Function (Holmgren, A , , Branden, C.I.. Jornvall, H . and Sjoberg, S . , eds.) pp. 369-376. Raven Press, New York. 62. Holmgren, A. (1986) In: Thioredoxin and Glutaredoxin Systems: Structure and Function (Holmgren, A , , Branden, C.I., Jornvall, H. and Sjoberg, S., eds.), pp. 1-9, Raven Press, New York. 63. Sawada, K., Hummel, B.C.W. and Walfish, P.G. (1986) Biochem. J . 234. 391-398. 64. Goswami, A. and Rosenberg. I.N. (1987) Endocrinology 121, 1937-1945. 65. Goswami, A. and Rosenberg, I.N. (1984) J. Clin. Invest. 74, 2097-2106. 66. Chopra, I.J. (1978) Science 199, 904-905. 67. Isaacs, J. and Binkley, F. (1977) Biochim. Biophys. Acta 497, 192-204. 68. Isaacs, J. and Binkley, F. (1977) Biochim. Biophys. Acta 498, 29-38. 69. Greenbaum. A.L., Gumaa, K.A. and McLean. P. (1971) Arch. Biochem. Biophys. 143, 617-663. 70. Van der Heyden, J.T.M.. Docter, R., Van Toor, H ., Wilson, J.H.P., Hennemann, G. and Krenning, E.P. (1986) Am. J. Physiol. 251, E156E163. 71. Kaplan. M.M. and Yaskoski, K.A. (1980) J. Clin. Invest. 66. 551-562. 72. Kaplan, M.M. (1980) Endocrinology 106. 567-576. 73. Leonard, J.L., Mellen, S.M. and Larsen, P.R. (1983) Endocrinology 112, 1153-1155. 74. Kaplan, M.M. and Shaw, E.A. (1984) J. Clin. Endocrinol. Metab. 59, 253-257. 75. Kaplan, M.M., McCann, U . D . , Yaskoski, K.A., Larsen, P.R. and Leonard. J.L. (1981) Endocrinology 109, 402. 76. Leonard. J.L. and Larsen, P.R. (1985) Brain Res. 327, 1-13. 77. St. Germain, D.L. (1986) Endocrinology 119, 84C-846. 78. Cavalieri, R.R., Gavin, L.A., Cole, R. and De Vellis, J. (1986) Brain Res. 364, 382-385. 79. Courtin, F., Chantoux, F. and Francon. J . (1986) Mol. Cell. Endocrinol. 48, 167-178.

103 80. Koenig. R . , Leonard, J.L., Senator, D.. Rappaport, N., Watson, A.Y. and Larsen, P . R . (1984) Endocrinology 115, 317-323. 81. St. Germain, D.L. (1985) J. Clin. Invest. 76. 89@893. 82. Visser, T.J., Leonard, J.L., Kaplan. M.M. and Larsen, P.R. (1982) Proc. Natl. Acad. Sci. U.S.A. 79. 508c-5084. 83. Visser, T.J., Kaplan, M.M., Leonard, J.L. and Larsen, P.R. (1983) J . Clin. Invest. 71. 992-1002. 84. Courtin, F., Pelletier. G. and Walker. P. (1985) Endocrinology 117. 2527-2533. 85. Leonard, J.L., Rennke, H.. Kaplan. M.M. and Larsen, P.R. (1982) Biochim. Biophys. Acta 718, 109-1 19. 86. Kaplan, M.M. (1979) Endocrinology 105, 548-554. 87. St. Germain, D.L. and Galton, V.A. (1985) J . Clin. Invest. 75. 679-688. 88. Leonard, J.L., Silva, J.E., Kaplan, M.M., Mellen, S.A.. Visser. T.J. and Larsen, P.R. (1984) Endocrinology 114, 998-1004. 89. Hidal, J.T. and Kaplan, M.M. (1985) J. Clin. Invest. 76. 947-955. 90. Silva, J.E. and Leonard, J.L. (1985) Endocrinology 116. 1627-1635. 91. Silva, J.E. and Larsen, P.R. (1986) J. Clin. Invest. 77. 1214-1223. 92. Silva, J.E. and Larsen, P.R. (1986) Am. J . Physiol. 251. E639-E643. 93. Melmed, S . , Nelson, M.. Kaplowitz. N . , Yamada. T. and Hershman, J . (1981) Endocrinology 108, 970-976. 94. Dickstein, Y . , Schwartz, H., Gross, J . and Gordon, A . (1980) Mol. Cell. Endocrinol. 20, 45-57. 95. Borges, M., LaBourene. J . and Ingbar, S.H. (1980) Endocrinology 107, 1751-1761. 96. Sato, K. and Robbins, J. (1980) J . Biol. Chcm. 255, 7347-7352. 97. Kaplan, M.M., Visser, T.J., Yaskoski. K.A. and Leonard, J.L. (1983) Endocrinology 112, 35-42. 98. Roti, E., Gnudi, A. and Braverman, L.E. (1983) Endocr. Rev. 4, 131-149. 99. Castro, M.I., Braverman, L.E., Alcx. S., Wu. C.F. and Emerson, C.H. (1985) J. Clin. Invest. 76, 1921-1926. 100. Huang, T.S., Chopra, I.J., Beredo. A.. Solomon, D . H . and Chua Teco. G.N. (1985) Endocrinology 117, 2106-2113. 101. Goswarni, A. and Rosenberg, I.N. (1986) Endocrinology 119, 916923. 102. Silva, J.E., Mellen, S. and Larsen, P.R. (1987) Endocrinology 121, 65C656. 103. Silva, J.E.. Gordon, M.B., Crantz. F.R.. Leonard, J.L. and Larsen, P.R. (1984) J. Clin. Invest. 73, 898-907. 104. Krenning, E.P. and Docter, R. (1986) I n : Thyroid Hormone Metabolism (Hennemann, G . . ed.) pp. 107-131. Marcel Dekker, New York. 105. Weisiger, R.A. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 1563-1567. 106. Pardridge, W.M. (1987) Am. J . Physiol. 252. E157-El64. 107. Mol, J.A., Krenning, E.P., Docter. R . . Rozing. J . and Hennemann, G. (1986) J. Biol. Chem. 216, 764c-7643. 108. Sarne, D.H. and Refetoff. S. (1985) J . Clin. Endocrinol. Metab. 61. 10461052. 109. Cheng, S-Y. (1983) Endocrinology 112. 17541762. 110. Docter, R., Krenning. E.P., Bernard. I1.F. and Hennemann, G. (1987) J. Clin. Endocrinol. Metab. 65, 624-628. 111. Halpern, J. and Hinkle. P.M. (1982) Endocrinology 110, 107&1072. 112. Pontecorvi, A . , Lakshmanan, M. and Robbins. J . (1987) Endocrinology 121, 2145-2152. 113. Oppenheimer, J.H. and Schwartz, H.L. (1985) J . Clin. Invest. 75, 147-154. 114. Bauer, A.G.C., Wilson. J.H.P., Larnherts. S.W.J., Docter, R . , Hennemann. G. and Visser, T.J. (1987) Acta Endocrinol. 116, 339-346. 115. Wartofsky, L. and Burman, K.D. (1982) Endocr. Rev. 3, 164-217. 116. Docter, R., Van der Heyden, J.T.M.. Krenning, E.P. and Hennemann, G . (1986) In: Frontiers in Thyroidology, Vol. 1 (Medeiros-Neto. G . and Gaitan, E.. eds.) pp. 423-426. Plenum, New York.

This Page Intentionally Left Blank

B.A. Cooke, R . J.B. King and H.J. van der Molen (cd5.) Hormones und their Actions, Purl I 01988 Elsevier Science Publishers BV (Biomedical Division)

105 CHAPTER 7

Characterization of membrane receptors: some general considerations LEO E. REICHERT, J r . Department of Biochemistry, Albany Medical College, Albany, NY 12208, U.S.A.

1. Introduction In subsequent chapters of this section on ‘Specific Action of Protein Hormones’, the mechanism of action of a variety of individual hormones will be considered in detail. The purpose of this ‘introduction’ is to discuss, in general terms, some considerations common to the study of ligands that bind to membrane receptors. Due to constraints of space, it should be understood that the reference selection in this chapter is highly eclectic and citations are chosen to illustrate particular points of discussion. Our own efforts have dealt with the study of membrane receptors for follitropin (follicle stimulating hormone, FSH). These have been summarized elsewhere [1,2] and so will not be dealt with in any detail at this time. A consideration of the membrane receptor-binding characteristics and post-binding effects of FSH, however, serves to illustrate several perplexing problems associated with this field of study. FSH is thought to act through the classical adenyl cyclase-CAMP 2nd messenger mechanism, although these sequelae may not be the only, or perhaps not even the major, mechanism of hormone action. FSH is a heterodimeric glycoprotein hormone. It binds to a multimeric glycoprotein membrane receptor [3] which, at the time of this writing, has not been chemically purified. When FSH is deglycosylated, receptor binding occurs, but post-binding events do not [4,5]. This suggests either an essential conformational change in the hormone as a result of the deglycosylation [6], or the presence of a critical receptor or parareceptor lectin-like binding domain necessary for signal transduction. Similar results have been reported with the other pituitary glycoprotein hormones, TSH [7] and LH [8]. In contrast, quite small molecules, such as the catecholamines, or simple peptides lacking a carbohydrate component, such as glucagon, also trigger the CAMP response. LH has been reported to stimulate the phosphatidylinositol2nd messenger pathway [9], whereas hCG, a molecule similar to LH in structure and biological

106 function, has, after deglycosylation, been reported to stimulate steroidogenesis, despite an apparent failure to stimulate CAMP accumulation [ 101. Clearly, major differences seem suggested in the mechanisms of hormonemembrane receptor interactions leading to various post-binding events [ 111. For FSH, as an example, there have been reports of hormone effects on membrane potentials of cultured Sertoli cells [12] and on conductance of artificial lipid membranes [13], as well as on amino acid transport systems [14]. Indeed, there has been one report of receptor-mediated gonadotropin action without receptor occupancy ~51. It may be paradoxical that for many investigators it is assumed the most powerful approach to answer questions related to function of membrane receptors is to remove them from their in situ membrane environment through use of various detergents. Demonstration of the functional nature of solubilized ligand-binding membrane proteins requires reconstitution of the receptor and related catalytic system components (G-protein, adenylate cyclase) into artificial membranes or liposomes, with consequent demonstration of hormone stimulatable post-binding effects, such as nucleotide binding or adenylate cyclase enzyme activation. Nevertheless, the cell membrane represents a complex, highly organized and quite specific environment for the hormone receptor and associated membrane proteins and, therefore, one must always be cautious in assuming an equivalency between in situ and ex situ experimental systems. Even between artificial lipid bilayers, it has been reported that ‘fully functional receptors’ show different binding characteristics depending on the environment of the reconstituted receptor [16].

2. Preparation of receptor probe Membrane bound receptors are generally characterized through use of radiolabelled ligands. It is absolutely essential, of course, that such ligands be chemically pure, for reasons related to specificity of the system under study (see Section 4.1). For protein or peptide hormones, this usually entails labeling with ‘*‘I, a relatively high energy isotope of reasonable stability and ready availability. A number of innovative approaches to radioiodination have appeared over the years, including such techniques as acylating lysyl residues on proteins through use of the N-hydroxysuccinimidyl ester of radioiodinated 3-(4-hydroxphenyl) propionic acid, popularly termed the Bolton-Hunter Reagent or through use of reductive alkylation. In most laboratories, however, radiolabelling is accomplished through use of either chloramine T or lactoperoxidase, and HzOzleading to ‘activation’ of iodine, probably as H,OI+. Although chloramine T labelling is considered the harsher approach and one that is more likely to damage protein, it should be noted that even with the enzymatic procedure, an oxidizing agent is required (hydrogen peroxide instead of chloramine T) and there is no theoretical reason to assume that one is less dam-

107 aging to susceptible proteins than the other, presumably through oxidation of accessible sulfhydryl, tryptophanyl or methionyl residues. For most protein or peptide ligands, the development of suitable conditions of iodination is one of trial and error, requiring manipulation of concentrations of all components of the system as well as of time and temperature of incubation. A detailed consideration of these various factors is beyond the scope of this introductory section. However, several caveats should be kept in mind. 1) A radiolabelling procedure suitable for use with a ligand intended for application in radioimmunoassay may not be equally suitable for preparation of radioligand for receptor studies [17]. 2) When using '"1, it is not possible to assume that the quality of the radioiodine as obtained from commercial reactors will always be the same [l8]. We have seen periodic variations in quality of radioiodine leading to damaged hormone and expressed as low specific binding to receptor. 3) Because of the possibility of variable damage to proteins during the iodination procedure, it may be necessary to fractionate after radiolabelling, such as by PAGE [ 171 in the absence of SDS or reducing agents. 4) Since there may be variations in quality of receptor as well as of radioligand it is necessary to develop rigorous and standardized control criteria for acceptability of radioligand after each labelling procedure, with each new batch of iodine, and with each new batch of membrane receptor. For example, one should always calculate specific radioactivity (pCi iodine incorporated per mg of ligand, as by the 'self-displacement' method) [19]. degree of bindability to excess receptor and extent of specific and non-specific binding t o a limited receptor concentration. Here, non-specific binding is defined as that which will occur in the presence of excess unlabelled hormone. Variation from normal ranges for any of these parameters should be a warning of complications in interpretation of any resulting data.

3. Preparation of membrane receptors 3.1. General considerations In studies of membrane receptors, a n important decision to be made is identification of a membrane source appropriate t o the goals of the project. Receptors, under the best of circumstances, arc present in vanishingly small amounts, with the added complication of instability and alteration of properties in some systems. Cuatrecasas [20] pointed out the advantages o f working with relatively crude homogenates in membrane receptor characterization studies, at least initially and for comparative purposes, to avoid the problem of damage or alteration of receptor during separation procedures which might not otherwise be detected. The usual choice to be made in terms o f receptor source is between membranes derived from whole cells in culture or membranes derived from whole tissue (organ) homogenates. As indicated tihove, the goal of the project will be important

here. If it is intended to obtain inferential information through use of such techniques as chemical cross-linking or photoaffinity labelling (there has been a report [21] of UV-induced labelling of iodinated hormone and receptor) with analysis of solubilized hormone receptor complex by polyacrylamide gel electrophoresis, then use of purified membrane preparations enriched with receptor, as from cells in culture, albeit more limited in quantity, may be preferable. If the goal is eventual purification of significant quantities of receptor for chemical and structural studies then, even with the availability of modern analytical microtechniques, relatively larger amounts of whole tissue will be needed to provide the required amounts of demonstrably purified receptor. A concern with receptors in membranes derived from cultured cells is that their properties may be altered by the culturing process, that is, the artificial conditions of their synthesis may concomitantly generate artifical receptor characteristics relative to the in vivo situation. A concern with use of whole tissue or organ homogenates, especially if collected at the abattoir, is that they are often collected under less than sterile conditions, requiring an awareness of possible consequences of bacterial contamination. Also, since it is often necessary to store such tissue frozen for prolonged periods, problems related to the time between collection and freezing (such as lytic enzyme action), conditions of freezing (temperature, time) and problems related to disruption of receptor-related membrane domains upon later thawing prior to use, must be considered (see Section 3.3).

3.2. Membranes f r o m cell culture Advances in cell culture techniques have rendered receptors present in membranes of almost any type of cell, potentially available. Again, caveats relate to the limited amounts of receptor available from membranes derived from cells cultured in most non-industrial laboratory settings, as well as possible effects of the culture conditions on receptor properties. Since cell culture techniques remain somewhat of an art, it is not possible to recommend general conditions for culture of all cell types. However, cell culture in the absence of serum may be desirable, since serum often contains factors that complicate interpretation of results, such as, ligand-binding inhibitors [22] that may be of significance in vitro, even if their in vivo role is uncertain. After an appropriate period of culture, the cells may be collected (often through use of a rubber policeman), suspended in an appropriate buffer, homogenized and membranes collected by differential centrifugation. Receptor-rich membrane fractions may then usually be stored safely at low temperatures (-80°C) in the presence of stabilizing agents such as glycerol [23] for indefinite periods.

109 3.3. Membranes from tissue hornogenates If the goals of the program require homogenates of whole organs as the source of membrane receptors, the initial choice is between freshly excised organs or those that, generally for logistical reasons, have been stored frozen, as is generally the case with abattoir material. With organs collected and stored frozen under circumstances not directly controlled by the investigator, it is essential to be aware of all aspects of the collection procedure, especially conditions of organ excision, time between excision and freezing, and the conditions of freezing itself (i.e., slow vs. flash freezing, temperature of storage, etc.). Further, the approach to thawing of frozen organ specimens must also be standardized (time required for thawing, temperature of the thawing process, etc.), in view of the well known effects that the freeze-thaw cycle may have on membrane components, their activities and properties. Conditions of homogenization of tissues should be determined to maximize yield of functional membrane receptors and minimize loss due to physical aspects of the homogenization procedure, such as too vigorous homogenization, extremes of pH and so forth. Special problems can exist in studying membrane receptors for protein hormones when the latter are derived from fresh whole tissue homogenates. It is, of course, virtually impossible to be confident that degradative damage to the membrane receptor has not occurred as a consequence of action of endogenous enzymes. Such effects occur quite rapidly and probably cannot be completely prevented by enzyme inhibitors or subsequent conditions of incubation, such as low temperatures. Also, a number of protein hormone receptors have been reported to have characteristics other than that of integral membrane proteins and have been recovered in the absence of detergents (for example, see Refs. 24, 25). Presumably such ‘receptors’ represent soluble components of transmembrane receptor proteins, or are hormone-binding proteins ‘shed’ by the membrane under conditions of buffer extraction.This focuses on the essential distinction between a hormone-binding protein and a functional hormone receptor. The latter can be defined as a receptor which, once complexed with its specific ligand, is associated with specific post-binding events. T o date, the presence of receptors in the circulation has not been convincingly demonstrated and, even if such receptors exist, it is likely that they represent structurally altered forms of the receptor compared to the functional receptor present in membranes. There has, however, been a report of autoantibodies to a peptide hormone receptor spontaneously developing as the anti-idiotype to the hormone antibody itself [26]. Another caveat in the use of whole tissue homogenates is the possible presence of hormone-binding inhibitors in such preparations [27]. Such inhibitors are extremely intriguing and of considerable potential interest, especially in terms of possible paracrine implications. Regardless of whether such factors have physiological

110 significance, their presence and effects must be considered when evaluating ligand binding characteristics of membranes in tissue homogenates. It is not possible here to detail various approaches to purification of membranes from tissue homogenates, as these are detailed in many specialty volumes and, in any event, must be developed in accord with the special requirements of the research project being pursued. However, in general, dialysis of membrane homogenates to remove possible low molecular weight binding inhibitors and transfer of membranes into a buffer medium compatible with subsequent binding studies seems prudent. An initial crude fractionation of whole tissue homogenate by centrifugation to separate heavier tissue components from lighter membrane fragments (at 7000 X g) is often helpful [28]. Assuming that binding studies indicate most membrane receptors are unsediniented at low g forces, a concentration step using available commercial ultrafiltration apparatus, such as the Amicon DC-10 unit fitted with various hollow fiber filtration cartridges (as the HI-50 fiber), can provide a rapid means for preparation of relatively large volumes of tissue homogenate for further processing [28]. A widely utilized approach to fractionation of membrane mixtures is by sucrose gradient centrifugation which separates membranes on the basis of their density. It should be clear that routine processing of large volumes of homogenates to satisfy the goals of a membrane receptor study will require, unavoidably, access to a large capacity ultracentrifuge and appropriate-sized rotor to allow processing of practical amounts of homogenate within reasonable time frames. In many cases, the need for purified membranes cannot be satisfied by even the largest sized rotors available, and pooling of successive runs will be necessary. Again, due to unavoidable dilution of membranes by sucrose density gradient centrifugation, as well as the need to remove sucrose prior to many types of subsequent experiments, it is often necessary to further dialyse and concentrate membranes, and this can also be accomplished through use of the previously mentioned ultrafiltration apparatus. The need for pooling of membrane fractions from successive runs is made more acute by the many advantages of utilizing a single large homogeneous batch of membranes for any particular series of membrane receptor studies. It is preferable to risk problems associated with pooling and storage of a single batch of membrane than to risk the almost inevitable variations in properties between successive batches of membranes and the problems the likelihood of such variations impose on interpretation of experimental data. If possible, membrane receptor preparations should be stored at maximal practical concentration, at lowest feasible temperature and, if needed, in the presence of a stabilizing agent such as glycerol. As mentioned earlier, their eventual warming for experimentation should be done under carefully standardized conditions.

111

4. Hormone-binding characteristics of the membrane receptor 4.1. Specificity

Assuming availability of an appropriately labelled ligand and suitable receptor preparation, it then becomes important to validate the specificity of the system. By this is meant developing experimental assurance that the putative receptor is interacting only with the hormone of interest. The latter is essential, since a fundamental aspect of receptor characterization is related to measurement of specific ligand binding and assessment of the ratio of ligand specifically bound to that which is unbound. The ratio of bound to unbound ligand allows assessment of affinity of binding and number and types of receptors present; essential criteria with respect to quality control aspects and when extending the data to conclusions regarding properties of the receptor (see Section 4.3 and 5 ) . Some confusion is possible when discussing ‘specific binding’ and ‘specificity of the system’. Specific binding is a measure of ligand binding to a particular receptor, whereas non-specific binding reflects binding of ligand to non-receptor components of the membrane under study or, in the case of the solubilized systems, non-receptor proteins that co-solubilize from the membrane under the selected experimental conditions. Specific binding used in this sense is often defined as that which can be blocked by the presence of large excesses of unlabelled ligand. Nonspecific binding being defined by the corollary, or that which occurs in the presence of large excess of unlabelled ligand. However, within the context of this section, specificity refers to competition for receptor binding by ligands other than the one of choice. The goal is to demonstrate that only Ligand A will bind to what is predicted to be Receptor A. This is usually done by demonstrating that relatively small concentrations of Ligand A inhibit binding of radiolabelled (for example) Ligand A to Receptor A, whereas much larger concentrations of Ligand B will not. In interpreting experiments of this type, it is important to keep in mind the problems of variable receptor affinities of different ligands, and also that very large amounts of ligands other than Ligand A may inhibit binding of radiolabelled Ligand A to receptor for reasons other than related to structural homologies and partial affinities, such as charge effects, alterations of pH, non-specific interactions with parareceptor regions effecting ligand-receptor interactions and so forth. Further, it should be noted that systems wherein ligand A-receptor A interactions are non-specific under one set of conditions, can be made specific by experimental adjustment of the binding system to exclude interfering substances. Care must be taken here, however, not to vary too much from physiologic subsequent results to the in vivo situation. Probably no aspect of membrane receptor study is subject to easy neglect more than that of specificity.

112

4.2. Selection of appropriate in vitro system 4.2.1. Effects of time, temperature, buffer Determining the optimal conditions for study of ligand membrane receptor interactions remains essentially an empirical enterprise, although some basic guidelines seem generally applicable. An accommodation is often necessary betweem time and temperature. At the lower temperatures, the period of incubations are generally relatively long and, therefore, prevent rapid conduct of an experimental series. At higher temperatures, binding proceeds much more rapidly, but there is danger of accelerated system degradation, either of the ligand or of the receptor. With many protein ligands, however, binding is generally more reversible at lower temperatures, becoming progressively less reversible as the temperature is elevated [29], thereby immeasurably complicating subsequent calculations of affinity, receptor concentrations or extensions into assessment of thermodynamic properties of the system, all of which assume a reversible ligand receptor system. Although seemingly mundane, it is critical to thoroughly understand the function of each component of the buffer system being utilized in the membrane receptor characterization studies. Many inorganic ions, such as sodium or phosphate, have profound effects of hormone receptor interactions and a a systematic study of the effects of various concentrations of inorganic ions on receptor properties is advisable [30]. One particularly noteworthy example of the care that must be taken in this regard is the observation that rapid down regulation of insulin receptors in adipocytes. rather than reflecting a profound observation of potential physiologic significance was, instead, due to the presence of Tris buffer in the incubation medium [311. 4.2.2. Steady-state (equilibrium) conditions It is self evident that the system under study must be at equilibrium or steady state for meaningful studies to be developed. This is particularly true when assessing the ability of various analogs or suspected binding regulators to effect hormonereceptor interactions. For example. it is possible that due to a lesser affinity, an inhibitor or analog may require 10 h to reach equilibrium with the receptor, whereas the ligand or precursor to the analog may require only 2 h. After that time, a conclusion may be drawn that the putative inhibitor does not inhibit binding of ligand to receptor, whereas, after 10 hours, binding inhibition could be significant. An example of this can be found in assessing the potency of FSH binding inhibitors secreted by the bacteria ‘Serratia’ [32] (see Fig. 2 of that report). It is, therefore, important to consider, in a systematic fashion, the kinetics of binding of the ligand and analogs under study before drawing conclusions regarding receptor affinity. This is the reverse concept from that previously mentioned (Section 4. l), where experimental conditions are altered to allow only a particular ligand to bind to receptor.

113

5. Molecular properties of the membrane receptor The molecular properties of membrane receptors have been studied by a variety of techniques, but the general approaches most often taken include characterization of free receptor or radiolabelled hormone-receptor complexes on the basis of hydrodynamic and/or charge properties, after column filtration or electrophoresis. A prerequisite to such studies is solubilization of the receptor (free or complexed) prior to analysis. When utilizing column chromatography through gels of calibrated pore sizes, it is common to first bind the radiolabelled ligand to the receptor in the presence (control) and absence of large excesses of unlabelled ligand. Each ligand-membrane preparation is then solubilized and the extracts filtered separately through the same calibrated column of an appropriate molecular sieve gel. One should detect a decrease in incorporation of radiolabelling in the putative receptor fraction when the incubation is done in the presence of excess unlabelled ligand. The elution volume of the ligand-receptor complex can then be utilized to estimate molecular weight of the receptor, making suitable adjustments for the molecular weight of the ligand. An important assumption in use of this common approach is that the hormone-membrane receptor complex is of sufficiently high affinity to preclude dissociation during gel-filtration, and this must be verified experimentally. In addition to simple monitoring of radioactivity, there are other techniques available for detection of the receptor. One, for example, would be to precipitate the ligand-receptor complex using antibodies to either the receptor or (more likely) to the ligand as a means of detection of the elution volume of the complex upon gel filtration. Analysis of hormone-membrane receptor complexes are now routinely done after polyacrylamide gel electrophoresis of the detergent-solubilized complex either in the presence and absence of reduction agents, or by isoelectric focusing in acrylamide gels with specific pH gradients. Thus, separations can be achieved using both size and charge properties of the receptor or ligand receptor complex. Often, the ligand is covalently linked to the receptor to prevent dissociation of the complex as a result of experimental manipulations and to facilitate identification of the complex. This can be done either through use of a wide range of commercially available chemical cross-linking reagents or by photoaffinity labelling of the receptor. The receptor can be identified on the basis of mobility of radiolabelled bands. This technique usually requires band detection after autoradiography and has, as a prerequisite, inclusion in the series of appropriate controls to distinguish between specific and non-specific binding of radiolabel to solubilized non-receptor membrane components. One disadvantage of this technique is the length of time it may require to develop usable autoradiographs, and the differences that may occur if autoradiographs are viewed after varying times of development. Further, it often appears that creative viewing is encouraged in the detection of extremely faint or diffuse bands on the gel. Because of the usually small gel size, relatively

114 minor variations in mobility can cause significant differences in estimates of molecular weights or other calculated parameters, or real differences in mobility may not be apparent due to limitation of the resolving power of the technique or the system. Because of this, replicate analyses are usually necessary and one should guard against overly ambitious interpretations of what often is an extremely subjective analysis of PAGE results.

6. Solubilization of the membrane receptor According to the model developed by Helenius and Simons [33], solubilization of integral membrane proteins is a consequence of detergent penetration into the membrane, which is dependent on the size and chemical structure of the detergent and which results in increased curvature of the membrane lipid bilayer. The effectiveness of the detergent, therefore, is related to its hydrophile-lipophile balance (HLB) value. Detergents with HLB values that fall between 10 and 15 have been found to represent a particularly effective group of surfactants that have been widely used to facilitate removal (solubilization) of membrane proteins. These include such detergents as Nonidet P-40, Lubrol PX, Lubrol WX, Triton X-100 and N-octyl glucoside. Whereas only 1% of membrane-bound adenylate cyclase activity from calf testes, for example, could be solubilized by buffer alone, from 3% (Lubrol WX) to 25% (Nonidet P-40) of membrane-bound adenylate cyclase activity could be solubilized in the presence of non-ionic detergents. Failure to achieve a higher recovery of activity after solubilization can be partly attributed to destabilization of the integral membrane protein under study, probably via denaturation. Others have reported an ‘imperfect’ substitution of detergent for phospholipid in contact with the hydrophobic domain of membrane proteins, leading to modification (loss) of activity. Addition of glycerol (30% wivol) usually results in increased recovery and stability, probably reflecting its role as a ‘thermodynamic booster’ and a general protein-stabilizing agent [34]. Clearly, stability of solubilized membrane receptor is of critical importance, since the reconstitution experiments and especially purification protocols will no doubt require receptor preparations that can be effectively utilized over a period of at least several days. Triton X-100 is a reagent used successfully by many investigators for the sobulization of integral membrane proteins, such as membrane-bound hormone receptors. An excellent treatment of this topic can be found in the review dealing with solubilization of membranes by detergents written by Helenius and Simons [33]. A serious problem in the use of Triton X-100 and other types of detergents is the lack of availability of chemically analysed and purified preparations. Common contaminants that complicate solubilization of sensitive membrane proteins by Triton X100 or related polyoxyethylene alcohols are powerful oxidizing agents, found in one

115 study to range between 0.04 to 0.22% H,O, equivalents [35].Indeed, differences between membrane-bound and isolated acetylcholine receptors have been explained on the basis of sulfhydryl group oxidation in the presence of Triton X-100. It is advisable, therefore, to purify commercially available non-ionic detergents prior to their use in hormone-receptor solubilization and related studies. Detergents such as Triton X-100 or deoxycholate tend to bind more readily to integral membrane proteins than to soluble proteins. When membrane proteins are freed of both detergents and lipids their exposed hydrophobic surfaces often cause aggregation and possibly denaturation.

Summary As indicated earlier, the intent of this section was not to be global with respect to the scope of its coverage, but rather to discuss in general terms some considerations common to the study of ligands which interact with membrane receptors and, thereby, elicit post-binding events. Many of the examples chosen have been drawn from my experience with the follitropin-gonadal receptor system, but they provide instances of problems, concerns and caveats in use of techniques and interpretation of results that are common to this particular field of study. The reader is referred to the specific examples of hormone receptor interactions to follow, wherein aspects of the problems not germaine to this section, such as, for example, techniques for purification of solubilized receptors, are considered in detail.

References 1. Reichert, L . E . , Jr., Dias, J . A . , Fletcher. P.W. and O’Neill. W.C. (1982) In: The Cell Biology of the Testis (Bardin, C.W. and Skarens. R.J.. eds.) pp. 135-150. Ncw York Acadcmy of Sciences,

New York. 2. Reichert, L.E.. Jr., Andersen. T . T . , Dias. J.A.. Fletcher, P.W., Sluss, P.M.. O’Neill. W Smith. R.A. (1984) In: Hormones and Receptors in Growth and Reproduction (Saxena, B.B.. Catt. C.. Birnbaumer, L. and Martini, L.. eds.) pp. 87-101. Raven Press, New York. 3. Smith, R . A . . Branca, A . A . and Reichert. L.E.. Jr. (1966) J . Biol. Chem. 261, 985k9853. 4. Calvo, F.O., Keutmann, H . T . , Bergert. E.R. and Ryan, R.J. (1986) Biochemistry 25, 3938-3943. 5. Manjunath, P . , Sairam, M . R . and Sairam. J . (1982) Mol. Cell. Endocrinol. 28, 125-138. 6. Keutmann, H.T., Johnson. L. and Ryan, R.J. (1985) FEBS Lett. 185, 333-338. 7. Amir. S.M.. Kubota. K . , Tramontano. D . , Ingbar. S.H. and Keutrnann, H . T . (1987) Endocrinology 120, 345-352. 8. Sairam, M.R. and Fleshner, P. (1981) Mol. Cell. Endocrinol. 22, 41-54. 9. Davis. J . S . , Weakland, L.L., West, L . A . and Farese, R.V. (1986) Biochem. J . 238, 597-604. 10. Bahl. O.P. (1977) Fed. Proc. 36. 2119-2127. 11. Themmen. A.P.N.. Hoogerbrugge, J.W., Rommerts, F . E . G . and van der Molen. H.J. (1985) Biochem. Biophys. Res. Commun. 128. 1163-1172. 12. Roche. A . and Joffre. M . (1984) IRCS Mcd. Sci. 12, 57&571.

116 13. Deleers, M . , Chatelain, P . , Poss. A . and Ruysschaert, J.M. (1979) Biochem. Biophys. Res. Commun. 89, 1102-1106. 14. da Cruz Curte, A. and Wassermann, G.F. (1985) J. Endocrinol. 106, 291-294. 15. Fletcher, W . H . and Greenan, J.R.T. (1985) Endocrinology 116, 166C1662. 16. Sargent, D . F . , and Schwyzer, R. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 57745778. 17. Schneyer, A.L., Sluss, P.M., Bosukonda. D. and Reichert, L.E. (1986) Endocrinology 119, 14461453. 18. Melson, B . E . , Sluss, P.M. and Reichert, L . E . (1987) Anal. Biochem. 160, 434439. 19. Calvo, J.C., Radicella, J.P. and Charreau. E.H. (1983) Biochem. J . 212. 259-264. 20. Cuatrecasas, P. (1974) In: Annual Reviews of Biochemistry (Snell. E.E.. ed.) pp. 169-214. Annual Reviews, Inc., Palo Alto, California. 21. Iwanij, V. and Hur, K.C. (1985) Proc. Natl. Acad. Sci. U.S.A. 82. 325-329. 22. Sanzo, M. and Reichert, L.E.. Jr. (1982) J . Biol. Chem. 257, 6033-6040. 23. Dias, J.A., Huston, J.S. and Reichert, L.E. (1981) Endocrinology 109. 73&742. 24. Dias, J . and Reichert, L.E., Jr. (1982) J . Biol. Chem. 257, 613-620. 25. Wimalesena, J . and Dufau, M.L. (1982) Endocrinology 110, 1004-1012. 26. Shechter, Y . . Maron. R.. Elias, D. and Cohen. I.R. (1982) Science 216, 542-545. 27. Krishnan, K . A . , Sluss. P.M. and Reichert. L.E. (1986) J . Androl. 7, 42-48. 28. Dattatreyamurty. B., Schneyer, A . and Reichert. L.E. (1986) J . Biol. Chem. 261, 1310413113. 29. Andersen, T.T., Curatolo, L.M. and Reichert. L.E., Jr. (1983) Mol. Cell. Endocrinol. 33, 37-52. 30. Andersen. T.T. and Reichert. L.E.. Jr. (1982) J. Biol. Chem. 257, 11551-11557. 31. Rennie. P. and Cliemann, J . (1981) Biochem. Biophys. Res. Commun. 102, 824-831. 32. Sluss, P.M., Ewing. J . F . , Melson. B.E. and Reichert, L.E. (1985) Biol. Reprod. 33, 925-933. 33. Helenius. A . and Simons. K. (1975) Biochim. Biophys. Acta 415. 29-79. 34. Timasheff, S.N.. Lee. N.J., Pitz. E.T. and Tweedy, N . (1976) J . Colloid Interface Sci. 55, 658,665. 35. Ashani, Y . and Catravas, G . E . (1980) Anal. Biochem. 109. 55-62,

B.A. Cooke. R . J . B . King and H.J. van dcr Molcn (cd\.) Hormones utid tlwir Actions. Purl 1 01988 Elsevicr Scicncc Publishers BV ( B i o m e d d Division)

117 CHAPTER X

Metabolism and intracellular processing of protein hormones ASHA SINGH KHANNA and DAVID MORTON WAISMAN Cell Regulation Research G r o u p , Deptirtment of Medical Biochemistry, T h e University of’ C d g a r y , Calgary, Alberta, Canada T 2 N 4N1

1. Introduction The polypeptide or protein hormones consist of a specific group of regulatory molecules whose functions are to convey specific information among cells and organs. These functions of communication probably arose early in the development of life and subsequently evolved into a complex system designed for the control of growth, development and reproduction and for the maintenance of metabolic homeostasis. The major sources of peptide hormones are the pituitary (hypophysis), the hypothalamus, the pancreatic islets, the placenta, the parathyroids and the gastrointestinal tract. The representatives of the protein hormones are shown in Table I. The peptide hormones consist of small peptides ranging in size from as small as three amino acids (thyrotropin releasing hormone) to 199 amino acids (prolactin), and larger glycoproteins. The glycoprotein hormones include a carbohydrate moiety that plays a role in mechanism of action and the physiological disposition of the compound. The two pituitary gonadotropins (FSH and LH), the placental gonadotropin (hCG), and thyrotropin (TSH), all share a common a-chain but have a distinct Pchain that confers biological specificity to the hormone. Isolated subunits are inactive but can be combined to give fully active molecules which have the biological specificity of the psubunit [1,2]. The a-subunits each contain two oligosaccharides, @subunits contain one or two oligosaccharides. The carbohydrate substitutions contribute to both the folding of a-and Psubunits into mature hormone [3] and to the relatively long circulating half-lives of glycoprotein hormones, compared to those of peptide hormones. Half-lives range from 30 min for LH to many hours for hCG. This chapter deals with various aspects of metabolism of protein hormones, in particular, how these regulatory molecules are synthesized and stored as precursor molecules and proteolytically processed to generate biologically active hormones. The mechanisms involved in secretion and degradation are also described.

118 TABLE I Protein hormones Hormone

Site of origin

Adrenocorticotropin (ACTH) Anterior a-Melanocyte stimulating Hormone (MSH) Anterior PEndorphin Anterior Thyroid stimulating hormone (TSH) Anterior

pituitary pituitary pituitary pituitary

Follicle stimulating hormone (FSH)

Anterior pituitary

Luteinizing hormone (1.13)

Anterior pituitary

Somatotropin (growth hormone) Prolactin Vasopressin Oxytocin Thyrotropin releasing hormone Somatostatin Insulin Glucagon Parathyroid hormones Calcitonin Gastrin Cholecystokinin Secrctin Chorionic gonadotropin (hCG)

Antcrior pituitary Antcrior pituitary Posterior pituitary Posterior pituitary Hypothalamus Hypothalamus Pancreas Pancreas Parathyroid Thyroid Gastro-intestinal tract Gastro-intestinal tract Gastro-intestinal tract Placcnta

Epidermal growth factor Nerve growth factor

Unknown Submaxillary glands

Structure 139 amino acids 13 amino acids 3 I amino acids glycoprotein: a chain 92 aa p chain 112 aa glycoprotein: a chain 92 aa /3 chain 118 aa glycoprotein: 01 chain 92 aa p chain 115 aa I91 amino acids 199 amino acids 9 amino acids 9 amino acids 3 amino acids 14 amino acids 51 amino acids 29 amino acids 84 amino acids 32 amino acids 17 or 34 amino acids 33 or 39 amino acids 28 amino acids glycoprotein: a chain 92 aa p chain 142 aa 53 amino acids I18 amino acids

2. Biosynthesis of protein hormones As mentioned above, the protein hormones contain one or two peptide chains and range in size from three amino acid residues (thyrotropin-releasing hormone) to 199 residues (prolactin). Glycoprotein hormones such as the gonadotropins have two peptide chains, a and p, containing respectively 92 and 100-140 residues. The synthesis of these protein hormones is directed by one or more genes - with one or more genes coding for the amino acid sequence and other genes being responsible for alterations of the peptide to its final form.

2.1. Transcription and translation The biosynthesis of the protein hormones involves the synthesis of molecular forms larger than the polypeptide secreted from the cell [4,5]. The genetic information in DNA which codes for peptide hormones is discontinuous with intervening se-

119 quences between regions which code for protein (Fig. 1). The gene is initially transcribed into a large molecular weight messenger RNA precursor. Intervening sequences (introns) are progressively removed, and the coding portions (exons) are spliced together into mature messenger RNA, which is capped and tailed. This genetic arrangement is thought to have facilitated evolution through recombination. The mRNA specific for the polypeptide hormone attaches to a free ribosome, and translocation at an AUG initiation codon is initiated.

intion

1 DNA

rr_

I Transcription messenger RNA precursor

Ill

1

Excision

1

Splicing

n

I

A-A

t a i l mature messenger RNA

Translation

mRNA

I /

- An

Bound ribosome

n

.

Signal p e p t i d e

ER lumen

1 f

Processing

80 0 m

Endoplasmic reticulum

A" Processing

o,/

Secretory vesicles

Fig. 1. Diagrammatic representation of mechanism of synthesis and processing of protein hormones.

120

2.2. Interaction of signal peptide with RER membrane The initial signal peptide or presequence, which is composed of many hydrophobic amino acids, causes the mRNA-ribosome complex to attach to the endoplasmic reticulum and induces a pore through the endoplasmic reticulum membrane which allows entry into the lumen [6]. Another theory to explain the mechanism of translocation of the peptide into the endoplasmic reticulum [7,8] is explained by the recurrent structural features of the signal peptides, particularly their strongly hydrophobic central regions (usually residues 7-17), as well as the non-random disposition of amino acids having charged side chains, especially within the more hydrophilic N-terminal region. According to this model, the signal peptides enter directly into the lipid bilayer leaving the more hydrophilic N-terminal region on the ribosomal side and with their hydrophobic central regions, extend across the membrane so as to form a loop in the entering peptide chain. The hydrophobic signal peptide is believed to form Ppleated sheet structure with preexisting membrane proteins, which function as non-specific receptors having secondary structure interactions [ 11. A third theory [9] suggests that the genetic information contained in the conformation of the newly synthesized (pre) secretory polypeptide hormone, induces a conformational change in the membrane on contact, which triggers the entry of the peptide into the membrane and its subsequent insertion or passage across the membrane. Recent studies based on post-translational analysis indicate that an essential feature of translocation may have to do with preventing the folding of the protein into a tertiary structure [lo]. Other workers have shown that high concentrations of the reducing agent dithiothreitol in cell-free systems enabled the post-translational translocation of preprolactin [ 111, thus implying that, by chemically preventing the formation of disulfide bonds, folding is prevented and translocation can occur posttranslationally across the membranes. The synthesized molecule including the signal peptide is referred to as the preprohormone. The signal protein [ 121 or leader sequence [ 131 therefore serves the important function of moving the protein hormone from the surface of the rough endoplasmic reticulum (RER) to the intracisternal space from which it will move to the Golgi apparatus. The preprohormone is short-lived as the signal peptide is removed by proteolytic cleavage shortly after synthesis [ 14-16]. 2.3. Cleavage of signal peptide The protease making the initial cleavage appears to be an endopeptidase which acts preferentially on amino acids having small neutral side chains such as Ala, Ser or Cys. Experiments with heterologous reconstituted systems suggest that it is unlikely that highly specific converting enzymes exist for each presecretory hormone [15]. Some evidence has been reported that the converting enzyme may be a metalloprotease [17,18]. Removal of the ‘pre’ sequence leaves the hormonal polypeptide as prohormone.

121

3. Processing of prohormones Proteolytic processing of precursor protein hormones is a widely used mechanism for producing biologically active hormones, with the exception of a few hormones such as growth hormone and prolactin [19,20]. Proteolytic cleavage of prohormones begins only after synthesis and folding of the prohormone peptide chain has been completed in the endoplasmic reticulum and the peptides have been transferred to and concentrated in the Golgi area [21,22]. The movement from endoplasmic reticulum to the Golgi is an energy requiring process, mediated by small transport vesicles which appear to bud from smooth regions of the endoplasmic rePROINSULIN (9K) 6-chain

A-chain

C

N

PROGASTRIN (-10K)

PROSOMA TO STAT1N (-1 2.5 K)

PROGLUCAGON (-18K)

PRO OPIOMELANOCORTIN (29K)

Y-MSH

B-LPH

"-MSH

a-MSH

PROENKEPHALIN (LEU) ( 2 9 K )

_ - -- - _--

- -- - -_

cx -Neo-Endorphin

--

I

I

C

N L ---_ - ---------Leu-Enkephalin

PROPARATHYROID HORMONE ( 1 0 K )

. ... PTH

Nr---~ L--A

C

Fig 2 Schematic structures of some known prohormones showing amino acid residues at malor cleavage sites 0, lysine, 0. arginine, 0. carbohydr'ite moicty The shaded areas indicate regions which appear as biologically active hormone?

122 ticulum, migrate to the Golgi area and fuse with the cis elements of Golgi complex [21,23] (Fig. 1).

3.1. Structures of prohormones The structures of some of the prohormones is summarized in Fig. 2. Half-lives of these prohormones range from about 20 min to several hours [24,25]. The site of cleavage of the precursor hormone is usually marked by a pair of basic amino acids, such as Arg-Arg or Lys-Arg. Apart from the highly characteristic feature, the ‘pro’ sequences vary greatly in length and structure and may occur in any part of the molecule.

3.1 . I . Pro-opiomelanocortin ( P O M C ) peptide family One of the highly characterized prohormones is the pro-opiomelanocortin (POMC) also known as proadrenocorticotropin and 31 K ACTH/endorphin. The POMC family consists of peptides which act as hormones (ACTH. LPH, MSH) and others which may serve as neurotransmitters or neuromodulators (endorphins). POMC is synthesized as a precursor molecule of about 285 amino acids and is processed differently in various regions of the pituitary. One gene is responsible for POMC expression. 3.1.1.1. The POMC Gene There is remarkable structural similarity between the POMC genes of various species and between human genes for POMC and proenkephalin. There are general similarities between the organization of these genes and those which encode other hormones, including corticotropin-releasing hormone, nerve growth factor, glucagon and calcitonin. These genes have large 3‘ end exons containing the coding sequences for the active peptides and have repeated regions of intrasequence homology. In POMC there are three such regions, each about 50 nucleotides in length, which code for a-, pand y-MSH. POMC and proenkaphalin have small exons, located about 3 kb upstream from the large 3’ exon which code for the N-terminal peptide, the signal peptide and a 5’ non-translated region. 3.1.1.2. Distribution and processing of POMC gene products The POMC gene is expressed in the anterior and intermediate lobes of the pituitary. POMC or related products are found in several other vertebrate tissues including the brain, placenta, gastro-intestinal tract, reproductive tract, lung and lymphocytes. This is presumably due to gene expression in these tissue (rather than to absorption from plasma), but has only been proved for brain, placenta and testes. The POMC protein is processed differently in the anterior lobe than in the intermediate lobe. The intermediate lobe is rudimentary in adult humans, but is active in human fetuses, pregnant women during late gestation and in many animal species. Processing of POMC protein in peripheral tissues resembles that in intermediate lobe. There are three basic peptide groups in the POMC polypeptide: (1)

123 ACTH, which can give rise to a-MSH and corticotropin-like intermediate lobe peptide (CLIP); (2)p-lipotropin (P-LPH), which can yield y-LPH, P-MSH and pendorphin (and thus a- and y-endorphins); and (3) a large N-terminal peptide which generates y-MSH. The diversity of these products is due to the many dibasic amino acid clusters that are potential cleavage sites for trypsin-like enzymes. The signal peptide is cleaved, and post-translational modifications such as glycosylation, acetylation and phosphorylation occur. The next cleavage in both anterior and intermediate lobes is between ACTH and P L P H , resulting in an N-terminal peptide with ACTH and a p-LPH segment. ACTH is subsequently cleaved from the N-terminal peptide, and in the anterior lobe essentially no further cleavages occur. In the intermediate lobe ACTH is cleaved into a-MSH and CLIP, P-LPH is converted to yLPH and p-endorphin. p-MSH is derived from y-LPH. 3.1.1.3. Additional modifications of POMC peptide family There are extensive additional modifications of these peptides. Much of t h e N-terminal peptides of POMC as well as ACTH are glycosylated in the anterior pituitary. a-MSH is found predominantly in an N-acetylated and carboxy terminal amidated form. p-Endorphin is rapidly acetylated in the intermediate lobe and made less active. In the hypothalamus p-endorphin is not acetylated and is presumably active. @-Endorphin is also trimmed in the intermediate lobe at the C-terminal end to form a- and yendorphin. The large N-terminal fragment is also extensively cleaved but not much is known about this fragment.

3.2. Significance of ‘pro’ sequence The significance of the ‘pro’ sequence remains obscure. It would appear that most of the prohormones have little if any biological activity. Possible roles for the ‘pro’ region include ensuring the correct folding of the hormone (as in proinsulin), providing a minimum critical length for segregation and transport through the secretory apparatus, or enabling the released peptides to act as signals [26]. It has also been suggested that the ‘pro’ regions reflect the evolutionary origin of the polypeptide, namely from the more primitive process of lysosomal digestion [27,28].

3.3. Cleavuge at dibasic amino-ucids Evidence suggests that most prohormones are processed by a special group of intracellular proteases located in the Golgi apparatus and developing secretory granules. There appears to be a requirement for dibasic residues as a signal for this converting activity, and Arg rather than Lys may be preferred on the carboxyl side of the pair [26]. One exception to this is the Lys-Lys pair in the ACTH/endorphin precursor which is located just before the P-MSH sequence. However, this bond is believed to be cleaved only in the pars intermedia during the maturation of the product [29,30]. The conversion of G34 gastrin also requires cleavage at a Lys-Lys pair,

124 but this also seems to be a relatively slow cleavage which allows G34 to accumulate as an intermediate form [31]. Abnormal proinsulin, in which mutation changes convert Arg to neutral residue, shows prevention or slowing down of the normal cleavage process [32].

3.4. Cleavage at monobasic amino acids Several other precursors have been identified that undergo cleavage at a single basic amino acid (usually arginine), such as dog proinsulin [33], the vasopressin-neurohypophysin precursor [34] and cholecystokinin [35]. Recent work has shown that in approximately one-third of the cases of monobasic cleavage, there is an occurrence of proline residues either immediately before or after the arginine [36]. Presence of proline residues is believed to be important for the cleavage mechanism as it influences the three-dimensional orientation of the peptide backbone, due to its a-nitrogen atom being part of the rigid pyrrolidine ring, which does not allow rotation. The special conformational constraint induced by proline residues can slow down the proteolysis or it can make the peptide chain more susceptible to proteolysis by presenting the Arg in an optimal way to the processing enzyme. It is still unclear whether these sites are processed by the same protease(s) which normally recognizes the dibasic pairs or by the other trypsin-like proteases associated with the endoplasmic reticulum or storage granules. These and other variations suggest that a number of different proteolytic enzymes may normally be present in the Golgiisecretory granule system and participate to varying degrees in prohormone processing [34]. The selection of cleavage sites thus may be dictated both by nature and relative activities of these enzymes as well as by the primary, secondary and tertiary structure of the precursors.

3.5. Processing enzymes The processing enzymes can be divided into three classes: (a) endopeptidases which cleave the intact precursor into peptide fragments; (b) exopeptidases which remove amino acids from the amino or carboxy termini of the peptides; and ( c ) other enzymes involved in the post-translational modification of the peptide.

3.5.1. Endopeptidases Various trypsin-like serine proteases such as kallikreins [37] and plasmin or plasminogen activator [38] have been suggested to be involved in the cleavage of hormone precursors. However, it is not clearly known whether any of these proteases can correctly process proinsulin to generate both native insulin and C-peptide. Cleavage of proinsulin by plasmin is shown to result in equal amounts of insulin and a form of insulin with a residual Arg residue before the Gly at the N-terminus of the A-chain [39]. There may possibly be a novel amino peptidase for removing this

125 Arg residue, but none, as yet, has been demonstrated. Trypsin cleavage of proinsulin generates small amounts (< 5%) of Arg-(A)-insulin, the form which is present in native insulin at very low levels (< 0.5%). In the absence of carboxypeptidase-B, both plasmin and trypsin tend to cleave Lys-Ala or Lys-Thr bonds in proinsulin, resulting in a partially degraded form of insulin that does not occur naturally (401. These serine proteases therefore seem to carry out the rare cleavage at single basic residue sites. Thiol proteases resembling lysosomal cathepsin-B have been implicated as the main processing enzymes [41,42]. Several thiol-containing proteases have been identified in partially purified islet granule fractions [42,43]. These have been shown to include a major 31.5 kDa component which is probably identical to cathepsin B and a 39 kDa component which on treatment with pepsin gives a form which is immunologically similar to cathepsin B [34]. Both the high and low molecular mass forms of this cathepsin B-like protein are present in both lysosomes and secretory granules [44]. However, it is suggested that ordinary lysosomal cathepsin-B is unlikely to be the converting protease for proinsulin or other prohormones because of its broad specificity [ 5 ] . Recent studies have found that cathepsin B is derived from a larger precursor of approximately 44 kDa, suggesting that an active precursor form or processive intermediate of procathepsin B, or a related thiol protease may be cosegregated into secretion granules along with proinsulin and participate in its conversion, while the bulk of the mature cathepsin B may be directed mainly to the lysosomes [34]. A similar model has been proposed [45] which assumes that the converting enzyme is a trypsin-like protease which is stored as a zymogen, and that such an enzyme cannot be active within the granule prior to secretion. The converting enzyme is stored with the prohormone and is activated during secretion, and the active enzyme converts the prohormone to hormone.

3.5.2. Exopeptidases In addition to endoproteolytic processing of prohormones, a carboxypeptidase-B like activity is required to remove the C-terminal basic amino acids left behind on these peptidases by the trypsin-like enzymes [40]. Earlier studies showed that proinsulin labelled with [3H]arginine was processed in isolated secretion granules to insulin and C-peptide accompanied by liberation of free [3H]arginine [46]. Dipeptides of basic amino acids such as Arg-Arg- or Lys-Arg were not found to be released. Lysed granule preparations were also demonstrated to release C-terminal Arg residues from [3H]Arg-labelled proinsulin cleaved by trypsin before addition to the granule extracts [34]. A soluble carboxypeptidase B, requiring a metal ion such as zinc and having a pH optimum of 5.5, has also been identified in islets, and similar carboxypeptidase B-like activities have been recently described in adrenal homogenates [47] and purified chromaffin granules [48]. The available data thus suggests that a distinct carboxypeptidase B-like enzyme is utilized in secretory granules for prohormone processing. However, it remains to be seen whether the

126 same enzyme is active in all tissues and whether it is similar to the acidic lysosomal carboxypeptidase. An aminopeptidase activity has been demonstrated to cleave the N-terminal Arg residue from bovine Plipotropin [49]. This activity appears to be specific for basic residues, but is blocked by the presence of proline residues on the carboxyl side of the N-terminal arginine [22]. However, it is not yet clear whether this aminopeptidase is involved in precursor processing in situ.

3.6. Post-translational modifications Several post-translational processing steps occur during the translocation of the prohormone through the membrane systems of the cell. These include asparaginelinked glycosylation of the prohormone and disulfide bond formation in RER, and more complex glycosylation of the prohormone in the Golgi [5@53]. The oligosaccharide chains are attached to the polypeptide backbone of glycoproteins at one of five amino acid residues: asparagine (Asn), serine (Ser), threonine (Thr), hydroxylysine (Hyl) or hydroxyproline (Hyp). There are two types of chemical bonds that provide the attachment sites, 0-glycosidic links and N-glycosidic links. In each glycoprotein hormone (FSH, LH, TSH and hCG), the mubunit has either one or two. The complex N-linked oligosaccharide contains the PMan-di-N-acetylchitobiose (Man-pl,4-Glc N Ac-P-l,bGlc N Ac-Asn) core structure and also consists of a variable number of outer chains containing sialic acid (sial), galactose (Gal) and fucose (fuc) residues linked to the core. Usually, 2a-mannose (Man) residues are at-

Fig. 3. Formation of complex N-linked oligosaccharides. A t least 11 discrete enzymes in three organelles act sequentially to modify the common precursor of N-linked oligosaccharides. The numbered steps in the diagram are as follows: 2.3. glucose removed; 4, four mannose removed; 5 , one N-acetylglucosamine (GlcNAc) added; 6, two mannose removed; 7, GlcNAc + fucose added; 8. GlcNAc added; 9. three galactose and three sialic acid added. N , N-acetylglucosamine; 0 . galactose; 0,mannose; A , fucose; A, glucose; sialic acid. Reproduced with permission from Mol. Cell Biol. (1986) Darnell, Lodish and Baltimore, p. 964, W . H . Freeman & Co.. U.S.A.

+.

127 tached directly to the P-Man-N-acetylchitobiose structure. The outer chains most often consist of sial-Gal-Glc-NAc. The generation of the oligosaccharide dolichol precursor and the transfer of its oligosaccharide moiety occurs in the rough endoplasmic reticulum. At that site, glycosylated secretory proteins which contain the high-mannose oligosaccharide moieties, migrate to the Golgi complex. The oligosaccharide moieties of these glycoproteins are further modified to form the complex oligosaccharides (Fig. 3 ) . This modification process involves glycosidases and glycosyltransferase enzymes. The elongation process generating the complex type oligosaccharides of glycoprotein hormones occurs exclusively in the Golgi complex. Many biologically active secreted peptides are also amidated at their carboxyl terminal, and acetylated at their amino-terminal. The consequences of these modifications are (a) to reduce the susceptibility of these peptides to degradative actions of extracellular aminopeptidases and carboxypeptidases after their secretion and (b) to influence the biological activity of the peptides. Corticotropin-releasing factor, gastrin, cholecystokinin and vasopressin require the C-terminal amide for full activity [54-561. Acetylation of the N-terminus of a-MSH is necessary for activity, whereas acetylation of P-endorphin inhibits its opioid activity [57].The enzymes responsible for acetylation have been identified from bovine and rat intermediate lobes [57]and enzymes with a-amidation activity have been reported in preparations of pituitary secretory granules [54,55].

