The Effects of Recombinant Human Growth Hormone on Skin, Muscle

Home , Somatrem

2807713108

THE EFFECTS OF RECOMBINANT

HUMAN GROWTH HORMONE ON SKIN,

MUSCLE AND NERVE.

Thesis submitted to the University of London for the degree of

Doctor of Philosophy in the Faculty of Medicine.

WOHAIB HASAN MSc BSc (Hons)

Division of Endocrinology and Department of Anatomy, Royal

Free Hospital School of Medicine

February 1997 ProQuest Number: 10046192

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest.

ProQuest 10046192

Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author.

ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 a Abstract

This study has investigated the mechanism of growth hormone (GH) action in the skin and its innervation as well as in skeletal muscle. Two main issues have been addressed: (1) To what extent are GH effects on extrahepatic targets the result of direct actions of this hormone, or, indirect effects mediated by local production of insulin-like growth factor I (IGF-I)? (2) Can GH, acting either directly or indirectly, be regarded as a key mediator of neurotrophic functions? These issues have been examined using three approaches: i) Muscle biopsies were taken from adult growth hormone-deficient patients before and after six to twelve months of recombinant human growth hormone (r-hGH) treatment, as part of a double-blind placebo controlled clinical trial. These patients exhibit reductions in muscle bulk and strength which can be reversed with r-hGH treatment. Muscle fibre types and fibre size were examined by light microscopy for an influence of r-hGH therapy. IGF-I message was localised to muscle fibre cytoplasm and fibre- type specific mRNA was quantified by combined non-radioactive in situ hybridisation with computerised image analysis. Circulating IGF-I levels were also assayed, making it possible to compare morphological changes in muscle fibre size with changes in circulating and muscle-derived IGF-I. The results obtained are consistent with a direct effect of GH on skeletal muscle resulting in generation of IGF-I by the muscle fibres. However, the influence of circulating IGF-I could not be dismissed from this scheme. ii) The innervation, morphology and function of eccrine sweat glands was examined in GH-deficient subjects who display reduced sweating, in patients with acromegaly (a GH hypersecretory state with excessive sweating) and in control subjects. Pilocarpine iontophoresis sweat tests were used to assess changes in sweating. An increase in the cholinergic innervation accompanied restoration of sweat rates in r-hGH treated subjects, whilst acromegalics displayed hypertrophy of sweat glands and elevated sweat gland innervation. These data imply an effect of GH on the sweat glands with repercussions for their innervation. A GH-mediated influence on neurotrophic support to the sweat gland nerves can thus be envisaged. iii) Aged rats were utilised to investigate direct effects of r-hGH administration in the vicinity of the footpad eccrine sweat glands, with the aim of minimising the confounding influence of serum IGF-I. Ageing is a state of relative GH deficiency and sweat glands are believed to be targets for GH action. Further aged rats display reductions in sweat gland size and innervation. An immunohistochemical method was utilised to stain for nerves and neurotrophin receptor sites around sweat glands. Due to injection-induced inflammatory effects, the results from this study were not clear, aside fi-om an increase in sweat gland acinar size on r-hGH administration and the demonstration of parallel alterations in neurotrophin receptor expression and periacinar nerve density. These results allow the hypothesis of a direct effect of GH on target tissues to be advanced, resulting in local generation of IGF-I and, perhaps, of other target-derived neurotrophic factors. Acknowledgements

In the name o f Allah, the Creator and Sustainer o f the Worlds "Over all those endowed with knowledge is the All-Knowing." "We will show them Our Signs in the universe, and in their own selves, until it becomes manifest to them that this is the Truth." (Al-Qur'an 12:76 & 41:53)

I am indebted and extremely grateful to Dr Tim Cowen whose guidance and support were essential throughout this study. Thanks are also due to Dr Pierre Bouloux for his joint supervision and financial support without which this thesis would not have been possible. I acknowledge the help of Dr Philip Barnett and Mr Christopher Thrasivoulou throughout the past few years. I thank "them upstairs" (Patsy Coskeran, Richard Quinton, Priyal D'Zoysa and Véronique Duke) and "them downstairs" (Isabella Gavazzi, Agus Santoso, Marion Leibl, Richard Johnson, Clare Underwood and David Kelly) for their help and intellectual discourse, and for keeping me sane! Dr Shiyu Yang and Mr Tjeu Gysbers assisted with in situ hybridisation and photography respectively, and their expertise is gratefully appreciated.

This thesis is dedicated to my parents Dr Suhaib Hasan and Mrs Shakila Khatoon whose constant encouragement has been my inspiration. I thank Api, Liaquat bhai, Hamdi, Shazra, Usama, Salma, Hafsa, Shahar and Mujahid for being the best family ever. The kindness and support of my parents-in-law, Mr & Mrs Bashir Butt, is duly acknowledged. In addition, I express my gratitude to Naeem, Saeed, Abdul Haq, Abdul Malik, and all the members of the Royal Free Islamic Society. Finally, a huge thanks to my wife Rizwana and my daughters Sana and Safana for putting up with a husband and father whose fleeting appearances were so rarely accompanied by a smile! Contents

Page Title 1 Abstract 2 Acknowledgements 4 Contents 5 List of figures and legends 12 Tables and Charts 12 Figures 12 Plates 18 List of abbreviations 19

Chapter 1-General Introduction 22 AI 1. Growth hormone; a historical perspective 23 AI 2 Adult growth hormone deficiency syndrome 23 AI.3. Acromegaly 24 AI.4. Growth hormone; structure, binding proteins and receptor 25 AI.5. Growth hormone; secretion and control 27 AI.6. Insulin-like growth factor I; structure, secretion, function and control 30 AI.7. Growth hormone, insulin-like growth factor I and the nervous system 33 AI.7.1. Growth hormone and the nervous system 33 AI.7.2. Insulin-like growth factor I in neuronal development 35 AI.7.3. Insulin-like growth factor I and the mature nervous system 36

AI . 8 . Aims of the investigation 38

Chapter 2-An investigation into recombinant human growth hormone treatment of growth hormone-deficient adults. 40 Chapter 2-Tntroduction BI. 1. Diagnosis of adult growth hormone deficiency 41 BI.2. Growth hormone, insulin-like growth factor I and thyroid hormones 41 BI 3 Growth hormone, insulin-like growth factor I and bone metabolism 42 BI 4. Aims of the investigation 42 Methods BM. 1. Growth hormone-deficient patients 44 BM.2. Administration of recombinant human growth hormone 44 BM.3. Serum insulin-like growth factor I assay 45 BM.4. Serum free 3,5,3'-triiodothyronine and free thyroxine assay 45 BM.5. Urinary total pyridinoline crosslinks assay 45 BM 5 1. Collagen pyridinium as a marker of bone metabolism 45 BM.5.2. Methodology 45 BM 6 Acknowledgments 46 Results BR. 1. Patient responses to recombinant human growth hormone therapy 47 BR.2. Serum insulin-like growth factor I 47 BR.3. Thyroid hormones 55 BR.3.1. Serum free 3,5,3'-triiodothyronine 55 BR.3.2. Serum free thyroxine 56 BR.4. Urinary total pyridinoline crosslinks 65 Discussion BD. 1. Patient responses to recombinant human growth hormone therapy 70 BD.2. Serum insulin-like growth factor I 70 BD 3 Serum free 3,5,3'-triiodothyronine and free thyroxine 72 BD.4. Urinary total pyridinoline crosslinks 74 BD 5 Synopsis of the investigation 76 Chapter 3-An investigation into the thermoregulatory apparatus in adult growth hormone deficiency and acromegaly. 77 Introduction Cl. The regulation of body temperature 78 CI.l. Central control of thermoregulation 78 CL2. Sweat glands; eccrine and apocrine 78 CL 3. Eccrine sweat gland innervation, neurotransmitters and functional tests 80 CI.4. Developmental changes in the innervation to eccrine sweat glands 84 CL 5. Neuropeptides in skin 85 CL5.1. Vasoactive intestinal polypeptide; distribution and function in the skin 86 CL5.2. Other neuropeptides in the skin 89 CL6. Abnormalities of sudomotor innervation in clinical conditions 90 CL 7. Growth hormone and thermoregulation 90 CL 8 . Growth hormone and the skin 91 CL 9. Aims of the investigation 92 Methods CM 1. Patient characteristics 95 CM.2. Pilocarpine iontophoresis-induced sweat production 95 CM. 3. Sweat gland morphology and innervation 95 CM 3.1. Skin biopsy and tissue processing 95 CM.3.2. Acetylcholinesterase histochemistry 96 CM 3 3 Immunohistochemistry 96 CM 3 4 Microscopy and image analysis for skin innervation 97 CM.3.4.1. Microscopy and image input 97 CM.3.4.2. Sweat gland size and innervation measurements 98 CM.3.4.3. Epidermal-dermal junction innervation measurements 98 Results CR. 1. Pilocarpine iontophoresis-induced sweat production 99 CR.2. Acetylcholinesterase histochemistry 104 CR 3 Vasoactive intestinal polypeptide immunoreactive nerve density 110 Chapter 3-Results (contd.) CR 4 Protein gene product 9.5 immunoreactive nerve density 114 CR 5 Protein gene product 9.5 immunoreactive nerves in the epidermal-dermal junction 118 CR 6 Eccrine sweat gland acinus perimeter 121 CR 7 Tyrosine hydroxylase distribution 124 Discussion CD. 1. Growth hormone deficiency; sweating and skin innervation 125 CD.2. Acromegaly; sweating, sweat gland size and innervation 127 CD.3. Gender differences; sweating, sweat gland size and innervation 127 CD.4. Mechanisms; growth hormone deficiency v acromegaly 129 CD.5. Mechanisms; central v peripheral 129

Chapter 4-An investigation into the effects of recombinant human growth hormone administration into footpad skin of ageing rats. 132 Introduction DI. 1. Ageing and sweating/thermoregulation 133 DI.2. Growth hormone and ageing; the 'somatopause' 134 DI.3. Nerve growth factor, sweat gland innervation and growth hormone 137 DI.3.1. Nerve growth factor and its receptors 138 DI.3.2. Nerve growth factor and ageing 138 DI.3.3. Nerve growth factor and sweat gland innervation 139 DI.3.4. Nerve growth factor, growth hormone and insulin-like growth factor I 139 DI.4. Aims of the investigation 141 Methods DM. 1. Animals and injection regime 142 DM.2. Immunohistochemistry 142 Results DR. 1. The effects of recombinant human growth hormone administration into ageing rat footpad skin. 144 DR. 1.1. Sweat gland acinar size 144 Chapter 4-Results (contd.) DR. 1.2. Sweat gland innervation 146 DR. 1.2.1. Protein gene product immunohistochemistry 146 DR. 1.2.2. Nerve growth factor receptor (p75) immunohistochemistry 148 Discussion 152

Chapter 5- An investigation into skeletal muscle in growth hormone- deficient subjects, before and after recombinant human growth hormone therapy. 157 Introduction ELI. Muscle; function and composition 158 EI.2. Growth hormone and muscle in human studies 160 EL 3. Growth hormone, insulin-like growth factor I and muscle innervation 162 EI.4. Growth hormone, insulin-like growth factor I and muscle in animal studies 162 EL 5. Growth hormone, insulin-like growth factor I and muscle fibre types and myosin expression 163 EL 6. Aims of the investigation 164 Methods EM. 1. Introduction 166 EM.2. Biopsy and muscle processing 166 EM. 3 Immunohistochemistry for myosin heavy chains 167 EM.4. In situ hybridisation with the digoxigenin system 168 EM.4.1. Introduction 168 EM.4.2. Probe characteristics 169 EM.4.3. Probe labelling 169 EM.4.4. In situ hybridisation protocol 170 EM 4.4.1. Pretreatment of sections 170 EM.4.4.2. Hybridisation 170 EM.4.4.3. Post-hybridisation washes 171 EM.4.4.4. Immunological detection 171 Chapter 5-Methods (contd.) EM. 5. Imaging methods and quantitation 171 EM.6. Reverse-transcriptase polymerase chain reaction methodology 172 Results ER. 1. Muscle fibre proportions 173 ER.2. Muscle fibre area 175 ER.3. In situ hybridisation 180 ER 3 1. In situ hybridisation signal localisation and distribution 180 ER.3.2. In situ hybridisation signal data expression, distribution and statistical analysis 180 ER. 3 .2.1. Data expression 180 ER.3.2.2. Distribution of the data 183 ER.3.2.3. Results of statistical analysis 183 ER.3.2.3.1. Length of sample storage analysis 183 ER.3.2.3.2. Trial group differences 187 ER.3.2.3.3. Fibre type differences 191 ER.3.2.3.4. Gender differences 191 ER.3.2.3.5. Interactions between variables 191 ER 3.2.3.6. Other parameters 191 ER 3.2.37. Results summary 192 ER.4. Reverse transcriptase polymerase chain reaction results and comparison with in situ hybridisation signal 198 ER.4.1. Patient A 198 ER.4.2. Patient B 198 ER.4.3. Patient C 198 Discussion ED. 1. Muscle fibre proportions 203 ED.2. Muscle fibre size 207 ED 3. Insulin-like growth factor I mRNA expression in muscle 209 ED 3 1. In situ hybridisation 209 ED .3.2. Reverse-transcriptase polymerase chain reaction 211 ED.3.3. Function of muscle insulin-like growth factor I 211

10 Chapter 5-Discussion (contd.) ED.4. Conclusion 213

Chapter 6-General Discussion 215 F. 1. Introduction 216 F.2. Direct or indirect? Growth hormone and insulin-like growth factor I 216 F.2.1. Direct or indirect? Growth hormone, insulin-like growth factor I and bone 217 F.2.2. Direct or indirect? Growth hormone, insulin-like growth factor I and muscle 218 F.3. Growth hormone and the nervous system; a re-examination 221 F.3 .1. Neurotrophins and growth hormone 221 F.3 .2. The influence of growth hormone on its regulatory neurons 222 F.4. Growth hormone, insulin-like growth factor I and the sudomotor apparatus 223 F.5. Gender, thyroid hormones and growth hormone interactions 227 F.6. Conclusion 230

Appendix G. 1. Data presentation and statistical methods 232 G 2 Assays for serum insulin-like growth factor I, free tri iodothyronine, free thyroxine, and urinary total pyridinoline crosslinks 234 G 3 Gelatin coating of slides and acetylcholinesterase methodology 236 G.3 .1. Gelatin-chrom alum coating of slides 236 G 3 2 Acetylcholinesterase methodology 236 G.4. In situ hybridisation procedures 237 G.4.1. Silane coating of slides 237 G.4.2. Probe labelling for in situ hybridisation 237 G.4.3. Reagents and chemicals 241

Bibliography 244

11 List of figures and legends

TABLES and CHARTS Page Table BR. 1. Individual patient characteristics, clinical history and trial groups. 49 Table CL 1. Methods for evaluating sweat gland function and innervation. 83 Table DI. 1. Changes in circulating growth hormone and insulin-like growth factor I levels with age in humans. 136 Chart ER. 3a: Flow chart demonstrating the two methods adopted for data expression. 182 Table ER. 3a: Results from an ordinary least squares regression for length of sample storage. 186 Table ER.4a: In situ hybridisation signal for patient A. 201 Table ER.4b: In situ hybridisation signal for patients B and C. 201 Table ED. 1 : Summary table for investigations into the percentage of type I fibres in GH-deficient subjects. 206

FIGURES-CHAPTER 2 Figure BR.2a: Graph showing data distribution for serum insulin-like growth factor I. 52 Figure BR.2b: Graph showing data distribution for serum insulin-like growth factor I following correction. 52 Figure BR.2c: Histogram showing serum insulin-like growth factor I levels. 53 Figure BR.2d: Histogram showing gender differences for serum insulin -like growth factor I levels. 53 Figure BR.2e: Histogram showing age differences for serum insulin-like growth factor I levels. 54 Figure BR. 3. la: Graph showing data distribution for triiodothyronine. 58 Figure BR.3. lb: Graph showing data distribution for triiodothyronine following correction. 58 Figure BR.3.Ic: Histogram showing triiodothyronine concentration. 59

12 Figures-Chapter 2 (contd.) Figure BR. 3. Id. Histogram showing gender differences for triiodothyronine concentration. 59 Figure BR.3.le. Histogram showing age differences for triiodothyronine and thyroxine. 60 Figure BR.3.If. Histogram showing triiodothyronine concentration for euthyroid patients and those on thyroxine replacement. 60 Figure BR.3.1g. Histogram showing gender differences for triiodothyronine concentration; patients divided into euthyroid and those receiving thyroxine replacement. 61 Figure BR.3.2a\ Graph showing data distribution for thyroxine. 62 Figure BR.3.2b: Graph showing data distribution for thyroxine following correction. 62 Figure BR.3.2c: Histogram showing thyroxine concentration. 63 Figure BR.3.2d: Histogram showing gender differences for thyroxine concentration. 63 Figure BR.3.2e: Histogram showing thyroxine concentration for euthyroid patients and those on thyroxine replacement. 64 Figure BR.3.2/: Histogram showing gender differences for thyroxine concentration; patients divided into euthyroid and those receiving thyroxine replacement. 64 Figure BR. 4a: Graph showing data distribution for total pyridinoline. 67 Figure BR.4b: Graph showing data distribution for total pyridinoline following correction. 67 Figure BR.4c: Histogram showing total pyridinoline crosslinks. 68 Figure BR. 4d: Histogram showing gender differences for total pyridinoline. 68 Figure BR.4e: Histogram showing gender differences for total pyridinoline expressed as a percentage of the baseline group. 69 Figure BR. 4f: Histogram showing age differences for total pyridinoline. 69

13 FIGURES CHAPTER 3 Figure C R I a: Graph showing data distribution for the initial rate of pilocarpine-induced sweating. 101 Figure CRlh .Histogram showing the initial rate of pilocarpine-induced sweating. 102 Figure CRlc\ Histogram showing the initial rate of sweating (percentage baseline). 102 Figure CR Id. Histogram showing gender differences for the initial rate of sweating. 103 Figure CRle. Histogram showing gender differences for the initial rate of sweating (percentage baseline). 103 Figure CR2a. Graph showing data distribution for acetylcholinesterase histochemical staining. 107 Figure CR2b\ Graph showing data distribution for acetylcholinesterase histochemical staining (square root corrected). 107 Figure CR2c: Histogram showing acetylcholinesterase histochemical staining. 108 Figure CR 2d. Histogram showing acetylcholinesterase histochemical staining (percentage baseline). 108 Figure CR2e\ Histogram showing gender differences for acetylcholinesterase histochemical staining. 109 Figure CRSa: Graph showing data distribution for vasoactive intestinal polypeptide immunoreactivity. 112 Figure CR3b\ Graph showing data distribution for vasoactive intestinal polypeptide immunoreactivity (square root corrected). 112 Figure CR3c: Histogram showing vasoactive intestinal polypeptide immunoreactivity. 113 Figure CR 3d. Histogram showing gender differences for vasoactive intestinal polypeptide immunoreactivity. 113 Figure CR4a\ Graph showing data distribution for protein gene product immunoreactivity. 116

14 Figures-Chapter 3 (contd.) Figure CR.4h\ Histogram showing protein gene product immunoreactivity. 117 Figure CR.4c: Histogram showing gender differences for protein gene product immunoreactivity. 117 Figure CR.5a: Graph showing data distribution for protein gene product immunoreactivity at the epidermal-dermal junction. 120 Figure CR.5b: Histogram showing epidermal-dermal protein gene product immunoreactivity. 120 Figure CR.6a: Graph showing data distribution for sweat gland acinar perimeter. 122 Figure CR.6b. Graph showing data distribution for sweat gland acinar perimeter (square root corrected). 122 Figure CR.6c: Histogram showing sweat gland acinar perimeter. 123 Figure CR. 6d. Histogram showing gender differences for sweat gland acinar perimeter. 123

FIGURES-CHAPTER 4 Figure DR. la\ Graph showing data distribution for rat sweat gland perimeter. 145 Figure DR. lb. Histogram showing sweat gland acinar perimeter in young and old rats. 145 Figure DR.1.2.1a\ Graph showing data distribution for protein gene product immunoreactivity in rat sudomotor nerves. 147 Figure DR. 1.2.1b: Histogram showing protein gene product immunoreactivity in young and old rats. 147 Figure DR. 1.2.2a: Graph showing data distribution for nerve growth factor (p75) immunoreactivity in rat sudomotor nerves. 149 Figure DR. 1.2.2b: Histogram showing nerve growth factor (p75) immunoreactivity in young and old rats. 149

15 FIGURES-CHAPTER 5 Figure ER. la: Graph showing data distribution for the proportion of type I fibres. 174 Figure ER.lb :Histogram showing the proportion of type I fibres for all patients and for patients according to gender. 174 Figure ER.la: Graph showing data distribution for type I and II muscle fibre area. 177 Figure ER.2b: Graph showing data distribution for type I and II muscle fibre area (square root correction). 177 Figure ER.2c: Histogram showing muscle fibre area. 178 Figure ER.2d: Histogram showing fibre type and gender differences for fibre area. 178 Figure ER.2e: Histogram showing age differences for fibre area. 179 Figure ER. 3a: Graph showing data distribution for in situ hybridisation signal in both type I and II fibres, expressed as signal per fibre. 185 Figure ER.3b: Graph showing data distribution for in situ hybridisation signal in both type I and II fibres, expressed as signal per fibre area. 185 Figure ER.3c: Histogram showing differences in in situ hybridisation signal with length of storage. 186 Figure ER. 3d: Histogram showing in situ hybridisation signal for both type I and II fibres (signal per fibre area). 188 Figure ER.3e: Histogram showing in situ hybridisation signal for both type I and II fibres (signal per fibre). 189 Figure ER.3f: Histogram showing in situ hybridisation signal for both type I and II fibres (signal per fibre area following logarithm correction). 190 Figure ER.3g: Histogram showing in situ hybridisation signal for fibre types (signal per fibre area). 193 Figure ER.3h: Histogram showing in situ hybridisation signal for fibre types (signal per fibre area expressed as a percentage of baseline). 193 Figure ER.3i: Histogram summarising overall fibre type and gender differences for in situ hybridisation signal (signal per fibre area following logarithm transformation). 194

16 Figures-Chapter 5 (contd.) Figure ER. 4a: Result of reverse transcriptase polymerase chain reaction for patient A. 200 Figure ER. 4a: Result of reverse transcriptase polymerase chain reaction for patients B and C 202

17 PLATES Plate 1. Photomicrograph of eccrine sweat glands in the dermis of normal male human skin. 94 Plate 2i\ Photomicrograph of eccrine sweat glands from normal male human skin stained with acetylcholinesterase or incubated with cholinesterase inhibitor. 105 Plate 2ii : Photomicrograph of eccrine sweat glands from trial patients stained with acetylcholinesterase. 106 Plate 3 : Photomicrograph of eccrine sweat glands from clinical trial, control and acromegalic subjects, stained for vasoactive intestinal polypeptide. I l l Plate 4\ Photomicrograph of eccrine sweat glands from GH-deficient and acromegalic subjects, stained for protein gene product. 115 Plate 5: Photomicrograph of skin epidermis and dermis staining for protein gene product and eccrine sweat glands immunoreactive for tyrosine hydroxylase. 119 Plate 6i: Photomicrograph of eccrine sweat glands from rat footpad skin, stained for protein gene product and nerve growth factor receptor. 150 Plate 6ii\ Photomicrograph of eccrine sweat glands from rat footpad skin, stained for protein gene product. 151 Plate 7i\ Photomicrograph of human skeletal muscle demonstrating insulin-like growth factor I message expression by in situ hybridisation. 195 Plate 7ii\ Photomicrograph of skeletal muscle from a baseline GH -deficient subject demonstrating myosin heavy chain and insulin-like growth factor I message expression. 196 Plate 7iii\ Photomicrograph of skeletal muscle from trial subjects following twelve months r-hGH therapy (demonstrating myosin heavy chain and insulin-like growth factor I message expression). 197 Plate 8: Photomicrograph of skin from a trial subject, following twelve months r-hGH therapy, demonstrating insulin-like growth factor I message expression by in situ hybridisation. 231

18 List of abbreviations

General terms For the purposes of this study, the use of the following terms will denote the meaning given below: Trial subjects/patients; adult GH-deficient patients taking part in the present investigation. Baseline subjects/patients', adult GH-deficient patients at the start of the clinical trial. Placebo subjects/patients', adult GH-deficient patients at the six month trial stage on placebo administration for this period. r-hGH treated subjects/patients', adult GH-deficient patients at either the six month or twelve month trial stage, having received six or twelve months r-hGH substitution. Euthyroid subjects/patients', adult GH-deficient patients who were not on thyroxine substitution at any stage of the trial.

Alphabetical list of abbreviations

ACh acetylcholine AChE acetylcholinesterase ACR acromegalic ANOVA analysis of variance ATP adenosine triphosphate bp base pair cAMP cyclic adenosine monophosphate cDNA complementary deoxyribonucleic acid CGRP calcitonin gene-related peptide CNS central nervous system CON control DEPC diethylpyrocarbonate DIG digoxigenin DNA deoxyribonucleic acid ESG eccrine sweat gland

19 GAP growth associated protein GH growth hormone GHBP growth hormone binding protein GHD growth hormone deficiency GHR growth hormone receptor GHRH growth hormone release hormone IGF-I insulin-like growth factor I IGFBP insulin-like growth factor I binding protein IGFR insulin-like growth factor I receptor me immunohistochemistry IR immunoreactivity ISH in situ hybridisation K-W ANOVA Kruskal-Wallis analysis of variance MHC myosin heavy chain min minute/s MLC myosin light chain mRNA messenger ribonucleic acid M-W U test Mann-Whitney U test NGF nerve growth factor NGFR nerve growth factor receptor NPY neuropeptide Y pV SN G F R nerve growth factor receptor (75-80kDa) PBS phosphate buffered saline PGP protein gene product 9.5 PNS peripheral nervous system R receptor r-hGH recombinant human growth hormone RNA ribonucleic acid RSA rat serum albumin RT-PCR reverse transcriptase polymerase chain reac sc subcutaneous SDS sodium dodecyl sulphate

20 SEM standard error of the mean SS somatostatin ssc sodium chloride-sodium citrate buffer SubP Substance P T3 3,5,3 '-triiodothyronine T4 thyroxine TH tyrosine hydroxylase trk A nerve growth factor receptor (140kDa) T-W ANOVA two-way ANOVA UTP uridine triphosphate VIP vasoactive intestinal polypeptide y year/s

21 CHAPTER 1

GENERAL INTRODUCTION

22 ALI. Growth hormone; a historical perspective Growth hormone (GH) is a polypeptide hormone produced from the anterior pituitary gland which has widespread efiects in the body, particularly on growth and metabolism. Abnormalities of GH secretion or action may result in conditions of growth hormone deficiency (GHD), growth hormone resistance (Laron syndrome) or excess (acromegaly). The study of these growth hormone actions has its origins in the classic descriptions of acromegaly by Pierre-Marie in 1886 ^ and the subsequent link which was established between gigantism and pituitary secretion The use of GH in the treatment of short stature in children (due to GHD) followed the isolation of the hormone from human pituitaries in the mid-1950's However, reports of Creutzfeldt- Jakob disease in some patients following this therapy precipitated the introduction of recombinant human GH (r-hGH) in 1985, the cell lines expressing the protein being bacterial or mammalian in origin The increased availability of the recombinant hormone allowed more frequent doses to be administered and also offered the possibility of treatment for children with less severe GH insufficiency “ and other forms of short stature (Turner's syndrome, uraemia in childhood, Prader-Willi syndrome, familial short stature etc) . In addition, clinical trials were initiated to treat patients with GHD in adulthood. At present, wider uses of r-hGH are being actively investigated including such diverse areas as ageing, thermal injury, parenteral nutrition, osteoporosis, obesity, heart failure and muscle metabolism

AI.2. Adult growth hormone deficiency syndrome The action of GH on skeletal growth in development and childhood constitute its primary function This is reflected in the GH profile which is high during foetal life and achieves its maximum values during puberty Though GH levels decline with age thereafter, the production of this hormone continues throughout life The functions of GH in adults have become apparent from studies in GH-deficient adults. These patients can be divided into those with onset of GHD in childhood and those with GHD acquired in adulthood. The latter acquire GHD following pituitary or peripituitary tumours or their treatment Adults with hypopituitarism traditionally received replacement therapy with thyroid hormones, adrenal and sex steroids but not GH However, conventional replacement therapy has failed to rectify several physiological

23 variables. Hence these patients exhibited alterations in body composition (including increased fat, decreased muscle volume, increased serum total- and LDL-cholesterol), psychosocial maladjustments, reduced muscle strength and exercise capacity, subnormal kidney function, reduced bone mineral content, defective thermoregulation, reduced energy expenditure and basal metabolic rate, abnormal thyroid hormone metabolism, premature atherosclerosis and reduced cardiac function, and an increased risk of mortality from cardiovascular disease Following several clinical trials of six to twelve months duration, the consensus was that most, if not all, of the above abnormalities were improved following r-hGH substitution 12-14,17,18,22-36 patients demonstrated reductions in adipose tissue mass, increases in lean body mass (including muscle volume by elevation of protein synthesis), a decrease in plasma LDL-cholesterol and triglycerides, increased mineral content of bone, an increase in extracellular water by normalisation of kidney function, improved exercise capacity, increased sweating, improvement of dry atrophic skin, an increase in cardiac muscle mass and hence cardiac function, and an increased general well-being. Hence the beneficial effects of replacement therapy suggest the extension of the uses of r-hGH to other situations such as ageing. However, with the supraphysiological doses being administered at present the likelihood is that extensions of the uses for GH therapy will only occur once beneficial effects with lower doses (and therefore threshold doses) have become established. There are potential drawbacks of therapy that could limit the benefits of GH substitution. Negative factors include the expense and reported side effects of treatment. The latter include fluid retention, carpal tunnel syndrome, and arthralgia whilst in addition hyperglycemia is a major concern in GH- treated elderly or burn patients, due to the insulin resistance caused by GH

AI.3. Acromegaly Chronic hypersecretion of GH by pituitary tumours results in the acromegaly syndrome. The incidence of the condition is 3-5 per million per year with clinical presenting features including headaches, visual problems, changes in appearance, carpal tunnel syndrome, amenorrhoea and diabetes. Elevated circulating GH and IGF-I levels, rather than the enlarged tumour, are believed to be responsible for the alterations in appearance such as hand and foot enlargement and coarsening of facial features. Other

24 dermatological manifestations include hyperhidrosis, oily skin, pigmented skin tags, and acanthosis nigricans. Cardiovascular problems, degenerative arthritis, alterations in lipid metabolism and an increased mortality are also consequences of acromegaly. Treatment of the condition include transsphenoidal pituitary surgery. Alternatively, therapy may involve administration of the dopamine agonist bromocriptine or the somatostatin analogue octreotide.