4 . Storage of protein hormones Once formed, hormones are either stored or secreted. Amounts of protein hormones sufficient to maintain normal secretory rates for hours to a day, are stored in the secretory granules. The limited capacity of the synthesizing tissues to store hormones is a chemical consequence of their unsuitability for incorporation into any of the three main storage compartments of the body (lipids, glycogen or protein). As a consequence of these factors the body pools of most hormones tend to be small.

5. Release of protein hormones Because of the limited capacity for storage, most hormones are released into plasma as a reflection of the rates of synthesis. The storage vesicles containing the hormone migrate to the plasma membrane by a process that involves microfilaments. The membrane of the secretory vesicles fuses with the cell membrane, the intervening membrane breaks down and the contents of the storage vesicle are released into the extracellular space. This mechanism is termed exocytosis. The trigger for secretion of the hormone is generally an influx of Ca2+,which is accompanied by changes in the electrical potential across the cell membrane.

128

In general, secretion of hormones does not occur at a constant rate. In some cases, secretion is pulsatile, occurring in short bursts, as for many pituitary hormones. The biochemical basis of this pulsatile release is not fully understood, but the release pattern may have profound effects on hormone function, i.e., the pulsatile administration of luteinizing hormone releasing hormone (LHRH) stimulates the release of L H by the pituitary, whereas the constant infusion of the same amount of hormone per unit time has the opposite effect.

6. Circulation in blood After secretion, the hormones circulate in blood for periods ranging from minutes (insulin) to a few hours (glycoprotein hormones). The protein hormones, being water-soluble, circulate in free form and are not bound to specific proteins. Their concentrations in blood are very low, the resting concentrations being lo-" M to 10-l" M. Under stimulated conditions the concentrations of peptide hormones in blood may rise 5- to 100-fold.

7. Degradation of protein hormones Once in circulation, protein hormones interact with their target tissues. The interaction of hormone and target tissues can be passive or involve active uptake. U1timately hormones are degraded by their target tissues. These processes of uptake and degradation are responsible for removing biologically active hormones from the circulation and therefore terminating the hormone signal. Hormone binding to receptors, while initially reversible, often leads to irreversible formation of the hormone-receptor complex, internalization and degradation of the hormone, presumably in the lysosomes. Receptor-mediated degradation is probably the major pathway for the degradation of most protein hormones in vivo, but other cellular degradation mechanisms are also present which are not completely understood. 7.1. Degradation of glycoprotein hormones Peptide hormones are, in general inactivated by proteases largely in the target tissues. The glycoprotein hormones (TSH, LH, FSH and hCG) have a different mechanism for degradation. These hormones typically have a sialic acid at the peripheral end of their carbohydrate side chain. Removal of the sialic acid, the mechanism of which is not understood, results in the exposure of a free galactose which is specifically recognized and bound by a receptor present on the hepatocyte. This receptor-binding is followed by internalization and degradation of the hormone. The asialo forms of the glycoprotein hormones are rapidly removed from plasma and destroyed.

129 7.2. Internalization of protein hormones Briefly, the process of internalization appears to involve several steps [58-611. Hormone initially binds in a reversible manner to receptors distributed diffusely over the cell surface. Within one to two hours, up to 50% of bound hormone becomes irreversibly associated with the receptor due at least in some cases to covalent bond formation [62]. In most systems studied so far, the initial steps of ligand-receptor internalization have certain common morphological properties. Most commonly, ligand-receptor aggregates accumulate over specialized, coated regions of the plasma membrane [63-651, although in some cases the ligand is internalized by non-coated imaginations [64]. These receptor-rich regions of the plasma membrane invaginate and pinch off, and are presumed to form clathrin-coated vesicles-specialized cytoplasmic structures implicated in the intracellular transport of proteins [66]. In most cases, these vesicles are rapidly translocated to the Golgi region of the cells. After a delay of 15-30 min, the hormone accumulates in large, multivesicular structures resembling secondary lysosomes. Only later the protein hormone accumulates in lysosomes, and in some cases in other intracellular organelles. The time course for internalization varies, depending on the hormone. Uptake of insulin, E G F and NGF is rapid (2-30 min) [67-691, whereas that of gonadotropins is much slower [63]. Internalization of protein hormones is a principal means of hormone degradation in lysosomes through a receptor-mediated pathway [70]. There may also be membrane-associated degradative activities for insulin (insulinase) and other protein hormones, however for EGF, degradation occurs exclusively through an intracellular process [68,71]. The degraded products of protein hormones are not retained intracellularly to any appreciable extent but are secreted as low-molecular weight fragments. The overall scheme for internalization also includes secondary pathways of hormonereceptor complexes to other target organelles besides lysosomes, as evident from the high affinity binding sites for insulin in nuclei and smooth endoplasmic reticulum [72]. Although the biological importance of these intracellular binding sites is unknown, it is possible that they are precursors of those found on plasma membranes. The binding of the hormone to its internal receptor is believed to mediate the organelle’s effects through alterations in the organelle membrane, and to mediate intracellular functions leading to increased synthesis of protein, RNA, and DNA and culminating in cell growth, cell differentiation or enzyme activation. Primarily, internalization is a means of removing excess, extracellular hormones by an active degradative pathway in lysosomes, thus terminating the hormone signal. Some other mechanisms of clearance of protein hormones are the following. a. Rapid degradation or conversion in blood or plasma, is observed with a number of smaller peptides like angiotensin 11, enkephalins and somatostatin as a result of the action of circulating aminopeptidases etc. However, this is not one of the major mechanisms of inactivation in vivo.

130 b. Conversion or degradation in transit through a tissue bed occurs in the lung, kidney and probably liver, for example, extensive conversion of angiotensin I to angiotensin I1 and the inactivation of bradykinin in the pulmonary circulation by the angiotensin converting enzyme. c. Equilibration with tissues and possible degradation as in the case of insulin. Equilibration and degradation refers to a process in which a peptide hormone can enter and leave sites within a tissue at which degradation occurs, for example, extensive degradation of ACTH occurs in the rat during uptake into tissues such as muscle and skin in the first minute after injection. Subsequently, the rate of degradation in the tissues falls and ACTH and peptide fragments return to the circulation.

8. Conclusions 1. Protein hormones are synthesized by the ribosomes in the form of a preprohormone which is biologically inactive.

2. The N-terminal of the preprohormone (the signal peptide) which may be involved in the penetration of the protein to the lumen of the ER, is cleaved in the Golgi-secretory vesicle system. 3. The remaining prohormone contains stretches of amino acids sequences unrelated to biological activity. These sequences are removed by proteolytic enzymes such as endopeptidases which act on dibasic or monobasic amino acid residues, and exopeptidases (e.g., carboxypeptidases and aminopeptidases) which remove amino acid residues from the C-terminal and N-terminal ends, respectively. A single prohormone may yield one or more biologically active and distinct hormones. 4. Several post-translational processing steps occur as the prohormone moves through the ER/Golgi system. These include asparagine linked glycosylation of the prohormone and disulfide bond formation, as well as extensive glycosylation in the Golgi. Furthermore, many hormones are amidated and/or acetylated.

5. The biological activity of the hormones is terminated at the target tissues chiefly by internalization and degradation in the lysosomes.

References 1. Pierce, J.G., Liao. T . H . . Howard, S . M . . Shome, B. and Cornell, J.S. (1971) Recent Prog. Horm.

Res. 27. 165-212. 2. Pierce, J.G. and Parsons, T.F. (1981) Ann. Rev. Biochem. 5 0 , 465-495.

131 3. Weintraub, B.D., Stannard, B.S., Linnekin. D. and Marshall, M . (1980) J. Biol. Chem. 255, 57 15-5723. 4. Steiner, D . F . , Quinn. P.S., Chan, S . J . . Marsh. J . and Tager. H.S. (1980) Ann. N.Y. Acad. Sci. 343, 1-16. 5. Docherty. K. and Steiner, D.F. (19x2) Ann. Rev. Physiol. 44, 625-638. 6 . Blobel. G . , Dobberstein, B. (1975) J. Cell Biol. 67, 835-851. 7. Steiner, D . F . , Quinn, P.S., Patzelt. C., Chan, S.J., Marsh, J. and Tager, H.S. (1980) In: Cell Biology: A Comprehensive Treatise, Vol. 4 (Goldstein, L. and Prescott, D.M., eds.) pp. 175-201. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

Academic Press, New York. Di Rienzo, J.M., Nakamura, K. and Inouye, M. (1978) Ann. Rev. Biochem. 47, 481-532. Wickner, W . (1979) Ann. Rev. Biochem. 48, 23-45. Zimmerman, R . and Meyer. D.I. (1986) TIBS 11. 512-515. Maher, P.A. and Singer, S.J. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 9001-9005. Habener. J.F. and Potts. J . T . , Jr. (1978) N. Engl. J. Med. 299, 58@585; 635-644. Blobel, G . and Sabatini, D. (1971) In: Biomembranes, Vol. 2 (Manson, L.A., ed.) pp. 193-195. Plenum Press, New York. Dorner, A.J. and Kemper, B. (1978) Biochemistry 17, 5550-5555. Shields, B . and Blobel. G . (1978) J . B i d . Chem. 253, 3753-3756. Patzelt, C . , Labrecque, A . D . , Duiguid. J.R., Carroll, R.J., Keim, P . S . , Heinrikson, R.L. and Steiner. D . F . (1978) Proc. Natl. Acad. Sci. U.S.A. 75. 126G1264. Zimmerman, M.. Ashe, B.M., Alberts. A.W.. Pierzchala. P.A., Powers, J.C., Nishino. N.. Strauss, A.W. and Mumford, R . A . (1980) Ann. N.Y. Acad. Sci. 343, 405-413. Zwinzinski, C. and Wickner, W . (1980) J . Biol. Chem. 255, 7973-7977. Seeburg. P.H., Shine, J . Martial, J . A . . Baxter. J . D . and Goodman, H.M. (1977) Nature 270,486-494. Cooke, N.E., Coit, D . . Weiner, R . I . , Baxter. J.D. and Martial, J . A . (1980) J. Biol. Chem. 255.

6502-6510. 21. Jamieson, J . D . , Palade, G . E . (1977) In: International Cell Biology (Brinkley, B.R. and Porles, H.R., eds.) pp. 308-317. New York University Press, New York. 22. Gainer, H . , Russell, J.T. and Loh, Y.P. (1985) Neuroendocrinology 40, 171-184. 23. Farquhar. M . G . and Palade, G . E . (1981) J. Cell Biol. 91, 77s-103s. 24. Steiner. D . F . , Cunningham. D.D., Spigelman, L. and Aten, B. (1967) Science 157, 697-700. 25. Steiner, D.F. (1967) Trans. N.Y. Acad. Sci. Ser. 11, 30, 6(&68. 26. Steiner, D.F.. Kemmler. W . , Tager, H.S.. Rubenstein, A . H . , Lernmark, A . and Zuhlke, H . (1975) In: Proteases and Biological Control (Reich. E.. Rifkin, D . and Shaw, E., eds.) pp. 531-549. Cold Spring Harbour Lab., Cold Spring Harbor N.Y. 27. Hales, C.N. (1978) FEBS Lett. 94, 1C-15. 28. Hales, C.N. and Docherty, K. (1979) In: Protcases and Hormones (Agarwal, M.K., ed.) pp. 19-46. ElsevieriNorth Holland Biomedical Press. Amsterdam. 29. Scott, A.P., Ratcliffe, J.G., Rees, L . H . . Landon. J.. Bennett. H . P . J . , Lowry, P.J. and McMartin, C. (1973) Nature 244, 65-67. 30. Eipper. B . A . and Mains, R.E. (1980) Endocrine Rev. 1. 1-27. 31. Gregory, R . and Tracy, H . (1975) In: Gastrointestinal Hormones (Thompson, J . , ed.) pp. 13-24, Univcrsity of Texas Press, Austin. 32. Robbins. D . C . , Blix, P.M., Rubenstein. A . H . , Kanazawa, Y.. Kosaha, K. andTager, H.S. (1981) Nature 291. 679-681. 33. Kwok, S.C.M., Chan, S.J. and Steiner, D.F. (1983) J . Biol. Chem. 258, 2357-2363. 34. Steiner, D . F . , Docherty, K. and Carroll. R.J. (1984) J. Cell Biochem. 24. 121-130. 35. Ryder. S.W., Straus, E. and Yalow. R.S. (1980) Proc. Natl. Acad. Sci. U.S.A. 77, 3669-3671. 36. Schwartz, T . W . (1986) FEBS Lett. 200, 1-10. 37. Ole-Moi Yoi. O . , Seldin, D . C . , Spragg. J . , Pinkus, G.J. and Austen, K.F. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 3612-3616.

132 38. Virji, M.A.G., Vassalli, J.D., Estensen, R.D. and Reich. E . (1980) Proc. Natl. Acad. Sci. U.S.A. 71, 875-879. 39. Markusson, J . (1979) In: Proinsulin, Insulin and C-peptide (Baba. S., Kaneko, T. and Yanaihara, N., eds.) pp. 50-61. Excerpta Medica, Amsterdam. 40. Kemmler, W., Peterson, J.D. and Steiner, D.F. (1971) J . Biol. Chem. 246, 67866791. 41. Chang, T.L. and Loh, Y.P. (1983) Endocrinology 112, 1832-1838. 42. Docherty, K., Carroll, R.J. and Steiner, D.F. (1982) Proc. Natl. Acad. Sci. U.S.A. 79,4613-4617. 43. Fletcher, D.J., Quigley, J.P., Bauer, G.E. and Noe, B.D. (1981) J . Cell Biol. 90, 312-322. 44. Hutton, J.C., Penn, E.J. and Peshavaria, M. (1982) Diabetologia 23. 365-373. 45. Green, D.P.L. (1984) Medical Hypotheses 15, 47-59. 46. Kemmler, W., Steiner, D.F. and Borg, J . (1973) J . Biol. Chem. 248, 4544-4551. 47. Hook, V.Y.H., Eiden, L.E. and Brownstein, M.J. (1982) Nature, 295, 341-342. 48. Fricker, L.D. and Snyder, S.H. (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 38863890. 49. Gainer, H . , Russell, J.T. and Loh, Y.P. (1984) FEBS Lett. 175, 135-139. 50. Freedman, R.B. and Hawkins, H.C. (1980) In: Enzymology of Post Translational Modification of Proteins, Vol. 1, p. 456. Academic Press, London. 51. Loh. Y.P. and Gainer, (1983) In: Brain Peptides (Krieger, D., Brownstein, M.J. and Martin, J . , eds.) pp. 79-116. Wiley, New York. 52. Loh, Y.P.. Brownstein, M.J. and Gainer, H. (1984) Ann. Rev. Neurosci. 7, 189-222. 53. Zimmerman, M., Mumford, R.A. and Seiner, D.F. (1980) Ann. N.Y. Acad. Sci, 343. 1-449. 54. Eipper. B.A., Glembotski, C.C. and Mains, F.E. (1983) Peptides 4, 921-928. 55. Eipper, B.A., Mains, R.E. and Glembotski, C.C. (1983) Proc. Natl. Acad. Sci. U.S.A. 80,51445148. 56. Manning, M., Olma, A., Klis, W., Kolodziejczyk, A , , Nawrocka, E., Misicka, A,, Seto, J . and Sawyer, W. (1984) Nature (London) 308, 558-560. 57. Glembotski. C.G. (1982) J . Biol. Chem. 257, 10493-10509. 58. Schlessinger, J., Schechter, Y., Willingham, M.C. and Pastan, I(1978) Proc. Natl. Acad. Sci. U.S.A. 75. 2659-2663. 59. Das. M. and Fox, C.F. (1978) Proc. Natl. Acad. Sci. U.S.A. 75,2644-2648. 60. Baldwin, D., Prince, M . , Marshall, S., Davies. P. and Olefsky, J.M. (1980) Proc. Natl. Acad. Sci. U.S.A. 77, 5975-5978. 61. Gorden, P . , Carpentier, J.L., Freychet, P. and Orci, L. (1980) Diabetologia 18, 263-274. 62. Saviolakis, G.A., Harrison, L.C. and Roth, J . (1981) J . Biol. Chem. 256. 4924-4930. 63. Amsterdam. A , , Nimrod. A , , Lamprecht, S.A., Burnstein. L. and Lindner, H . R . (1979) Am. J . Physiol. 236, E129-138. 64. Haigler, H.T., McKanna. J.A. and Cohen, S. (1979) J . Cell Biol. 81, 382-395. 65. McKanna. J.A., Haigler, H.T. and Cohen, S. (1979) Proc. Natl. Acad. Sci. U.S.A. 76,5689-5693. 66. Pearse, B.M.F. (1976) Proc. Natl. Acad. Sci. U.S.A. 73, 1255-1259. 67. Carpentier, J.L., Gordon, P., Freychet. P. and Le Cam. A. (1979) J. Clin. Invest. 63, 1249-1261. 68. Carpenter, G . and Cohen, S. (1976) J. Cell Biol. 71, 159-171. 69. Johnson, E . M . , Jr., Andres, R.Y. and Bradshaw, R.A. (1978) Brain Res. 150, 319-331. 70. Desbuquois, B., Willeput. J. and Huet de Froberville, A. (1979) FEBS Lett. 106, 338-344. 71. King, A.C., Hernaez-Davis, L. and Cuatrecasas, P. (1980) Proc. Natl. Acad. Sci. U.S.A. 77. 3283-3287. 72. King. A.C. and Cuatrecasas, P. (1981) New Engl. J. Med. 305, 77-78.

B . A . Cookc. R.J.B. King and H.J. van der Molcn (eda.) Hormones wid their Actions. Purr I 01988 Elsevicr Science Publishers BV (Biomedical Diviaion)

133 CHAPTER 9

Internalization of peptide hormones and hormone receptors DEBORAH L. SEGALOFF and MARIO ASCOLI The Population Council, 1230 York Avenue, New York, N Y 10021, U.S.A.

1. Introduction Peptide hormones are one class of many agents present in the bloodstream that affect the multiplication and differentiated functions of mammalian cells. The ability of a particular peptide hormone to elicit an effect in the appropriate target cell is dictated by the presence of receptors on the surface of the target cell which specifically bind that hormone. Although the cellular responses to the different peptide hormones vary, as do many of the mechanisms of signal transduction that translate the binding of the hormone to the cellular response, there is one salient feature that all peptide hormones studied to date share. This is the receptor-mediated endocytosis (RME) of the hormone. The idea that proteins could be internalized by a receptor-mediated mechanism by their target cells was sparked by the pioneering studies of Goldstein and coworkers [l]and by Cohen and co-workers [2,3], who obtained evidence for the receptor-mediated internalization and degradation of low-density lipoprotein (LDL) and epidermal growth factor (EGF), respectively, in the mid 1970s. Although endocytosis of a non-specific nature had been described by then, the concept of endocytosis of a specific ligand being mediated by the binding of that ligand to a cell surface receptor was unprecedented. These investigators were one of the first to study the binding of 12sI-labelledligands to intact cells (as opposed to studying the binding of the ligand to membranes, which was the prevailing approach at the time). Interestingly, their studies showed that when the binding studies on the cultured cells were performed at 37"C, but not at 4"C, there was a time-dependent accumulation of degradation products

Abbreviafions and trivial names used are: RME, receptor-mediated endocytosis: LDL, low density lipoprotein; EGF. epidermal growth factor; SDS, sodium dodecyl sulfate; LH, luteinizing hormone; hCG, human chorionic gonadotropin; and G protein. guanine nucleotide binding protein.

of the ligand in the culture medium. That the degradation of these ligands was occurring as a result of internalization of the ligand into the cell was suggested by observations that the accumulation of degradation products in the medium was both energy- and temperature-dependent and that it could be inhibited by agents known to inhibit lysosomal function. By using specific treatments to release the surfacebound ["'IILDL or [12sII]EGF,it was possible to document the appearance of intracellular radioactivity (representing intact or partially degraded ligand) prior to the release of degradation products into the medium. Furthermore, it was found that some compounds (such as metabolic inhibitors) prevented the accumulation of intracellular ligand (presumably by inhibiting internalization); whereas other compounds known to inhibit lysosomal function (such as NH4Cl or chloroquine) allowed internalization, but prevented degradation of the ligand [3-71. Concomitant morphological studies by electron microscopy on the fates of receptor-bound LDL and E G F (using ligands covalently attached to electron-dense ferritin) elegantly confirmed the inferences from the biochemical data that these ligands were internalized and degraded in the lysosomes [S-111. Since the internalization and degradation of ligand was strictly dependent upon binding of the ligand to the cell surface receptor, this process was called receptor-mediated endocytosis (RME). RME has since been shown to occur with other transport proteins, other growth factors, and with peptide hormones (for reviews see Refs. 12-16). The general features of RME as they are understood today from biochemical and morphological studies on a variety of ligands are discussed below as they pertain to peptide hormones.

2. General features of receptor-mediated endocytosis A schematic overview of RME is shown in Fig. 1. The cell surface receptors for a particular hormone are either located in areas of the plasma membrane referred to as coated pits or they are randomly distributed throughout the cell surface and migrate to the coated pits upon binding of the hormone. Coated pits are indented areas of the plasma membrane where there is an intracellular 'lining' of the membrane of the total with the protein clathrin and they constitute a small percentage (4%) area of the plasma membrane [8,17,18]. In the cases where the hormone-receptor complexes migrate to coated pits, there often is a microaggregation of the complexes (two to four per group) during this redistribution [19]. Following this microaggregation there is a more massive clustering of hormone-receptor complexes in the coated pits. Coated pits containing receptor-bound hormones become invaginated and pinch off intracellularly to form what are called coated vesicles. The coated vesicles still have clathrin associated with them, forming basket-like structures around the ves-

135

e d pit

\

f receptor sequestered

coated vesicle

/ O r

\

\endosome

0 -

lysosome

hormone and receptor degraded

i

\

receptor recycled

GD CURL

@

0’ /

y

x

receptor hormone clathrin

hormone degraded

Fig. 1 . Schematic representation of the possible routes of receptor and hormone during RME

icles [20]. The lumen (fluid-filled interior) of the coated vesicles does not have any free hormone. A t this stage, the hormone is still bound to the receptor, facing the lumen [lo]. With time, the coated vesicles shed their clathrin coats and fuse with other similar vesicles; all this time these vesicles are moving further into the interior of the cell [15]. The prelysosomal vesicles resulting from these fusions are called endosomes or endocytic vesicles and have a critical role in RME due to the acidic environment of their lumen. Although not as acidic as lysosomes (with an intra-compartmental pH of 4.5, see Ref. 21), the pH 5.5 environment of the endosome [22] is sufficiently low to cause the dissociation of some hormones from their receptors. When this occurs, there is a subsequent sequestering of the free hormone from the receptor in a related vesicle and tubule compartment called CURL (compartment for uncoupling of receptor from ligand, see Ref. 23), where the free hormone is sequestered into the vesicular structure while the receptor accumulates in the membrane of the tubule structure. A subsequent physical separation of these compartments allows for the differential processing of the hormone versus the receptor. Thus, while the free hormone is ultimately delivered (via vesicle fusion) to the lysosome where it is degraded, the free receptor may be recycled (via the Golgi compartment) to the cell

136 surface, where it can rebind hormone and repeat the whole process of RME. The free receptor may also remain sequestered intracellularly. It should be pointed out, however, that not all hormones dissociate from their receptor in the pH 5.5 environment of the endosome [24]. Some hormone-receptor complexes require much lower pH values for dissociation to occur. Although not a peptide hormone, the iron-transport protein transferrin is a peculiar example of this phenomenon and should be pointed out. In this case, at the neutral pH of the extracellular fluid transferrin containing bound iron binds to its cell surface receptor and is internalized. In the low pH environment of the endosome, iron becomes dissociated from transferrin, but transferrin remains bound to its receptor. The transferrin receptor, with bound transferrin, is then recycled to the cell surface. With iron no longer bound to the transferrin, the transferrin readily dissociates from its receptor at the neutral pH of the extracellular fluid [25,26]. This mechanism provides for an efficient continual uptake of iron into cells. Unlike transferrin, however, in those instances where peptide hormones have been documented not to be dissociated from their receptor in the endosome compartment, the hormone and receptor are delivered to the lysosomes via fusion of the endosomes with lysosomes, where both hormone and receptor are degraded [24,27]. The continuous degradation of the receptor with each round of RME eventually leads to a decrease in the number of receptors o n the cell surface, a phenomenon called down-regulation. The distinction between a given receptor being recycled versus degraded is not always an all-or-none phenomenon. In fact, in many cases both processes occur to different degrees. Thus, even though the majority of receptors may be recycled, each round of endocytosis can result in the degradation of a small percentage of receptors. If the rate of synthesis of new receptors plus the rate of recycling of internalized receptors is slower than the rate at which receptors are degraded with each round of RME, there will eventually be a down-regulation of the cell surface receptors. Another factor to be taken into account is that some receptors may be spared degradation, but they may not be immediately recycled back to the cell surface (i.e., they may be sequestered intracellularly). These possible routes of receptor disappearance and appearance on the cell surface are summarized schematically in Fig. 2. Whether a given hormone receptor is recycled or not during RME depends not only upon which hormone the receptor binds, but also upon the cell type and stage of differentiation of a given cell. Thus, the insulin receptor has been shown to be recycled during RME in rat adipocytes [28,29], but not in lymphocytes [30]; and it is down-regulated in the adult rat liver [31], but not in the fetal rat liver [31].

137

(intracellular pool) degradation

synthesis

Fig. 2. Possible routes of receptor appearance and disappearance from the cell surface.

3. Methods used to assess receptor-mediated endocytosis 3.1. Morphological approaches

On a light microscopic level it is possible to visualize the binding of fluorescently labelled hormones to intact cells or to visualize the native hormone with fluorescent antibodies [32-351. Using fluorescently labelled hormones, investigators have observed a band of fluorescence defining the circumference of each cell when the binding of the fluorescently labelled hormone to the cells was performed under conditions where internalization was inhibited (such as at 4°C). When the cells were allowed to bind hormone at 4"C, washed to remove unbound hormone, and then incubated at 37°C to allow the surface-bound hormone to be internalized, it was possible to observe a concentration of the fluorescence into small patches on the cell surface and a subsequent increase in diffuse fluorescence located inside the cell. This experimental approach is powerful in that it allows one to visually determine whether under different conditions a hormone is bound to the cell surface or is internalized, and therefore it has been widely used. In order to identify the particular organelles with which the internalized hormone becomes associated, however, it is necessary to examine the cells using an electron microscope. Using electron microscopy, one can 'follow' the fate of a given peptide hormone in its target cell by using preparations of hormone that have been coupled to electron dense particles, such as ferritin or colloidal gold; or by using hormone preparations that have been radiolabelled to a high specific activity (typically with 12sI) and performing autoradiography [8,9,11]. Alternatively, one can bind the unaltered hormone to the cell, prepare the sample for electron microscopy and then bind an electron-dense anti-hormone antibody to the sample to visualize the hormone [23].The latter approach is generally preferable in that one need not be concerned that the electron dense or radiolabelled hormone is handled by the cell differently than the native hormone. Since colloidal gold is available in a range of sizes, if an antibody to the receptor is available (that can recognize the receptor even when hormone is bound to it), then by using an anti-receptor antibody coupled to colloidal gold of one diameter and an anti-hormone antibody couple to colloidal gold of a different diameter, one can simultaneously follow the fate of both the hormone and the receptor during RME [ 2 3 ] .

138 Using these approaches, it has been found that when the binding of the hormone to the cells is done at 4°C the hormone is associated with the plasma membrane only. If the binding is done at 4"C, the cells washed to remove unbound hormone and then subsequently warmed to 3TC, there is a decrease in the cell surface-bound hormone and a concomitant increase in intracellular hormone. By morphological appearances and by enzymatic or immunological staining, it is possible to identify the intracellular compartments with which the hormone is associated. Furthermore, if one uses a hormone (or antibody to the hormone) made electron dense, the resolution is usually fine enough that one can assess whether the hormone is associated with the organelle membrane (and thus probably receptor-bound) or is free in the lumen [lo]. Typically, a morphometric analysis is performed where a large number of micrographs taken at each time point are examined and the number of grains of ferritin or gold particles associated with a given cellular organelle (plasma membrane, coated vesicle, endosome, lysosome, etc.) are tabulated. As such, one can calculate the percentage of grains or particles associated with a given organelle at each time point and arrive at a statistically valid conclusion as to the route of the hormone (and/or receptor) during RME [9,11,36].

3.2. Biochemical approaches

In order to study the RME of a peptide hormone biochemically, it is necessary to be able to radiolabel the hormone to a high specific activity with Iz5I, while retaining the normal binding and biological properties of the hormone. As discussed in the introduction, if one binds the iodinated hormone to intact cells at 37°C and detects ligand degradation products in the medium, that is an indication that RME of the hormone may be occurring. Ligand degradation can be ascertained by analyzing the molecular size of the radioactive products by gel filtration or by testing the precipitability of the radioactivity by trichloroacetic acid [37]. Since single amino acids and small peptides are not precipitable by trichloroacetic acid, the percentage of acid-soluble radioactivity in the medium represents the percentage of degraded ligand. It is then necessary to document that the accumulation of acid-soluble radioactivity in the medium is dependent upon the extent of hormone binding, the length of the incubation, and the temperature (such that degradation of the hormone should not be apparent at 4°C). Furthermore, one should be able to inhibit the appearance of acid-soluble radioactivity with metabolic inhibitors (such as NaN,) or with compounds that inhibit the delivery of the hormone to the lysosomes or inhibit lysosomal function (such as leupeptin, NH,Cl, chloroquine, or monensin; see Refs. 38,39). When one measures the amount of hormone bound to an intact cell at 37"C, this represents a sum of surface-bound hormone plus hormone that has since been internalized (and is in an intact or partially degraded form). It should be noted that

139 once an internalized protein has been degraded to free amino acids, these are rapidly released from the cell, and thus are not detected to an appreciable extent within the cell. In order to measure the level of surface-bound versus internalized hormone, it is necessary to develop a method that will quantitatively release the surface-bound hormone. Many peptide hormones can be dissociated from their receptor under conditions of low pH (pH 3 4 ) and thus this has been a commonly used method [6,36]. An advantage of this method is that it is a mild treatment and thus in some cases one can treat the cells with acid to remove the surface-bound hormone and then rebind fresh hormone and observe a cellular response [36]. Another method that is generally applicable is to degrade the surface-bound hormone by adding proteases using conditions that do not lyse the cells or allow penetration of the added enzyme [3]. It should be noted, however, that this treatment may also damage the receptor and thus cannot be used if one wishes to subsequently rebind fresh hormone to the cells. Lastly, a variety of other methods tailored to the binding characteristics of a given ligand have also been used [5,40,41]. With any given treatment, however, it is necessary to document that one is indeed releasing most (or all) of the surface-bound hormone. This can be done by saturating the binding sites of the cell with radiolabelled hormone under conditions where no internalization should occur (such as at 4°C) and then testing if the treatment releases all the cell-associated radioactivity. Thus, by measuring total cell-associated radioactivity in one set of cells and releasable radioactivity in another set of cells, one can calculate the amount of internalized hormone by subtraction. Therefore, one can in fact measure hormone binding to an intact cell at 37°C and construct a time course of cell surface-bound hormone, internalized hormone and degraded hormone. Under these conditions, cells are continuously exposed to hormone in the medium and thus are undergoing many rounds of RME. If the internalized receptor is not recycled back to the cell surface, the cell surface receptor will become down-regulated. A schematic example of a time course of hormone binding and internalization to intact cells where the receptor is down-regulated is shown in Fig. 3A. In contrast, Fig. 3B depicts a representation of such a time course when the internalized receptor is not down-regulated. It should be pointed out that the maximal amount of hormone internalized andlor degraded will vary depending upon the extent of receptor recycling. Thus, if the receptors do not recycle, the maximal amount of hormone internalized and/or degraded should be less than or equal to the number of cell surface receptors. Therefore, the amount of hormone that is processed in this case is dictated by the number of hormone receptors. If the receptors do recycle, then the maximal amount of hormone internalized and/or degraded should exceed the number of cell surface receptors. In this case the cells can theoretically degrade all the added hormone, regardless of the number of hormone receptors. From the biochemical approaches discussed thus far, one can conclude that a given hormone may be internalized by RME. Conclusive evidence for such internalization, however, can only be obtained by concurrent morphological data as described

140

W

A

2

0

5

surface-bound

I

K

0

rI

--

m N

L E N G T H O F INCUBATION ( h o u r s )

Fig. 3. Distribution of hormone bound to cells during many rounds of endocytosis. Cells are incubated with radiolabelled hormone for increasing lengths of time at 37°C. At various time points, hormone that is surface-bound, internalized o r degraded and released into the medium is determined as described in the text. Panel A represents a case where the cell surface receptor becomes down-regulated; Panel B represents a case where the cell surface receptor is not down-regulated, but instead is recycled.

above. For data on the rate of internalization of the hormone and possible downregulation and/or recycling of the receptor one must again use biochemical approaches. A commonly used method to calculate the rate of internalization of a hormone is to bind the hormone to intact cells at 4°C (where no internalization should occur), wash the cells to remove unbound hormone and then measure the amount of surface-bound radioactivity remaining as a function of time after warming the cells. Unlike the experimental approach described above where the cells are allowed to continuously bind and internalize hormone at 37°C and thus undergo many rounds of RME (see Fig. 3 ) , under these conditions the cells are internalizing only the prebound hormone and thus are undergoing only one round of RME. A schematic example of results of this kind of experiment is shown in Fig. 4. Typically, one observes a loss of surface-bound radioactivity with a concomitant increase in the levels of internalized radioactivity. Since the internalized hormone is degraded, the levels of internalized radioactivity subsequently decline and there is an increase in the levels of degradation products in the medium. It should be noted that when these experiments are done it is difficult to detect a lag in the appearance of the internalized radioactivity; however, there is a lag in the appearance of degradation products in the medium [36]. This lag is a composite of the rate of accumulation of hormone in the lysosomes, the rate of hormone degradation and the rate of release of degradation products. Among these processes, the rate of hormone degradation appears to be limiting [36]. The use of the loss of cell surface-bound radioactivity as a measure of the rate of hormone internalization is valid, though, only when there is little or no dissociation of the hormone from the receptor during the 37°C incubation (which can be assessed by the appearance of trichloroacetic acid-insoluble radioactivity in the medium). Otherwise, the rate of loss of surface-bound hormone would reflect both the rate of internalization of receptor-bound hormone and the rate of dissociation of the hormone from the cell surface receptor [42].

141

8

TIME AFTER W A R M I N G ( m i n u t e s )

Fig. 4. Distribution of hormone bound to cells during one round of endocytosis. Cells are incubated with hormone at 4°C to saturate the cell surface receptor. At f = O , the cells are washed to remove unbound hormone and warmed to 37°C. At various time points after warming, the hormone that is surface-bound, internalized, o r degraded and released into the medium is determined as described in the text.

A more valid approach to calculating the rate of internalization of a hormone is to use a steady-state approach as originally described by Wiley and Cunningham [4345]. In this approach, the cells are incubated with the hormone at 37°C under conditions where the cells undergo many rounds of RME as they continuously bind and internalize hormone (c.f., Fig. 3). Under these experimental conditions, the rate of internalization can be calculated from the ratio of internalized to surfacebound hormone provided that (i) the time course chosen is shorter than the observed lag of appearance of degradation products in the medium (see above); and (ii) the level of surface-bound radioactivity is at a steady state [43]. Although the first of these two criteria must always be met, one can also perform this experiment while the surface-bound radioactivity is approaching a steady state. If this is done, however, the rate of internalization is calculated from t h e ratio of the internalized radioactivity (which is, by definition, an integral since the experiment is done before any degradation products are released into the medium) versus the integral of the surface bound radioactivity [42,44,45]. In addition to the rate of internalization, the steady-state analysis described by Wiley and Cunningham allows one to calculate many other parameters pertaining to the hormone-receptor interaction during RME. These include the steady-state association constant for the hormone-receptor complex (a steady state equivalent of the K , calculated by Scatchard analysis), the number of cell surface receptors, the rate of receptor appearance at the cell surface, the rate constant for the internalization of occupied receptors, the rate constant for the internalization of unoccupied receptors and the rate constant for the degradation of the internalized hormone. Furthermore, if the receptor for the hor-

142 mone is down-regulated, one can use this steady-state model to determine if the down-regulation is due to an increase in the rate of internalization of occupied versus unoccupied receptors or to a decrease in the rate of appearance of receptors on the cell surface. Using this approach, both hCG and EGF have been shown to downregulate their respective receptors by increasing the rate of internalization of the occupied versus the unoccupied receptor [44,46]. Another important aspect of RME that one would want to determine is the intracellular route of the hormone and receptor. As discussed above, one can use morphological approaches to address this question. The morphological approach is particularly elegant if one has an antibody to the receptor such that one can simultaneously detect both the hormone and its receptor. One can, however, also address this question biochemically. Indeed, it is possible to fractionate cell extracts on Percoll gradients into fractions composed primarily of plasma membrane, endosomes or lysosomes [24,47,48]. Thus, one can bind radiolabelled hormone to the cells, allow the cells to internalize the hormone for a given length of time, and then fractionate the cells to determine in which intracellular compartment the hormone (i.e., radioactivity) is located. One can determine whether the internalized hormone is free or receptor-bound by precipitation of the internalized radioactivity by polyethylene glycol or ammonium sulfate [24,49]. Furthermore, by analyzing the internalized radioactivity in the different compartments on SDS-polyacrylamide gels, one can assess whether the hormone is intact or partially degraded [24,48]. Thus, one can determine in which compartment the hormone dissociates from its receptor and in which compartment degradation of the hormone occurs. Using these tools, it has been possible to document that unlike many other hormones which dissociate from their receptor in the endosome, hCG remains receptor bound. Thus, the hCGreceptor complex is delivered to the lysosome intact, whereupon the complex is dissociated [24]. Although it can only be directly ascertained that the hormone is then degraded (since it is the hormone which is radiolabelled), it is assumed that delivery of the hCG receptor to the lysosome also results in its degradation (which is consistent with the down-regulation of the hCG receptor in these cells). Another frequently used tool to assess the intracellular route of internalized hormones and their receptors is the use of compounds or conditions that allow hormone binding and internalization to occur, but impede the intracellular route of the internalized hormone receptor. By using an inhibitor of lysosomal enzymes, such as leupeptin, one can ‘trap’ undegraded hormone (and possibly receptor) in the lysosome [24]. By performing the experiment at 16-20°C, it is possible to internalize receptor-bound hormone, but ‘trap’ it in the endosome compartment [5O]. Other compounds such as monensin and NH4CI can be used to raise the pH in intracellular organelles [39]. Unfortunately, since pH gradients exist in both endosomes and lysosomes (and other intracellular organelles), these compounds may impede any one (or many) of the steps in the transit of the hormone and receptor, and thus one must use additional approaches (as outlined above) to determine in which organelle

143 the hormone (or receptor) has been trapped. Thus, although it has been shown that in many cases monensin and NH,CI trap the ligand-receptor complex in the endosomes [47,48,51], it has been documented that in murine Leydig tumor cells these compounds allow the delivery of hCG (bound to its receptor) from the endosome to the lysosome but inhibit the subsequent dissociation of hCG from its receptor and degradation of the hormone [24]. Once the hormone-receptor complex has been internalized, the receptor may be degraded, sequestered intracellularly, and/or recycled back to the cell surface. If the receptor were sequestered intracellularly, then one should be able to allow cells to internalize hormone and then detect a pool of intracellular receptors in a detergent extract of the cells. To do this, one would allow the cells to bind and internalize unlabelled hormone and then measure the binding of radiolabelled hormone to the intact cells versus a detergent extract of the cells (where both the cells and extract have been treated with acid to remove the unlabelled hormone prior to adding the radiolabelled hormone). Since the detergent extract would be composed of both cell surface and intracellular receptors, an increase in binding activity of the detergent extract and a decrease in binding to the intact cells would be indicative that the receptors internalized during RME of the unlabelled hormone were being sequestered intracellularly . Alternatively, if one detected a decrease in the binding activity in the intact cell and in the detergent extract, this would indicate that the internalized receptor was being degraded (or inactivated). Indications that receptor recycling may be occurring are (i) if the level of cell surface binding and internalized hormone attain a steady state instead of decreasing with increasing time of exposure to the hormone; (ii) if the amount of hormone degraded over a period of time far exceeds the steady-state level of surface-bound and internalized hormone (c.f., Fig. 3B). To document receptor recycling, one typically allows the cells to bind and internalize unlabelled hormone, washes the cells free of unbound hormone, and then proteolyzes the cells to destroy the cell surface receptors. The rapid reappearance of cell surface receptors (as measured by binding of radiolabelled hormone), especially if observed under conditions where de novo synthesis of new receptors has been inhibited, is suggestive of receptor recycling [29]. One possible explanation for these data other than receptor recycling, however, is that there may exist a preformed intracellular pool of receptors which can be rapidly mobilized to the cell surface. Whether such an intracellular pool exists can be determined by comparing the binding activity of intact cells that had not been incubated with hormone versus a detergent extract of similarly unexposed cells [52]. It must be stressed again that, even if receptor recycling is documented, a hormone receptor may still become down-regulated if with each round of RME there is some receptor degradation also (the rate of which exceeds the rate of de novo synthesis). The relative degrees of recycling versus degradation of the receptor will determine the rapidity with which down-regulation occurs.

144

4. Biological consequences of receptor-mediated endocytosis It has become clear in recent years that different peptide hormones may use different signal transduction systems (a few examples being the activation of adenylate cyclase, the stimulation of protein kinase C, the stimulation of the breakdown of polyphosphoinositides or, as in the case of insulin and EGF, the stimulation of the receptor tyrosine kinase) to translate the binding of the hormone into a cellular response. Furthermore, a given hormone may activate more than one transducing system. It is becoming increasingly apparent, however, that many of these transducing systems are stimulated upon hormone binding to its receptor via the intermediary action of GTP-binding regulatory proteins (called G proteins), which are associated with the plasma membrane. Thus, in most (or all) cases, the process of signal transduction appears to occur at the cell surface. What then is the role of RME in the stimulation of a cellular pathway in response to hormone binding? 4.1. Microaggregation

One of the first consequences of hormone binding to receptors that are not already located in coated pits is a microaggregation of the hormone-receptor complexes (two to four complexes per group) [19]. In a classic study by Kahn and co-workers [53] it was found that certain antibodies to the insulin receptor were able to mimic the actions of insulin in that they stimulated glucose oxidation when added to adipocytes. Interestingly, although bivalent F(ab), antibody fragments retained this insulin-like action, monovalent F(ab) antibody fragments were not stimulatory. However, when secondary bivalent IgGs directed towards the monovalent F(ab) fragments were subsequently added, a cellular response was observed. These results suggested that sheer occupancy of the insulin receptor was not sufficient to elicit a response, and that a microaggregation of the receptors (caused in this instance either by cross-linking receptors with bivalent anti-receptor antibodies or by crosslinking the monovalent of the F(ab)-receptor complexes amongst each other by the bivalent anti-F(ab) antibodies) was necessary to provoke the cellular response. These studies were also significant in that they suggested that the ‘information’ for evoking a cellular response lies not in the hormone, but in the receptor; and that the role of the hormone is to confer a given configuration to the receptor that allows it to transduce the signal. Similar kinds of studies with E G F [34,35], GnRH [54], prolactin [55] and LH [32,33] have also suggested an important role for microaggregation in signal transduction. This is intriguing in light of the fact that the structures and functions of some of these receptors are quite different. Thus, although the EGF and insulin receptors are known to contain intrinsic tyrosine protein kinase activity, no such activity has been described for the LH, prolactin or GnRH receptors. Furthermore, although both the L H and the GnRH receptor interact with G proteins. it is not yet clear if either the E G F or insulin receptors also do so. Clearly more will

145 have to be learned about the precise mechanisms by which microaggregation ‘stimulates’ these different peptide hormone receptors in order to better understand its role in hormonal stimulation.

4.2. Internalized and degraded hormone As discussed above, once the hormone-receptor complex has been sequestered into the coated pits of the plasma membrane, it is rapidly internalized, and the hormone is ultimately degraded in the lysosome. It has long been debated (and still is) whether the internalized and/or degraded hormone has a role in provoking a cellular response. The question as to the role of the degraded hormone has been easier to address in that it has been possible in some cases to inhibit the degradation of the hormone (using inhibitors such as NH,CI and chloroquine) without inhibiting the hormone-provoked biological response [56].Thus, it appears that degradation of the hormone per se is probably not necessary to provoke a cellular response. The question as to whether the internalized hormone has a biological role has not been as easy to answer, however, for technical reasons. Thus, the compounds typically used to block internalization (such as metabolic inhibitors) have general toxic effects on the cells; and although 4°C conditions inhibit internalization, these conditions also ‘slow’ the biological responses one is measuring. Another means of addressing this question has been to immobilize the hormone onto beads that cannot be internalized and to ask whether the hormone is able to provoke a response. A drawback to this approach, however, is that given positive results, it is difficult to rule out the possibility that free hormone has ‘leaked’ off the resin and that it is the free hormone (able to be internalized) which is eliciting the response. Given this caveat, however, this approach has been used by some investigators and their results suggest that immobilized hormone is indeed capable of provoking the appropriate response (see Ref. 57 and references therein). Other indirect methods have also been used to investigate the role of the internalized hormone. One approach has been to ask what happens to the cellular response when the surface-bound hormone is removed or neutralized. Thus, it was found that if one removes surface-bound LH or hCG from Leydig cells, even after a significant amount of hormone has been internalized, there is a rapid cessation of stimulated CAMP and steroid production [58,59]. Similarly, it was found that the addition of neutralizing antibodies to E G F up to 8 h after the addition of E GF (at which time a considerable amount of E G F has been internalized and degraded) prevented the EGF-induced increase in DNA synthesis in human fibroblasts [6O,61]. Thus, these kinds of studies have also suggested that it is the hormone-receptor complex on the cell surface that provokes the cellular response, and that the internalization of the complex (with the eventual degradation of the hormone) serves as a means of terminating the response. For those receptors that do interact with G proteins (such as those that stimulate

146 adenylate cyclase), internalization of the hormone-receptor complex may be a means of segregating the hormone-receptor from the G proteins (if the G proteins are not co-internalized), and thus terminating the response. Furthermore, even if the G proteins were co-internalized, they would now be facing the lumen of the endocytic vesicle and may not have access to the appropriate substrates. Thus, although earlier events may functionally ‘uncouple’ the receptor-G protein interactions, internalization of the hormone-receptor may provide a ‘failsafe’ mechanism for terminating the hormonal response arising from interactions with G proteins. There is recent data to suggest that there may in fact be a biological role for the internalized insulin and E G F receptors (both of which are themselves tyrosine kinases). Thus, microinjection of insulin-occupied insulin receptors into Xenopus oocytes causes the increased phosphorylation of ribosomal protein S6 (a known substrate for the insulin receptor/kinase) [62]; and the E GF receptor in endocytic vesicles has been shown to retain its kinase activity [63]. Whether the internalized insulin receptodkinase or E G F receptorikinase has a physiological role or not is as yet unknown. Clearly, though, these data suggest that there is much more to be learned about the role of internalized hormone-receptor complexes, especially those where the receptor possesses intrinsic enzymatic activity.

4.3. Receptor down-regulation Although the hormone internalized by RME is ultimately degraded in the lysosomes, the internalized receptor may be sequestered intracellularly, recycled back to the cell surface or delivered to the lysosomes and degraded. A consequence of the sequestration of the receptor or the degradation of the internalized receptor is the ‘down-regulation’ of the cell surface receptor. As discussed above, even if most of the receptor is recycled, if some is degraded with each round of RME, there will still eventually be a down-regulation of the cell surface receptor. There are two possible consequences of receptor down-regulation. If there are ‘spare’ receptors (i.e., if only a small percentage of the cell surface receptors need to be occupied in order to provoke a response), then a decrease in the number of cell surface receptors will not decrease the maximal possible response, but it will decrease the sensitivity of the cell such that a higher concentration of hormone will be needed to evoke a half-maximal response. If there are few or no spare receptors, then downregulation of the cell surface receptors will lead to both a decreased sensitivity to hormone and to a decreased maximal response by the cell [64]. The down-regulation of the hormone receptor may therefore constitute one mechanism (usually of many) by which the cell becomes refractory (i.e., less responsive) to a continuous hormonal challenge or to a re-challenge of hormone.

147

5. Conclusion In summary, once peptide hormones are bound to their receptors on the appropriate target cells, they are internalized and degraded by the process of RME. It is hoped that this review has provided a general overview of the salient features of this pathway and how it is related to the biological actions of peptide hormones. It cannot be overstressed, however, that within this general pathway there are many variatians on the possible routes and fates of the internalized hormone and receptor. Thus, whether a given hormone receptor is down-regulated or not depends not only upon the hormone, but upon the particular cell types that have receptors for this hormone. Furthermore, a given cell may both degrade and recycle internalized receptors. Whether the cell surface receptor is down-regulated in this case would depend upon the quantitative contributions of each of these routes. Although there is as yet no compelling evidence to suggest that internalized peptide hormones and/or their receptors have a role in the signal transduction processes involved in the hormonal cellular responses, it cannot categorically be ruled out that such a possibility exists. T o date, however, most of the studies done on the RME of peptide hormones suggest that the internalization of the hormone serves as a mechanism for terminating the actions of the hormone. Ongoing studies in many laboratories are still aimed at gathering more information on the role of RME in hormone action.

References 1. 2. 3. 4.

5. 6.

7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17.

Brown, M.S. and Goldstein, J.L. (1979) Proc. Natl. Acad. Sci. U.S.A. 76. 3330-3337. Carpenter, G . , Lembach, K.J., Morrison, M.M. and Cohen. S . (1975) J . Biol. Chem. 250.4297-4304. Carpenter, G. and Cohen, S. (1976) J . Cell Biol. 71, IS-171. Goldstein. J.L., Brunschede, G.Y. and Brown. M.S. (1975) J . Biol. Chem. 250, 7854-7862. Goldstein. J.L.. Basu. S.K.. Brunschedc. G . Y . and Brown. M.S. (1976) Cell 7, 85-95. Haigler. H . . Maxfield. F.R., Willingham. M.C. and Pastan, I . (1980) J . B i d . Chem. 255, 123Ck1241. Haigler, H.T.. Willingham, M.C. and Pastan. I . (1980) Biochem. Biophys. Res. Commun. 94. 63CL637. Anderson, R.G.W., Brown, M.S. and Goldstein. J.L. (1977) Cell 10. 351-354. Haigler. H., McKanna, J . A . and Cohen. S. (1979) J . Cell Biol. 81, 382-395. McKanna. J.A.. Haigler, H.T. and Cohcn. S. ( 1979) Proc. Natl. Acad. Sci. U.S.A. 76,5689-5693. Gorden. P., Carpentier. J.-L., Cohen. S. and Orci, L. (1978) Proc. Natl. Acad. Sci. U.S.A. 75. 5025-5029. Goldstein. J.L.. Anderson, G . W . and Brown. M.S. (1979) Nature 279. 679-685. Mellman. I . . Fuchs, R . and Helenius. A . (1986) Ann. Rev. Biochem. 5 5 , 663-700. Steinman. R . M . . Mellman, I.S.. Muller. W.A. and Cohn. Z . A . (1983) J . Cell Biol. 96. 1-27. Helenius. A . . Mellman, I . , Wall. D. and Huhhard. A. (1983) Trends Biochem. Sci. 8, 245-250. Limbird, L.E. (1986) Cell Surface Receptors: A Short Course on Theory and Methods. Ch. 6, pp. 159-194. Martinus Nijhoff Publishing. Boston. Fawcett, D.W. (1965) Cytochemistry 13, 75-91.

148 18. 19. 20. 21. 22. 23. 24. 25.

Roth, T.F. and Porter, K.R. (1964) J . Cell Biol. 20. 313-332. Schlessinger, J. (1980) Trends Biochem. Sci. 5. 21C-214. Heuser, J. (1980) J . Cell Biol. 84, 56G-583. Ohkuma, S. and Poole. B. (1978) Proc. Natl. Acad. Sci. U.S.A. 75, 3327-3331. Tycko, B. and Maxfield, F.R. (1982) Cell 28, 643-651. Geuze, H.J., Slot, J.W., Strous, F.J.A.M.. Lodish, H.F. and Schwartz, A.L. (1983) Cell 32,277-287. Ascoli, M. (1984) J . Cell Biol. 99. 1242-1250. Dautry-Varsat, A . , Ciechanover. A . and Lodish, H.L. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 2258-2262. 26. Klausner, R . D . , Ashwell. G., van Renswoude. J., Harford, J.B. and Bridges. K.R. (1983) Proc. Natl. Acad. Sci. U.S.A. 80. 2263-2266. 27. Stoscheck, C.M. and Carpenter, G . (1984) J . Cell Biol. 98, 1048-1053. 28. Marshall, S. (1985) J . Biol. Chem. 260, 41364144. 29. Arsenis, G., Hayes, G.R. and Livingston, J.N. (1985) J. Biol. Chem. 260, 2202-2207. 30. Kosmakow, F.C. and Roth, J . (1980) J . Biol. Chem. 255, 986(k9869. 31. Caliendo, A.M. and Patel, M.S. (1983) Arch. Biochem. Biophys. 227, 552-561. 32. Amsterdam, A , , Berkowitz, A., Nimrod. A. and Kohen, F. (1980) Proc. Natl. Acad. Sci. U.S.A. 77. 344C-3444. 33. Podesta, I.J., Solano, A.R.. Attar, R . , Sanchez. M.L. and Vedia, L.M.Y. (1983) Proc. Natl. Acad. Sci. U.S.A. SO, 3986-3990. 34. Schrieber, A . B . , Libermann, T . A . , Lax, I . , Yarden. Y. and Schlessinger, J. (1983) J . Biol. Chem. 258. 846853. 35. Schechter. Y . ,Hernaez, L.. Schlessinger, J . and Cuatrecasas, P. (1979) Nature 278, 835-838. 36, Ascoli, M. (1982) J . Biol. Chem. 257, 1330613311. 37. Ascoli, M. and Puett, D. (1978) J. Biol. Chem. 253, 4892-4899. 38. Ascoli. M. and Puett, D . (1978) J . Biol. Chem. 253, 7832-7838. 39. Maxfield, F . R . (1982) J. Ccll Biol. 95, 676-hX1. 40. Ashwell, G. and Harford, J. (1982) Ann. Rev. Biochem. 51, 531-554. 41. Gonzalez-Noriega. A.. Grubh. J.H., Talkad. V. and Sly. W.S. (1980) J . Cell Biol. 85, 839-852. 42. Ascoli. M. and Segaloff, D.L., (1987) Endocrinology 120, 1161-1172. 43. Wiley, H.S. and Cunningham, D.D. (1981) Cell 25. 433-440. 44. Wiley. H.S. and Cunningham, D.D. (1982) J . Biol. Chem. 257, 4222-4229. 45. Wiley. H.S., VanNostrand, W., McKinley, D.N. and Cunningham, D.D. (1985) J . Biol. Chem. 260. 5290-5295. 46. Lloyd. C. and Ascoli, M. (1983) J . Cell Biol. 96. 521-526. 47. Merion. M. and Sly. W.S. (1983) J . Biol. Chem. 96, 644650. 48. Harford, J . , Bridges, K., Ashwell, G . and Klausner. R.D. (1983) J. Biol. Chem. 285, 3191-3197. 49. Bridges. K . , Harford. J . . Ashwell, G. and Klausner, R . D . (1982) Proc. Natl. Acad. Sci. U.S.A. 79. 35C-354. 50. Dunn. W.A., Hubbard, A.L. and Aronson. N.N. (1980) J . Biol. Chem. 255, 5971-5978. 51. Harford, J.. Wolkoff, A.W., Ashwell. G . and Klausner, R.D. (1983) J. Cell Biol. 96, 1824-1828. 52. Deutsch. P.J.. Rosen. O.M. and Rubin. C.S. (1982) J . Biol. Chem. 257, 535C-5358. 53. Kahn, C.R.. Baird, K.L., Jarrett, D.B. and Flier, J.S. (1978) Proc. Natl. Acad. Sci. U.S.A. 75, 4209-42 13. 54. Conn. P.M.. Rogers, D.C.. Stewart, J.M., Niedel. J . and Sheffield, T. (1982) Nature 296, 653-655. 55. Dusanter-Fourt. I.. Djiane, J.. Kelly, P.A.. Houdebine. L.-M. and Teyssot, B. (1984) Endocrinology 114. 1021-1027. 56. Ascoli, M. (1978) J . Biol. Chem. 253, 7839-7843. 57. Conn. P.M., Smith, R.G. and Rogers, D.C. (1981) J . Biol. Chem. 256, 1098-1100. 58. Segaloff, D.L. and Ascoli, M. (1981) J. Biol. Chcm. 2.56, 1142&11423.

149 Segaloff, D.L., Puett, D. and Ascoli. M. (1081) Endocrinology 108, 632-638. Haigler. H.T. and Carpenter, G. (1980) Biochirn. Biophys. Acta 598, 314325. Schechter, Y., Hernaez, L. and Cuatrecasas. P. (1978) Proc. Natl. Acad. Sci. U.S.A. 75, 5788-5791. Maller. J.L., Pike, L.J., Freidenberg, G.R.. Cordera, R., Stith, B.J., Olefsky, J.M. and Krebs. E.G. (1986) Nature 320, 459-461. 63. Sawyer, S.T. and Cohen. S. (1985) J . Biol. Chern. 260. 8233-8236. 64. Ascoli. M. (1985) In: Luteinizing Hormone Action and Receptors. Ch. 6 (Ascoli, M., ed.) pp. 199-217. CRC Press, Boca Raton.

59. 60. 61. 62.

This Page Intentionally Left Blank

B.A. Cooke. R.J.B. King and H.J. van der Molen (ed\.) Hormones and (heir Actions. Part I 01988 Elsevier Science Publishers BV (Biomedical Division)

151 CHAPTER 10

Physiological aspects of luteinizing hormone releasing factor and sex steroid actions: the interrelationship of agonist and antagonist activities A.E . WAKELING Research Department I , Imperial Chemical Industries PLC, Pharmaceuticals Division, Mereside, Alderley Park, Macclesfield, Cheshire SKlO 4TG, England

1. Introduction Classical endocrinology has defined the nature and interrelationships of two major classes of hormones, the peptides and steroids, and has elucidated the physiological mechanisms which control their secretion. Neuroendocrine surveillance of circulating hormone concentrations leads to changes in the secretion of hypothalamic peptides (the releasing factors) which control the activity of the anterior pituitary gland. The pituitary peptide hormones, in turn, modulate the production of steroid hormones by the gonads and adrenal glands to complete the homeostatic loop. The steroid hormones themselves produce extremely diverse physiological effects ranging from the regulation of blood pressure to control of secondary sexual characteristics. The predominant cellular characteristic which provides a common denominator in mechanisms for the control of the synthesis, secretion and action of both peptide and steroid hormones, is the presence or absence of specific hormone receptors. Cells which respond to peptide hormones are distinguished from unresponsive cells by the presence of specific, high-affinity receptors at the cell surface. Similarly steroid-responsive cells possess intra-cellular high affinity receptors. Analysis of the structure and function of receptors. in particular the application of molecular biological techniques, has begun to reveal the complexity of the events linking receptor occupancy and physiological effects. In the particular case of steroid hormones, formation of the steroid-receptor complex facilitates recognition of specific DNA sequences by the receptor and subsequent activation of the transcription of hormone-regulated genes. For peptide hormones coupling of receptor occupancy and

152 biological response is achieved indirectly by a spectrum of mechanisms, often involving the formation of second messengers as the first step in a complex cascade of events. The peptide hormones may serve as a model for the cellular actions of growth factors which are thought to function as autocrine or paracrine regulators of both normal and abnormal growth and differentiation. Furthermore, there is increasing evidence that steroid hormone modulation of growth factor synthesis provides another important link in physiological growth control. Molecular mechanisms of hormone action are considered in detail elsewhere in this volume and are not further exemplified here. The focus of attention in this chapter is directed towards the exploration of the physiological effects of modulation of hormone secretion and action. In particular, the utility of steroid hormone antagonists and of peptide hormone agonists in exploring fundamental biological mechanisms is considered. It is to be hoped that the reader will recognize the continuing importance of discovering novel means of antagonizing hormone action, not only for therapeutic applications. but in providing unique tools to bridge the gulf which sometimes exists between those interested in hormone action at the macroscopic and molecular levels. Detailed discussion is confined to luteinizing hormone releasing hormone (LHRH) analogues to exemplify peptide agonists and to antiandrogens and antioestrogens t o exemplify steroid antagonists. A more extensive discussion of inhibitors of hormone action and secretion is available [l].

2. L H R H and L H R H analogues 2. I. Physiology LHRH is one of several neuropeptides secreted by the hypothalamus which are transported directly to the anterior pituitary gland by the hypothalamo-hypophysialportal system (see Fig. 1).Secretion of both luteinizing hormone (LH) and folliclestimulating hormone (FSH) by specific pituitary cells, the gonadotrophes, is directly stimulated by LHRH binding to specific membrane receptors on these cells. In the male, FSH stimulates spermatogenesis and LH stimulates androgen synthesis by the testicular Leydig cells. In the female, FSH stimulates ovarian follicular development and oestrogen secretion and L H triggers ovulation, corpus luteum formation and progesterone synthesis. The gonadal steroids in turn modulate the activity of the hypothalamic-pituitary system by both negative and positive feedback. In general, there is a reciprocal relationship between serum steroid concentration and L H and FSH secretion. The most notable exception is the ovulatory surge of L H which is thought to depend on increased follicular oestrogen secretion. Reproductive function is thus dynamically controlled by the activity of the hypothalamic-pituitary-gonadal axis. Clearly, LHRH plays a role of pivotal importance and this is reflected by the enormous growth of the literature following the isolation and

153

HYPOTHALAMUS

0

(Hypophysial/portal blood)

MALE

1.

LH

1

FSH

I:a"^y

(Peripheral blood)

SEMINAL VESICLES

u UTERUS BREAST

Fig. 1. Hormone secretion and control of the hypothalamic-pituitary-gonadal axis. Plus and minus signs indicate stimulation and inhibition respectively. In both male and femalc animals the dominant feedback effect of gonadal steroids on the hypothalainic-pituitary axis is inhibitory. The only major exception is in the female in which a rapid FSH-induced increase in the concentration of oestradiol triggers a large increase of the secretion of L H to induce ovulation. Note that L H R H secreted by the hypothalamus is carried directly to the pituitary gland without entering the peripheral circulation. Since the pituitary and gonadal hormones are secreted into the peripheral blood stream the functional response to hormone agonists or antagonists may be assessed by simple measurements of circulating hormone concentration (s) .

characterization of LHRH in 1971 [ 2 , 3 ] .In a review published in the same year Schally and Kastin [4] predicted the development of potent LHRH agonists and antagonists and foresaw their clinical application to the control of fertility, and to the treatment of endocrine disorders. The predictions have been in large part fulfilled but, as yet, the clinical application of LHRH analogues is restricted to the use of LHRH agonists in the therapy of prostate carcinoma. The most important physiological feature of LHRH secretion is its pulsatile nature. Episodic secretion of LHRH occurs at intervals of 20-60 min, leading, in turn, to episodic release of LH and FSH from the pituitary gland. Both the frequency and amplitude of LHRH pulses vary according to the physiological state of the animal. The origin and nature of the hypothalamic LHRH pulse generator are unknown (see Ref. 5 and references therein for review). Disruption of this normal pulsatile pattern by the chronic administration of exogenous LHRH or LHRH analogues has profound effects on the pituitary gland and consequently upon gonadal steroid secretion [6,7]. Treatment with LHRH leads to a loss of pituitary LHRH receptors [8], a process of down-regulation commonly observed for peptide hor-

154 mones. Many other mechanisms, however, including steroid-induced changes in pituitary LHRH receptor concentration, also contribute to the regulation of pituitary LH secretion [8]. Despite this complexity a number of simple in vitro and in vivo test systems have been described which permit facile comparisons of the potency of LHRH analogues [9,10]. The combination of in vitro effects on LH or FSH secretion by pituitary cultures and the induction of ovulation in female rats allows comparative evaluation of both potency and efficacy. Such studies have been used extensively to document the structure-activity relationships described in the following section. 2.2. Biological activity of LHRH analogues The relatively small size of LHRH, Gln-His-Trp-Ser-Tyr-Gly-Leu-Arg-Pro-GlyNH,, provided an attractive target for medicinal chemistry to modify the natural peptide sequence providing analogues with enhanced physiological properties. A further powerful impetus for the synthesis of highly potent agonists was the relatively short in vivo half-life of the natural peptide. The first report of compounds with a biological potency greater than that of LHRH appeared rapidly after the isolation of LHRH [ 111. The compound [Des-G1y'"-NH,, proethylamidel-LHRH which is 5-fold more potent than LHRH in ovulation induction in the rat, was the prototype for a large number of carboxy-terminal amide superagonists containing nine or ten amino acids. In the 15 years since this compound was described many hundreds of analogues have been synthesized permitting a detailed analysis of structureiactivity relationships (see Refs. 12-14 for reviews). The potent analogues of LHRH share certain characteristic modifications which combine one or more substitutions at positions 6 and 10 whilst retaining the native Gln1-Tyr5sequence. Substitution of Gly6 with a D-amino acid [D-Ala', D-Leu', D-Trp6, D-ser(t-Bu)6, D-His(Bzl)6, ~-Na1(2)"]and of Proy-G1y"'-NH, with a Proyethylamide to form LHRH [ 1-91 nonapeptide ethylamide analogues or Aza-Gly'" in LHRH [l-101 decapeptide analogues produces increases of in vitro and in vivo potency of up to 200-fold. The increased potency of these compounds is due, not only to increased affinity for pituitary LHRH receptors, but also to a prolonged association with the receptor compared to LHRH [12]. Daily administration of potent LHRH agonists initially stimulates LH and FSH release in male and female rats, but continued daily dosing for at least 7 days produces a fall in serum LH and FSH through pituitary desensitization associated with the loss of LHRH receptors. In the longer-term ovarian oestrogen and testicular androgen production declines causing regression of the accessory sex organs equivalent to that following surgical castration. Nicholson and Maynard [15] showed that continuous treatment with the agonist [D-Ser(t-Bu)6Aza-Gly"']-LHRH (goserelin, Zoladex") causes regression of rat mammary tumours, similar to the effect of ovar*Zoladex is a trade mark, property of Imperial Chemical Industries PLC

155 iectomy. This compound and a similar potent analogue [ D - T ~ ~ ~ I - L Halso R Hproduce a castration-like reduction in the growth of Dunning prostate tumours in male rats [16,17]. These studies in animals provided the basis for the therapeutic application of LHRH analogues to prostate cancer and to breast cancer in premenopausal patients [18]. Clinical studies have shown that chronic daily injection of LHRH analogues causes a daily rise of serum LH concentration. This is not, however, accompanied by an increase of serum testosterone concentration which reaches castrate values after one week and remains suppressed throughout treatment [19]. This apparent paradox is explained by a difference between the concentration of bioactive and immunoreactive LH. The former is reduced by more than 90%, closely paralleling serum androgen concentration, whereas the latter decreased by only 50% [20]. Diurnal stimulation of endocrine parameters in patients is entirely eliminated by the use of depot formulations of LHRH analogues designed to release the drug in a constant manner for one month or more [21,22]. This suggests that in patients, as observed in animals, pituitary LHRH receptors are continuously down-regulated with no recovery of receptor number normally associated with once daily dosing [7]. Experimental studies with potent LHRH agonists in animals have revealed the potential for direct stimulatory [23] and inhibitory [24] effects on gonadal steroidogenesis. The significance of these observations remains uncertain because LHRH is normally confined to the hypophysial-portal system and peripheral serum concentration is less than lo-" M. Also, the direct gonadal actions are highly variable between species being, for example, entirely absent in the mouse and probably also in man. These direct actions of LHRH agonists which appear to be mediated by gonadal LHRH-like receptors, may be indicative of the existence of endogenous gonadal LHRH-like peptides. Such peptides may directly affect steroidogenic activity by an autocrine or paracrine mechanism [23]. Reports of tumour remission in postmenopausal breast cancer patients treated with LHRH agonists are also suggestive of possible direct effects on tumour cells [la]. The complexity of the physiological actions of LHRH agonists on LH/FSH secretion and gonadal steroidogenesis presents major difficulties for practical contraceptive applications in both males and females. Many of these problems could, in pri'nciple, be eliminated by the use of LHRH antagonists. Identification of LHRH analogues capable of blocking LHRH-induced LH release in vivo has provided difficult. Although in vitro antagonist activity was first reported for [Des-His2]-LHRH in 1972 [25], the first compound with in vivo activity [Des-His2, Des-Gly"']-LHRH ethylamide was not discovered until 1974 [26]. Since that time more potent antagonists with receptor affinities up to two orders of magnitude greater than that of LHRH have been described. In contrast to the agonists which retain the natural Gln'-Tyr5 sequence, the antagonists contain modifications in the amino terminal region at residues 1, 2 or 3, or combinations thereof, as well as at residues 6 and 10. The development and pharmacology of LHRH antagonists has been reviewed extensively [12,27].

156 The mode of action of LHRH antagonists at the LHRH receptor is quite different from that of the agonists. Although the antagonists bind to LHRH receptors on the pituitary gonadotrophes the receptor down-regulation characteristic of agonist action does not occur [28]. For reasons which remain unclear, microaggregation of receptors occupied by an LHRH antagonist on the plasma membrane, as a prelude to internalization, does not take place [29]. The physiological implications are profound since successful blockade of LHRH action in vivo requires continuous antagonist occupancy of a large proportion of the receptors. That this can be achieved with a single daily injection of potent antagonists has been demonstrated in rats, dogs and monkeys [27]. Clinical application has been delayed by the occurrence of significant toxicity following parenteral administration of many of the most potent LHRH antagonists. Many compounds, particularly those with a D-Arg6 substitution, cause release of inflammatory mediators. The recent discovery [30] of a compound [Ac-A3, Pro', ~ F - D - T ~ ~ " . " ] - L Hwhich R H suppresses serum LH and testosterone in man for at least 18 h after a single subcutaneous injection, without adverse reactions, offers hope that clinical trials of antagonists for both fertility control and cancer therapy can proceed in the near future.

3. Steroid antagonists 3.1. Physiology The modern era of pharmaceutical research on steroid antagonists was founded almost 30 years ago on the observation by Lerner et al. [31] that the non-steroidal compound, MER 25, will antagonize the effect of exogenous of endogenous oestradiol in castrated or intact rats, mice, monkeys, chicks and rabbits. The clearest index of physiological response was provided by the uterus in mice. The compound administered alone provoked a small and transient increase of uterine weight but on chronic administration of high doses, together with oestradiol, almost completely blocked the trophic effect of oestradiol on the uterus. A further interesting observation in these studies was the antagonism by MER 25 of the uterotrophic action of non-steroidal oestrogens like diethylstilbestrol and chlorotrianisene (TACE) in mice. MER 25 was rapidly supplanted by the more potent compound clomiphene (MRL-41). The structures of these compounds are illustrated in Fig. 2. Since both MER 25 and clomiphene are structural homologues of the non-steroidal oestrogen chlorotriphenylethylene, Lerner's observations exemplified for the first time several important features of antioestrogen action which provided a common theme in subsequent studies. These include the conversion of agonists to antagonists by the addition of a basic ether side-chain, the subtle differences between biological responses to closely related structures and the implication from pharmacological studies that the action of these divergent chemical structures are mediated through a common receptor mechanism.

157 OESTROGENS

OH TRIPHENYLCHLOROETHYLENE

CHLOROTRIANISENE CTACE)

DIETHYLSTILBESTROL (DES)

ANTIOESTROGENS

ETHAMOXYTRIPHETOL (MER 25)

CHLOMIPHENE (MRL-41)

Fig. 2. Chemical structures of nonsteroidal oestrogens and the first antioestrogens to be described

Although the antioestrogenic activity of MER 25 in patients was rapidly demonstrated, its poor therapeutic ratio, associated with the occurence of central nervous system disturbances [32] prevented its widespread clinical use. The search for more potent antioestrogens was, however, stimulated by the realization of their potential for treatment of both malignant (breast cancer) and non-malignant (endometriosis, benign breast disease) oestrogen-dependent conditions. The reports that MER 25 had antifertility activity in both rats [33]and rabbits [34] also highlighted the possible use of antioestrogens as contraceptive agents. Similar considerations also provided the impetus for seeking antiandrogens which might fulfil parallel objectives in the male. Thus, although the discovery of the first antiandrogen, cyproterone acetate, was fortuitous (see Ref. 35 for review) the search for novel compounds has continued throughout the succeeding 25 years. The chemical structures of the parent hormones and some examples of antagonists for each of the major classes of steroid hormones are illustrated in Fig. 3. It is important to recognize that each of these compounds blocks hormone action at the receptor level. In in vivo tests for hormonal or antihormonal activity the ap-

158

PARENT STEROID A

OH

@

0

CYPROTERONE ACETATE

TESTOSTERONE

FLUTAMIDE

RU 23908

ICI 176.334

3

OCHzCHzN

I

B OH

HO OESTRADIOL NAFOXlDlNE

TAMOXIFEN

3

OH

OCHzCH2N

CH3 HO @OH /

ICI 164.384 LY 117,018

Fig. 3. Chemical structures and common names of parent steroids and some examples of steroid antagonists.

159 parent inhibition of steroid action may occur by a variety of mechanisms which include interference with gonadotrophin secretion by blockade of central receptors and inhibition of steroidogenic enzymes in the gonads as well as blockade of the receptor at the target tissue of interest (see Fig. 1). For example, the effects of chemical castration, with an LHRH analogue or long-term antiandrogen treatment, on prostate or seminal vesicle weight may be identical. Thus, the unequivocal identification of a receptor antagonist requires tests to exclude alternative modes of action. Such tests would normally include in vitro and/or in vivo inhibition of specific binding of the authentic radiolabelled steroid. Although the compounds shown in Fig. 3 are not intended as an exhaustive listing, two general statements may be made. Firstly, that steroidal antagonists have been described for all classes of steroids except, until recently, for oestrogens (see ICI 164, 384 later) and, secondly, that currently available non-steroidal antagonists are confined to antioestrogens and antiandrogens. Some clinical uses of steroid antagonists are listed in Table I . Note that earlier hopes for the application of the latter two classes of compounds to contraception have not been fulfilled but that maPARENT STEROID 7H3

C

0

ANTAGONISTS

c=o

& PROGESTERONE

RU 486 MIFEPRISTONE

so D

CHzOH

‘@C=CH

ZK 98734

o&zH

CHZOH

IllllOH

0

CORTISOL

0

RU 25,593

E

OH l

CHzOH

CORTEXOLONE

-COOK

0 ALDOSTERONE

Fig. 3 Contd.

SPIRONOLALTONE

SC 14266 CANRENOATE POTASSliJM

160 TABLE I Clinical uses of steroid hormone antagonists Antioestrogens

Breast cancer Benign breast disease Endometriosis Uterine disorders

Antiandrogens

Prostate cancer Benign prostatic hypertrophy Acne

Antiprogestins

Contraception

Antiglucocorticoids

Cushing’s syndrome Adrenocortical carcinoma

Antimineralocorticoids

Essential hypertension Oedemaiascites

jor interest in this respect is now focussed on the progesterone antagonist mifepristone (RU486) [36]. Although RU486 has a high affinity for the progesterone receptor it has a similarly high affinity for the glucocorticoid receptor and is a potent antiglucocorticoid. This does not appear to limit its clinical use but further separation of receptor specificity in analogues of RU486 has been described recently for example ZK97734, including compounds with reversed D-ring configuration (13a-methyl, 17a-hydroxy) [37]. The remainder of the discussion in this section is concerned with antiandrogens and antioestrogens to exemplify the pharmacology of steroid antagonists.

3.2. Antiandrogens Androgens secreted by the testis promote the growth and secretory activity of the male accessory sex-organs, particularly the seminal vesicles and the prostate gland, and are essential for the maintenance of libido and secondary sexual characteristics. Androgen synthesis is dependent on pituitary FSH and LH secretion which in turn is controlled by hypothalamic LHRH secretion. Circulating androgens control hypothalamic/pituitary activity by negative feedback (Fig. 1). It is assumed that both the central and peripheral actions of androgens are mediated by specific androgen receptors. It is thus clear that a non-selective antagonist may be expected to produce effects as diverse as suppression of libido and regression of the prostate gland. Cyproterone acetate (see Fig. 3 for structure) exemplifies this diversity. Although this compound has been used successfully in the treatment of prostate cancer it has also achieved widespread application for the treatment of sexual disorders in men because it suppresses libido [35]. Cyproterone acetate is not a pure antiandrogen:

161 it is a potent progestin and is also antigonadotrophic. The latter property is clinically important because it opposes the tendency towards increased LH secretion as a result of antiandrogen-induced blockade of the central negative feedback action of androgens: this is particularly relevant when compared to the actions of the original non-steroidal antiandrogen, flutamide [38], and of a more recently reported compound, R U 23,908 [39]. Both antiandrogens cause regression of the prostate gland but also lead to increased LH and testosterone secretion. In this laboratory we set out to discover a peripherally selective antiandrogen, that is a compound which produces the desired regression of the prostate gland without blocking central negative feedback. A non-steroidal analogue of flutamide designated, ICI 176,334, with these properties has recently been described [40]. The structures of flutamide, RU23908 and ICI 176,334 are illustrated in Fig. 3. In adult male rats treated for 14 days with ICI 176,334 or flutamide the weight of the seminal vesicles and prostate gland decreased in a dose-dependent manner. Direct comparison of the potency of ICI 176,334 with that of flutamide showed that ICI 176,334 is 5-fold more potent than flutamide [40]. The central antiandrogenic effect of the two compounds was compared in this study by measuring the serum concentration of LH and testosterone. Flutamide significantly increased both LH and testosterone, whereas ICI 176,334 had no effect on either parameter. Thus ICI 176,334, unlike flutamide, is described as a peripherally selective pure antiandrogen. This selectivity appears paradoxical since no tissue-specific differences in the structure or pharmacology of androgen receptors have been described. Similar paradoxes, some of which are referred to in the next section, are even more evident among tissue responses to antioestrogens and have been ascribed to differential gene control by receptor-antihormone complexes [41].

3.3. Antioestrogens The description of the first antioestrogen MER 25 [31] was quickly followed by the discovery of clomiphene [42] and, somewhat later, of tamoxifen [43]. These compounds are structural homologues of the non-steroidal oestrogen chlorotriphenylethylene, which was discovered almost 50 years ago [44]. Tamoxifen, the only antioestrogen in widespread clinical use for the treatment of breast cancer, serves to exemplify the complex pharmacology of the non-steroidal antioestrogens. Physiological responses to tamoxifen are species-, target-organ- and response-specific (see Ref. 45 for review). For example, if we consider the rodent uterus, tamoxifen alone will stimulate uterine growth in immature rats to a lesser extent than oestradiol, but will block partially the uterotrophic action of exogenous oestradiol. These effects are illustrated in Fig. 4. Thus. in the rat tamoxifen is classified as an antagonist with partial agonist activity. In the mouse, however, tamoxifen fully stimulates uterine growth and does not block the action of oestradiol; it is thus classified as an agonist in the mouse.

162

B

300-

0" 0

> c

-E

200-

0.01

0.1

1.0

10.0

Dose (mg/kg) Fig. 4 . Uterotrophic and antiuterotrophic activity of tamoxifen in the immature rat. Animals received three daily doses of vehicle alone (open bar), 0.5 pg oestradiol benzoate S.C. alone (hatched bar) or increasing oral doses of tamoxifen alone (dotted line) or together with oestradiol benzoate (continuous line).

It is widely assumed that breast tumour regression during Nolvadex* (tamoxifen) therapy is due to antagonism of oestrogen-stimulated cell growth. Although in vitro studies on oestradiol-responsive breast cancer cells support this hypothesis [46] Nolvadex can, in some circumstances, manifest specific oestrogenic activity in patients. For example, both oestrogens and Nolvadex stimulate tumour progesterone receptor synthesis [47] and hepatic synthesis of serum proteins [48]. There remains some uncertainty about the significance of this partial agonist activity in breast cancer patients, for example, whether it may affect significantly the qualitative or quantitative aspects of clinical response to Nolvadex therapy. Ultimately this dilemma will be resolved by clinical studies with a pure antioestrogen. Although no pure antioestrogen has yet reached clinical trial the discovery of compounds with reduced agonist activity [49], and more recently of the first pure antagonists [50], promises to resolve this question. Although the benzothiophene antioestrogens LY 117018 (see Fig. 3) and LY 139481 (piperidine analogue of LY 117018) described by Jones et al. [49] are less uterotrophic than tamoxifen in the rat they are generally less effective anti-tumour agents in rats bearing mammary tumours induced by treatment with dimethylbenzanthracene [51]. This may be due, in part, to the more rapid conjugation and excretion of the LY compounds compared with tamoxifen [52],but Nicholson and Gotting [53] have drawn attention to the major difference between uterine and mammary growth promoting effects of these compounds in the rat. Both tamoxifen and LY 117018 promote mammary ductal growth as effectively as oestradiol, a complete rather than a partial agonist effect. This is consistent with a high rate of cell proliferation observed in mammary terminal end buds. In contrast, the devel*Nolvadex is a trademark, the property of Imperial Chemical Industries PLC

163 opment of mammary lobular structure is not fully stimulated by the antioestrogens [53]. These differential effects on the different cell populations of the mammary gland are analogous to those previously described in the uterus [45]. Differential responses to antioestrogens extend from the cellular to the sub-cellular level. For example, in MCF-7 human breast cancer cells, synthesis of both plasminogen activator and progesterone receptor is induced by oestradiol, whereas tamoxifen fully induces the latter protein, fails to stimulate the former and will block its induction by oestradiol [41]. Although it is widely assumed that the trophic effects of oestrogens are largely mediated at the transcriptional level by specific receptors, it is difficult to account for the differential effects of partial-agonist antioestrogens without invoking additional alternative or indirect receptor-mediated mechanisms. Several potential alternative mechanisms for the growth-inhibitory actions of tamoxifen have been described including inhibition of calmodulin action [54] and of protein kinase C [55]. Recent studies on oestrogen-mediated cell growth also indicate that oestradiol induces the secretion of specific growth factors [56]. These growth factors may, in turn, affect cell growth by autocrine or paracrine mechanisms and form an obligate link between activation of transcription and oestrogen-induced cellular replication. A further recent development of considerable interest is the observation that LY 117018 and tamoxifen stimulate the synthesis of transforming growth factor-p (TGF-P) by oestrogen-receptor positive breast cancer cells [57]. This protein inhibits the growth of both receptor positive and negative breast cancer cells. Its synthesis is suppressed by oestradiol. These changes in TGF-P occur in the absence of major changes in the rate of transcription of the TGF-P gene suggesting that modulation of post-transcriptional mechanisms may also play a role in antioestrogen action. The existence of multiple growth effector mechanisms and multiple levels of control offers numerous possibilities for rationalizing the apparently diver-

\

I

1

, 10 Dose (mg/kg)

I

30

Fig. 5. Uterotrophic and antiuterotrophic activity of ICI 164.384 in the immature rat. Legend as for Fig. 4 except that both oral (A)and subcutaneous ( 0 ) administration was employed.

164 gent effects of both oestrogen agonists and antagonists in steroid responsive cells. A new class of steroidal oestrogen antagonists [50]may provide a novel approach to further analysis of steroid action at both the physiological and molecular level. These compounds, for example ICI 164,384 (see Fig. 3 for structure), appear to be pure antagonists as exemplified for the rat uterus in Fig. 5 , and are unique in this respect. The trophic actions of exogenous and endogenous oestradiol, and of partial agonist antioestrogens, are blocked completely by ICI 164,384 whilst the compound itself is devoid of stimulatory activity [50,58].These pharmacological studies are consistent with competitive antagonism at the level of the specific receptor. Direct proof of this is provided by the observation that both oestradiol and ICI 164,384 bind with a similar affinity to the same number of sites in a partially purified preparation of the human oestradiol receptor [59]. Thus, in contrast to antioestrogens like tamoxifen, the ICI 164,384-oestrogen-receptorcomplex appears to encode consistently an 'off' message within those pathways which normally respond to the oestrogens. The progression of a compound of this kind to the therapy of breast cancer offers, for the first time, the opportunity to determine the therapeutic effects of complete oestrogen withdrawal. It should be noted that alternative current strategies like surgical or chemical ovariectomy or treatment with aromatase inhibitors do not achieve this objective.

References 1. Furr, B.J.A. and Wakeling, A.E. (1987) Pharmacology and Clinical Uses of Inhibitors of Hormone Secretion and Action. Bailliere Tindall, Eastbourne, England. 2. Matsuo, H., Baba, Y . , Nair, R.M.G., Arimura, A. and Schally, A.V. (1971) Biochem. Biophys. Res. Commun. 43, 1334-1339. 3. Burgos, R . , Butcher, M.. Ling. N.. Monahan, M., Rivier, J . , Fellows, R . . Amoss, M., Blackwell. R.. Vale, W. and Guillemin. R. (1971) C.R. Acad. Sci. Serie D. 273, 1611-1613. 4. Schally, A.V. and Kastin, A.J. (1971) Drug Ther. 1, 29-32. 5 . Karsch, F.J. (1987) Ann. Rev. Physiol. 49, 365-382. 6. Rabin, D. and McNeil, L.W. (1980) J . Clin. Endocrinol. Metab. 51, 873-876. 7. Clayton. R.N. and Catt. K.J. (1981) Endocr. Rev. 2, 186209. 8. Labrie, F.. Lefebvre, F.A., Marchetti. B., Kelly, P.A., Reeves. J . J . , Proulx-Ferland, L.. Giguere, V., Godbout, M.. Pelletier, G . . Raymond. V. and Leung, P. (1985) In: Polypeptide Hormone Receptors, Chapter 13 (Posner, B.I., ed.) pp, 481-505. Marcel Dekker, New York. 9. Hahn, D.W., McGuire, J.L., Vale. W. and Rivier, J . (1984) I n : LHRH and Its Analogs, Chapter 5 (Vickery. B.H.. Nestor, J.J. and Hafcz, E.S.E., eds.) pp. 49-60. MTP Press Limited, Lancaster. England. 10. Labrie, F . , Belanger, A., Cusan, L.. Seguin, C., Pelletier. G.. Kelly, P.A., Reeves, J.J., Lefebvre, F. A. , Lemay, A , , Gomdeau, Y . and Raynaud, J.P. (1980) J . Androl. 1, 209-228. 11. Fujino, J.. Kobayashi. S . . Obayashi, M., Fukuda. T.. Shinagawa, S . , Yamazaki, I . , Nakayama, R . , White, W.F. and Rippel, R.H. (1972) Biochem. Biophys. Res. Commun. 49, 863-869. 12. Schally, A.V., Arimura, A. and Coy. D . H . (1980) Vit. Horm. 38, 257-263. 13. Nestor. J.J.Jr. (1984) In: LHRH and Its Analogues, Chapter 1 (Vickery, B.H., Nestor, J.J.Jr. and Hafez, E.S.E., eds.) pp. 3-10, NlTP Press Limited, Lancaster. England.

165 14. Sandow, J. (1987) In: Pharmacology and Clinical Uses of Inhibitors of Hormone Secretion and Action, Chapter 20 (Furr. B.J.A. and Wakcling, A . E . , eds.) pp. 365-384. Bailliere Tindall. Eastbourne, England. 15. Nicholson, R.I. and Maynard, P.V. (1979) Br. J . Cancer 39, 268-273. 16. Redding, T.W. and Schally, A.V. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 6509-6512. 17. Furr. B.J.A. and Nicholson, R.I. (1982) J . Reprod. Fertil. 64, 529-539. 18. Nicholson, R . I . , Walker. K.J. and Davies, P. (1986) Cancer Surveys 5 , 463-486. 19. Ahmed. S.R., Grant, J.B.F., Shalet, S.M. Howell, A , , Costello, C.B., Weatherson, T. and Blacklock, N.J. (1986) Br. J. Urol. 5 8 , 534-538. 20. St. Arnaud, R . , Lachance, R . , Kelly, S.J., Belanger, A , , Dupont, A . and Labrie, F. (1986) Clin. Endocrinol. (Oxford) 24, 21-30. 21. Hutchinson, F.G. and Furr, B.J.A. (1987) In: Pharmacology and Clinical Uses of Inhibitors of Hormone Secretion and Action, Chapter 22 (Furr. B.J.A. and Wakeling, A.E., eds.), pp. 409-431. Bailiere Tindall, Eastbourne, England. 22. Grant, J.B.F., Ahmed. S.R., Shalet. S . M . . Costello. C.B., Howell, A. and Blacklock, N.J. (1986) Br. J. Urol. 58, 539-544. 23. Jones, P.B.C. and Hsueh, A.J.W. (1984) In: LHRH and Its Analogs, Chapter 13 (Vickery, B.H., Nestor, J.J. Jr. and Hafez, E.S.E. eds.) pp. 163-179. MTP Press Limited, Lancaster, England. 24. Fraser, H.M., Sharpe, R.M. and Popkin. R.M. (1984) In: LHRH and Its Analogs, Chapter 14 (Vickery, B.H., Nestor, J.J. Jr. and Hafez, E.S.E., eds.), pp. 181-195. MTP Press Limited, Lancaster, England. 25. Vale, W . , Grant, G . , Rivier, J., Monahan, M., Amoss, M., Blackwell, R., Burgus, R. and Guillemin, R. (1972) Science 176, 933-934. 26. Vilchez-Martinez, J.A., Schally, A.V., Coy, D.H., Coy, E.J., Debeljuk, L. and Arimura, A. (1974) Endocrinology 95, 2 13-2 18. 27. Vickery, B.H. (1987) In: Pharmacology and Clinical Uses of Inhibitors of Hormone Secretion and Action. Chapter 21 (Furr. B.J.A. and Wakeling, A . E . , eds.) pp. 385-408, Bailliere Tindall, Eastbourne. England. 28. Clayton, R.N. (1984) In: LHRH and Its Analogs, Chapter 4 (Vickery, B.H., Nestor. J.J. Jr. and Hafez. E.S.E., eds.) pp. 35-46. MTP Press Limited. Lancaster, England. 29. Conn, P.M., Rogers, D.C., Stewart. J.M., Neidel, J. and Sheffield, T. (1982) Nature 296,653-655. 30. Pavlou, S.N., Debold, C. W., Island, D.P., Wakefield, G., Rivier, J., Vale, W. and Rabin, D . (1986) J . Clin. Endocrinol. Metab. 63, 303-308. 31. Lerner, L.J.. Holthaus, F.J. Jr. and Thompson, C.R. (1958) Endocrinology 63, 295-318. 32. Kistner, R.W. and Smith, O.W. (1961) Fertil. Steril. 12, 121-141. 33. Segal, J.S. and Nelson, W.O. (1958) Proc. SOC.Exp. Biol. Med. 98, 431-436. 34. Chang, M.C. (1959) Endocrinology 65 35. Neumann, F. (1987) In: Pharmacology and Clinical Uses of Inhibitors of Hormone Secretion and Action, Chapter 7 (Furr, B.J.A. and Wakeling, A.E., eds.) pp. 132-159. Bailliere Tindall. Eastbourne, England. 36. Baulieu. E.E. (1985) In: Abortion: Medical Progress and Social Implications. Ciba Foundation Symposium 115, (Porter, R . and O’Connor, M., eds.) pp. 192-210. Pitman Press, London. 37. Henderson. D . (1987) In: Pharmacology and Clinical Uses of Inhibitors of Hormone Secretion and Action, Chapter 10 (Furr. B.J.A. and Wakeling. A . E . , eds.) pp. 184-211. Bailliere Tindall, Eastbourne, England. 38. Neri. R.O., Florance, K., Koziol. P. and Vancleave. S. (1972) Endocrinology 91, 427-437. 39. Moguilewsky, M . , Fiet, J., Tourncmine. C. and Raynaud. J.-P. (1986) J. Steroid Biochem. 24, 139- 146. 40. Furr, B.J.A.. Valcaccia, B . , Curry, B., Woodburn, J . R . , Chesterton, G . and Tucker, H. (1987) J. Endocrinol. 113. R7-R9.

166 41. Wakeling, A.E. and Slater, S.R. (1981) In: Mechanisms of Steroid Action, Chapter 12 (Lewis, G.P. and Ginsburg, M., eds.) pp. 159-171. Macmillan Press, London. 42. Holtkamp, D.E., Greslin, S.C., Root, C.A. and Lerner, L.J. (1960) Proc. SOC.Exp. Biol. Med. 105, 197-201. 43. Harper, M.J.K. and Walpole, A.L. (1967) J . Reprod. Fertil. 13, 101-119. 44. Robson, J.M.. Schonberg. A. and Fahin, H.A. (1938) Nature 142,292-293. 45. Furr. B.J.A. and Jordan, V.C. (1984) Pharmacol. Therapeut. 24, 127-205. 46. Lippman, M.E.. Bolan, G. and Huff, K. (1976) Cancer Res. 36, 4595-4601. 47. Waseda, M., Kato, Y . , Imara, H. and Kurata, M. (1981) Cancer Res. 41, 1984-1988. 48. Fex, G.. Adielsson, G . and Mattson, M. (1981) Acta Endocrinol. 27, 109-113. 49. Jones, C.D., Jevnikar, M.G., Pike. A.J., Peters, M.K., Black, L.J., Thompson, A.R., Falcone, J.F. and Clemens, J.A. (1984) J. Med. Chem. 27, 1057-1066. 50. Wakeling, A.E. and Bowler, J. (1987) J . Endocrinol. 12, R7-Rl0. 51. Wakeling, A.E. and Valcaccia, B. (1983) J. Endocrinol. 99, 455-464. 52. Wakeling. A.E., Valcaccia, B., Newboult. E . and Green, L.R. (1984) J. Steroid Biochem. 20, I 1 1-120. 53. Nicholson, R.1. and Gotting, K. (1986) Rev. Endocr. Rel. Cancer (Sappl.) 19, 49-62. 54. Lam, H-Y.P. (1984) Biochem. Riophys. Res. Commun. 118,27-32. 55. O’Brian, C.A., Liskamp, R.M., Solomon, D.H. and Weinstein, I.B. (1986) J . Natl. Cancer Inst. 16, 1243-1246. 56. Lippman, M.E., Dickson, R.B.. Kasid. A . . Gelmann, E., Davidson, N., McManaway, M., Huff, K., Bronzert, D., Bates, S.. Swain, S. and Knabbe, K. (1986) J. Steroid Biochern 24, 147-154. 57. Knabbe, C., Lippman, M.E., Wakefield, L.M., Flanders, K.C., Kasid, A , , Derynck, R. and Dickson, R.B. (1987) Cell 48, 417-428. 58. Wakeling, A.E. and Bowler, J. (1988) J . Steroid Biochem. (in press). 59. Weatherill, P.J., Wilson, A.P.M., Nicholson, R.I., Davies. P. and Wakeling, A.E: (1988) J . Steroid Biochem. (in press).

SECTION I1

Specific actions of steroid hormones

This Page Intentionally Left Blank

B . A . Cooke, R . J . B . King and t 1 . J . van dcr Molen (eds.) Hurmones and their Actions, Purt 1 01988 Elsevier Science Publishcrs BV (Biomedical Division)

169 CHAPTER 11

The functions of testosterone and its metabolites W. IAN P. MAINWARING, SHONA A. HAINING and BARBARA HARPER Department of Biochemistry, University of Leeds, Leeds LS2 YJT, England

1. Introduction A distinctive feature of the androgens is their diversity of function. The early history and investigation of their mode of action has been reviewed [l]and an excellent contemporary update, especially from the clinical standpoint, has been published [2]. In higher eukaryotes and particularly mammals, androgens have a powerful influence on the male at all stages of development, from the embryo, fetus, neonate to sexually mature adult. The secretion of the principal androgen, testosterone, occurs in three critical periods of development; two small bursts in the embryo and neonate, with very conspicuous production at the time of puberty. The biosynthesis of testosterone will be mentioned only in passing here as it is most ably described in the accompanying chapter by David Gower (see Chapter 1). Each phase of secretion elicits crucial morphological and biochemical changes. In certain situations, notably in malignant conditions, the secretion of androgens by the adrenal cortex, zonae reticularis and fasiculata, assumes prominence. The synthesis of testosterone is primarily in the testis, and based on the concept elucidated originally by Jensen et al. [ 3 ] ,it is also the principal ‘target’ organ for androgens. It should not be overlooked that the vital function of androgens, in biological terms, is to promote and then maintain spermatogenesis. Muscle is the major androgen target in terms of sheer mass in male animals; paradoxically, and despite the widespread use of steroids by athletes, the molecular basis of this anabolic function of androgens remains unknown. Many other functions depend on the selective metabolism of testosterone in different target organs and, again, this is a notable aspect of an-

Abbreviations: ABP, androgen-binding protein: Scu-DHT. Scu-dihydrotestosterone; SP-DHT, Sp-dihydrotestosterone; EGF, epidermal growth lactor: FSH, folitropin; KAP. kidney androgen-induced protein; LH, lutropin; NGF, nerve growth factor; SGF, testis-specific growth factor.

170 Function

Structure

Function

Structure

OH

?H

I

5 , Dihydrotestosterone

Enhancement of

Inductions of cenaln liver enzymes

haem biosynthes6

Androstenedione

5 1-Androslanediols

Fig. 1. The structure and function of testosterone and its active metabolites.

drogen action. A brief summary of the nature and function of these various metabolites is presented in Fig. I . Remarkably, the consequent functions range from mitogenicity, haemopoiesis, development of the urogenital tract to sexual attraction. It should also be stressed that testosterone works directly, not through metabolites, in spermatogenesis and its anabolic function in skeletal muscle. Nonetheless, during the course of differentiation, varying androgen target organs are imparted with the enzymic capability for producing the appropriate metabolites of testosterone as and when required. This important concept was established by the innovative work of Farnsworth and Brown [4] and enforced by other investigators [5,6].Some findings, such as the vital conversion of testosterone to the primary oestrogen, oestradiol-17/?, were unexpected but now authenticated [7,8]. In general terms, androgens work primarily through selective changes in gene expression. However, as described later, this statement is not true for all the functions of testosterone and its metabolites. The explosive impact of recombinant DNA technology has rapidly accelerated our understanding of androgen action, as well described in the accompanying chapter by Malcolm Parker (see Chapter 3). Nonetheless, it must be said that much remains to be learned about the mechanism of action of androgens in molecular terms. In the ensuing account, the effects of an-

171 TABLE I The biological effects of androgens on target organs of male animals Organ

Species

Developmental Effect period

External genitalia Liver Blastoderm Brain

All Most Birds only Most

Liver Testis Accessory sex glands

Most All All

Embryonic Embryonic Embryonic Fetal o r neonatal Neonatal Pubertal Pubertal

Muscle Hair follicles in specific areas Sebaceous glands Vocal cords Antlers Brain Brain

Most Most Most Most Deer All Birds and fish Most

Bone marrow Brain

DOP

Sexual differentiation Haemoglobin synthesis Haemoglobin synthesis Sexual differentiation

Pubertal Pubertal Pubertal Pubertal Adult Adult Adult

Synthesis of enzymes Spermatogenesis Rapid growth and stimulation of secretions Positive nitrogen balance and growth Hair growth Sehum production Thickening of cords Growth and strengthening Male libido Courtship behaviour

Adult Adult

Structural maintenance Social behaviour

drogens on certain organs will be highlighted and we conclude by speculating how future progress may be accomplished. A brief summary of the functions of androgens is presented in Table I; this is by no means a comprehensive list and additional information may be found in the extensive review by Mooradian et al. [2]. Certain aspects of the function of androgens warrant further comment. First, some responses are extremely organ- and even species-specific. Second, similar responses in different species illustrate significant differences in temporal terms: for example. sexual differentiation of the brain is complex and in certain rodents and primates this occurs in the neonate but in the fetus in other species. Third, androgenic responses are not simply a reflection of circulating concentrations of testosterone and metabolites in the blood; hair growth and deepness of voice are certainly controlled by androgens but oestrogens are needed as well. Fourth, the actual concentration of free, biologically active testosterone in itself does not provide exclusive control of androgenic responses; many ageing men remain fertile despite a decline in their secretion of androgens. Finally, ethnic differences are apparent even in the face of a universal similarity of testosterone secretion; for example, Orientals have significantly less pubic and bodily hair than Caucasians. Since the distinctive metabolism of testosterone provides the most plausible explanation of the diversity of androgenic responses, brief reference should be made to the enzymes involved. Such a summary is presented in Table 2. Crucial enzymes include microsomal and nuclear NADPH-dependent 5a-reductase, the microsomal

TABLE I1 The formation of metabolites of testosterone in different target organs

r 4

h,

Target organ

Metabolite

Enzyme involved

Main response

Embryonic anlagen

Sa-DHT

Sa-Reductase

Differentiation

Accessory sex organs

Sa-DHT

Sa-Reductase

Dog prostate

Sa-Androstan-3a. 17a-diol

Hair follicles and sebaceous glands

Sa-DHT and oestradiol17P

Sa-Reductase and dehydrogenascs Sa-Reductase and aromatase

Liver

Androstencdione

17P-Dehydrogenase

Brain

Sa-DHT and oestradiol17P 5P-DHT

Sa-Reductase and aromatase SP-Reductase

Comb and wattles Musclc

Various Sa-androstanediols Sa-DHT None

Sa-Reductase and dehydrogenascs Sa-Reductase No metabolism

Bone marrow Testis

None

Sol-DHT

No metabolism Sa-Reductasc

Spleen, liver, blastoderm Kidney

Comments

Certain anlagen respond to testosterone directly. not to 5aDHT Growth and secretion The enzyme activity has extreme species variation and may demonstrate marked developmental change Growth Novel metabolite, not reported elsewhere Growth of sexual hair and Complex process, involving fluctuation in hormone production glandular secretions developmentally and with age Imprinting of catabolic enComplex process, not yet fully zymes understood Sexual differentiation Temporally subject to species variation, late fetus to neonate Haemoglobin synthesis The response does not primarily involve genetic transcription Induction of enzymes Some induced proteins have yet to have a prescribed function Growth Profound historical importance Anabolic effect Mechanism of response not yet elucidated in molecular terms R N A and protein synthesis Mechanism of response uncertain Growth and spermatogenesis Enzyme needed only during attainment of sexual maturation

173 aromatase enzyme complex, cytosolic and microsomal dehydrogenases and cytosolic, NADH-dependent 5p-reductase. The metabolic outcome of these activities is the generation of metabolites necessary for the function of distinctive target organs. The situation is made more complex by fluctuations in the activity of these enzymes during male development. Nonetheless, these fluctuations are not haphazard and reflect a subtle means of regulation, albeit by mechanisms currently unknown. Crucial metabolites include 5a-dihydrotestosterone, Sp-dihydrotestosterone, oestradiol-l7a, androstenedione and various Sp-androstanediols. Of these metabolites, 5P-DHT is notable because it has an extremely angular geometric shape, resulting from the fusion of the A and B rings in the cis orientation. The 5aandrostan-3a, 17a-diol identified in the dog prostate is also of extreme interest; only

Testosterone

H

LJ

5~-Dihydrotestosterone

5R-Dihydrotestosterone

J Oestradiol

Ring structures (side view)

M Q Boat

Chair

Aromatic

Fig. 2. The shape of testosterone and some of its metaholites. The six-carbon rings in steroids can be in three forms (lower diagram). All presented here are in the chair configuration except for the planar aromatic ring in oestradiol; this particular steroid is not easy to present diagrammatically.

174 the dog shares a common and high incidence of prostate cancer with the human male, yet investigations on this novel metabolite appear to have come to a halt. As illustrated in Fig. 2, the importance of the different stereochemical shapes of the various metabolites of testosterone is self-evident. The receptor concept depends on a close geometric fit or structural recognition between testosterone, its metabolites and appropriate receptor proteins. The current consensus suggests that androgen target cells have the ability to produce and subsequently bind differently shaped metabolites and so explains how testosterone can elicit such a diversity of biological effects.

2 . The functions of androgens in various target cells 2.1. Testis Higher animals, notably Mammalia, are intrinsically female. Establishment of the male phenotype essentially depends on a genetic diversion. Little of the genetic information encoded in the distinctive male or Y chromosome is ever expressed. Nonetheless, the Y chromosome dictates the sexual differentiation of the primary sex organs, ovary or testis, from a common pool of uncommitted cells in the embryo. The overall process is unclear, but persuasive evidence suggests that a crucial role is played by a glycoprotein, the H-Y antigen, induced only in the presence of the Y chromosome [9]. In the absence of H-Y antigen, the embryonic gonad develops inevitably into the ovary; in the presence of the Y chromosome, the testis is produced. As reviewed by Bellve [lo], development of the primordial testis needs harmonious division and differentiation of both somatic and germinal cells. The somatic cells are derived from the involuted mesonephros, whereas the germinal cells originate from the yolk sac endoderm. In essence, two primordial structures, or anlagen, are involved in testicular development. Once developed, the embryonic testis begins to secrete small but crucial amounts of testosterone and anti-Mullerian duct hormones. At puberty, the testis significantly enlarges under the influence of a dramatic surge in the pituitary hormones, L H and FSH. As a consequence, the conversion of cholesterol to testosterone is strikingly enhanced and, in turn, testosterone serves as a powerful autocrine factor in this organ. The testis is a complex organ and an understanding of its normal control and function emanated from the technique of maintaining constituent and purified cells in culture [ 113. The two gonadotropins interact with different cell types, but work by a common mechanism involving the second messenger, cyclic AMP. LH interacts specifically with the Leydig cells, thus promoting the secretion of testosterone. FSH interacts specifically with the Sertoli cells and, together with testosterone, promotes spermatogenesis and the induction of certain proteins. The mechanisms whereby the combination of FSH and testos-

175 terone maintains spermatogenesis remains to be clarified, but the secretion of ABP by the Sertoli cells of the seminiferous tubules is clearly important. This protein enables the semen to accumulate testosterone and thus provides the appropriate hormonal milieu in the epididymis for ensuring the viability, motility and maturation of the spermatozoa. Testosterone itself, rather than metabolites, regulates spermatogenesis. In keeping with this postulate, testosterone-specific receptors have been identified in the testis and they are primarily concentrated in the Sertoli cells. Importantly, a close temporal relationship has been established between the nuclear accumulation of testosterone and enhanced transcription of certain genes in the Sertoli cells by RNA polymerase B. Two exciting aspects of testicular biochemistry deserve particular mention. First, after years of controversy, the enigmatic polypeptide, inhibin, has now been purified and characterized [12]. These discoveries help to clarify how negative feedback loops control the biosynthesis of testosterone in the testis. Second, contemporary research has led to the identification of a testis-specific growth factor, SGF [13]. This mitogenic protein (hydrophilic; M , 15 700) is highly conserved and predominantly found in the Sertoli cells, where FSH and testosterone are selectively concentrated. SGF enhances DNA synthesis and induces meiotic division in the avascular, central compartment of the seminiferous epithelium. Fundamental interest aside, these pioneering studies by Bellve and others could lead to the raising of monoclonal antibodies and thereby a safe and specific means of male contraception. 2.2. Urogenital tract Classical studies by Jost provided the understanding of how the distinctive urogenital tract of the two sexes develops in the mammalian embryo [14]. As in the case of the testis, embryonic development of the male urogenital tract requires a genetic diversion. Some of the male and female accessory sex organs arise from common, uncommitted anlagen derived from primitive Mullerian and Wolffian ducts, whereas the male and female external genitalia originate from common anlagen found outside the duct systems. In the absence of further stimuli, the Mullerian duct system persists and differentiates as female; female external genitalia develop inevitably from the remaining anlagen. The Wolffian duct system contains a ‘suicide message’ based on very active lysosomes, which if unchecked by androgens, leads to rapid atrophy. Under other circumstances to be soon described, the Mullerian duct system can be prone to dissolution. The biochemical explanation of correct direction and sexual dimorphism of the urogenital tract emanated largely from the work of Wilson and Gloyna [15] and Josso [16]. The crucial triggers are the secretion of testosterone and anti-Mullerian hormone by the embryonic testis. As a further part of this complex developmental diversion, certain anlagen contain an active 5a-reductase system whereas those in the Wolffian duct system do not. In the hormonal environment generated by the embryonic testis, the Wolffian duct system proliferates

176 and differentiates into the male urogenital tract, some anlagen responding specifically to testosterone and others to 5a-DHT. These profound morphological changes are illustrated in Fig. 3. Specific receptors for the appropriate androgens, distributed within the correct primordial sites, have been identified. Corroboration of the developmental scheme proposed above has come from both experimental and clinical sources. In all instances, the subjects of study have been cases of testicular feminization syndromes or pseudohermaphroditism. Lyon and Hawkes first described a mutant Tfm mouse, which possesses a normal male XY karyotype, yet has a female phenotype [17]. On subsequent genetic analysis, it was established that a deletion in a specific Tfm locus on the X chromosome resulted in loss of the androgen receptor system. Animals carrying the mutation were refractory to massive injections of all androgens, including all natural androgens and their synthetic counterparts. Similar findings apply to another mutant, the Ps rat. Using the elegant technique of embryo aggregation, chimaeric mice can be produced by the fusion of early embryos of contrasting genotype. Fusion of androgen-resistant (TfmiY) and normal (XU) embryos produced mice active in spermatogenesis when adult. Testosterone then is essential for sperm maturation, with no direct effect on b DEFINITIVEMALE

Primitive urogenital system

srire, " O "1 iPlmllei0i e ~ u i d n MulCran i Ih","0

elm,,

enibvon c y ~ n n n

Preouce

Scrotum

I

Testis enlarged and descended into scrotum Vagina'

a. DEflNmVE FEMALE

c. SEXUALLYMATURE MALE

Fig. 3. The involvement of androgens in the development of the male urogenical tract. The organs identified in heavy type in the definitive male (b) differentiate in the presence of testosterone; the remainder require So-DHT for their correct establishment.

177 the germ cells but rather the Sertoli cells. Imperato-McGinley et al. made fascinating observations in Dominica on a form of intersexuality caused by inborn errors in 5a-reductase activity [18]. The individuals had normal circulating concentrations of testosterone and consequently had an essentially male phenotype. However, since all developmental processes dependent on 5a-DHT are seriously retarded, at birth the external genitalia are hypoplastic and the penis malformed (hypospadia) or absent. In many such individuals, the male pattern of facial and bodily hair may fail to occur; in others, the surge of testosterone secretion at puberty may stimulate penile growth and muscular stature. These are physically and psychologically disturbing symptoms, for infants may mistakenly be raised as girls when really they are boys. The condition may be detected early by conducting assays for 5a-reductase on fibroblasts cultured from perineal or other erotogenic areas of skin, but there is no cure other than surgically initiating a sex change.

2.3. Haemopoietic organs The regulation of erythropoiesis is complex and vital. Various forms of haemoglobin are produced in many organs at different stages of development. The switching from one form to another clearly involves changes in the expression of specific globin genes [19], but the molecular basis of regulation remains unclear. If the appropriate switch fails, notably in the human from fetal haemoglobin (HbF) to adult haemoglobin (HbA), then the consequences are clinically dire. Certain clues to the control of erythropoiesis are emerging. The secretion of erythropoietin, primarily by the kidney, but also from extrarenal sites, is clearly critical since it controls the production of colony-forming units and so-called burst-forming units. It has been known for many years that androgens enhance erythropoiesis. Testosterone elevates the circulating concentration of erythropoietin and markedly enhances erythropoiesis in marrow cells maintained in culture. The effect of androgens is promoted by various metabolites in a very subtle manner, subject to significant developmental change. It is currently advocated that 5preduced metabolites act directly on erythropoietic stem cells, whereas 5a-reduced derivatives enhance the level of erythropoietin. Interestingly, mutant mice with testicular feminization ( TfmiY) actively maintain erythropoiesis, suggesting that androgenic effects on red cell formation are not mediated by conventional receptor mechanisms; antiandrogens are either of variable or no influence. As well reviewed elsewhere [2], the control exerted by Sa-reduced steroids on erythropoiesis is best explained by transcriptional control of the erythropoietin gene. Certainly, such androgens are actively concentrated in the nuclear chromatin of bone marrow, both in vivo and in vitro. By contrast, the situation is entirely different with respect to Spreduced steroids. In the embryonic and fetal organs of many animals, aetiocholanolone or 5 P D H T regulate the appearance of a 5’ cap-recognition protein, without which the translation of the mRNAs for haemoglobin E (embry-

178 onic) and haemoglobin P (primitive) cannot proceed [20]. One contentious aspect of this area of research remains; the identification of receptors for Spsteroids reported by one research group [21] has not been corroborated by others.

2.4. Salivary glands It was recognized some years ago that the submaxillary gland in many strains of mice was sexually dimorphic [22]. In male and androgen-treated females, the gland has a characteristic morphology and biochemistry (Fig. 4). Many male-specific proteases and other proteins are evoked in the gland by testosterone. Of remarkable interest is the androgenic control of NGF in this ‘non-neural’ site. The identification and characterization of NGF by Levi-Montalcini was a milestone in our understanding of proteinaceous rather than steroidal mitogenic factors. In other species, especially the Gottingen miniature pig, Booth established that androgens promote the abundant appearance of distinctive metabolites of testosterone, often exceeding 2 mmol/g wet weight [23]. The primary products are odorous steroids of the unusual l6a-androstene series, especially 3a-androstenol, 5a-

Fig. 4. Sexual dimorphism in the mouse submaxillary salivary gland

179

H 5 ~-Androst.l6.en.3.01

5 =Androst.l6.en.3.one

Fig. 5. Novel androgenic pheromones idcntitied i n thc hoar subinaxillary gland

androst-16-en-3a-ol and 5a-androstenone, 5a-androst-16-en-3-one (Fig. 5 ) . These metabolites are extremely powerful pheromones, exciting female pigs in oestrous to adopt the mating stance. Boar meat is seldom sold because of the pungent smell of these steroids. A specific binding protein for these pheromones is also present in the glands and is subject to androgenic regulation.