AI.4. Growth hormone; structure, binding proteins and receptor Human GH (hGH) is a 191 amino acid single chain polypeptide with two disulphide bonds and constitutes up to 10% (5-lOmg) of the dry weight of the pituitary gland. The somatotrophic cells in the pituitary produce not only the major 22 kDa form of the hormone but also several variants (molecular forms derived from the same primary structure) including a 20kDa variant which constitutes 10-20% of hGH in the pituitary. Coding for hGH is carried out by a gene cluster on the long arm of chromosome 17. Two separate genes code for the various forms of hGH, hGH-N gene (for normal) and hGH-V gene (for variant, primarily expressed in the placenta). The 20kDa variant differs by 13 amino acids from the normal hGH and is produced in the pituitary as a result of post-transcriptional events. These result in the production of two mature mRNAs from the hGH-N gene, encoding the two main forms of hGH. Other splicing products of the hGH-N and hGH-V transcripts also exist, though their biological significance is uncertain. The exception is the placental GH coded by the hGH-V gene which is a potent somatogen and has a role in pregnancy. The function of the 20kDa variant of hGH is similar to that of the native hGH but it lacks the carbohydrate- regulating properties and hence the diabetogenic properties of the 22kDa form. In addition the 20kDa variant has reduced affinity for GH receptors when compared with the native form. Though hGH is synthesised as a pre-hormone with a hydrophobic tail, this is short-lived and post-translational events govern the ultimate form of the hormone. Following secretion, GH rapidly associates with circulating binding proteins (Bps); these may be of a high or low affinity. The low affinity BP is a lOOkDa molecule which forms a 125kDa complex with GH. It is unrelated to the high affinity BP or the GH receptor and binds the 20kDa variant specifically with equal or greater affinity to the 22kDa

25 form Only 5-8% of GH is complexed by this BP which is not saturable at GH levels occurring in vivo The high af&nity GH binding protein is a glycoprotein of 61 kDa which binds the 22kDa hGH (but not the 20kDa) with high affinity to form a complex of 80-85kDa The soluble extracellular domain of the GH receptor is structurally identical to this GH-BP and, in humans, the BP arises from hepatic GH receptor turnover by a process of proteolytic cleavage . In rats and mice, however, GH receptor and BP arise from two alternatively spliced mRNAs . The link between BP and receptor is demonstrated in the rare autosomal recessive growth disorder, Laron syndrome, which is characterised by normal or elevated levels of circulating GH and low levels of IGF-I In this condition of GH resistance, not only is there a GH receptor defect but also the high affinity BP is absent or nonfunctional Though the high affinity BP complexes about 40-45% of circulating GH, at high GH concentrations (above 10-15ng/ml), the BPs are partially saturated and hence acromegalics (ACRs) have a lower proportion of the GH-BP complex Levels of the BPs in patients with GHD are however normal and no significant sex difference is observed in health or disease, though women tend to have slightly higher values Both BP and receptor are regulated by GH, age and nutritional status The function of the BP includes providing a circulating reservoir of GH which is not so easily cleared or degraded In addition, the BP could block GH actions by preventing fiirther binding to receptors However, the affinity of GH for its receptor is greater than that for the BP ensuring a net positive role of the BP such that the half life of GH is increased from 7min for free GH to 27min for the bound GH Since the GH and BPs are in a continuous association-dissociation state, the reservoir property of the BPs serves to dampen the GH oscillations caused by pulsatile release of the hormone (see AI. 5.) from the pituitary The GH receptor is coded for by a single gene and consists of an extracellular hormone- binding domain (24.6kDa), a single transmembrane region (24kDa), and a cytoplasmic domain (3 5kDa) Receptor levels are reduced in rats following fasting and streptozotocin-induced diabetes, and are normalised after feeding and insulin treatment, respectively Also, adipocytes from hypophysectomised rats have low receptor levels which are increased on GH treatment in these and normal rats Conversely, in

26 hypophysectomised male rats GH receptor sites are increased, whilst in culture, GH treatment has been shown to decrease binding sites An inverse correlation between BP/receptor and GH secretion rates has been demonstrated. Hence, the level of GH secretion is indirectly regulated by the BP/receptor through negative feedback by the mediator of many GH actions, insulin-like growth factor I (IGF-I) (see AI.6.) However, the influence of IGF-I may depend on GH status as IGF-I increases GH receptor in normal rats and decreases the elevated receptor levels present in hypophysectomised rats The binding of GH to its receptor involves two receptor sites on the hormone where each site binds to the same region of the GH receptor. A sequential complex is formed where after binding of one site to a receptor, a second receptor binds to the other receptor site thus forming a homodimer It has been proposed that dimérisation of the extracellular portion of the GH receptor could bring the intracellular domains together for signal transduction to proceed This is followed by induction of a non-receptor intracellular protein from the Janus family of tyrosine kinases, JAK2, by physical association of the dimerised receptor with JAK2 The proteins thus tyrosine phosphorylated include the receptor itself and mitogen activated protein kinases which in turn phosphorylate other intracellular enzymes and transcription factors Ultimately the induction of nuclear transcripts occurs including increases of c-fos, c-jun^ JunB, the serine protease inhibitors, Spi 2.1 and 2.2, and IGF-I Gene transcription of IGF-I involves a GH-regulated chromatin region within the IGF-I locus which is reversibly altered in response to GH A parallel system to regulate gene transcription at the local level has also been proposed after the demonstration in the nucleus of GH, BP and full-length GH-receptor The nuclear translocation of GH following internalisation is similar to that of insulin, nerve growth factor, epidermal growth factor etc and is receptor-dependent via the endocytotic lysosomal pathway

AI.5. Growth hormone; secretion and control The production of GH is carried out throughout life with peak production in late puberty and thereafter a gradual decline occurs such that many individuals over sixty years of age have very low circulating levels of GH Throughout the day, GH is released as a series of discrete pulses, about six to twelve per twenty-four hours, with

27 very low concentrations in between pulses. Other stimulatory factors include sleep, fasting, exercise, protein ingestion, hypoglycaemia, pregnancy and psychological stress GH production can be stimulated by oestrogen production with females generally producing more GH than males Obesity, carbohydrate ingestion and high serum free fatty acid levels are known to suppress endogenous GH production The secretion of GH from the pituitary is under the direct influence of the hypothalamus via the stimulatory peptide growth hormone release hormone (GHRH) and the inhibitory peptide somatostatin (SS, or somatotropin release-inhibiting hormone). These peptides are released from nerve terminals in the median eminence into the hypophyseal portal blood and thence reach the pituitary somatotrophs The control exerted by GHRH and SS relates not only to the amount of GH secreted from the somatotrophs but also governs the ultradian, pulsatile pattern of GH release. In humans, this is achieved by a combination of decreased SS secretion and a burst of GHRH release. SS and its withdrawal sets the timing of each GH pulse, whilst GHRH secretion determines the magnitude of the subsequent GH peak The importance of pulsatile release has been demonstrated in children where slow growth can result from a disordered pattern of GH secretion GHRH-expressing neurosecretory cells exert a stimulating effect on GH secretion from their main location in the arcuate nucleus of the hypothalamus. The mechanism for GH secretion comprises GHRH-receptor mediated G-protein activation of adenylate cyclase in the somatotroph, increased cyclic adenosine monophosphate (cAMP) generation and eventually the activation of a pituitary-specific transcription factor which activates the GH gene Somatostatin, the inhibitory constituent to the dual control of GH secretion, is predominantly located in the periventricular nucleus of the hypothalamus with some expression in the arcuate as well The classic mechanism for SS inhibition of GH release involves the binding of SS to its receptor on the somatotroph, which initiates a sequence of events including the inhibition of adenylate cyclase, and hence GH release The demonstration of reciprocal innervation between GHRH neurons of the arcuate nucleus and SS neurons of the periventricular nucleus suggests a hypothalamic mechanism by which GH secretion is coordinated The presence of additional SS neurons in the arcuate, which do not project to the median eminence, may also reduce the steady state levels of GHRH by a local influence

28 GH exerts a negative feedback on GHRH production and hence GH-deficient animal models exhibit elevated levels of GHRH mRNA whilst the converse is true in models of GH hypersecretion including elevated levels of SS mRNA This effect is believed to be direct and independent of a similar action by central nervous system derived IGF-I GH action in the hypothalamus may be through its receptor on hypothalamic neurons leading to the induction of c-fos expression (a proto-oncogene involved in neural plasticity and other responses) Double labelling studies suggest that GH stimulates SS neurons in the periventricular nucleus and neuropeptide Y (NPY) neurons in the arcuate nucleus Hence, NPY neurons may inhibit GH secretion by projecting to GHRH neurons and modulating their activity Other stimulatory influences on GHRH release include noradrenaline and serotonin which are present in the arcuate nucleus Dopaminergic neurons, also in the arcuate, stimulate GH production in normal subjects; however dopamine agonists are used to suppress GH in ACR patients where their effect appears to be mediated by a direct effect on the somatotrophs The synthetic peptide, growth hormone releasing peptide- 6, acts synergistically with GHRH not only as a direct GH secretagogue in the pituitary, but also by activating GHRH neurons in the hypothalamus. This effect is sensitive to inhibition by SS in both the hypothalamus and pituitary Additionally, the neuropeptide vasoactive intestinal polypeptide (VIP) directly stimulates GH release in the human pituitary (through activation of the adenylate cyclase-cAMP system) and indirectly in the hypothalamus (by inhibiting SS release) Conversely, the neuropeptide calcitonin gene-related peptide (CGRP) causes a suppression of GH release in the pituitary and in the hypothalamus by stimulating SS secretion Catecholaminergic projections to the hypothalamus can also influence GH secretion. Stimulation of (%2-^drenoreceptors in the hypothalamus results in GH release through GHRH secretion whilst p-adrenoreceptor or muscarinic cholinergic receptor activation causes SS stimulation and inhibition respectively A functioning central serotonin (5HT) system is required by the a2-adrenoreceptors in order to stimulate GH release The significance of these brain transmitter molecules in the control of GH release could be an influence on the pulsatile release of GH since pulse generation is believed to be generated by explosive increases in the spike activity of neurons In addition, the recognition that some forms of GHD in children may be due to catecholaminergic

29 disturbances has resulted in successful treatment with noradrenergic-acting drugs Similarly, cholinergic drugs enhance GHRH treatment by reducing the influence of SS

AI.6. Insulin-like growth factor I; structure, secretion, function and control The observations by Salmon and Daughaday in 1957 on the stimulation by GH of a serum-derived sulfation factor, opened the way for the subsequent proposal that GH actions on tissue growth are mediated by an intermediate compound. This factor was purified in 1972, termed somatomedin-C, and was later re-named insulin-like growth factor I (IGF-I) There are two IGFs, known as IGF-I (generally GH-dependent) and IGF-n (generally GH-independent) The human serum IGF-I is a 70 amino acid single chain polypeptide whilst IGF-II has 67 amino acids; they share about 70% identity in their amino acid sequence and have about 50% structural homology to proinsulin IGF-II appears to be important in foetal life though it retains some functions in the adult The traditional mode of action for IGF-I involves GH-dependent release from the liver with the circulating IGF-I thus formed serving a classic endocrine function These general functions of IGF-I can be divided into those involving stimulation of growth and those that are insulin-like, such as lipid and glucose metabolism Anabolic effects include promotion of skeletal, renal, thymic and splenic growth, and the promotion of skeletal muscle and hepatic protein synthesis In addition, many tissues in the body synthesise IGF-I; hence there are paracrine and autocrine functions of IGF-I which are generally under GH control. Exceptions exist as in the gonadal production of IGF-I ” , and the modulation of IGF-I by additional factors in the uterus, aorta, thyroid, muscle and in fibroblasts The biological actions of the IGFs also encompass the terminal differentiation of many cell types In the circulation and extracellular spaces, IGFs exist almost always bound to their binding proteins (IGFBPs). There are six in number and they fulfill five major fimctions; they are transport proteins, they prolong the half-life of IGFs (from lOmin to 12-20h), they provide a means of tissue- and cell-type specific localisation, they modulate the interaction of IGFs with their receptors, and they can have direct effects on cellular functions In serum, about 75% of IGFs are associated with the ~50kDa IGFBP-3, and an acid-labile subunit, to form a 150kDa complex (ratio of Imol IGF per mol BP-3)

30 The association of the IGFBP with proteins on the ceil surface or in the extracellular matrix results not only in a lowering of the IGFBP affinity to IGF, but also in the increase in the local concentration of IGF in the vicinity of its receptor and hence the potentiating of IGF action The control of IGF-BP3 and the 150kDa complex appears to be from GH, with low levels in GHD and high levels in acromegaly Since IGFBPs are produced in many tissues, it would appear that they may also be closely linked to the autocrine and/or paracrine actions of IGF-I The IGFs bind to three classes of receptor, the type-I IGF receptor (high affinity IGF-I and low afSnity IGF-II), the type-II IGF receptor (high affinity IGF-II) and the insulin receptor (low affinity IGF-I and IGF-II) In addition, IGF/insulin hybrid receptors have also been demonstrated in cells expressing both receptor genes The type-II receptor consists of a single polypeptide chain whilst the other receptor types consist of two extracellular a-subunits and two transmembrane p-subunits Signal transduction through the type-I receptor is carried out by both tyrosine kinase activity and receptor autophosphorylation The substrate for phosphorylation is termed insulin receptor substrate 1 This ultimately activates the guanine triphosphate-binding protein Ras which in turn stimulates mitogen activated protein kinases and these transmit a signal to the nucleus Additional pathways identified include those with phosphatidylinositol-3 phosphate generation, activation of protein kinase C, and possibly cAMP and serine/threonine protein kinases pathways A variety of growth factors have been shown to upregulate IGF-I receptors, including platelet-derived growth factor, basic fibroblast growth factor, oestrogen, angiotensin II, and thrombin This is not an incidental function of these growth factors since a functional IGF-I-IGF-I receptor autocrine loop is suspected to be essential for many of their mitogenic actions The exact level of IGF-I regulation may involve both transcriptional and post- transcriptional mechanisms The IGF-I gene in rats is composed of six exons separated by five introns, with the mature peptide-coding sequence contained within exons 3 and 4, which form a core sequence to which alternate 5' and 3' sequences are spliced Two promoters control transcription from multiple start sites on exons 1 and 2 The resultant transcripts are expressed differentially in a pattern that is dependent on tissue type, age, nutritional status and GH status Hence, GH has been shown to

31 induce accumulation of transcripts through both IGF-I gene promoters Also, exon 2 IGF-I mRNA is more GH-dependent than exon 1 mRNA Exon 2 transcripts are expressed primarily in the liver, whilst exon 1 transcripts are the major transcripts found in all tissues, including liver Le Roith and Roberts suggest that exon 2 transcripts encode the 'endocrine' form of IGF-I whilst exon 1 transcripts may be considered to encode the paracrine/autocrine form of IGF-I. The latter may be regulated by additional factors such as nutritional status, oestrogens, tissue-type etc The generation of these transcripts may occur through the encoding of different sized signal peptides such that the IGF-I molecule generated by signal peptidase would be incomplete for the exon 1-derived molecule (proIGF-I) and be ready for secretion for the exon 2-specified form

The negative feedback by GH on IGF-I production is at the transcriptional level and involves a series of DNA-protein interactions through the major promoter of the IGF-I gene Conversely, there is increasing evidence that the short-loop feedback mechanism for GH secretion involves not only an effect of GH/IGF-I on the hypothalamus, but may also involve local autocrine/paracrine actions of IGF-I in the pituitary. Hence, the presence of IGF-I and -II mRNAs as well as IGF receptor mRNA in rat brain including the hypothalamus and pituitary suggests dual sites of action for IGF-I regulation of GH production In rats at least, the mechanism in the hypothalamus may include a synergy between IGF-I and IGF-II to modulate SS/GHRH secretion Qi^ect pituitary inhibition of GH gene expression at the level of transcription has been demonstrated Data from a rat model of somatomammotropic tumours suggests that this autoregulation may involve the mutually interactive regulation of GH and IGF-I genes in either neighbouring cells or at a single cell level The demonstration of GH-receptor downregulation by locally produced IGF-I in osteoblasts suggests the presence of a peripheral negative feedback loop for GH in extrahepatic tissues The uses of r-hlGF-I have been explored in clinical studies to treat catabolic states, such as ageing and cachexia The administration of r-hlGF-I to normal adults causes an inhibition of insulin secretion without causing hypoglycaemia In other studies hypoglycaemia does occur due to an effect of IGF-I on glucose transport, perhaps at high doses. Given that GH causes insulin resistance it has been suggested that the

32 administration of both GH and IGF-I may be more beneficial in the treatment of patients with catabolic diseases In this context, the finding that IGF-II acts as a functional inhibitor of the anti-catabolic actions of IGF-I, suggests a role for IGF-II in adulthood Additional uses of IGF-I may be in the treatment of hyperglycaemic states such as diabetes in the treatment of GH-insensitivity conditions such as Laron syndrome and in recovery from renal disease Aside fi*om conditions such as Laron syndrome where a GHR defect prevents GH action, the clinical use of r- hlGF-I would avoid beneficial effects of direct acting GH. The finding of GH-secretion suppression in Laron syndrome subjects following r-hlGF-I treatment is also an indication of the possible undesirable consequences of such an approach Further concerns over r-hlGF-I treatment include the inducing of hypoglycaemia, and the lack of prevention of protein catabolism in normal adults

AI.7. Growth hormone, insulin-like growth factor I and the nervous system AL7.1. Growth hormone and the nervous system GH binding sites have been identified in adult human and rat brain regions such as choroid plexus, hippocampus, thalamus, hypothalamus, and pituitary Lobie et al have demonstrated protein and mRNA for the GHR/BP in a number of CNS sites in neonatal rat and rabbit brain. These include the cerebral cortex, neurones of the thalamus, hypothalamus, colliculi and brainstem, pyramidal cells of the hippocampus, Purkinje cells of the cerebellum, glomeruli and mitral cells of the olfactory bulbs, and retinal ganglion cells. Glial cells, particularly astrocytes and ependymal cells also display GHR/BP immunoreactivity. These sites of expression correlate well with the distribution of IGF-I protein and message; sites of common distribution include cerebellar Purkinje and glial cells, the hippocampus, thalamus, hypothalamo- hypophyseal system, spinal cord and olfactory bulb. Treatment with GH of the GH- deficient Little mouse results in glial proliferation, and hence increased myelinogenesis, restoring normal weight of the cerebrum and cerebellum GHR/BP mRNA levels, as well as GH binding sites, in the brain decline with age in humans, rats and rabbits 111,112,114 addition, GH appears in the brain before its ontogenic appearance in the adenohypophysis, indicating a role of GH in brain development and maturation GH appears to be important to the nervous system during post-natal life and in maturity:

33 Excess states, for example acromegaly, may lead to the development of a generalised peripheral neuropathy which has its origins in an initial demyelination process Patients with isolated GHD in childhood, and Laron syndrome subjects, have a small head circumference and show mental retardation; r-hGH therapy in these patients aids increases in IQ “ 5,119,120 syndrome patients also have visuo-motor disturbances suggesting that receptor defects in the CNS are responsible GH-deficient adults have decreased psychosocial well-being whilst r-hGH therapy results in an improvement in the quality of life and increased vitality and mental alertness in these subjects 27,116,121-123 improvements may be secondary to changes in body composition etc; however, GH effects on the nervous system cannot be ruled out It has been suggested that GHRs in the choroid plexus may be involved in the transfer of the hormone over the blood-brain barrier; however, other authorities suggest an endogenous GH/IGF-I axis in the brain with local GH synthesis in the CNS Evidence from mature rats implicates both the latter suggestions; intraperitoneal injections of ^^^I-labelled GH result in accumulation of GH in brain areas, whilst GH mRNA is present in several brain areas, being broadly co-distributed with its receptor sites “ 2,114,116,124 Yhe demonstration of increased levels of GH and IGF-I, in the cerebrospinal fluid of r-hGH treated adult GH-deficient subjects, was the first indication that r-hGH can cross the blood-cerebrospinal fluid barrier in humans This study also showed no correlation between serum IGF-I levels and those in the cerebrospinal fluid Johansson et al suggest a CNS origin for the IGF-I which presumably also implies that r-hGH crossed the blood-brain barrier. An intriguing finding from the latter study was an increase in p-endorphin levels as well as a decline in concentration of VIP and the dopamine metabolite homovanillic acid following r-hGH therapy These changes, taken together with GH action in the hippocampus, may constitute the biological basis of the improved psychological well-being observed following r-hGH treatment in GH-deficient subjects Dwarf mice have been used extensively to investigate the influence of GH on its regulatory neurons in the hypothalamus '^^>“ 5-129 ^ggg Phelps, C.J. for review). In these mice, a reduction occurs in the hypothalamus, of both SS neurons (in the periventricular nucleus) and dopamine neurons (in the arcuate nucleus). Conversely, GHRH neurons in the arcuate nucleus are increased in the same region where dopamine

34 neuronal cell loss occurs. These results suggest a lack of pituitary GH feedback control over these neurons in the hypothalamus. Further, a reduction occurs in GHRH and SS terminations in the median eminence which is indicative of a role of pituitary signals in axon guidance during development. The mechanism for these actions may be indirect via IGF-I, though GH may also pass through the blood-brain barrier which is permeable in the ventral arcuate nucleus. Further, GH may be carried to the hypothalamus by retrograde flow from the hypophyseal portal vasculature The intercellular mechanism for these changes has been investigated in r-hGH treated hypophysectomised rats The process appears to involve activation of the c-fos gene (an indicator of neuronal plasticity), in SS neurons of the periventricular nucleus, and also in NPY neurons in the arcuate nucleus; the latter may project either to SS neurons in the periventricular nucleus or to GHRH neurons in the arcuate nucleus GH may thus be regarded as a neurotrophic factor for its regulatory neurons, influencing survival, connectivity and phenotypic plasticity of its developing regulatory 'hypophysiotropic' neurons The role of GH in the PNS has not attracted much research interest; however, sciatic nerve regeneration is known to be positively influenced by r-hGH treatment in normal and hypophysectomised rats Since the r-hGH was administered systemically, it could not be ascertained if this was a direct effect of GH or an indirect one mediated by IGF-I. Ullman et al investigated muscle regenerative responses in normal adult rats and showed that treatment with r-hGH resulted in satellite cell proliferation and muscle fibre hypertrophy. Following denervation, r-hGH administration prevented muscle fibre atrophy, thus indicating that this effect was independent of innervation Once again, the influence of systemic or local IGF-I could not be dismissed in this experimental setup.

AI.7.2. Insulin-like growth factor I in neuronal development The activity of IGFs in promoting neurite outgrowth and survival was initially described in cultured (neuroblastoma, sympathetic, sensory) cell lines by Recio-Pinto et al Subsequently, IGF-I was shown to promote motoneuron survival during the period of developmentally regulated neuronal death in chick embryos Further, axotomy- induced death of motoneurons was prevented by IGF-I in developing mouse spinal cord

35 as was atrophy of surviving motoneurons In embryonic rats, IGF-I mRNA increased in muscle at the same time as polyneuronal innervation of single motor endplates was occurring Synapse elimination in rat development occurs once motoneurons reach their targets and synaptogenesis has occurred This muscle-activity dependent decline is accompanied by downregulation of muscle IGF-I, neuronal growth associated protein (GAP) and tubulin which can be prevented by local IGF-I treatment Ishii, D.N. suggests that the development of neuromuscular synapses, regeneration of synapses, and sprouting of motor nerve terminals are modulated by the presumably autocrine and/or paracrine neurotrophic effects of IGF-I. The transient localisation of IGF-I and its receptor to many CNS neurons and glia in development also indicate a neurotrophic action of this growth factor IGF-I was shown to induce oligodendrocyte and astroglial proliferation, and oligodendrocyte differentiation Further actions included those on neuronal proliferation, differentiation, elaboration and maturation during development In transgenic mice overexpressing IGF-I, brain neuronal cell size and number, and total myelin content, was increased significantly over control mice Conversely, transgenic mice lacking the IGF-I gene were shown to have distinct neurological aberrations including decreases in neuronal size Preliminary evidence has been presented suggesting that r-hlGF-I treatment to a Laron syndrome subject resulted in an increased myelination of the optic nerve

AI.7.3. Insulin-like growth factor I and the mature nervous system The detection of IGF-I and IGF-I receptor (binding sites, protein and/or message) in adult rat neurons (hypothalamo-hypophyseal system, olfactory bulb, cerebral cortex, choroid plexus, hippocampus, motoneurons and PNS neurons), as well as the characterisation of a neuronal structural subtype of IGF-I receptor in mature rat brain, indicated a function of IGF-I in adulthood 141,142,147,i6i-i64 ^ variant IGF-I peptide has been demonstrated in both foetal and adult human brain This IGF-I molecule arises out of post-translational modification and, due to its reduced affinity to IGFBPs, is more potent than the native peptide IGF-I in the mature nervous system may support synapse turnover and formation. Consistent with the latter role, there are

36 relatively high levels of IGF-I and its receptor in the olfactory bulb, which constitutes an area of the rat brain with a high synapse turnover IGFs can be regarded as soluble nerve- and muscle-derived neurotrophic factors able to regulate the rate of peripheral nerve regeneration Hence, IGF-I infusion was found to increase the rate of adult rat and mice sciatic nerve regeneration and sprouting 167-169 pur|;hei-^ perfusion with specific antibodies against IGF-I impaired the regenerative response IGF-I immunoreactivity was found to be present in regenerating sciatic nerve fibres with expression elevated in the Schwann cells Further, axoplasmic transport of IGF-I was shown to be in both retrograde and anterograde directions in adult rat sciatic nerves following crush The sequestering of IGF-I by its BPs close to the surface of muscle fibres during regeneration and development, is suggested by Ishii, D.N. to provide a mechanism for the chemotactic attraction, support and maintenance of nerve fibres in these situations. The latter proposal is supported by observations on rat sciatic nerve regeneration; IGF-I was produced by both Schwann cells and invading macrophages, whilst transcripts for IGFBP-5 and the IGF-I receptor were additionally generated by Schwann cells A similar finding for regenerating rat CNS motoneurons (following facial nerve axotomy) has been observed In the latter study, the production of IGF-I and IGFBP-2 by astrocytes is preceded by possible activation of the IGF-I response by microglial cells IGF-I treatment enhances the rate of functional recovery in peripheral nerve regeneration suggesting a possible influence on neurotransmitters in neuromuscular innervation The release by IGF-I of acetylcholine from rat cortical slices, and the inducement of choline acetyl transferase activity in neuroblastoma cell lines, also suggest that IGF-I affects neurotransmitter synthesis or release In this context, the prevention of lumbar motoneuron death during development by IGF-I is accompanied by maintenance of choline acetyltransferase levels Induction of catecholaminergic phenotype by IGF-I, in dorsal root ganglion cultures, was the first direct evidence to show an involvement of IGF-I in the developmental changes which occur in peripheral neurons Aside fi*om neurotransmitter/neuropeptide regulation, Patterson, P.H. holds the opinion that many cytokines (including IGF-I, nerve growth factor and brain-derived neurotrophic factor) may influence intercellular

37 signalling directly, through protein kinase activity for example, and thus serve a neurotransmitter function.