2.5. Kidney In most rodents of similar age, but not inbred rats, the kidneys in the male are larger and structurally different from those in the female. Androgens are responsible for this sexual dimorphism. This contrast is not mimicked in the human. Many renal enzymes are subject to androgenic control [2], but the elegant studies by Paigen on P-glucuronidase are particularly noteworthy. His imaginative research demonstrated the real complexity of the control of gene expression in higher eukaryotes [24]. The six genes which clearly regulate the biosynthesis of this enzyme have been identified and mapped, but others are also involved. Together, the known genes regulate the molecular size of the enzyme, its intracellular location, its inducibility by androgens and its storage (Fig. 6). Paigen classified the genes into four categories depending on their function. The basic subunit of p-glucuronidase is a polypeptide, M , 70000, but the six functional forms of the enzyme are tetramers, M , range 260000 - 470000. The various forms may be present in the cytosol, lysosomes or endoplasmic reticulum. These may be summarized as follows: type L (lysosomes, M , 260000), type M (four species, microsomes, M , 310000 - 470000) and type X (cytosol and microsomes, inducible by androgens, M , 260000, but distinct from L). The four species of M enzyme contain varying amounts of another protein, egasyn, which enables the tetramer to bind to the membranes of the endoplasmic reticulum. The six genetic loci engaged in this process, plus their classification, are: (a) Gus, structural gene controlling the synthesis of the mRNA encoding the enzyme monomer; (b) Tfm, structural gene controlling the mRNA for the androgen receptor protein; (c) Cur,

180

Fig. 6 . The genes cngaged in the expression of 0-glucuronidase in mouse kidney

regulatory gene responsible for the androgenic induction of the enzyme; (d) C u t , temporal gene necessary for controlling the appearance of the enzyme during development; (e) Eg, processing gene responsible for the synthesis of egasyn; (f) Bg, processing gene necessary for regulating the entry of the enzyme into the lysosomes and its secretion. Two points warrant further mention. First, only one gene of this set is responsible for the sensitivity of the overall process to androgens. Second, the results were achieved by standard techniques, rather than contemporary means of recombinant DNA technology. Other studies have shown that mouse kidney contains a receptor system specific for testosterone and this is lost in the testicular feminization mutant. Certainly, androgens are primarily responsible for the induction of pglucuronidase, but part of the response could be attributable to progesterone receptors. In this context, var-

181

ious progestins have been cited as androgenic, synandrogenic or antiandrogenic, depending on whether they simulate, potentiate or negate the effects of androgen on the synthesis of this hydrolase in mouse kidney [25]. More recently, studies on androgen action on mouse kidney have centred on two other proteins, namely KAP and ornithine decarboxylase, but work on p-glucuronidase has also been extended (for a review see Ref. 26). After only a single exposure to testosterone, the mouse kidney accumulates the Poly A+-mRNAs for all of these three proteins (Fig. 7). The three genes encoding these proteins have diverse characteristics both in terms of structure and mode of regulation. Ornithine decarboxylase genes are present as a cohort or family of several types, whereas those for KAP and Pglucuronidase are seemingly unique in the genomes of most species of mice. The transcription product of the KAP gene is a typical representative of the abundant class of mRNA, but the mRNAs for ornithine decarboxylase and pglucuronidase are members of the non-abundant mRNA class. After androgenic stimulation, the expression of these three genes is not identical. Following rapid occupancy of the nuclear receptor sites, the appearance of ornithine decarboxylase mRNA is quick and profound; the responses of the other two genes are slower and less extensive. The dramatic response in ornithine decarboxylase is perhaps not too

8-

7-

6-

t

;

5-

0

.-

'c1 O

4.

3-

2-

1-

00

10

20

30

40

50

Hours

Fig. 7. Changes in the contents of specific mRNAs and nuclear receptors in mouse kidney after exposure to testosterone. Female NCS mice were given a single intraperitoneal dose of 10 mg of testosterone at time zero. Thereafter, nuclear androgen receptors were measured, -0-. The specific mRNAs of three proteins were also assayed, namely for ornithine decarboxylase (m),KAP (-o-) and P-glucuronidasc ( - 0 . ) . Redrawn from Catterall et al. [26].

182 surprising; the enzyme is controlled by many effectors in all mammalian cells and has the highest turnover rate of any protein in higher organisms. The mouse kidney is unique in being subject to control by androgens. This response is elicited by testosterone itself rather than by metabolites. Despite their elegant work on these three genes, Catterall et al. [26] conclude that the modulation of gene expression remains somewhat enigmatic. While it is clear that androgens can enhance the rate of genetic transcription in many target cells, this alone is not the ultimate means of hormonal control. To help finally solve the basis of androgen action, future work needs to be directed towards the measurement of absolute rates of transcription and mRNA turnover, as adroitly advocated by Brock and Shapiro [27] from their studies on the expression of vitellogenin genes in Xenopus luevis. What is becoming clear is that the selective control of gene expression in the kidney will prove to be complex; certain genes are partially regulated by pituitary hormones wheieas others are not. In addition, renal responses have a marked and differential sensitivity to blockade by antiandrogens [26,28]. 2.6. Muscle Classical studies by Kochakian [29] on the dog established that androgens possess a powerful anabolic function, namely the maintenance of a positive nitrogen balance in skeletal muscle. This important observation was subsequently confirmed in the human. This anabolic function results in the accumulation of muscle mass and the mobilization of fat reserves; the overall result is a strong yet lithe physique. The anabolic function of androgens, under normal circumstances, must be attributed to testosterone itself as skeletal muscles have a minimal, if not negligible ability to metabolize steroids. In keeping with this concept, it was later demonstrated that testosterone stimulated mitotic activity in cultures of myoblasts more effectively than all other naturally occurring androgens. In one sense, however, this observation is not beyond question because androgens do not change the numbers of myofibrils in muscle but significantly increase their size by hypertrophy rather than hyperplasia [2]. At first, investigators encountered serious difficulties in demonstrating the presence of androgen receptors in muscle [l];earlier indications of their presence [30] have now been amply vindicated by many research groups. Despite intensive research effort, however, the anabolic function of androgens remains essentially unknown in molecular terms (for a review see Ref. 31). The need has been recognized for some time to develop steroids with a higher anabolic function, for use in the treatment of muscular dystrophy, weak stature, obesity, certain forms of eczema and psoriasis to cite but a few common problems of health. The phenotypic effects of testosterone can be arbitrarily subdivided into two classes of response: anabolic, essentially meaning an increase in the strength and size of skeletal muscle; and strictly adrogenic, referring basically to the establishment and maintenance of the male reproductive apparatus and associated sex-

183 OH

Nandrolone

OH

Norethandrolone

OH

?H

Mesterolone

Stanozol

Fig. 8. The structures of several commonly used anabolic steroids.

ual characteristics, such as libido. By making subtle substantial and stereochemical changes to testosterone, it has proved possible to produce drugs with a very high anabo1ic:androgenic ratio. A selection of some of these powerful synthetic analogues of testosterone is presented in Fig. 8. The most commonly prescribed anabolic steroid is probably nandrolonc. closely followed by stanazol. Aside from their authentic use, it is regrettable that anabolic steroids are now commonly used by athletes of both sexes in their preparation for power field events. While the taking of anabolic steroids may certainly aid the acquisition of muscular mass. their protracted and illicit use incurs serious hazards to health. Many effects of anabolic steroids are relatively mild and reversible but, in the long term, they can have dire consequences. Ultimately, all takers suffer from acute retention of Na' ions, often with subsequent oedema, heart diseases and renal failure. In women, anabolic steroids cause thickening of the vocal cords, hirsutism, male pattern baldness, severe enlargement of the clitoris, reduced ovulation and irregularity in menstruation. All of these symptoms are very predictable, a s the woman is exposing herself to an androgenic hormonal milieu; while many synthetic steroids are powerfully anabolic, they nonetheless have residual androgenic properties as well. In men, the effects of anabolic steroids are more difficult to explain but sterility is not uncommon. There is considerable evidence that anabolic steroids promote the conversion of testosterone to oestradiol-l7P, the principle female sex hormone. This aromatization process is general, but occurs principally in adipose tissue. The for-

184 mation of peripheral oestrogens enhances the risk of cardiovascular disorders and gynaecomastia. As an overview, the taking of anabolic steroids may well have serious consequences; disqualification from competition is one thing, but irreparable bodily damage is another. Using samples from blood, urine or saliva. detection of excess administration of anabolic steroids or testosterone presents no problems with contemporary knowledge and technology. For synthetic anabolic steroids, wide-ranging antibodies can be used in radioimmunological assays for 17a-methyl and l7a-ethyl derivatives. Where necessary, confirmation of a positive test can be readily gained by gas liquid chromatography, coupled with mass spectrometry. For detection of excessive abuse of testosterone itself, two approaches of assay are available. The first depends on measuring the ratio of circulating testosterone to LH. The rationale underlying this approach is that testosterone inhibits the release of LH from the hypothalamic-pituitary axis, this being an example of classical 'down regulation' or negative feedback. The second method is surprisingly simple and based on the ratio of circulating testosterone to epitestosterone; the latter is a weak androgen because the Cl7-hydroxyl group is in the a- or lower face configuration. Other studies have shown that epitestosterone is formed in a variety of peripheral tissues from androstenedione, in a stable, constant manner, irrespective of the presence of other administered compounds, including anabolic steroids. Accordingly, dramatic increases in the testoster0ne:epitestosterone ratio of an individual reflect the illicit administration of testosterone. Currently, checks on the use of anabolic steroids by all athletes are frequent and rigorously accurate. For more details, see the review by di Pasquale [311. The effects of androgens on smooth muscle have not been studied in great depth, but certain interesting observations have been made. After castration, the muscle weight and cell size decline, but there are no changes in the synthesis of collagen or cell number. All of these changes are reversed by administration of testosterone. Testosterone enhances the stimulatory effects of oestrogens on the smooth muscle in certain accessory sex organs without changing the function of the oestrogen receptors present in any way. An explanation of this novel example of hormonal cooperativity is awaited with interest and could be of clinical value. Androgen receptors have been identified and characterized in cardiac muscle of many species. By means currently unknown, androgens evoke significant biochemical and morphological changes in cardiac muscle. and enhance contractile function, notably in the ventricles, with a beneficial accumulation of Ca2+ ions in the sarcoplasmic reticulum and sarcolemma. The size of mycjSb 'Is and other fibrous elements is increased in response to androgens. These observai ?ns may explain why, in the short term, men have a more powerful cardiac outp i t than women. This physical advantage diminishes with age when men become fa1 more prone to atherosclerosis and other cardiovascular problems [2].

185 2.7. Liver Kochakian was the first to demonstrate the effects of androgens of the liver, not only in terms of size and morphology, but also with respect to enzymic composition; fumarase, catalase and D-amino acid oxidase activities were particularly sensitive to androgenic manipulation [32]. Subsequent studies confirmed and extended our understanding of the androgenic responses in liver, but illustrated that they were subject to extreme species variation. No common pattern of hepatic responses to androgens has emerged. It is not easy to review the involvement of androgens in the physiology and biochemistry of the liver. At first, many investigators failed to demonstrate the presence of androgen receptors, but subsequently, using improved assay procedures, they were unequivocally identified. Interestingly, Gustafsson et al. identified a high and specific binding of androstenedione in liver cytosol [33]; other investigators subsequently used synthetic androgens as model ligand in their assays. As cogently argued by Mooradian et al. [2], many effects of androgens on liver are probably mediated by receptor-dependent mechanisms, but others are elicited directly by androgens, with the cooperation of hormones from the anterior pituitary, especially somatotropin and somatostatin. The enzymes engaged in steroid catabolism demonstrate significant sexual dimorphism. The complement of hepatic enzymes is such as to provide protection against excessive androgens or oestrogens in the male and female, respectively. Testosterone secretion in the neonate and at puberty induces or imprints the malepattern system of enzymes. Without androgenic stimulation, the female-pattern system of enzymes is present. Among this complex assembly of enzymes, 16a-hydroxylase appears to be male-specific, whereas l5p-hydroxylase is female-specific [33]. This brief summary is something of an over simplification, because the hormonal control of these catabolic enzymes is complex and varies enormously from one species to another. Androgens are involved in regulating the activity of microsomal enzymes which metabolize drugs and other chemicals, of both endogenous and exogenous origin. Androgens have a regulatory dichotomy o n these hepatic enzyme systems; a direct, non-receptor control leading to an increase in liver mass and microsomal membranes. plus a receptor-mediated control over the amount of cytochromes P-466 and P-450 [34]. Again, it must be emphasized that the hormonal control of these cytochrome systems is difficult to summarize because of known inter-species variations. This confusion is well illustrated by sulphatase enzyme activities [35]. The liver of several species of inbred rats has provided a novel experimental system for studying the mechanism of action of androgens. The protein, a-2u-globulin, constitutes 50% of the urinary protein in male rats and its specific mRNA represents a remarkable 2% of the total hepatic mRNA; the protein is virtually undetectable in female rats. In response to androgens, male rats synthesize a unique

186 receptor system, one specific for androgens, the other for oestrogens [36]. The biosynthesis of a-2u-globulin is induced by androgens, specifically 5 a-DHT, and curtailed by oestrogens. In the hands of Roy and his co-workers, this system has provided penetrating confirmation of the selective change in gene expression promoted by androgens, using all the techniques of contemporary molecular biology [36]. Even in this instance, however, overall hormonal control is not elicited by androgens alone, because insulin plays a crucial role as well [37]. It is significant, indeed distinctive, that many androgenic responses in liver cannot be suppressed by the administration of currently available antiandrogens.

2.8. Central nervous system It has been known since the time of the Greek Civilization that the gonads have a profound effect on male sexual behaviour. About 400 B.C., for example, Aristotle commented on the diminished aggression and stunted comb of the castrate cockerel or capon. The effects of androgens on the brain are profound, encompassing a diverse range of morphological and biochemical changes, from the late fetus and neonate onwards. In the human, any serious reduction in the secretion of testosterone before and immediately after birth can result in the subsequent curtailment of maletype behavioural patterns at puberty. The male brain is imprinted seemingly irrevocably during the neonatal phase of human development. An insight into the regulation of brain function came from classical studies on NGF by Levi-Montalcini [38,39] and corroborated by research on E GF by Cohen [40,41]. Work on nerve growth factor showed that neurofibrils are produced from activated neurites, ultimately fusing with specific, homologous fibrils thereby establishing neuronal arborization. On these foundations, it became possible to begin to unravel the complex effects of androgens on the brain. In a crucial paper, Phoenix et al. provided the first invaluable clues [42]. Their studies on the guinea pig indicated that the effects of androgens on the brain were two-fold and temporally distinct. During early development, androgens promoted the differentiation and correct assembly of the neuronal pathways and then later in the adult, activated the previously organized nervous system and provided long-term maintenance of the brain. Another incisive breakthrough came from the enterprising work of Naftolin [7]. His research team established that testosterone, but not 5a-DHT, could be converted into oestradiol-17P in certain areas of the brain. This aromatization or ‘conversion’ concept has helped enormously in furthering our understanding of how androgens influence nervous and behavioural patterns. Later, it was found that 5aD H T could also be produced in the brain, as very ably reviewed by McEwen [43] and Arnold and Gorski [44]. Aromatization of testosterone does not occur in all areas of the brain; it is primarily in the preoptic area, hippocampus and amygdala, but absent in the cerebral cortex and pituitary. 5a-Reductase is evenly distributed and thus its location does not appear to play a dominant role in the regulation of

187 TABLE 111 Selected examples of the distinctive accumulation of testosterone and its metabolites in areas or nuclei of the central nervous system Location of area or nuclei

Probable triggers and specific receptors

Function

Preoptic area

Probably oestradiol alone Testosterone plus oestradiol- 178

Sexual dimorphism: masculinization Regulation of gonadotropin secretion Sexual dimorphism: masculinization Increased neuronal size

Amygdala

Superior cervical ganglion Lumbar spinal ganglion

Testosterone plus ~(u-DHT Probably testosterone alone

Increased neuronal size Control of bulbocarvernosus and levator ani muscles

brain function. Androgen receptors have been identified in many areas of the central nervous system and are somewhat distinctive in that they bind testosterone and Sa-DHT with equal affinity. Receptors for all other classes of steroid hormones have been demonstrated throughout the central nervous system, with notable concentrations in certain areas; for example, glucocorticoid receptors are primarily located in the hippocampus. A brief summary of the action of testosterone and its metabolites on nuclei in various areas of the brain is presented in Table 111. Of these areas sensitive to androgens, the aromatization of testosterone and the subsequent binding of oestradiol- 17p to specific receptors is of particular significance. This vital activity during neonatal development ensures that the response to the surge of testosterone during puberty will result in a relatively constant, certainly non-cyclical, secretion of gonadotropins in the adult male. This is an important example of neonatal programming. It is important to emphasize that the responses in the brain to both exogenous testosterone or oestradiol can be manipulated in most species, particularly rodents and Old World Primates (for a review see Ref. 8). The crucial time is in the early neonate. As superbly reviewed by Mooradian et al [ 2 ] , testosterone and metabolites generated within the central nervous system control a complex range of behavioural and physiological patterns in male mammals. Oestradiol stimulates physical activity in males and this response cannot be mimicked by Sa-DHT. Oestradiol synthesis is also responsible for initiating the drive and physical preparation of the male prior to copulation. However, testosterone is also involved and the factors responsible for the male psyche and libido remain to be fully elucidated. The regulation of body weight and the establishment of eating habits is a complex process involving many hormones and neurotransmitters. On present evidence this nutritional response re-

quires the generation of a combination of oestradiol and 5a-DHT. From the foregoing account, it follows that many responses will be impaired by antioestrogens, such as tamoxifen and nafoxidene, or inhibitors of 5a-reductase, such as aminoglutethamide and 4-aza-5-reduced steroids (451. 2.9. Anterior pituitary

Testosterone controls the secretion of gonadotropins in the male through an extensive series of negative feedback loops [45]. The pituitary has no aromatization system and is is now clear that androgen rather than oestrogen receptors are engaged in suppressing the release of L H [43]. The gland does contain 5a-reductase, however, but the involvement of 5a-DHT in regulating gonadotropin secretion has yet to be settled. There are certainly androgen receptors in the anterior pituitary which bind both testosterone and 5tu-DHT with equal affinity. Some investigators suggest 5a-DHT is deeply implicated but the regulatory system is not impaired by even the most powerful inhibitors of 5cr-reductase [46]. Testosterone also accelerates the secretion of growth hormone and thyroid stimulating hormone but to date reports on the molecular mechanisms involved have been very conflicting. In addition, the influence of testosterone on the secretion of these two hormones can be diametrically opposite, depending on the species.

2.10. Breast Testosterone depresses growth of the breast in males; there is considerable evidence for this from both experimental and clinical sources [2]. Gynaecomastia is often the result of an imbalance in the ratio of testosterone to oestradiol, but the condition has a complex aetiology and does not invariably reflect an increasing concentration of circulating oestradiol. The molecular mechanisms involved have not been elucidated; a decreased testosterone to oestradiol ratio is found in some patients, but certainly not all. Gynaecomastia becomes more common in senescence and this again could be attributed to many factors including a decreased secretion of testosterone or chahges in the transport of androgens in the blood. 2.11. Hair

Work by Hamilton established that a combination of circulating androgens and inherited factors control the length and density of facial hair in the human male [47]. Somewhat later it was shown that testosterone evoked the differentiation of hair follicles in cultured explants of embryonic skin. Male human hair may be classified into three basic types. First, male sexual hair, the obvious being the beard, which grows only under the stimulation of high concentrations of circulating androgens. Second, ambosexual hair, for example the pubis, which grows in response to lower

189 concentrations of androgens. Finally, asexual hair, such as the eyebrows. Hair follicles and the hair itself are active sites of testosterone metabolism, from which are produced 5a-DHT and androstenedione; the latter may also be readily aromatized to oestrone. Including testosterone itself there are therefore these potential hormonal triggers in the hair follicle. Of these, 5a-DHT appears to be the most important. Receptors for 5a-DHT have been identified in the follicle and 5a-reductase activity is more active during the anagen phase of rapid growth than during the telogen resting phase. This increase in 5a-reductase activity is evoked by the stimulation of the pentose phosphate cycle by testosterone, thereby producing increased amounts of reducing energy in the form of NADPH. Despite a great amount of cosmetic interest in the growth of hair, more work is clearly needed to explain the involvement of androgens and oestrogens in molecular terms. There is no doubt that elevated testosterone levels are responsible for hirsutism, especially in women [2], but the basis of idiopathic hirsutism remains an enigma.

2.12. Sebaceous glands Acne is such a common and socially distressing disease that considerable effort has been directed towards a better understanding of androgen action in the sebaceous glands. During the initiation of this condition, usually in the face, chest and back, excessive amounts of testosterone and its precursor, dehydroepiandrosterone sulphate, are present in plasma. In women, acne is promoted by 3a-androstanediol. In both sexes, the predisposing trigger is Scu-DHT and acne-bearing skin has up to 20-times the 5a-reductase activity of normal skin. The secretion of sebum from lytic sebaceous glands is the prime response to androgenic stimulation. The secretion of sebum involves another metabolite, 5a-androstan-3/3,17/3-diol[48]. The disintegration of the glands caused by excess circulating androgens provides an ideal environment for the multiplication of bacteria with ensuing facial disfigurement. Currently available antiandrogens have had little success in the arrest of acne and certainly none are active topically [45]. However, it is to be hoped that newly developed 5a-reductase inhibitors will be active in this regard. The path to a real cure may not be that simple because Ebling has convincingly demonstrated the involvement of the anterior pituitary in the onset of acne [48]. 2.13. Skin

In view of the importance of the skin, it is surprising how little we know of the mechanism of androgen action on this large and complex organ. Androgen concentrations are higher in areas of genital skin in both sexes, such i j the. scrotum in men and the labia in women. There is considerable 5a-reductase activity in androgen-sensitive areas of skin, but this enzyme system is replaced by 3a-reductase in androgen-insensitive skin [2]. Because of clinical significance to dermatology, more research is needed on the effects of hormones on skin.

190

2.14. Bone In a manner yet to be fully ascertained, androgens exert an influence on the biochemistry of mineralization and the laying down of bone mass. There is a firm correlation between bone volume and circulating levels of testosterone in ageing men. Additional clinical evidence suggests that osteoporosis can be a feature of hypogonadism and certain testicular feminization syndromes [2]. The marrow can convert testosterone to 5a-DHT and androstenedione, thereby eliciting the overall response in bone. The mechanism is currently unclear as collagen synthesis and bone resorption are not influenced by testosterone.

2.15. Lymphocytic organs High levels of androgens lead to atrophy of the thymus and clinical studies suggest that autoimmune diseases, especially lupus erythematosis, may be promoted by either high levels of oestrogen or low levels of androgens. Despite these observations, the current literature on the effects of androgens on the immune system is somewhat sparse and confusing, even using model systems akin to autoimmune diseases in the human [2]. Much more work remains to be done on this particular aspect of androgen action.

2.16. Accessory sexual glands 2.16.1. Prostate Research on the prostate of several species, including the human, continues at a high intensity to find explanations for the high incidence of hypertrophy and hyperplastic adenoma in the dog and human glands. This important aspect of androgen action has recently been reviewed in detail elsewhere (491. Progress over the last few years has been remarkable. After years of disappointing toil by many investigators, the research team of Bruchovsky and Rennie has succeeded in purifying a nuclear form of the prostate androgen receptor to apparent homogeneity [50]. Although presently available in only small quantity, the technical capability is at hand to purify the receptor on a preparative scale. Investigation on the nuclear acceptor sites continues, but a definition of their precise nature has still to be achieved. The critical androgen in the prostate is 5a-DHT; recently, powerful inhibitors of 5a-reductase have been shown to promote atrophy of the prostate [51]. Using such drugs it has proved possible to distinguish between the responses distinctively evoked by testosterone or 5a-DHT. After critical methodology for the isolation of prostate mRNA had been developed [52],it proved possible to clone androgen-sensitive genes, notably the three encoding the subunits of the binding protein, prostatein. This protein constitutes about half the soluble protein of the prostate and binds many androgens and certain drugs. Progestins and prolactin are also involved in the androgenic responses in the prostate [53].

191 2.16.2. Seminal vesicle The effect of androgens on this gland is rather subtle. Oestrogens and androgens cooperate synergistically to maintain the stroma, whereas epithelial height is maintained by testosterone but not Sa-DHT. Interestingly, oestrogens antagonize the trophic response of androgens in the epithelium [1,2]. The seminal secretion is rich in two proteins, now termed S and F, which help form the vaginal plug in female rodents after copulation [54]. The genes encoding proteins S and F have now been cloned and largely sequenced [SS]. 2.16.3. Epididymis The epithelium of the epididymis is acutely sensitive to androgenic stimulation and maintenance. After castration or treatment with antiandrogen, the epithelium involutes, accompanied by rapid loss of Sa-reductase activity and androgen receptors. Administration of testosterone alone cannot fully support the epididymis epithelium. Evidence is accumulating that the seminal fluid contains factors, including androgens, which are necessary for the full maintenance of the epithelium [56].This is an interesting example of 'functional cooperation between two androgen target organs. Certainly, seminal fluid supplies 5a-DHT to the epididymis and this is the favoured androgen for the epididymis receptors. Antiandrogens have been reported to impair the translocation of the receptor complex to the nucleus. Sperm maturation occurs in the epididymis and this process is promoted by the induction of two glycoproteins, termed C and DE, under the influence of androgens [S6]. 2.17. Exotic systems Throughout the historical development of the biological sciences, studies on unusual, even exotic, species and systems have provided insights into fundamental processes. Studies on the giant oocytes of Xenopiis luevis, with their amplified rRNA genes and remarkable translational and transcriptional efficiency, are a good case in point. With specific respect to the androgens, three forms of animal warrant mention. The male of many wading and swimming birds possesses modified sebaceous glands, organized into a composite organ, the preen or uropygial gland. This organ plays a distinctive role in the male courtship behaviour. The preen gland exhibits almost total sexual dimorphism and contains both Sa-reductase activity and receptors for 5 a-DHT. Seasonal changes in the secretion of testosterone activate the glands at the breeding season. Bird song is restricted exclusively to males in certain species and this process requires a remarkably complex interplay between several areas of the brain and metabolites of testosterone. Despite the complexity of this response, the morphological and biochemical aspects of the process are known in remarkable detail [ 2 ] . As the most Northerly living reptile, t h e garter snake has a bizarre existence but a remarkably efficient means of reproduction. The snakes necessarily have to spend nine months underground to endure the winter cold. I n

192

the brief period of higher temperature and stronger light intensity of summer, the snakes emerge, both sexes produce the appropriate sex hormones and enter into a large ‘mating ball’ to produce offspring in the short time of two months. This is a striking example of biological adaptation to a hostile environment.

3. Concluding remarks The overall mechanism of action of androgens is clearly complex but far better understood than a decade ago. Certainly, predictions of the importance of aromatase reactions made by Wilson and others [57] have been fully vindicated. The involvement of oestradiol and androstenedione in androgenic responses opens up possibilities for the development of a wider and more specific range of antiandrogens for clinical use. The structures of certain antiandrogens are presented in Fig. 9, some of which are well known, others under development. A review of antiandrogens has recently been published [45], but the underlying principles for new antiandrogens are efficacy, low toxicity and minimal contraindication. The new antiandrogen, ICI 176,334, has interesting properties and may soon be on trial against prostate adenocarcinoma [%I. The compound is a novel variant of flutamide, possessing the distinct advantage of successful blockade of androgen receptors in peripheral organs such as the prostate, without the concomitant LH surge [59,60] seen with other non-steroidal pure antiandrogens [61]. CH3

I co

CF3‘ Flutamide

RU 23908

Fig. 9. The structures of selected antiandrogens.

ICI 176334

193

I

LTR

I

I

LTR

I

Fig. 10. The structure of mouse DNA containing the integrated mouse mammary tumour provirus. The integrated viral DNA can be present in many copies. The provirus contains two long terminal repeat sequences, LTR, only one of which (left) is shown i n detail here. Each LTR contains sequences termed U3, R and US (for details, see Ref. 67). Interspersed between the LTRs are the genes named gag, pol and env, encoding viral coat proteins, reverse transcriptase and envelope proteins, respectively. The glucocorticoid-binding sequence is represented by the black box, and the transcriptional initiation area is indicated by the hatched box.

Most current evidence supports the general view that the mechanisms of action of androgens is best explained by selective changes in gene expression under the overall control of specific receptors. Nonetheless, the biosynthesis of active species of RNA in eukaryotes is remarkably sophisticated, suggesting other means of control. Regulation of androgen action at the levels of RNA processing, mRNA stabilization and enhanced translational efficiency are further possibilities which should not be discounted. Many androgen-regulated proteins are secreted and again, the possibility of androgenic control of the secretory process may prove to be of profound importance. As described earlier in this chapter, purified receptors and cloned genes open up exciting prospects for future investigation. In addition, factors are being isolated which conduct the interconversion of the various forms of receptors [62]. Modernday techniques are rapid, specific and sophisticated, proving portents for a rapid acceleration in our progress towards the understanding of androgen action in molecular terms. The outstanding success of recent research on the mechanisms of action of glucocorticoids may provide the vital impetus for future investigations on androgens (for a review see Ref. 63). Considerable interest has centred on the possibility that the glucocorticoid receptor is a protein kinase, but this no longer seems to be true [64]. As reviewed by others [65], the importance of enhancer elements in the control of gene expression in eukaryotes is now universally accepted. It has been elegantly demonstrated that the human glucocorticoid receptor binds strongly to defined consensus sequences, one upstream and the other downstream, of the transcriptional initiation point of the genes encoding growth hormone and placental lactogen [66]. As lucidly advocated by Rousseau [67], the DNA from mouse mammary tumour virus provides an exciting model for studies on the molecular biology of glucocorticoids. Again, the glucocorticoid receptor conspicuously recognizes base sequences in and around the mouse mammary tumour DNA [68]. The regions where the receptor binds are certainly promoter sites [69] and probably constitute enhancer elements [70]. A simplified diagram of this viral DNA is presented in Fig. 10.

Against this background of technical advance, it has been possible to transfer androgen-sensitive genes into a variety of recipient cells [71], and in a strikingly important paper, Darbre et al. [72] have shown that androgens can also regulate the expression of mouse mammary tumour DNA and genes surrounding it. In more specific detail, androgens appear to activate the long terminal repeat regions of the viral DNA. On these exciting foundations, the future progress of studies on androgen action looks assured.

Acknowledgements The authors wish to thank Sandra Gray and Anne Turner for their painstaking help in preparing the manuscript, Allan Haigh for the art work and the Yorkshire Cancer Research Campaign for a Postdoctoral Fellowship to Shona Haining.

References 1. Mainwaring, W.I.P. (1977) The Mechanism of Action of Androgens. Springer-Verlag. New York. 2. Mooradian. A . D . , Morley, J.E. and Korenman, S.G. (1987) Endocr. Rev. 8, 1-28. 3. Jensen, E.V., Suzuki. T.. Stumpf. W.E.. Jungblut, P. and de Sombre, E . R . (1968) Proc. Natl. Acad. Sci. U.S.A. 59, 632-638. 4. Farnsworth. W.E. and Brown, J.R. (1963) Natl. Cancer Inst. Monogr. 12. 323-329. 5. Bruchovsky, N. and Wilson. J . D . (1968) J. B i d . Chem 243, 2012-2021. 6. Anderson, K.M. and Liao, S. (1968) Nature (London) 219, 277-279. 7. Naftolin, F., Ryan. K.J., Davies. I.J., Reddy, V.V., Flores, F., Petro, A , , Kuhn, M., White, R.J., Takaoka Y . and Wolin. L. (1975) Recent Progr. Horm. Res. 31, 295-319. 8. Beyer, C. and Feder. H.H. (1987) Annu. Rev. Physiol. 49, 349-364. 9 , Ohno, S. (1977) Major Sex-Determining Genes. Springer-Verlag. Berlin. 10. Bellvt. A . R . (1977) In: Oxford Reviews of Reproductive Biology (Finn, C.A.. ed.) pp. 159-261. Oxford University Press, Oxford. 11. Dorrington, J . H . , Roller, N.F. and Fritz, I.B. (1975) Mol. Cell. Endocrinol. 3, 57-70. 12. de Jong, F.H. and Robertson, D.M. (1985) Mol. Cell. Endocrinol. 42, 95-105. 13. Bellvt, A.R. and Feig, L.A. (1984) Recent Progr. Horm. Res. 40, 531-567. 14. Jost, A. (1953) Recent Progr. Horm Res. 8. 379-418. 15. Wilson. J.D. and Gloyna, R.E. (1970) Recent Progr. Horm. Res. 26, 309-336. 16. Josso, N . (1977) Recent Progr. IHorm. Res. 33, 117-167. 17. Lyon, M.F. and Hawkes, S.G. (1970) Nature (London) 227. 1217-1219. 18. Imperator-McGinley. J., Guerro, J.. Gautier T. and Peterson, R . E . (1974) Science 186, 1213-1215. 19. Brown, J.L. and Ingram. V.M. (1974) J. Biol. Chern. 249. 3960-3972. 20. Mainwaring, W.I.P. (1986) In: The Role of Receptors in Biology and Medicine (Gott, A.M. and O’Malley, B.W.. eds.) pp. 79-90. Raven Press, New York. 21. Lane. S . E . , Gidari. A.S. and Levere. R . D . (1975) J . B i d . Chem. 250, 820%8213. 22. Booth, W . D . , Hay, M.F. and Dott. H . M . (1973) J . Reprod. Fertil. 33, 163-166. 23. Booth, W.D. (19x4) J . Endocrinol. 100, 195-202. 24. Paigen, K.. Swank, R.T., Tomino, S. and Gamshow, R . E . (1975) J. Cell Physiol. 85, 379-392. 25. Bullock, L.P., Bardin, C.W. and Sherman, M.R. (1978) Endocrinology 103, 176S1782.

.

.

195 26. Catterall, J.F., Kontula, K.K., Watson, C.S., Seppaner, P.J., Funkelstein, B., Melanitou, E . , Hickok, N.J., Bardin, C . W . and Janne, O . A . (1986) Recent Progr. Horm. Res. 42, 71-109. 27. Brock, M.L. and Shapiro, D.J. (1983) J . Biol. Chem. 258, 5449-5455. 28. Kontula, K.K., Torkelli, T.K.. Bardin, C.W. and Janne, O.A. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 731-735. 29. Kockakian. C.D. (1935) Proc. Soc. Exp. B i d . Med. 32, 1064. 30. Jung, I. and Baulieu, E-E. (1972) New Biol. Nature (London) 237, 2 4 2 6 . 31. di Pasquale, M . G . (1984) In: Drug Use and Detection in Amateur Sports (di Pasquale, M.G., e d . ) pp. 41-65. M G D Press, Ontario. 32. Kockakian. C.D. (1959) Lab. Invest. 8, 538-556. 33. Gustafsson, J-A.. Pousette, A . , Steinberg, A . and Wrange, 0. (1975) Biochemistry 14,3942-3948. 34. Brown, T . R . , Greene, F.E. and Bardin. C.W. (1976) Endocrinology 99, 1353-1362. 35. Schneider, G . , French, A , , Bullock. L.P. and Bardin, C.W. (1971) Endocrinology 89, 308-315. 36. Roy, A . K . , Chatterjee, B., Demyan, W.F., Milin. B.S., Motwani. N.M., Nath, T.S. and Schiop, M.J. (1983) Recent Progr. Horm. Res. 39, 425-461. 37. Roy, A.K.. Chatterjee, B., Prasad, M.S.K. and Unakar, N.J. (1980) J . Biol. Chem. 255,11614-11618. 38. Levi-Montalchini, R. and Calissano. P. (1979) Sci. Am. 240, 68-77. 39. Aloe, L . and Levi-Montalcini, R . (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 12461250. 40. Carpenter, G . and Cohen, S . (1979) Annu. Rev. Biochem. 48, 193-216. 41. Haigler, H . T . and Cohen. S . (1979) Trends Biochem. Sci. 4, 132-134. 42. Phoenix, C.H., Goy, R . W., Gerall, A . A . and Young, W.C. (1959) Endocrinology 65, 369-382. 43. McEwen, B.S. (1980) Annu. Rev. Physiol. 42. 97-110. 44. Arnold, A.P. and Gorski, R . A . (1984) Annu. Rev. Neurosci. 7, 413-442. 45. Mainwaring. W.I.P., Freeman, S.N. and Harper, B. (1987) In: Pharmacology and Clinical Use of Inhibitors of Hormone Secretion and Action, Ch. 6 (Furr. B.J.A. and Wakeling, A . E . , eds.) pp. 1 0 6 1 3 1 . Bailliere Tindall, London. 46. Liang, T . , Brady, E.J., Cheung, A . and Saperstein, R . (1984) Endocrinology 115, 2311-2317. 47. Hamilton, J.B. (1958) In: The Biology o f Hair Growth (Mantagna, W. and Ellis, R . A . , eds.) pp. 399-437. Academic Press. New York. 48. Ebling, F.J. (1958) J. Endocrinol. 15. 297-306. 49. Mainwaring, W.I.P. (1986) In: Introduction to the Cellular and Molecular Biology of Cancer, Ch. 4 (Franks, L.M. and Teich, N . , eds.) pp. 277-305. Oxford University Press, Oxford. 5 0 . Bruchovsky, N . , Rennie, P . S . , To. M . T . , Snoek. R., Lefebvre, Y.A. and Golsteyn. E.T. (1987) Prostate 10, 207-222. 51. Wenderoth, U . K . , George, F.W. and Wilson, J . D . (1983) Endocrinology 113, 569-573. 52. Parker, M.G. and Mainwaring. W.I.P. (1977) Cell 12, 401-407. 53. Bardin, C.W.. Brown. T., Isomaa, V.V. and Janne, O.A. (1984) Pharmacol. Ther. 23, 443-459. 54. Higgins, S.J., Burchell, J . M . and Mainwaring. W.I.P. (1976) Biochem. J . 160, 43-48. 55. McDonald, C . , Williams, L . , McTurk, P., Fuller. F . , McIntosh, E . and Higgins, S.J. (1983) Nuclcic Acid Res. 11, 917-930. 56. Garheri. J.C.. Kohanc, A.C.. Cameo. M.S. and Blaquier, J . A . (1979) Mol. Cell. Endocrinol. 13. 73-82. 57. Schweikert, H.U. and Wilson, J . D . (1975) J . Clin. Endocrinol. Metah. 40, 413-417. 58. Mainwaring, W.I.P., Furr, B.J.A. and Freeman. S.N. (1987) In Press: Proceedings of Symposium of the Clinical Management of Prostate Cancer, Nice. Raven Press, New York. 59. Freeman, S.N., Mainwaring, W.I.P. and Furr, B.J.A. (1986) J . Endocrinol. 111, S155. 60. Furr, B.J.A., Valcaccio, B., Curry, B.. Woodburn, J.R.. Chesterson, G . and Tucker, H . (1987) J . Endocrinol. 113. R7. 61. Neumann, F., Graf. K.-J., Hasan. S . H . . Schenck, B. and Steinbeck, H . In: Antiandrogens (Maritini, L . and Motta, M., eds.) pp. 163-177. Raven Press. New York.

196 62. Wilson, E.M. (1985) J . Biol. Chem. 260, 868S8689. 63. Gustafsson, J.-A., Carlstedt-Duke, J . , Poellinger, L., Okret, S . , Wilstrom, A,-C., Bronnegard, M. Gillner, M., Dong, Y., Fuxe, K., Cintra, A,. Harfstrand, A. and Agnati, L. (1987) Endocrinol. Rev. 8, 185-234. 64. Sanchez, E.R. and Pratt, W.B. (1986) Biochemistry 25, 1378-1382. 65. Khoury, G . and Gruss, P. (1983) Cell 33, 31S314. 66. Eliard, P.H., Marchand, M.J., Rousseau, G . G . , Formstecher, P., Mathy-Hartert, M., Belayew, A. and Martial, J.A. (1985) DNA 4, 409-417. 67. Rousseau, G.G. (1984) Biochem. J . 224, 1-12. 68. Geisse, S . , Scheidereit, C., Westphal, H.M., Hynes, N.E., Groner, B. and Beato, M. (1982) EMBO J. 1, 1613-1619. 69. Pfahl, M. (1982) Cell 31, 475-485. 70. Zaret, K.S. and Yamamoto, K. (1984) Cell 38, 29-38. 71. Parker, M.G. and Page, M.J. (1984) Mol. Cell. Endocrinol. 34, 159-168. 72. Darbre, P., Page, M. and King, R.J.B. (1986) Mol. Cell. Endocrinol. 6. 2847-2854.

B . A . Cookc. R.J.B. King and H.J. van dcr Molen (eds.) Hormones und their Actiuns. Purr I 01988 Elsevier Science Publishers BV (Biomcdical Division)

197 CHAPTER 12

Oestrogen actions ROBERT L. SUTHERLAND, COLIN K.W. WATTS* and CHRISTINE L. CLARKE** Garvan Institute of Medical Research, St Vincent’s Hospital, Sydney, New South Wales 2010, Australia

1. Introduction Oestrogens like the other major groups of steroid hormones, the androgens, progestins, glucocorticoids, mineralocorticoids, 1,25-dihydroxyvitamin D, and the ecdysones, are relatively rigid, liphophilic molecules, with molecular masses of approximately 300 Da. Oestrogens are distinguished from other steroids by the presence of a phenolic hydroxyl group at the 3 position of the steroid A ring (Fig. 1): this structure facilitates interaction with the oestrogen receptor (to the exclusion of other steroid hormone receptors) and thus determines oestrogenic activity. Evidence that oestrogens are evolutionarily ancient molecules is provided by the recent demonstration that such molecules are present in yeasts [l]. The major physiological role of oestrogenic steroids is in the control of reproduction in the female. Such a role covers a broad range of biological activities which in mammals include: the development of the reproductive tract and secondary sex organs, the regulation of the oestrus cycle including complex feedback loops with hypothalamic regulatory peptides and pituitary hormones, the development of the mammary gland and control of lactation, and the control of female reproductive behaviour. These steroids are also intimately involved in the development and control of replication of a number of common human cancers such as those of the breast and endometrium. A deeper knowledge of the mechanisms of oestrogen action has been provided by a number of useful models, e.g., growth responses of uteri in immature or ovariectomized animals, control of prolactin and growth hormone gene

*Present address: Hormone Biochemistry Department, Imperial Cancer Research Fund Laboratories, Lincoln’s Inn Fields, London, WC2A 3PX, United Kingdom. **Dr Clarke is an MLC-Life Research Fellow

198

HO

HO oestradiol

diethylstilboestrol (DES)

trianisylchlorethylene

4 hydroxytamoxifen (OH-TAM)

(TACEI

Fig. 1. Structures of the major physiological oestrogen, 17poestradio1, two synthetic oestrogens, diethylstilboestrol and trianisylchloroethylene and a hydroxylated high affinity triarylethylene antioestrogen, 4hydroxytamoxifen. Note that all compounds have a phenolic hydroxyl group in a position equivalent to the C3 hydroxyl of oestradiol (in the case of T A C E this occurs following demethylation in vivo). Despite the similarities between the triphenylethylene structure5 of T A C E and OH-TAM the latter molecule is an oestrogen antagonist, this property being conferred by the presence of the basic aminoether side chain common to all antioestrogens of this class. The clinically most important antioestrogen, tamoxifen, lacks the hydroxyl group of OH-TAM with a resultant 10-fold decrease in affinity for the oestrogen receptor.

expression in cultured normal and neoplastic pituitary cells, and the control of cell replication in human breast cancer cells. In reptiles, amphibia and birds oestrogens play a fundamental role in the control of egg production. In these groups of vertebrates oestrogens are intimately involved in the growth and development of the oviduct, the organ responsible for the synthesis of egg white proteins [ 2 ] , and in the control of synthesis of egg yolk proteins in the liver [3]. Because the relative abundance of these proteins has facilitated their purification, characterization and subsequent cloning and structural analysis of their respective genes, these models have not only provided valuable insight into the molecular mechanisms by which steroid hormones act but have also made considerable contributions to current knowledge on the control of eukaryotic gene expression [4,5]. The modern era of studies in oestrogen action began more than 25 years ago with two fundamental discoveries. First, Mueller and his colleagues at the University of Wisconsin demonstrated that the actions of oestrogens in promoting growth of the immature rat uterus were dependent upon new RNA and protein synthesis (reviewed in Ref. 6). These observations led to the hypothesis that the primary function of oestrogenic hormones was to regulate gene expression, a phenomenon which is only being fully investigated with the availability of the molecular genetic tools

199 of the 1980s. These observations, however, could not explain the tissue selectivity of oestrogen action. Considerable insight into this phenomenon was also forthcoming at about the same time with the synthesis of high specific activity tritiated derivatives of many small molecules. Jensen and his colleagues were the first to synthesize tritiated oestradiol, which was administered to immature rats and the accumulation into various organs monitored [7,8]. These experiments showed that tritiated oestradiol appeared in almost all tissues studied during the first hour of administration but was only retained for a protracted period by tissues like uterus and vagina which had long been known to be target organs for oestrogens. Furthermore, the radioactive steroid was retained in these tissues against a concentration gradient with blood. This and similar findings by others using tritiated derivatives of the synthetic oestrogen, hexoestrol, led to the postulate that cellular and tissue retention of oestrogens was mediated by a high affinity intracellular molecule, the oestrogen receptor (ER). The existence of this molecule was soon confirmed by a number of groups [7-lo], making it the first hormone receptor to be identified. Because of this historical precedent and the enthusiastic investigation by a number of impressive research groups the E R became a paradigm for the study of hormone receptors in general and steroid hormone receptors in particular. Prelabelling the E R by administration of ['H]oestradiol to immature rats followed by homogenization and differential centrifugation of target tissues demonstrated that the ER was located predominantly in the nuclear and cytosol fractions (Fig. 2). It was further demonstrated that the binding molecule had different sedimentation characteristics dependent upon which subcellular compartment it occupied. Thus the cytosol ER sedimented at about 8-10s upon sucrose gradient centrifugation and could be converted to 4s in the presence of 9 0.3 M KCl. In contrast, the form extracted from the crude nuclear-myofibrillar pellet sedimented at 5s in buffers containing concentrations of KC1 3 0.3 M. To explain these observations, Gorski and Jensen independently developed the 'two-step model' of oestrogen action which involved the cytoplasmic localization of receptor in the absence of hormone, with the administration of hormone resulting in binding to the 8s cytoplasmic receptor and a hormone-induced conformational change which resulted in translocation of the hormone receptor complex to the nuclear compartment [7,8,lo]. Thus by the late 1960s we had the basis for our current understanding of oestrogen action i.e. oestrogens were small lipophilic molecules that could readily diffuse into cells where they were retained if that cell was an oestrogen target cell, such a cell being defined by the presence of a cytosoluble protein, the ER, with high affinity for oestrogenic molecules. Following interaction with cytoplasmic E R the molecule underwent a conformational change which was associated with a change in sedimentation coefficient and increased affinity for polyanions including DNA. These properties resulted in translocation of ER from the cytoplasm to the nuclear compartment where, in some undefined manner, it regulated the rate of gene transcription leading to the production of new proteins which were responsible for changed

200 I

. TWO STEP MODEL

! NUCLEAR LOCALIFATION

ALTERED C i U . FUNCTION

Fig. 2. Schematic models of oestrogen action. Both the ‘two-step’ model (top left) and the ‘nuclear localization’ model (top right) are presented. In the ‘two-step’ model the 8s unoccupied receptor is located in the cytoplasm of the cell. Upon interaction with the ligand it undergoes transformation to a 5s form and translocation to the nuclear compartment where it is bound tightly to chromatin. In the ‘nuclear localization’ model the unoccupied receptor is present in the nuclear compartment and upon interaction with the ligand increases its affinity for chromatin and becomes tightly bound. In both cases tight nuclear binding leads to interactions with oestrogen regulatory elements 5’ to oestrogen responsive genes, leading t o changes in the rate of gene transcription. Messenger RNA is then processed and translated into new protein. It is the changes in the concentration of these intracellular proteins that changes the proliferation rate and/or differentiation state of the cell, which characterize the oestrogenic response.

cellular function including changes in cell proliferation and differentiation (Fig. 2). Since the time of these initial observations and the development of hypothetical models of oestrogen action several groups of workers have added a wealth of experimental data to further delineate the molecular mechanisms involved. Most of the data accumulated during the 1970s have been reviewed in detail elsewhere [4-121 and only the more important observations will be reiterated here. With the wider application of monoclonal antibody and molecular genetic techniques in the 1980s and their direct application to studies of oestrogen action, the past few years has seen the publication of major new information on the structure and subcellular distribution of the ER, and on the ways in which oestrogens control gene expression and cellular proliferation. It is these recent developments that will form the major emphasis of this chapter.

2. Oestrogen receptors It has been well established that oestrogen binds to an ER complex sedimenting at 8-10s in low-salt extracts of oestrogen responsive cells. Dissociation of this com-

201 plex and purification of E R has revealed a protein of molecular weight around 65-70000 in all species so far examined, which include human, bovine, porcine and rodent [13]. Covalent labelling of ER with tamoxifen aziridine has shown that this protein represents the steroid-binding moiety of E R (see Ref. 14 and references therein), and monoclonal antibodies directed against E R recognize a protein of molecular weight 66 000 in MCF-7 human breast cancer cells and in rat uterine cells [14]. The molecular weight of the E R detected by these biochemical methods is in good agreement with the molecular weight of around 66000 predicted more recently from the sequencing of E R cDNA clones [15-181. Furthermore, when ER cDNA was expressed in HeLa cells and immunoprecipitated with anti-ER monoclonal antibodies, a protein of molecular weight 65000 was detected [16]. The oestrogen binding site in E R has been examined recently: limited proteolysis of tamoxifen aziridine-labelled human, rat and porcine E R generated a hydrophobic fragment of E R with a molecular weight around 30 000 and full oestrogen binding activity [14,19]. This ‘meroreceptor’ form had been identified previously as the smallest receptor fragment to contain the steroid binding site [20]. These biochemical data have been supported by the recent construction and expression of deletion mutants of E R , which have localized the oestrogen binding domain to a 28000 molecular weight region in the C-terminal half of the E R molecule [21]. This relatively large domain has been analysed further by limited proteolysis of the tamoxifen aziridine-labelled 30 000 molecular weight fragment, yielding a fragment of molecular weight 6000 still containing the oestrogen binding site [14]. The way in which the well-characterized oestrogen binding protein of molecular weight 65-70 000 detected by denaturing gel electrophoresis assembles in solution to form the native, 8-10s complex is poorly understood, despite extensive efforts directed towards the elucidation of its structure. A recent suggestion has been that the oestrogen binding protein associates with non steroid binding proteins to form the 8-10s oligomer measured in vitro. These other proteins could serve to regulate the binding of ER to responsive elements of oestrogen responsive genes [22], and candidates for this role include p29, a human ER-associated 29000 molecular weight protein (see Ref. 23 and references therein) and the 90000 molecular weight heatshock protein which has been shown to associate with several steroid hormone receptors [24]. E R cDNA expressed in CHO cells also sediments at 8-9S, suggesting either that the complex is an oligomer of the hormone binding 60-70000 molecu’ fr weight protein, or that the putative non steroid binding proteins associated with the native ER are present in non target cells such as CHO cells [15]. Subsequent to oestrogen binding, ER undergoes a process (termed transformation) which is detectable in vitro as a change in its sedimentation coefficient from 8-10s to 4 5 S , resulting in increased affinity for DNA and polyanions (Fig. 2). The biological significance of this in vitro conversion is unknown, and the events surrounding transformation have been the subject of numerous studies. Conformational modification has been proposed as a consequence of transformation, al-

202 though recent studies have shown the structure of transformed and non-transformed ER to be similar [14]. Nevertheless, present evidence suggests that oestrogen binding results in the exposure of a pre-existing DNA binding domain (see Section 3). Enzymic modification of E R by phosphorylation has also been implicated as a consequence of oestrogen binding, and two models have been developed to investigate potential influences of protein kinases and phosphatases on E R . Auricchio and coworkers (see Ref. 25 and references therein) have postulated that a calmodulin stimulated protein kinase regulates oestrogen binding to calf and rat uterine E R by phosphorylation of E R on tyrosine. They have further described a nuclear phosphatase activity in oestrogen target tissues which inactivates oestrogen binding to ER, presumably leading to cessation of oestrogen action and recycling of E R . The second model has been developed by Smith and co-workers (see Ref. 26 and references therein) in the chick oviduct. Two forms of ER (designated Rx and Ry) with different affinities for oestrogen have been identified in this tissue, and different functional roles in oestrogen action have been ascribed to each form (see Ref. 27 and references therein). The antioestrogen 4-hydroxytamoxifen binds to the Rx form only, and fails to induce ovalbumin synthesis, a function ascribed to the Ry form. Interconversion between the Rx and Ry forms, and between the Ry and a non oestrogen binding form (Rnb), has been demonstrated after ATP/Mg2+treatment, presumably due to a phosphorylation reaction. The Rx and Ry forms of ER have also been identified in the human uterus. The intracellular localization of ER has been a much debated issue in recent years. Although it was generally accepted that ER was tightly bound to the nucleus in the presence of its ligand, most biochemical studies indicated that unoccupied E R was cytoplasmic (reviewed in Ref. 28). However, the enucleation of rat pituitary GH3 cells with cytochalasin B showed that ER was principally confined to the nucleoplast fraction: these observations were confirmed by the immunocytochemical demonstration that ER was exclusively nuclear regardless of the hormone status (reviewed in Ref. 28). The presence of ER in cytosol fractions is now generally regarded as an homogenization artefact resulting from dissociation of loosely bound nuclear ER. In the current model, ligand binding produces a conversion of the E R from a form with low affinity for nuclear components to one with high affinity (Fig. 2 ) . However, the nature of the association between ER and various nuclear components is poorly understood. Unoccupied E R has been localized to the dispersed euchromatin, and proposed intranuclear acceptor sites for ER include specific DNA sequences, nuclear matrix [29], basic and acidic non histone proteins and ribonucleoprotein (see Ref. 30 and references therein). The binding of ER to specific DNA sequences has been the subject of a number of studies, which will be discussed in Section 4.

203

3. Oestrogen receptor genes Understanding of E R structure and function has been greatly extended by the recent cloning and structural manipulation of ER cDNAs from a number of species including: human [15-17,21,22,31], chicken [17,18], rat [32], Xenopus [33] and mouse [ 341. Human E R (hER) cDNA clones were first isolated from Agtll cDNA libraries prepared from poly(Af) mRNA fractions enriched in ER mRNA from the human breast cancer cell line, MCF 7. Southern analysis of MCF 7 DNA probed with a 2.1 kb cDNA clone, OR8, showed that the E R gene was contained within approximately 40 kb of genomic sequences (indicating the presence of extensive intron sequences) and was most probably present as a single copy gene localized on chromosome 6 [31]. The OR8 cDNA clone contained an open reading frame of 1785 nucleotides coding for a 585 amino acid protein which corresponded to a molecular mass of 66200 Da. The complete hER mRNA contains a 5‘ untranslated region of 232 nucleotides, which includes a 20 codon open reading frame of unknown significance. Transfection of OR8-containing plasmids into C H O cells resulted in expression of an oestradiol binding, immunoreactive, 8-9s protein, confirming that the O R8 clone coded for the entire hER protein sequence [15]. Overall homology between the sequences of hER and cloned chicken E R (cER) is 80%. while that between hER and rat, mouse and Xenopus ER is 88, 89 and 69%, respectively. Three structural domains of the hER were initially defined from hydropathicity studies of the deduced amino acid sequence. These included an N terminal region of 120 amino acids, a hydrophilic region rich in cysteine, lysine and arginine residues postulated to contain a nucleic acid binding domain, and the C terminal half of the molecule which was mainly hydrophobic and thus a potential steroid binding domain. Chambon’s group has now defined six ‘functional’ regions within E R designated A-F (Fig. 3). Three regions were found to be highly conserved: region A , at the amino terminus, region C , the DNA binding site which is >98% conserved in all species studied and region E , the steroid binding domain which is >82% conserved between species. Regions B, D and F, the last forming the carboxy terminal end of the receptor, are less well conserved.

IAi

--

180

1 38

NH2

B

j

263302

C ID1

DNA BINDING

553 595CoOH

E

IF

hER

OESTROGEN BINDING

Fig. 3. Functional domains of human E R . Schematic rcpresentation of human E R , divided into six domains based o n the homology between human and chicken ER (see Ref. 21 for further details). Regions A . C and E are highly conserved in all specics \tuclied. Region C contains the DNA binding site and region E the oestrogen binding site. Thc numhcrs indicate the deduced amino acid positions at the boundary of each domain.

The only receptor regions to which functions have been clearly ascribed are those that contain the DNA (region C) and hormone (region E) binding sites. Region C (66 amino acids) is characterised by its high cysteine and other basic amino acid content, the number and location of which is extremely well conserved in all steroid hormone receptors. Similar sequence arrangements of cysteine, histidine and basic amino acids are found in the transcription factor TFIIIA, and these form Zn*+-coordinating loop structures (‘zinc fingers’) which constitute a nucleic acid binding protein domain, as described in detail elsewhere in this volume [35]. By analogy, sequences defining two such putative loops (non-identical) have been identified in region C of all steroid hormone receptors, although these sequences differ substantially from those of TFIIIA, containing for example two pairs of cysteine residues instead of one pair of cysteine and one pair of histidine residues. There has yet to be direct demonstration that such loop structures are formed in steroid hormone receptors, or that they are coordinated to Zn2+or other metals, or are involved in DNA binding. However, mutation studies with hER cDNA have shown that any deletion in this region abolishes DNA binding [21]. Similar studies with insertional mutants of the glucocorticoid receptor (GR) gene show abolition of receptor function [36]: given that the zinc fingers might still be able to form under such conditions these results may indicate that zinc fingers are not involved in DNA binding of steroid hormone receptors. However, it is more likely that even minor structural changes in this region cause disruption of function. Even though region C is highly conserved between different classes of steroid hormone receptors, chimeric receptors, in which region C from G R replaced region C of the ER, produced activation of a glucocorticoid-inducible gene (but not an oestradiol-inducible gene) upon oestradiol binding, indicating that receptor specificity resides in region C and that the hormone and DNA binding regions are functionally independent [21]. The identification of G R fragments which have tight nuclear binding activity and constitutive enhancer function in the absence of the hormone binding domain [37-391 provides further support for these observations. Region E is also highly conserved between E R of different species and between different classes of steroid hormone receptors. In hER region E is located between amino acids 301-552 (28 kDa). Mutation studies show that this is the only region required for oestradiol binding and that binding to the ligand with the expected affinity can take place in the absence of the N-terminal half of the molecule [21]. The amino acid sequence of this region predicts a secondary structure containing hydrophobic pockets, and presumably ligand specificity among the different classes of steroid hormone receptors is determined by the non-conserved amino acids in this domain. Knowledge of E R gene structure has led to some understanding of the mechanisms whereby hormone binding to receptor leads to tight nuclear binding of the hormone-receptor complex (receptor transformation). Present evidence suggests transformation corresponds to a conformational change occurring upon ligand

205 binding which serves to expose the pre-existing DNA binding domain. This conformational change could be mediated by region D . It has been postulated that this poorly conserved domain could function as a hinge between the flanking regions C and E [ 171. Amino acid deletions or insertions in this region can abolish tight nuclear binding in both E R and G R [21,36]. An alternative hypothesis is that ligand binding might unmask the DNA binding domain by causing dissociation of a separate receptor-associated protein. Candidates for this role have been suggested as outlined above (Section 2). Deletion mutations in either regions A or B have no effect on either oestradiol or nuclear binding [21]. Although the conservation of region A between species implies importance for receptor function, studies on the functional significance of these regions in E R have yet to be performed. However, mutant glucocorticoid receptors having amino acid insertions in the N-terminal sequence are defective in enhancer activation despite retention of DNA and hormone binding properties [36]. Similar observations have been made in the case of ‘increased nuclear transfer’ (nt’) glucocorticoid receptor mutants which are apparently truncated at the N-terminal and have been identified in glucocorticoid-resistant cell lines [40]. It is not yet clear whether such observations confirm a direct functional role of the A region in regulating transcription, or are instead the result of mutations in this region producing indirect effects on receptor structure and €unction.

4. Oestrogen control of gene expression Oestrogen control of specific gene expression in target cells is believed to be mediated through the E R . From studies on E R cDNA and others on structurally related genes for other steroid hormone receptors [41-44], the avian erythroblastosis virus v-erb A oncogene [4S] and its cellular thyroid hormone-binding counterpart, c-erb A [46,47], it is apparent that the ER is one of a family of transcriptional regulatory proteins. These proteins have in common a highly conserved DNA binding domain by virtue of which they are able to interact with cis-acting DNA promoter elements resulting in the activation of gene transcription. Steroid hormone receptors have been shown to bind preferentially to the 5’flanking portions of steroid sensitive genes, and recent work has focussed on the identification of 5’-sequences responsible for oestrogen control of gene expression. The cloning of oestrogen-sensitive genes, such as the chicken ovalbumin gene and the Xenopus laevis vitellogenin genes, and their transfer into heterologous expression systems has led to the more detailed identification of DNA sequences within the 5’-regions responsible for oestrogen sensitivity. When the chicken ovalbumin gene was transfected into the ER-containing, oestrogen-responsive human breast cancer cell line MCF-7, ovalbumin expression was increased 8-10-fold by oestrogen. Transfection of MCF-7 cells with a hybrid gene, consisting of the bacterial

206 xanthine-guanine phosphoribosyltransferase gene attached to a 4 kilobase chicken DNA which included 226 base pairs of the 5’-terminus of the ovalbumin gene and its upstream flanking sequence, showed that oestrogen responsiveness resides within the 5’-terminus and flanking sequence of the ovalbumin gene [48]. A short sequence located 140 base pairs upstream from the ovalbumin gene was shown to be conserved in several oestrogen responsive chicken egg white proteins, and a consensus sequence 5’-AAAAT(G/T)G(G/A)C-3’ is also present in all oestrogen-inducible genes in the chicken oviduct [48]. The role of this sequence in oestrogen control of gene expression is yet to be elucidated, and the dual influence of oestrogen and progesterone on egg white protein synthesis has been a complicating factor. The egg yolk precursor vitellogenin genes in the Xenopus laevis liver are under strict oestrogen control, and therefore have served as useful models to study oestrogen control of gene expression. Sequence analysis has shown that significant homology exists in the region upstream of the initiation site in the vitellogenin genes A l , A2, B1 and B2 [49]. Four blocks of sequence homology have also been demonstrated between these genes and the oestrogen-regulated vitellogenin and very low density apolipoprotein I1 (apo-VLDII) genes from the chicken [49]. A palindromic consensus sequence 5’-GGTCANNNTGACC-3’ is located in all cases within the region of highest homology (block 4). An ER binding site has been identified upstream of the chicken vitellogenin gene, between position -597 and -620, which also contains an oestrogen-dependent hypomethylation site and DNAse I sensitive sites [XI]. Transcription in embryonic liver in vitro of a hybrid gene containing the 5‘ region of the chicken vitellogenin gene showed modulation of transcription by ER [51]. Recent work on the Xenopus vitellogenin genes from two groups has extended these observations and identified the oestrogen-responsive DNA elements in two of the vitellogenin genes. In the study by Klein-HitBass et al. [52], the 5‘-flanking sequences (-821 to +14) of the vitellogenin A2 gene were inserted 5’ to sequences coding for the enzyme chloramphenicol acetyltransferase (CAT) and transfected into MCF-7 cells. CAT activity was shown to be under oestrogen control whether directed by the homologous vitellogenin promoter, or by a heterologous thymidine kinase promoter linked to vitellogenin A2 5’-flanking sequences. These studies showed that positions -482 to -87 of the vitellogenin sequence were sufficient to confer oestrogen responsiveness. Transfection of constructs, containing progressive 5’ and 3’ deletions of this region, linked to the thymidine kinase promoter and the CAT structural sequences showed that 35 base pairs located at positions -331 to -297 were sufficient for oestrogen inducibility of CAT activity. This sequence, termed the Estrogen Regulatory Element (ERE), is located within homology block 4 previously identified in Xelzopus and chicken vitellogenin genes [49], and contains the palindromic core 5’-GGTCACAGTGACC-3’. These results demonstrated that only one of the sequence homology blocks previously described within vitellogenin

207 genes [49] is required for oestrogen regulation of vitellogenin gene expression. Similar results have been obtained by Seiler-Tuyns et al. in a recent study [53],in which MCF-7 cells were transfected with vitellogenin gene B1 5'4anking sequences linked to CAT structural sequences. A region from position -334 to -297 in the vitellogenin B1 ti'-flanking sequence was shown to confer oestrogen responsiveness. Apart from its proposed role as a transcriptional regulatory factor, E R may also be implicated in post-transcriptional regulation. The administration of oestrogen increases the half-life of chicken ovalbumin, vitellogenin, apo-VLDII mRNAs and Xenopus vitellogenin mRNA from 16 h to 500 h [54]. There is concomitant destabilization of Xenopus serum albumin mRNA by oestrogen. These effects are independent of protein synthesis, and are probably mediated by ER as evidenced by the demonstration that the antioestrogen, 4-hydroxytamoxifen, prevents these oestrogenic effects.

5. Oestrogen control of cell proliferation The proliferation of secondary sex organs in immature or oophorectomised female vertebrates in response to oestrogens has been known for decades, indeed this response in the mouse uterus formed the basis of the first accurate bioassay for compounds with oestrogenic activity. It has also been known since the end of the last century that the ovaries, and oestrogens in particular, are intimately involved in the regulation of proliferation of human breast cancers. Despite the early development

Fig. 4. Schematic model of the mechanisms of oestrogen control of cell proliferation. Three different mechanisms are illustrated. In (1) the interaction of oestrogen (E) with ER leads to increased transcription of genes whose products are directly involved in the control of cell replication. The mechanism illustrated in (2) postulates that oestrogens modulate the production of autocrine growth factors which in turn bind to growth factor receptors at the cell surface and mitogenesis occurs as a consequence of growth factor-activated metabolic pathways. The underlying hypothesis in ( 3 ) is that cells are under inhibitory (I) control by undefined molecules in the extracellular Huid and that oestrogens block the effects of these inhibitory molecules.

of this knowledge (reviewed in Ref. 5 5 ) the molecular mechanisms responsible for these proliferative responses have not been delineated and there is now considerable debate as to whether oestrogenic effects on proliferation are mediated directly through ER-mediated changes in the transcription of specific genes critical to cell cycle progression and DNA synthesis, indirectly through endocrine, autocrine or paracrine mechanisms [56,57], or indirectly by the release from inhibition by serumborne growth inhibitory factors [58] (Fig. 4). In the 1960s and 70s studies were concentrated on the effects of oestrogens on DNA synthesis and cell cycle kinetic parameters in rodent reproductive organs utilizing the techniques of tritiated thymidine incorporation in vivo and autoradiography. These studies provide the current basis of our knowledge and lead to the general conclusion that in reproductive tissues oestrogens increase the rate of cell proliferation by: increasing the size of the proliferating cellular pool due to recruitment of non-cycling, quiescent or G,, cells, into the cell cycle thereby increasing the growth fraction; shortening the overall cell cycle time by a reduction in the length of G I phase and, to a lesser extent, of S phase and perhaps G, + M; and by a decrease in the cell death rate [SS]. The first evidence, that oestrogenic effects similar to those seen in vivo could be reproduced in vitro, came from studies on primary cultures of rabbit endometrium grown in chemically defined medium and demonstrated that diethylstilboestrol increased the proportion of cycling cells, and also significantly shortened the generation time by decreasing the length of both GI and S phases. Such a preliminary result does not support the view that oestrogenic effects are mediated by secondary endocrine or paracrine effects or by influences on serum-borne growth inhibitory factors. The demonstration of a relationship between the presence of ER in human breast tumours and their ability to respond to some form of endocrine therapy, and the subsequent widespread use of the synthetic nonsteroidal antioestrogen, tamoxifen, in the treatment of human breast cancer, led to a renaissance in studies of the oestrogenic control of cell proliferation especially in breast cancer cells. This was facilitated by the development of a number of human breast cancer cell lines, some of which responded to oestrogen in vitro with increased cell proliferation rates. The few published studies on the effects of oestrogen on breast cancer cell cycle progression in vitro, and the more detailed studies of the cell cycle kinetic effects of antioestrogens [59], support the view that the major control of cell cycle progression by oestrogens in tissue culture is in G I phase, in agreement with earlier studies with in vivo model systems. The concept that oestrogens stimulate cell proliferation directly arises mainly from the observation that physiological concentrations of oestrogens stimulate both the de novo and salvage pathways of DNA synthesis as well as inducing a number of enzymes intimately involved in DNA synthesis and including: DNA polymerase, thymidine and uridine kinases, thymidilate synthetase and dihydrofolate reductase. There is evidence that some of these enzymes may be regulated at the transcrip-

209

tional level [57].Such data do not, however, allow distinction of a direct effect of oestrogen on the transcription of genes critical to cell cycle progression from effects secondary to activation of some other second messenger system, including autocrine growth factor production. Recent studies on the regulation of eukaryotic cell replication indicate that control of entry into S phase may be controlled by a small number of highly conserved genes [60]. To date the effects of oestrogen on the expression of such genes have not been investigated. Clearly, the type of proof required to substantiate a direct effect of oestrogen on cell replication would be demonstration of binding of ER or an ER-induced gene product, to regulatory elements of such genes, leading to increased transcription of these genes and increased rates of cell cycle progression and cell replication. Until such proof is forthcoming the direct effects of oestrogens on cell replication will remain speculative and controversial. With the recent explosion of knowledge in the area of growth factor control of cell replication, especially in fibroblasts where distinct competence and progression factors have been described, and the prior knowledge that oestrogens and antioestrogens control cell cycle progression especially through G , phase, it is therefore not surprising that attention has recently been focussed on the involvement of autocrine and paracrine factors in the proliferative response to oestrogens. It must be appreciated, however, that the growth factor control of the cell cycle of epithelial cells is poorly understood and may differ appreciably from that in fibroblasts. Despite these potential limitations, considerable exciting work has emanated from the laboratory of Lippman where he and his co-workers have demonstrated that human breast cancer cells produce a number of autocrine growth factors with properties similar to epidermal growth factor (EGF), insulin-like growth factors (IGF), platelet-derived growth factor (PDGF) and an epithelial cell colony stimulating factor. Furthermore, it has been demonstrated that the production of some of these factors is regulated by oestrogens and antioestrogens. Thus it seems apparent that the growth of human breast cancer cells in vitro is determined to a large extent by the delicate balance between growth stimulatory (aTGF,IGF-I) and growth inhibitory (pTGF) peptides secreted by the cells themselves and that the balance of production of these autocrine factors is controlled in part by oestrogens and antioestrogens. Lippman has demonstrated that oestrogens stimulate the production of aTGF and IGF-I and inhibit the production of /3TGF while antioestrogens have the opposite effects facilitating production of the growth inhibitory factor at the expense of growth stimulatory factors [57].The relevance of these types of growth control mechanisms in the proliferative response of normal cells to oestrogens has yet to be elucidated. In fact the observation that conditioned medium from MCF 7 cells treated with oestrogen, increased the growth of MCF 7 tumours in nude mice without stimulating proliferation in the uterus argues against the same mechanisms being operative in normal oestrogen target tissues [61]. In addition to the regulation of growth factors, oestrogens also regulate the pro-

210 duction of many other proteins some of which, for example, plasminogen activator and other proteases, may have roles in proliferative responses especially in liberating growth factors from large molecular weight precursors or carrier proteins. Finally, Sonnenschein and colleagues have postulated that the effects of oestrogen on cell proliferation are mediated by releasing cells from the growth inhibitory effects of specific growth inhibitory molecules normally present in serum (Fig. 4). Although such growth inhibitory molecules have yet to be identified this group has accumulated significant experimental data which have recently been reviewed [ 5 8 ] .

6. Antioestrogen actions Oestrogen antagonists have been invaluable tools for defining the mechanisms of oestrogen action at the molecular level. Although apparently ER-mediated, a precise understanding of the molecular events underlying oestrogen antagonism is far from complete. The action of antioestrogenic compounds is made extremely complex by the expression of partial or full agonist as well as antagonist properties; the activity expressed by a given compound is determined by the species, tissue or oestrogen-responsive parameter studied. Synthetic nonsteroidal compounds, in particular triarylalkene and triarylalkane derivatives such as the antitumour agent tamoxifen (Fig. l ) , comprise the most studied group of oestrogen antagonists. The biological and pharmacological properties of these compounds and the new generation of steroidal antioestrogens are described elsewhere in this volume [62]. Structure-activity studies have defined those structural features of the antioestrogenic ligand responsible for the expression of oestrogen antagonist and agonist activities and those which are responsible for potency [63]. Potency is largely determined by binding affinity for E R . Compounds such as the triphenylethylenes have planar trans-stilbene structures similar to the conformation of oestradiol and thus have relatively high affinity. Affinity is especially influenced by the presence of a phenolic hydroxyl group at an equivalent position to the C3 phenol of oestradiol (see Fig. 1). Agonist and antagonist activities are determined by certain side chain substitutions extending out of the plane of the molecule. Unsubstituted compounds are full oestrogen agonists whereas nonsteroidal compounds with basic alkylamino ether side chains in the correct orientation are almost invariably antioestrogens (Fig. 1). This is of necessity an over-simplification, and fails, for example, to provide any explanation for species, tissue and in vivo versus in vitro differences in pharmacological activity. At the molecular level antioestrogens act as competitive inhibitors of oestrogen binding to the E R with consequent formation of antioestrogen-ER complexes that no longer possess proper receptor function [62-661. Initially it was believed that oestrogen antagonism arose from low affinity interaction of the antioestrogen with

21 1 ER and consequent premature dissociation of ligand resulting in failure to express agonist activity. Antioestrogens must act through other mechanisms, however, given that potent hydroxylated nonsteroidal antioestrogens have affinities for ER at least as high as that of oestradiol. In fact, no kinetic differences between agonist and antagonist binding to E R have yet been found that could explain antioestrogenic activity. Many early studies showed failure of antioestrogens to stimulate cytoplasmic E R replenishment and it was proposed that oestrogen antagonism was therefore due to the inhibition of E R resynthesis. However, recent studies using dense amino acid labelling techniques to follow synthesis and degradation of E R in MCF 7 cells demonstrated that antioestrogens do not prevent E R synthesis nor d o they accelerate or block receptor degradation [67]. Current thinking is that differences in the conformation of oestrogen-ER and antioestrogen-ER complexes could account for antioestrogenic activity, with the implication being that oestrogen and antioestrogen promote different molecular orientations of the E R protein. Polyclonal antibodies to calf uterine E R decrease affinity of unlabelled E R for oestradiol but not 4-hydroxytamoxifen suggesting that the antibody prevents the conformational change which occurs upon oestradiol but not 4-hydroxytamoxifen binding [63,66]. Differences in the physical properties of the nuclear salt-extracted ligand-ER complexes from MCF 7 cells have also been noted, high affinity antioestrogens forming higher molecular weight complexes with E R than oestradiol [68,69]. Digestion of chick oviduct chromatin with micrococcal nuclease has shown that oestradiol, but not tamoxifen or 4-hydroxytamoxifen, induces a specific 13-14s E R peak o n sucrose gradients [70].Thus, by this criterion, nuclear binding of antioestrogen-ER complexes to chromatin appears qualitatively different to that of oestrogen-ER complexes. Hopefully, the recent cloning and in vitro expression of the E R cDNA will soon provide sufficient E R protein to enable direct studies of the tertiary structure of ligand-ER complexes and provide a deeper understanding of antioestrogen action at the molecular level. Little is known about the variety and quantity of mRNA transcripts produced in the presence of the antioestrogen-ER complex. Tamoxifen inhibits accumulation of ovalbumin and conalbumin mRNA after administration to oestrogen stimulated chickens or upon simultaneous administration with oestrogen [71,72]. In contrast the mRNA synthesis induced by dexamethasone and progesterone is amplified by tamoxifen. A good correlation is found between levels of gene transcription and rates of conalbumin and ovalbumin protein synthesis, indicating that these actions of tamoxifen involve effects at the transcriptional level, although tamoxifen alone has no effect on transcription of these genes. Similarly tamoxifen alone fails to induce oestrogen-responsive pS2 and thymidine kinase mRNAs in breast cancer cells [73,74]. In other cases post-transcriptional effects of antioestrogens have been observed. Antioestrogens induce low levels of vitellogenin mRNA in Xenopus liver even

212 though they are able to suppress oestrogen-induced transcription. Intracellular but not serum vitellogenin levels are increased indicating effects on post-translational or secretory processes [75].In MCF 7 cells the antioestrogen induction of P T G F secretion, while apparently ER-mediated, does not seem to occur at the level of mRNA which is unchanged by antioestrogen treatment. The point of control appears to be post-transcriptional and possibly involves changes to the rate of translation [57]. A number of the diverse biological effects of antioestrogens are apparently not mediated by E R , for example the cytotoxic activity of high concentrations of drug in vitro, and may result instead from the binding of these compounds to a variety of other cellular components. Several interactions are potentially important within the range of serum concentrations achieved during antioestrogen therapy. A high affinity (Kd for tamoxifen = 1 nM), saturable binding site for nonsteroidal antioestrogens (AEBS) has been widely studied (see Ref. 76 and references therein) and shown to be an intrinsic microsomal membrane protein, with wide tissue distribution, binding not only nonsteroidal antioestrogens substituted with side chains terminating in basic alkyl amino groups, but also a wide variety of structurally related compounds with a range of biological activities. The AEBS has no affinity for oestradiol or other natural or synthetic steroids. Putative low affinity natural ligands for this site have been partially characterized. There has been considerable interest in the function of the AEBS especially as a potential mediator of antioestrogenic activity although there now appears to be little evidence that it is directly involved in such activity. Antioestrogens are also known from in vitro studies to have high affinity interactions (acting as antagonists) with histamine H, ( K , = 10-' M), dopamine D, (k, for inhibition of spiperone binding = lo-' M), and muscarinic cholinergic receptors (Kd = M) (see Ref. 77 and references therein). Nonspecific (local anaestheticor detergent-like) interactions with plasma or intracellular membranes resulting in membrane disruption and changes in permeability have also been described [78,79]. Intracellularly, antioestrogens interact with cytochrome P-450 resulting in antioestrogen metabolism and both inhibition ( k , = lops M) and induction of the enzyme [80,81].Tamoxifen and related antioestrogens have also been shown to inhibit both protein kinase C and calmodulin at low micromolar concentrations [82-841. These interactions may underlie oestrogen-irreversible effects on cell growth and cytotoxicity observed in vitro at antioestrogen concentrations > lN.

7. Conclusions Although the basis of our current thinking on the mechanisms of action of oestrogens has been with us for over 20 years, the molecular genetic techniques developed in the past decade have complemented more classical biochemical approaches

213 and facilitated a greatly enhanced understanding of the E R and oestrogen action. In particular, the cloning of E R from a number of species has shown that these genes belong to a much larger family of genes coding for transcriptional regulatory proteins, and the identification of specific regulatory sequences that bind ER and confer oestrogen specificity to transcription of these genes, has markedly increased our understanding of oestrogen action at the molecular level. Extension of this technology will allow expression of the ER in bacterial and mammalian cell hosts in vitro and guarantee supply of sufficient quantities of protein for detailed structural analysis. This should soon allow a much deeper understanding of both ER structure and the specific interactions between oestrogenic and antioestrogenic ligands, of known agonist and antagonist activity, and specific amino acid sequences within the protein. Such information is likely to be of immense importance, not only in developing tools for basic studies on oestrogen action, but also in the rational design of molecules of potential importance in the treatment of human diseases including cancer.

Ac kno wledgernents Research in this laboratory is supported by grants from the National Health and Medical Research Council of Australia, the New South Wales State Cancer Council, Leo and Jenny Leukaemia and Cancer Foundation of Australia and MLC-Life Ltd.

References

.

1. Feldman, D.. Tokes, L.G. Stathis, P.A.. Miller, S.C., Kurz. W. and Harvey, D. (1984) Proc. Natl. Acad. Sci. U.S. A . 8 1, 4722-4726. 2. Schimke, R.T.. McKnight, G.S.. Shapiro, D . J . . Sullivan, D. and Palacios, R . (1975) Rec. Progr. Horm. Res. 31, 175-211. 3. Clemens, M.J. (1974) Progr. Biophys. Mol. Biol. 28, 71-107. 4. Yamamoto. K.R. and Alberts, B.M. (1976) Ann. Rev. Biochem. 45,721-746. 5 . Yamamoto, K.R. (1985) Ann. Rev. Genet. 19. 209-252. 6. Mueller, G.C., Vonderhaar, B.. Kim, U.H. and Lc Mahieu. M. (1972) Rec. Prog. Horm. Res. 28. 1-49. 7. Jensen, E.V. and DeSombre, E . R . (1972) Ann. Rev. Biochem. 41. 203-230. 8. Jensen. E . V . and DeSombre, E . R . (1973) Science 182. 12G.133. 9. Baulieu, E.-E.,Atger, M . , Best-Belpomme, M.. Corvol, P., Courvalin, J.-C., Mester, J . , Milgrom, E., Robel, P., Rochefort, H. and De Catalognc. D. (1975) Vitamins Hormones 33, 649-736. 10. Gorski, J . and Cannon, F. (1976) Ann. Rev. Physiol. 38. 425-450. 11. Chan, L. and O’Malley. B.W. (1976) N. Engl J. Med. 294. 1322-1328, 1372-1381, 143K-1437. 12. Katzenellenbogen, B.S. (1980) Ann. Rev. Physiol. 42. 17-35. 13. Lubahn, D.B., McCarty. K.S. J r and McCarty. K.S. Sr. (1985) J . Biol. Chem. 260,2515-2526. 14. Katzenellenbogen, B.S., Elliston. J.F.. Monsma, F.J.. Springer, P. A , , Ziegler, Y .S. and Greene, G.L. (1987) Biochemistry 26, 23642373.

214 15. Greene, G.L., Gilna, P . , Waterfield, M., Baker, A . , Hort. Y . and Shine, J . (1986) Science 231, 115~1154. 16. Green, S . , Walter, P . , Kumar, V.. Krust, A . , Bornert. J.-M., Argos. P . and Chambon, P. (1986) Nature 320, 134-139. 17. Krust, A., Green, S., Argos, P . , Kumar. V., Walter, P., Bornert, J.-M. and ChLmbon, P. (1986) EMBO J. 5 , 891-897. 18. Maxwell, B.L., McDonnell, D.P., Conneely, O.M., Schultz, T.Z., Greene, G.L. and O’Malley, B.W. (1987) Mol. Endocrinol. 1, 25-35. 19. Koike. S., Nii. A.. Sakai, M. and Muramatsu, M. (1987) Biochemistry 26, 2563-2568. 20. Sherman, M.R. and Stevens. J . (1984) Ann. Rev. Physiol. 46, 83-105. 21. Kumar, V . , Green, S . , Staub. A . and Chambon, P. (1986) EMBO J . 5 , 2231-2236. 22. Green, S. and Chambon. P. (1986) Nature 324, 615-617. 23. King, R.J.B., Cano, A . , Finley, J . and Coffer, A.I. (1986) J . Steroid Biochem. 24, 369-372. 24. Joab, I., Radanyi, C . , Renoir, M., Buchou, T., Catelli, M.-G., Binart, N.. Mester, J . and Baulieu, E.-E. (1984) Nature 308, 850-853. 25. Migliaccio. A . , Rotondi. A . and Auricchio, F. (1986) EMBO J. 5 , 2867-2872. 26. McNaught. R.W., Raymoure. W.J. and Smith, R . G . (1986) J . Biol. Chem. 261, 17011-17017. 27. Raymoure, W.J., McNaught, R.W., Greene. G . L . and Smith, R.G. (1986) J. Biol. Chem. 261, 170 18- 17025. 28. Gorski. J., Welshons, W.V., Sakai, D., Hansen, J., Walent, J., Kassis, J . . Shull, J., Stack, G. and Campen, C. (1986) Rec. Progr. Horm. Res. 42. 297-329. 29. Alexander, R.B., Greene, G.L. and Barrack, E.R. (1987) Endocrinology 120, 1851-1857. 30. Greene, G.L. and Press, M.F. (1986) J . Steroid Biochem. 24, 1-7. 31. Walter, P., Green. S . , Greene, G . , Krust. A.. Bornert, J.-M., Jeltsch, J.-M., Staub, A , , Jensen, E., Scrace, G., Waterfield. M. and Chambon, P. (1985) Proc. Natl. Acad. Sci. U.S.A. 82,7889-7893. 32. Koike, S., Sakai, M. and Muramatsu, M. (1987) Nucleic Acids Res. 15, 2499-2513. 33. Weiler, I.J.. Lew, D . and Shapiro, D.J. (1987) Mol. Endocrinol. 1, 355-362. 34. White, R . , Lees. J.A., Needham. M., Ham. J. and Parker, M. (1987) Mol. Endocrinol. 1,735-744. 35. Parker, M.G. (1988) This volume. 36. Giguere. V.. Hollenberg. S.M.. Rosenfeld, M.G. and Evans, R.M. (1986) Cell 46, 645-652. 37. Godowski, P.J., Rusconi. S., Miesfeld, R. and Yamamoto, K.R. (1987) Nature 325, 365-368. 38. Rusconi, S. and Yamamoto, K.R. (1987) EMBO J . 6. 1309-131s. 39. Miesfeld, R . , Godowski, P.J., Maler, B.A. and Yamamoto, K.R. (1987) Science 236, 423-427. 40. Northrop, J.P.. Danielsen, M. and Ringold, G.M. (1986) J. Biol. Chem. 261, 11064-11070. 41. Hollenburg, S.M., Weinberger. C., Ong, E.S., Cerelli, G . , Oro, A , , Lebo, R., Thompson, E.B., Rosenfeld, M.G. and Evans, R.M. (1985) Nature 318, 635-641. 42. Conneley, O.M., Sullivan, W.P., Toft. D.O., Birnbaumer, M., Cook, R.G., Maxwell, B.L., Zarucki-Schulz, T., Greene, G.L., Schrader, W.T. and O’Malley, B.W. (1986) Science 233,767-770. 43. Jeltsch, J.M., Krozowski. Z., Quirin-Stricker. C., Gronemeyer, H . , Simpson, R.J., Gamier, J.M., Krust, A , , Jacob, F. and Chambon, P. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 5424-5428. 44. McDonnell, D.P., Mangelsdorf, D.J., Pike, J.W., Haussler, M.R. and O’Malley, B.W. (1987) Science 235, 12141217. 45. Debuire. B., Henry, C., Benaissa, M., Biserte, G., Claverie, J.M., Saule, S . , Martin, P. and Stehelin, D. (1984) Science 224, 145g1459. 46. Sap, J . , Munoz, A , . Damm, K.. Goldberg. Y., Ghysdael, J., Leutz, A , , Beug, H. and Vennstrom, B. (1986) Nature 324, 635-640. 47. Weinberger, C., Thompson, C.C.. Ong, E.S., Lebo, R . , Gruol, D.J. and Evans, R.M. (1986) Nature 324, 641-646. 48. Lai, E.C., Riser, M.E. and O’Malley, B.W. (1983) J. Biol. Chem. 258, 12693-12701. 49. Walker, P., Germond, J.-E., Brown-Luedi. M. Givel, F. and Wahli, W. (1984) Nucleic Acids Res. 12. 8611-8626.

215 50. 51. 52. 53.

Jost, J.-P., Seldran, M. and Geiser, M. (1984) Proc. Natl. Acad. Sci. U.S.A. 81,429-433. Jost, J.-P., Geiser, M. and Seldran, M. (1985) Proc. Natl. Acad. Sci. U.S.A. 82. 988-991. Klein-HitRass, L . , Schorpp, M., Wagner, U. and Ryffel, G.U. (1986) Cell 46, 1053-1061. Seiler-Tuyns, A., Walker, P., Martinez, E . , Mcrillat, A.-M., Givel. F. and Wahli, W. (1986) Nucleic Acids Res. 14, 8755-8770. 54. Riegel, A.T., Aitkin, S.C., Martin, M.B. and Schoenburg, D.R. (1987) Mol. Endocrinol. 1, 160-166. 55. Sutherland, R.L., Reddel, R . R . and Green, M.D. (1983) Eur. J . Cancer Clin. Oncol. 19,307-318. 56. Lippman, M.E., Dickson, R.B.. Bates, S., Knabbe, C . , Huff, K., Swain, S., McManaway, M.. Bronzert, D . , Kasid, A. and Gelmann. E.P. (1986) Breast Cancer Res. Treat. 7, 59-70. 57. Dickson, R.B. and Lippman, M.E. (1987) Endocr. Rev. 8,29-43. 58. Sato, A.M. and Sonnenschein, C. (1987) Endocr. Rev. 8, 44-52. 59. Sutherland, R.L., Murphy, L.C., Hall, R.E.. Reddel, R.R., Watts, C.K.W. and Taylor, I.W. (1984) Progr. Cancer Res. Ther. 31, 193-212. 60. Murray, A.W. (1987) Nature 327, 14-15, 61. Dickson, R.B., McManaway, M. and Lippman, M.E. (1986) Science 232, 1540-1542. 62. Wakeling, A.E. (1988) this volume. 63. Jordan, V.C. (1984) Pharmacol. Rev. 36. 245-276. 64. Sutherland, R.L. and Jordan, V.C. (1981) Non-Steroidal Antioestrogens. Academic Press, Sydney. 65. Sutherland, R.L. and Murphy, L.C. (1982) Mol. Cell. Endocrinol. 25, 5-23. 66. Jordan, V.C. (1986) EstrogeniAntiestrogen Action and Breast Cancer Therapy. University of Wisconsin Press, Madison. 67. Eckcrt, R . L . , Mullick, A., Rorke, E . A . and Katzenellenbogen, B.S. (1984) J. Biol. Chem. 114, 629-637. 68. Eckert, R.L. and Katzenellenhogen. B.S. (1982) J . Biol. Chem. 257. 8840-8846. 69. Miller, M.A., Mullick, A . , Greene. G.L. and Katzenellenbogen, B.S. (1985) Endocrinology 117, 5 15-522. 70. Lebeau. M.C., Massol, N. and Baulieu, E.-E. (1982) Biochem. J . 204, 653-662. 71. Palmiter, R.D., Mulvihill. E.R.. McKnight. G.S. and Senear. A.W. (1977) Cold Spring Harbor Symp. Quant. Biol. 42, 639-647. 72. Schweizer, G., Cadepond-Vincent, F. and Baulieu, E . - E . (1985) Biochemistry 24, 1742-1749. 73. Westley, B.. May, F.E., Brown. A.M.C., Krust, A . , Chamhon, P., Lippman, M.E. and Rochefort, H . (1984) J. Biol. Chem. 259, 1003(L10035. 74. Kasid, A . , Davidson, N.E., Gelmann. E.P. and Lippman. M.E. (1986) J . Biol. Chem. 261, 5562-5567. 75. Riegel, A.T., Jordan, V.C., Bain. R . R . and Schoenhcrg, D . R . (1986) J . Steroid Biochem. 24, 1141- 1149. 76. Watts, C.K.W.. Murphy, L.C. and Sutherland. R.L. (1986) In: EstrogeniAntiestrogen Action and Breast Cancer Therapy (Jordan, V.C. ed.) pp. 93-1 14. University of Wisconsin Press. Madison. 77. Watts, C.K.W. and Sutherland, R.L. (1987) Mol. Pharmacol. 31, 541-551. 78. Morris, I.D. (1985) Life Sci. 37, 273-278. 79. Ramu. A , , Glaubiger, D . and Fuks, Z . (1984) Cancer Res. 44, 4392-4395. 80. Ruenitz. P.C. and Toledo, M.M. (1980) Biochem. Pharmacol. 29. 1583-1587. 81. Meltzer. N.M., Stang, P., Sternson, L.A. and Wade, A.E. (1984) Biochem. Pharmacol. 33, 115-123. 82. Lam, H.-Y.P. (1984) Biochem. Biophys. Res. Commun. 118, 27-32. 83. Tallant, A.E. and Wallace, R.W. (1985) Biochem. Biophys. Res. Commun. 131, 370-377. 84. O’Brian, C.A., Liskamp, R.M., Solomon, D.H. and Weinstein, I.B. (1985) Cancer Res. 45, 2462-2465.

This Page Intentionally Left Blank

B . A . Cookc, R.J.B. King and H.J. van der Molcn ( c & ) Hurlnones und their Actions. Purr 1 01988 Elsevier Science Publishers BV (Bionicdical Division)

217 CHAPTER 13

Glucocorticoid receptor actions ULRICH GEHRING lnstitut f u r Biologische Cheniie der Universitat, lm Neuenheimer Feld 501, 6900 Heidelberg, F. R. G.

1. Introduction The physiological actions of glucocorticoids in mammalian organisms are so manifold that a complete treatment of this subject would certainly go much beyond the scope of a single contribution. In fact, there are probably only few cell types which do not respond in one way or another to glucocorticoids while the effects of other steroids, notably the sex hormones are much more limited in their ranges of target tissues. The reader is therefore referred to a series of review articles [l-151 including a book on glucocorticoid hormone action [16] by which various aspects of the subject are being covered. The main emphasis of this chapter is on the specific receptors which are the mediators for glucocorticoids in target cells. Classically, there are two major targets for glucocorticoids: liver and the lymphatic system. In hepatic cells certain enzymes involved in gluconeogenesis (hence the name glucocorticoid) and in amino acid degradation are inducible by these corticosteroids while in lymphoid cells a host of inhibitory reactions may occur which eventually may culminate in cell death. This latter type of response has been used for obtaining variant cells which no longer respond to the hormone and which harbour defective receptors. The study of mutant receptors obtained in this way has contributed significantly to our present understanding of glucocorticoid receptor actions.

2. Glucocorticoid induced lymphocytolysis Various cells of lymphatic origin, most prominently immature thymocytes and certain leukaemia and lymphoma cells respond to high physiological or pharmacological doses of glucocorticoids by growth inhibition and cell lysis [8,12,17-251. In some cells the response is limited to the inhibition of cell proliferation, in others an initial phase of diminished growth with preferential accumulation of cells in the G , phase

218 of the cell cycle is followed by cell death. Even though this lymphocytolytic effect has been known for a long time [26] the biochemistry behind it is not yet understood. Various effects may precede growth inhibition, for example the inhibition of transport of nutrients into cells, decreased activities of polyamine synthesizing enzymes, decreased intracellular levels of polyamines and inhibition of RNA and protein synthesis, but their cause-effect relationships - if any - are not clear. The primary biochemical event which triggers all these cell inhibitory effects has not been identified unequivocally. In principle, one might envisage two completely different mechanisms: The induction of a specific gene may be involved with the gene product expressing suicidal lysis effects or the transcription of one or several genes of vital function may be turned off [ 12,231. There is in fact precedence for hormonally controlled repression of genes; thus glucocorticoids cause decreased expression of the pituitary specific gene product prolactin [27,28] and pro-opiomelanocortin [29-311. If genes coding for vital functions in lymphoid cells were turned off by the hormone, mutations in such genes would be detrimental to the cells and could therefore not be recovered, except perhaps as conditionally lethal mutants. On the other hand, there is evidence, although still limited, for a lysis function to exist. Mouse lymphoma cells of line SAK8 normally respond to glucocorticoids by growth inhibition without lysis but after treatment with 5-azacytidine, an inhibitor of DNA methylation, cells were obtained in which the hormone did induce lysis [25,32]. Hypomethylation of a lysis gene may thus make it susceptible to hormonal induction. Recent experiments show that nuclease activities appear in these cells upon glucocorticoid treatment suggesting that nucleolytic breakdown may be central to lymphocytolysis [33]. There is in fact convincing evidence for DNA fragmentation in rat and mouse thymocytes and mouse lymphoma cells in response to glucocorticoids [34-41]. Destruction of the genome occurs by cleavage at internucleosomal linker sites. It is, however, not yet clear whether preexisting nucleolytic acti d i e s are being released by a glucocorticoid receptor-mediated pathway or whethei new enzymes are being induced. Also, fragmentation of nuclear DNA may only b; part of the lymphocytolytic mechanism. Cell hybridization studies with several murine lymphoma cell lines revealed complex genetic regulations of hormone responsiveness suggesting that multiple genes may be involved [42]. The recently developed methods for differential cDNA cloning will in the future help to solve this complex problem.

3. Lymphoid cell variants with altered hormone responsiveness Despite the uncertainty in the mechanism of hormone-induced lymphocytolysis, cultured cells of mouse and human origin have been very useful for obtaining resistant cell variants. Mostly, three cell culture lines have been used: S49.1 and WEHI-7 mouse lymphoma cells of thymic origin [43,44] and human CEM-C7 lym-

219 phoblastic leukaemia cells [45]. The mouse cell lines differ from each other in that WEHI-7 has two wild-type receptor alleles while S49.1 is functionally hemizygous [ 4 H 9 ] . In S49.1 cells only one allele gives rise to wild-type receptor, whereas the other carries a mutation which results in the production of a receptor polypeptide devoid of steroid-binding ability [SO,51]. This defect is due to a spontaneous mutation which had occurred either during the early history of the cell line or in the S49 tumour from which the cell line had been established. The difference in genetic constitution explains why WEHI-7 cells have a higher receptor content per cell and are thus affected by lower glucocorticoid concentrations than S49.1 cells [46,52]. Also the frequency at which hormone resistance occurs in WEHI-7 is dramatically lower than in S49.1 cells, however, resistant WEHI-7 sublines could be obtained by using a two-step selection procedure in which cells of intermediate sensitivity similar to S49.1 were preselected (46,471. The above mentioned cell lines can be cloned under appropriate conditions in soft agar. Upon addition of high concentrations of a potent glucocorticoid, for example, dexamethasone at 0.1-1 pM concentrations the vast majority of S49.1 or CEM-7 cells die but a few cells form colonies. These clones are completely resistant to the cytolytic hormone effect. Resistance is a stable heritable trait which arises randomly in the cell population; treatment with mutagens increases the rate at which resistant cells come up [20,53]. The same cloning procedure has also been used for obtaining cell variants of decreased sensitivity [S4]. In this case S49.1 subclones were isolated which grew in the presence of much lower concentrations, i.e., 10 nM dexamethasone. These variants turned out to contain significantly reduced levels of glucocorticoid receptors with unaltered affinity for the steroid. This observation suggests that receptors are not only a necessary requirement for hormone action but also play a quantitative role. A relationship between cellular receptor content and the extent of hormonal sensitivity has in fact been seen in a series of mouse lymphomas of independent origin

PI.

Other interesting observations concern the hormonal regulation of receptor levels and hence glucocorticoid sensitivity. Certain lymphoid cells in culture may first respond to glucocorticoid treatment with cytolysis but the surviving cells may resume cell proliferation even in the presence of high hormone concentrations; thus exposure to the steroid induces resistance [22]. In some variants of WEHI-7 cells selected for dexamethasone resistance a similar transient inhibitory hormone effect was observed [55]. In these cells the steroid caused a reversible down-regulation of receptor levels which then allowed the cells to escape from the lethal effect. With receptor-specific cDNA probes now available it is feasible to search for changes in receptor mRNA levels in cells under different physiological conditions. Glucocorticoid treatment was recently found to produce a significant decrease in receptor specific mRNA in rat hepatoma cells and in rat liver [56].

4. Glucocorticoid receptor defects A large number of glucocorticoid-resistant subclones of the lymphoid cell lines S49.1, WEHI-7 and CEM-7 have been isolated in the way described above. Receptor defects of different types were found to be the prevailing cause for resistance. For classifying the receptor abnormalities a whole-cell hormone-binding assay followed by crude cell fractionation has been employed [20,57]. The resistance phenotype recovered most frequently is characterized by very low or undetectable glucocorticoid binding activity. It was called r- (‘receptor deficient’) [20]. However, this term may be somewhat misleading because in some S49.1 variants of the r- type a polypeptide has been detected which is of the size of the wild-type glucocorticoid receptor, cross-reacts with anti-receptor antibodies but is nevertheless unable to bind hormone [50,51]. In these cells the wild-type receptor allele appears to be shut off such that no gene product or receptor-specific mRNA originate from it [12,58,59]. The molecular mechanism causing this silent state of a receptor gene is not known but gross DNA deletions or rearrangements could be excluded [60]. Two r- subclones of WEHI-7 have been characterized in detail [51,60]. One of them displayed a parallel decrease in receptor activity, cross-reacting material, and receptor message. In the other r- clone no hormone-binding activity or immunoreactive material were detected but the cells still expressed a low level of receptor mRNA. Thus multiple controls of receptor expression may be in effect. Unresponsive variants of the r- type have also been derived from rat hepatoma cells of the HTC cell lines by use of a very specific selection technique [61]. The cells had been infected with mouse mammary tumour virus and contained multiple proviruses integrated into their genomes, at least some of which were still glucocorticoid inducible. The major viral antigen, glycoprotein gp 52 on the cell surface is controlled by glucocorticoid and can be detected by a specific antiserum. Selection of variants was accomplished by several rounds of enrichment on a fluorescence-activated cell sorter of cells with reduced expression of gp 52 in the presence of glucocorticoid. Such variants again showed a correlation between decreased hormone binding activity, immunoreactive material and receptor mRNA content but, in contrast to resistant lymphoma cells, they contain lower levels of receptor genomic sequences than the parental wild-type cells [51,60]. In two other types of receptor abnormalities discovered in resistant S49.1 variants the binding of hormone is normal or almost normal but the subcellular receptor distribution differs from the wild-type. In the phenotype called nt- the receptors are ‘nuclear transfer deficient’ in the sense of diminished nuclear binding and decreased affinity for chromatin and general DNA. I n the nt’ (‘increased nuclear transfer’) type the receptors show increased binding to nuclei or chromatin and abnormally high affinity for DNA [12,20,23]. This latter phenotype is quite rare; it has only been recovered in two instances in S49.1 subclones and in glucocorticoid-re-

22 1 sistant P1798 mouse lymphoma cells [62,63]. However, the biochemical properties of nt' mutant receptors are of particular interest. The distinction between the receptor types of abnormal nuclear binding was clarified by the use of DNA-cellulose chromatography [20]. As compared to the wildtype, the nt- receptor complexes eluted with lower salt concentrations and the nt' receptors with higher salt (Table I) corresponding to lower and higher affinities for general DNA, respectively. Interestingly, one of the nt- variants of S49.1 (clone 22R, Table I) contains receptors which also have a decreased affinity for the hormonal ligand [66]. In this receptor two amino acid exchanges have occurred in separate domains of the molecule each affecting different receptor properties [67]. Amongst resistant variants of CEM-7 cells an additional type of receptor defect was detected which has so far not been found in other cell systems [68,69]. It was called 'activation labile' because these receptors are very unstable under conditions which normally activate receptor complexes to forms with an exposed DNA binding site. These defective receptors have recently been shown to behave normally in terms of biochemical and immunochemical properties if the hormonal ligand is covalently attached by affinity labelling and therefore is unable to dissociate upon activation [70].

5. Molecular weights of glucocorticoid receptor polypeptides The molecular weights of receptors may depend heavily on the conditions of investigation. High molecular weight forms are detected under very mild conditions which avoid denaturation (see below). On the other hand, the hormone-binding polypeptides are most easily analysed by SDS gel electrophoresis if previously laTABLE I DNA binding properties and polypeptide molecular weights of receptors DNA-cellulose chromatography (mM KCI required lor elution)

Molecular weight of hormone-labelled polypeptide

native

after chymotrypsin

native

after chvmotrvnsin

175 7s 86 22s 210

230 I20 87 229 209

94 700 94 000 94 000 40 300 40 800

38 000 37 400 37 400 39 000 4 1 000

WEHI-7 wild-type

186

233

94 000

38 000

CEM-C7 wild-type

172

22 I

95 000

-

P1798 wild-type PI798 nt'

193 212

Receptor type

S39.1 wild-type S39.1 nt- (clone 22R) S39.I nt (clone 83R) S49.1 nt' (clone SSR) S49. I nt' (clone 143R)

Data from Refs. 62, 64, 65

93 000 38 500

222 belled covalently. The method of affinity labelling has in recent years become a useful tool for analysing steroid hormone receptors [71,72]. It has the advantage that crude receptor preparations can be used because following removal of excess free ligand the hormone is only bound to specific receptor sites. An exciting observation was made when wild-type and mutant receptors of the nt- and nt' types were investigated by this technique [62,64,73]. Wild-type and nt- receptor polypeptides were found to have a molecular weight of 94000 while nt' receptors have an M, of 40000 independent of whether they are of S49.1 or P1798 origin (Table I).

6 . Partial proteolysis of glucocorticoid receptors Steroid hormone receptors are very sensitive to proteolytic degradation and quite a few receptor forms have been described over the years which now have to be regarded as products of degradation by endogenous proteases. However, proteolysis studies also provided useful insight into the organization of receptor structures [62,64,65,74-781. Mild treatment of S49.1, WEHI-7 and CEM-C7 wild-type receptors with chymotrypsin or several other proteases produced receptor fragments with properties similar to nt' mutant receptors (Table I). Hormone-labelled polypeptides of molecular weights of about 38000 were detected by photoaffinity labelling and SDS gel electrophoresis and the receptor fragments bound more tightly to DNAcellulose than before proteolysis. Receptor fragments of similar sizes were also obtained from nt- receptors but these did not show dramatic increases in their affinities for DNA (Table I). When trypsin was used instead of chymotrypsin it generated hormone binding fragments of molecular weights 29000 and 27000 which did not bind to DNA-cellulose [62]. Tryptic digestion, therefore, is able to sever the domains for hormone and for DNA binding: chymotrypsin leaves these domains combined while cleaving off another part of wild-type and nt- receptors. The cleavage site for trypsin and a pair of sites for chymotrypsin have now been determined in the rat liver receptor by amino acid sequence determination [78].

7. Functional domains of glucocorticoid receptors Proteolysis studies as well as the comparison between wild-type and nt' mutant receptors suggest a domain model for glucocorticoid receptors [12,62,64]. This is presented in schematic form in Fig. 1. It shows three distinct domains which are thought to be linearly arranged along the receptor polypeptide chain. In addition to the domains for steroid binding and for interaction with DNA a third domain designated M is shown. This was thought to function by modulating the receptor's interaction with DNA or chromatin. In the absence of the M domain, truncated nt' receptors and partially proteolysed wild-type receptors bind with abnormally high affinities

223

Fig. 1. Domain model of the receptor. The DNA binding domain is in the centre and the steroid hormone binding domain (S) is to the right. The domains are not drawn to proportion. The arrows point at hinge regions between the domains which are easily cleaved by proteases like chymotrypsin (+) or trypsin (0).

to general DNA. Resistance to the lymphocytolytic glucocorticoid effect of nt' variant cells may be explained by too tight receptor binding to the vast excess of general DNA in the cell with only few receptor molecules left for interaction with specific glucocorticoid response elements (see below). It is worth noting that wild-type receptors have a lower isoelectric pH than chymotrypsin degraded or ntl receptors. Thus the ionic interaction with DNA is altered significantly by elimination of the M domain; the modulating effect may therefore be due - at least in part - to limiting the receptor's electrostatic attraction by DNA. In addition to three functional domains the scheme of Fig. 1 also shows connecting polypeptide sections between the domains which may be viewed as hinge regions. These are the major targets for proteolysis [78]. While biochemical studies clearly demonstrated that the domains for hormone and for DNA binding are closely linked they did not provide evidence for the order of the domains along the receptor polypeptide chain. This information became available by investigating receptor cDNA clones [59,67,79-811. The primary structures of most steroid and thyroid hormone receptors have recently been deduced from the respective nucleotide sequences (for a review see Ref. 15) and the amino acid sequences of glucocorticoid receptors of man [79], mouse [67] and rat 1811 are now available. They contain 777, 783 and 795 amino acid residues, respectively, and are highly homologous. The M domain of the glucocorticoid receptor occupies the amino terminal half of the polypeptide chain. It is followed by the DNA binding domain in the centre of the molecule and the hormone binding domain in the carboxy terminal portion. The only striking difference between the receptor sequences of different species is the presence of a stretch of 8 and 19 glutamine residues within the M domains of the mouse and rat receptors, respectively, which is absent from the human receptor. However, this sequence is probably of no biological significance.

224 The availability of receptor-specific cDNAs has opened up the possibility of producing new mutations in the receptor domains by in vitro DNA manipulations and to check the products for functional activities. Thus far, mostly deletions have been studied. They were tested following transfection into r- variants of HTC cells (see above) or the monkey cell line CV-1 and its derivative COS-7 which have low levels of glucocorticoid receptors. I n the case of r- HTC cells the hormonal induction of either an endogenous liver specific gene or of integrated genomes of the mouse mammary tumour virus were assayed [81,82]. In the experiments with simian cells an artificial inducible gene was introduced into the cells by cotransfection. This indicator gene consisted of the regulatory region (long terminal repeat or LTR) of the mouse mammary tumour virus coupled to the prokaryotic gene for chloramphenicol acetyltransferase which is easy to assay and has no eukaryotic counterpart. 7.1. The M domain

Several laboratories have used the approach outlined above and have constructed selective deletions in the M domains of human [83], mouse [67,84] and rat [82] glucocorticoid receptors. Elimination of either parts or the whole domain resulted in receptor polypeptides of correspondingly reduced sizes which, nevertheless, were able to mediate hormonal induction. Depending on the deletion, however, inducibility was significantly decreased. These data suggest that the M domain is not a necessity of receptor function but facilitates transcriptional activation by improving the ability to discriminate between general DNA and glucocorticoid response elements in the genome. Interestingly, a stretch of about 100 amino acids with a preponderance of acidic residues is largely responsible for the enhancing activity of the entire M domain [84]. The M domain has also been called ‘immunogenic’ or ‘immunoactive’ domain [80,85] because most monoclonal and polyclonal antibodies raised against purified glucocorticoid receptors recognize epitopes within the M domain. Thus, neither nt’ mutant receptors of S49.1 and P1798 origin [50,63,86], nor chymotrypsin degraded wild-type receptors [85,87] react with these antibodies. There is, however, a monoclonal antibody, designated BUGR-1, which does recognize receptors from which the M domain has been eliminated by limited proteolysis [88,89]. This epitope has recently been mapped to a sequence about 20 amino acid residues upstream from the amino terminal end of the DNA binding domain [90], i.e., very close to the chymotrypsin-sensitive hinge region linking the domains [78]. A monoclonal antibody which reacts with an epitope within the M domain has recently been used to check whether it influences the receptor’s DNA binding properties [65]. The immune complex with wild-type receptors was indeed found to chromatograph on DNA-cellulose similar to ntl mutant receptors or the chymotrypsin-degraded wild-type. Reaction with the antibody mimics the removal of the M domain which suggests that biochemical modifications of the M domain could

225 affect hormone responsiveness within the cell by modifying the DNA binding properties of receptors. The origin of nt' receptors is still puzzling (for a discussion see Ref. 12). It is now known that these abnormal receptors are synthesized from truncated messages which are missing about 1.5 kb of 5' sequences and that the genes coding for them do not contain gross rearrangements or deletions [58-601. Therefore, the truncated ntl mRNA appears to be the product of abnormal splicing due to a mutational change. Recent studies with nt' cDNA clones indeed show sequence divergence upstream of amino acid 406 of the wild-type mouse receptor [91]. It is not clear, however, whether synthesis of the nt' receptor polypeptide starts with methionine corresponding to position 406 of the wild-type or from a new initiation site located upstream and within the diverging nucleotide sequence. Consequently, ntl receptors are either missing a long stretch of amino terminal residues or they have new sequences at their amino termini which are not present in the wild-type. In either case, the region of divergence corresponds to the BUGR-1 epitope mentioned above which explains why nt' receptors do not react with this monoclonal antibody [51]. As pointed out above, nt' receptors have been discovered in lymphoma cells selected for resistance to the cytolytic glucocorticoid effect. Since receptors from which the M domain had been eliminated by cDNA manipulation still function to some extent in transfection studies it was important to find out whether nt' receptors would also be able to mediate some hormonal response. This was in fact observed when nt' lymphoma variants were transfected with a DNA construct consisting of the LTR region of the mouse mammary tumour virus coupled to the gene for chloramphenicol acetyltransferase (U. Gehring and H. Losert, unpublished experiments). Hormonal induction of enzyme activity was consistently observed but was low, as one might expect. Information on the structural organization of the glucocorticoid receptor gene now also becomes available. The entire M domain of the rat receptor, i.e., amino acids 1-415 are encoded by a single exon of about 1.2 kb [82.91]. This is separated by an unusually large intron of more than 30 kb from the rest of the coding sequence which is organized in multiple small exons. The most diverging areas both in length and sequence are the amino terminal parts of the various steroid and thyroid hormone receptors (cf. Ref. 15) but there is sequence conservation in the M domains of receptors for the same hormone, for example glucocorticoid receptors of three species. This suggests that during evolution the various receptor genes have picked up very different DNA segments which then gave rise to the respective M domains. It is reasonable to assume that the common ancestral gene for steroid and thyroid hormone receptors was missing such 5' sequences. 7.2. The DNA binding domain This is the centre part of steroid hormone receptors and perhaps the most interesting domain. A comparison of the presently available amino acid sequences of

226 steroid and thyroid hormone receptors shows striking homologies in this region (cf. Ref. 15). Completely conserved throughout all the sequenced members of this multigene family are nine cysteine residues and a series of basic and hydrophobic amino acids. In fact, when the primary structure of the first steroid receptor, the human glucocorticoid receptor [79] was deduced from the nucleotide sequence the homology of this central region with the product of the v-erb-A oncogene of the avian erythroblastosis virus immediately became conspicuous [80]. This similarity subsequently led to the identification of c-erb-A, the cellular counterpart of v-erb-A, as the gene for a thyroid hormone receptor [92,93]. The kinship amongst these sequences has also greatly helped to identify cDNA clones for other steroid hormone receptors. Within the mid portion of 66 amino acids the glucocorticoid receptors of human, mouse and rat [67,79,81] are identical except for valine at position 437 instead of glycine in the WEHI-7 as compared to the S49.1 wild-type receptor. Although this exchange is of no functional consequence it is an interesting case of allelic variation since both lymphoma cell lines are derived from the same mouse strain [43,44]. The human and chicken oestrogen receptors [94-961 still have 40 amino acids identical with the glucocorticoid receptors over this stretch of 66 residues. The similarity between glucocorticoid, mineralocorticoid and progesterone receptors is even greater. Over this sequence of 66 residues there are only four differences between glucocorticoid receptors and the human mineralocorticoid receptor [97] and six between glucocorticoid and progesterone receptors [98-1001. This homology fits well to the observation that both glucocorticoid and progesterone receptors can bind to the same nucleotide sequences [ 1011, and, perhaps more interestingly, that glucocorticoid, mineralocorticoid and progesterone receptors can elicit transcriptional activation from the same glucocorticoid response elements [97,102,103]. Sequence studies with cDNAs from one of the nt- variants of S49.1 mouse lymphoma cells were successful in localizing the defect responsible for the decreased DNA binding ability of the nt- receptor [67]. It was found to be due to the replacement of arginine in position 484 by histidine within the DNA binding domain (Fig. 2). Interestingly, all the receptors sequenced to date (cf. Ref. 15) have arginine in this very position. It is therefore reasonable to assume that this basic amino acid is either directly and specifically involved in interacting with DNA or profoundly influences the overall folding of the DNA binding domain. Replacement of a tetrapeptide sequence which includes this arginine by an unrelated pentapeptide likewise abolished DNA binding activity [90]. The exchange of individual amino acids in this domain by site-directed mutagenesis will provide more insight into the structural requirements for interaction with DNA. Thus far a series of deletion mutations have been introduced into glucocorticoid receptor cDNAs in the region coding for the DNA binding domain [83,84,90,104]. These were tested either by in vitro transcription and translation followed by a spe-

227 cific D N A binding assay or for transcriptional activation in transfected cells. Deletions removing the entire domain or parts of it eradicated both activities but a relatively short fragment which comprises the DNA binding domain retains these functions (see below). Several insertions introduced into this domain also destroyed the receptor’s potential for transcriptional activation [ 1051. How are the DNA binding domains of steroid hormone receptors folded for establishing the appropriate contacts with DNA sequences? This problem has recently evoked quite some interest and it has been proposed that the DNA binding domain of glucocorticoid receptors may fold into ‘DNA binding fingers’ [lo61 similar to those of the transcription factor TFIIIA of Xenopus oocytes [107,108]. In TFIIIA, tandem repeats of about 30 amino acids have been detected, each of which contains invariant pairs of cysteine and histidine residues. These are thought to coordinate tetrahedrally around a central zinc ion such that a 12 amino acid loop may be formed which makes the contact with DNA. Similar sequence motifs have been found in several DNA binding proteins and steroid hormone receptors are thought to contain two such structures [82,90,99]. A close consideration of the sequences involved, however, shows that the detailed structure of the DNA binding domains of glucocorticoid and other steroid hormone receptors must be quite different from the finger-like arrangements in TFIIIA [ 151. Also unequivocal proof for the involvement of Zn2+ in glucocorticoid receptor action is still missing.