AI.8. Aims of the investigation The present study was made possible by a clinical trial initiated in 1993 to explore the effects of r-hGH in adult GH-deficient subjects. This involved a double-blind placebo controlled investigation with r-hGH treatment for six to twelve months. Despite a positive finding by Raben,M.S. in 1962, it has only been in the last decade that a number of studies have examined the repercussions of GH-substitution in the hypopituitary adult. Since the initial trials 22-24,29,180-182^ the focus of subsequent studies has been increasingly targeted at determining the underlying mechanisms which govern the beneficial effects of r-hGH therapy. Hence, the mode of r-hGH replacement (eg pulsatile or continuous, frequency of doses) has been evaluated and changes associated with inter-patient variation were noted by Johannsson et al Further, target tissues such as the CNS heart and muscle have been examined for their responsiveness to r-hGH therapy. In this context, the present investigation has attempted not only a broad monitoring of 'classic' endocrine responses, but has additionally utilised morphological, biochemical, histological and molecular approaches in an attempt to understand some aspects of the mechanism of GH action. These studies have focussed on three parallel questions. To what extent are beneficial effects of GH the result of direct actions of this hormone, or, of indirect effects mediated by IGF-I? In indirect responses, what role does target- derived IGF-I have as compared with circulating IGF-I? Finally, does GH either directly or indirectly have a neurotrophic function? The two target systems evaluated in this study, sweat glands and muscle, are known to show altered function in GHD. Abnormal thermoregulation and hypohidrosis are noted consequences of GHD whilst ACRs classically display hyperhidrosis (see chapter 3). Muscle mass and strength are reduced in GHD and revert to normal levels following r-hGH therapy (see chapter 5). The relative ease of obtaining biopsy specimens from the skin and muscle have allowed the exploration of end-organ changes before, during and at the end of the clinical trial {chapters 3 and 5). The methodology was chosen to facilitate careful examination of data in a quantitative or semi-quantitative fashion (see

38 CM. 3.4. and EM. 5.). Further, biochemical analysis of key circulating hormones (eg serum IGF-I and thyroid hormones, see chapter 2) was carried out to allow comparison with the target-organ data from skin and muscle. Part of the process of understanding GH actions has been the increasing use of animal models such as mutant dwarf animals (eg. dw, d fm d lit strains of mice, dw strains of chicken, Lewis dwarf rat), transgenic mice, and hypophysectomised rats The horizons for the usage of r-hGH have, however, recently expanded to include ageing, malnourishment and maintenance/increase of the anabolic status in a wide variety of situations (for increasing muscle bulk in sports and for patients undergoing trauma, severe infections or glucocorticoid therapy) Part of the present study has expanded on the sudomotor results from Chapter 3 using r-hGH treated aged rats (see chapter 4) in order to evaluate the direct actions of GH on the sweating apparatus in ageing, a state of physiological GHD. The results from these complimentary approaches have generated novel hypotheses governing the mechanism of GH action, particularly in relation to interactions of GH with peripheral nerves and their targets.

39 CHAPTER 2

AN INVESTIGA TION INTO RECOMBINANT

HUMAN GROWTH HORMONE TREA TMENT OF

GROWTH HORMONE-DEFICIENT ADUL TS.

40 Chapter 2 Introduction

B1.1. Diagnosis of adult growth hormone deficiency The recognition of the GHD syndrome as a distinct clinical entity in adults followed from the early work of Cuneo et al in the late 1980s. Diagnosis is usually based on both clinical and biochemical features. The former includes known pituitary pathology, reduced strength and exercise capacity, increased abdominal adiposity and impaired psychological well-being There are a range of GH stimulation tests including the use of insulin, GHRH, clonidine, levodopa, arginine etc The assessment of spontaneous GH secretion by 24-hour blood GH measurements, and the evaluation of serum IGF-I and IGFBP3 levels, have also been used to diagnose GHD. The relative infancy of this field is reflected in a consensus on the criteria, for diagnosis of profound adult GHD, emerging only in 1995 Thomer et al suggest a peak GH response of less than 3pg/l (9mIU/l, with the insulin tolerance test) as one of the main indications which merit r-hGH substitution.

BI.2. Growth hormone, insulin-like growth factor I and thyroid hormones The hypothalamic influences of thyrotropin releasing hormone (stimulatory) and SS (inhibitory) govern the secretion of thyroid stimulating hormone (TSH or thyrotropin) from the anterior pituitary The feedback inhibition of TSH by the prohormone thyroxine (T4) is an additional control mechanism, whilst peripheral T4 metabolism to 3,5,3'-triiodothyronine (T3) results in the production of the active hormone Hypothyroidism is associated with growth impairment coincident with reduced basal and stimulated GH release, as well as low circulating levels of IGF-I Colonna et al have shown that an early response to r-hGH treatment of hypothyroid rats is a direct feedback of GH on GHRH neurons. However, at the onset of hypothyroidism, rats displayed reductions in GH levels prior to changes in GHRH neurons and no changes in SS levels were observed Hence, the mechanism of thyroid hormones action on GH release may not be mediated by the hypothalamic regulatory system for GH During development, T3 binding to nuclear receptors is known to stimulate GH synthesis (as well as NGF, epidermal growth factor) and consequently IGF-I production

41 It is thus likely that the anabolic effects of thyroid hormones are mediated by other growth factors IGF-I has been shown to increase the rate of T4 conversion to T3 and, conversely, T3 can inhibit IGF-I gene expression in the pituitary 104,202,207,20s interactions between the GH/IGF-I axis and thyroid hormones in the adult exist in many tissue systems and will be further clarified at appropriate junctures as well as in the general discussion.

BI.3. Growth hormone, insulin-like growth factor I and bone metabolism A major anabolic effect of GH is to stimulate bone turnover and increase bone mass In addition, GH stimulates type I and III collagen metabolism GH markedly influences phosphocalcium homeostasis and bone metabolism and continues to play a role in bone physiology in adulthood In the adult skeleton, bone formation is regulated by the coupling of formation to resorption so as to achieve a mechanically optimal bone structure; IGF-I is believed to mediate this process The mechanism involves GH stimulation of osteoclast production which then promotes and governs osteoblast-mediated cavity re-filling. GH further directly activates osteoblasts to increase local synthesis of IGF-I hence enhancing bone matrix apposition, suggesting a role in the preservation of bone mass. Though the growth-promoting activities of GH are mediated by IGF-I, bone resorption appears to be a direct action of GH Further aspects will be covered in the discussion (BD.4.) and general discussion chapters (F2.1.).

BI.4. Aims of the investigation Aside from providing patient characteristics and an overview of the design of the clinical trial, this chapter presents results from assays for serum IGF-I, serum thyroid hormones (T4 and T3) and urinary total pyridinoline crosslinks. The clinical trial covered many aspects of r-hGH therapy including body composition, heart function, quality of life etc. However, the primary purpose of presenting the former three parameters is to allow an appreciation that r-hGH therapy has consequences for various target organs. Hence, r-hGH effects in liver, thyroid, and bone not only have influences for themselves, but also impinge on other tissue systems (eg skeletal). For serum IGF-I

42 and thyroid hormones, an additional consideration was the interaction between these hormones with possible repercussions for their target organs. Data presentation and statistical methods in this, and subsequent chapters, are explained in Appendix G. 1.

43 Chapter 2 Methods

BM.l. Growth hormone-deficient patients Twenty subjects (10 males, 10 females; aged 25 to 53 years, mean age 39 years) with GHD (defined as peak GH <10.0 mIU/1 during an insulin hypoglycaemia test) for more than 12 months were selected. Where necessary, patients had been receiving other pituitary hormone replacement therapy for more than six months. Subjects were randomly assigned to receive either r-hGH (Genotropin R: Pharmacia, Sweden) or placebo for 6 months, following which all patients received r-hGH for 6 months. Therefore, of the GH-deficient subjects, 8 patients (5 males, 3 females) received r-hGH for 12 months and 12 patients (5 males, 7 females) received placebo then r-hGH for six months each. At the end of the clinical trial, patients were asked whether they wished to continue with r-hGH replacement therapy, (see table BR. 1. for patient details and responses) The present study was part of a broader investigation exploring the effects of r-hGH on quality of life, body composition and numerous parameters of metabolism in GH- deficient adults as well as assessing its short term safety. The protocol was approved by the Royal Free NHS Trust Ethic's Committee and informed written consent was obtained from all subjects.

BM.2. Administration of recombinant human growth hormone The study drug was supplied as sterile powder for injection with solvent in a two- compartment cartridge with a customised 'pen' for injection purposes. Each cartridge after reconstitution contained; 16 lU r-hGH, 2.0mg amino acetic acid, 41mg mannitol, 0.29mg anhydrous sodium phosphate, 0.28mg anhydrous disodium phosphate, 3mg M- cresol, and 1ml water for injection. Placebo was supplied in identical cartridges for reconstitution containing 1ml water for injection with 3.0mg M-cresol. The doses were 0.125 lU per kg sc/week during the first four weeks of each 6 month period and thereafter 0.25 lU per kg sc/week for five months. The weekly dose was divided into seven daily subcutaneous injections. The mean dosage per day was 1.5 lU (first month) or 3.0 lU (for subsequent five months).

44 BM.3. Serum insulin-like growth factor I assay Fasted early morning blood samples were taken for measurement of serum IGF-I after acid-ethanol extraction. The IGF-I levels were measured by polyethylene glycol-assisted second antibody (polyclonal anti-rabbit IgG antiserum) radioimmunoassay as described in Bang et.al. The detection limit of the assay was 20 ng/ml. The intra-assay and inter-assay coefficient of variation at 202 ng/ml were 3.1% and 10.0% respectively, (for further details of the protocol see appendix G.2.)

BM.4. Serum free 3,5,3-triiodothyronine and free thyroxine assay T4 and T3 were determined by enzyme immunoassay using ES 700 analyser (Boehringer Mannheim, UK). The intra-assay coefficient of variation were as follows; T4 6.0% at 14 pmol/1, 3.7% at 57.1 pmol/1; T3 3.3% at 6.5 pmol/1, 1.3% at 21.6 pmol/1. The inter-assay coefficient of variation were; T4 2.7% at 15.7 pmol/1, 2.1% at 43.5 pmol/1; T3 3.8% at 6.9 pmol/1, 2.9% at 17.2 pmol/1. (for further details of the protocol see appendix G.2.)

BM.5. Urinary total pyridinoline crosslinks assay BM.5.1. Collagen pyridinium as a marker of bone metabolism The pyridinium cross-links of bone, pyridinoline (also present in articular cartilage) and deoxypyridinoline provide structural integrity and rigidity to mature, insoluble collagen fibrils. These naturally fluorescent molecules are derived from lysine and hydroxylysine. During bone degradation (resorption), the cross-links are released into the circulation, not metabolised and eventually excreted in urine. Pyridinium cross-link measurements in urine have been shown to be sensitive and specific indicators of resorption by both established HPLC and newer enzyme immunoassay techniques.^^^

BM.5.2. Methodology An isocratic ion-paired reverse-phase high-performance chromatographic assay using narrow-bore columns and high sensitivity fluorescence detection was utilised, with

45 coefficient of variation for extraction and analysis being 7.96%. (for further details of the protocol see appendix G.2.)

BM.6. Acknowledgements The serum IGF-I assay was carried out by Pharmacia Peptide Hormones UK, the T4 and T3 assays by the Department of Chemical Pathology, Royal Free NHS Trust, and the urinary total pyridinoline courtesy of Dr D Perrett, Department of Rheumatology, St. Bartholomew's Hospital Medical College, London.

46 Chapter 2 Results

BR I. Patient responses to recombinant human growth hormone therapy The r-hGH substitution was generally well-tolerated; however, five patients experienced peripheral (hands and feet) oedema. In the patient group as a whole, there was no significant improvement in quality of life while on active drug, as assessed by quality of life questionnaires. Individual patients did, however, experience improvements in life style with positive changes in weight, body composition and exercise capacity. Eleven patients (six males and five females) chose to continue with the r-hGH therapy (see table BR.1).

BR.2. Serum insulin-like growth factor ! The distribution of the data were not normal (figure BR.2a); however, square root transformation resulted in a near normal distribution (Shapiro-Wilk statistic=0.897; figure BR.2b). Original data are presented in the figures below; however, statistical analysis were carried out on square root transformed values. The mean serum IGF-I value for the GH-deficient subjects was 150+/-20ng/ml. Following r-hGH treatment, patients showed a statistically significant increase (52-61%, x^

47 A negative correlation was observed between patients serum IGF-I values, and their levels of T4 (Spearman correlation=-0.53, p<0.001). There was no correlation observed between T3 values and the circulating IGF-I levels.

Summary ► Serum IGF-I levels are low in GH-deficient patients and increase markedly within six months of r-hGH substitution. There is no further rise in serum IGF-I in patients treated with r-hGH for twelve months. ► Male subjects have greater levels of serum IGF-I than female patients. ► The subjects aged 35-44 years show elevated serum IGF-I levels as compared with older and younger patients. ► A negative correlation exists between serum IGF-I and serum free T4.

48 Table BR I. Individual patient characteristics, clinical history and trial groups.

patient age sex cause of durat additional r-hGH r-hGH at (male GHD ion of substitution or continued start or GHD during trial placebo at end of of female) at for first trial (yes trial start of six or no)

(y) trial months (y) 1 45 male pituitary 2 testosterone r-hGH yes adenoma surgery 2 38 male pituitary 2 thyroxine r-hGH no apoplexy hydrocortis one testosterone 3 53 male chromoph 5 thyroxine placebo yes obe hydrocortis adenoma one surgery testosterone 4 37 female pituitary I none placebo no tumour surgery

5 27 female craniophar 2 none placebo no yngioma surgery

6 29 male prolactin 12 thyroxine r-hGH yes oma surgery hydrocortis one testosterone

49 7 28 female craniophar 15 thyroxine placebo yes yngioma hydrocortis surgery one Estrapak vasopressin 8 25 male juvenile 1 none placebo yes astrocytoma surgery 9 53 male Cushing's hydrocortis placebo yes disease; one surgery fludrocortis one 10 48 female chromoph 1 thyroxine placebo yes obe hydrocortis adenoma one surgery Prempak C 11 48 male chromoph 1 thyroxine placebo no obe hydrocortis adenoma one surgery testosterone 12 39 female prolactin 1 hydrocortis placebo yes oma surgery one Prempak C 13 46 female prolactin 2 Prempak C r-hGH no oma surgery

14 46 male prolactin 3 thyroxine placebo yes oma; prednisolone bromocript testosterone ine treatment

50 15 37 female prolactin 13 none r-hGH yes oma; bromocript ine treatment 16 32 female craniophar 2 thyroxine placebo no yngioma hydrocortis surgery one microgynon 17 46 female pituitary 1 thyroxine placebo yes tumour hydrocortis surgery one Estrapak 18 35 male Cushing's 1 thyroxine r-hGH no disease; hydrocortis surgery one testosterone 19 51 male chromoph 1 thyroxine r-hGH yes obe hydrocortis adenoma one testosterone 20 34 female Sheehan's 2 thyroxine r-hGH yes syndrome prednisolone ethinyloestra- diol norethisterone

51 12

200 300 500 800 serum IGF-I (ng/ml) observed expected (normal)

Figure BR.2a: Graph showing data distribution for serum IGF-I (ng/ml). Samples from patients at the three trial stages. Expected normal distribution also shown.

9 (/) c 8 Q) m 7 Q.

Q) 5 3E C 4

0 10 20 serum IGF-I (corrected) obsen/ed expected (normal)

Figure BR2b; Graph showing data distribution for serum IGF-I following square root transformation. Samples from patients at the three trial stages Expected normal distribution also shown

52 ***

rj) C 200

[ GHD-Om P-6m GH-6m GH-12m GH-«/12m ]

Figure BR.2c: Histogram showing serum IGF-I levels. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH. Error bars show standard error of the mean (SEM), ***p<0.001.

400

350

300

,250

200

150

100

0 male female

I ^ GHD-Om ^ P-6m ^ GH-6/12m |

Figure BR.2d: Histogram showing gender differences for serum IGF-I levels. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH. Error bars show SEM, ***p<0.005.

53 350

300

250

I 200 q E 150 2 (/)0) 100

I ffl 25-34y % 35^y 1:^ 45-53y |

Figure BR.2e: Histogram showing age differences for serum IGF-I levels. Error bars show SEM, *p<0.05.

54 BR.3. Thyroid hormones BR.3.1. Serum free 3,5,3-triiodothyronine The data were not normally distributed (figure BR.3.la); however, logarithmic transformation resulted in a near normal distribution (Shapiro-Wilk statistic=0.865, figure BR.3.lb). Original data is presented in the figures below; however, statistical analysis was carried out on logarithm transformed values. The serum levels of free T3 decreased (15.4%, ^=0.0A1, figure BR.3. Id) in the placebo arm of the study. There were no statistical changes in T3 levels with r-hGH therapy, though a trend existed (9.5-\5.2Vo, figure BR.3.Id) for T3 levels to be reduced as compared with the baseline untreated group. Expressing data relative to the baseline GH-deficient group revealed a similarly significant decrease of T3 in the placebo group (data not graphically presented, p=0.013). The reduction in T3 observed in the placebo group, as compared with the baseline group, was particularly prominent for female patients (23.7%, p=0.043, figure BR.3.Id). Following two-way ANOVA analysis, there were no interactions found between gender and r-hGH therapy. Age of subjects was found not to be a significant variable for T3 levels (figure BR.3.le). Data were examined with regard to those patients on maintenance T4 replacement. There were no significant differences in T3 levels overall between those patients receiving T4 and those without. Higher levels of T3 in the baseline GH-deficient group, as compared with the placebo group, were observed for patients on T4 substitution (20.5%, p=0.041,//gwre BR.3.If). The latter changes could not be ascertained for euthyroid subjects, due to small numbers (figure BR.3.If). The placebo group, for patients on T4 replacement, appeared to exhibit lower levels of T3 than the placebo group in the euthyroid subjects, however, this was not a significant finding (7%, t-test p=0.182). A positive correlation was observed overall between serum T3 and T4 levels (Spearman correlation=0.353, p=0.006). A distinct gender difference was observed between patients on T4 replacement and those without. Hence, though euthyroid patients displayed higher levels of T3 in males than females (18.8%, p=0.0\9, figure BR.3.1g), those on T4 maintenance were shown to have greater levels of T3 in females (12.6%, ^=0.016, figure BR.3.1g).

55 Summary ► T3 concentrations are reduced in the placebo period as compared to the baseline levels and, following r-hGH therapy, values appear to remain below the baseline group. ► The reduction of T3 levels in the placebo group is particularly prominent in females. ► There is a positive correlation between serum free T3 and T4 values. ► For patients receiving T4 replacement, T3 levels are higher in females as compared to males. Conversely, patients not receiving T4 substitution display greater concentration of T3 in males as compared to females.

BR.3.2. Serum free thyroxine Following data distribution analysis, a normal distribution of the data were not established {figure BR.3.2a). Square root transformation of the data resulted in a near normal distribution (Shapiro-Wilk statistic=0.894,/zgz/rg BR.3.2b). Original data are presented in the figures below; however, statistical analysis was carried out on square root transformed values. The levels of T4 were higher in the baseline GH-deficient group (figure BR.3.2c) as compared to placebo (21.3%, t-test p=0.073), six month r- hGH treated (24.8%, p=0.0015), and twelve month r-hGH treated (40.3%, p=0.0015) patients. Male subjects demonstrated reduced levels of T4 in the r-hGH treated group as compared to both the baseline subjects (31%, i^=Q.QOA\ figure BR.3.2d) and the placebo group (12%, t-test p=0.019). In females, after data were expressed as a percentage of the baseline group, both placebo (11%, t-test p=0.027) and r-hGH treated subjects (11%, t-test p=0.028) exhibited reduced concentration of T4 as compared to the baseline patients (data not graphically presented for percent baseline; see figure BR.3.2d). Age of subjects was shown to influence T4 levels. Patients in the 25-34y bracket, and those in the 45-53y bracket, had greater T4 levels than those in the 35-44y group (30.2% and 33.8% respectively, figure BR.3.le). Two way ANOVA analysis failed to reveal any interaction between either r-hGH treatment and age or gender, or between age and gender against each other.

56 Examination of the data to distinguish those patients on T4 replacement and those without was carried out. Patients on T4 maintenance had a greater level of this hormone in their serum (26.9%, figure BR.3.2e) than in euthyroid patients. Higher levels of serum T4 in the baseline GH-deficient group, as compared with all other groups, were observed for patients on T4 substitution (27.4-46.3%, p<0.05, figure BR.3.2e), Euthyroid subjects displayed similar changes, though of a reduced amplitude; however, these results were only significant for r-hGH treated subjects (13.8-22.3%, p=0.051, figure BR.3.2e). For patients on T4 replacement, females exhibited greater (22.1%, \)=0.024, figure BR.3.2J) levels of serum T4 than males, an effect not observed for patients not receiving T4 substitution. A positive correlation was observed between serum T3 and T4 levels (Spearman correlation=0.353, p=0.006). Conversely, a negative correlation was observed for T4 and serum IGF-I (Spearman correlation=-0.53, p<0.001).

Summary ► T4 levels are higher in the baseline GH-deficient subjects as compared to the r- hGH treated and possibly placebo group. ► Both males and females exhibit reduced T4 concentrations following r-hGH therapy, as compared to the baseline subjects. ► Only females exhibit significantly lower T4 values after six months placebo. ► Patients in the 25-34y bracket, and those aged 45-53y, have higher levels of T4 than subjects aged 35-44y. ► Patients on T4 substitution exhibit greater circulating levels of this hormone as compared with subjects not on T4 therapy. ► The reduction in T4 levels following r-hGH treatment is particularly prominent in patients receiving T4 substitution. ► Only patients on T4 therapy exhibit a statistical reduction in T4 values following placebo administration. ► Females have greater levels than men of T4 in patients undergoing T4 maintenance. ► There is a positive correlation between serum free T3 and T4 values and a negative correlation between T4 and serum IGF-I levels.

57 %

12 c Q) 10 CL 8 0) 3E C

0 5 6 7 e 9 10 T3 pmol/l observed expected (normal)

Figure BR.3.1a: Graph showing data distribution for triiodothyronine (T3, pmol/l) Samples taken from patients at the three trial stages. Expected normal distribution also shown.

u (/) c 12 0) re to Q-

0) 8 3E C 8

0.6 0 .7 0.8 0.8 1.0 log T3 (corrected) observed expected (normal)

Figure BR3.1b: Graph showing data distribution for triiodothyronine (T3), following logarithm transformation. Samples taken from patients at the three trial stages. Expected normal distribution also shown.

58 10

0 1 6 COI s (A

I ^ GHD-Om P-6m GH-6m GH-12m GH-6/12m ]

Figure BR.3.1c: Histogram showing triiodothyronine (T3, pmol/l) concentration, GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/l2m =patients after six or twelve months r-hGH Error bars show SEM, *p<0.05.

o E Q. 6 CO I- z 4

0 male female I GHD-Om P-6m GH-6/12m ]

Figure BR.3.Id; Histogram showing gender differences for triiodothyronine (T3, pmol/l) concentration GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH Error bars show SEM, *p<0.05.

59 25

I- c o 15 Ic

E E %

T3 T4

r # 2W4y % 3&44y ^ 45-53y |

Figure BR.3.le: Histogram showing age differences for triiodothyronine (T3, pmol/l) and thyroxine (T4, pmol/l) concentrations. Error bars show SEM, ***p<0.001.

0 E Q. HCO 1 E 2 4) (/>

no T4 replacement T4 replacement [ ^ P-6tn GH-6m GH-12m g GH-6/12m |GHD-Om

Figure BR.3.If: Histogram showing triiodothyronine (T3, pmol/l) concentration for euthyroid patients and those receiving thyroxine (T4) replacement GHD-Om-patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH Error bars show SEM, *p<0.05.

60 10

0 Ë 6 P

I- 1

no T4 replacement T4 replacement

I maie ^ female |

Figure BR.3.1 g; Histogram showing gender differences for triiodothyronine (T3, pmol/l) concentration, patients divided into euthyroid and those receiving thyroxine (T4) replacement. Error bars show SEM, *p<0.05.

61 22

m 14 Q. 12

JH 10 E 3 8 C

0 10 20 30 40 T4 pmol/l Observed expected (normal)

Figure BR.3.2a: Graph showing data distribution for thyroxine (T4, pmol/l). Samples from patients at the three trial stages. Expected normal distribution also shown.

ro 12 CL

10 XJ0) E 8 C3

0 3 5 6 T4 (corrected) Observed expected (normal)

Figure BR3.2b; Graph showing data distribution for thyroxine (T4) following square root correction. Samples taken from patients at the three trial stages. Expected normal distribution also shown.

62 [ % GHDOm P-«m GH-6m GH-12m E GH-6/12m |

Figure BR.3.2c: Histogram showing thyroxine (T4, pmol/l) concentration, GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH Error bars show SEM, ***p<0.005.

25 ***

=6 20

0 male female

79 GHDOm P-6m y GH-6/12m |

Figure BR.3.2d: Histogram showing gender differences for thyroxine (T4, pmol/l) concentration. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH Error bars show SEM, ***p<0.005, *p<0.05; dotted line refers to significant differences with data expressed as a percentage of the baseline group (see text BR.3.2).

63 no T4 replacement T4 replacement [ GHD-Om P-«m GhWm GH-12m g GH-6/12m j

Figure BR3.2e: Histogram showing thyroxine (T4, pmol/l) concentration for euthyroid patients and those receiving thyroxine replacement. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH. Error bars show SEM, ***p<0.005, *p<0.05.

Ô 20 E a. 15

E 2 10

no T4 replacement T4 replacement Œ. male femfWe 1

Figure BR3.2f: Histogram showing gender differences for thyroxine (T4, pmol/l) concentration; patients divided into euthyroid subjects and those receiving thyroxine replacement Error bars show SEM, *p<0.05.

64 BR.4. Urinary total pyridinoline crosslinks Data distribution analysis revealed a non-normal distribution for total pyridinoline (figure BR4a). Logarithm transformation of the data resulted in a near normal distribution (Shapiro-Wilk statistic=0.907,//gwrg BR.4b). Original data is presented in the figures below; however, statistical analysis was carried out on logarithm transformed data. Following six months r-hGH therapy there was a significant increase in total pyridinoline over the baseline GH-deficient and placebo groups (48.6% and 46.9% respectively, figure BR.4c). There appeared to be a plateauing effect at twelve months of r-hGH therapy for total pyridinoline. However, the combined r- hGH treated group (six and twelve months) displayed an overall increase over the baseline GH-deficient and placebo groups (47.1% and 45.4% respectively, p<0.001, figure BR. 4c). On examination of gender differences, the increase in total pyridinoline observed overall, on r-hGH treatment, appeared to be demonstrated in males only (59.3% for GH-Om, 64.1% for P-6m, p<0.001,//gwrg BR.4d). However, expression of the data as a percentage of the baseline patient values, showed an increase in both sexes for total pyridinoline (males 54-55%, p<0.001; females 39-40%, p=0.043,//gwre BR.4e). Pyridinoline levels were also shown to be influenced by age of the patient. Trial subjects in the 25-34y group and those in the 35-44y bracket displayed a greater (25-34y 11.5%, t-test p=0.022; 35-44y 13.7%, figure BR.4f) concentration of pyridinoline than patients aged 45-53y. Analysis of data by two way ANOVA failed to reveal any interaction between r-hGH treatment and gender or age, or between age and gender against each other. There was a positive correlation between total pyridinoline levels and serum IGF-I (Spearman correlation for serum IGF-I =0.415, p=0.001) and a negative correlation with T4 values (Spearman correlation for T4 =-0.35, p=0.006). ANOVA analysis failed to reveal any link between total pyridinoline levels and either subjects on T4 replacement or wishing to continue on r-hGH substitution therapy.

Summary > Total pyridinoline levels are elevated on r-hGH therapy as compared to baseline and placebo groups.

65 Both males and females demonstrate an increase in pyridinoline concentration following r-hGH substitution. The older subjects (45-53y) have reduced levels of total pyridinoline as compared to younger patients (25-44y). A positive correlation exists between serum IGF-I and pyridinoline levels, whilst a negative correlation for pyridinoline is found with T4 values.

66 22 1

(/) (DC

Q. O m =jE c

0 40 80 120 total pyridinoline(nmol/mmolCr) observed expected (normal)

Figure BR.4a: Graph showing data distribution for urinary pyridinoline (total crosslinks in nmol/mmol creatinine). Samples from patients at the three trial stages. Expected normal distribution also shown.

(/) c 0) CD Q.

0) E C3

total pyridinoline (corrected) observed expected (normal)

Figure BR.4b: Graph showing data distribution for urinary pyridinoline (total crosslinks) following logarithm correction. Samples taken from patients at the three trial stages Expected normal distribution also shown.

67 GHCXOm 1:^ P-6m GH-6m GH-12m g GH-6/12m |

Figure BR4c; Histogram showing urinary pyridinoline (total crosslinks in nmol/mmol creatinine). GHD-Oni=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH Error bars show SEM, ***p<0.001.

male female

I / y GHD-Om ^ P-€m | g GH-6/12m |

Figure BR.4d: Histogram showing gender differences for urinary pyridinoline (total crosslinks in nmol/mmol creatinine) concentration GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH. Error bars show SEM, ***p<0.001.

68 250 ***

200

■E 150

male female

GHCXkn P-6m ^ G h^12m |

Figure BR.4e: Histogram showing gender differences for total urinary pyridinoline crosslinks expressed as a percentage of the baseline group GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH Error bars show SEM, *p<0.05, ***p<0.001.

I ^ 25-34y ^ 45~53y |

Figure BR.4f; Histogram showing age differences for urinary pyridinoline (total crosslinks in nmol/mmol creatinine) concentration. Error bars show SEM, *p<0.05, **p<0.01.

69 Chapter 2 Discussion

BD.l. Patient responses to recombinant human growth hormone therapy The normal treatment regime consisting of daily subcutaneous r-hGH injections in the evening results in a serum IGF-I profile which is intermediate between pulsatile and continuous delivery The mean dosage of 3.0 lU/day (see BM.2.) was above the 2.0 lU/day ceiling suggested by De Boer et al The latter group identify doses of r-hGH above 2.0 lU/day as constituting supraphysiologic replacement, and attribute most symptoms of fluid retention to dosages above 4 lU/day. Hence, the fluid retention exhibited by trial patients can be ascribed to the high dose of r-hGH utilised in the present investigation.

BD.2. Serum insulin-like growth factor I Serum IGF-I levels in the GH-deficient subjects before r-hGH therapy were comparable with those in other studies on GHD (see However, IGF-I concentrations before treatment were not below the normal range (normal range 114-314ng/ml; see 22,i88,2i6,2i7y results from the present study reinforce the increased awareness that serum IGF-I cannot be regarded as a completely reliable index of the GH-deficient status of the individual Thorner et al suggest that the weak relationship between GHD status and circulating IGF-I levels may be due to the multifactorial regulation of IGF-I. Following r-hGH therapy, a prominent increase in serum IGF-I occurred which resulted in levels in the upper normal range. This elevation in serum IGF-I to values approaching supraphysiologic concentrations (mean IGF-I in ACRs 420ng/ml 712ng/ml is indicative of the need to carefully monitor the IGF-I profile during clinical trials on GH-replacement. Daughaday,W.H. and Rotwein,P. have shown serum IGF-I levels in males (mean 184ng/ml) to be greater than those in females (mean 120ng/ml) in I8-40y normal subjects. Similar values for serum IGF-I were obtained prior to treatment in this study, with the gender difference maintained during r-hGH substitution. Johannsson et al demonstrated a 33.5% increase in serum IGF-I levels overall on r-hGH therapy, and reported that this elevation was more pronounced in men. Though results concerning

70 the latter finding were not shown, this sex difference was apparently maintained even after adjusting values for body composition Further aspects to gender-specific responses during r-hGH therapy will be evaluated in the general discussion. The reduction in serum IGF-I levels in the post-45 y bracket, as compared to the 35- 44y group, confirms the known reductions in serum IGF-I levels with age (see chapter 4\ DI.1.2 & table DI.l). Hence, after peak levels of IGF-I are achieved during puberty, there is a gradual decline in serum IGF-I levels with age However, the lower serum IGF-I levels in the 25-34y bracket, as compared to the 35-44y bracket, are against the expected pattern. Apart from Johannsson et al who reported an inverse correlation between age and increases in lean body mass, there do not appear to be many studies which have investigated age-related responsiveness to r-hGH therapy in GH-deficient adults. The peculiar finding of lower IGF-I levels in the 25-34y age group can be understood as a general feature of GHD. An alternative explanation may be that skewness in the patient age brackets occurred during the course of the study. The mean serum IGF-I levels in the 25-34y age group were in the low-normal range for serum IGF-I, whilst those in the 35-44y bracket were in the high-normal range. Hence, if a number of patients receiving r- hGH therapy crossed from the 25-34y bracket to the 35-44y group during the course of the trial, then this would reflect in the IGF-I profile. An examination of patient characteristics indicated only one patient made such a transition, thus, this explanation is not sufficient. There is the consideration, however, concerning amalgamation of all data points, at differing trial stages, into one analysis. Despite, two-way ANOVA indicating otherwise, the possibility of r-hGH treatment influencing parameters in this type of analysis cannot be dismissed. Further, due to small patient numbers, these analyses are only practical if patient values are incorporated from each of the trial stages. There is a concern, with the latter methodology, of unduly weighting individual patient responses which may not be reflective of the overall situation. The data were thus re-examined with age groups assessed according to patient GH status. Only patients before r-hGH therapy exhibited significantly lower levels of serum IGF-I in the 25-34y group as compared to the 35-44y subjects (32%, p =0.048). It can be assumed, therefore, for trial subjects in the present investigation, that the severity of GHD (if serum IGF-I levels

71 are taken as an indicator), is particularly prominent in the 25-34y group. Further studies are warranted to establish if the latter result was particular to this group of patients, or, whether it constitutes a normal feature of adult GHD.

BD.3. Serum free 3,5,3-triiodothyronine and free thyroxine Both serum free T3 and T4 concentrations exhibited a similar decline following r-hGH therapy when compared to baseline levels. This finding was in contrast to several reports which suggest that increased extrathyroidal conversion of T4 to T3 occurs following r-hGH therapy in adult GH-deficient subjects 10-18,22,202,207,208,222 latter studies have indicated that the enhanced conversion rate of T4 to T3 is mediated through an increase in levels of IGF-I. Conversely, it has been shown that adult hypopituitary patients with T4-substituted central hypothyroidism, exhibit low T3 levels despite T4 concentrations well within the normal range, and following r-hGH therapy T3 levels are normalised An alternative process in subjects with intact pituitaries may involve SS feedback inhibition of pituitary thyroid stimulating hormone production However, the precise mechanism for GH influence on thyroid hormones is yet to be elucidated. Further, in the present study, not only was there a decline in both T3 and T4 levels with r-hGH substitution, but this reduction appeared to occur in the placebo period as well. Examination of the data revealed two apparent influencing variables. Females appeared to reflect the decrease in T3 and T4 rather than males. In addition, patients on T4 substitution showed a reduction in T3 and T4 which was not observed in euthyroid subjects. Jorgensen et al ^ also demonstrated, for free T4 and T3 levels in their group of GH- deficient adults, more marked changes in patients on T4 replacement. Hence, following four months r-hGH therapy, they noted decreases in free T4 of 6% (euthyroid) and 24% (on T4 replacement), and increases in free T3 of 10% (euthyroid) and 16% (on T4 replacement). It can thus be suggested that the primary reason for declines in free T4 and T3 levels during the placebo period were due to altered sensitivity towards T4 substitution. An influence of sex hormones on this process can also be postulated given the reduction in females only, during the placebo period, of both T4 and T3 levels. Also relevant to the latter suggestion are the greater levels of T3 and T4 in female patients on T4 replacement.

72 The finding of reduced T4 levels in the 35-44y age group probably reflects the negative correlation of T4 with serum IGF-I values; the latter being elevated in subjects aged 35-44y (cf. figure BR.3.le 8l figure BR.le). What are the physiological consequences of the overall changes in thyroid hormones? The initial finding by Lippe et al of declining free T4 and T3 levels in GH-deficient children, were assumed to reflect a GH-mediated suppression of thyroid function. However, this notion was difficult to reconcile with the growth promoting and calorigenic actions of GH, some of which are believed to be inter-related with thyroid hormones function Subsequent studies have shown that free T3 levels are increased rather than decreased by r-hGH treatment, with T4 concentrations demonstrating a decline. Monson et al have reconciled these apparent discrepancies by suggesting that GH-induced elevated T3 levels only occur in patients with an intact thyroid stimulating hormone reserve, whilst subjects with hypothyroidism showed a less marked fall in T4 and no alteration in T3 levels. Of relevance to the present investigation may be the observation that fluctuations in levels of T3 and T4, including for patients on T4 replacement, occurred within the normal range. A decline in free T4 below 10 pmol/l or fi-ee T3 levels below 4.5 pmol/l would be considered to be indicators of hypothyroid status Therefore, the explanation by Monson et al would not appear to apply in this study given the presence of normal T4 levels. Part of the alterations may have been due to thyroid hormones BP influences which would have become apparent from an examination of the total T4 and/or T3 profile (not carried out in this study). A further explanation may lie in the kinetics of GH action on thyroid hormones. De Boer et al have surmised from the available evidence that maximum inducement of thyroid hormones, following r-hGH therapy, occurs within six weeks and that after this period thyroid hormone levels tend to decline. Though Salomon et al demonstrated maximal levels of free T4 and T3 at one month of r-hGH therapy, the other studies mentioned by Boer et al fail to confirm these data. Bengtsson et al were able to show that free T3 levels remained elevated over a six and twenty-six week treatment period, whilst free T4 concentrations were reduced at six weeks and had returned to pre-treatment levels by twenty-six weeks. A similar picture was shown by Binnerts et al with T4 levels reduced at six months and T3 levels elevated. Nevertheless, it is apparent that more

73 frequent evaluations of thyroid hormone levels during the course of r-hGH substitution trials, will provide a more complete understanding of the interactions between GH and thyroid hormones. In summary, the decline in T4 and T3 concentrations following six and twelve months of r-hGH therapy, can be ascribed to increased utilisation of these hormones. These decreases were within the normal range and may also indicate a trend towards normalisation as suggested by De Boer et al The parallel reduction in these hormones during the placebo period may be due to, as yet unclear, interactions between administered T4 and sex hormones. The relatively high-normal levels of thyroid hormones in all trial patients at the start of the trial, may have been why further enhancement of T3 concentration did not occur during the course of this investigation, though measurements at more frequent time points may have provided evidence of a transient, IGF-I mediated, T4 to T3 conversion.

BD.4. Urinary total pyridinoline crosslinks The concentration of total pyridinoline in the urine of adult GH-deficient subjects was low as compared to control values obtained using the same protocol (mean baseline =26nmol/mmolCr; mean control =37.5nmol/mmolCr) and increased following r-hGH therapy. Bone resorption is known to reflect increased bone turnover and hence total pyridinium assays can serve as indices of both processes; however, assessment of skeletal metabolism should include a combination of different markers so that the balance between formative and resorptive events can be adequately evaluated. Unfortunately, only pyridinium cross-links were performed as part of the present investigation; the inclusion of bone formation markers, such as serum alkaline phosphatase and osteocalcin, could have provided complementary information. Since r-hGH administration is known to stimulate both bone formation and resorption, thus, it can be assumed that the present investigation has confirmed previous observations concerning the beneficial effects of r-hGH therapy on bone turnover 18,28,211,225,226 However, no changes in bone mineral density (data not shown) were observed in this study following r-hGH therapy. Bone mineral density is known to be low in adult GH- deficient subjects Wuster,C. has interpreted the failure to demonstrate increases in bone mineral density in studies with six months r-hGH substitution, as a

74 consequence of the increased bone resorption. Further, Vandeweghe,M. points out that each cycle of bone turnover takes about six months, and suggests that observations on four or five cycles is needed before the effect on bone mass can be assessed. As such, only after thirty months of r-hGH therapy was bone mineral density increased in childhood-onset GH-deficient adults Similarly, both Degerblad et al ^^^and Rosen et al ^ were able to demonstrate significant increases in bone mineral density in adult- onset GH-deficient patients only after eighteen (and not after twelve) months r-hGH treatment. Total pyridinoline levels were similar between males and females before r-hGH therapy, and both sexes exhibited a significant increase following r-hGH treatment. However, the increase in males was more pronounced than in females. Peak bone mass in men is 25% to 30% higher than in women If bone resorption increases can be regarded as indicative of a net positive influence on bone formation, then the gender difference in pyridinoline levels may be explained as a result of the known greater increments in lean body mass in male, as compared to female, r-hGH treated subjects The finding of low pyridinoline levels in 45-53y subjects suggests that bone resorption in this age group was lower than that of younger subjects. Peak bone mass is reached between the thirty and forty years of age and is then followed by a gradual loss of bone mass in both sexes This process involves an alteration in the balance of bone turnover, resulting in more bone resorption than formation Due to lack of information on bone formation in the present study, it is difficult to explain the low levels of pyridinoline in older subjects. However, a substantial decrease in the older patients of bone formation markers may have indicated an overall decline in bone turnover in this age bracket. The positive correlation between serum IGF-I levels and pyridinoline concentrations is consistent with GH-mediated bone turnover. However, the precise mechanism (see general discussion, F.2.1.) is believed to partly involve IGF-I action on bone formation and direct GH actions on bone resorption A positive correlation between bone density and serum IGF-I levels is known however, the lack of correlation with T4 in the present study is a little surprising. Both T3 and T4 are known to stimulate bone resorption directly in a dose-dependent manner Hence, hyperthyroidism is

75 associated with an increase in bone turnover The lack of correlation with T4 may relate to the altered influences of GH on shared pathways for bone resorption.

BD.5. Synopsis of the investigation The findings from this chapter include the demonstration of decreased thyroid hormones during the course of the clinical trial. These results indicate the need for careful evaluation of data from this study for the possible influence of these hormones. Further aspects to this investigation concern the positive influences of r-hGH therapy on serum IGF-I and on a marker of bone resorption. Hence, an elevation of circulating IGF-I concentrations has been correlated with increased fimction in a particular target of GH action. It should thus be possible to link further targets of GH with the levels of the known mediator for GH. In so doing, part of the mechanism of GH effects may be clarified. In the subsequent chapters an attempt at evaluating the pathway of GH action will be attempted in skin and muscle, two targets of GH with known functional changes inGHD.

76 CHAPTER 3

AN INVESTIGA TION INTO THE

THERMOREGULATORY APPARATUS IN ADULT

GRO WTH HORMONE DEFICIENCY AND

ACROMEGALY.

77 Chapter 3 Introduction

Cl. The regulation of body temperature CI.1. Central control of thermoregulation Temperature changes in the arterial blood flowing through the preoptic area and anterior hypothalamus are detected by neurons which send impulses to the posterior hypothalamus. The latter area also receives impulses from cutaneous and spinal thermoreceptors allowing the integration and activation of thermoregulatory responses to maintain a constant body temperature. To the concept of a thermoregulatory centre in the hypothalamus needs to be added the more recent discovery of integrative responses involving many other regions of the brain. Thermosensitive neurons may be cold- or warm-sensitive, with low and high positive responses to temperature respectively. The activity of opposite sets of thermo sensitive neurons may be dependent on the degree to which they register differences from a hypothalamic 'set-point' temperature. Alternatively, it has been proposed that these neurons may alter their sensitivity following a change in body temperature. Hence, a rise in mean skin temperature decreases the core temperature threshold of thermal sweating. The converging excitatory and inhibitory influences determine the subsequent pathways of thermogenesis. These can be broadly divided into the behavioural and autonomic responses. Behavioural thermoregulation involves conscious temperature sensations as well as emotional feelings of thermal status and the subsequent steps taken to adjust to the environment. Autonomic responses encompass shivering mediated by somatic motor nerves, non-shivering metabolic thermogenesis mediated by sympathetic and sympathoadrenal pathways, as well as cutaneous blood flow, piloerection, endocrine responses, and sweating.

CI.2. Sweat glands; eccrine and apocrine There are two classes of glands in mammalian skin, the epitrichial glands (associated with hair follicles) which usually secrete by a process of partial (apocrine) or complete (sebaceous) cellular disintegration, and the sweat secreting atrichial or eccrine gland The precise role of the apocrine 'sweat' glands in humans is debatable; their innervation

78 is sympathetic cholinergic but they also have an adrenergic component The apocrine glands secrete a hypotonic fluid which contains protein, reducing sugar and ammonia, and these may provide the components of the human pheromones The principal fimction of the human eccrine sweat glands (ESGs) is thermoregulation with an ability to secrete as much as 1 litre per hour of sweat when exposed to a hot environment, which theoretically allows the dissipation of 675 W of thermal energy 235,237 ^ increase in core temperature is about nine times more efficient than changes in skin temperature in stimulating the sweating aspect of thermogenesis In addition to heat, sweating is also induced by emotional state, exercise, fainting, vomiting and gustatory factors ESGs are distributed throughout the body surface (totalling more than 2x10^). However, those on the palms and soles are primarily responsive to mental, emotional and sensory stimuli Density of the ESGs in adults varies from 64 glands/cm^ on the back, 108 glands/cm^ on the forearm, and 600 glands/cm^ on the palms and soles The actual number of ESGs does not change after birth hence the density of glands decreases with growth The structure of the ESG consists of a tightly spiralled secretory coil in the reticular dermis and a duct which comprises a coiled segment, a straight segment and a corkscrew terminal channel which leads to an opening on the epidermal surface (see p/a^e 1) Secretory coil diameter varies from 60-80pm with a length of 2-5mm in normal adults A myoepithelial layer, consisting of spindle shaped cells, separates the epithelium from the basement membrane, provides structural support in the secretory coil and may be actively involved in sweating responses The pseudostratified columnar epithelium consists of secretory/serous 'clear' cells and mucoid 'dark' cells Sweat is a dilute electrolyte solution that contains mainly NaCl, K, bicarbonate, and inorganic compounds such as lactate, urea and ammonium An isotonic serous precursor of eccrine sweat is derived from the clear cells whilst muco sub stance in eccrine sweat is believed to originate from the dark cells Generation of the sweat precursor requires active transport of sodium from secretory cells involving a Na-K- ATPase and oxidative phosphorylation of plasma glucose though other pathways utilising glycolysis and lactate production also exist With heat acclimatisation, the NaCl concentration of the sweat decreases so that NaCl is conserved

79 The duct has luminal and basal cuboidal cell layers. The luminal cells contain folds and microvilli and are principally involved with sodium and chloride reabsorption from the sweat precursor In cystic fibrosis, an autosomal recessive disorder characterised by dysfunction of epithelial transport processes, a target organ is the ESG Hence, a defect in the ductal segment results in abnormally high levels of sodium chloride in sweat from these patients The normal reabsorption mechanism involves a cAMP dependent pathway which regulates chloride ions and a calcium based mechanism for active transport of sodium ions The major regulatory defect in cystic fibrosis is in the cAMP pathway About 1% of ESGs have branched ductal segments joining several secretory coils

CI.3. Eccrine sweat gland innervation, neurotransmitters and functional tests Central control of sweating is mediated via efferents from the hypothalamus to the medulla which mostly cross at the level of pons, thence to the lateral horn of the spinal cord where they synapse onto preganglionic sympathetic neurons. Preganglionic sweat fibres innervating the head, neck and chest are supplied by C8 to T4 (C8 to T1 more important), the arms by T2 and T3 and perhaps T4 to T7, the trunk by T5 to T12, and the legs by T9 to L3. Axons leaving the lateral horn via the ventral roots synapse in the pre- and para- vertebral sympathetic ganglia. The postganglionic axons are nonmyelinated C fibres which join the peripheral mixed nerves, and eventually reach the ESGs."""'"""'""' Nerve fibres to the ESGs branch repeatedly in the skin and innervate the secretory coil 237,249,250 ^ (Jensc nctwork or 'ground plexus' of very fine fibres is found encircling the ESG with individual fibres exhibiting a series of varicosities along their length, these being the prejunctional endings """. Each fibre innervates many secretory coils and every secretory coil is probably innervated by several different sudomotor neurones """. One or two axons accompany the sweat ducts toward the skin surface Acetylcholine (ACh) was recognised as the terminal neurotransmitter in the sympathetic innervation to the sweat glands in 1934 following the demonstration of ACh release on stimulation of nerves to rat footpads. This was subsequently confirmed in humans by the histochemical visualization of acetylcholinesterase (AChE), the enzyme responsible for ACh hydrolysis at nerve terminals More recently, abundant

80 choline acetyltransferase, one of the key enzymes in ACh synthesis, was demonstrated in ESG nerve fibres by immunohistochemical techniques Sweating was found to be modulated by cholinergic agonists or antagonists acting via muscarinic cholinergic terminals (see table CI.l.) with physiological sweating completely inhibited by atropine, a muscarinic cholinergic antagonist ACh is released from nerve terminals close to the ESG capsule and it enters the gland through the basal lamina and interstitial channels Binding of ACh to its muscarinic receptor (M2 subtype) on ESG secretory coils initiates second messenger systems which initially involve phospholipid hydrolysis (via a G-protein and phospholipase C) and the generation of inositol triphosphate The latter stimulate calcium release and subsequent activation of potassium and chloride currents; the secretory effluent fi'om rat ESGs is a high potassium and low sodium fluid Denervation of rat ESGs results in absence of muscarinic agonist-induced sweating However, muscarinic receptor affinity and numbers are unchanged suggesting that cholinergic innervation does not regulate receptor expression Similarly, sweating defects in aged and Alzheimer's disease subjects are not accompanied by changes in the density of ESG muscarinic receptors Hence, Ryer et al suggest that ageing and Alzheimer's disease may involve alterations in second messenger pathways. Alternative sweating responses mediated by spinal- and axon-reflex arcs also exist in humans Transection of the spinal cord produces hypohidrosis below the level of interruption; however, sweating on the non-interrupted side can occur through spinal reflexes This polysynaptic reflex is used in the sympathetic skin response test of dysautonomia The spinal reflex involves diverse afferents, a common efferent pathway through the spinal cord, pre- and post-ganglionic sympathetic fibres and with ESGs as effectors Part of the afferent pathway may involve VTP-releasing thermosensitive cutaneous sensory fibres Axon reflex sweating occurs through ACh binding to postganglionic sudomotor axons via nicotinic cholinergic receptors; these are characterised by their fast response times (1-2 milliseconds) The nerve impulse generated travels antidromically to a branch point and then orthodromically to release ACh at the nerve terminal where it binds to muscarinic receptors Local sweat production at the edge of ulcers, or in inflammatory skin conditions, may be a consequence of stimulation of the axon reflex

81 pathway 237,248,255 sweating induced by local application of nicotine (and some of the ACh response) is also due to axon reflex pathways (see table CI.L). The quantitative sudomotor axon reflex test utilises ACh iontophoresis as an indicator of the integrity of the postganglionic sudomotor axon 248,266,257 Adrenergic innervation to human ESGs was demonstrated by Uno in 1977 using histochemical visualisation of adrenergic nerve fibres and an electron microscopic demonstration of small dense-cored vesicles (300-600Â) to indicate adrenergic terminals. Adrenergic agonists can evoke sweating; however, the maximal secretory rate differs when compared with cholinergic drugs. Hence, in vitro, the sweat rate expressed as a proportion is 4:1:2 for cholinergic (methacoline), a-adrenergic (phenylephrine), and p-adrenergic (isoproterenol) stimulations in monkey palm secretory coils, and 5:1:1 in human ESGs in vivo and in vitro 234,233,245,255 Second messenger systems for these neurotransmitters have been elucidated. For cholinergic and a-adrenergic mechanisms, a critical role of calcium and the phosphatidylinositol-protein kinase cascade in signal transduction has been demonstrated, though other factors may also be involved 234,259 Following incubation of isolated ESGs with isoproterenol, cAMP accumulation occurs. Hence, sweat induction via a p-adrenergic process is dependent on cAMP, though a minor role of calcium in this process has also been observed 2^9,270 addition, methacholine has been shown to markedly augment the isoproterenol induced cAMP accumulation, via a calcium dependent pathway, though it has no effect on cAMP levels by itself In addition to the cholinergic and adrenergic components of sudomotor innervation, evidence for the presence of neuropeptides in the innervation to ESGs has been mainly obtained from immunohistochemistry studies. These include the neuropeptides VIP 272-274^ calcitonin gene-related peptide (CGRP) 275-277 ^ galanin atrial natriuretic peptide neurotensin peptide histidine methionine 279,280^ and substance P (Sub P)

281,282

To elucidate the functional significance of these various transmitters/peptides it is important to understand the changes which occur in the innervation of the ESGs during development.

82 a) Pharmacologic tests Local intradermal/subcutaneous injections or iontophoresis of: Nicotine; stimulates local intra-sympathetic axon reflex (via nicotinic receptors) Methacholine, muscarine or pilocarpine; stimulate ESGs directly (via muscarinic receptors) ACh; stimulates directly and by axon reflex Atropine; inhibits muscarinic cholinergic sweating Phenylephrine; stimulate cc-adrenergic sweating Isoproterenol; stimulate p-adrenergic sweating Prazosin; inhibit a-adrenergic sweating Propanolol ox alprenolol; inhibit p-adrenergic sweating 6-hydroxydopamine or guanethidine; adrenergic neurotoxins b) Histochemical and immunohistochemical methods AChE histochemistry and immunohistochemistry; cholinergic fibres ChAT immunohistochemistry; cholinergic fibres Formaldehyde vapour histofluorescence; noradrenergic fibres Tyrosine hydroxylase {TH) immunohistochemistry; noradrenergic fibres Dopamine P-hydroxylase immunohistochemistry; noradrenergic fibres

Table C I.l. Methods for evaluating sweat gland function and innervation (adapted from Collins,KJ. and Habecker and Landis ^*^).