7.3. The hormone binding domain Our present information about the hormone binding domain, although still limited, comes from a combination of biochemical and genetic studies. The amino acid which reacts with the chemical affinity label dexamethasone 21-mesylate has now been identified as cysteine in position 656 of the rat glucocorticoid receptor [109]. This cysteine is therefore located within or very close to the hormone binding site. It corresponds to cysteine 644 of the mouse receptor (Fig. 2). Evidence for the participation of two additional amino acids in hormone binding has come from sequence studies with receptor cDNAs obtained from variant S49.1 cells [67]. The receptor polypeptide described above which is unable to bind hormone carries a replacement of glutamic acid 546 by glycine which is responsible for the defect (Fig. 2). Also, tyrosine in position 770 contributes to the hormone binding ability. This amino acid was found to be exchanged for by asparagine in the double mutant described above which is defective in DNA binding and has decreased affinity for the hormone (Fig. 2). These observations suggest that the hormone binding domain is intricately folded such that amino acids 546, 644 and 770 of the murine receptor, and probably several more, contribute directly or indirectly to steroid binding. In this context a general problem needs to be pointed out concerning mutant gene products that are devoid of biological function. In principle, any alteration in the primary structure of a protein, i.e., deletions, insertions or

228 DNA binding

hormone binding

J

V//?

783

1

wild - type

484

546

644

770

Arg

Glu

Cys

TY r

GlY

hormone binding deficient mutant double mutant n t - and decreased a f f i n i t y f o r hormone

His

Asn

Fig. 2. Wild-type and mutant receptors of murine origin. I n the wild-type receptor, amino acids Arg 484, Glu 546, Cys 644 and Tyr 770 are indicated. The amino terminus is to the left and the carboxy terminus to the right. Data from Ref. 67.

amino acid exchanges might lead to changes in the overall conformation of the protein which could affect activity rather indirectly. Likewise, protein turnover might be altered. However, this appears not to be so in the case of these S49.1 mutant receptors since they did not behave differently from the wild-type with respect to stability. The hormone binding domain was also studied by deletion analyses employing cDNAs. In one approach cDNAs were transcribed and translated in vitro and the products were tested for steroid binding [90,104]. Deletion of 29 amino acids from the carboxy terminus of the rat receptor reduced hormone binding by about 100fold. No binding was detected with more extensive carboxy terminal deletions or internal deletions. However, truncated polypeptides containing just the hormone binding domain retained the ability for specific hormone binding. In another approach deletions were tested for functional activity by transfection into r- cells. The elimination of as few as 27 carboxy terminal residues resulted in the complete loss of biological activity and longer deletions did not restore any hormonal effect [104]. However, truncation of the entire hormone binding domain or internal portions thereof allows for gene activation independent of hormone (cf. Section 7.4). The striking homology between glucocorticoid, mineralocorticoid and progesterone receptors in their DNA binding domains has been pointed out above. The hormone binding domains of these receptors likewise show significant sequence homologies, for example, 57% amino acid identity between the human glucocorticoid and mineralocorticoid receptors [97]. A glutamic acid residue analogous to that in position 546 of the murine glucocorticoid receptor mentioned above is, in fact, located in mineralocorticoid and progesterone receptors within similar amino acid sequences and might well serve a similar function in hormone binding. The homology between these three receptors is reflected by the long known fact of cross-binding activity of the respective steroids which amongst themselves are also quite related in structure. For example, progesterone competes for glucocorticoid binding to its

229 receptor and acts as an antagonist to glucocorticoid induction of tyrosine aminotransferase, a liver-specific enzyme [110,111]. Aldosterone also binds to the glucocorticoid receptor but with lower affinity than glucocorticoids; in contrast to progesterone it is able to induce tyrosine aminotransferase, albeit less efficiently than glucocorticoids [110,111]. The mineralocorticoid receptor for its part binds glucocorticoids with high affinity, in particular naturally occurring steroids. Using the glucocorticoid inducible promoter region of the mouse mammary tumour virus it was now shown that glucocorticoids can elicit transcriptional activation via the mineralocorticoid receptor [97]. The mineralocorticoid receptor may therefore function under certain conditions as a bona fide glucocorticoid receptor. One of the major problems of corticosteroid physiology then is to understand how mineralocorticoid action in specific target tissues, for example, the kidney comes about in the presence of circulating glucocorticoid levels which by far exceed those of aldosterone [1121.

7.4. H o r m o n e independent gene activation by truncated receptors Very interesting information about the function of receptor domains came to light when progressive carboxy terminal deletions in the hormone binding domain were constructed and assayed for biological activity in transfected cells [ 1041. Receptor derivatives with up to 180 amino acids removed from the carboxy terminus did not stimulate transcriptional activation either in the presence or absence of glucocorticoid. By contrast, deletions of 20G270 residues produced activation independent of the addition of hormone. Similarly, deletions in the amino terminal portion of the hormone binding domain resulted in constitutive gene activation [83,84]. A relatively short portion of about 60 amino acids is largely responsible for this hormone independent effect [84]; this region shows sequence homology between various steroid hormone receptors. These observations suggest a mechanism for hormonal action. In the unliganded state the hormone binding domain acts by suppressing receptor function. In this situation the inherent ability of the DNA binding domain for interacting with specific nucleotide sequences and for transcriptional activation remains dormant. Following hormone binding this inhibition may be released by a profound change in the structure of the receptor molecule. Evidence for the dissociation of receptor multimeric complexes upon activation to the DNA binding state is discussed below (cf. Section 9). As an extension to the deletion studies decribed above several double mutations have been constructed which, upon transfection, produced receptor derivatives truncated at both ends of the polypeptide chain [82,83,90]. These are missing both the M domain and the hormone binding domain or parts thereof but are still able to bind DNA. A receptor fragment of only 150 amino acids was sufficient for constitutive transcriptional activation. This suggests that the DNA binding domain of

the receptor independently folds into a conformation which is both stable and functional. The 150 amino acids fragment spans a core of 86 residues which has been delineated by maximum deletions on either end of the molecule. Both functions, specific DNA binding and transcriptional activation colocalize to this core region. This includes the highly conserved sequence of 66 residues (cf. Section 7.2) and harbours a signal for nuclear localization. Interestingly, the same receptor core region which is required for transcriptional enhancement is also sufficient for down regulation of the prolactin operator by glucocorticoid [91].

7.5. A chimaeric receptor Some of the above mentioned results suggest that the domains of glucocorticoid receptors may function independently of each other. This view is supported by recent results obtained with a chimaeric receptor molecule which was constructed by replacing on the cDNA level the DNA binding domain of the oestrogen receptor by that of the glucocorticoid receptor [113]. Cotransfection of this DNA construct into HeLa cells, together with a glucocorticoid inducible indicator gene, resulted in oestrogen dependent expression. By contrast, a similarly constructed oestrogen inducible indicator gene was not activated by the chimaeric receptor. These data clearly show that domains can be swapped between steroid hormone receptors. The specificity for gene activation, however, is inherent to the respective DNA binding domain.

8. Glucocorticoid response elements The interaction of receptor-glucocorticoid complexes with specific sequences in the DNA - so called response elements - is a pivotal step in the mechanism of glucocorticoid action. Such specific nucleotide sequences have first been identified in the control region of the mouse mammary tumour virus in its integrated form and have subsequently been detected in several glucocorticoid controlled genes (for reviews see Refs. 7, 10, 11, 14, 114). Preferential DNA binding assays can be carried out using several approaches. The DNase footprinting technique is perhaps the most direct and informative method for detecting protein-DNA interactions as it discloses the nucleotide sequence which is covered by the receptor and thus protected against nucleolytic degradation. Specific protection of guanine residues against methylation by dimethyl sulfate can similarly be detected if these are in intimate contact with the receptor protein. Using highly purified receptor preparations, a series of specific receptor binding sequences have been characterized in this way. They may be located at very different distances upstream of the start site of transcription or within the gene and at either orientation thus resembling transcriptional enhancer elements. They may bind

23 1 receptor complexes with different affinities and appear to preferentially occur in clusters. Sequence comparison has resulted in the consensus sequence 5’-G G T A C A N N N T G T T C T - 3 ’ [114] in which the hexanucleotide TGTTCT is most highly conserved. It needs to be noted that this 15 base pair long sequence has a partial dyad symmetry which is compatible with the adjacent binding of two receptor monomers thus forming a dimer on the DNA. A model of this type has been proposed [114]. Electron micrographs indeed suggest a dimeric structure in association with DNA [115]. An important question concerns the functional relevance of multiple sites for in vitro receptor binding in relation to inducibility. Studies with the mouse mammary tumour virus promoter showed that the number of glucocorticoid response elements may affect the level of hormone-inducible gene expression in an additive way [116]. In the case of the tyrosine aminotransferase gene three binding regions have been identified by footprinting which are located in series far upstream of the cap site [117]. Interestingly, one of these can be deleted without impairment of activity. Of the others, one specific region is absolutely required for hormone inducibility while the presence of the other one strongly enhances the effect. These experiments suggest cooperativity between response elements and clearly demonstrate that binding regions may exist which are nonfunctional. The functional response elements of the tyrosine aminotransferase gene were also detected in intact hepatoma cells by use of the genomic footprinting technique [118]. Changes in the reactivity of specific guanine residues with dimethyl sulfate were detected in response to glucocorticoid treatment; this supports the notion that hormone binding increases the affinity of the receptor for its target sequences. A straightforward test of whether a DNA sequence is capable of exerting inducibility is to couple it to an appropriate indicator gene and to transfect the construct into cells containing the receptor. In this way the above consensus sequence was found to be functional [ 1031. Moreover, an oligonucleotide in which the partially symmetric sequence of the element responsible for tyrosine aminotransferase induction was changed to the perfect palindrome 5 ’-AGAACATGATGTTCT-3’ likewise conferred glucocorticoid inducibility. A related, but distinct palindrome of 15 base pairs was recently found to produce inducibility by oestrogen [119]. Overlapping DNA binding activities of glucocorticoid and progestin receptors have briefly been mentioned (Section 7.2). Both receptors elicit responses from the same control regions of the mouse mammary tumour virus [lo21 and the chicken lysozyme gene [ 1201 and from identical synthetic oligonucleotides [103]. Interestingly, nucleotide substitutions within the above mentioned consensus sequence similarly affected inducibility by glucocorticoids and progesterone [ 1031. Nevertheless, there is evidence that both receptors may make contact to the same DNA sequence in different ways [121]. It also needs to be stressed that the group of genes

232 controlled by glucocorticoid and progestins are certainly not identical despite the similarities in both the DNA binding domains of the receptors and the nucleotide sequences recognized by them. It appears reasonable to assume that the M domains of the different receptors which are very different in size and structure (cf. Section 7.1) may play some as yet unidentified role in relation to specificity, for example, by interacting with other components. In this respect the observation may be relevant that the glucocorticoid receptor from which the M domain had been eliminated by chymotrypsin treatment (cf. Section 6) binds to glucocorticoid response elements with greatly decreased selectivity as compared to the intact receptor [ 1221. Binding of receptor-steroid complexes to specific sites on the DNA is a necessary prerequisite for affecting transcription but may not be sufficient. The above mentioned fact that glucocorticoid response elements may operate from varying distances from a promoter is suggestive for the involvement of other factors. A polypeptide of molecular weight 72 000 which was copurified with the glucocorticoid receptor from rat liver was found to increase the selectivity of sequence-specific DNA binding of the intact receptor but not the chymotryptic receptor fragment [122]. Specific nucleotide sequences in the vicinity of glucocorticoid response elements are also important. Thus, destruction of the binding sequence for nuclear factor 1 in the control region of the mouse mammary tumour virus resulted in the loss of hormone inducibility and reinsertion of one or two binding sites restored inducibility [123]. Moreover, the binding sites for nuclear factor 1 as well as that for another hitherto unidentified factor within the control region of the mammary tumour virus were protected against nuclease digestion in nuclei isolated from hormone treated cells but not from control cells [124]. Since nuclear factor 1 has a high affinity for its binding sequence on naked DNA it appears likely that this region is masked in the absence of hormone. Sequence specific phasing of nucleosomes over the promoter region has in fact been observed and it was found to change upon hormone treatment in just the region containing the glucocorticoid response elements [125]. Binding of the receptor-hormone complex to its target sequence could therefore result in opening up the chromatin structure followed by the recruitment of other protein factors to the promoter. In summary, multiple factors are expected to act in concert with hormone receptors and their respective response elements in order to bring about transcriptional activation or repression of specific genes. Many facets of this complex network still need to be worked out. A particularly interesting although puzzling situation concerns steroid hormone induction of certain genes which requires de novo protein synthesis. The best studied system appears to be that of 0,-acid glycoprotein [12&128], a member of the acute phase reactant proteins of inflammation. The level of specific mRNA increases substantially after glucocorticoid treatment in liver and hepatoma cells but this increase is prevented by protein synthesis inhibitors. The S'-flanking region of the gene was shown to confer both hormone responsiveness and the dependence on ongoing protein synthesis onto the expression of heterol-

233 ogous genes. Inhibition studies with cycloheximide suggest that a pre-existing protein is required for efficient hormone induction and that this is quite labile. Nothing is yet known about this putative regulatory protein or its possible interaction with the glucocorticoid receptor. However, the upstream sequences required for hormone responsiveness did not contain a glucocorticoid response element that corresponded to the above given consensus sequence. By use of the in vivo footprinting technique it should be possible to shed some light on this complicated problem.

9. Higher order structures of glucocorticoid receptors As briefly mentioned above (Section 5) glucocorticoid receptor forms of high molecular weight can be detected in extracts of target cells if dissociation and denaturation are carefully avoided. Molecular weights as large as 300000 to 350000 have been reported (for reviews see Refs. 5, 9) by combined measurements of Stokes radii and sedimentation coefficients. This brings up the question of the way in which the glucocorticoid binding polypeptide of molecular weight 94 000 is organized within such large receptor structures. One possibility is a tetramer consisting solely of hormone binding subunits. The extreme alternative is a heteromeric complex containing one hormone binding polypeptide which is associated with other macromolecules. It has been argued that a comparison of the hydrodynamic properties of wild-type and ntl mutant receptors of mouse lymphoma cells should yield information about the subunit structure [ 129,1301. If the large form of ntl receptors were a homotetramer of 40000 molecular weight polypeptides (Section 5 ) , it should obviously exhibit a molecular weight of about 160000. On the other hand, if the large receptor complexes of wild-type and nt' mutant contained only one hormone binding subunit they should differ by a molecular mass of about 50000 provided, of course, that the associated components are the same. T A B L E I1 Molecular properties of receptors Receptor type (treatment)

KCI (mM) added Stokes radius Sedimentation to cell extracts (A) coefficient (S)

Molecular weight

S49.1 wild-type (native)

0 300

80.6 62.0

9.6 4.9

328000 128000

S49.1 wild-type 0 (after chymotrypsin) 300

68.9 40.7

10.0 3.8

291 000 65 000

-

S49.1 nt' (clone 55R) (native)

71.9 38.4

9.8 3.9

298000 63000

-

0 300

Data from Ref. 130. "Binding to DNA-cellulose was tested atter adjusting KCI concentrations to 20 mM

Binding to DNA" ~

+

+

+

Table I1 lists the properties of wild-type and nt' receptors under different salt conditions. At low ionic strength, i.e., conditions which preserve large structures the data show a difference in molecular weights of about 30000. This result is compatible only with a higher order receptor structure containing one rather than two or more hormone binding subunits. Moreover, if the wild-type complex was subjected to mild treatment with chymotrypsin a receptor form was recovered which is very similar to the ntl mutant (Table 11). This suggests that, in the wild-type, high molecular weight receptor form the component most susceptible to proteolysis is the hormone binding polypeptide itself. Interestingly, a heteromeric structure has also been proposed for the rat liver receptor using immunochemical techniques [131]. Taken together, these studies suggest a higher order receptor structure in which one hormone binding polypeptide of molecular weight 94 000 or 40 000, respectively, is associated with other subunits which are identical in the wild-type and the ntl mutant. The large receptor structures are known not to interact with DNA (for a recent review see Ref. 132). By increasing the salt concentration to 300 mM KCl or higher, dissociation to receptor forms of lower molecular weight occurs (Table 11). Concomitantly, DNA binding ability is generated. This suggests that subunit dissociation leads to the exposure of the DNA binding domain which appears to be a preformed structure within the receptor complex (cf. Section 7.4). It needs to be mentioned that activation of receptor-glucocorticoid complexes can be achieved not only by increasing the ionic strength but also by other means, for example, by warming to 20°C; subunit dissociation has also been observed in this case [130]. Interestingly, receptor activation can be prevented if dissociation is made impossible by covalently cross-linking the receptor subunits [133]. The molecular weights of the dissociated receptor forms, about 130000 and 6.5 000 for the wild-type and the nt' mutant (Table 11), are somewhat higher than those determined in SDS containing gels for the respective steroid binding polypeptides (cf. Section 5 ) . It is not clear at present whether this is merely due to the use of different analytical procedures or whether the hormone binding polypeptides are associated under non-denaturing conditions with some as yet unidentified component(s) of about 30 000 molecular weight [130]. The observation that undissociated wild-type and ntl receptor complexes bind tightly to DEAE-cellulose and require the same salt conditions for eluting [64,130] suggests that the associated components are highly negatively charged. Two types of macromolecules need to be discussed in this context. A non-hormone binding phosphoprotein of molecular weight 90 000 has been detected which is common to the higher order structures of various steroid hormone receptors; this has been identified as heat-shock protein hsp 90 (for a review see Ref. 132). An increasing body of evidence also points to the participation of RNA in large receptor structures (for a review see Ref. 134). This RNA is of low molecular weight similar to tRNA. Most significantly, it can be covalently cross-linked to the steroid binding

235 polypeptide in high molecular weight receptor complexes [135]. It is tempting to simply add up the molecular masses of these components: a complex consisting of one steroid binding polypeptide, two units of hsp 90 and one tRNA molecule would fit the molecular weight data quite well [ 1301, but appropriate reconstitution experiments would certainly be needed to unequivocally prove such a structure. However, there is also evidence that the large glucocorticoid receptor form of molecular weight of about 300000 does not contain RNA [136], but that a receptor form of intermediate size contains tRNA [137,138]. In any event, the biological functions of the hsp 90 and RNA molecules in relation to the mechanism of glucocorticoid receptor actions are not yet clear. One possibility is that they function by covering up the hormone binding polypeptide, in particular the DNA binding domain in the large structures (cf. Section 7.4).

References 1. Higgins, S.J. and Gehring, U. (1978) Adv. Cancer Res. 28, 313-397. 2. Harrison, R.W. (1983) In: International Review o f Cytology, Suppl. 15 (Danielli, J.F., ed.) pp, 1-16. Academic Press, New York. 3. Rousseau, G.G. (1984) Biochem. J . 224, 1-12. 4. Rousseau, G.G. (1984) Mol. Cell. Endocrinol. 38. 1-11. 5. Sherman, M.R. and Stevens, J . (1984) Ann. Rev. Physiol. 46, 83-105. 6. Housley, P . R . , Grippo, J.F. and Pratt. W.B. (1985) In: Hormonally Responsive Tumours (Hollander, V.P., ed.) pp. 259-285. Academic Press. New York. 7. Ringold. G.M. (1985) Ann. Rev. Pharmacol. Toxicol. 25. 529-566. 8. Stevens, J . and Stevens, Y.-W. (1985) In: Hormonally Responsive Tumours (Hollander, V.P., ed.) pp. 155-183. Academic Press, New York. 9. Vedeckis, W.V. (1985) In: Hormonally Responsive Tumors (Hollander, V.P., ed.) pp. 3-61. Academic Press, New York. 10. Yamamoto, K.R. (1985) Ann. Rev. Genet. 19. 209-252. 11. Beato, M. (1986) In: Oncogenes and Growth Control (Kahn. P. and Graf, T . , eds.) pp. 219-225. Springer-Verlag, Berlin. 12. Gehring, U. (1986) Mol. Cell. Endocrinol. 48, 89-96. 13. King, R.J.B. (1986) J. Steroid Biochem. 25, 451-454. 14. Pfahl, M. (1986) In: Biochemical Actions of Hormones, Vol. XI11 (Litwack. G.. ed.) pp. 325-357. Academic Press. New York. 15. Gehring, U. (1987) Trends in Biochem. Sci. 12. 399-402. 16. Baxter. J.D. and Rousseau, G.G. (1079) Glucocorticoid Hormone Action, Monographs on Endocrinology, Vol. 12, Springer-Verlag, Berlin. 17. Claman, H.N. (1972) N. Engl. J. Med. 287. 388-397. 18. Weissman, I.L. (1973) J. Exp. Med. 137, 504510. 19. Huggins, C.B. and Uematsu, K. (1976) Cancer 37, 177-180. 20. Yamamoto, K.R., Gehring, U.. Stampfcr. M.R. and Sibley, C.H. (1976) Recent Progr. Horm. Res. 32, 3-32. 21. Craddock, C.G. (1978) Ann. Intern. Med. 88. 564566. 22. Harris, A.W. and Baxter, J.D. (1979) In: Glucocorticoid Hormone Action (Baxter, J.D. and Rousseau, G.G., eds.) Monographs on Endocrinology, Vol. 12, pp. 423-448. Springer-Verlag, Berlin.

23. Gehring, U . (1980) In: Biochemical Actions of Hormones, Vol. VII (Litwack, G . , ed.) pp. 205-232. Academic Press, New York. 24. Munck, A . and Crabtree, G . R . (1981) In: Cell Death in Biology and Pathology (Bowen, I . D . and Lockshin, R . A . , eds.) pp. 329-359. Chapman and Hall, London. 25. Bourgeois, S. and Gasson, J . C . (1985) In: Biochemical Actions of Hormones, Vol. XI1 (Litwack, G . , ed.) pp. 311-351. Academic Press, New York. 26. Dougherty, T.F. (1952) Physiol. Rev. 32, 379-401. 27. Ivarie. R.D., Morris, J . A . and Martial, J . A . (1982) Mol. Cell. Biol. 2, 179-189. 28. Camper, S.A., Yao, Y.A.S. and Rottman, F.M. (1985) J. Biol. Chem. 260, 12246-12251. 29. Eberwine, J.H. and Roberts, J . L . (1984) J . Biol. Chem. 259. 21662170. 30. Gagner, J.-P. and Drouin, J . (1985) Mol. Cell. Endocrinol. 40, 25-32. 31. Israel, A . and Cohen, S.N. (1985) Mol. Cell. Biol. 5, 2443-2453. 32. Gasson, J.C., Ryden, T. and Bourgeois, S. (1983) Nature 302, 621-623. 33. Bourgeois, S., Crepin, M . and Dean, D . C . (1987) J. Cell. Biochem. Suppl. 11A, 15. 34. Ivannik, B.P., Golubeva, R.V. and Ryabchenko, N.I. (1977) Biochemistry (Translation of Biokhimiya) 42, 771-776. 35. Wyllie, A . H . (1980) Nature 284, 555-556. 36. Vedeckis, W . V . and Bradshaw, H . D . (1983) Mol. Cell. Endocrinol. 30. 215-227. 37. Cohen, J.J. and Duke, R.C. (1984) J . Immunol. 132, 38-42. 38. Wyllie, A . H . , Morris. R . G . , Smith, A . L . and Dunlop, D. (1984) J . Pathol. 142, 67-77. 39. Compton. M.M. and Cidlowski. J . A . (1986) Endocrinology 118. 38-45. 40. Wielckens, K. and Delfs, T. (1986) Endocrinology 119,2383-2392. 41. Compton. M.M. and Cidlowski, J.A. (1987) J . Biol. Chem. 262, 82%-8292. 42. Weinroth, S.E.. MacLeod, C . L . . Minning, L. and Hays, E . F . (1985) Cancer Res. 45, 4804-4809. 43. Harris, A.W. (1970) Exp. Cell Res. 60, 341-353. 44. Harris, A.W., Bankhurst, A.D.. Mason, S. and Warner, N.L. (1973) J . Immunol. 110, 431-438. 45. Norman. M.R. and Thompson, E . B . (1977) Cancer Res. 37, 3785-3791. 46. Bourgeois, S. and Newby, R.F. (1977) Cell 11. 423-430. 47. Bourgeois, S., Newby. R.F. and Huet, M. (1978) Cancer Res. 38, 4279-4284. 48. Gehring, U . (1980) Mol. Cell. Endocrinol. 20.261-274. 49. Francke, U . and Gchring. U. (1980) Cell 22, 657-664. 50. Westphal. H . M . , Mugele. K., Beato, M. and Gehring, U . (1984) EMBO J . , 3, 1493-1498. 51. Northrop, J.P.. Gametchu. B., Harrison, R . W . and Ringold. G . M . (1985) J . Biol. Chem. 260, 6398-6403. 52. Gehring, U . , Mugele, K. and Ulrich, J . (1984) Mol. Cell. Endocrinol. 36. 107-113. 53. Harmon, J.M. and Thompson. E.B. (1981) Mol. Cell. Biol. 1, 512-521. 54. Gehring. U.. Ulrich, J . and Segnitz. B. (1982) Mol. Cell. Endocrinol. 28, 605-611. 55. Danielsen, M. and Stallcup. M.R. (1984) Mol. Cell. Biol. 4, 449-453. 56. Okret, S., Poellinger. L.. Dong, Y. and Gustafsson. J.-A. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 5899-5903. 57. Sibley, C . H . and Tomkins. G.M. (1974) Cell 2. 221-227. 58. Miesfeld. R . . Okret. S.. Wikstr6m, A.-C.. Wrange, 0..Gustafsson, J.-A. and Yamamoto, K.R. (1984) Nature 312. 779-781. 59. Miesfeld, R., Rusconi, S.. Okret. S.. Wikstrom, A.-C.. Gustafsson, J.-A. and Yamamoto, K.R. (1985) In: Sequence Specificity in Transcription and Translation. UCLA Symposium on Molecular and Cellular Biology, Vol. 30. (Calendar, R . and Gold, L . , eds.) pp. 535-545. Alan R . Liss, New York. 60. Northrop, J.P., Danielsen, M. and Ringold, G . M . (1986) J . Biol. Chem. 261, 11064-11070. 61. Grove, J.R.. Dieckmann, B.S., Schroer, T . A . and Ringold, G.M. (1980) Cell 21, 47-56. 62. Gehring, U . and Hotz, A . (1983) Biochemistry 22, 4013-4018.

237 63. Stevens, J., Stevens, Y.-W. and Haubenstock, H . (1983) In: Biochemical Actions of Hormones, Vol. X (Litwack, G . , ed.) pp. 383-446. Academic Press. New York. 64. Dellweg, H.-G.. Hotz, A., Mugele, K. and Gehring, U . (1982) EMBO J. 1, 285-289. 65. Gehring, U . and Segnitz, B. (1988) Mol. Cell. Endocrinol. 56, 245-254. 66. Spindler-Barth, M. and Gehring, U . (1982) FEBS Lett. 138, 91-94. 67. Danielsen, M., Northrop, J.P. and Ringold, G.M. (1986) EMBO J. 5, 2513-2522. 68. Schmidt, T.J., Harmon. J.M. and Thompson. E.B. (1980) Nature 286, 507-510. 69. Harmon. J.M., Schmidt, T . J . and Thompson, E . B . (1984) J . Steroid Biochem. 21, 227-236. 70. Eisen, L.P., Elsasser, M.S. and Harmon. J.M. (1987) J . Cell. Biochem. Suppl. 11A, 105. 71. Simons, S.S. and Thompson, E.B. (1982) In: Biochemical Actions of Hormones, Vol. IX (Litwack, G., ed.) pp. 221-254. Academic Press, New York. 72. Gronemeyer, H. and Govindan, M.V. (1986) Mol. Cell. Endocrinoi. 46, 1-19. 73. Nordeen, S.K., Lan. N.C., Showers, M.O. and Baxter, J.D. (1981) J. Biol. Chem. 256, 1050?-10508. 74. Wrange, 0. and Gustafsson. J.-A. (1978) J . Biol. Chem. 253, 856865. 75. Andreasen, P.A. and Gehring. U. (1981) Eur. J . Biochem. 120, 443-449. 76. Sherman, M.R., Moran, M.C., Tuazon. F.B. and Stevens, Y.-W. (1983) J. Biol. Chem. 258, 1 0 3 6 610377. 77. Reichman, M.E., Foster. C . M . , Eisen. L.P., Eisen. H.J.. Torain, B.F. and Simons, S . S . (1984) Biochemistry 23, 53765384. 78. Carlstedt-Duke, J . , Stromstedt, P.-E., Wrange, O . , Bergman, T., Gustafsson, J.-A. and Jornvall. H. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 4437-4440. 79. Hollenberg, S.M., Weinberger, C., Ong. E.S.. Ccrelli, G . , Oro, A , , Lebo, R . , Thompson, E . B . , Rosenfeld, M.G. and Evans, R.M. (1985) Nature 318, 635-641. 80. Weinberger. C . , Hollenberg, S.M., Rosenfeld. M.G. and Evans. R.M. (1985) Nature 318.670-672. 81. Miesfeld, R., Rusconi, S., Godowski. P.J.. Maler, B.A., Okret. S., Wikstrom, A,-C., Gustafsson. J.-A. and Yamamoto, K.R. (1986) Ccll 46. 389-399. 82. Miesfeld, R . , Godowski, P.J., Maler, B.A. and Yamamoto, K.R. (1987) Science 236. 423-427. 83. Hollenberg. S.M., Giguere. V.. Segui. P. and Evans. R.M. (1987) Cell 49, 39-46. 84. Danielsen, M., Northrop, J.P., Jonklaas. J . and Ringold. G.M. (1987) Mol. Endocrinol. I , 816822. 85. Carlstedt-Duke, J . , Okret. S . , Wrange. 0 . and Gustafsson, J.-A. (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 4260-4264. 86. Okret, S., Stevens, Y.-W.. Carlstedt-Duke, J . , Wrange. O., Gustafsson, J.-A. and Stevens, J . (1983) Cancer Res. 43, 3127-3131. 87. Eisen, H . J . (1982) In: Biochemical Actions of Hormones, Vol. IX (Litwack. G . . ed.) pp, 255-270. Academic Press, New York. 88. Eisen, L.P.. Reichman, M.E.. Thompson. E.B., Gametchu. B.. Harrison. R . W . and Eisen, H.J. (1985) J. Biol. Chem. 260, 11805-11810. 89. Harrison, R.W., Hendry, W.J., Turney, M . . Kunkel, E., Thompson. E.B., Denton. R.A. and Gametchu, B. (1987) In: Recent Advances in Steroid Hormone Action (Moudgil, V.K., ed.) W. de Gruyter, Berlin, 467-475. 90. Rusconi, S. and Yamamoto, K.R. (1987) EMBO J . 6, 1309-1315. 9 1. Miesfeld, R . , Sakai. D . , Inoue. A . , Schena. M.. Godowski. P.J. and Yamamoto, K.R. (1988) In: Steroid Hormone Action, UCLA Symposium on Molecular and Cellular Biology (Ringold, G.M.. ed.) Alan R . Liss, New York. in press. 92. Sap. J . , Muboz, A . , Damm, K., Goldberg. Y.. Ghysdael, J . , Leutz. A , , Beug, H. and Vennstrom. B. (1986) Nature 324, 635-640. 93. Weinberger, C., Thompson, C.C., Ong, E.S., Lcbo, R. Gruol, D.J. and Evans, R.M. (1986) Nature 324, 641-646. 94. Green, S . , Walter. P . , Kumar, V . , Krust, A . , Bornert. J.-M., Argos, P. and Chambon, P. (1986) Nature 320, 134139.

238 95. Greene, G.L., Gilna, P., Waterfield, M., Baker, A . , Hort, Y. and Shine, J. (1986) Science 231, 115C-11S4. 96. Krust, A., Green, S . , Argos, P . , Kumar, V., Walter, P . , Bornert, J.-M. and Chambon, P. (1986) EMBO J . 5 , 891-897. 97. Arriza, J.L., Weinberger, C.. Cerelli, G . , Glaser, T.M., Handelin, B.L.. Housman, D.E. and Evans, R.M. (1987) Science 237, 26R275. 98. Loosfelt. H., Atger, M., Misrahi, M.. Guiochon-Mantel, A , , Meriel, C., Logeat, F., Benarous, R. and Milgrom, E . (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 9045-9049. 99. Jeltsch, J.M., Krozowski, Z . , Quirin-Stricker, C., Gronemeyer, H., Simpson, R.J., Granier, J.M., Krust, A . , Jacob, F. and Charnbon, P. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 5424-5428. 100. Conneely, O.M., Sullivan, W.P.. Toft, D . O . , Birnbaumer, M., Cook, R.G., Maxwell, B.L., Zarucki-Schulz, T . , Greene, G.L., Schrader, W.T. and O’Malley, B.W. (1986) Science 233,767-770. 101. Ahe, D.v.d., Janich, S . , Scheidereit, C., Renkawitz, R., Schiitz, G . and Beato, M. (1985) Nature 313, 706709. 102. Cato, A.C.B., Miksicek, R . , Schutz, G . , Arnemann. J . and Beato. M. (1986) EMBO J . 5 , 2237-2240. 103. Strahle. U., Klock, G . and Schutz, G. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 7871-7875. 104. Godowski, P.J., Rusconi, S . . Miesfeld. R . and Yamamoto, K.R. (1987) Nature 325, 365-368. 105. Gigukre, V., Hollenberg, S.M., Rosenfeld, M.G. and Evans, R.M. (1986) Cell 46, 645-652. 106. Berg, J.M. (1986) Nature 319. 264-265. 107. Miller, J.. McLachlan, A.D. and Klug, A. (1985) EMBO J . 4, 1609-1614. 108. Brown, R.S., Sander, C. and Argos, P. (1985) FEBS Lett. 186, 271-274. 109. Simons, S . S . , Purnphrey, J.G., Rudikoff, S. and Eisen, H.J. (1987) J . Biol. Chem. 262,96769680. 110. Rousseau, G.G., Baxter. J.D. and Tomkins. G.M. (1972) J . Mol. Biol. 67, 99-115. 111. Rousseau, G . G . and Schmit, J.-P. (1977) J. Steroid Biochern. 8, 911-919. 112. Funder, J.W. and Sheppard, K. (1987) Ann. Rev. Physiol. 49. 397-411. 113. Green, S. and Chambon, P. (1987) Nature 325, 75-78. 114. Beato, M. (1986) In: Advances in Gene Technology: Molecular Biology of the Endocrine System (Puett, D . , Ahmad, F., Black, S . , Lopez, D.M., Melner, M.H., Scott, W.A. and Whelm, W.J., eds.) pp. 25C-254. Cambridge University Press, Cambridge. 115. Govindan, M.V., Spiess. E . and Majors. J . (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 5157-5161. 116. Toohey, M.G., Morley, K.L. and Peterson, D.O. (1986) Mol. Cell. Biol. 6, 45264538. 117. Jantzen, H.-M., Strahle, U., Gloss, B., Stewart, F.. Schmid, W., Boshart, M., Miksicek, R. and Schutz, G. (1987) Cell 49, 29-38. 118. Becker. P.B., Gloss, B., Schmidt, W., Strihle, U . and Schutz, G . (1986) Nature 324, 686688. 119. Klock, G.. Strahle, U . and Schutz, G . (1987) Nature 329, 734736. 120. Renkawitz, R., Schutz. G . , Ahc. D.v.d. and Beato, M. (1984) Cell 37. 503-510. 121. Ahe, D.v.d., Renoir. J.-M., Buchou, T., Baulieu. E.-E. and Beato, M. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 2817-2821. 122. Payvar, F. and Wrange, 0 . (1983) In: Steroid Hormone Receptors: Structure and Function (Eriksson, H. and Gustafsson, J.-A, eds.) pp. 267-282. Elsevier, Amsterdam. 123. Miksicek, R., Borgmeyer, U. and Nowock, J . (1987) EMBO J . 6, 1355-1360. 124. Cordingley, M., Riegel, A.T. and Hager, G.L. (1987) Cell 48, 261-270. 125. Richard-Foy, H. and Hager, G.L. (1987) EMBO J . 6,2321-2328. 126. Reinke, R . and Feigelson, P. (1985) J . Biol. Chern. 260, 4397-4403. 127. Baumann, H . and Maquat, L.E. (1986) Mol. Cell. Biol. 6,2551-2561. 128. Klein, E.S., Reinke. R . , Feigelson, P. and Ringold, G.M. (1987) J . Biol. Chern. 262, 520-523. 129. Gehring, U . and Arndt, H . (1985) FEBS Lett. 179, 138-142. 130. Gehring, U., Mugele, K., Arndt, H. and Busch, W. (1987) Mol. Cell. Endocrinol. 53, 33-44. 131. Okret, S . , Wikstrom, A.-C. and Gustafsson, J.-A. (1985) Biochemistry 24, 6581-6586.

23 9 132. Pratt, W.B. (1988) In: Steroid Hormone Action, UCLA Symposium on Molecular and Cellular Biology (Ringold, G.M., ed.) Alan R . Liss, New York, in press. 133. Rexin, M. and Gehring, U. (1987) Eur. J . Cell Biol. 43, Suppl. 17, 47. 134. Webb, M.L. and Litwack, G . (1986) In: Biochemical Actions of Hormones, Vol. XI11 (Litwack, G . , ed.) pp. 379-402. Academic Press, New York. 135. Economidis, I . V . and Rousseau, G . G . (1985) FEBS Lett. 181,47-52. 136. KovaEiE-Milivojevit, B. and Vedeckis. W.V. (1986) Biochemistry 25, 826623273. 137. Ali, M . and Vedeckis, W.V. (1987) J . Biol. Chem. 262, 6771-6777. 138. Ah. M . and Vedeckis, W.V. (1987) J . Biol. Chem. 262, 6778-6784.

This Page Intentionally Left Blank

B.A. Cooke. R.J.B. King and H.J. van der Molen (ed5.) Honnunes and their Actions, Part I 01988 Elsevier Science Publishers BV (Biomedical Division)

241 CHAPTER 14

Progesterone action and receptors NANCY L. KRETT", DEAN P. EDWARDS' and KATHRYN B. HORWITZ".b Departments of "Medicine and hPathology, University of Colorado Health Sciences Center, 4200 East Ninth Avenue, Denver, C O 80262, U.S.A.

1. Introduction Corner and Allen reported in 1929 [ l ] that alcoholic extracts of the corpus luteum, but not of the follicular fluid, contain a hormone which has for one of its functions the preparation of the uterus for implantation, by inducing 'progestational proliferation' of the endometrium. Besides the uterus, other known targets of progesterone are the ciliated and secretory epithelium of the fallopian tube and oviduct; the cornified epithelium of the cervix; the mammary gland secretory alveoli; and the central nervous system where the hormone appears to affect hypothalamic appetite, temperature and respiratory centers, and pituitary gonadotropin secretion. Progesterone generally antagonizes the actions of estrogens in various tissues. During pregnancy, the corpus luteum is maintained, menstruation and ovulation are suspended, and uterine motility is inhibited, all under progesterone control. In this chapter we briefly review the biological actions of progesterone and progestin derivatives in normal physiology, cancer treatment, and contraception, then focus o n recent biochemical and molecular work aimed at an understanding of the mechanisms of action of this hormone.

2. Physiology and clinical uses Progesterone and estrogen act synergistically on the endometrial lining of the uterine cavity. In women following menses, the proliferative phase is under estrogen control and is characterized by true glandular growth. After ovulation, progesterone leads to further endometrial thickening characterized by an increase in cell size Address correspondence to: Dr. K.B. Horwitz. Department of MedicineiEndocrinology -B-151, 4200 East Ninth Avenue. Denver, CO 80262, U.S.A.

242 rather than number, increased tortuosity of the glands and arterioles, and glandular secretion. As the corpus luteum wanes, progesterone withdrawal and local prostaglandin synthesis lead to increased uterine contractility, vasospasm, endometrial desquamation and bleeding [2]. Progesterone also acts synergistically with estrogen in the normal development of the breast. Estrogen stimulates cell mitosis and growth of the ductal system, while lobular development and differentiation is dependent on progesterone. When estrogen is administered in the absence of progesterone, the tubular system proliferates and the ducts dilate resulting in the formation of cysts and fibroses. These changes are comparable to those observed in fibrocystic disease and are suppressed by progestins, so that normal breast development requires that estrogen and progesterone be administered together [ 2 4 ] . Depending on the physiological state, progesterone may antagonize estrogen action. One effect of estradiol is to increase the levels of progesterone receptors (PR). Binding of progesterone to its receptors then leads not only to progestational effects, but also to antiestrogenic effects by causing a reduction in estrogen secretion into the systemic circulation; by stimulating the enzyme 17phydroxysteroid dehydrogenase which converts estradiol to the less active estrogen estrone; and by lowering the levels of estrogen receptors in cells thereby decreasing the ability of the target tissue to respond to estradiol (see Ref. 4 for a review). Progesterone also acts as an antiestrogen in cultured normal and malignant breast cells. Normal human epithelial cell cultures obtained from reduction mammoplasty proliferate in response to estradiol, while R5020, a synthetic progestin, slows their growth [4]. The effects of progestins in breast cancer have been studied mainly in T47D,, cells, a human breast cancer cell line that expresses high levels of PR in the absence of estrogen receptors, which permits the study of pure progestin effects without interference by estradiol. In these cells R5020 suppresses growth in a steroid specific manner [ 5 ] . One treatment paradigm for tumors of steroid responsive tissues (the endometrium, ovary and breast) is therefore the use of progestins to inhibit growth. The presence of PR in the tumor is a good indication for therapy success. Data from studies of 105 cases of endometrial tumors indicate that PR-positive tumors have an 89% rate of response to progestin therapy [6]. Similarly, breast tumors that test positive for PR have a 70% objective response to progestin therapy (see Ref. 7 for a review). Progesterone as a contraceptive has several effects but it acts mainly by inhibiting ovulation and creating a uterine environment hostile to implantation [S]. Exogenous progesterone disrupts the hypothalamic-pituitary gonadotropin axis and blunts the midcycle FSH and LH surge causing an inhibition of ovulation. In the reproductive tract, progestins thicken the cervical mucus thereby impairing sperm transport, and decrease the motility and secretory activity of the fallopian tubes and uterus. Based on the observation that progesterone is necessary for embryo implantation and maintenance of pregnancy, it has long been suspected that an agent

243 which blocks progesterone’s action could prevent conception. In recent studies, a synthetic antiprogestin called R U 486, has shown promise as a contraceptive agent [9]. This progesterone analog has high affinity for PR but has not agonist actions, and competitively inhibits the binding of progesterone to receptors [lo]. In a recent clinical trial, R U 486 was administered to 11 women who were 6-8 weeks pregnant and in nine cases the pregnancy was interrupted. Another potential use of RU 486 is as a postcoital contraceptive; when given at the end of the menstrual cycle on days 26 to 27, it simulates uterine progesterone withdrawal and induces menstruation whether or not there is a pregnancy. However, in a clinical trial using monkeys [ 111, the first cycle following the use of RU 486 was lengthened from 15 days to 31 days, which is clearly an undesirable side-effect. Further studies on the efficacy and dosage of R U 486 are warranted.

3. Mechanisms of action Progesterone is a small, hydrophobic molecule, that crosses the plasma membrane without impediment, and once inside the cells, binds tightly but reversibly to receptor (PR) proteins. Binding of the hormone rapidly transforms the proteins by triggering a conformational change, so that they bind tightly to DNA and regulate the transcription of a small number of genes [12]. This is accomplished by the specific interaction of the hormone-receptor complexes with a class of transcriptional enhancers termed hormone regulatory elements. These are characterized by a wellconserved consensus sequence typically located upstream of the transcription initiation site, though they are occasionally also found in the coding region [13-151. From these functional properties of the receptors it can be deduced that at the very least, the proteins must contain domains involved in hormone binding and DNA binding; deductions that recent cDNA cloning data confirm [16].

3.1. Recent technological developmerits There have been a number of technological advances in recent years which are responsible for much of the progress made in our understanding of steroid receptor mechanisms. These include the ability to purify receptors, affinity labeling techniques, production of antireceptor antibodies for direct detection of receptor proteins, and isolation and characterization of hormone responsive genes as well as the genes for receptors themselves.

3.1.1. Receptor purification Two steroid affinity resins have proven to be most useful in the purification of PR. One, developed by Grandics et al. [17], is a deoxycorticosterone derivative modified through the 21-carbon and linked to Sepharose 2B through an epoxide and a

244 stable spacer arm. The other developed by Renoir et al. [18] is a carboxylic acid derivative of deoxycorticosterone coupled to Sepharose 4B through a diamino spacer arm. Both resins have been used to purify PR from chicken oviducts in nearly homogeneous form. More recently, Logeat et al. [19] used a monoclonal antibody (MAb) prepared against chick PR for immunoaffinity purification of receptors from rabbit uterus. The MAb was chemically cross linked to protein A Sepharose and PR were isolated by a single immunoaffinity chromatography step. Monoclonal antibody affinity chromatography has the advantage of giving higher receptor yields than steroid affinity chromatography and can be used to isolate liganded as well as unliganded receptors.

3.1.2. Affinity labeling of receptors The synthetic photoactive progestin, R5020, was first shown to be a suitable affinity ligand for progesterone receptors by Dure et al. [20]. Exposure of receptor[3H]R5020 complexes to UV light in vitro resulted in covalent coupling of the ligand to PR, and allowed detection of binding proteins by denaturing SDS-gel electrophoresis. Horwitz and Alexander [21] later improved the efficiency of covalent attachment by using UV light at 300 nm, and developed an in situ photoaffinity labeling method for intact cells that involves UV irradiation of cell monolayers after their incubation with [3H]R5020. This method gives a more efficient cross linking of [3H]R5020 and minimizes proteolytic artifacts that might occur during labeling in vitro. The synthetic progestin Org 2058 and the antiprogestin RU 486 are also useful as photoaffinity ligands for PR [22,23]. 3.1.3. Anti-receptor antibodies Both polyclonal serum antibodies and monoclonal antibodies have now been produced against PR. Highly specific polyclonal antibodies have been raised against purified preparations of rabbit uterine and chicken oviduct PR [24,25]. Gronemeyer et al. [26] produced rabbit antisera to chick oviduct A and B receptors isolated separately from electrophoretic gels. Feil [27] was able to produce antibodies to PR by immunization of guinea pigs with partially purified PR from rabbit uterus. The degree of cross reaction of these guinea pig antibodies with other antigens was not reported so the specificity of this antiserum is uncertain. Logeat and co-workers [28] were the first to produce monoclonal antibodies to PR. Five mouse MAbs were obtained using rabbit uterine PR as immunogen. All cross-react with other mammalian PR including rat, guinea pig and human, but fail to detect avian receptors. More recently Toft and co-workers [29] produced five MAbs to chicken oviduct PR. Two of these antibodies designated aPR-6 and aPR-22 cross react with certain mammalian PR; the others react only with avian receptors. The aPR-6 antibody cross reacts with receptors from human breast cancer cells and was used by us for immunoaffinity purification of human PR. This material was used as immunogen for production of three MAbs to human PR [30].

245

3.1.4. Cloning of the PR c D N A Over the years there has been considerable debate whether steroid receptors regulate gene expression by directly binding to DNA of hormone responsive genes. A number of hormone responsive genes have now been shown to contain specific sequences that bind receptors preferentially in vitro and that are also required in vivo for biological response. These DNA regions, which are separate from and usually located upstream of inducible promoters, have become known as hormone response elements (HRE). Evidence that steroid receptors function as DNA binding proteins is further supported by structural information obtained from recently cloned cDNAs for estrogen [ 16,311, glucocorticoid [32,33], and progesterone receptors [34-371. Two regions of high sequence homology have been identified among the different steroid receptors. One is a cysteine rich region in the central portion of the protein. This is predicted to be a DNA binding domain because of similarities with cysteine repeat structures in other DNA binding proteins. The second region near the C-terminus is rich in hydrophobic sequences. and is predicted to be the steroid binding domain. The N-terminus shows the greatest divergence in sequence among steroid receptors, is most antigenic, and may be involved in protein-protein interactions on chromatin. Mutagenesis studies with a human glucocorticoid receptor (GR) expression vector have confirmed that these structural regions indeed represent discrete functional domains for DNA and steroid binding [38]. In this respect steroid receptors resemble certain ptokaryotic gene regulatory proteins which also contain separate domains for DNA a n d effector binding. Thus it appears that steroid receptor binding to sequence specific DNA is in part responsible for hormonal control of gene expression, though other nuclear factors are also likely to be involved (see below).

3.2. Progesterone receptor structitre We review below the major controversies dealing with PR structure including the functional significance of two hormone binding forms, the proteins comprising native holoreceptors, and the role of phosphorylation. 3.2.1. The A - and B-receptor questiori Structural analyses of steroid receptor binding proteins either by affinity labeling techniques or by immunoblotting with antireceptor antibodies show a single class of steroid binding proteins for estrogen (ER), glucocorticoid (GR) and androgen receptors (AR). By contrast receptors for progesterone, at least in chicken oviduct and in human breast cancer cells, appear to be composed of two types of steroid binding proteins differing in molecular mass by = 25000 Da (see Ref. 39 for a review). The two forms have been termed A- and B-receptors. In the chicken oviduct A and B are 79 and 108 kDa, respectively as estimated by SDS-polyacrylamide gel electrophoresis. The rabbit B-protein is also = 108 kDa. Human A and B receptors

246 are larger with apparent molecular masses respectively of 94 and 120 kDa. The differences in molecular mass among PR of different species may be either an artifact of electrophoresis or due to secondary differences in covalent modification, since recent cDNA cloning data predict a molecular mass for the B-protein of rabbit and human PR of = 99 kDa [36,37]. The A-protein, if it is synthesized from a downstream methionine codon on the PR mRNA (see below), would be = 81 kDa for both rabbit and human. The natural cellular existence of two steroid binding proteins for PR remains controversial. Milgrom and co-workers have argued that the A-receptor is an artifical degradation product of B that forms in vitro by proteolysis. This argument is based on the finding that use of proteolysis inhibitors in preparation of cytosols and in purification of PR from rabbit uterus, results in a proportionate increase in the larger 108 kDa receptors and a simultaneous decrease in smaller receptor fragments, including one of a size similar to that of A-receptors [19]. Furthermore, cellfree translation of rabbit uterine mRNA and immunoprecipitation of translation products with anti-receptor antibodies pulls down a single protein of 108 kDa [40]. Unfortunately, the antibody used for these experiments has not been fully characterized - receptor purification data are unpublished to our knowledge, and the antibody used for the blotting studies and immunoprecipitation failed to bind to 50% of native PR in salt-containing sucrose gradients [28], suggesting that it may have very different affinities for proteins A and B, and that it belongs to the class of Bspecific antibodies commonly generated from partially purified PR. Rabbit uterine receptors may differ, however, from PR of other target tissues. It may indeed contain only a single steroid binding protein or, the second may be particularly susceptible to in vitro degradation. The data with PR isolated from other species are quite compelling in favor of the existence of two steroid binding proteins. For example, Horwitz and co-workers using in situ photoaffinity labeling methods, which minimize in vitro proteolysis artifacts, consistently find equal molar amounts of A- and B-proteins in T47D human breast cancer cells [41]. This is true for both cytosolic, untransformed PR (obtained by incubation of cells at 0°C with ['H]R5020) and nuclear bound transformed PR (obtained by incubation of cells with ['H]R5020 at 37°C). Moreover, both A and B proteins are extremely stable to in vitro proteolysis. No degradation of B to A or formation of smaller receptor fragments (in the absence of protease inhibitors) could be generated by incubation of photoaffinity labeled cytosol PR at 37°C for as long as one hour. The two proteins are also covalently modified in different ways: in the absence of hormone, B is a doublet or triplet while A is a singlet. In contrast, 30 min resident chromatin-bound B-receptors are singlets while A-receptors are doublets [41]. These different forms cannot arise from the actions of a single protease. In the chicken oviduct. several investigators have also demonstrated equivalent amounts of A- and B-proteins and each has been purified and characterized. These two receptor forms of human PR can be identified by in situ photoaffinity

247 labeling [21], and by immunoblotting with a monoclonal antibody that cross-reacts with both proteins [42]. Results from such an experiment using T47D human breast cancer cells are shown in Fig. 1. In this study, a viable suspension of T47D cells was incubated for 4 h at 0°C with 20 nM ['H]R5020 to bind cytosolic receptors. Cells were then irradiated for 2 min with UV light at 300 nm. An aliquot of cytosol was separated by SDS-polyacrylamide gel electrophoresis, proteins were electroblotted to nitrocellulose sheet, and the sheet was first probed with the A- and B-specific antibody AB-52 [30], then dried and t h e radioactive proteins were visualized by fluorography [43]. ['H]R5020 and AB-52 bind specifically to two major proteins of 120 and 94 kDa. No other proteins in cytosol are coupled to R5020 or cross-react with the antibody. Each band represents PR since ['HH]R5020 binding is displaceable by excess cold progestins, but by no other steroids [39]. Human B-receptors characteristically migrate as doublets or triplets on SDS gels; these may represent covalent modifications, particularly differential phosphorylation [41,43], or microC3H] R5020 Photoaffinity Label

AB-52 lmmunoblot

120 kDa-

-

9-receptors

94 kDa-

-

A-receptors

Fig. 1. Comparison ot human PR detectcd by photoaffinity labeling and inimunoblotting. Intact T47D cells were incubated for 4 hours at 0°C with 20 n M ['H]RS020. The cells were irradiated 2 min with 300 nm U V to photoaffinity label the receptors in sit,, then homogenized, and a cytosol was prepared which was subjected to SDS-polyacrylamide gel electrophoresis. Proteins were blotted onto riitrocellulose and P R in the right lane were visualized by immune analysis with MAb AE-52; radioactive proteins in the left lane were visualized by fluorography.

heterogeneity of the primary amino acid structure arising from translation of multiple messages [42]. Close structural similarities between A- and B-proteins have been demonstrated by peptide mapping of photoaffinity labeled PR [41,44,45]. Partial proteolysis yields a similar pattern of hormone binding peptides for A- and B-proteins below the molecular mass of A (94 kDa in human and 79 kDa in chick), but no A-like fragment is generated by partial proteolysis of B. Similar data were obtained when receptor proteolytic fragments were detected by immunoblotting with antibodies prepared against gel-purified chick A- and B-proteins [26]. That is, antigenic peptides below the size of the A-protein were similar for both A- and B-receptors. Since antigenic and hormone binding peptides below the size of the A-protein are similar for both A- and B-receptors, since B and A bind hormone (a site at the C-terminal end of the molecule) and DNA (in the center of the molecule). and since no A-specific antibodies exist, we postulate that the sequences of A are contained in B. and that B-receptors differ from A by having additional N-terminal residues. Possible mechanisms for the synthesis of such partially homologous proteins include gene duplication, alternate transcription from a single gene, synthesis of multiple messages by alternate processing of a single precursor RNA, or use of alternate translation start sites from a single message. Two putative in-frame AUG codons, satisfying the Kozak [46] rule for translation initiation sites, are present 165 bases apart in the rabbit and human P R messages [37]. These could theoretically code for two homologous proteins with the shorter protein truncated at the N-terminal end and 18000 Da lighter. Preliminary evidence of such a mechanism for chick A- and B-protein synthesis has been reported [47]. The functional roles of A- and B-proteins remains largely unresolved. A- and Breceptors separated and purified from chicken oviduct have been reported to bind with different nuclear components (see reviews in Refs. 48,49). The B-protein binds poorly to DNA but well to nonhistone chromosomal proteins: A-receptors on the other hand bound strongly to DNA. These data led to the hypothesis that B-receptors might function as a specifier, or directional protein, while A-receptors function as the effector or DNA binding gene regulatory protein. Other studies have reported, albeit by use of different receptor purification methods, that chick A- and B-proteins bind equally well to DNA-cellulose [26]. Purified chick B-receptors have also been reported to bind to specific sequences in promoter distal regions of the lysozyme gene as demonstrated by DNAse I footprinting analysis [50].We have purified human B-receptors from T47D cells by immunoaffinity chromatography. Under the conditions of our purification methods, human B-receptors bind efficiently to DNA-cellulose requiring 0.25 M salt for their elution [30]. Thus, most recent data suggests that A and B are both good DNA binding proteins. Evidence that both A- and B-receptors are important for biological response in the chick is supported by studies from Spelsberg and co-workers [51-531. I t was shown that failure of the chick to respond to progesterone during seasonal and maI -

249 turational changes was correlated with the loss of one or the other of the two forms of PR. On the other hand we find (Krett and Horwitz, unpublished) that transfected cells synthesizing only human A-receptors can respond to progestin treatment, clearly suggesting that they can function as independent receptors.