83 CI.4. Developmental changes in the innervation to eccrine sweat glands Classical studies by Landis et al have established the developmental history of ESG neurotransmitter changes in rat footpads When sympathetic axons first contact ESG primordia at 7 days after birth, they stain for catecholamines and for the noradrenaline synthesising enzymes dopamine p-hydroxylase, and tyrosine hydroxylase (TH) which catalyzes the rate-limiting step in the synthesis of noradrenaline. These fibres display no staining for ChAT or VIP and only limited AChE activity. However, as the glands and their innervation mature there is a gradual increase in staining for ChAT followed by the appearance of both VTPergic and AChE fibres at day 14. A corresponding decrease in staining for adrenergic markers occurs such that by day 21 there are no endogenous catecholamines present, and very few fibres staining for TH and dopamine P-hydroxylase. The increases in cholinergic markers are necessary for nerve-evoked sweating to take place. Several elegant experiments have demonstrated that the switch in neurotransmitter phenotype is mediated by a soluble cholinergic differentiation factor (CDF) secreted by the ESGs. The role of the ESGs, in specifying the properties of their innervating neurons, forms part of the broad neurotrophic hypothesis. Hence, innervated tissues produce a signal for their innervating neurons in development which may result in (i) selective neuronal survival, (ii) determination of neurite growth and size of dendritic arbors, and (iii) specification of neurotransmitters and neuropeptides. It is being increasingly accepted that the survival, maintenance and plasticity of adult and ageing neurones continue to depend on the availability of neurotrophic factors and other target-derived molecules (for reviews see 29o-296y The retrograde specification of ESG innervation by CDF has recently been found to be preceded by anterograde signalling, resulting in the establishment of functional synapses. Hence, stimulation of CDF production is dependent on the release of noradrenaline from sudomotor nerves prior to their switch to a cholinergic phenotype Also, the acquisition and maintenance of secretory function, including muscarinic receptor activation, is conditional on the release of ACh from the glandular innervation, and not VIP or CGRP In humans, ESGs begin to develop around the third month of gestation as a cord of epithelial cells growing from the epidermal ridge on the palms and soles and at about

84 five months on the rest of the body From their initial appearance, the ESGs increase in number and differentiate morphologically until term Though studies on innervation characteristics are limited, an early appearance of neuronal fibres immunoreactive for the general neuronal marker PGP9.5 (PGP) is observed in the subepidermal plexus at six weeks gestation, followed by the appearance of CGRP fibres in the dermis at seven weeks Primitive ESGs are first seen at seventeen weeks with PGP and VTP staining around them. CGRP fibres are now prominent in the subepidermal plexus, as are fibres in the dermis and around blood vessels which are immunoreactive for Sub P and NPY Thus, sensory innervation is assumed to precede autonomic innervation to the skin. However, staining for adrenergic markers or other cholinergic markers was not carried out in the above study. It has been shown in human axillary skin that the appearence of adrenergic innervation may be due to cholinergic nerves which have retained their ability to take up exogenous catecholamines Hence, it can be assumed that the cholinergic innervation in human skin may, like the innervation of rat ESGs, start out as catecholaminergic, though direct evidence is at present lacking in this regard. It is also important to remember the differences between sweating in humans and rats. Hence, though ESGs are present in all mammals, usually in the footpads, most mammals are dependent on their pelage to protect them from heat As such, in humans, ESGs respond to thermal and emotive stimulations; however, only the latter is capable of inducing sweating in rats

Cl.5. Neuropeptides in skin There is both a sensory and an autonomic component to the innervation of mammalian skin. The sensory nerve fibres share certain locations with the autonomic fibres such as blood vessels and hair follicles whilst the epidermal/dermal free or specialised nerve endings are peculiarly of sensory origin Autonomic fibres also innervate lymphatics, sebaceous glands, erector pili muscles and ESGs Though there are some sensory fibres present around ESGs, these do not directly innervate the glands

85 CI.5.1. Vasoactive intestinal polypeptide; distribution and function in the skin The neuropeptide VTP is a 28 amino acid single chain linear peptide which has a wide distribution in the central (cortex, pineal, suprachiasmatic nucleus and pituitary) and peripheral (autonomic parasympathetic and sympathetic) nervous systems The primary vasodilatory function of VTP is due to its relaxation of vascular and non- vascular smooth muscle and in addition VTP induces secretion directly from a number of glands Hence, VTP stimulates pancreatic secretion and salivary secretion by the parotid, submaxillary and lacrimal glands, promotes water and chloride secretion by the intestine and bronchi, and bicarbonate and water secretion by the pancreas There are two classes of the membrane-bound VTP receptor these are coupled to adenylate cyclase and, on stimulation, receptor mediated endocytosis of VIP occurs as well as a rapid elevation in cAMP levels In addition to the adenylate cyclase pathway alternate second messenger systems may utilise inositol phosphate, protein kinase C and calcium generation following VTP receptor stimulation \TP is the most abundant neuropeptide in sudomotor nerves where it is co-localised in the same terminals as ACh, the VTP being present in a few large granules (diameter 700-1000 A) and the ACh in more numerous small agranular vesicles (500Â) The similar distribution of VIP and AChE fibres around ESGs was initially observed in the cat with any difference in the staining pattern attributed to the presence of VIP in storage vesicles within the nerve, whilst AChE was situated on the outside surface of the nerve terminal membrane. Following lumbar sympathectomy both types of fibres disappeared indicating a sympathetic origin. In salivary glands and nasal mucosa, secretion and vasodilation are under cholinergic control by ACh and VIP respectively, such that concomitant release of ACh and VIP from the same nerve terminal results in potentiation of both secretion and vasodilation Coexistence of AChE and VTP was later confirmed in the sudomotor nerves of human ESGs Both neurotensin and VIP fibres were observed in contact with ESG glandular, myoepithelial and occasional duct cells, whilst the demonstration of both classes of VIP receptor on isolated and sectioned human ESGs suggested a direct action on sweating 278 ,280,307-309 ^ similar distribution of VIP and the structurally related peptide PHI (peptide with N-terminal histidine and C-terminal isoleucine amide) was observed around ESGs in human skin and specific binding sites for both were also demonstrated on the secretory cells of ESGs The

86 staining pattern for the VIP receptor on secretory coils showed scattered cells with a broad apical portion facing the lumen and only a narrow process reaching the basal lamina This suggests that myoepithelial cells are not stained for the VIP receptor hence indicating an influence on sweat production for this peptide. In addition, the prominent VIP receptor immunoreactivity on the ductal portion is restricted to the luminal inner cell layer whereas the innervation to the ducts is known to be sparse and obviously restricted to the basal side. However, receptors on the lumen are known to exist for other peptides and adenylate cyclase may also be present on the luminal membrane Contact between VTP positive nerve varicosities or terminals and blood vessels in the dermis was demonstrated by light and electron microscopy in both rats and humans 272,274 size of blood vessel influenced the pattern of innervation, with arteries and veins supplied by VIP and neurotensin fibres, and Sub P supplying microvessels of the papillary dermis. However, small blood vessels around ESGs also stained for VIP In rat lip, cutaneous blood vessels were innervated by fibres containing both AChE and VIP which were shown to be of parasympathetic origin by removal of the otic ganglion There has been some debate on the control of cutaneous vasodilation with increasing speculation on the role of sudomotor neurons. The latter, through their release of VIP, may control vasoconstrictor tone, rather than putative vasodilator nerves whose transmitter has yet to be identified The demonstration of spinal cord excitability in rats following thermal stimuli and intrathecal VIP injection suggested a sensory role for VTP This was also suggested by the presence of VTP in dorsal horn regions associated with sensory afferents innervating the skin Infi-equent VIP fibres were subsequently observed as non-vessel associated nerve fibres which terminated in the dermal papillae without reaching the dermo-epidermal junction. These fibres were presumed to be of sensory origin and were co-localised with Sub P 263,277 ,278,307

Occasional VTP fibres were also found to be present around the hair shaft, in contact with dermal smooth muscle strands and around sebaceous and apocrine glands in hairy skin Merkel cells, which are found in the suprabasal epidermis of hairy skin and represent a slowly acting mechanoreceptor, were found to contain non-neuronal VTP VIP can stimulate TH release in both the cell bodies and axon terminals of

87 noradrenergic neurons; hence, VIPergic nerve endings may modulate catecholamine synthesis These actions are not dependent on calcium levels but are via the cAMP pathway Though most of the above descriptions of VIP distribution are common to many mammals, some studies have suggested a more limited distribution for human skin. Hence, VTP-IR was found to be confined to staining around ESGs and in some fibres in the dermis In addition, VTP-receptor immunoreactivity was not present in human skin around apocrine and sebaceous glands, nor did blood vessels and hair follicles exhibit staining for the VTP-receptor Also, VTP-receptor IR is absent from vascular smooth muscle which receives VIPergic nerves and which responds to VTP in vitro thus suggesting VIP may act through alternative receptor subtypes VIP is known to enhance the affinity of muscarinic receptors for ACh, and VTP release may be under inhibitory control by ACh via muscarinic, pre-synaptic receptors as atropine administration results in increased VIP release In cat submandibular salivary glands, though it has no secretion-inducing activity, VTP enhances the secretory response of ACh, hence this may be a result of increased blood flow but could also be a direct action of VTP on secretory elements including ACh receptors The role of VIP and noradrenaline in sweating was further investigated; neither elicited a sweat response when injected into 16 day rats nor did VTP enhance muscarine-induced sweating In adult rats, however, VIP was shown to enhance adrenergic sweating via a cholinergic mechanism The most prominent biochemical repercussion of both adrenergic and VTPergic innervation to ESGs is the synergistic amplification of cAMP levels, which are also augmented by ACh Hence, the possible effects of VIP on sweating may be inextricably linked with the extent of adrenergic innervation and perhaps the input of the cholinergic nerves on cAMP accumulation. The localisation of adenylate cyclase activity, stimulated by VTP, to plasma membranes of glandular and myoepithelial cells of human ESGs suggests a direct action on these cells Some authorities have limited the role of cAMP accumulation to promoting glandular growth or protecting the cells from abnormal rises in calcium levels In addition, cAMP generation has been linked to the vasodilatory actions of VIP on vascular smooth muscle The defective P- adrenergic sweating in cystic fibrosis could be explained by a failure to generate cAMP.

88 Further, cAMP is thought to regulate chloride channels in the secretory cells as well as in the duct As such, cAMP can be regarded as an important component of cholinergic sweating, aside from its role in p-adrenergic and VIP-induced sweating.

CI.5.2. Other neuropeptides in the skin CGRP, a 37 amino acid peptide, is an alternate gene product of the calcitonin gene, being the most potent vasodilator among endogenous peptides The peptide is present in several cholinergic motor systems and sensory neurons, including the epidermis and dermis as free nerve endings and around blood vessels and hair follicles of rat and human skin 275.277,318 nerve endings and perivascular fibres which contain CGRP may also contain SS, neurokinin A, galanin and Sub P, and are known to be of sensory origin as retrograde tracing experiments indicate their origin from the dorsal root ganglion 276,277 ,301,319 sensory nociceptive fibres, CGRP regulates Sub P levels by inhibiting its degradation NPY and SS are preferentially expressed in noradrenergic sympathetic neurons including around blood vessels of the skin VTP, PHI, Sub P and CGRP are vasodilators and NPY causes vasoconstriction The widespread and diverse distribution in the skin suggests functions of a broader nature for CGRP including (i) sensory roles of mechanoreception, chemosensation and nociception, (ii) efferent modulation of cutaneous blood flow, (iii) local calcium regulation, and hence modulation of calcium dependent functions, such as those of mast cells in injury and epidermal cell differentiation, and (iv) the direct stimulation of fibroblast and epidermal proliferation in normal skin and in wound healing In addition, CGRP may have a role in a peripheral feedback control loop for thermoregulation where cutaneous thermoreceptors may interact with ESGs to maintain body temperature

CGRP-IR is co-localised with VIP in the cholinergic sympathetic fibres around rat and human ESGs 225.277 p ^^s also shown to be present in these fibres in studies on cat skin which were shown to originate in the stellate, lower lumbar and sacral sympathetic ganglia Cell counts in these ganglia indicated that almost ten times as many VIP immunoreactive neurons were present as those staining for CGRP SS immunoreactivity (IR) has been localised to basal cells with luminal processes in the ductal portion of human ESGs prolactin staining to both luminar duct cells and

89 secretory coil cells of human ESGs, and CGRP-IR has also been observed in secretory coil basal cells of rat ESGs with an apex facing the glandular lumen

CI.6. Abnormalities of sudomotor innervation in clinical conditions In cystic fibrosis, the ESGs display an abnormal transmitter content. Hence, VTP, CGRP and NPY innervation is very sparse; however. Sub P levels are normal as is the cutaneous vasodilation These glands are non-responsive to p-adrenergic stimulation but have normal cholinergically evoked sweating despite an abnormal cell permeability which is cholinergic-sensitive In leprosy, cutaneous nerves are damaged resulting in a loss of sweating, thermal and touch sensation and an almost total loss of sensory and autonomic neuropeptides including VIP, NPY, CGRP and Sub P 325,326 Yn Raynaud's phenomenon PGP-IR and CGRP-IR is decreased in the epidermis/dermis whilst in systemic sclerosis ER for PGP, CGRP and VTP is decreased in the skin including sweat glands An increase in VTP-IR has been demonstrated in eczema and psoriasis but not in axillary hyperhidrosis In patients with cholinergic urticaria, muscarinic ACh receptors are reduced around sweat glands in skin taken from unaffected areas whilst a transient increase followed by a decrease has been demonstrated for the muscarinic ACh receptor in skin biopsies of cholinergic wheals The sweating deficiency of diabetes mellitus results from polyneuropathy of the autonomic nervous system hence the sympathetic skin response has been used as a marker for degree of neuropathy Skin biopsies from diabetic patients demonstrate a reduction in immunoreactivity for CGRP (including around sweat glands). Sub P VIP, NPY and PGP

CI.7. Growth hormone and thermoregulation Impaired thermoregulation is now a well-established consequence of GHD yet little attention has been paid to the mechanism by which this may occur. The first study to explore the relationship between GHD and hypohydrosis described patients as having lower sweat production rates than control subjects. During treatment with r- hGH, sweat secretion rates increased In addition, a correlation between serum IGF-I and sweat mass was shown in GH-deficient patients before r-hGH treatment. A role of GH in normal ESG function and hence in thermoregulation was postulated

90 reinforced by reports of unexplained hyperthermia in a number of GH-deficient patients Subsequently, a number of other studies have confirmed the link between GH and thermoregulation. Sweat secretion is reduced in children with GHD ^ and in Laron syndrome Treatment of patients affected by Laron syndrome, with either r-hGH or with r-hlGF-I failed to alter sweating rates. Since Laron syndrome results from GH insensitivity, it is not surprising that r-hGH treatment was ineffective. However, the lack of effect with IGF-I suggests that GH may use alternative pathways to influence thermoregulation. A more comprehensive study in GH-deficient adults investigated both heat exposure and exercise effects; an increase in GH levels was observed in both situations in the control but not patient groups After moderate heat exposure, evaporation was less, heat storage greater, sweat secretion rate lower, and skin temperature greater in the GH-deficient subjects compared to controls. Following exercise, GH-deficient patients attained higher core temperatures without the concurrent drop in skin temperature observed in controls. This was attributable to lack of evaporative cooling in GHD The same group has also reported the demonstration of reduced sweating together with higher core temperatures, following exercise in the heat, in r-hGH treated GH-deficient subjects with normalised levels of serum IGF-I and IGFBP-3 In acromegaly, hyperhydrosis occurs in 50% to 88% of patients and is rapidly reversed following successful transphenoidal adenomectomy, as well as after bromocriptine and octreotide therapy Stimuli that raise GH in normal adult subjects include raising body temperature by external heating exposure to hot air exercise sauna and swimming in different water temperatures An inhibition of GH secretion during cooling has also been demonstrated by some though not all investigators Ruch et al have observed a correlation between GH release and thermoregulatory responses following infusion of cat brains with noradrenaline and L-dopa. Further, cooling of the preoptic anterior hypothalamus raised GH in association with shivering, vasoconstriction and alterations in core body temperature

CI.8. Growth hormone and the skin Patients with isolated growth hormone deficiency exhibit a fine, wrinkled skin similar to that of premature ageing Abramovici et al have investigated this phenomenon

91 and found a reduction in the skin elastin fibres of GH-deficient patients. However, elastin fibre numbers were normal in Laron syndrome subjects, suggesting an effect of GH in the skin without generation of IGF-I Treatment with r-hGH of osteoporotic patients has been shown to stimulate blood vessel proliferation and increase the numbers of mast cells and fibrocytes Further, a re-orientation of collagen and elastin bundles occurs so that they become parallel with the surface of the skin An increase in collagen synthesis, coupled to greater mechanical strength, has been observed in rat skin following r-hGH injection Growth hormone receptors are present on human and animal fibroblasts hence collagen production is believed to be a direct action of this hormone Cook et al have also shown that local IGF-I production occurs following GH-stimulation of cultured human fibroblasts IGF-I derived from fibroblasts also stimulates keratinocyte growth and thus the role of GH and IGF-I in wound healing is under active clinical investigation IGFBPs have been localised in both the epidermis and dermis of human skin. Hence, basal kératinocytes in the epidermis express mRNA for IGFBP-3 whilst dermal papilla express mRNA for both IGFBP-3 and -5 In addition, dermal sheaths around hair follicles express ISH signal for IGFBP -4 and -5, and both sebaceous glands and eccrine sweat gland display message for IGFBP -2 and -4 Patients with acromegaly display an increase in coarse body hair and skin thickness, as well as thickening and hardening of nails Sebum excretion and sweat production are increased as is cutaneous mucinosis Both hypertrophy and hyperplasia of epidermal cells in acromegaly occur However, Matsuoka et al noted a thinning of the epidermis in ACR subjects and attributed most of the increased skin thickness in these patients to elevated deposition of glycosaminoglycans in the dermis.

CI.9. Aims of the investigation It is apparent that a link between thermoregulation and GH can be envisaged with GH secretion an important consequence of elevated internal/external temperature. Alterations of thermoregulation in GH-deficient subjects are accompanied by reduced sweating responses whilst ACRs have elevated sweat production. How may GH influence the sweating apparatus? A central 'feedback' influence of GH on the thermoregulatory centre can be hypothesised. In addition, morphological changes to the

92 ESGs may occur, a suggestion supported by the structural changes which are a feature of GH-deficient and ACR skin (see AI. 3. & CI.8.). Some of the GH effects in the skin are known to be direct. However, ESGs are also under nervous control fi'om their sympathetic innervation. Given the plastic nature of the sudomotor innervation (see CI.4. & CI.6.) it was hypothesised that, irrespective of central or peripheral effects of GH on sweating, an alteration in the innervation to the ESGs may occur. A study was undertaken to examine morphology, function and innervation of ESGs. In addition to the GHD trial subjects, 7 ACR and 12 normal age- and sex-matched control (CON) subjects have been examined. For each of these study groups, ESG periacinar innervation, ESG acinar size, and sweat secretion rates following pilocarpine iontophoresis, have been evaluated. A histochemical method for AChE was employed to visualise sympathetic cholinergic sudomotor fibres. Immunohistochemical techniques were utilised to study distribution of the neuropeptide VIP which is colocalised in sudomotor cholinergic terminals with ACh. Adrenergic nerves to ESGs were immunohistochemically defined with an antibody to TH. Innervation around ESGs and in the dermis/epidermis has been investigated with a general marker of peripheral nerve fibres, protein gene product 9.5 (PGP) ESG acinar size and levels of staining for AChE, VIP, and PGP were determined by computerised image analysis.

93 Plate 1 Photomicrograph of eccrine sweat glands in the dermis of normal male human skin.

A: Secretory coil (sc). B: Secretory coil-duct junction. C: Duct segment (d). All fields stained for PGP.

Note: Periacinar nerves (n). Lipofuscin pigment (1) in secretory coils only. Magnification for all fields =x20. Scale bar =50pm.

94 cl Chapter 3 Methods

CM.l. Patient characteristics Aside from the GH-deficient trial patients (see BM.L), seven subjects (6 males, 1 female; aged 36 to 82 years, mean age 54 years) with active acromegaly (clinically established with elevated IGF-I levels and a failed GH suppression following glucose challenge) and CON subjects (8 males, 4 females; aged 25 to 53 years, mean age 44 years), age and sex matched to the GHD patients, were also selected. Sweat rates were measured and skin biopsies taken from the trial patients at the start of the study, after six months of r-hGH or placebo treatment, and six months after that (six or twelve months r-hGH treatment). ACR and CON subjects had skin biopsies taken once following sweat rate analysis. In addition, 14 extra normal subjects (8 males, 6 females; aged 24 to 57 years, mean age 36 years) were recruited to establish normal values for sweat secretion following pilocarpine iontophoresis.

CM.2. Pilocarpine iontophoresis-induced sweat production The rate of sweat collection on the flexor aspect of the forearm was measured using a Macroduct sweat collection system (Wescor, Model 3700-SYS) following 0.5% pilocarpine nitrate iontophoresis This involved maximal stimulation of the ESGs by an electrical current for 5 min following which the sweat was funnelled into a plastic capillary tube. The distance travelled in this tubing was measured at 5 min intervals for 45 min. The initial rate of sweat production was calculated, following bivariate regression and polynomial curve fitting, from the gradient of the curve (cm min'^).

CM.3. Sweat gland morphology and innervation CM.3.1. Skin biopsy and tissue processing 5mm punch skin biopsies were taken from the flexor aspect of the forearm after local infiltration of 2% lignocaine. Tissue orientation was preserved by laying the skin longitudinally on a piece of card. The skin was fixed in 4% paraformaldehyde for 120 min, washed in Hepes buffer (3xlOmin), and dehydrated overnight in 20% sucrose in

95 Hepes buffer. The tissue was subsequently embedded in a cryoprotectant (O C T compound, BDH Ltd., UK) and frozen in liquid nitrogen. 7pm longitudinal cryostat sections were cut at -25 °C and thaw-mounted onto gelatinized slides (see appendix G.3.; 3-4 sections per slide). Once sections had been dried at room temperature (15- 30min), slides were wrapped in aluminium foil then plastic bags, and stored at -20°C. Slides were warmed up to room temperature prior to unwrapping and subsequent processing

CM.3.2. Acetylcholinesterase histochemistry The 'direct-colouring' thiocholine method for cholinesterase modified in a previous study using a shorter incubation time and inhibitors to control non-specific staining, was used The sections were equilibriated in PBS buffer (2xl0min) then pre-incubated for Ih at room temperature in lO'^M tetraisopropyl-pyrophosphoramide (iso-OMPA in phosphate buffer, see appendix G.3.) to abolish non-specific staining. The sections were then incubated in fresh acetylthiocholine incubation medium (see appendix G.3) for 20 min at 37°C. After thorough rinsing in PBS (4xl0min), sections were counterstained with 0.5% light green in 0.5% acetic acid (10 sec), dehydrated through 40% alcohol (30 sec), 80% alcohol (30 sec), 100% alcohol (2x30 sec), cleared in xylene (2x30 sec) and mounted in DPX after wiping off excess fluid. Control pre-incubations in eserine (lO '^M physotigmine, see appendix G3.) and B.W.284C51 (10"^ 1:5 (bis-allyl-dimethyl-ammonium-phenyl) pentan-3-one-diiodide, see appendix G.3.) resulted in complete abolishing of staining (see plate 2i).

CM.3.3. Immunohistochemistry All stages were carried out at room temperature with an established indirect immunofluorescence procedure (see 227,372-378 ^ polyclonal antibodies utilised in this study were well-characterised for the PNS in the aforementioned studies. Further, IHC with VIP and PGP had been carried out in human skin with a similar protocol by Abdel- Rahman et al Primary layer omission controls were incorporated into every experimental run. Polyclonal antibodies from the same batch were utilised throughout the study, and different trial stage tissue was processed together where possible. Four

96 slides per marker were processed for each biopsy with a thickness of about 500pm separating each group. Sections were equilibriated in Hepes buffer (IxlOmin), then incubated in a humid chamber with 5% swine serum (SOpl per section; 60 min) to block non-specific binding sites. Excess fluid was drawn off slides followed by incubation overnight with rabbit polyclonal antisera raised against TH (Eugene Tech, New Jersey, USA; clone TElOl, 1:100), VTP (Incstar, UK; 1:1000 dilution) and PGP (Ultraclone Ltd, UK; 1:800 dilution). Following incubations, the sections were rinsed in phosphate buffered saline (PBS; 3x10min) and incubated with swine anti-rabbit fluorescein isothiocyanate (Dako Ltd, UK; 1:80 dilution) for 90 min. Both primary and secondary antibody diluant solutions contained triton X-100 (0.1%). The X-100 is a detergent which breaks down lipid components of the cell membrane and hence aids antibody penetration Sections were washed in PBS (3x10min), counterstained for 5 min in 0.05% pontamine sky blue (BDH Chemicals Ltd, UK) to reduce background auto-fluorescence and rinsed again in PBS (2xl0min). Excess fluid was removed, slides were mounted in antifade mountant (Citifluor, UK), and sealed with varnish.

CM.3.4. Microscopy and image analysis for skin innervation CM.3.4.1. Microscopy and image input Densities of AChE, VTP and PGP immunoreactive nerves and stain intensities were measured using established methods of image analysis A Kontron IPS (Kontron, UK) image analyser was interfaced to an Olympus Vanox AH-2 fluorescence microscope by a low-light video camera (model VL350, PCO computer optics, Germany). A x20 oil objective lens was used to visualise appropriate fields which were then transferred into the image analyser. The gain setting and integration time of the camera were set at every session so that the intensity of fluorescent light passing through a uranyl glass standard was always at a maximum gray value of 220. For AChE measurements, transmitted light input was passed through a standard neutral density filter, and the gain setting and integration time adjusted for a maximum gray value of 180.

97 CM.3.4.2. Sweat gland size and innervation measurements Three fields containing ESG secretory coils were selected at random from one section per slide. Acini were counted and the acinar perimeters were measured by interactive drawing on the analyser screen. Autofluorescent lipofuscin (age-pigment) was concentrated in gland cells, distinct from ESG duct segments (see plate 1). Therefore, the defined glandular areas were masked out of the image before subsequent analysis of nerves. Following established procedures of background subtraction, image enhancement and thresholding specific periacinar staining was binarised, measured and expressed as the intercept density per acinus (nerves/acinus). The intercept density per acinus measures the number of nerve fibres that intercept a standard computer-generated grid grid and expresses this after taking account of the numbers of ESG acini associated with the nerves.

CM.3.4.3. Epidermal-dermal junction innervation measurements Quantification of presumably sensory fibres in the upper dermis, and spiralling into the epidermis, was carried out with the same image analysis system as described above (CM.3.4.1.). A computer-generated frame (120pmx250pm) was placed at three standard locations along the section. The top of this frame was placed just underneath the stratum comeum with the length of the frame chosen so as not to allow deeper nerves (eg perivascular nerves) to be included in the measurement. The nerve density was expressed as the intercept density (intercepts per mm) which is a measure of the number of nerves crossing a computer-generated grid.

98 Chapter 3 Results

CR.1. Pilocarpine iontophoresis-induced sweat production The data were found not to follow a normal distribution (figure CR.1 a). Logarithm or square root transformation (see appendix G.l.) failed to normalise data spread; hence, analysis of results was carried out with the non-parametric Kruskal-Wallis (K-W) ANOVA. The initial rate of sweating was reduced (45%, ^=0.03\', figure CR.Ib) in the baseline GH-deficient patients compared to control subjects. Following twelve months of r-hGH therapy, there was a 51% increase in sweating rate, however, this was not statistically significant (figure CR.lb). When data were expressed relative to the baseline GH-deficient group; sweating rates in the twelve month r-hGH treated subjects were greater (57%, 03^5, figure CR.lc) than those in the baseline and placebo groups. Acromegalics unexpectedly showed no evidence of increased sweating over control subjects (figure CR.Ib). Sweat rates for women in the trial groups were less than those in men (females 37.5%

Summary ► The pilocarpine-induced initial rate of sweating is reduced in GH-deficient patients and is restored to control levels afl;er twelve months r-hGH therapy. ► Acromegalic subjects do not have a stimulated rate of sweat production above that of controls.

99 A gender difference exists in sweating rates for the trial subjects, with males demonstrating a higher rate of sweating than females. Following r-hGH treatment, females display a greater increase in sweating rates as compared to males.

100 21 22

20 18

I t re Q. 12 0) JD 10 3E C 8 t 4

0 0 2 3 initial rate of sweating (cm/min) Observed expected (normal)

Figure C R .la: Graph showing data distribution for the initial rate of pilocarpine- induced sweating (cm min’’). Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown.

101 1.2 c 1

0.6 !'o I 0.4 2 c 0.2

GHOOm ^ P-6m M ^ GH-12m E GI+6/12m Ü CON M a c r

Figure CR.lb: Histogram showing the initial rate of pilocarpine-induced sweating. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH, CON=control subjects, ACR=acromegalic subjects. Error bars show SEM, *p<0.05.

350

300

O)250 c

200 0 1 150 +3m 100

0 GHD-Om P-6m GH-12m m

Figure C R lc: Histogram showing the initial rate of sweating; data expressed relative to the baseline GH-deficient group (GHD-Om). Data originally in cm min'V GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH. Error bars show SEM, *p<0.05.

102 O) I 0 0.4 1 <0 0.2 c

male female

GHMm P-6m GH-6/12m H CON

Figure CR.ld; Histogram showing gender differences for the initial rate of sweating (cm min '). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/!2m ^patients after six or twelve months r-hGH, CON=control subjects. Error bars show SEM, *p<0.05.

300

250

E 200

1 150

100

male female

GHD-Om P-6m GH-6/12m

Figure CR.le: Histogram showing gender differences for the initial rate of sweating expressed as a percentage of the baseline GH-deficient group (originally in cm min ') GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH. Error bars show SEM, *p<0.05.

103 CR.2. Acetylcholinesterase histochemistry The data were found not to be normally distributed {figure CR.2d); however, square- root transformation resulted in a near normal data spread (Shapiro-Wilk statistic =0.79, figure CR2b). A significant reduction in periacinar AChE staining was observed in the baseline GH-deficient patients as compared with all other trial groups (21-37%, p

Summary ► Staining for AChE around ESGs is reduced in GH-deficient subjects and is increased to normal control levels following r-hGH therapy. ► Acromegalic patients display greater sudomotor staining for AChE than control subjects. ► Following r-hGH therapy, female GH-deficient subjects have a more marked increase in AChE staining as compared with males. ► AChE stain data is positively correlated to circulating IGF-I levels, and negatively correlated with thyroid hormones.

104 B

Plate 2i Photomicrograph of eccrine sweat glands from normal male human skin stained with acetylcholinesterase or incubated with cholinesterase inhibitor.

A: ESGs stained with AChE. B: ESGs incubated with AChE medium and with BW284C51 (an inhibitor of AChE). Note: A & B photographed with different filters hence the contrasting background colour. Periacinar nerves (arrowhead). Magnification for all fields =x20. Scale bar =50pm.

105 Plate 2ii Photomicrograph of eccrine sweat glands from clinical trial patients stained with acetylcholinesterase.

A: ESGs from a male baseline GH-deficient patient. B: ESGs from a GH-deficient female patient after six months placebo. C: ESGs from a female trial patient after twelve months r-hGH therapy.

Note: Periacinar nerves (arrowhead). ESG size in male ESGs (A) is greater than that in females (B & C). Magnification for all fields =x20. Scale bar =50pm.

106

26 -

c0) (0 CL o w E C3

12 2.0 2 4 2.8 3.2 AChE; intercepts per acinus observed expected (normal)

Figure CR.2a: Graph showing data distribution for AChE histochemical staining expressed as intercepts per acinus. Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown.

24 - (I) 22 - c OJ m Q. o m E =3 C

0 10 20 30 40 AChE (corrected) observed ^ expected (normal)

Figure CR.lb; Graph showing data distribution for AChE histochemical staining following square root correction. Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown.

107 25

***

f20 I 8 15

^10

Q. S I 5 « 'il'nll'ife

GHD-Om ^ P-em 0 GH-6m ^ GH-12m 3 GH-6/12m H CON ■ 1 a c r

Figure CR.2c: Histogram showing AChE histochemical staining expressed as intercepts per acinus (square root corrected). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH, CON=control subjects, ACR=acromegalic subjects. Error bars show SEM, *p<0.05.

o) 10

Î 8 LU i s

11 i i i

GHDOm P-6m Gl+6m GH-12m GH-6/12m

Figure CR.2d; Histogram showing AChE histochemical staining expressed as a percentage of baseline (square root transformed, intercepts per acinus data). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m -patients after six or twelve months r-hGH. Error bars show SEM, *p<0.05.

108 25

? 2 0

c N ® 10 t g JBc 5

male female

GHCXOrn P-6m GH-6/12m m CON

Figure CR2e: Histogram showing gender differences for AChE histochemical staining expressed as intercepts per acinus (square root transformed). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH, CON^control subjects. Error bars show SEM, *p<0.05.

109 CR.3. Vasoactive intestinal polypeptide immunoreactive nerve density Data distribution analysis indicated non-normality {figure CR.3a)', however, distribution was near-normal following square root transformation (Shapiro-Wilk statistic =0.87, 36, figure CR.Sb). Density of periacinar nerves staining for the neuropeptide VIP was reduced in placebo and baseline GH-deficient patients compared to control subjects (22-25%, ^=6.605, figure CR.3c, plate3). Following six or twelve months r-hGH treatment, VIP immunoreactive nerve density increased (12-18%; t-test figure CR.3c, plate3) in the trial subjects to reach levels similar to those in controls. Acromegalic subjects displayed greater VIP periacinar staining than the baseline GH- deficient patients (28%, ^=6.605, figure CR3c, plate3). When males and females were analysed separately, r-hGH treatment showed no significant effects, probably because of the small group sizes (figure CR.3d). However, the percentage increase on r-hGH therapy was greater in females (baseline 18%, placebo 14%) than in males (baseline 10%, placebo 4%). In general, the staining for VIP in males was significantly greater (13%, p=O.OOA, figure CR3d) than that in females and therefore compared favourably with gender differences seen for sweat production. Two-way ANOVA analysis failed to demonstrate an influence of GH status on gender differences for VTP immunoreactivity. VTP immunohistochemistry data displayed a negative correlation with thyroxine levels (Spearman correlation =-0.255, p=0.035) and a strong positive correlation with serum IGF-I values (Spearman correlation =0.513, p<0.001).

Summary ► VIP immunoreactivity is reduced in GH-deficient subjects and is increased towards control values following r-hGH therapy. ► Acromegalics have similar VIP staining to control patients, with greater nerve fibres than in baseline GH-deficient patients. ► Male trial subjects have elevated VIP immunostaining as compared to female trial patients. ► Thyroxine levels are negatively correlated with VIP immunoreactivity, whilst serum IGF-I levels are positively correlated with VIP nerve density.

110 Plate 3 Photomicrograph of eccrine sweat glands from clinical trial, control and acromegalic subjects, stained for VIP.

A: ESGs from a female baseline GH-deficient patient. B: ESGs from a female trial patient after six months r-hGH treatment. C: ESGs from a female control subject. D: ESGs from a male ACR subject.

Notei Periacinar nerves (arrow). ESG size is greater in ACRs when compared to all other groups (cf. male acini in plate 1), Magnification for all fields =x20. Scale bar =50pm.

I l l

24 -I

20 -

0)C ro CL o m JQ 3E C

0 too 200 3 0 0 4 0 0 5 0 0 VIP-IR; intercepts per acinus observed expected (normal)

Figure CR.3a: Graph showing data distribution for VIP immunoreactivity (IR) expressed as intercepts per acinus. Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown.

22 -

(/) c 0)

o nm 3E C

8 t2 to 20 VIP-IR (corrected) Observed expected (normal)

Figure CR.3b: Graph showing VIP-IR following square root correction. Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown.

112 16

i " &10 (A I 8 (0 & 6

8 4 I

GHD-Om ^ P-8m 0 GH-6m ^ GH-12m m GH-6/12m Ü CON ACR

Figure CR.3c; Histogram showing VIP immunohistochemical staining expressed as intercepts per acinus (square root corrected). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH, CON=control subjects, ACR^acromegalic subjects. Error bars show SEM, *p<0.05.

g 8 10 (/> I 8 (D & 6 & § 4 I

male female

GHD-Om P-6m GH-6/12m m CON

Figure CR.3d: Histogram showing gender differences for VIP immunohistochemical staining expressed as intercepts per acinus (square root transformed). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH, CON=control subjects. Error bars show SEM, *p<0.05.

113 CR.4. Protein gene product 9.5 immunoreactive nerve density The data were found to follow a normal distribution (Shapiro-Wilk statistic =0.86; figure CR4a). Density of nerves around ESGs staining for the structural, pan-neuronal marker, PGP, was greater in acromegalic subjects (37-40%, p=0.009; figure CR.4b, plate 4) than in the trial groups, both before and after r-hGH therapy. GH-deficient patients were not different from control subjects either at the begining or end of the trial (figure CR4h). A trend appeared to exist, in the trial and control groups, for PGP- immunoreactive nerve density in males to be greater than in females, however, this was not significant (13%, ^=0.0551; figure CR4c). T-W ANOVA analysis did not reveal any interaction between trial groups and gender. No correlation was observed between PGP immunostaining nerve density and serum IGF-I or thyroid hormone values.

Summary ► There is no difference in total periacinar nerve density, as assessed by PGP, in GH-deficient patients before and after r-hGH treatment. ► Acromegalic subjects display elevated PGP immunoreactivity when compared with trial subjects.

114 Plate 4 Photomicrograph of eccrine sweat glands from GH-deficient and acromegalic subjects, stained for PGP.

A: ESGs from a female baseline GH-deficient patient. B: ESGs from a male ACR subject. C: Hypertrophic ESG from a male ACR subject.

Note: Periacinar nerves (arrow). ESG size is greater in ACRs when compared to all other groups (cf. male acini in plate i). Magnification for all fields =x20. Scale bar =50pm.

115

CLm 12 10 0) 8 DE C 6

0 100 200 300 400 500 600 700 800 PGP-IR; intercepts per acinus observed ^ expected (normal)

Figure CR4a; Graph showing data distribution for PGP-IR expressed as intercepts per acinus. Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown.

116 5.300

E 200

GHDOm ^ P.6m 0 GH-6m GH-12m GH-eiMm ü CON ü ACR

Figure CR.4b: Histogram showing PGP immunohistochemical staining expressed as intercepts per acinus. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH, CON=control subjects, ACR=acromegalic subjects. Error bars show SEM, *p<0.05

500

male female

GHDOm P-6m GH-6/12m H CON

Figure CR.4c: Histogram showing gender differences for PGP immunohistochemical staining expressed as intercepts per acinus. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH, CON=control subjects. Error bars show SEM, *p<0.05.

117 CR.5. Protein gene product 9.5 immunoreactive nerves in the epidermal-dermal junction Distribution of the data were found to be normal (Shapiro-Wilk statistic =0.74, figure CR.5 a). PGP immunostained nerves in the epidermal-dermal junction displayed no difference in intercept density between baseline GH-deficient patients and trial subjects after r-hGH therapy (figure CR.5h). However, in the six months placebo period there was a reduction in PGP nerve density which was significant when compared with the twelve month r-hGH treated group (38%, t-test figure CR.5b, plate 5). There were no gender differences observed (data not shown). A positive correlation was found between thyroid hormone levels and PGP immunoreactive nerve density (Pearson correlation for T3 =0.348, p=0.009, for T4 =0.382, p=0.003). Serum IGF-I levels were not correlated with PGP immunostaining.

Summary ► Epidermal-dermal nerves immunoreactive for PGP display a reduction in density following six months placebo administration, possibly due to progressive effects ofGHD. ► A positive correlation between PGP immunoreactive nerves in the epidermal- dermal region and thyroid hormones exists.

118 Plate 5 Photomicrograph of skin epidermis and dermis staining for PGP and eccrine sweat glands immunoreactive for TH.

A: Epidermal-dermal junction in skin from a trial patient after six months r-hGH therapy, stained for PGP. Note: Nerve fibres (large arrow), nerve bundle (small arrow). Epidermis (e), dermis (d). B: Epidermal-dermal junction in skin from a trial patient after six months placebo, stained for PGP. Note: Nerve fibres (arrow). C: ESGs from a female baseline GH-deficient patient, stained for TH. Note: Nerve fibres (arrow). D: ESGs from a male r-hGH treated subject, stained for TH. Note. Periacinar staining (small arrow), perivascular staining (large arrow). Blood vessel (bv).

Magnification for all fields =x20. Scale bar =50pm.

119

26 -

0)c

Q. O m JD E C3

0 8 12 PGP (dermal-epidermal) observed * expected (normal)

Figure CR.5a; Graph showing data distribution for PGP immunoreactivity at the epidermal-dermal junction Data expressed as the intercept density Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown.

a g 2 Bc

0 GHD-Om ^ P-6m m GH-6m ^ GH-12m m GH-e/12m Ü CON ACR

Figure CR.Sb; Histogram showing epidermal-dermal PGP immunohistochemical staining expressed as the intercept density. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/l2m ^patients after six or twelve months r-hGH, CON=control subjects, ACR=acromegalic subjects. Error bars show SEM, *p<0.05.

120 CR.6. Eccrine sweat gland acinus perimeter The data were not found to be normally distributed (figure CR.6a). Following square root transformation, data achieved a near normal spread (Shapiro-Wilk statistic =0.88, ^=032,figure CR.6b). Acromegalic subjects showed larger (11-17%, p<0.001,//gwrg CR.6c , plates 3 & 4) ESG acini compared to all other groups. There were no significant differences between control and GH-deficient subjects, nor did r-hGH treatment influence acinar size. However, gender had a significant effect on ESG perimeter (p<0.001, figure CR.6d) with acini in males some 8% larger than those in females. Further, in males, a small reduction was observed in acinar size following r- hGH treatment, as compared with the male control group (7%, t-test p=0.034, figure CR.6d). T-W ANOVA suggested an interaction between the clinical trial patients and gender; however, this was not-significant (trial patients x gender =0.061). Acinar perimeter was not correlated with thyroid hormones but a positive correlation was present with serum IGF-I levels (Spearman correlation =0.276, p=0.026).

Summary ► Acinar size is not influenced by GHD or r-hGH treatment. A possible exception may be in male subjects whose secretory coil perimeter is reduced following r- hGH therapy. ► Acromegalics display an elevation in secretory coil perimeter as compared with all other groups. ► Males have larger acinar perimeters than females. ► A positive correlation is found between ESG size and circulating IGF-I levels.

121 22 -

CL o m 3E C

80 120 100 acinar perimeter (micrometers) Observed expected (normal)

Figure CR*6a: Graph showing data distribution for sweat gland acinar perimeter (pm). Samples from trial subjects, controls and acromegalics. Expected normal distribution also shown

20 -

(/) "0) c ro CL o flû E C3

7 8 9 10 11 1213 14 acinar perimeter (corrected) Observed expected (normal)

Figure CR.6b; Graph showing data distribution for sweat gland acinar perimeter following square root transformation. Samples from trial subjects, controls and acromegalics Expected normal distribution also shown.

122 12

1 .

S0) 6 E •c

ro •e(ü

GHDOm GH-6m ^ GH-12m a GH-6/12m Ü CON 0 ACR

Figure CR.6c: Histogram showing sweat gland acinar perimeter (square root transformed, originally in pm). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m ^patients after six or twelve months r-hGH, CON=control subjects, ACR=acromegalic subjects. Error bars show SEM, ***p<0.001

12 ***

I 8 I I 6 E

^Q. 4

male female

GHD-Om P-6m GH-6/12m m CON

Figure CR6d; Histogram showing gender differences for sweat gland acinar perimeter (square root transformed, originally pm). GHD-Om^patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH, CON=control subjects. Error bars show SEM, *p<0.05.

123 CR.7. Tyrosine hydroxylase distribution Immunoreactivity for TH was observed around blood vessels of the skin (see plate 5). Occasional TH fibres were observed around ESG secretory coils (see plate 5). These were not of a level to allow quantification and did not show obvious differences between trial, control or acromegalic groups.

124 Chapter 3 Discussion

The influence of deficiency, or excess of GH, on thermoregulation has been explored in this study. The effectiveness of treatment in reversing sweating deficits in adult GHD patients has also been investigated. Evidence is provided to suggest that hypohidrosis in GHD is accompanied by a decrease in the functional innervation to the ESGs, whilst in acromegaly, hypertrophy of ESGs and an increase in total innervation occurs. Following r-hGH therapy in GH-deficient adults, an increase in the sweat rate occurs which is accompanied by elevations in AChE- and VIP-stained nerves, indicating GH- dependent plasticity of the sudomotor nerves. Gender differences have been observed in sweating and in the sudomotor apparatus. Possible mechanisms of GH action will be explored below whilst other influences such as thyroid hormones, IGF-I, and possible gender-specific responses to GH, will be evaluated in the general discussion.

CD.l. Growth hormone deficiency; sweating and skin innervation Reduced sweating in adult GHD has been demonstrated in several studies by Skakkebaek and colleagues 335,339,391.382 sweating rates following pilocarpine iontophoresis 33-45% lower than in control subjects. Data from the present investigation has confirmed this sweat deficit in GHD with similar figures. The only other study to examine the pilocarpine response, before and after r-hGH treatment in GH-deficient patients noted that sweat production in their placebo group were 39% lower than control values. Following four months r-hGH treatment, levels increased non-significantly by 22%, however, patients noticed an increase in sweating during r- hGH treatment This discrepancy between observed and subjective assessments of sweating suggests that the sweating assay used by Skakkebaek and colleagues 335,339,381,382 not have been of sufiScient sensitivity. This is reflected in the finding that the pilocarpine method employed in their investigations is subject to substantial day-to- day variations of up to 20% Further, Hjortskov et al showed, in GH-deficient and control subjects, that the amount of sweat collected after exercise and heat exposure was 61% greater than that obtained by pilocarpine iontophoresis whilst resting at room

125 temperature The inference made from these observations was that the pilocarpine test did not induce maximal sweating. However, the low level of sweating in GH- deficient patients, both before and after treatment as compared to controls, suggests that the relative insensitivity of the pilocarpine method may not be entirely responsible for the lack of change observed in these investigations following r-hGH therapy 335,339,381,382 methodology in the latter studies involved sweat collection on filter papers, which was expressed as a cumulative total at the end of 30min. In the present investigation, the calculation of the initial rate of sweat production probably constitutes a better method of evaluating sweat output as it seems to be responsive to the most efficient period of ESG sweat production. An aspect of sweating that has received little attention in the context of GHD is the autonomic nerve supply of ESGs. The sudomotor nerves influence sweating through a cholinergic effect on acinar activity, combined with a local vasodilation (see CL 3.). VIP is a potent vasodilator and enhances acetlycholine-receptor binding both of which stimulate sweat secretion (see Cl. 5.1.) 234,316,383 AChE and VTP staining appear to provide a sensitive index of the effects of GHD and of replacement therapy. Thus, AChE- and VIP-stained nerves were reduced in the GHD subjects compared to the control and ACR groups. Following r-hGH treatment, levels of AChE-positive and VIPergic nerves increased to values comparable with the controls. In contrast, adrenergic nerves, as assessed by TH immunostaining, were limited to occasional fibres around ESGs, both before and after r-hGH therapy. Hence, the hypothesis that VTP is linked to the adrenergic component of sudomotor nerves (see Cl. 5.1.) appears unlikely. Based on the increased sweating rate and parallelled elevations in AChE and VTP levels, it appears that the function of VTP is inextricably linked to cholinergic regulation of sweating. The demonstration of a reduction in dermal/epidermal innervation during the placebo period suggests that extended periods of GHD can result in differential changes in separate nerve populations. Hence, sympathetic nerve fibres around sweat glands did not display these alterations. The basis for these differences in presumptive sensory nerves may relate to the reduced elasticity and thinner skin in growth hormone deficient subjects Replacement therapy with r-hGH in aged subjects has been shown to reverse age-related declines in epidermal thickness However, alterations in the

126 skin are common features of both hypothyroidism and hyperthyroidism Hence, oedema, epidermal thickening/thinning and hyperkeratosis occurs in the skin of hypothyroid patients, whilst increased cutaneous vasodilation and glycosaminoglycan accumulation is seen in hyperthyroidism Since the changes in dermal/epidermal innervation were positively correlated with thyroxine levels, it is tempting to speculate that the reduced levels of thyroid hormones in the placebo period may have influenced innervation to the skin. The mechanism for this process warrants further investigation.

CD.2. Acromegaly; sweating, sweat gland size and innervation Classic studies by Sato and Sato established a positive correlation between the size of SG secretory coil units and their activity in vitro for normal subjects. Though numbers of individuals examined were small, a difference in ESG acinar diameter of -25% appeared to differentiate, in that study, between 'slow' and 'fast' sweaters. Our observation that secretory coil diameter in ACRs was 11-17% greater than other groups suggest a similar correlation between gland size in the theoretically more active sweating group. Though earlier workers have hinted at larger ESGs in acromegaly the present study is the first to provide data confirming such a change. However, no increase in sweating over controls was demonstrated in these subjects, generally considered hyperhidrotic Lack of hyperhidrosis may reflect the methodological problems associated with the pilocarpine method (see CD. 1.), particularly as reports in the literature of excess sweating in ACRs are from visual, or patient questionnaire, assessments. Also, the cutaneous changes seen in acromegaly, including classic dermal thickening may have interfered with the normal assumptions for the pilocarpine method which presume a preferential transport of drug ions through the sweat ducts It is also possible that hyperhidrosis is more profuse in regions, such as the hands, which display greater cutaneous changes in acromegaly

CD.3. Gender differences; sweating, sweat gland size and innervation Sex differences in the response to agents such as pilocarpine have been well- documented 297,386-390 men having a higher maximal sweat secretory rate than women, and boys showing higher values than girls. Calorimetric investigations have confirmed that men begin to sweat at a lower temperature than women and at any given

127 temperature, men sweat more than women The present study has confirmed such a gender difference for the initial rate of sweating in GH-deficient subjects. However, the control group of women unexpectedly had similar rates of sweating compared to men. This may reflect sampling inaccuracies as only a single sweat test was performed on the control group, whilst multiple measurements were taken with the trial group patients. However, the possibility that the control population was skewed in terms of gender responses cannot be dismissed. Gender specific dififerences in sweating have been explained in other studies on the basis of dissimilar body weight and surface area, or due to the degree of sweat gland training (acclimation) Electrolytes in sweat increase with age particularly in males, whilst males generally display higher concentrations of electrolytes than females The localisation of androgen receptor immunoreactivity in cell nuclei of skin, including ESGs suggests that sex hormones may also act at the level of the ESGs. Collins, K.J. suggests that at puberty androgens induce a sex-dependent change in ESGs. Hence, adult males with hypogonadism of pre-pubertal origin have a low sweat rate which is increased by androgen replacement Sex hormones may continue to have a function in adulthood as hyperandrogénie women frequently have skin problems and androgen receptor immunoreactivity has been quantified in the skin with greater staining present in adult males than females Hence, androgens appear to have a role in the normal maintenance of skin including ESGs, and the increased levels of androgens in males may explain their higher levels of sweating. The presence of oestradiol receptor immunoreactivity has also been investigated in skin from premenopausal women Immunohistochemical staining for the oestradiol receptor was demonstrated in the epidermis, sebaceous glands, hair follicles and ducts but not secretory coils of the ESGs Hence, there may be some influence of oestrogens on the sweating apparatus; however, the stage of menstrual cycle was not found to influence the level of staining intensity This agrees with observations that though progesterone is thermogenic, and body temperature alterations occur during the menstrual cycle, no appreciable change in the sweat output occurs Novel findings were the demonstration of larger secretory units and innervation markers (VIP and possibly PGP) in males compared to females, for the trial subjects. These correlated with the demonstration of greater sweating rates in male trial subjects. Since

128 these observations are independent of GH levels (T-W ANOVA analysis, see CR.3. & CR4.), it is proposed that sex differences in the size and innervation of ESGs may be a normal reflection of the known greater rates of sweating in males. Further, the demonstration of sex hormone receptor sites (see above) on ESGs is suggestive of a direct effect of these hormones on the sweating apparatus. It can thus be proposed that the basis for gender-specific sweating responses may be at the level of ESG morphology and innervation.

CD.4. Mechanisms; growth hormone deficiency v acromegaly The total nerve density around the ESG acini of the GHD subjects, as expressed by PGP immunoreactivity, was similar to values in controls, and was unchanged after r- hGH therapy. In contrast, ACRs exhibited an increased density of PGP-stained nerve fibres in parallel with hypertrophy of their ESGs. Hence, an association appears to exist between the overall nerve supply of ESG acini, as demonstrated by the pan-neuronal marker, PGP, and the size of the ESG as measured in microscopical section. This suggests that in situations where GH levels are elevated but still within physiological limits, as occurs during r-hGH therapy in GHD, the ESGs increase their functional innervation through increased expression of AChE and VTP. Alternatively, when GH levels are above physiological levels, as happens in acromegaly, increased sweating may involve an increase in the size of glandular elements, and a corresponding growth of the periacinar innervation, as well as altered levels of neurotransmitters. Thus, over shorter time scales of GH changes, neurotransmitter changes may be the predominant form of nerve plasticity, whilst over longer periods or greater extremes of hormonal elevation, nerve growth in the form of collateral sprouting may occur.

CD.5. Mechanisms; central v peripheral Hypothalamic injury may occur secondary to pituitary tumours, with repercussions for the thermoregulatory centres. Hence, hypothalamic dysfunction contributes to accidental hypothermia in elderly subjects The central temperature-regulating system is inaccessible to experimental studies in humans; however, the improvements in sweating and sudomotor parameters, following r-hGH therapy in this study, suggest that any influence of GHD on the hypothalamic centre were probably reversible. The

129 use of spinal and/or axon reflex methods (see table CI.l.) for assessment of autonomic dysfiinction would probably have been a better option than the ESG-acting pilocarpine iontophoresis technique; however, the data from the total innervation marker suggest that these pathways are maintained in GHD. The mechanism of GH action on thermoregulation can be envisaged to involve several possible pathways. A general increase in GH-induced body metabolism may provide for optimum energy status for thermoregulation to proceed. However, altered thermoregulation in obese subjects is not a general feature suggesting that GH may have a more subtle fianction. GH-induced alterations in thyroid hormones may provide for an indirect influence on thermoregulation, given the altered sweating in hyperthyroid patients Further possibilities concern either a direct effect of the GH/IGF-I axis on the sudomotor apparatus, and/or actions in the thermoregulatory centres in the CNS. An effect of GH on the thermoregulatory centres would presumably involve GH and/or IGF-I crossing the blood-brain barrier. Though the latter suggestion is controversial, evidence from the literature (see AI. 7.) suggests that this may not be a limiting factor. Once into the brain circulation, GH and/or IGF-I would exert their effects through appropriate receptors. The general distribution of these receptor sites has already been described (see AI. 7.). A more thorough evaluation suggests an influence of IGF-I on the thermoregulatory centre. Hence, IGF-I, IGF-I receptor (protein and/or message) and/or binding sites are present in the rat hypothalamus in median preoptic nucleus, supraoptic nucleus, paraventricular nucleus, periventricular nucleus, preoptic suprachiasmatic nucleus, and median eminence 147,103,396 jQp_j mRNA is also present in the anterior hypothalamus, ventromedial nucleus, medial preoptic area and arcuate nucleus IGF-I immunoreactivity has been localised in 'accessory' neurons of the lateral and anterior hypothalamic areas and in the medial and lateral preoptic areas The preoptic anterior hypothalamus area and posterior hypothalamus are important subregions of the thermoregulatory centre. Further, the role of the suprachiasmatic nucleus in circadian temperature alterations is also known. Hence, the distribution of IGF-I provides circumstantial evidence for a direct role of this growth factor on the thermoregulatory centre. A similar evaluation for GH was not possible as only general distributions in the hypothalamus are mentioned by Zhai et al and Lobie et al

130 Immunohistochemical and mRNA localization of GH receptor/binding protein in animal and human skin including ESGs has been demonstrated. IGF-I protein and mRNA expression, and IGF binding protein mRNA, have also been shown in human skin GH and/or IGF-I may therefore influence the sweating apparatus directly. Further aspects of these interactions will be explored in the general discussion.

In conclusion, an influence of GH on the eccrine sweat glands has been established. Patients with GHD display reduced sweating responses which are accompanied by a decrease in AChE- and VTP- positive nerves in the glandular innervation. Following r- hGH therapy, a parallel increase in sweating rates and periacinar cholinergic markers occurs. Conversely, in acromegaly, hypertrophic secretory coils have a correspondingly increased nerve supply. These results suggest that GH may have a direct effect on the ESGs and raise the possibility that GH influences trophic support to the peri-glandular nerves.

131 CHAPTER 4 AN INVESTIGATION INTO THE EFFECTS OF RECOMBINANT HUMAN GROWTH HORMONE ADMINISTRATION INTO FOOTPAD SKIN OF AGEING RATS.

132 Chapter 4 Introduction

This investigation has explored the consequences of r-hGH infusion into the region around ESGs, in footpad tissue from young and old rats. As a preliminary study, part of the objective has also been to establish the value of the rat footpad model, for the assessment of local administration of test substance on sudomotor nerves. The size of ESGs and alterations in innervation have been monitored following a two-week r-hGH injection regimen. The aim of this study was to try to replicate changes observed in GH- deficient and ACR patients (see chapter 3) by local injections of supraphysiologic doses of r-hGH. The choice of aged rats, rather than animal models for GHD or excess (see AI.8.), was to allow exploration of the role of GH in ageing (a state of reduced GH production). In so doing, it was hoped to shed some light on the mechanism of GH action on ESGs. Two main questions were addressed: Is GH acting in a direct manner on the ESGs? and if so, are the changes in innervation due to GH-mediated alterations in the trophic supply to the sudomotor neurons?