3.2.2. Native PR structure: purification studies Native PR exist in at least two structural forms, each associated with different functional activities. Untransformed receptors have low affinity for nuclei and DNA and are extractable in hypotonic buffers. Transformed receptors bind DNA and chromatin with high affinity and require high salt for their extraction. Receptor transformation can be induced in vitro by a number of manipulations including increased temperature, treatment with high ionic strength buffers, incubation with heparin or ATP, or dilution of cytosol. There are a number of properties characteristic of transformed PR. First. they sediment as small 4 s molecules on sucrose density gradients compared with untransformed PR which sediment at 8-10s. The conversion from 8-10s to 4 s is presumed to be due to dissocation of subunits that form the oligomeric untransformed receptors. Second, transformed PR have higher affinity for DNA and other polyanions due perhaps to changes in surface charge groups. Third, the receptor-ligand complex after transformation is kinetically more stable than untransformed PR. Fourth, receptor transformation does not appear to be accompanied by proteolysis of the steroid binding proteins since these remain intact. Fifth, transformation is an irreversible process (see Ref. 54 for review). Finally, in vivo both A- and B-proteins undergo transformation and tight nuclear binding equally well in response to hormone [21]. Untransformed 8-10s receptors are stabilized in vitro by sodium molybdate. Several laboratories have purified native PR in both transformed and untransformed states and have examined their protein composition. Molybdate-stabilized PR contain, in addition to A- or B-proteins, a non-steroid binding of 90 kDa, which is a heat-shock protein (hsp). This has been observed for PR of different target tissues and species [55-581 and also for molybdate stabilized glucocorticoid receptors [59,60]. Since the 90 kDa hsp associates only with molybdate-stabilized receptors it has been suggested that transformation and receptor conversion from &10S to 4 s is due to 90 kDa dissociation and unmasking of receptor DNA binding sites. If so, the 90 kDa component of 8s receptors would function to maintain receptors in an inactive state in the absence of hormone. As with other hsps, 90 kDa is a ubiquitous and abundant protein and only a small fraction is found associated with 8-10s receptors. Because of this there is concern that 90 kDa-receptor associations may be in vitro artifacts; it is not known whether these associations occur in vivo and whether these interactions are of physiological relevance to receptor function. Another non-steroid binding protein of 72 kDa has recently been described to be associated in equimolar amounts with transformed 94 kDa glucocorticoid receptors [61]. The 72 kDa component is unrelated to the 94 kDa receptor since it does not

250 TABLE I Purification of human B receptors from T47D cells by single step immunoaffinity chromatography Total protein

Cytosol pH eluate

B receptors

(mgW

(mg)

pmol.'

Yield (%)

14.40

1152.0

814

-

221'

27.1

0.021

0.132

Sp. act. (pmolimg)

Purityh (%)

0.71 1915.00

23

"B receptors in starting cytosol (estimated to be 50% of total receptors) were measured by steroid binding assay. hTheoretical purity for B receptors based on M , of 120000 = 8352 pmolimg protein. 'Receptor-hormone binding of the purified product was measured by hydroxylapatite assay with inclusion of 1% carrier albumin. Non-specific binding was determined with carrier albumin alone. Values are average determinations from three separate purifications.

bind hormone, fails to react with an antireceptor antibody, and generates a different proteolytic peptide pattern from that of the 94 kDa receptor. Like 90 kDa, the 72 kDa protein is an hsp [62]. Transformed GR are proposed to be heterodimers of one steroid-binding subunit of 94 kDa and one non-steroid binding subunit of 72 kDa. The role of the 72 kDa protein is not known, but it may function in DNA binding since addition of the 72 kDa protein to purified GR leads to increased binding of G R to specific MMTV sequences [61]. We have purified human P R B-receptors from T47D cells by immunoaffinity chromatography using an antibody designated PR-6 that binds B-receptors of chick oviducts [29] and cross-reacts with human B-proteins [30]. Single step immunoaffinity chromatography resulted in enrichment of human B-receptors (identified by immunoblotting with antibody PR-6 and by photoaffinity labeling with [3H]R5020) from a specific activity of 0.71 to 1915 pmolimg protein (or 23% purity) with 27% yield (Table I). Purity and yields as judged by gel electrophoresis and densitometric scanning of the silver-stained products were approximately 1.7-fold higher than by ligand binding due to partial loss of hormone binding activity at the elution step. A second purification step using D E A E chromatography gave further enrichment to 3720 pmolimg protein (or 44% purity) to yield only B-receptors plus one other component; a 72 kDa protein present in approximately equivalent amounts. Based on physiochemical properties, single-step immunoaffinity purified B-receptors were in the native transformed state. Isolated receptors were maintained as undegraded 120 kDa doublets (Fig. 2); they retained their hormone binding activity; and they displayed the correct steroid specificity for PR. Isolated B-receptors also bound efficiently to DNA-cellulose requiring 0.25 M salt for their release. They sedimented as 4 s monomers on salt-containing sucrose gradients and as a 6 s peak in the absence of salt. All these confirm their transformed state. In addition, under these conditions, purified B-receptors were free of the 90 kDa receptor-associated heat shock protein that always copurifies with 8s untransformed receptors in other

25 1

200-

B-Receptors

116-

97-

Silver

Stain

Immunoblot

Fig. 2. Analysis of human B-receptors purified by single step immunoaffinity chromatography from T47D cytosols. B-receptors were purified by circulating cytosols from T47D cells over a column prepared by cross-linking PR-6 to protein-A-Sepharose. Proteins bound to the column were eluted with buffer at pH 11.5. Eluates were separated by SDS-PAGE. and the gels were silver stained or blotted to nitrocellulose and analyzed with PR-6.

systems, and they were free also of the nonhormone binding 108 kDa antigen that copurifies with chick oviduct PR (Fig. 2). However, in addition to the 120 kDa Bdoublet, three silver stained bands at 72, 62 and 58 kDa are detected in the purified preparations. The proteins are not reactive with PR-6 by immunoblotting, nor do they bind ['HH]R5020. Thus they are not receptor fragments but unrelated proteins. The two smaller proteins are abundant cellular proteins that bind the immunomatrix nonspecifically, but the 72 kDa protein is receptor-associated since it can not be purified from receptor-depleted cytosols. It appears to be similar to the GR-associated 72 kDa protein described by Gustaffson et al. [61].

3.2.3. Native PR structure: immune analyses The partially purified B-receptors from T47D cells were used to immunize mice, and three monoclonal antibodies (MAbs) against human PR were produced. They

252 have been subcloned and are clonally stable. Their identifying codes and other properties are summarized in Table 11. Although mice were injected only with Breceptors, the production of one antibody (AB-52) that cross-reacts with both Aand B-receptors again attests to the structural homology of these two proteins. The three IgGs are capable of shifting labeled PR on sucrose gradients. We have not

Gradient Fraction NS

PR-6

18

15

kDa - 200

12

-

- 97.4

c>

0

- 68

9

X

k

- 43

0

- 6 U C

3

- 25.7

0

m

E

0

3 - 18.4

m

11I

- 12.3

I

0

- 0

1 5 10 Fraction Number

15

20

(TOP)

Fluorogram

Fig. 3. Separation of untransformed 8s B-receptors from 8s A-receptors by PR-6 in the presence of molybdate. Suspended T47D cells were incubated at 0°C with 80 nM [‘H]R5020. photoaffinity labeled in situ, and a labeled cytosol was prepared in buffer containing molybdate. Aliquots (200 pl) of the cytosol were incubated with antibody PR-6 or a nonspecific control antibody (NS); both sets were then incubated with a second anti-mouse IgG, and the mixtures were layered on 5-20%, molybdate-containing sucrose gradients and centrifuged. Gradient fractions (200 p1) were collected; 25 pl were counted to obtain the radioactive profiles (left panel), and 175 pI from the fractions of interest (fractions labeled a-d) wcre subjected to electrophoresis on SDS-PAGE. transferred to nitrocellulose, and the filter was used for combined immunoblotting (not shown) and fluorography (right panel). 3.28 mg proteini200 pl and 71 1 166 cpm specific countsi200 pl were loaded on the gradient. “C-Labeled BSA was included as a 4s scdimentation marker on gradients, and its position is marked by the arrow. I4C-Labeled molecular weight markers were run in a parallel electrophoretic lane and their position is shown on the right of the fluorogram.

253

yet determined whether B-30 and B-64 bind the same or different epitopes unique to B; clearly AB-52 binds a region common to both A- and B-proteins. The synthesis of the monoclonals has been described in detail elsewhere [30]. Because it is B-specific and binds to native human PR, MAb PR-6 was used to study the nature of the association between the A- and B-proteins in the untransformed 8s state, and in the transformed 4 s state [43]. This is of interest since there are conflicting models for the molecular interaction of the A- and B-proteins. One model holds that A and B dimerize [48,49] and that they are subunits of a larger holoreceptor; the other model holds that A and B exist as separate 8s molecules [54,631. Using PR-6 to immunoprecipitate B-receptors, we have been unable to co-precipitate A-receptors even in the presence of molybdate. This should have been possible if A and B are tightly associated. We therefore tested the association of A and B more extensively using sucrose density gradient analysis and in situ photoaffinity labeling. We reasoned that if the dimeric subunit model is correct, addition of PR6 to receptors stabilized in the 8s state would shift both A- and B-proteins to the bottom of sucrose density gradients, but that A would not be shifted if the two proteins form independent 8s holoreceptor complexes. Figure 3 shows a study in which human PR, covalently labeled in situ with ['H]RS020 by UV irradiation, were incubated with PR-6 or a control antibody and then sedimented on sucrose gradients in the continuous presence of molybdate and protease inhibitors to maintain intact and native PR conformation. Aliquots of every gradient fraction were counted to obtain the ['H]R5020-binding profile (Fig. 3, left) and additional aliquots of the bottom fractions (a,c) and peak 8s fractions (b,d) were analyzable by gel electrophoresis and fluorography since they were photoaffinity labeled (Fig. 3, right). The control antibody had no effect on 8s sedimentation of PR, and electrophoretic analysis showed that all B- and A-receptors remained in the 8s peak. In contrast, after addition of PR-6, half of the 8s radioactivity was shifted to heavier sedimenting forms. Electrophoretic analysis of the shifted photoaffinity labeled receptors at the bottom of the gradient (fraction c) as well as the residual receptors in the 8s peak (fraction d), showed only B present in the antibody-shifted fraction while A, separated from B, remained at 8s. We conclude that A and B do not dimerize but that each exists as a separate 8s receptor either because of self-association, or because of association with non-hormone binding proteins. Two such proteins of 90 and 72 kDa co-purify with untransformed human B-receptors [30]. PR-6 can also separate B-receptors from A-receptors that have been transformed to the 4 s species by treatment with salt. Approximately half of the radioactivity seen in the 4 s peak in the presence of control antibody is shifted to heavier aggregates upon addition of PR-6 and a secondary antibody. The control 4s peak contains both A and B, but only A remain at 4 s after PR-6 addition while all the B-receptors shift to heavier sedimenting fractions. It appears then, that antibody PR-6 cross-reacts both with the native 8s as well as the transformed 4 s form of B-receptors, and that

like the 8s receptors, there are two types of 4s species containing either A-protein or B-protein, but not both. The 72 kDa non-hormone binding protein co-purifies with transformed 4s B-receptor [30]. Our working model for the structure of native PR is that the untransformed 8s receptors are complexed to two non-steroid binding proteins of 72 and 90 kDa. Transformation results from dissociation of the 90 kDa protein exposing DNA binding sites masked in 8s receptors. Transformed 4s receptors which bind tightly to DNA may be heterodimers composed of one steroid binding protein and the 72 kDa protein. The A- and B-steroid binding proteins do not form A-B dimers but exist as separate 8s and 4s molecules that function independently as DNA binding proteins to mount a biological response.

3.2.4. Native receptor structure: phosphorylation That steroid receptors are phosphoproteins is now indisputable in view of recent publications that attest to this fact for estrogen [64], glucocorticoid [65], vitaminD, [66], and progesterone receptors [43,67]. Phosphorylation is thought to be critically important for functional activity, just as the phosphorylation and endogenous kinase activity of the cell surface receptors is crucial for their actions [68,69]. However, just what role this covalent modification plays in steroid receptor action is still the subject of speculation. Among others, phosphorylation/dephosphorylation reactions are thought to: (a) convert receptors from an ‘inactive’ nonhormone binding state to an ‘active’ hormone binding state [70]; (b) ‘transform’ receptors to forms that bind tightly to chromatidDNA [71]; ( c ) be involved in the association or dissociation of nonhormone binding proteins to receptors (721; (d) modulate the DNA binding capacity of receptors and/or their nuclear down-regulation [73]. The strongest evidence that receptor activation to the hormone binding state is dependent on phosphorylation comes from the work of Pratt and co-workers with GR. They show that the receptors are phosphoproteins [65], that conversion to an inactive form can result from phosphatase activity [74,75], or energy depletion [76], and that ATP can restore hormone binding activity [77]. Auricchio et al. [78] have described a similar ATP-dependent activator for ER. Hormone-dependent transformation is very rapid; PR in breast cancer cells are transformed within 1-2 min of progestin treatment, for example. Though this transition occurs for all steroid receptors, there is no agreement on a mechanism. For estrogen receptors, the mechanism is thought to involve dimerization of hormone-binding subunits [79], but another mechanism involving receptor proteolysis by endocatalytic activity was recently proposed [go]. For G R (the most extensively studied system) and PR, the proposed mechanisms include dephosphorylation [74,75,81-841, binding of a heat-stable cytoplasmic modulator protein coupled to subunit dissociation, ATP binding and phosphorylation [71,86] interaction with a tRNA [87], and subunit dissociation [88]. While some of these are not necessarily mutually exclusive (i.e., binding of a modulator plus phosphorylation), others clearly are (receptors probably do not simultaneously phosphorylate and dephosphorylate). Furthermore, most of the studies were done

255 in vitro and their relevance to endogenous mechanisms is not addressed. For example, it is well known that steroid receptors are very sensitive to proteolytic attack, and it is conceivable that limited proteolysis occurring in vitro could expose a DNA binding site through a mechanism different from that operating in vivo. Transformed receptors bind tightly to chromatin or DNA and regulate expression of specific genes. What happens then? There must be a mechanism to terminate nuclear receptor action but virtually nothing is known about this stage. We have shown that for both E R and PR, nuclear binding is followed several hours later by a receptor down-regulation step [89,90] and have shown, by immunoblot analysis, that down-regulation is due to a real decrease in the amount of protein [42] and not simply loss of hormone binding or exchange capacity [92]. Auricchio et al. [92] have shown that a nuclear phosphatase can decrease binding of estradiol to ER in vitro and suggest that a similar mechanism accounts for down regulation. However, dephosphorylation alone cannot explain the observed loss of immunoassayable protein. Our studies for PR show that both A- and B-receptors undergo a second round of phosphorylation while they are bound to chromatin, which precedes down-regulation [42]. This phosphorylation step may be involved in subsequent down-regulation, perhaps by reducing receptor affinity for DNA or by promoting the loss of the 72 kDa protein thought to be a partner in receptor-DNA binding [61]. Similar hormone-dependent phosphorylation, probably also occurring on chromatin, has been described for rabbit uterine PR [73] and the vitamin D, receptors [66]. However, Garcia et al. [93], using chick oviduct PR, have questioned the existence of this nuclear phosphorylation step and, they suggest further that Areceptors are not phosphorylated.

3.3. lntracellular localization It has long been thought that unoccupied steroid receptors are cytoplasmic proteins that translocate to nuclei upon binding with hormone. This is based on the fact that steroid-free receptors are present in the cytosols of most target tissues, but become tightly bound to nuclei after hormone addition. This concept has increasingly come under question. For example, (1) cytosol-nuclear compartmentalization of unoccupied receptors may be dependent in part on the method of cell lysis and subcellular fractionation [94-96]. (2) Substantial amounts of steroid-free ER sites are found in nuclei of breast tumor cells [97-991. (3) Gorski and co-workers prepared cytoplasts (enucleated cells) and nucleoplasts (nucleus plus a small ring of cytoplasm) from GH3 pituitary cells and measured the distribution of unoccupied receptors for estrogen, progesterone and glucocorticoids between these two fractions. Unoccupied receptors for all three steroids were found almost exclusively in nucleoplasts [100,101]. This suggests that receptors are nuclear in the absence of hormone and that cytosol localization of unoccupied receptors after cell disruption is due to a redistribution artifact. (4) Immunocytochemical studies with antireceptor monoclonal

256 antibodies to E R [lo21 and PR [103,104] also show that unoccupied receptors are predominantly nuclear in a number of different target tissues. Milgrom and co-workers [ 1051 recently developed an immunogold method for detection of PR in the rabbit uterus and have examined the effect of hormone addition on receptor localization at the ultrastructural level. PR were found to be predominantly nuclear in the presence and absence of hormone, but a small amount was detectable in the cytoplasm which was not apparent at the light microscopical level. These cytoplasmic PR were localized over endoplasmic reticulum and clusters of free ribosomes and may likely represent newly synthesized protein. No PR were located in the plasma membrane. Within the nucleus, unoccupied PR were associated with condensed chromatin which became more dispersed after hormone addition. These ultrastructural studies indicate that steroid-free PR translocate from their site of synthesis in the cytoplasm to the nucleus in a hormone independent manner, and that addition of hormone changes their intranuclear localization. Whether unoccupied G R are totally nuclear is less clear. Immunocytochemical studies with anti-GR antibodies show both cytoplasmic and nuclear localization at the light level, and an increase in nuclear staining simultaneous with decreased cytoplasmic staining in response to addition of glucocorticoids [106-109]. It may well be that G R have a different intracellular distribution from the sex-steroid receptors. We have performed immunocytochemistry of PR in human breast cancer cells with Toft’s aPR-6 MAb and with the MAbs produced in our laboratory against human PR. Typical immunoperoxidase staining of formalin fixed hormone-free T47D cells with aPR-6 is shown in Fig. 4A. Peroxidase staining is predominantly nuclear and there is some cell to cell heterogeneity. Some nuclei stain strongly, others moderately, and some nuclei react weakly. There is also weak cytoplasmic staining which is above the background seen with control antibodies. This may represent the small fraction of PR seen on endoplasmic reticulum and free ribosomes at the EM level [105]. Control incubations with a nonspecific IgG eliminated all staining (Fig. 4B). Omitting either secondary biotinylated antibody or the avidin-biotin complex also gave no staining reaction. Finally, pre-incubation of the anti-receptor antibody with highly purified human PR blocked essentially all staining. The immunohistochemical assay, therefore, appears to be specific for PR. The three MAbs produced to human PR (Table 11) gave a similar nuclear staining pattern (not shown). We found that a stronger staining signal was obtained by use of all three MAbs as a mix than was obtained with any single MAb alone. This suggests that at least some of these MAbs are against different epitopes and are able to bind receptors simultaneously. The same immunocytochemical procedures have also been used to stain PR in frozen tissue sections. Specific nuclear staining was obtained in ductal and glandular epithelium of human endometrium, smooth muscle cells of human myometrium, ductal epithelial cells of normal breast and fibroadenoma, and cancer cells of breast tumor biopsies. These MAbs should therefore be of practical value for clinical immunocytochemical detection of receptors in breast tumors.

257

Fig. 4. Immunocytochemistry of formalin tixed 1 4 7 0 cell cultures using a-PR-6 anti-receptor MAb (left) and a control MAb (right). Immunocytocheinistry was performed with T47D breast canccr cells grown lixed for IS min with 3.7% buffered formalin. followed by as monolayers in chamber slides. Cells WCI-c a permcabilization step with Triton X- 100 (O.lC+) for antibody penetration into the cell. Immunocytochemistry was performed by thc indirect avidin-biotin immunoperoxidase method using diaminobenzidine as the chromagen.

3.4. Receptor function: regulution o f gene expression Steroid receptors control gene expression in target cells. An understanding of the molecular interactions between receptors and specific nuclear components is crucial for elucidating the mechanism of hormone action. Receptor binding to at least three structural elements of nuclei has been described. These include the nuclear matrix, nucleoacidic chromatin proteins (acceptor proteins), and specific DNA sequences in 5‘ upstream elements of hormone responsive genes. T A B L E I1 Monoclonal antibodies to human PR MAb

Antibody subtype

Reactivity” A

B

B-30 B-64 AB-52 “Reactivity with 94 kDa A- or 120 kDa B-receptor was determined by immunoblotting. See Ref. 30 for details.

258

3.4.1. Nuclear matrix Specific saturable binding sites for estrogens and androgens have been identified in the nuclear matrix of avian liver and prostate, respectively [110,111]. A biological role for nuclear matrix binding sites is suggested by the studies of Simmen et al. [112]. They observed that the presence of specific binding sites for estrogens in the nuclear matrix of liver cells was estrogen dependent and was prevented by the addition of antiestrogens. Recent studies have directly demonstrated, by use of antireceptor MAbs. that nuclear matrix-associated estradiol binding sites are indistinguishable from E R found in cytosol and salt nuclear extracts [113]. Since actively transcribed genes and RNA processing are associated with the nuclear matrix (see Ref. 114 for review) one could speculate that steroid receptor binding to nuclear matrix of target cells may be of functional importance by condensing receptors in regions of actively transcribed genes. However, nuclear matrix binding sites for progestins have not been described. 3.4.2. Acceptor proteins Spelsberg and co-workers have proposed that specific chromosomal proteins or ‘acceptors’ for steroid receptors are complexed to DNA. Chromatin isolated from chicken oviduct contains sites that bind PR saturably, specifically and with high affinity [115]. Nuclear acceptor sites specific for estrogen receptors have also been demonstrated in appropriate target tissues [116-1181. Acceptor site activity has been reconstituted in vitro by reannealing partially purified nucleoacidic proteins from chromatin of chick oviducts, to pure hen DNA. Reconstituted DNA-protein complexes bind PR with properties indistinguishable from PR binding to intact chromatin. Since PR binding to naked DNA is non-saturable and of low affinity, Spelsberg suggests that P R bind with the protein components of acceptor DNA-protein complexes. This is supported by the finding that receptor binding to reconstituted nuclear acceptors is destroyed by proteases (1191. Acceptor proteins do not confer target cell specificity since these proteins are found in non-target tissues as well. The failure of PR to bind specifically to non-target chromatin is explained by the existence of masking proteins. That is, acceptors may be present in all tissues but are unmasked only in target cells. Spelsberg and colleagues recently reported the production of monoclonal antibodies against acceptor proteins isolated from hen oviduct [ 1201. The antibodies inhibit the binding of PR to nuclear acceptor sites in a cell free system and they appear to bind directly to acceptor proteins and not to DNA or receptors. These antibodies will be used to isolate and identify specific acceptor proteins and to further characterize their biological and chemical nature. 3.4.3. D N A hormone response elements Most recent studies of steroid receptor interactions with nuclear sites have focused on DNA sequences called hormone response elements adjacent to the promoter of

2.59 hormonally induced genes. The progestin responsive genes studied most widely are the chick oviduct lysozyme and ovalbumin genes. In both genes 5' flanking sequences in a region ;= 220 bp upstream from the transcription initiation site have been shown by gene transfer and deletion analysis to be required for hormone induced transcription in vivo [121,122]. Sequences in the same regions bind purified P R in vitro [121,123-12.51. The rabbit uteroglobin gene also contains preferential binding sequences for P R . These progestin response elements (PREs) are much further upstream from the transcription start site than the PREs of the chick oviduct genes. One site is between -2946 and -2608, another at -2568 and -1842, and a third in the first intron between + 197 and + 1054. Progestins have been described to induce the synthesis of E G F and insulin receptors [5], and of sets of secreted and intracellular proteins in human breast cancer cells [126,127]. The cDNAs for EGF, insulin and one of the progestin-induced proteins, a 2.50 kDa cellular protein, have been cloned and partially characterized [128-1301, but have not been analyzed for potential PREs. The most extensively studied hormone reponsive elements are found -- 200 bp from the transcription start site in the long terminal repeat (LTR) of the mouse mammary tumor virus (MMTV). They are responsive to G R as shown by in vitro DNA binding studies with purified G R [ 131-1331 and by in vivo response data with deletion mutants [ 134-1391. From these studies three important functional domains in the LTR of MMTV have been identified: two binding sites for G R - a promoter distal binding site and a promoter proximal binding site - and a third site which binds nuclear factor-one (NF-1). NF-1 is a protein found in nuclei of most eukaryotic cells which binds with high affinity to consensus sequences of TGG(A/C)N,GCCAA. NF-1 binding sequences, either actual or potential, have been identified near the origin of replication of adenovirus [140] and adjacent to transcriptional control regions of a number of inducible genes including two hormone responsive genes, MMTV [ 1411 and the chicken lysozyme gene [142]. A functional role for NF-1 in replication of adenovirus has been established and it is also believed to be a basal transcription factor. In MMTV, NF-1 recognition sequences are located within a region from -80 to -60 bp upstream from the transcription start site. This places NF-1 binding sequences between the proximal G R binding site and promoter elements. Systematic linker scanning mutagenesis studies of MMTV have shown that this NF-1 region is important for full biological response to glucocorticoids. Linker scanning mutants in the NF-1 site (constructed without altering G R binding sites) reduce response to glucocorticoids compared to wild type MMTV sequences [ 1431. These mutants showed parallel reduction in NF-1 binding measured in vitro, suggesting that glucocorticoid regulation of MMTV transcription requires binding of both G R and NF1. Further evidence that glucocorticoid regulation of MMTV transcription involves cooperative binding between G R and NF-1 was recently reported by Cordingly et al. [144] who showed that treatment of cells with glucocorticoids and activation of

MMTV transcription is accompanied by recruitment of NF-1 binding to a region between -82 and +12 bp as delimited by in vivo exonuclease protection. No binding was detectable in this region in the absence of hormone. Recent reports have demonstrated that the LTR of MMTV binds not only GR, but PR as well [125,133]. PR used in these DNA binding studies were isolated from rabbit uterus and binding was demonstrated by exonuclease DNA footprinting. Nuclear protection patterns indicate that PR and G R bind to overlapping but not entirely identical nucleotide sequences. While earlier data had suggested that MMTV transcription was not influenced by progesterone [ 145-1471, experiments from Beatos’ group [148] and from our laboratories [148]. have shown that transcription from the MMTV promoter can be induced by physiological levels of progestins in PRcontaining T47D human breast cancer cells. Cells transfected with the MMTV promoterienhancer linked to the reporter gene chloramphenicol acetyltransferase are inducible in a dose depedent manner with physiological levels of progestins. The PREs are contained in the same -- 200 bp region upstream of the transcription start site required for glucocorticoid response. Since the incidence of mammary tumors in mice is highest during pregnancy and occurs almost exclusively in females, a role for progesterone in regulation of MMTV is not entirely unexpected. Despite the earlier inability to observe direct progestin effects on MMTV transcription, it would seem logical that a sex steroid might be the primary regulatory hormone of MMTV in mammary tissues. Our recent report that the human PR gene and the human homologue of the MMTV-induced mammary oncogene int-2 share a similar chromosome location also functionally links the hormone and the virus [150]. As with MMTV there are a number of other genes that are under multiple hormonal control. The human metallothionein IIA gene is regulated by heavy metals and by glucocorticoids; the chicken lysozyme gene is induced by four classes of steroid hormones; transcription of human growth hormone is activated by glucocorticoids and thyroid hormone; and the uteroglobin gene is regulated by different hormones depending on the tissue. Expression of uteroglobin is under control of progesterone in the rabbit uterus and of glucocorticoids in the lung (see Ref. 151 for review). In the androgen-sensitive S115 mouse mammary tumor cell line. MMTV transcription is regulated by androgens as well as by glucocorticoids [152]. Although an estrogen effect has not been demonstrated. the MMTV gene may well be regulatable by all steroid hormones given the proper cellular environment. It is well established from gene transfer and expression studies using wild type and mutant DNA sequences that cis-acting HREs are an important and necessary component for hormonal control of gene transcription. It is less clear from in vitro binding data that steroid receptors function as trans-activators by direct binding to HREs. There are a number of problems with in vitro receptor-DNA binding data. First, the affinity of receptors for sequence specific DNA is little higher than for non-specific DNA. No quantitative data for steroid receptor binding with sequence

26 I specific DNA have been reported. Most studies assume that the ability to generate a DNAse protection footprint with purified receptors used in vast excess is equivalent to high affinity DNA binding. In the few studies that have attempted to measure binding affinities relative to random DNA, receptors were found to bind sequence specific DNA with only a 5-10-fold higher afinity compared to nonspecific DNA. This is much lower affinity than reported for other nuclear transcription factors. It has been argued that for a protein in the cell nucleus to distinguish between specific and nonspecific DNA, its affinity for specific and nonspecific DNAs must differ by at least three orders of magnitude [153-1551. Based on available in vitro DNA binding data this is clearly not the case for steroid receptors. Second, not all in vitro receptor binding sites correlate with mutagenesis and in vivo biological response data. Finally, it has not been possible to demonstrate an effect of the hormone on in vitro receptor-DNA binding. Two recent studies, one which assessed binding of G R to MMTV [156] and another which assessed binding of PR to the uteroglobin gene [157], reported no differences in the DNA binding properties of unliganded and liganded receptors (or agonist vs. antagonist). This calls into question the significance of receptor-DNA binding sites measured in vitro since the hormone is clearly required in vivo. It may be that the hormone does not promote transformation and binding of receptors to specific recognition sequences as once thought, but that it alters the kinetics of pre-existing receptor-DNA binding complexes. Quantitative and kinetic binding analysis have not been performed to explore this possibility. DNA binding experiments may be fawed because receptors are damaged during purification, or because associated proteins or factors required for high affinity DNA binding are removed by purification. Functional receptor binding may also require the cooperative binding of other DNA binding proteins at some distance along the DNA. Recall that cooperative binding occurs between G R bound at the HRE of MMTV, and NF-1 further downstream [ 1441. In summary, receptor-binding specific DNA sequences may be necessary but not sufficient for activation of transcription. Receptor binding to nuclear proteins, acceptors, or the nuclear matrix may be required for full functional activity. In the final analysis the role of receptor binding to DNA will be elucidated when a hormonally regulated gene transcription system can be fully assembled in vitro.

4. Conclusions Considerable progress has been made in the past few years in our understanding of the mechanism of progesterone action. Advancements have been due in large part to the availability of new techniques for structural analysis, in particular the use of affinity, immunologic and molecular probes. Interesting questions remain. For example, PR are unique among steroid receptors in that they are composed of two

steroid binding proteins. The molecular origin, functional role, and structural arrangement of the two PR proteins in the native receptor molecule remain unresolved. Questions have been raised by the demonstration that PR are associated with heat shock proteins. It will be necessary to determine whether receptor-heat shock protein interactions exist in vivo and if so, what physiological role stress proteins might play in receptor actions. No functional significance has yet been ascribed to the demonstration that PR are targets for multiple phosphorylation reactions. Our concept of hormone induced nuclear translocation of steroid receptors must be re-evaluated based on recent immunocytochemical studies with anti-receptor monoclonal antibodies. If receptors are nuclear proteins in the presence and absence of hormone, what are the nuclear translocation signals for nascent steroid receptors? And are intranuclear binding sites of unoccupied and hormone-occupied receptors the same or different? It has long been a goal to identify specific nuclear sites in target cells that bind steroid receptors. Progestin response elements upstream of inducible promoters of target genes have gained a great deal of attention since they bind receptors in vitro and are indispensable for biologic response. Although receptor binding to 5’ upstream recognition sequences may be necessary and important, it may not be sufficient to explain progestin action since binding is not of high affinity and an effect of hormone on in vitro DNA binding has not been demonstrated. Thus PR interaction with other nuclear proteins may be required for hormonal induction in vivo. Answers to these and other questions should be forthcoming in the next several years and will provide more complete insight into the action of these interesting proteins.

Acknowledgements Our studies are supported by grants from the National Institutes of Health, the National Science Foundation, and the National Foundation for Cancer Research. We are grateful to David Toft for his generous gift of antibody PR-6, and to Nancy Hart and Clairene Mraz for preparation of this manuscript.

References 1. Corner. G . W . and Allen, W . H . (1929) Am. J . Physiol. 88. 321-339. 2. Ross, G.T. (1985) In: Textbook of Endocrinology (Wilson. J.D. and Foster. D.W., eds.) pp. 206-258. W.B. Saunders, Pennsylvania. 3. Mauvais-Jarvis, P.. Kuttenn. F. and Gompel. A . (1986) Ann. N.Y. Acad. Sci. 464. 152-167. 4. Mauvais-Jarvis, P.. Kuttenn. F. and Gonipel. A . (1986) Breast Cancer Res. Treat. 8. 179-187. 5. Horwitz, K.B. and Freidenberg. G . R . (1985) Cancer Res. 45. 167-173. 6. Kauppila, A . (1984) Acta Obstet. Gynecol. Scand. 63, 41-450. 7. Sedlacek. S.M. and Horwitz, K.B. (1984) Steroids 44. 467-484.

263 8. Smith, M.A. and Youngkin, E.Q. (1984) Clin. Pharm. 3, 485-496. 9. Herrmann, W., Wyss, R., Riondel, A.. Philibert, D . , Teutsch, G . . Sakiz, E. and Baulieu, E. (1982) C.R. Acad. Sci. Paris 294, 933-938. 10. Horwitz, K.B. (1985) Endocrinology 116, 22362245. 11. Hodgen, G.D. (1985) Fertil. Steril. 44. 263-267. 12. Yamamoto, K.R. (1985) Ann. Rev. Genet. 19. 209-252. 13. Jantzen, H.-M., Styahle, U . , Gloss, B.. Stewart. F.. Schmidt, W., Boshart, M., Miksicek. R . and Schutz, G. (1987) Cell 49, 29-38. 14. Chandler, V.L., Males, B.A. and Yamamoto. K.R. (1983) Cell 33, 489-499. 15. Karin. H., Haslinger, A , , Holtgreve. A . , Richards. R.I.. Krauter, P., Westphal, H.M. and Beato, M. (1984) Nature 308, 513-519. 16. Maxwell, B.L., McDonnell, D.P., Conneely, O.M., Schulz, T.Z., Greene, G.L. and O’Malley, B.W. (1987) Mol. Endocrinol. 1, 25-35. 17. Grandics, P., Puri, R.J. and Toft, D . O . (1982) Endocrinology 110. 1055-1057. 18. Renoir, J.-M., Yang, C.-R., Formstecher, P., Lustenberger, P.. Wolfson, A , , Redeuilh, G . , Mester, J . , Richard-Foy, H. and Baulieu, E.-E. (1982) Eur. J . Biochem. 127, 71-79. 19. Logeat, F.. Pamphile. R.. Loosfelt. H . . Jolivet. A . , Fournier, A . and Milgrom, E. (1985) Biochemistry 24, 1029-1033. 20. Dure, L.S., Schrader, W.T. and O’Malley. B.W. (1980) Nature 283, 784-786. 21. Horwitz, K.B. and Alexander, P.S. (1983) Endocrinology 113, 2195-2201. 22. Lamb, D.J.. Holmes, S.D., Smith. R . G . and Bullock, D.W. (1982) Biochem. Biophys. Res. Commun. 108, 1131-1135. 23. Horwitz, K.B. (1985) Endocrinology 116. 22362245. 24. Logeat, F., Vu Hai, M.T. and Milgrom. E. (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 14261430. 25. Touhimaa, P., Renoir, J.-M., Radanyi. C., Mester, J . , Joab, I., Buchou, T. and Baulieu, E.-E. (1984) Biochem. Biophys. Res. Commun. 119. 433-439. 26. Gronemeyer, H . , Govindan, M.V. and Chambon, P. (1985) J . Biol. Chem. 260, 69166925. 27. Feil, P.D. (1983) Endocrinology 112. 39G399. 28. Logeat, F., Vu Hai, M.T., Fournier, A . , Legrain, P . , Buttin, G . and Milgrom, E. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 6456-6459. 29. Sullivan, W.P., Beito, T . G . , Propcr. J . , Krco. C. and Toft, D . O . (1986) Endocrinology 119, 1549-1557. 30. Estes, P . A . , Suba, E.J., Lawler-Heavner, J . , Elashry-Stowers, D., Sullivan, W.P., Toft, D.O., Horwitz, K.B. and Edwards, D.P. (1987) Biochemistry 26, 6250-6262. 31. Green, S., Walter, P., Kuman, R., Krust. A.. Bornert, J.-M., Argos. P. and Chambon, P. (1986) Nature 320. 134139. 32. Miesfeld, R.. Okret. S.. Wilkstrom, A.-C.. Wrange, O . , Gustafsson, J.-A. and Yamamoto. K.R. (1984) Nature 312, 779-781. 33. Weinberger, C., Hollenberg, S.M.. Ong. E.S., Harmon, J.M.. Brower. S.T., Cidlowski, J . , Thompson, E.B., Rosenfeld, M.G. and Evans, R.M. (1985) Science 228. 740-742. 34. Jeltsch, J.M., Krozowski, S . , Stricker, 0..Gronemeyer, H . , Simpson, R.J.. Gamier, J.M., Krust, A.. Jacob, F. and Chambon, P. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 5424-5428. 35. Conneely, O . , Sullivan, W.P., Toft, D.PO.. Birnbaumer, M. Cook, R.G., Maxwell, B.L., ZaruckiSchulz, E . , Greene, G., Schrader, W.T. and O’Malley, B.W. (1986) Science 233, 767-770. 36. Loosfelt, H . , Atger, M., Misrahi, M.. Guichon-Mantel. A . , Meriel, C . , Logeat, F . , Benarous, R . and Milgrom, E . (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 9045-9049. 37. Misrahi, M., Atger, M., d’Auriol, L.. Loosfelt, H . , Meriel, C., Fridlansky, F . , Guichon-Mautel, A . , Galibert, F. and Milgrom, E. (1987) Biochem. Biophys. Res. Commun. 143, 740-748. 38. Giguere, B., Hollenberg, S.M., Rosenfeld, M.G. and Evans, R.M. (1986) Cell 46, 645-652. 39. Horwitz, K.B., Wei. L.L., Sedlacek, S.M. and d’Arville, C.N. (1985) Rec. Progr. Horrn. Res. 41, 249-316.

40. Loosfelt. H . . Logeat. F.. Hai Vu. T.T. and Milgroni, E. (1984) J . Biol. Chem. 259, 1419614202. 41. Horwitz, K.B., Francis. M.D. and Wei, L.L. (1985) DNA 4, 451-460. 42. Wei, L.L.. Krett, N.L.. Gordon. D.F.. Wood. W.M. and Horwitz, K.B. (1987) Endocr. Soc. Abstr. 989. 43. Wei, L.L., Sheridan. P.L.. Krett. N.L.. Francis. M . D . , Toft, D.O., Edwards, D.P. and Horwitz, K.B. (1987) Biochemistry 26, 6262-6272. 44. Gronemeyer. H . . Harry. P. and Charnbon. P. (1983) FEBS Lett. 156, 287-292. 45. Birnbaumer. M.. Schrader, W.T. and O’Malley. B.W. (1983) J . Biol. Chem. 258. 7331-7337. 46. Kozak, M. (1986) Cell 47. 481-483. 47. Conneely. O . M . , Dobson. A.D. W.. Carson. M.A.. Beattie. W.. Huckaby, C., Maxwell, B.L., Toft, D.O.. Tsai. M.J., Schrader, W.T. and O’Mallcy. B.W. (1987) Endocr. Soc. Abstr. 988. 48. Minghetti, P.P.. Weigel, N.L., Schrader, W.T. and O’Malley, B.W. (1983)In: Gene Regulation by Steroid Hormones I1 (Roy. A . K . and Clark. J . H . . eds.) pp. 182-191. Springer-Verlag, New York. 49. Schrader, W.T.. Birnbaumer. M.. Hughes. M.. Weigel. N.L., Grody, W.W. and O’Malley. B.W. (1981) Rec. Progr. Hormone Res. 37, 583-624. 50. Von der Ahe. D., Renoir, J.-M.. Buchou, T., Baulieu, E.E. and Beato, M. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 2817-2821. S1. Spelsberg, T . C . and Halberg, F. (1980) Endocrinology 107, 12341243. 52. Leinen-Boyd. P . . Fournicr. D . and Spelsberg. T . C . (1982) Endocrinology 111. 30-36. 53. Leinen-Boyd. P., Gossc. B.. Rasmussen. K.. Dani-Martin. G . and Spelsberg, T . C . (1984) J. Biol. Chem. 259. 24111-2421. 54. Renoir. J.-M. and Mester, J . (1984) Mol. Cell. Endocrinol. 37. 1-13. 55. Joab. I . . Radanyi. C . . Renoir. M . . Buchou. T.. Catelli. M.-G.. Binart, N.. Mester. J . and Baulieu. E.-E. (1984) Nature 308. 85k853. 56. Catelli. M . G . . Binart, N . . Testas, 1.J Renoir. J . M . . Baulieu. E.E.. Feramisco. J.R. and Welch, W.J. (1985) EMBO J . 4.3131-3135. 57. Schuh. S.. Yonemoto, W.. Bruggc, J . . Bauer, V . . Riehl, R.M., Sullivan, W.P. and Toft, D.O. (1985) J. Biol. Chem. 260. 14292-14296. 58. Renoir. J.-M.. Buchou, T. and Baulieu, E . - E . (1986) Biochemistry 25, 6405-6413. 59. Sanchez. E . R . , Toft. D . O . . Schclsingcr. M.J. and Pratt. W.B. (1985) J . Biol. Chem. 260. 12398- 12401 60. Housley. P . R . , Sanchez. E . R . . Westphal. H.M.. Bcato, M. and Pratt. W . B . (1985) J . B i d . Cheni. 260. 1381S13817. 61. Gustafsson, J.-A.. Carlstedt-Duke. J . . Wrange, 0 . .Okret. S. and Wilkstrom, A.-C. (1986) J. Steroid Biochem. 24, 63-68. 62. Welch. W.J. and Feramisco, J . R . (1984) J . Biol. Chcm. 259. 4501-4513. 63. Dougherty. J . J . and Toft, D . O . (1982) J . Biol. Chem. 257. 3113-3119. 64. Migliaccio. A , , Rotondi, A . and Auricchio, F. (1986) EMBO J . 5, 2867-2872. 65. Housley. P.R. and Pratt. W.B. (1983) J . Biol. Chcm. 258, 4630-4635. 66. Pike, J.W. and Sleator. N.M. (1985) Biochem. Biophys. Res. Commun. 131, 378-385. 67. Puri, R.K. and Toft. D . O . (1986) J . Biol. Chem. 261. 565 1-5657. 68. Stadel. J . M . , Nambi. P., Shorr. R . G . L . , Sawyer. D . F . , Caron. M . and Lefkowitz. R.J. (1983) Proc. Natl. Acad. Sci. U.S.A. XU. 3173-3177. 69. Haring. H.-U.. Kasuga. M.. White. M.F.. Crettaz. M. and Kahn, C . R . (1984) Biochemistry 23, 3298-3306. 70. Nielscn, C.J.. Sando J . and Pratt, W.B. (1977) Proc. Natl. Acad. Sci. U.S.A. 74. 1398-1402. 71. John. J.K. and Moudgil. V.K. (1979) Biochem. Biophys. Res. Commun. 90, 1242-1247. 72. Mendel. D . . Bodwell. J . E . . Garnetchu, B . , Harrison. W. and Munck. A. (1986) J . Biol. Chem. 261. 3758-3763. 73. Logcat. F.. LeCunff, M.. Pamphile. R . and Milgrom. E.. Biochem. Biophys. Res. Commun. 131. 421-427.

.

265 74. 75. 76. 77.

Housley, P . R . , Dahmer. M.K. and Pratt. W.B. (1982) J. Biol. Chem. 257. 8615-8618. Nielsen, C.J., Sando, J.J. and Pratt. W.B. (1977) Proc. Natl. Acad. Sci. U.S.A. 74, 1398-1402. Ishii, D . N . , Pratt, W.B. and Aronow, L. (1972) Biochemistry 1 I . 3896-3904. Carter-Su. C. and Pratt, W.B. (1984) In: The Receptors. Vol. 1 (P.M. Conn. ed.) pp. 541-585. Academic Press, NY. 78. Auricchio. F., Migliaccio, A . , Castoria, G . . Lastoria, S. and Schiavone, E. (1981) Biochcm. Biophys. Res. Commun. 101, 1171-1178. 79. Notides, A.C., Sasson, S . and Callison. S . (1985). In: Molecular Mechanisms of Steroid Hormone Action (Moudgil, V.K., ed.) pp. 17.1-197. W. DeGruyter and Co., Berlin. 80. Puca. G . A . , Abbondanza, C . , Nigro. V.. Armetta. I.. Medici. N. and Molinari, A.M. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 5367-5371. 81. Nishigori. H . and Toft, D . (1980) Biochemistry 19. 77-83. 82. Leach, K.L., Dahmer, M.K., Hamniond. N.D., Sando, J.J. and Pratt. W.B. (1979) J . Biol. Chem. 254, 1188411890. 83. Shyamala, G . and Leonard, L. (1980) I . Biol. Chcm. 255, 6028-6033. 84. Barnett, C . A . , Schmidt, T.J. and Litwack. G . (1980) Biochemistry 19, 5446-5455. 85. Schmidt. T.J., Miller-Diener. A . . Wehh. M.L. and Litwack, G. (1985) .I.Biol. Chem. 260, 16255-16262. 86. Kurl. R . N . and Jacob, S.T. (1984) Biochem. Biophys. Res. Commun. 90. 1242-1248. 87. Tymoczko, J.L. and Lee. J . Fed. Proc. 44, 867, Abstr. No. 22667. 88. Vedeckies, W.V. (1983) Biochemiatry 22, 1983-1989. 89. Horwitz, K.B. and McGuire. W.L. (1978) J . Biol. Chem. 253, 2223-2228. 90. Mockus. M.B. and Horwitz, K.B. (1983) J . Biol. Chem. 258. 4778-4783. 91. Strobl. J.S. and Lippman. M.E. (1979) Cancer Res. 39. 3319-3327. 92. Auricchio. F., Migliaccio, A . and Castoria. G. (1981) Biochem. J . 198, 699-702. 93. Garcia, T., Testas, I.J. and Baulieu. E.E. (1986) Proc. Natl. Acad. Sci. U.S.A. 83. 7S73-7577. 94. Sheridan, P.J., Buchanan. J.M., Anselmo. V.C. and Martin. P.M. (1979) Nature 282. 579-580. 95. Sheridan. P.J., Buchanan. J.M.. Anselmo. V.C. and Martin. P.M. (1981) Endocrinology 108, 1533-1537. 96. Edwards. D . P . , Martin, P.M., Horwitz. K.B., Chamness, G.C. and McGuire. W.L. (1980) Exp. Cell Res. 127, 197-213. 97. Zava, D . P . and McGuire, W.L. (1977) J . Biol. Chem. 252. 3703-3708. 98. Panko. W.B. and MacLeod, R.M. (1978) Cancer Res. 38, 1948-1951. 99. Geier, A,. Ginzburg, R . , Stauber, M . and Lunefeld, B. (1979) J . Endocr. 80, 281-288. 100. Welshons, W.V., Krummel, B.M. and Gorski. J . (1985) Endocrinology 117, 214C2147. 101. Welshons. W.V. Lieberman. M.E. and Gorski. J . (1984) Nature 307, 747-749. 102. King. W.J. and Greene, G . L . (1984) Nature 307. 745-747. 103. Gasc. J.-M., Renoir, J.-M.. Radanyi. C.. Joab. I . , Touhimaa, P. and Baulieu, E.-E. (1984) J . Cell Biol. 99, 1193-1201. 104. Perrot-Applanat, M., Logeat, F., Groycr-Picard, M.T. and Milgrom, E. (1985) Endocrinology 116, 1473- 1484. 105. Perrot-Applanat. M., Groyer-Picard, M., Logcat, F. and Milgrom, E. (1986) J . Cell. Biol. 102, 1191-1 199. 106. Govindan. M.V. (1980) Ex. Cell Rcs. 127. 293-297. 107. Papamicail. M., Tsokos, G.. Tsawdaroglou. N. and Sekeris. C . E . (1Y80) Exp. Cell Rcs. 125, 490-493, 108. Antakly. T . and Eisen, H.J. (1984) Endocrinology 115. 19861989. 109. Wilkstrom, A.-C., Bakke, O., Okret, S., Bronnegard, M . and Gustafsson, J.-A. (1987) Endocrinology 120, 1232-1242. 110. Barrack, E.R. and Coffey, D.S. (1980) J . Biol. Chem. 255. 7265-7275.

111. 112. 113. 114. 115.

Barrack, E . R . (1983) Endocrinology 113. 43&432. Simmen, R.C.M., Means, A . R . and Clark, J.H. (1984) Endocrinology 115, 1197-1202. Alexander, R.B., Greene, G.L. and Barrack, E . R . (1987) Endocrinology 120, 1851-1857. O'Malley, B.W. (1984) J . Clin. Invest. 74, 307-312. Spelsberg, T.C., Littlefield, B.A., Seelke. R.. Dani-Martin, G . , Toyoda, H., Boyd-Leinen, P., Thrall, C. and Kon, O . L . (1983) Rec. Progr. Hormone Res. 39,463-513. 116. Ruh. T.S. and Spelsberg, T.C. (1983) Biochem. J. 210, 905-912. 117. Singh. R.K., Ruh, M.F., Butler, W.B. and Ruh, T.S. (1986) Endocrinology 118, 1087-1095. 118. Ross, P. and Ruh, T.S. (1984) Biochim. Biophys. Acta 782, 18-25. 119. Spelsberg, T.C., Gosse, B.J., Littlefield. B.A., Toyoda, H. and Seelke, R. (1984) Biochemistry 23, 5103-5113. 120. Goldberger, A . , Littlefield. B . A . , Katzmann, J . and Spelsberg, T.C. (1986) Endocrinology 118, 2235-2241. 121. Rcnkawitz, R . , Shutz, G . , von der Ahe. D. and Beato, M. (1984) Cell 37, 503-510. 122. Dean. D.C., Knoll, B . J . , Riser, M.E. and O'Malley, B.W. (1983) Nature 305, 551-554. 123. Compton, J.G., Schrader. W.T. and O'Malley, B.W. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 16 2 0 . 124. Mulvihill, E.R., LePennac, J.-P. and Charnbon, P. (1982) Cell 24, 621-632. 125. Von der Ahe, D . , Renoir, J.-M., Buchou. T.. Baulieu, E . E . and Beato, M. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 2817-2821. 126. Chalbos. D. and Rochefort, H . (1984) J . Biol. Chem. 259, 1231-1238. 127. Chalbos, D. and Rochefort, H . (1984) Biochem. Biophys. Res. Commun. 121, 421-427. 128. Chalbos, D . , Westley, B., May, F.. Aliberg. C. and Rochefort, H. (1986) Nucl. Acid Res. 14, 965-987. 129. Ullrich, A . et al. (1984) Nature 309. 418-425. 130. IJllrich, A . , Bell, J . R . , Chen, E.Y., Heuvera, R.. Petruzzelli, L.M., Dull, T.J., Gray, A , , Coussens, L., Liao, Y.C., Tsubokawa, M.. Mason. A . , Seeburg, P . H . , Greenfeld, C . , Rosen, O.M. and Ramachaadran, J . (1985) Nature 313, 756-761. 131. Payvar, F., deFranco, D . . Firestone. G . L . . Edgar, B., Wrange, 0.. Okret, S., Gustafsson, J.-A. and Yamamoto. K. (1983) Cell 35, 381-392. 132. Scheidereit, C. and Beato. M. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 3029-3033. 133. Von der Ahe, D., Janich, S., Scheidereit, C., Renkawitz, R.. Shutz, G. and Beato, M. (1985) Nature 313, 706709. 134. Chandler, V.L., Maler, B.A. and Yamarnoto, K. (1983) Cell 33, 489-499. 135. Hynes, N., van Doyen, A.J.J., Kennedy, N.. Herrlich, P., Ponta, H . and Groner, B. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 3637-3641. 136. Majors, J. and Varmus, H.E. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 58665870. 137. Ponta, H . , Kennedy, N., Skroch, P., Hynes. N.E. and Groner, B. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 102C-1024. 138. Buetti, E . and Kuhnel, B. (1986) J. Mol. Biol. 190. 41-54. 139. Kuhnel. B . , Buetti. E . and Diggelmann. H. (1986) J. Mol. Biol. 190. 57-68. 140. Nagata, K., Guggenheimer, R.A., Enornoto, T . , Lichy, J . H . and Hurwitz, J . (1982) Proc. Natl. Acad. Sci. U.S.A. 79, 634841442. 141. Nowock, J . , Borgrneyer. U . , Puschel. A.W., Rupp, R.A.W. and Sippel, A.E. (1985) Nucl. Acid Res. 13, 2045-2061. 142. Nowock. J . and Sippel. A . E . (1982) Cell 30. 607-615. 143. Kuhnel. B. (1986) Doctoral Dissertation. Catholic University of Nijmegen. 144. C'ordingly, M.G., Reigel. A.T. and Hager, G.L. (1987) Cell 48. 261-270. 145. Shyarnala. G . and Dickson, C. (1976) Nature 262. 107-112. 146. Young. H.A., Scolnick. E.M. and Parks, W.P. (1875) J . Biol. Chem. 250, 3337-3343.

267 147. Cardiff, R.D., Young, L.J.T. and Ashley, R.L. (1976) J . Toxicol. Environ. Health 1, 117-129. 148. Cato, A.C.B., Miksicek, R., Schutz, G . and Beato, M. (1986) EMBO J . 5 , 2237-2240. 149. Nordeen, S., Lawler-Heavner, J . , Barber. D. and Edwards, D.P. (1987) Endocr. SOC.69th Annual Meeting, p. 273. 150. Law, M.L., Kao, F.T., Wei, Q . , Hartz, J . A , . Greene, G.L., Zarucki-Schulz, T., Conneely, O.M., Jones, C . . Puck, T.T., O’Malley, B.W. and Horwitz, K.B. (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 2877-2881. 151. Scheidereit, C . , Krauter, P., von der Ahe, D., Janich, S., Rabenau, O . , Cato, A.C.B., Suske, G . , Westphal, H.M. and Beato, M. (1986) J . Steroid Biochem. 24, 19-24. 152. Darbre, P.D., Moriarity. A.. Curtis. S . A . and King, R.J.B. (1985) J. Steroid Biochem. 23, 379-384. 153. Lin, S. and Riggs, A . D . (1975) Cell 4, 107-111. 154. Emerson. B.M., Lewis, C. and Felsenfeld, G . (1985) Cell 41, 21-30. 155. Travers, T. (1983) Nature 303, 755. 156. Willmann, T. and Beato, M. (1986) Nature 324, 688-691. 157. Bailly, A., LePage, C., Rauch, M. and Milgrom. E. (1986) EMBO J . 5 , 3235-3241.