D1.1. Ageing and sweating/thermoregulation Alterations in thermoregulation are believed to contribute to the increased mortality in the elderly population due to hypothermia/hyperthermia A general rise in the temperature threshold for sweating and vasodilation occurs with age The sweating response to thermal and local stimulation is reduced in the aged, particularly in women, and a widening threshold for warmth and cold perception occurs in individuals over 65y 267,379,402,405-408 studles suggest that impaired sweating responses in the elderly may be a manifestation of more generalised autonomic dysfunction, given the abnormal sweating responses in many conditions involving autonomic failure Changes in ESGs include a marked reduction in secretory coil density and size, a disorganisaton of secretory cells, an increase and coagulation of elastic fibres around secretory coils, and an increased accumulation within secretory cells of agyrophilic, lipofijscin-like granules that contain no iron 379,407,409,410 further, concentrations of electrolytes, glucose and urea increase with age in sweat obtained fi'om men Both aged human and rat studies have indicated a reduction in overall nerve density, as well as in cholinergic and

133 neuropeptide content, of the sudomotor nerves In aged mice, the number of ESGs responsive to cholinergic stimulation is decreased Also, the capacity for both regeneration and collateral sprouting of sudomotor efferents is reduced with age in the mouse This is believed to be due to the progressive loss of axons, and reduced plasticity of remaining sudomotor neurons, with age Hence, a reduction in the complement of ESGs per peripheral nerve with age occurs, which may also reflect retraction of terminal axon branching These changes in cholinergic innervation appear not to be accompanied by alterations in muscarinic ESG receptors, whose density remains similar in aged subjects as compared to young controls 259

DI.2. Growth hormone and ageing; the somatopause* The anterior pituitary gland is decreased in size with age in humans and shows evidence of fibrosis, necrosis, and accumulation of deposits Both the size and number of somatotrophs decline from youth to middle age Circulating levels of IGF-I are relatively low at birth and during the first fifteen months of life, they increase until the onset of puberty when levels peak; a subsequent decline occurs as a function of age (see table DI.J) *^ 220,221 q jj serum levels in the human fetus rise from around 2nM at 12 weeks gestation to values of 8nM or more by 22 weeks before declining to l-2nM at birth The integrated serum GH level rises to its highest values at puberty. Thereafter, GH levels decline (see table D I.l) in a similar manner to IGF-I in both men and women Hence, compared with young adult men (20-34y), peak GH levels with the insulin-induced hypoglycaemia test were 30% lower in middle-aged men (35- 49y), and 60% lower in elderly men (53-84y) The amplitude of GH pulses is also reduced in senescence in both men and women In rats, a 10% decrease in somatomedin activity within adulthood (fi'om 2-6 months to 12-16 months) is followed by a 20-40% reduction in old rats (24-28 months) GH secretion also declines in senescent (18-20 months) as compared to young rats (4-5 months) The mechanism for the geriatric decline in GH secretion has been explored and involves

an initial decrease in hypothalamic neurotransmitters (dopamine and noradrenaline), a consequent reduction in GHRH and increase in SS production, resulting in a lower GH, and subsequent IGF-I release In aged rats, the increased SS release is accompanied by a preferential production of SS-28, which is a more active form than its counterpart

134 s S-14 The signal transduction pathway for GH has been shown to be impaired in livers of senescent rats despite increased levels of GH receptors, suggesting a cause for the decline in IGF-I gene expression with age Rudman et al have investigated the effects of r-hGH administration to healthy elderly subjects (61-73y) with low circulating IGF-I levels, and found an increase in IGF-I concentration to levels comparable to young controls. This was accompanied by increases in indices of lean body mass and bone density A number of other studies have also been carried out on ageing subjects, differing in their patient criteria, the duration of the treatment (7 days to 6 months), and dosage I3,i5,385,4i6,4i8 though beneficial effects have been reported (improvement in lean body mass, decrease in fat mass, increase in bone turnover, decreased excretion of urinary nitrogen and sodium, greater liver and spleen volume etc), side-effects such as hyperglycaemia, carpal tunnel syndrome and gynecomastia have also been observed. A lowering of the dosage may allow physiological advantages to predominate However, a recent report by Brown- Borg et al suggesting that GH-deficient mice have a longer life-span than their normal siblings is bound to add to the debate on the benefits, or otherwise, of r-hGH therapy to GH-deficient, but otherwise healthy, aged individuals.

135 Age male serum IGF-I (lU/ml) female serum IGF-I (lU/ml) birth to 15 months 0.3 0.5

ly to puberty 1.0 1.5 puberty 2.2 (13 to 15y) 2.6(11 to 13y)

Age male and male and female Age male and female female serum nocturnal GH daily GH IGF-I peak (ng/ml) secretion (lU/ml) (pg/I/24 hours) 20 to 25y 1.2 21 to 31y 80 25 to 35y 0.9 20 35 to 50y 0.8 9.7 50 to 60y 0.51 7.2 29 to 59y 32

60 to 70 0.41 5.3 61 to 71y 21

70 to 80 0.32 3.2

Table DLL Changes in circulating GH and IGF-I levels with age in humans (adapted from Shetty and Duthie and Hammerman,M.R.

136 DI.3. Nerve growth factor, sweat gland innervation and growth hormone The basis of the neurotrophic theory concerns the production of a trophic agent from target tissues, which is retrogradely transported from the nerve terminals to the soma, and which determines the survival of innervating sympathetic, sensory, and other neurones The role of nerve growth factor (NGF; 2.5S P-NGF) which was the first target-derived neurotrophic substance to be characterised is well-established. Subsequent studies have signalled a broader role for one family of neurotrophic factors -the neurotrophins- in general, and NGF in particular. Hence, basal forebrain and septal CNS neurones require NGF. In addition, the definition of neurotrophic factors and the scope of their functions has broadened. For example, NGF promotes growth, differentiation and branching of axons and dendrites It increases neurotransmitter synthesis, acts on neuronal stem cells and glia, and can alter the electrophysiological properties of neurons Further, NGF has actions beyond the nervous system, particularly on the immune and endocrine systems and may be regarded as a co ordinator between these systems NGF is no longer limited to a role in development; it is now believed to be a key modulator of neuronal plasticity in maturity and perhaps even in ageing Neurotrophic factors are defined as those molecules which promote neuronal survival or neurite outgrowth In addition to the classical neurotrophic factors, a wide range of hormones, proteins and oncogenes are now included as substances which can be considered to have a neurotrophic function Additional requirements would have to be fulfilled if these substances are to be regarded as true neurotrophic factors. These include the demonstration of neuronal receptors and neurotrophic factor availability in the neuronal environment Walicke,P. A. has evaluated the role of proteins such as apolipoprotein E and laminin, of growth factors such as IGF-I and epidermal growth factor, and of oncogenes such as c-fos and ras, and has ascribed neurotrophic functions to these various substances. A similar approach by Strand et al has identified adrenocorticotropic hormone, thyrotropin-releasing hormone, and neuropeptides such as SS and substance P, as having neurotrophic properties. Finally, both GH and prolactin, through their actions on their respective regulatory neurons, may be regarded as having neurotrophic functions in addition to their other roles

137 DI.3.1. Nerve growth factor and its receptors There are two NGF receptors (NGFRs); (75-80kDa ) and trk A (HOkDa, pHO'^* with tyrosine kinase activity Though p75^°*^ lacks tyrosine kinase activity, its involvement in neurotrophin signal transduction is evident from its expression in every neuronal cell type known to be responsive to NGF, and the fact that all classical neurotrophins can bind to it 294,423,424 signalling machinery for p75^°^ may be trk independent and p75^°^ is also believed to associate with trk k ox ^ for certain functions However, p75^*^^ is not essential for some ^A-mediated responses to neurotrophins and is believed to have a specialised role in facilitating retrograde signalling In addition, p75^^^^ may prevent or induce apoptotic cell death

DI.3.2. Nerve growth factor and ageing In the mature nervous system, neuronal connectivity is believed to be in a state of dynamic equilibrium, with continuous growth and retraction in response to altered functional demands, and altered availability of target-derived neurotrophic factors The effects of ageing may thus be understood as a distortion of the interactions between targets and their innervation Though some evidence from the central and peripheral nervous systems suggest that NGF levels are not decreased in ageing, altered availability of NGF is believed to influence the extent of neurodegeneration occurring with age NGF infusion into cognitively impaired aged rats results in improved spatial memory as well as a reduction in neuronal atrophy These results suggest that a higher synthesis (or availability) of NGF may be required to counteract decreased neuronal plasticity in ageing. A reduction in both NGFR sites have been demonstrated in the basal forebrain and neostriatum neurons of aged rats; these are accompanied by decreases in the cholinergic activity of these regions Similarly, a reduction in p75^°*^ expression has been observed in superior cervical ganglion of aged mice, which may be correlated with the reduced innervation of some targets of the superior cervical ganglion, including the pineal gland and cerebral arteries ^^ 5,425

138 DI.3.3. Nerve growth factor and sweat gland innervation The role of NGF in the collateral sprouting response (but not regeneration) of sensory and sympathetic neurites is well-established In the skin, NGF is released from kératinocytes, fibroblasts, muscle cells, and blood vessels, and influences the function of mast cells, melanocytes and Langerhans cells, as well as kératinocytes and Merkel cells, both of which have p75^^^ 428,429 A receptor has been shown to be present in human skin, not only in neurones but also on cells in the basal layer of the epidermis, on dendritic cells in the basal and suprabasal layers, on ESG ductal and secretory coil cells (more prominent in the duct), and on cutaneous arteries Despite the presence of trk A on ESG cells, production of NGF by these cells has not been demonstrated The latter study was also unable to demonstrate trk A staining on sudomotor neurites. However, treatment of neonatal rats with an anti-NGF antibody results in a lack of axons and terminals in the periglandular innervation and absence of choline acetyltransferase immunoreactivity occurs Similarly, neurons plated in medium deficient in NGF, but containing sweat gland extracts (ie the cholinergic differentiation factor) lack the ability to survive Hence, NGF appears to act as a survival factor for sudomotor neurones. This has been further explored by Lee et al who noted, in mice bearing a null mutation in the p75^°*^ locus, a deficient innervation of lateral interdigital footpads as opposed to the medial pads which had a normal cholinergic innervation (as assessed by VIP-IR). Further, the lateral interdigital pads displayed a smaller number of active ESGs, than the medial pads whose response to pilocarpine injection was normal The explanation given was of a differential guiding effect of p75^°™ on the lateral and medial plantar nerves. However, this suggestion merits further study.

DI.3.4. Nerve growth factor, growth hormone and insulin-like growth factor I Though experimental evidence for a direct influence of GH on the PNS is lacking, an interaction between GH and NGF has been shown to exist in the pancreatic P-cell line INS-1 GHR sites are present on these cells and r-hGH treatment results in upregulation of both classes of NGFR The pancreatic p-cells are believed to be similar to neuronal cells and indeed, under NGF treatment, they extend neurite-like

139 processes and initiate NGF signal transduction Hence, a similar influence of GH on neuronal NGFRs may exist. The importance of IGF-I in development is emphasised by its stimulation of the initial stage of neurite outgrowth of chick sympathetic neurons in culture The latter effect is replicated by NGF at a later developmental stage and is continued in maturity In the developing mouse following sciatic nerve axotomy, IGF-I can prevent death for 86%, and NGF 61%, of motoneurons which would otherwise die Both IGF-I and NGF also prevented atrophy of surviving motoneurons in the latter study IGF-I increases chick sympathetic neuron proliferation in serum-free cultures However, the role oflGFs in sympathetic neuron survival is not believed to be as important as that of NGF Ishii, D.N. suggests that cooperation between NGF and IGFs may be required for neuronal plasticity in development and maturity. In support of this proposal, IGF-I immunoreactivity has been co-localised with p75^°^ staining in human adult cerebellum purkinje cells Aguado et al suggested that the IGF-I was probably of endogenous origin, based on previous ultrastructural demonstrations of IGF-I associated with rough endoplasmic reticulum. However, the possibility that IGF-I had been transported retrogradely could not be dismissed How would IGF-I and NGF interact in their effects on nerves ? Part of the answer to this question may lie in the ability of IGF-I to potentiate NGF receptor activity. In the human neuroblastoma cell line (SH-SY5Y), trk A receptors are unresponsive to NGF unless IGFs are present in the culture medium Common intracellular actions of NGF and IGF-I include the expression of tubulin and neurofilament genes Interestingly, both NGF and IGF-I stabilise tubulin mRNA transcripts whilst neurofilament mRNA levels are increased through an elevation in transcription rates In primary sensory neurons, NGF and IGF-I fail to induce appropriate neurite outgrowth responses in the absence of the protein kinase C substrate, GAP-43 It would appear that for certain functions a common biochemical pathway exists for these growth factors. Ishii, D.N. argues that retrograde transport of NGF and IGF-I may not be the critical events which govern their neurotrophic fianctions. The tyrosine kinase second messenger system is common to both NGF and IGF-I; hence, phosphorylation of the receptor kinase and subsequent protein kinase C generation may be the key steps for neurotrophic activity Though the latter point is unproven, it does offer some insight

140 into how the two growth factors may complement each other's actions. Due to the well-documented effects of NGF and IGF-I as neurotrophic factors, increasing attention is being focussed on their uses in neurodegenerative conditions. These approaches include the possible use of NGF in treatment of sensory neuropathies and Alzheimer's disease, whilst IGF-I therapy for motor neuron atrophy and chemotherapy-induced peripheral neuropathy, is being actively investigated

DI.4. Aims of the investigation Ageing is accompanied by reduced secretion of GH from the anterior pituitary and by defects in thermoregulation In states of altered GH secretion, abnormalities of thermoregulation exist (see chapter 5). The innervation to ESGs is compromised in ageing It is proposed that the ESGs are a target-organ of GH action. Based on this assumption, the injection of r-hGH into footpad skin of rats is predicted to result in morphological changes to the ESG secretory coils. These alterations, it is proposed, may be particularly evident in aged rats because of their state of relative GHD. It is further suggested that the reduced functional innervation to the sudomotor apparatus of both ageing humans and rats and GH-deficient adults (see chapter 3), is accompanied by a reduction in the neurotrophic suport provided by ESGs to their innervating neurons. Hence, r-hGH treatment may reverse these deficits. As a consequence of r-hGH treatment, it is speculated that p75^*^^ expression will be altered around the ESGs Immunohistochemistry with antibodies against pTS^*^^ and PGP will be carried out on lateral interdigital footpad tissue following a two week period of local r-hGH treatment with supraphysiologic doses. The lateral interdigital footpads have been chosen because of their particular susceptibility in p75^°*^ null mutany mice, as noted above (see DI.3.3.)

141 Chapter 4 Methods

DM.l. Animals and injection regime A total of 6 old (2 years; 700g) and 10 young (2 months; 200g) male rats were used. The animals were from an inbred colony of Sprague-Dawley rats which were kept in a controlled environment, 2-4 rats per cage. The ageing rats were senile with mortality ranging from 30-50% of the population. Young animals were given food and water ad libitum, whilst the ageing rats were kept on a restricted diet (30g day'^) from the age of 3 months. Limited food intake keeps the animals healthier and presenile mortality is decreased Animals were anaesthetized with a mixture of 2% halothane and oxygen. They received three injections per week for two weeks of either r-hGH (Genotropin R; Pharmacia, Sweden) or rat serum albumin (RSA, 1 mg ml-1 in PBS; Sigma, UK). The subcutaneous injections were made with fine, lymphangiography needles and Hamilton syringes into the paw, adjacent to the lateral interdigital footpads. The volume of injected solution was lOpl for old rats and 5 pi for young rats because of the larger size of the paw in old rats. Uninjected paws from the treated animals served as controls. Hence, 3 old rats and 5 young rats were injected with r-hGH, and 3 old rats and 5 young rats were injected with RSA. Doses were calculated on the basis of the total volume of blood in a rat being 60ml k g \ the serum levels in young rats being 35pgU and the half-life of GH in a rat being 60min after subcutaneous injection *^*>'‘38-440 Hence, in order to achieve ten times systemic concentrations for the duration of the study in the footpad, 3.5pg injection^ r-hGH in RSA for young rats, and 12.3pg injection ^ r-hGH in RSA for old rats, were administered. Three days after the last injection animals were killed by an overdose of pentobarbitone sodium (200mg kg'\ Euthatal, RBM Animal Health Ltd, UK). The lateral interdigital footpads were dissected, fixed in 4% paraformaldehyde and processed as for human skin samples (see CM. 3.1).

DM.2. Immunohistochemistry The anti-nerve growth factor-receptor antibody (p75^°^) was a mouse monoclonal (Boehringer Mannheim, UK, clone 192; 1:10 dilution) whilst the PGP antibody was a

142 rabbit polyclonal (Ultraclone Ltd, UK; 1:800 dilution). The immunohistochemistry protocol was similar to that already detailed (see CM. 3.3.). However, for the p75^°^ staining non-specific binding sites were blocked with 5% goat serum. Incubation with primary antiserum was for 40 hours, and secondary antibody was a goat anti-mouse fluorescein isothiocyanate (Dako Ltd, UK; 1:80 dilution). Image analysis methods were the same as those described earlier (see CM.3.4).

143 Chapter 4 Results

DR.1. The effects of recombinant human growth hormone administration into ageing rat footpad skin A prominent but focal inflammatory infiltrate was observed in most (r-hGH and RSA) injected footpad sections (plate 6i). The infiltrate was highly cellular, composed of polymorph cells as well as lymphocytes, and seemed characteristic of an acute response. The exudate appeared to be limited to the injection site, being prominent in the upper dermis and at the epidermal-dermal junction. There was evidence of some infiltrate in the vicinity of ESG secretory coils. Where possible, ESGs chosen for analysis were away from areas of obvious inflammation.

DR. 1.1. Sweat gland acinar size For sweat gland acinar perimeter, the data were shown to follow a non-normal distribution (figure DR. la) which was not normalised following logarithmic or square root transformation. Data was analysed with non-parametric tests. There was no difference in size between young and old ESGs (figure DR.Ib). A reduction in acinar size compared to controls was observed in old footpads, following RSA (19.1%, M-W U test p=0.004) or r-hGH (13.5%, M-W U test p=0.018) injection (figure DR.lb). ESGs from r-hGH injected footpads appeared to be slightly larger (6.5%, M-W U test \)=0.Q\2, figure DR.lb) then those from RSA injected animals.

144 12

0 130 HO 150 160 170 180 190 200 210 220 perimeter of sweat glands Observed expected (normal)

Figure DR.la: Graph showing data distribution for rat sweat gland perimeter (pm). Samples from young and old rats Expected normal distribution also shown

200 ***

■o 150

^ 100

O-RSA 0-GH

Figure DR lb: Histogram showing sweat gland acinar perimeter in young and old rats. Y-UT=young rats untreated, Y-RSA=young rats RSA injected, Y-GH=young rats r- hGH injected, 0-UT=old rats untreated, 0-RSA=old rats RSA injected, 0-GH=old rats r-hGH injected Error bars show SEM, *p<0.05, ***p<0.005.

145 DR. 1.2. Sweat gland innervation Because there was a change in ESG size with RSA/r-hGH injection, presenting nerve data relative to individual acini was not considered appropriate, especially as the decrease in acinar size was to a dissimilar degree in RSA and r-hGH injected groups. Nerve parameters have therefore been presented as nerve area per pm acinar perimeter. However, statistical analysis was also carried out on data expressed as nerve area per acinus and where appropriate complementary data has been commented on. For the py^NGFR staining, due to technical problems, two data groups are missing; these are the young RSA injected and the old control groups.

DR. 1.2.1. Protein gene product immunohistochemistry Immunoreactivity with the general neuronal marker PGP was observed in footpads from young and old rats, being prominent in nerve fibres around ESGs {plates 6i &6ii). Distribution of the data were not normal (figure DK 1.2. la), and remained so following logarithmic, or square root transformation. Data were therefore analysed with non- parametric methods. Young groups generally displayed a greater (27-33%, K-W AND VA p=0.003, DR. 1.2. lb) sudomotor innervation than that present in footpads from old animals. There were no statistical differences between control and injected groups for either young or old rats. However, the r-hGH injected footpad samples appeared to have reduced sudomotor innervation as compared to both the control and RSA injected groups (young ESGs; control 17.3%>r-hGH, RSA 13.5%>r- hGH, old ESGs; control 19.6%>r-hGH, RSA 20AVo>v-\iGYi, figure DR.1.2.1b).

146 7

2 3 4 nerve area per acinar perimeter Observed ^ expected (normal)

Figure DR.12.la Graph showing data distribution for PGP immunoreactivity in rat sudomotor nerves, expressed as nerve area per pm acinar perimeter. Samples from young and old rats Expected normal distribution also shown.

2.5 I1 2 g 1.5

& I (D 0.5 0) c

Y-UT g Y-RSA Y-GH 0-UT ^ O-RSA m O-GH

Figure DR. 1.2.lb: Histogram showing PGP immunoreactivity in young and old rats. Data expressed as area of nerves per pm of sweat gland secretory coil. Y-UT=young rats untreated, Y-RSA=young rats RSA injected, Y-GH=young rats r- hGH injected, 0-UT=old rats untreated, 0-RSA=old rats RSA injected, 0-GH=old rats r-hGH injected. Error bars show SEM, ***p<0.005.

147 DR.1.2.2. Nerve growth factor receptor (p75) immunohistochemistry In tissue from both young and old rats, there was extensive staining of nerve fibres innervating ESGs with the antibody (plate 6i). The distribution of the data were not normal (figure DR. 1.2.2a) and were not normalised following transformation by taking logarithms or square roots. Analysis of data with the non-parametric K-W ANOVA revealed greater (26-50%, p=0.006, figure DR. 1.2.2b) p75^*^^ immunoreactivity in young sudomotor nerves, as compared with tissue from old rats. Aged rat footpads, injected with r-hGH, appeared to have reduced (25,4%, K-W ANOVA p=0.086,^gwrei)Æ7.2.20) p75^°^ staining around ESGs as compared with RSA injected aged rat tissue, however this was not significant. Interestingly, the nerve area occupied by p75^^^ immunoreactivity was roughly comparable with that for PGP staining suggesting that most (83.8% from all data pooled) sudomotor nerves contain p75^‘^*^.

148 i

B o I ‘ ?£

0 2 3 nerve area per acinar perimeter Observed expected (normal)

Figure DR. 1.2.2a: Graph showing data distribution for p75 immunoreactivity in rat sudomotor nerves, expressed as nerve area per pm acinar perimeter. Samples from young and old rats. Expected normal distribution also shown.

3.5 ** I E & 2.5 2 1 E ZL I

C O 1

c 0.5

Y-UT Y-GHO-RSA m O-GH

Figure DR. 1.2.2b: Histogram showing immunoreactivity in young and old rats. Data expressed as area of nerves per pm of sweat gland secretory coil. Y-UT=young rats untreated, Y-GH=young rats r-hGH injected, 0-RSA=old rats RSA injected, 0-GH=old rats r-hGH injected. Error bars show SEM, **p<0.01.

149 Plate 6i Photomicrograph of eccrine sweat glands from rat footpad skin, stained for PGP and p75"'^'™. A: ESGs from an old r-hGH injected footpad, stained for B: ESGs from an old RSA injected footpad, stained for p75^°*'^.

C; ESGs from a young uninjected footpad, stained for PGP. Note Periacinar nerves (arrow), inflammatory exudate (i). Scale bar ^50pm. Lipofuscin (1) pigment in old footpad secretory coils Magnification for all fields =x20. Due to technical problems the photomicrographs presented are not of a good standard

Additional photography was not possible due to fading of immunofluorescent sections. The photomicrographs do, however, illustrate the localisation of PGP and p75^^'™ staining on periacinar nerves. For any comparison between groups please refer to the quantitative data provided in Figure DR. 1.2. lb & Figure DR. 1 2.2b.

150 Plate 6ii Photomicrograph of eccrine sweat glands from rat footpad skin, stained for PGP.

A: ESGs from an old uninjected footpad. B: ESGs from an old r-hGH injected footpad. C: ESGs from a young RSA injected footpad.

Note: Periacinar nerves (arrow). Magnification for all fields =x20. Scale bar =50pm.

151 ' | : 5 - ' Chapter 4 Discussion

There was no difference of sweat gland size between young and old rats. In a previous study on lateral interdigital footpads in young and old rats, Abdel-Rahman and Cowen noted an increase in secretory coil perimeter in ageing (14%), accompanied by a reduction in numbers of acini (43%). For the control groups, in the present study, a trend existed for ESGs from old rats to be larger (8.7%) than those from young rats. Hence, a similar though reduced increase in aged ESG was observed in this study. Injection of either RSA or r-hGH resulted in a decrease in the size of the secretory coils in aged rats. These changes may be assumed to be due to the frequent injections provoking repeated acute inflammatory responses in the vicinity of ESGs. The larger acinar size in r-hGH treated aged rats, as compared to RSA injected animals, may be ascribed to a GH-dependent return towards normal parameters. Further aspects, in relation to inflammation, are explored below. A marked difference was observed in the overall innervation to ESGs as assessed by PGP, between young and old rats. This has been previously described by Abdel- Rahman and Cowen who noted a 50% decrease in PGP immunoreactivity in old rats as compared with young controls. In a similar investigation, Abdel-Rahman et al were able to demonstrate a comparable reduction in PGP immunoreactivity, around sudomotor nerves of aged but healthy human subjects. Interestingly, in the latter study, evidence from electron microscopy indicated that axonal loss, rather than reduced PGP expression in nerve fibres, was responsible for the reduced PGP immunoreactivity in ageing The results from the present investigation (31% decrease in PGP staining with age) were not as dramatic as in the Abdel-Rahman and Cowen study. One possible explanation, for the discrepancy between the two studies, concerns the nutritional status of the animals. Experiments currently underway in T.Cowen's laboratory indicate that enteric neuronal cell loss is particularly prominent in ad libitum fed aged rats, as compared to aged rats fed on a restricted (50% of normal) diet. The aged rats in the Abdel-Rahman and Cowen study were fed ad //7>z7w/w,whilst those in the present investigation were on a restricted diet of 30g protein a day. These preliminary observations suggest that a range of autonomic neurons are

152 affected by dietary conditions. In this context, the lack of a substantial increase in ESG size with ageing, seen in the present study, may also be due to the relatively 'less-aged' status of these animals. The inference from these observations is that unchecked feeding may accelerate age-related changes in the autonomic nervous system, suggesting that the mechanism involved may not be directly associated with digestion. D'Costa et al have reported that increased protein synthesis in moderately diet restricted rats is coupled to elevations in IGF receptors as well as increases in the amplitude of GH pulses Hence, further investigations of the trophic status of autonomic neurones from rats under different dietary regimes are warranted. The present study is the first to describe age-related changes in p75^°^ expression in sudomotor nerves. Most observations of neurotrophic influences on the sudomotor apparatus have concentrated on the role of the soluble cholinergic differentiation factor, which induces a switch in transmitter phenotype during development 289,443-445 role of NGF in supporting sympathetic neuron survival is well-established, and is believed to extend to sudomotor neurones Observations in the present study, that p 75 ^GFR nerve density was roughly comparable with that for PGP in young and aged rats, are consistent with a role of NGF in the survival of sudomotor neurones. Though P75 NGFR is capable of responding not only to NGF but also to a number of related neurotrophic factors the present focus on neuronal survival is particularly applicable to NGF. In this context, p 75 ^GFR may be involved in cell survival by preventing expression of apoptotic genes or in the regulation of collateral sprouting Treatment with r-hGH is known to upregulate p 75 ^GRR expression in the pancreatic P- cell line INS-1 This was achieved via a distinct GH-responsive sequence in the promoter region of the p 75 ’^GFR gene, resulting in increased transcription of p 75 ^GFR 433 The finding from the present study was of a reduction of both PGP and p 75 ^GFR expression on sudomotor nerves by injection of r-hGH into aged rat footpads. These results suggest that nerve fibre loss was accompanied by comparable decreases in p 75 NGFR levels. There is thus no indication from this preliminary study for an influence of GH on either p 75 ^GFR expression or nerve growth. The glandular acini appear to have a critical role in the present investigation. An inflammation-mediated atrophy in gland size was accompanied by a reduction in total innervation and with it p 75 ^GFR expression. In line with the initial hypothesis, acinar

153 hypertrophy would be expected to result in parallel increases in innervation. Only the first of these inferences were found to be correct. Analysis of nerve data as nerve area per acinus revealed a decrease in the PGP innervation of ESGs (in RSA injected aged rats) to levels comparable for those in the r-hGH treated footpads (data not shown). However, no increase in nerve density or p75^*^^^ expression was observed following r-hGH treatment. Hence, despite an influence on gland size, GH had no effect on nerve growth. If the direct hypothesis is only partly correct in this model, what other factors could be influencing GH action? There may have been a failure to generate IGF-I, or inflammatory processes may have interfered with the predicted function of GH. Both these suppositions will be discussed below. GH-induced sciatic nerve regeneration is believed to be mediated via paracrine actions of IGF-I The corollary, for the present investigation, would relate to a failure of IGF-I production which would be the mediator of neurotrophic functions of GH. The reason for a possible lack of IGF-I effect in the present study can only be speculated at, considering the limited data. Possible explanations, however, may lie in a failure of locally injected GH to stimulate paracrine IGF-I production. Alternatively, the neurotrophic effect may be dependent on circulating IGF-I levels which are unlikely to have been increased by the local injection regime in the present study. However, both GH and IGF-I are involved in tissue repair and regeneration during inflammation Further, IGF-I is expressed in granulation tissue and IGF-I and its BPs in wound fluid Hence, the presence of IGF-I in the local environment was probably increased following footpad injections. With regard to inflammation, is there any evidence which indicates a reduction in trophic support in the vicinity of the response? In rat lacrimal glands, the ageing process is associated with a reduction in cholinergic innervation, which may be a consequence of inflammation-mediated acinar destruction Similarly, glandular atrophy in inflammatory bowel disease is accompanied by a reduction in periglandular peptidergic nerves, even in resolving ulcerative colitis Alterations in skin neuropeptides have been documented in a wide range of human conditions with an inflammatory component 325,328,329,452 gg^iapati et al were able to demonstrate a reduction in neuropeptide content in the vicinity of wound healing in rat footpads for up to two weeks.