This Page Intentionally Left Blank

B.A. Cooke, R.J.B. King and H.J. van der Molen ( c d a . ) Hormones and their Actions. Purr I 01988 Elaevier Science Publishers BV (Biomedical Division)

269 CHAPTER 15

The pleiotropic vitamin D hormone LEONOR CANCELA, G. THEOFAN" and ANTHONY W. NORMAN Division of Biomedical Sciences, University of California, Riverside, CA 92521-0121, U.S.A .

1. Introduction Vitamin D, along with parathyroid hormone and calcitonin, plays a primary role in calcium and phosphorus homeostasis in the body. Intensive research efforts over the past several years have elucidated a role for vitamin D in many other physiological processes as well. The biological actions of this seco-steroid*" are mediated primarily through the action of its polar metabolite, 1,25-dihydroxyvitamin D, (1,25(OH),D,). There is emerging evidence that 1,25(OH),D, has many more target tissues than those involved in its classical role in the control of mineral metabolism. In addition, some of the actions of 1,25(OH),D3 may be mediated by mechanisms other than the classical steroid-receptor interaction. In this chapter we will provide a brief overview of the multiple actions of vitamin D, and the pleiotropic mechanisms by which these actions are accomplished.

2. Production and metabolism of vitamin D Vitamin D, can be obtained either through the diet or by the conversion of 7-dehydrocholesterol in the skin by the action of ultraviolet light. Extensive metabolism of vitamin D, can occur (no less than 30 metabolites of vitamin D have been isolated and chemically characterized). To obtain the most biologically active metabolite, vitamin D, is first hydroxylated in the liver to form 25-hydroxyvitamin D,, which is then converted to 1,25(OH),D3 by a cytochrome P-450 mitochondria1 en-

*Present address: Department of Biology, University of California, San Diego, La J o b . California 92093, U.S.A. *"The term 'seco' indicates that one of the rings of the steroid nucleus is broken; in the instance of vitamin D. the 9,lO carbon bond is broken thus generating a seco-B steroid.

270 zyme complex in the kidney (Fig. 1) (see Refs. 1 and 2 for review). There is recent evidence that 1,25(OH),D, may also be produced extrarenally in cells of hematopoietic origin (activated monocytes and macrophages) [ 3 ] and by the mammalian placenta [4]. Although most biological activities of vitamin D, are ascribed to 1,2S(OH),D,, some activity has been associated with a second metabolite produced by the kidney, 24R,2S-dihydroxyvitamin D, [l].Thus, at the present time it is not clear whether 1,2S(OH),D, acting alone can produce all the biological responses supported by the parent vitamin D, or whether in some instances the combined or sequential actions

VITAMIN

D

ENDOCRINE SYSTEM KIDNEY

7-DEHYDROCHOLESTEROL

ENDOCRINE MODULATORS

GROWTH HORMONE PROLACTIN

"0

(PRESENT IN SKIN)

HORMONE

30 CHEMICALLY CHARACTERIZED METABOLITES

&

VITAMIN D,

BLOOD

DIETARY SOURCES

24,25(OHl2D\ RECEPTORS CHONOROCYTES IBONEI PARATHYROID GLAND

'6lONEY PARAiHYROlOS PANCREAS PI7UITARY EGG SHELL GLANO CHORIOALLANTOIC MEMBRANE OVARY TESTES EPIOIDYMUS

'

1

PAROTID GLAND UTERUS HEART1Mur:le myoblosisl BONE THYMUS MONOCYTES MACROPHAGES B 8 T-LYMPHOCYTES loclaoled 1

L E UCOCITES COLON MAMMARY TISSUE SKIN CEREBELLUM PLACENTA CANCER CELL L l N E S I I a REABSORPTION-Ca" B PI -ABsORPT~ON-C~+*

8 PI M O B I L I Z A T I O N I A C C R E T I O N - Co*' 8 PIA W NORMAN

Fig. 1. Vitamin D endocrine system. Vitamin D,, obtained either through the diet or by conversion of 7-dehydrocholesterol in the skin, is sequentially hydroxylated in the liver and kidney to produce the active metabolites 1a,25(OH),D3 and 24R,25(OH)2D3.

271 of 24R,25(OH),D3/1 ,25(OH),D3 are required. Vitamin D was originally classified as a vitamin because of its presumed dietary requirement. However because of the capability of in vivo production and its mechanism of action, 1,25(OH),D3 can now more properly be considered a hormone in the steroid family.

3. Modes of action of 1,25(OH),D3 3.1. Introduction The mechanism by which most 1,25(0H),D3-mediated actions occur is analogous to that of the classical steroid hormones. 1,25(OH),D3 is known to interact with a specific receptor in its target cells resulting in transformation of the receptor to a form with high affinity for DNA. The hormone-receptor complex interacts with specific DNA sequences regulating the transcription rate of certain target genes to induce de novo synthesis of mRNA species. These mRNAs are translated into proteins which mediate the biological responses initiated by 1,25(OH),D,. Some recent studies [19-211 suggest that not all of the actions of 1,25(OH),D3 are explained by 1,25(OH),D3 receptor interactions with the genome. Rapid effects of 1,25(OH),D3 on stimulating intestinal calcium transport have been demonstrated which occur too quickly (within 4-6 minutes) to involve genome activation and have led to the hypothesis that some of the actions of 1,25(OH),D3 may be mediated at the membrane or by extranuclear subcellular components.

3.2. Receptor-mediated genomic interactions 3.2.1. 1,25(OH),D3 receptor characteristics The 1,25(OH),D3 receptor of the chick intestine has been extensively characterized by our laboratory and several others, and its biochemical properties reveal many similarities to classical steroid hormone receptors [5,6].The 1,25(OH),D3 receptor is a protein with a molecular mass of approximately 67 000 Da when occupied, unoccupied or in the absence of protease inhibitors. The receptor has a high affinity for 1,25(OH),D, with a K , in the range of 1-50 x lo-'' M. The specificity of the receptor for binding of 1,25(OH),D3 metabolites and analogues, as determined by competitive binding studies, parallels the biological activity of these compounds. Unlike the estrogen, progesterone and glucocorticoid receptors, it appears that the 1,25(OH),D3 steroid-receptor complex does not require a temperature-dependent 'transformation' step, i.e., the binding of 1,25(OH),D3 does not appear to result in a dramatic change in the shape and size of the receptor. It has recently been demonstrated, however, that the binding of 1,25(OH),D, to its receptor exhibits positive cooperativity; the binding of one molecule of 1,25(OH),D3 to the re-

272 ceptor leads to a conformational change in the receptor which results in increased affinity for the binding of a second molecule of 1,25(OH),D3 [7] (Fig. 2A). This conformational change may represent a dimerization of two identical receptor subunits or the site-site interaction of two binding sites on the same receptor molecule. Similar to other steroid receptors, the unoccupied 1,25(OH),D3 receptor has been shown to be primarily localized in the nucleus, and can be extracted using high salt buffers [8]. In addition to its classical target tissues, bone, intestine and kidney, 1,25(OH),D, receptors have been demonstrated in many tissues not previously recognized as targets for the hormone [2].

3.2.2. Evidence for the genoniic action of 1,25(OH),D3 One of the major effects of 1,25(OH),D, in its target tissues is the induction of a calcium binding protein, which is to date the major known protein product induced by 1,25(OH),D, [9]. Many of the tissues which possess 1,25(OH),D, receptors also express this calcium binding protein, called calbindin [2]. In the chick as well as the mammalian kidney and brain, a larger form of the protein, calbindin-D,,,, is expressed, while the mammalian intestine and placenta contains the smaller calbindin-& [lo]. The expression of calbindins in various tissues of different species appears to be regulated to different degrees by 1,25(OH),D,. In the past several years, our laboratory has been extensively studying the molecular biology of the regulation of calbindin-D,,, by 1,25(OH),D, in the chick, focusing on the duodenum where it is highly expressed, constituting 2-3% of the soluble protein of the cell. Early experiments demonstrated that actinomycin D and alpha-amanitin could block induction of calbindin-D,,, by 1,25(OH),D, [9]. It was later demonstrated that 1,25(OH),D, was able to stimulate general RNA synthesis in the chick intestine [ l l ] as well as specifically inducing calbindin-D,,, mRNA [12]. In addition, '

v A

'

1

02

06

1

I

I

I

1 4

18

22

1

06

0 2 -

Bound

[ 'H]

10

- 1.25 (OH), D, ( n M )

Fig. 2A. Scatchard plot of the specific binding of ['H]1,25(OH),D3 to chick duodenal chromatin receptor showing positive cooperativity.

273 1,25(OH),D3 receptor occupancy is correlated with calbindin-D,,, induction following administration of the hormone 1131. These results are consistent with the concept that the induction of Calbindin-D,,, by 1,25(OH),D3 involves the initiation of RNA and protein synthesis via a receptor mediated mechanism, analogous to that of other more classical steroid hormones. The production of cDNA probes to calbindin-D,,, mRNA [14] have allowed us to further elucidate the regulation of calbindin-D,,, by 1,25(OH)*D3. Northern analysis has demonstrated that there are three size species of calbindin-D,,, mRNA (2.1, 2.8 and 3.1 kb) which accumulate concurrently in the vitamin D-deficient chick intestine following a dose of 1,2S(OH),D3 [lS] (Fig. 2B). The levels of calbindinDZxKmRNA are expressed to different extents in different tissues of the chick, in accordance with the level of calbindin-D,,, synthesized, and are also regulated to different degrees by 1,25(OH),D,. For example, expression of calbindin-D,,, and calbindin-D,,, mRNA are totally dependent on the 1,2S(OH),D3 in the intestine, are present in the vitamin D-deficient kidney where its levels can be stimulated by 1,25(oH),D1, and are independent of 1,25(OH),D3 levels in the brain [lS] (Fig. 2B). Using nuclear transcription (run-off) assays, we have recently demonstrated the direct induction of transcription of the calbjndin-D,,, gene by 1,25(OH)2D3in the chick intestine [16] (Fig. 2C). 1 ,2S(OH),D3-induced transcription of the calbindin-DZXK gene occurs rapidly (significant stimulation by one hour), and is correlated with the level of occupied 1,25(OH),D1 receptors.

Fig. 2B. Northern analysis of poly A ( + ) mRNA from chick duodenum hybridized to a calbindin-Dz,, cDNA probe showing 3 mRNA species at 2100, 2800 and 3100 nucleotides.

274

Time a f t e r 1,25(0H)?Dg (hours)

Fig. 2C. Time course of calbindin-D2,, gene transcription in chick intestine following a dose of 1,25(OH),D3 to vitamin D-deficient chicks. Transcription was measured by nuclear run-off assay.

A series of experiments using vitamin D-replete chicks has demonstrated that dietary alteration of calcium and phosphorus can modulate calbindin-D,,, levels without effecting changes in steady state calbindin-D,,, mRNA levels [17], suggesting the 1,25(OH),D3 may have additional posttranscriptional effects in modulating calbindin-D,,, expression. This is further indicated by the lag time observed between the peak of 1,25(OH),D3 induced calbindin-D,,, gene transcription (3 h) and the accumulation of calbindin-D,,, mRNA (12 h). Posttranscriptional regulation of gene expression has also been demonstrated for other steroid hormones, and may involve effects on stability of the mRNAs. Along these lines, we have shown that cycloheximide treatment of vitamin D-replete chicks results in a very rapid degradation of calbindin-D,,, mRNA levels, indicating the need for continual protein synthesis, perhaps of an mRNA stabilizing protein, to allow expression of the message [18] (Fig. 2D). Taken together, these data lead to a model for the expression of calbindin-D,,, by 1,25(OH)2D3which is depicted in Fig. 3. The 1,25(OH),D3 hormone-receptor complex acts at the transcriptional level to initiate the synthesis of calbindin-D,,, mRNA, and also at the posttranscriptional level to maintain stability of the message through the production or stimulation of a stabilizing protein.

3.3. Evidence for non-genomic actions of 1,25(0H),D3 The 1,25(0H),D3-stimulated transport of calcium across the intestinal epithelium has been extensively studied in our laboratory. 1,25(OH),D, can stimulate rapid transport of calcium in vascularly perfused chick duodenal loops [19] before the ap-

275

M i n u t e s a f t e r cyclohexirnide (600u g )

Fig. 2D. Time course of degradation of calbindin-DZXK mRNA in vitamin D-replete chick intestine following administration of cycloheximide. RNA was measured by dot blot hybridization to a calbindinDZRK cDNA probe.

Fig. 3. Possible mechanism for regulation of calbindin-Dzx, gene expression by 1,25(OH)?D, in chick intestine.

276 pearance of calbindin-D,,,. This 1,25(0H),D,-mediated rapid transport of calcium has been termed ‘transcaltachia’ [20] and seems to be independent of genomic activation. Transcaltachia exhibits a polarity in that transport of calcium occurs when 1,25(OH),D, is presented to the basal lateral membrane of the intestinal epithelial cell but not the brush border membrane surface. 1,25(0H),D3-mediated transcaltachia is not inhibited by actinomycin D, but can be inhibited by the anti-microtubule agent colchicine, and by leupeptin, an antagonist of lysosomal cathepsin B. Differential centrifugation experiments reveal high specific activity of 4sCa localized in the lysosomal fraction and in pinocytic vesicles following 1,25(OH),D, treatment [21]. Therefore, 1,25(OH),D3 induced transcaltachia in the intestine appears to consist of the internalization of calcium at the brush border membrane in endocytic vesicles that subsequently fuse with lysosomes and are then transported along microtubules to the basal lateral membrane where exocytosis occurs [21].

4. Vitamin D and the maintenance of mineral homeostasis Vitamin D is essential for the development of normal bone and mineral metabolism, its primary function being to maintain normal serum calcium and phosphate levels through a direct stimulation of intestinal transport and mobilization of mineral from bone [22,23]. The classical target tissues for 1,25(OH),D,, the hormonally active form of vitamin D , are also those which have been shown to be directly involved in the regulation of mineral homeostasis, that is, the intestine, kidney, bone and parathyroid gland. In addition, during the reproductive stages in mammals, the fetoplacental unit and the mammary gland, which play an important role in the regulation of the fetusinewborn, have also been shown to contain receptors for 1,25(OH),D3 [24-27]. 4.1. The kidney

The kidney is the major, and until recently was thought to be the unique site of synthesis of 1,25(OH),D, as well as the major site of synthesis of other hydroxylated vitamin D derivatives. At low calcium levels in blood, and under the stimulating effect of parathyroid hormone [28], the 25(0H)D3-la-hydroxylase is stimulated and 25(OH)D3 is converted to la25(OH),D3 at high rate, increasing the circulating levels of the hormone. In contrast, in the presence of high calcium levels in the blood stream, the secretion of PTH is inhibited and the decrease in the circulating levels of this polypeptide hormone is paralleled by a decrease in the activity of the la-hydroxylase. At the same time, there is stimulation of the 25(OH)D,24-hydroxylase leading to the formation of increased amounts of 24,25(OH),D,. Under normal conditions, this process is tightly controlled so that the levels of 1,25(OH),D3 are maintained within a quite narrow range. Other hormones such as

277 calcitonin [29,30] and prolactin [ 3 I ] stimulate the 25(OH)D3-1a-hydroxylase activity, possibly through a direct effect upon this enzyme, although some controversy remains. O n the other hand, both insulin and growth hormone as well as other hormones from the steroid family (glucocorticoids, estrogens) exert only an indirect effect upon the activity of the renal la-hydroxylase [2]. 1,25(OH),D, has specific receptors localized in the renal tubule but it does not seem to have a direct effect upon renal calcium reabsorption. Under t h e stimulatory effect of PTH, 1.25(OH)2D, and, to a lesser extent, 25(OH)D7 and 24.25(OH),D3 interfere with the renal reabsorption of phosphate by inhibiting the synthesis of CAMPwithin the membrane of renal cells. Contradicting results have, however, been published as to the direction of this transport (32,331. Interestingly, the presence of calbindin in the mammalian. avian and even amphibian kidney has been recently described [34]. But the physiological relevance of this finding is as yet unknown. 4.2. The intestine 1,25(OH),D, is directly responsible for calcium absorption in the intestine. Furthermore, this calciotropic hormone is also involved in the synthesis by intestinal epithelial cells of calcium-binding proteins known as calbindins (abbreviated CaBP) through a direct effect on genomic expression. In mammals, 1,25(OH),D3 induces the synthesis of a 9000 D a protein (calbindin-D,,), whereas in birds there is synthesis of a 28000 Da protein (calbindin-D,,,) under the same stimulatory conditions. However, the exact relationship between calbindins and calcium transport remains obscure. In the vitamin D-deficient state, both mammals and birds have severely impaired intestinal calcium absorption with no detectable synthesis of calbindin. There appears, however, to be no synchronism between the increased cellular levels of calbindin and the stimulation of calcium transport. After administration of 1,25(OH),D, to vitamin D depleted chicks, the transport of calcium is apparent long before calbindin levels are significantly increased [20]. The complexity of the system is such that, despite extensive work, the exact mechanism involved in the vitamin D-dependent intestinal calcium absorption remains unknown (see Sections 3 and 4). Vitamin D also exerts an effect upon the intestinal absorption of phosphorus probably in relation with the stimulation of alkaline phosphatase synthesis [22,35]. 4.3. Bone Vitamin D has been known for more than half a century to be a powerful antirachitic agent, necessary for the development of normal bone. However, and oddly enough, its primary and most direct effect in this tissue is the stimulation of bone resorption [36] leading to an increase in the circulating levels of calcium and phos-

278 phorus. In response to even a slight decrease in circulating levels of calcium, parathyroid hormone synthesis is increased thus mediating the stimulation of the renal 25(OH)D3-1a-hydroxylase activity. The resulting increased levels of 1,25(OH),D3 will have a positive effect upon bone resorption concomitant with the resorptive effect of PTH itself. Considering the crucial importance of maintaining the concentration of calcium in blood constant, whether or not there is a regular intake of this mineral, the ready availability of calcium from its largest body store represented by the bone is of relevant importance. In order to understand this mechanism, it must be kept in mind that the bone is a dynamic structure which is continuously undergoing resorption by osteoclasts and apposition of new material by osteoblasts under regulation of a wide number of factors and hormones whose complex interactions are poorly understood despite intensive efforts made in that area by a large number of laboratories. Under normal physiological conditions, however, bone formation and bone resorption are tightly balanced so that bone mass is conserved. Despite the importance of the availability of mineral from bone, the major source of calcium remains dietary. The vitamin D-dependent intestinal calcium absorption is therefore essential in order to provide the necessary environment for normal bone growth and development, as well as maintenance of blood calcium levels. In contrast with the bone mineral mobilization promoted by 1,25(OH),D3, 24,25(OH),D3 seems to protect bone from demineralization. This observation has been noted both in vivo [37] and in vitro [38] situations. 24,25(OH),D, treatment has therefore been utilized to inhibit excessive 1,254OH),D,-dependent bone resorption in some cases of human pathology when high levels of 1,25(OH),D3 must be administered [39,40]. The vitamin D metabolites (1,25(OH),D3 and 24,25(OH),D3) are also known to exert an effect upon cartilage [33]. Receptors for both dihydroxylated vitamin D metabolites have been apparently found in this tissue, and they both can stimulate the uptake of 3 5 S 0 , into proteoglycans. In addition, 1,25(OH),D3 is involved in a number of genomic events related to the synthesis of a number of proteins found in bone cells (procollagen [41], osteocalcin [42], osteonectin [43], matrix Gla protein [44]). 4 . 4 . The reproductive stages

Pregnancy and lactation in mammals bring about considerable changes in both the hormonal status and the mineral and skeletal metabolism in mothers. The maternal demands for calcium and phosphorus are highly increased in order to promote normal fetal growth and bone mineralization and to induce a normal milk production. Accordingly, a number of adjustments must take place in maternal calciotropic hormones, including the vitamin D endocrine system, so that the amounts of calcium and phosphorus available are increased without leading to an overall state of depletion that would result in serious bone damage for the mother [45].

279 Large amounts of calcium are necessary for the developing fetus. These demands are particularly high during the last third of pregnancy in both humans and other mammals, where the actual mineralization of skeletal bone occurs in the fetus [46]. Accordingly, the transplacental mineral movements are dramatically increased at this particular period. In order to meet the increased requirements, the intestinal mineral absorption is highly stimulated in the mother, and this even in a vitamin Ddeficient status. In addition, maternal bone turnover is also dramatically increased, resulting in a rapid fragilization of maternal skeleton if adequate amounts of calcium are not available [45]. The circulating levels of 1,25(OH),D, rise progressively towards the end of pregnancy, probably in response to the increased mineral demands. At the same time, an extra-renal synthesis of 1,25(OH),D3 takes place in the fetoplacental unit [4,47,48]. This synthesis is thought to occur in the decidual cells rather than in the placenta itself, although some controversy still remains upon the exact location of this process. Placental tissues have also been shown to contain specific receptors for 1,25(OH),D, [49] as well as the faculty to synthesize the small calbindin (9 kDa) [50]. However, the transplacental calcium transport is independent of the overall maternal vitamin D status (511. The hypothesis of a differential regulation of calcium transport within the feto/placental unit may be in relation with the in situ synthesis of 1,25(OH),D3 or the fetal synthesis of this hormone. Receptors for 1,25(OH),D3 have also been described in the mammary gland [26,27], where a major calcium and phosphorus translocation must take place in order to provide the essential amounts of these minerals for milk synthesis. However, this process does not seem to be under the dependence of vitamin D since, at least in small mammals (rat, mouse) the mineral content of milk is not significantly affected by the vitamin D-deficiency of the mother [52]. During lactation, the losses of calcium and phosphorus through the human milk average approximately 220-340 mg/day for calcium and 110-170 mg/day for phosphorus, assuming that the total volume of milk secreted is between 650 and 1000 mVday [53]. These losses can even increase to 1000 mg/day (calcium) and 500 mg/day (phosphorus) in women producing large amounts of milk [54]. Accordingly, the amounts of calcium and phosphorus needed for milk synthesis in humans can be more important during three months of full breast-feeding than during the entire gestation period. This mineral loss is even more striking in small mammals, where the number of pups/litter can be high and there is a total dependence upon maternal milk during the first period of life. The daily loss of calcium in the milk of the lactating rat is approximately 30-times greater than its daily urinary excretion and during the 21 days of the lactation period, the amount of calcium transferred from the mother to its offspring through the milk represents 60% of its total skeletal calcium [46]. The hormonal adjustments involved in the regulation of mineral homeostasis during the reproductive stages are therefore critical and its comprehension is not yet fully achieved.

280

5. Non-classical vitamin D responsive systems In the past few years, due to a number of technological improvements, 1,25(OH),D, receptors were able to be identified in a very wide range of tissues and cell lines, extending by far the classical limits of the vitamin D actions upon calcium metabolism (see Ref. 2, page 507). In many of these non-classical target tissues, the reason for the presence of 1,25(OH),D, receptors is still under active research. We will describe here the possible action of 1,25(OH),D3 in some of those new target tissues in an effort to display the complexity of the vitamin D endocrine system.

5.1. The pancreas The involvement of vitamin D in endocrine pancreas function was first suggested by Boquist et al. [55], who described 1,25(0H),D3-induced morphological changes compatible with enhanced Pcell activity. Accordingly, vitamin D deficiency was shown to inhibit insulin but not glucagon secretion both in vivo and in vitro [56-591, whereas the administration of a maintenance dose of vitamin D, in rats of 2-3 weeks clearly increases insulin release from the isolated perfused pancreas, irrespective of the dietary intake and prevailing levels of serum calcium [60]. The existence of high affinity receptors for 1,25(OH),D3 in chick pancreas [61-631, the localization of [3H]1,25(OH)2D3in the nucleus of rat pcells [64] and the presence of pancreasassociated calcium-binding proteins in various species [65-691 suggest that vitamin D or its metabolites may have a direct effect on @cells of pancreas and seem to be essential for the processing of normal insulin secretion. Recently, it has been shown that in human patients with vitamin D deficiency, there is impairment of insulin secretion but not glucagon [70], confirming previous results obtained in animals [56-591. This effect is not due to a decrease in the circulating levels of calcium since in these patients the calcemia was normalized through an exogenous intake. These first results obtained in humans are important and support the hypothesis of a direct effect of 1,25(OH),D3 upon the pancreatic-P cells.

5.2. Reproductive organs

In addition to the organs responsible for the development and maintenance of the fetus and newborn, 1,25(OH),D3 receptors have also been localized in several organs from the reproductive apparatus such as the uterus [71], ovary [72] and testis [73]. Since these tissues are not directly associated with calcium translocations, the presence of 1,25(OH),D3 receptors may be related to a role of the hormone in cellular proliferation, differentiation and/or maturation. Accordingly, the levels of testicular 1,25(OH),D3 receptors have been found to correlate with the meiotic and mitotic development of the spermatogonia [73]. Clearly, more studies are needed in this area to clarify the role of the vitamin D hormone in these tissues. However,

281 the presence of 1,25(OH),D3 receptors within selected zones of these tissues supports the hypothesis of a specific in situ effect of the hormone.

5.3. Neural tissues Receptors for 1,25(OH),D3 have been detected in limited sections of the brain. However, the presumed effects of 1 .25(OH),D3 in brain are not well understood. Recent data have shown that administration of 1,25(OH),D3 to vitamin D-deficient rats leads to an increase in the activity of the choline acetyltransferase (CAT) in specific brain nuclei [76]. The bed nucleus of the stria terminalis and the nucleus centralis of the amygdala are the two regions of the brain in which the highest nuclear concentration of 1,25(OH),D3 has been described [74], and also those reported to have the largest increase in 1,25(0H),D,-stimulated CAT activity [76]. These authors have provided the first evidence of a 1,25(0H),D3-dependent activity in brain and therefore suggest that this hormone, like other steroid hormones, may selectively affect the metabolism of a specific neuronal population. The presence of vitamin D-dependent calcium-binding proteins (calbindins) in brain has been firmly established. However, no vitamin D-dependence has been detected for either the CaBP,,, nor the CaBPgKin brain [76,77]. 5.4. Contractile tissues

5.4.1. Skeletal muscle The presence of a muscle weakness or myopathy during metabolic bone diseases was mentioned in one of the first known reports of rickets [78]. This fact has been thereafter well documented using clinical and electromyographical as well as histological approaches [79-861, and they have emphasized the satisfactory results obtained during vitamin D therapy. These findings have been strengthened by the discovery of a 1,25(OH),D3 receptor in skeletal muscle myoblasts [87,88], as well as by studies showing evidence for a 1,25(OH),D,-dependent mechanism affecting muscle calcium metabolism and muscle contraction [71-741. Furthermore, there is also some evidence tending to link the action of vitamin D metabolites upon muscle calcium fluxes and the maintenance of calcium homeostasis in the whole organism. During vitamin D depletion, calcium tends to accumulate in muscle tissue, a quick release occurring after a single administration of vitamin D . Apparently, this seems to be directly related with the rapid increase in blood calcium levels [92]. However, this effect could also be mediated, at least partially, by PTH [93]. In addition, vitamin D seems also to be involved in the regulation of phosphate fluxes across the muscle membranes. In this case, 25(OH)D3seems to be the active metabolite, both in vivo and in vitro [94,95].

5.4.2. Cardiac muscle Calcium plays a major role in the function of cardiac muscle [96,97] although the exact mechanisms that mediate the calcium movements in heart remain to be fully elucidated. On the other hand, the absolute requirement for calcium of the cardiac muscle cells, together with their marked sensitivity to an excess of circulating calcium levels, indicate their need for a perfect calcium balance. The presence of 1,25(OH),D, receptors showing selective localization in a few cardiac muscle nuclei [98] call for an important specific action of the hormone within this very special calcium-dependent organ which may prove to be important in pathological conditions affecting heart function. The presence of at least one vitamin D-dependent calcium binding protein in heart [99] suggests that the 1,25(OH),D, receptors are functional and considerably increases its importance. In vitro studies using rat ventricular cardiac muscle cells have recently been performed confirming that 1,25(OH),D, does have a direct affect on these cells by stimulating their 45Ca2+uptake and addressing the possibility of a vitamin D-dependent regulation of intracellular calcium homeostasis in heart cells [loo]. Further studies in this area promise to be exciting.

6. Vitamin D and the immune system In the last few years, considerable evidence has accumulated linking 1,25(OH),D, to the hematopoietic system and possibly to the immune response. Non-classical target tissues for 1,25(OH),D,, determined by the presence of 1,25(OH),D3 receptors include thymus and bone marrow, as well as cells derived from these tissues [101-1041. Early studies using a leukemic cell line (HL-60) provided the first line of evidence of the positive effects of 1,25(OH),D3 upon the hematopoietic system. In fact, these cells differentiate into macrophage-like cells when submitted to 1,25(OH),D, treatment [lo51 and this finding led to extensive studies of the effects of vitamin D metabolites in different types of cells of hematopoietic origin. In addition, macrophages obtained from different tissues can synthesize 1,25(OH),D, [106], a phenomenon previously thought to occur only in kidney cells and the fetoiplacental unit during pregnancy [4,47,48] (see Section 4). Furthermore, pinterferon (yIFN),which is normally produced by activated T-lymphocytes, was found to highly stimulate the 1,25(OH),D, production by macrophages [106]. 1,25(OH),D3 being a powerful mediator of the increased calcium levels in the blood stream, these findings were then related to the existence, in some patients, of the so-called hypercalcemia of malignancy, particularly in sarcoidosis since these individuals are known to possess activated T-lymphocytes which secrete large amounts of yIFN. Supporting this hypothesis, high levels of 1,25(OH)2D3were detected in some cases of sarcoidosis, once the patient had undergone nephrectomy, thus suppressing the renal synthesis of the hormone [107,108]. At this point, the data available strongly support the idea that macrophages may be a normal physiological source of

283

Lymphoid precursor Lymphoblast

Pronormoblast Lymphocyte*

Prornyelocyte

.(

Normoblast

1 **

Monocyte

b

Retlculocyte

MARROW ---Erythrocyte B L OOD - ------

Macrophage**

TISSUE

Cells which h a v e been shown t o

* P o a r e s s 1.25(OH)20 r e c e p t o r 8

Q Produce 1,25(OHl2D3

Fig. 4. Schematic representation of the vitamin D-micro-endocrine system (stem cell differentiation) showing the localization of 1,25(OH)2D3receptors as well as cells which have been shown to synthesize 1,25(OH)D3.

1,25(OH),D3, thus providing in situ high levels of the hormone which could be responsible for (1) normal local bone resorption and (2) playing a role in stimulating the differentiation of monocytes along the macrophage pathway (Fig. 4). Furthermore, these findings support the hypothesis considering the existence of a local paracrine system for vitamin D, which is depicted in Fig. 5. Another line of evidence supporting the linkage between vitamin D and the immune system is derived from the fact that 1,25(OH),D, can suppress immunoglobulin production by activated

284 ANTIGEN

T-LY MPHOCYTES CFU-GM.

GRANULOCY

\

I L"

OSTEOCLASTS

PROLIFERATION

W

Fig. 5. Schematic representation of a possible vitamin D paracrine system in cells from hematopoietic lineage.

human peripheral blood mononuclear cells [ 1091. Furthermore, 1,25(OH),D, blocks phytohemagglutinin- and antigen-induced lymphocyte blast transformation, possibly through the attenuation of interleukin 2 production [110-1121. In conclusion, the relationship between vitamin D , through its hormonally active form 1,25(OH),D3 and the hematopoietic system and consequently the immune system has been unequivocally established. However, much work remains to be done in order to establish the exact role played by 1,25(OH),D, in vivo and its relevance to the treatment of related pathological states.

7. Clinical disorders related to vitamin D An increasing number of pathological disorders can be related either directly or indirectly to a wide type of malfunctions affecting one or several of the vitamin D metabolic pathways (Fig. 6). These disorders can be roughly related to four different situations: (a) decreased or lack of availability of vitamin D , (b) defective conversion of vitamin D into its major hydroxylated derivatives, (c) altered end-organ responsiveness to the vitamin D metabolites and (d) abnormal interactions between the vitamin D metabolites and other hormones, leading to impaired biological response. Because of the complexity of the endocrine and paracrine systems of vitamin D , it is our purpose in this brief review only to outline the different pathologic states resulting from a deficient or altered vitamin D metabolism.

285

Osteitis fibrosa cystica Osteomalacia Osteoporosis Osteopenia

BLOOD D3 25(OH)D3 lq25(OH)2&, 24R,25(OH&D3

Malabsorption syndrome Sarcoidosis Steatorrhea

Vitamin D deficiency remains the most common cause of rickets and osteomalacia in the world, with the exception of the United States and the Scandinavian countries where most dairy products are supplemented with this vitamin. This deficiency can be caused either by dietary habits or by insufficient exposure to ultraviolet light. The same type of symptoms can be observed when there is interruption of the normal vitamin D metabolic pathways due to a number of liver and/or kidney diseases. In addition, a number of inherited factors can lead to different types of vitamin D resistance which require massive supplements of vitamin D and/or minerals. Extensive reviews have been published depicting both the clinical features and their most likely causes, as well as the possible treatments of the different types of clinical disorders resulting from vitamin D deficiencies [ 113-1 171. The newly defined role for 1,25(OH),D3 upon the hematopoietic system could also have clinical relevance in bone disorders such as osteoporosis where patients have been shown to possess abnormal T-cell subsets [ 1181.

8. Summary Vitamin D is now clearly considered to be a pro-hormone. Its principal hormonally active derivative, 1,25(OH),D,, acts at the genomic level following the same path-

286 way previously described for other more classical steroid hormones. In addition, there is clear evidence for 1,25(OH)*D3being involved in a different type of receptor-mediated action not requiring genomic activation. Furthermore, both in vivo and in vitro data suggest that this pluripotent seco-steroid hormone is not only a major regulator of the mineral metabolism and calcium homeostasis, but is also involved in a number of other biological activities not yet fully understood but clearly related to cell proliferation and differentiation. The number of tissues and cell types possessing receptors for vitamin D derivatives has been greatly increased in the last few years. New target tissues including the heart, pancreas, ovary, testis, lung, as well as a possible involvement within the immunological response, bring forward a whole new aspect of the vitamin D endocrine and paracrine system which promise to lead to new and very exciting findings in the near future.

References 1. Norman. A.W., Roth, J . and Orci. L. (1982) Endocr. Rev. 3. 331-366. 2. Henry. H.L. and Norman, A . W . (1984) Ann. Rev. Nutr. 4. 493-520. 3 . Reichel. H., Bishop, J . E . , Koeftler. H.P. and Norman, A.W. (1987) J . Clin. Endocrinol. Metab. 64, 1-9. 4. Weisman. Y.. Harell. A., Edelstein, S . . David. M.. Spirer, Z . and Golander. A . (1979) Nature 281. 3 17-3 19. 5. Norman, A.W. (1983) In: Steroid Hormone Receptors: Structure and Function (Eriksson, H. and Gustafsson. J.-A., eds.) pp. 479-493. Elsevier Science Publishers, Amsterdam. 6. Haussler. M . R . (1986) Ann. Rev. Nutr. 6. 527-562. 7. Wilhelm. F. and Norman. A.W. (1985) J . B i d . Chem. 260, 10087-10092. 8. Walters. M.R. (1985) Endocr. Rev. 6. 512-543. 9. Corradino. R . A . and Wasserman. R . H . (1968) Arch. Biochem. Biophys. 126, 957-959. 10. Perret, C.. Desplan, C . , Brehier. A . and Thomasset, M. (1985) Eur. J. Biochem. 148, 61-66. 11. Tsai. H.C. and Norman, A.W. (1973) Biochem. Biophys. Res. Cominun. 54. 622-627. 12. Siebert, P . D . , Hunziker. W. and Norman, A.W. (1982) Arch. Biochem. Biophys. 219. 286-296. 13. Hunziker, W.. Walters, M.R.. Bishop. J.E. and Norman. A.W. (1982) J . Clin. Invest. 69,826434. 14. Hunziker, W.. Siebert. P.D.. King, M . W . , Stucki, P., Dugaiczyk, A . and Norman. A.W. (1983) Proc. Natl. Acad. Sci. U.S.A. 80, 422X-4232. 15. King. M.W. and Norman, A.W. (1986) Arch. Biochem. Biophys. 248,612-619. 16. Theofan, G . , Nguyen, A.P. and Norman. A.W. (1986) J . Biol. Chem. 261, 16943-16947. 17. Theofan, G . , King, M.W., Hall, A.K. and Norman. A.W. (1987) Mol. Cell. Endocrinol. 54, 135-140. 18. Theofan, G. and Norman, A . W . (1986) J . B i d . Chem. 261. 7311-7315. 19. Nemere, I. and Norman, A.W. (1986) Endocrinology 119. 140&1408. 20. Nemere. I. and Norman. A.W. (1987) J . Bone Mineral Res. 2. 99-107. 21. Nemere, I . . Leathers. V.L. and Norman, A.W. (1986) J . Biol. Cheni. 261. 16106-16114. 22. Putkey. J . and Norman. A.W. (1982) In: The Role of Calcium in Biological Systems. Vol. I1 (Anghileri. L.J. and Tuffet-Anghileri. A.M.. eds.) pp. 171-202, C R C Press, Boca Raton. FL. 23. Bordier. P.. Rasinussen. H . , Marie. P.J.. Miravet. L.. Gueris, J . and Ryckwaert, A . (1978) J . Clin. Endocrinol. Metab. 46, 284-294. 24. Pike, J.W., Gooze. L.L. and Haussler, M.R. (1980) Life Sci. 26. 407-414. 25. Kream, B.E., Jose, M . , Yamada. S. and DeLuca. H . F . (1977) Science 197. 1086-1(388.

287 26. Colston, K., Hirst, M. and Feldman. D . (1980) Endocrinology 107. 1916-1922. 27. Eisman, J.A., MacIntyre. I.. Martin, T.J.. Frampton, R.J. and King, R.J.B. (1980) Clin. Endocrinol. (Oxford) 13, 267-272. 28. Habener, J.F., Rosenblatt, M. and Potts. J.T., Jr. (1984) Physiol. Rev. 64, 985-1053. 29. Kawashima, H . , Torikai, S. and Kurokawa. K. (1981) Nature 291, 327-329. 30. Jaeger, P., Jones, W., Clemens. T.L. and Hayslett. J.P. (1986) J . Clin. Invest. 78, 456461. 31. Spanos. E . , Colston, K.W. and Evans. I.M.A. (1976) Mol. Cell. Endocrinol. 5, 163-169. 32. Kurnik. B.R.C., Huskey, M. and Hruska. K.A. (1987) Biocbem. Biophys. Acta 917, 81-85. 33. Garabedian, M. (1984) In: La vitamine D et les Maladies des 0 s et du Metabolisme Mineral (Fournier, A , , Garabedian, M., Sebert, J.L. and Meunier. P.. eds.) pp. 21-40. Masson, Paris. 34. Rhoten, W.B., Bruns, M.E. and Christakos, S. (1985) Endocrinology 117. 674683. 35. Lee, D.B.N., Walling, M.W. and Brauthar. N . (1986) Am. J . Physiol. 250, G369-373. 36. Marie, P.J. and Travers. R . (1983) Calcif. Tissue Int. 35, 418-425. 37. Pavlovitch, J . H . , Cournot-Witmcr. G . . Bourdeau. A , . Balsan. S., Fischer. J . A . and Heynen. G . (1981) J. Clin. Invest. 68. 803-810. 38. Sebert, J.L., Fournier, A . , Lambrey, G . , ct al. (1982) Nephrologic 3. 133-143. 39. Evans. R.A.. Hills, E.. Wong, S.Y.P.. t't al. (1982) In: Vitamin D: Chemical, Biochemical and Clinical Endocrinology of Calcium Metabolism (Norman, A.W.. ed.) pp. 835-840. Walter de Gruyter, Berlin. 40. Mahgoub, A . (1981) Calcif. Tissue I n t . 33. 663-666. 41. Rowe, D . W . and Kream, B.E. (1982) J . Biol. Chem. 257, 8009-8015. 42. Price, P.A. and Baukol, S.A. (1980) J . Biol. Chem. 255, 11660-11663. 43. Young, M.F.. Bolander. M.E.. Day. A.A., Ramis. C.I., Robey, P.G., Yamada, Y . and Termine, J.D. (1986) Nucl. Acid Res. 14, 4483-4497. 44. Fraser, J.D. and Price, P.A. (1987) Ninth Annual Scientific Meeting of the American Society for Bone and Mineral Research. J. Bone Min. Res. 2, abstr. 21. 45. Marie, P.J., Cancela, L., Leboulch, N . and Miravet, L. (1986) Am. J . Physiol. 251, E400-406. 46. Garel, J.M. (1987) Physiol. Rev. 67. 1-66. 47. Gray, T.K., Lester, G.E. and Lorenc, R.S. (1979) Science 204, 1311-1313. 48. Delvin, E.E.. Arabian. A . , Glorieux, F.H. and Mamer, O.A. (1985) J. Clin. Endocrinol. Metab. 60, 880-885. 49. Delvin, E.E., Pilon, A.M. and Vekemans. M. (1987) Pediatric Res. 21. 432-435. 50. MacManus. J.P., Watson, D.C. and Yaguchi. M . (1986) Biochem. J. 235. 585-595. 51. Brommage. R. and DeLuca. H.F. (1984) Am. J . Physiol. 246, F526529. 52. Toverud, S.U. (1963) In: The Transport of Calcium and Strontium Across Biological Membranes (Wasserman, R . H . , ed.) p. 341. Academic Press, New York. 53. Greer, F . R . , Tsang. R.C., Searcy, J.E.. Levin. R.S. and Steichen, J.J. (1982) Am. J . Clin. Nutr. 36, 431-437. 54. Hunscher, A.A. (1930) J. Biol. Chem. 86, 37-47. 55. Boquist, L., Hagstrom, S. and Strindlund, L. (1977) Acta Pathol. Microbiol. Scand. 85, 485-489. 56. Norman, A . W., Frankel, B.J., Heldt. A.M. and Grodsky, G.M. (1980) Science 209, 82S825. 57. Kadowaki, S. and Norman, A.W. (1984) J . Clin. Invest. 73. 759-766. 58. Tanaka, Y., Seino, Y., Ishida, M., Yamaoka, K.. Yabuuchi, H., Ishida, H., Seino, S., Seino, Y . and Imura, H . (1984) Acta Endocrinol. (Copenhagen) 105, 528-533. 59. Chertow, B.S., Sivitz, W.I., Baranetsky, N.G., Clark, S.A., Waite. A . and DeLuca, H.F. (1983) Endocrinology 113, 1511-1518. 60. Cade, C . and Norman, A.W. (1986) Endocrinology 119, 8 4 9 0 . 61. Christakos, S . and Norman, A.W. (1979) Biochem. Biophys. Res. Commun. 89, 5 6 6 3 . 62. Christakos, S. and Norman, A.W. (1981) Endocrinology 108, 14C149. 63. Pike, J.W. (1981) J. Steroid Biochem. 16, 385-395.

64. Clark, S.A., Stumpf, W.E., Sar, M., DeLuca, H . F . and Tanaka, Y. (1980) Cell Tissue Res. 209, 515-520. 65. Christakos, S., Friedlander, E.J.. Frandsen, B.R. and Norman, A.W. (1979) Endocrinology 104, 1495-1503. 66. Morrissey. R.L., Bucci, T.J., Empson, R.N. and Lufkin, E . G . (1975) Proc. Soc. Exp. Biol. Med. 149, 5&60. 67. Arnold, B.M., Kuttner, M., Willis, D.M., Hitchman. A.J., Harrison, J.E. and Murray, T.M. (1975) Can. J. Physiol. Pharmacol. 53, 1135-1140. 68. Kadowaki, S. and Norman, A.W. (1984) Arch. Biochem. Biophys. 233, 228-236. 69. Roth, J., Bonner-Weir, S., Norman, A.W. and Orci, L. (1982) Endocrinology 111, 22162218. 70. Gedik, 0. and Akalin, S. (1986) Diabetologia 29, 142-145. 71. Walters, M.R. (1981) Biochem. Biophys. Res. Commun. 103, 721-726. 72. Dokoh. S., Donaldson, C.A., Marion, S.L., Pike, J.W. and Haussler, M.R. (1983) Endocrinology 112, 20C-206. 73. Walters, M.R. (1984) Endocrinology 114, 2167-2174. 74. Stumpf. W.E., Sar, M., Clark, S.A. and DeLuca, H.F. (1982) Science 215, 1403-1405. 75. Clemens, T.L., Zhou, X.Y., Pike, J.W., Haussler, M.R. and Slouiter, R.S. (1985) In: Vitamin D. Chemical, Biochemical and Clinical Update (Norman, A.W., Schaefer, K., Grigoleit, H . G . and Herrath, D.V., eds.) pp. 95-96. DeGruyter. BerliniNew York. 76. Sonnenberg, J., Luine, V.N., Krey, L.C. and Christakos, S. Endocrinology 118, 1433-1439. 77. Thomasset, M . , Parkes, C.O. and Cuisinier-Gleizes. P. (1982) Am. J. Physiol. 243. E483-488. 78. Glisson, F. (1660) De Rachitide. Sadler, London. 79. Scott, A.C. (1916-1917) Indian J . Med. Res. Calcutta, iv, 14&168. 80. Hagenbach-Burckhardt, E. (1904) J . Kinderheilk., Berlin, LX, 471-487. 81. Peitsara, H . (1944) Acta Paediatr. Scand. 31, 1-244. 82. Prineas, J.W.. Mason. A S . and Henson, R.A. (1965) Br. Med. J. 5441, 1034-1036. 83. Smith, R . and Stern, G . (1967) Brain 90, 593-596. 84. Floyd, M.. Ayyar, D . R . , Barwick, D . D . , Hudgson, P. and Weightman, D . (1974) Q. J. Med. 43, 509-523. 85. Brickman. A.S., Coburn, J.W., Massry, S.G. and Norman, A.W. (1974) Ann. Intern. Med. 80, 161-168. 86. Schott, G . D . and Wills, M.R. (1976) Lancet i , 626-629. 87. Boland, R., Norman, A . W . , Ritz, E. and Hasselbach, W. (1985) Biochem. Biophys. Res. Commun. 128, 305-311. 88. Simpson, R . U . , Thomas, G.A. and Arnold, A.S. (1985) J. Biol. Chem. 260, 8882-8891. 89. Giuliani, D . L . and Boland, R.L. (1984) Calcif. Tissue Int. 36, 20&205. 90. de Boland, A.R. and Boland. R.L. (1985) Biochim. Biophys. Acta 845, 237-241. 91. Rodman, J.S. and Baker, T. (1978) Kidney Int. 13. 189-193. 92. Bauman, V.K., Valinietse, M.Y. and Babarykin. D.A. (1984) Arch. Biochem. Biophys. 231, 2 11-216. 93. Borle, A.B. (1981) Rev. Physiol. Biochem. Pharmacol. 90. 13-153. 94. Birge, S.J. and Haddad, J.G. (1975) J . Clin. Invest. 56, 110C-1107. 95. Bellido. T . and Boland, R.L. (1985) In: Vitamin D. A Chemical, Biochemical and Clinical Update (Norman, A.W., Schaefer, K . , Grigoleit, H.G. and Herrath, D.V., eds.) pp. 590. Walter de Gruyter. Berlin. 96. Winegard, S. (1982) Ann. Rev. Physiol. 44. 451-462. 97. Fleckenstein, A. (1983) Circul. Res. 52, 3-16. 98. Walters, M.R., Wicker. D.C. and Riggle, P.C. (1986) J. Mol. Cell Cardiol. 18, 67-72. 99. King, M.W., Hunziker, W., Siebert, P.W., Williams, G . and Norman, A.W. (1983) Proc. Soc. Bone Min. Res. 5 . A60 (abstract).

289 100. Walters, M.R., Ilenchuk, T.T. and Claycomb. W.C. (1987) J. Biol. Chem. 262, 25362541. 101. Reinhardt, T.A., Horst, R.L., Littledike, E.T. and Beitz, D.C. (1982) Biochem. Biophys. Res. Commun. 106, 1012-1018. 102. Prowedini. D.M., Tsoukas, C.D.. Deftos, L.J. and Manolagas, S.C. (1983) Science 221, 1181-1183. 103. Merke, J . , Senst, S . and Ritz. E . (1984) Biochem. Biophys. Res. Commun. 120, 199-205. 104. Tanaka. H . , Abe. E . , Miyaura, C.. Kuribayashi, T . , Konno, K., Nishii. Y. and Suda, T . (1982) Biochem. J . 204, 713-719. 105. Miyaura, C., Abe, E., Kuribayashi, T . , Tanaka. H., Konno. K., Nishii, Y. and Suda, T. (1981) Biochem. Biophys. Res. Commun. 102, 937-943. 106. Reichel, H., Koeffler, H.P. and Norman, A.W. (1987) J. Biol. Chem. 262, 10931-10937. 107. Barbour, G.L., Coburn, J . W., Slatopolsky, E . , Norman, A.W. and Horst, R.L. (1981) N. Engl. J . Med. 305, 44C-443. 108. Maesaka, J.K., B a t m a n , V.. Pablo, N.C. and Shakamuri, S. (1982)Arch. Int. Med. 142. 1206-1207. 109. Lemire, J.M., Adams, J.S., Sakai, R . and Jordan, S.C. (1984) J . Clin. Invest. 74, 657-661. 110. Lemire, J.M., Adams, J.S.. Kermani-Arab. V.. Bakke, A.C., Sakai, R. and Jordan, S.C. (1985) J . Immunol. 134, 3032-303s. 111. Rigby, W.F.C., Stacy, T. and Fanger, M.W. (1984) J . Clin. Invest. 74, 1451-1455. 112. Bhalla, A.K.. Amento, E . P . , Serog, B. and Glimcher, L.H. (1984) J . Immunol. 133, 1748-1754. 113. Marx, S.J.. Liberman, V.A. and Eil, C . (1983) Vitamins Hormones 40, 235-308. 114. Kumar, R . (1983) Ann. Intern. Med. 98, 662-663. 115. Dunnigan, M.G., McIntosh, W.B.. Ford. J . A . , et al. (1982) In: Calcium Disorders (Heath, D . A . and Marx, S.J. eds.) pp. 125-150. Butterworth Scientific, London. 116. Liberman. V.A., Samuel, R., Halabe, A., Kauli, R.. Edelstein, S . , Weisman. Y . ,Papapoulos, W.E.. Clemens, T.L., Fraher. L.J. and O’Riordan. J . L . (1980) Lancet i, 504-507. 117. Fournier, A . , Sebert, J.L., Boudailliez, B. and Moriniere. P. (1985) Ann. Med. Interne (Fr) 136, 1 6 4179. 118. Fujita. T., Matsui. T., Nakao, Y. and Watanahe, S. (1984) Min. Electrol. Metab. 10. 375-378.

This Page Intentionally Left Blank

29 1

Subject Index Accessory sexual glands, androgens 190 Action of androgens, 171 antioestrogens, 35, 161, 210 glucocorticoids, 217 LHRH, 152 oestrogens, 197 progesterone, 241 steroid hormones, 29, 30, 39, 169, 197, 217, 24 1 thyroid hormones, 61, 66 Adenylate cyclase, thyroid hormones, 70 Aldosterone, biosynthesis, 14 Anabolic steroids, 183 Androgens, accessory sexual glands, 190 anterior pituitary, 188 biological actions, 171 biosynthesis, 15 bone, 190 breast, 188 central nervous system. 186 gene expression, 193 haemopoietic organs, 177 hair, 188 kidney, 179 lymphocytic organs, 190 muscle, 182 salivary glands, 178 skin, 189 testis, 174 urogenital tract, 175 Antagonists to steroid hormones, 35, 156 Anterior pituitary, androgens, 188 Antiandrogens, 160, 192 Antioestrogens, 161, 210 Biosynthesis of aldosterone, 14 androgens, 15 corticosteroids, 11 oestrogens, 20 peptide hormones, 118

steroid hormones, 3 thyroid hormones, 63 vitamin D , 269 Bone, androgens, 190 vitamin D , 277 Brain development, thyroid hormones, 73 Breast, androgens, 188 oestrogens, 60, 153, 208, 242 progesterone, 242 Calcium, vitamin D, 274 Cell proliferation, oestrogen, 207 Central nervous system, androgens, 186 Cholesterol, 4 side chain cleavage, 8 transport, 6 Clinical disorders, vitamin D, 284 Conjugation, thyroid hormones, 84 Contractile tissues, vitamin D , 281 Corticosteroids, biosynthesis, 11 DNA binding of androgens, 193 glucocorticoids, 222, 226, 230 oestrogen, 203 progesterone, 258 steroid hormones, 31, 42, 46

FSH receptors, 105 Gene expression, androgens, 193 oestrogen, 205 glucocorticoids, 229 progesterone, 257 steroid hormones, 39 GH, thyroid hormones, 66 Glial cell differentiation, thyroid hormones, 75 Glucocorticoids, biological actions, 217 DNA binding, 222, 226. 230 gene activation, 229 lymphocytolysis, 217

292 receptor M,, 233 receptor defects, 220 receptors, 221 Haemopoietic organs, androgens, 177 Hair, androgens, 188 Immune system, vitamin D, 282 Internalization, peptide hormones, 129, 134, 140 Intestine, vitamin D , 277 Iodination. peptide hormones, 106 Iodothyronine deiodinases. 93 Kidney, androgens, 179 vitamin D , 276 LHRH, action, 152 analogues, 154 Lipoproteins, 4 Lipogenesis, thyroid hormones, 68 Lymphocytic organs, androgens, 190 Lymphocytolysis, glucocorticoids. 217 Membrane receptors, characterization. 112 isolation, 107 solubilization, 114 specificity, 1 1 1 Mineral hemostasis, vitamin D , 276 Muscle, androgens, 182 thyroid hormones, 72 Neural tissues, vitamin D , 281 Neuronal differentiation, thyroid hormones, 74 Oestrogen, actions, 197 biosynthesis, 20 breast, 60, 153, 208 cell proliferation, 207 D N A binding, 203 gene expression, 205 receptor, 200 receptor genes, 203 Pancreas, vitamin D , 280 Peptide hormones,

biosynthesis, 118 circulation, 128 degradation, 128 identity, I18 internalization, 129. 134 internalization, temperature effects, 140 iodination, 106 prohormones, 121 receptor down regulation. 146 receptors microaggregation, 144 recycling, 136 release, 127 storage, 127 Progesterone, action, 241 DNA binding, 258 gene expression, 257 reccptors, 243 receptor phosphorylation, 254 receptor localization, 255 receptor structure, 245 Prohormones, peptide hormones, 121 Receptor assay. steroid hormones. 50, 53 characterization, steroid hormones, 57 defects, glucocorticoids, 220 down regulation. peptide hormones, 146 glucocorticoids, 221 localization, progesterone, 255 microaggregation, peptide hormones, 144 phosphorylation, progesterone, 254 progesterone, 243 properties, steroid hormones, 52 purification. steroid hormones, 55 recycling, peptide hormones, 136 specificity, 34 steroid hormones, 30 structure. steroid hormones, 31, 39 structure, progesterone, 245 thyroid hormones, 64 vitamin D, 271 Reproduction, vitamin D , 278, 280 Salivary glands, androgens, 178 Skin, androgens, 189 Solubilization, membrane receptors, 114 Specificity, membrane receptors, 11 1 steroid hormones, 32, 36

293 Stereochemistry, testosterone, 173 Steroid hormones, action, 29, 30, 39, 169, 197, 217, 241 antagonists, 36, 156 biosynthesis, 3 DNA binding, 31, 42, 46 gene regulation, 39 receptor structure, 39 receptor characterization, 57 receptor assay, 50, 53 receptor properties, 52 receptor purification, 55 receptors, 30 secretion, 24 specificity, 32 Testis, androgens, 174 Testosterone, active metabolites, 170 stereochemistry, 173 Thyroid hormones, action, 61, 66 adenylate cyclase, 70 biosynthesis, 63 brain development, 73 cofactors for metabolism, 90 conjugation, 84 GH, 66 glial cell differentiation, 75 inhibitors of metabolism, 87

lipogenesis, 68 mechanisms of metabolism, 89 metabolism, 81 muscle, 72 neuronal differentiation, 74 receptors, 64 regulation of metabolism, 99 structures, 62 transport, 63, 97 TSH, 68 TSH, thyroid hormones, 68 Urogenital tract, androgens, 175 Vitamin D, biosynthesis and metabolism, 269 bone, 277 calcium, 274 clinical disorders, 284 contractile tissues, 281 endocrine system, 270 immune system, 282 intestine, 277 kidney, 276 mineral hemostasis, 276 neural tissues, 281 pancreas, 280 receptor, 271 reproduction, 278, 280

This Page Intentionally Left Blank