154 Conversely, Konttinen et al showed labial salivary gland atrophy, despite a normal cholinergic innervation, in subjects with local acute inflammation. These studies indicate that glandular atrophy is a normal consequence of inflammatory processes and that innervation to these structures may also regress in parallel with the glandular change. There is also evidence, however, of an increase in nerve activity in the vicinity of inflammatory responses. The efferent role of sensory fibres during inflammation is well documented and involves an antidromic release of neuropeptides from nerve endings 455-457 ^ critical role of mast cells in this process is evidenced by close contact of mast cells with these nerve fibres, with bidirectional interactions occurring Though mast cells express the NGF receptor trk A, and NGF is believed to influence mast cell mediator release, pJ 5^GFR present on mast cells Evidence for a direct role of sudomotor nerves in the inflammatory response is lacking. Therefore, a differential response of sympathetic and sensory nerves may be hypothesised, with sensory nerve ingrowth into the region of inflammation in the upper dermis, accompanied by a regression of sympathetic sudomotor fibres. Further, the critical role of NGF in inflammation, and the dependence of both sets of nerve populations on this factor, suggests that the instructive signal to the ESG innervation may be NGF. Hill et al have shown that sympathetic and sensory neurones compete for NGF produced by target tissues in the developing rat footpad skin. Further elucidation of this process is required, especially as to the neuropeptide content of the sudomotor nerves in the present study and regarding the interaction between GH and NGF. Quantification of NGF levels in the sudomotor nerves may confirm the proposals outlined above.

A difference was observed, in both acinar size and possibly p75^^^ expression of aged rats, presumably due to inflammation and r-hGH treatment, which were not apparent in young rats. These results would suggest a greater sensitivity of aged rats to inflammation, a conclusion also suggested by the data of Williams et al who noted a greater mast cell infiltrate in exorbital lacrimal glands from aged rats. Changes due to r-hGH therapy would thus be consequent on a greater inflammatory reaction in ageing footpad tissue. The results from the present investigation suggest that r-hGH therapy evoked a local glandular change and, to that extent, the footpad appears to be a suitable model for

155 investigating growth factor-target interactions. However, the sensitivity of the footpad skin to inflammation means that frequent injection regimes should be avoided in future studies. In summary, preliminary data, from r-hGH treatment of young and old rat footpad tissue, suggests that inflammation caused an atrophy to aged rat ESGs, an effect which was only partly reversed by r-hGH administration. The decrease in acinar size was accompanied by a reduction in the extent of innervation and neuronal p75^^^ expression around the ESGs. Local GH treatment had no effect on nerve growth or p 75 ^GFR expression possibly due to complications of the inflammatory response.

156 CHAPTER 5 AN INVESTIGA TION OF SKELETAL MUSCLE IN GRO WTH HORMONE-DEFICIENT SUBJECTS BEFORE AND AFTER RECOMBINANT HUMAN GROWTH HORMONE THERAPY.

157 Chapter 5 Introduction

£1.1. Muscle; function and composition Myosin and actin are the major (80%) myofibrillar proteins in muscle. Cross-linking of these proteins results in contraction via actin-activated myosin adenosine triphosphatase (ATPase) activity, and hence the transduction of ATP-hydrolysis free energy into mechanical work and heat. Other myofibrillar proteins, which have a mainly regulatory or structural function, are also present in muscle and these include tropomyosin, troponin, a-actinin, desmin, connectin, C protein and M protein. Myosin has a molecular weight of 500 kDa and consists of two heavy chains and two pairs of light chains. The length of the myosin molecule is about 150nM constituting a tail region and two globular heads. The tail is composed of two a-helical heavy chain segments which coil around each other, fi'om these the heads project forming a globular complex with one molecule of each type of light chain. The globular heads interact with actin and form the cross-bridges between the thick (myosin) and thin (actin) filaments of muscle. ATPase activity is restricted to the globular head whilst the tail has a structural role.'^^® Both myosin heavy chains (MHCs) and myosin light chains exist as multiple isoforms that are differentially distributed in several fibre types MHC and myosin light chain variants can associate in various combinations, generating a large number of myosin molecules The functional significance of myosin polymorphism lies in each muscle fibre having an isomyosin expression, coded for by a distinct gene, which determines its intrinsic contractile dynamics Individual muscle fibres also contain varying proportions of glycolytic, oxidative and phosphorylase enzymes which can be used to histochemically characterise the fibres. Hence, the myosin ATPase reaction allows the detection of at least four types of fibres in human muscle. These are the slow contracting type I fibres (oxidative, for sustained powerful activity), and the fast contracting type IIA (oxidative/glycolytic), IIB (anaerobic, for rapid intermittent activity) and IIC fibres The last has intermediate properties to type IIA and IIB fibres, and is found in foetal muscle and under pathological conditions

158 The isomyosin profile is increasingly being utilised as a marker of fibre type. Expression of myosin genes changes during development, as well as during post-natal life in response to physiological stimuli. Hence, changing functional demands may be detected by an alteration in the fibre type and hence isomyosin profile Analysis of single muscle fibres have demonstrated that specific MHCs I, Ila, and Ilb, correspond to the histochemically defined fiber types I, IIA, and IIB respectively The presence of two distinct groups of early myosin components is found in foetuses which correspond to the type IIC fibres Further, an additional MHC Ild has been characterised, and fibres containing this isoform have been designated HD (fast type), and found to be the same as the recently identified 2X fibres ^^^>^^2,464 containing Ild/IIxMHC have an aerobic oxidative capacity intermediate between that of the type IIB and IIA fibres, a similar shortening velocity to DA fibres, and share a similar power output to IIB fibres 461,465 MHC genes are located on two clusters in humans, the p/slow-MHC gene is closely linked to the cardiac a-MHC gene on chromosome 14, whilst at least six genes, including the embryonic, neonatal and adult fast MHCs, are clustered on chromosome 17 Aside from situations of development and muscle regeneration, where muscle hyperplasia occurs, fibre hypertrophy/atrophy and fibre type switching are believed to be the normal response of muscle fibres to physiologic stimuli The modulation of MHC genes in adult skeletal muscle can be influenced by a variety of factors, including innervation, mechanical loading/strength training, and hormonal (mainly thyroid hormone) alterations Hence MHC gene expression can transform fibre phenotype in the order I ^ I C ^ I I C ^ I IA ^ 2 X ^ I IB Thus, for example, chronic low- frequency nerve-stimulation in the rabbit induces pronounced fast-to-slow transitions for both MHCs and fibre types Evidence to suggest that similar mechanisms exist in humans are mainly circumstantial. Hence, the presence of both fast and slow type myosin chains in the same fibre, which is a characteristic of a fibre in an intermediate stage of fibre switching, has been demonstrated in mammalian developmental and cross innervation studies The presence of type DC fibres in human foetal muscle, which are the precursors of both the type I and the type IIA fibres, suggest that similar mechanisms of myosin gene activation exist in humans

159 The situations in which fibre type changes in adult humans are varied. Fibre type changes are known to occur in response to resistance training Also, the proportion of type I fibres has been shown to increase with age in the 22-65y bracket though other studies have demonstrated no such change in similar age groups However, longitudinal studies suggest that the proportion of type I fibres does indeed increase in normal adults with age, whist the percentage of type I fibres is greater in adult females than adult males In contrast, teenage boys and girls demonstrate a similar fibre type distribution Hence, fibre type may change under the hormonal influences of puberty. This is believed to occur by the switching-on of normally inactive genes and the switching-off of those that had been active In dystrophic muscle, in neuromuscular diseases, and in anorexia nervosa, an increase in the proportion of type I fibres occurs. This is attributed to muscle atrophy which results in surviving muscle fibres being more intensely used than normal The increase in type I fibres in Duchenne muscular dystrophy is accompanied by an increase in the slow myosin content of their muscle A similar pattern of fast to slow fibre type conversion occurs in hypothyroidism with the reverse occurring in hyperthyroidism

EI.2. Growth hormone and muscle in human studies Patients with acromegaly have an increase of muscle bulk and strenth in the earlier stages of the disease'^^^. The demonstration of myopathy in acromegaly by Mastaglia et al indicated an effect of GH on neuromuscular function. ACR patients were found to have proximal muscle weakness, easy fatiguability, short action potentials and hypertrophy of both type I and type II muscle fibres Atrophy of muscle fibres was also observed in ACR subjects during the later stages of the disease A functional failure of morphologically normal muscle fibres was postulated especially as animal studies also suggested that the contractile strength of long-term GH treated animals was reduced, despite hypertrophy of their muscle In addition, excess GH treatment in rats was shown to decrease ATPase activity, and membrane excitability, despite increased concentrations of actin and myosin GH-deficient adults frequently suffer fi'om general fatigue, low physical capacity and reduced muscle strength, size and force-generating capacity suggesting the presence of a proximal myopathy these shortcomings, adult patients with

160 childhood onset GHD have muscle force and peak power comparable to that recorded in controls In these subjects, following discontinuation of GH therapy for a year, a reduction in muscle strength, mass and fibre area occurs, whilst no change in fibre type composition is observed ^9.483,484 patients with GHD acquired as children or as adults, treatment with r-hGH, for six to eighteen months, results in an increase in skeletal muscle and lean body mass, a reduction in fat mass, an elevation in maximum isometric force, and an increase in strength is seen, but only in certain muscle groups (quadriceps and limb girdle muscles) 12,22,187 ,188 ,190,482 ^ positive relationship between peak IGF-I levels on GH and the changes in muscle strength is observed in these subjects In a similar study, on the effects of six months of r-hGH therapy in adult subjects with GHD, there was no elevation observed in fibre diameter or fibre proportions of either type I or type II muscle fibres Also, adult GH-deficient patients were shown to have comparable type I/II fibre areas, and analogous fibre proportions, to control subjects both before and after six months r-hGH treatment The lack of a substantial change in muscle fibre size and proportions suggests that the influence of r-hGH may become evident only when accompanied with other factors such as training. Evidence from animal studies indicate that muscle atrophy, in hindlimb-suspended young and old rats, can be reduced by a combination of resistance exercise and r-hGH therapy However, exercise-induced increases in muscle area and strength were not augmented by r-hGH therapy in elderly subjects 216,489 Since muscle mass was shown to increase in r-hGH treated GH-deficient adults the lack of increase in fibre area may suggest a flaw in the methodology which failed to detect fibre hypertrophy. Alternatively, non-contractile elements such as water, fat and connective tissue, may have increased, though no histological evidence of this was observed It is also possible that hyperplasia of fibres may have occurred. This is suggested by the data of Cuneo et al though not commented on by the authors (number of type I fibres per quadriceps femoris increased by 29.5% on r-hGH, p>0.05). In support of the latter proposal, r-hGH treatment is known to increase nuclear DNA synthesis and to stimulate satellite cell proliferation in rat skeletal muscle Treatment with r-hGH in normal and postoperative patients stimulates protein synthesis Also, six hours r-hGH administration to nutritionally depleted subjects has been shown to elevate MHC mRNA levels without changing serum IGF-I levels However, muscle

161 protein anabolism seen in humans, following r-hGH infusion, can be replicated by infusion with IGF-I

EI.3. Growth hormone, insulin-like growth factor I and muscle innervation Muscle fibres innervated by a single motor neurone possess identical histochemical characteristics and uniform physiological properties Cross-innervation studies have indicated the trophic influence of the innervation in determining the structure and physiological properties, and hence histochemical type, of skeletal muscle The role of IGFs as candidate muscle-derived neurotrophic factors is well-established and involves their action as survival and differentiation factors for motoneurons, as inducers of nerve sprouting, and as modulators of development and turnover of neuromuscular synapses ACh receptor clusters have been shown to be induced by IGF-I in cultured muscle cells, and motoneuron GAP downregulation in vivo can be prevented by IGF-I IGFBPs have been localized in intramuscular nerves and at the neuromuscular junction within the presynaptic nerve terminal, retrograde transport of the BP occuring along the axon The altered strength in GH-deficient patients may be due to an altered neural recruitment of muscle fibres. However, Sartorio et al were able to fully activate all muscle fibres even before GH therapy. Hence, the increase in force seen with GH treatment, due to fibre hypertrophy and/or hyperplasia, was produced by the same level of neural activation

EI.4. Growth hormone, insulin-like growth factor 1 and muscle in animal studies Immunoreactivity and mRNA for IGF-I in striated muscle has been demonstrated Receptors for IGF-I have been identified in skeletal muscle and IGF-I exerts an anabolic action on skeletal muscle through mitogenic action and stimulation of cell differentiation In addition, it elicits metabolic actions such as stimulation of glucose and amino acid transport, and promotes a net positive protein balance A number of different stimuli are known to induce IGF-I expression in animal muscle. These include stretch-induced skeletal muscle growth, partial denervation of muscle, work-induced compensatory hyperplasia/hypertrophy, electrically contracted muscle, and muscle regeneration following injury Muscle regeneration and hypertrophy, in normal and hypophysectomised rats, initially involves localisation of IGF-I in satellite

162 cells which are induced by IGF-I to proliferate and differentiate, particularly fibroblasts 99,503-506 satellite cells constitute the precursor cells for muscle growth and hypertrophy in the postnatal organism, and these display IGF-I immunoreactivity along with intramuscular nerves and blood vessels Hence, elevated intramuscular IGF-I may be an early signal from muscle fibres to initiate regenerative changes Subsequently, autocrine release of IGF-I and BPs from myofibres occurs Thus, during muscle regeneration, IGF-I stimulates proliferation, and promotes fiision, of myoblasts with increased immunoreactivity for IGF-I present in the cytoplasm of myoblasts, myotubes and mature muscle cells, the cytoplasmic location suggesting local synthesis 500,506,507 Regeneration can be impaired in mice by the injection of antibodies against IGF-I Following hypophysectomy or injury, skeletal muscle IGF-I level increases in rats were found to be dependent on exogenous r-hGH administration 5o»,509

£1.5. Growth hormone, insulin-like growth factor I and muscle fibre types and myosin expression Exercise is one of the most important stimuli for GH production in humans Following exercise, human and rat muscle display an increased expression of IGF-I in vascular and satellite cells The exercise evoked increase in IGF-I is of both liver and muscle IGF-I mRNA in rats GH suppression in rats results in a reduction of IGF-I message in the liver to a much greater extent than in the muscle. Training in these GH suppressed rats causes an increase of mainly muscle IGF-I mRNA, suggesting a GH-independent increase of IGF-I in muscle Similarly for both normal and GH- deficient adults, following exercise, an increase in serum IGF-I occurs which is independent of GH levels Muscle protein synthesis is decreased in hypophysectomised rats, including a reduction in type I fibres, both of which can be reversed on r-hGH treatment p^rther, message for the type I, type Ila, and embryonic MHCs are decreased on hypophysectomy, whilst those of type lib and neonatal MHCs increase dramatically Treatment with r-hGH reverses most of these changes, including increases in total DNA, RNA and protein content of the muscle 512-514 jype I MHC levels are the exception though these also return to pre-hypophysectomy levels in rats on a restricted diet The latter data suggests an involvement of IGF-I in the process, since IGF-I can

163 directly affect MHC gene expression in the heart, and nutrition is known to alter GH modulation of IGF-I levels An increase in muscle fibre diameter occurs on r-hGH treatment of normal rats, as does an increase in the DNA.protein ratio and satellite cells:muscle fibre ratio However, normal rats do not exhibit an alteration in muscle fibre phenotype on r-hGH administration in the same muscle groups that demonstrate changes in hypophysectomised rats (soleus and extensor digitorum longus muscles)

132,515

£1.6. Aims of the investigation The fatigue and weakness experienced by GH-deficient adults may be due to an increased proportion of fast, fatiguable, type II fibres in their skeletal muscle This would be in keeping with the above data from hypophysectomised rats which suggests a decrease in type I fibres A consequence of increased levels of type II fibres would be an elevation in the speed of contraction and relaxation of the muscle This was indeed found to be the case in a group of GH-deficient adults for their quadriceps muscle Further, there was a trend observed in these patients for a return to normal values for these parameters following GH replacement Though other studies have found no changes in either the proportions or size of type I/II fibres, it was important to verify these observations due to the recent indirect evidence of Rutherford et al suggesting an increase in type II fibres in GHD, and the study by Cuneo et al which indicated that hyperplasia of fibres was occurring on r-hGH therapy. The data from patients with lean body wasting (see EL2.) suggest either a direct action of GH on muscle, or an indirect influence via local production of IGF-I In support of this proposal, injection of anti-GH antibodies into prepubertal rats is shown to reduce the growth rate of the whole body, including muscles and bones, without affecting plasma IGF-I levels However, in adult rats serum IGF-I concentrations are substantially reduced on anti-GH treatment, muscle mass being similarly decreased Hence most of the actions of GH on muscle, in animal studies at least, are believed to occur as a combination of muscle- and liver-derived IGF-I, with the overall evidence suggesting autocrine IGF-I as a major mediator in the process. Since the situation in human studies is not so clear, an evaluation has been attempted of IGF-I mRNA expression in muscle from patients with GHD, both before and after r-hGH treatment.

164 A comparison of such data with the circulating levels of IGF-I should provide an indication for the importance of a local verses endocrine action of IGF-I. Rat muscle IGF-I expression is altered in paralyzed and stretched skeletal muscle The distribution of IGF-I message, in the latter studies, was particularly prominent in the cytoplasm of muscle fibres. However, the particular fibre types associated with these increases in IGF-I mRNA were not evaluated. It was hypothesised that beneficial effects of r-hGH therapy in adult GH-deficient patients would be accompanied by fibre type-specific changes in muscle IGF-I mRNA expression. The methodology adopted in the present study is a combined immunohistochemical (IHC) and in situ hybridisation (ISH) approach. A quantification of muscle fibre proportions and size has been carried out on tissue taken from GH-deficient patients at the three trial stages. With the same biopsy samples, mRNA for IGF-I has been semi quantified for particular muscle fibre types. In addition, preliminary data with a reverse transcriptase polymerase chain reaction (RT-PCR) method have been obtained, in order to provide complementary evidence to the ISH results.

165 Chapter 5 Methods

EM .l. Introduction In order to investigate the effects of GHD and r-hGH therapy on skeletal muscle, IHC, ISH and RT-PCR was used. Biopsies from the vastus lateralis muscle were obtained (see EM.2.) at the three trial stages. Due to incompatible fixation procedures, and patient non-cooperativity, tissue from only ten patients was available for analysis at all three trial stages. The ten patients consisted of 7 males (mean age 45.4y) and 3 females (mean age 38y). Three of the male patients received r-hGH for twelve months whilst all other patients were given placebo for six months, then r-hGH for the next six month period. There were no control patients recruited due to the unpleasant nature of the biopsy procedure. Cryostat sections were cut and IHC carried out with monoclonal antibodies to MHCs for type I or type II fibres. All fibres in muscle sections were stained with this procedure, allowing an assessment of fibre proportions (see EM. 3.). Adjacent slides to those stained for MHCs were processed for ISH utilising a cRNA probe to a general sequence of IGF-I (sQeEM.4.). The ISH signal was imaged in the light microscope and assessed semi-quantitatively using computerised image analysis (see EM. 5.), comparison with adjacent slides stained for MHCs allowing a differentiation between signal on type I or type II fibres. The image analysis also allowed the simultaneous measurement of fibre areas. Uncut biopsy material was processed for RNA analysis using a RT-PCR method (see EM.6.). The probes utilised in the RT-PCR procedure were for specific IGF-I transcripts, expressed normally, or in situations of muscle growth (see EM. 6. ).

EM.2. Biopsy and muscle processing All biopsies were carried out by a single physician with the Bergstrom method The vastus lateralis muscle was chosen due to its mixed fibre distribution and availability of reference data for morphological parameters. Following infiltration of the skin with 2% lignocaine, a small incision was made and a standard percutaneous muscle biopsy needle inserted into the muscle, with a similar mid-muscle depth attempted at every

166 biopsy. After several bites with the biopsy needle, about 50mg of muscle was obtained per biopsy. The muscle was divided into two portions. That for IHC and ISH was subdivided into several pieces, segments of which were mounted in cryo-embedding medium (Bright, UK) on cork blocks in a transverse orientation. These muscle blocks, and other pieces for RT-PCR analysis, were frozen in iso-pentane cooled by liquid nitrogen (to avoid ffeeze-ffacture artefacts) and subsequently stored at -80°C. Cryostat sections were cut at -20 °C as a sequence of 7 pm serial transverse sections, with three sections per slide. Sections were thaw mounted on slides coated with silane (see appendix G.4.\), and representative slides obtained from a thickness of 3mm muscle. For each biopsy, four groups of three slides were processed, separated from each group by a thickness of about 400-500pm. After cutting of slides, and room temperature drying for up to 15min, slide trays were wrapped in aluminium foil, then in plastic bags and stored at -80°C. Care was taken to avoid RNase contamination at this and all subsequent stages for ISH and RT-PCR analysis. This was accomplished by using sterile equipment, baking all glassware or using disposable plasticware, autoclaving appropriate solutions, using RNA-grade chemicals, treating solutions with diethyl pyrocarbonate (DEPC, an inhibitor of RNases), as well as elementary precautions of wearing gloves and maintaining a clean environment. In addition, before cutting sections, the cryostat blade was cleaned with acetone then with a 1% sodium dodecyl sulphate (SDS) solution containing DEPC.

EM.3. Immunohistochemistry for myosin heavy chains Preliminary experiments were carried out using mouse monoclonal antibodies to MHCs I and n (IgG-1, Clone No. MHCs and MHC/’respectively from Biogenesis Ltd, UK). These antibodies had previously been shown to stain for myosin with an indirect immunoperoxidase method An immunofluorescence procedure was substituted for the immunoperoxidase technique to allow double-immunolabelling. Results fi'om the double-labelling procedure revealed recipricol staining for the two antibodies with no overlap in staining observed. This was true for muscle taken from patients at all trial stages. It was important to verify the distinct staining pattern of both antibodies, as the coexistence of fast and slow myosins within a single fibre may occur during fibre type

167 switching Hence, for the subsequent experiments, only the MHC I antibody was used to stain sections. Unstained fibres were assumed to contain MHC II. The IHC procedure was broadly similar to that described earlier (see CM. 3.3.). Slides were thawed at room temperature before unwrapping, to avoid ffeeze-thaw artefacts, then fixed in 4% paraformaldehyde for 20 min. After washing in Hepes buffer, sections were incubated in 5% goat serum for 60 min. A rinse in Hepes buffer was followed by incubation overnight in a humid chamber with mouse antisera raised against MHC I (1:100 dilution). Following incubations, the sections were rinsed in PBS and incubated with goat anti-mouse fluorescein isothiocyanate (Dako Ltd, UK; 1:80 dilution) for 90 min at room temperature. After a final wash in PBS, slides were mounted in antifade mountant (Citifluor, UK).

£M.4. In situ hybridisation with the digoxigenin system ËM.4.1. Introduction The non-radioactive digoxigenin (DIG) system for labelling and detection of nucleic acids is simple, provides better spatial resolution, and is a safer alternative to the conventional radioisotope methods for in situ hybridisation The use of the DIG molecule (a steroid isolated from the plant Digitalis purpurea), linked to uridine- triphosphate (UTP), as a marker in cRNA probes for the detection of mRNA sequences in situ is reliant on subsequent detection by anti-DIG antibodies. The detection procedure involves an amplification stage (with alkaline phosphatase and colorimetric substrates) in order to increase the signal. Intensity of staining on a nylon membrane for a given probe can be linearly correlated with the concentration of control RNA Hence semi-quantitation of the stain intensity can be carried out if standardised protocols are followed in the hybridocytochemistry procedure A complementary approach is to measure the signal density. The latter method allows relative estimates of the area occupied by the signal grains To ensure semi-quantitation the following steps were incorporated into the design of the experiments: i) Since DIG-UTP cRNA probes are stable for months, the same labelled probe was used throughout the study hence limiting the influence of variable probe activity. ii) Tissue from patients at the three trial stages were processed together in the same in situ hybridisation experiment. Quenching of the reaction product was carried out at the

168 same time for all slides in a given run. Stopping the colour-development reaction was carried out at a time (15 hours) when, though the linear increase in stain intensity had not yet plateaued yet was of a measurable density without the background signal being of consequence. iii) Serial sections on separate slides were processed using either anti-sense or sense probes. Subsequent computerised image analysis was carried out on the same fibres in each section (see EM. 5.). Details for chemicals and reagents in the following sections are to be found in Appendix G.4.3.

EM.4.2. Probe characteristics The synthesis and characterisation of a 280 base pair (bp) IGF-I RNA probe had been previously described and established and did not constitute a part of this study. The previous work had shown, for rabbit muscle, the hybridisation of the probe with the two prominent IGF-I mRNA species, 1.2kb and 7.5kb long. The cDNA, covering exon 3 and part of exon 4 of the IGF-I gene, was subcloned into a phagemid vector (Stratagene) containing T 3 and T 7 promoters. It was now possible to synthesise labelled sense and anti-sense RNA probes by in vitro transcription, with RNA polymerase using DIG-UTP as substrate.

EM.4.3. Probe labelling The DNA to be transcribed was already subcloned into the polylinker site of a transcription vector (multiple cloning site adjacent to the RNA polymerase promoter in the pSPTlS or pSPT19 transcription vector). Digestion of the plasmid with a restriction enzyme linearized the DNA. After purification and establishing the concentration of linearized DNA, 'run off transcripts of uniform length were synthesised by using this linear DNA template. These transcripts have the DIG molecule incorporated enzymatically into the riboprobe during the in vitro transcription stage. The success of the transcription was verified by checking the concentration of RNA present, and DIG incorporation was established immunochemically with an anti- digoxigenin antibody conjugated to alkaline phosphatase. For a detailed description of the probe labelling procedure see appendix G.4.2.

169 EM.4.4. In situ hybridisation protocol EM.4.4.1. Pretreatment of sections Slides were thawed at room temperature before unwrapping to avoid ffeeze-thaw artefacts then fixed in fresh 4% paraformaldehyde (20 min at 4°C). Steps to partly denature the tissue and hence increase accessibility and improve signal to noise ratio were carried out. These consisted of incubations in 0.2N HCl (20 min at room temp.), in 2x sodium chloride-sodium citrate buffer (2xSSC; 15 min at 70°C), and in pronase (7 min at 37°C), with 2 min IxPBS washes in between. The digestion by pronase was stopped with glycine (30 sec at room temp ). The slides were then fixed once again in fi'esh 4% paraformaldehyde (5 min at 4°C), fixation being blocked by 3xPBS (5 min at room temp.) and IxPBS (2x30 sec at room temp ). After rinsing, slides were treated with acetic anhydride in order to cover the section with negative charge to repel nucleic acids. This was accomplished by initially equilibriating slides in fresh triethanolamine bufifer (O.IM, 2 min at room temp ), followed by incubations in 0.25% acetic anhydride containing fresh triethanolamine buffer (5 min at room temp ), and 0.5% acetic anhydride containing fresh triethanolamine buffer (5 min at room temp ). Acétylation was blocked by 2xSSC (5 min at room temp ). The above steps were carried out with gentle agitation. Preliminary experiments had established that a pre-hybridisation incubation, with hybridisation buffer but without probe, did not reduce background staining, hence, a pre-hybridisation step was not included in the protocol.

EM.4.4.2. Hybridisation Adjacent slides were divided for labelling with either sense or anti-sense probes and fi-om this stage processed separately. Additional slides were also included as negative controls. These were treated identically except for either no addition of probe or an addition of 100-fold excess of unlabelled probe. The previously labelled probe (see appendix G.4.2:, probe concentrations were 107.5ng/pl for anti-sense, 95.47ng/pl for sense) was diluted in hybridisation buffer to a concentration (lOOOng/ml) which had been shown to maximise the nucléation reaction. After drawing around the sections with a paraflSn pen (Dako, UK), diluted and vortexed hybridisation bufifer was applied (55 pi per section). Sections were coverslipped and slides placed in a container

170 humidified with 2xSSC. The humidified container was placed in an hybridisation oven at 68 °C for 60 min, afl;er which the oven temperature was lowered to the anneal temperature (43.5 °C, 15 hours).

EM.4.4.3. Post-hybridisation washes For the subsequent stages, though sterile solutions were used, it was no longer necessary to maintain RNase free conditions as the probe-mRNA duplex is stable. Coverslips were removed and slides washed in 2xSSC (15 min at room temp ). Steps to remove non-duplex sequences, unbound probe, and non-specific probe-mRNA sequences, were carried out. Incubation with RNase (30 min at 37°C) was followed by three 30 min incubations in SSC (IxSSC, 0.5xSSC, 0.5xSSC at room temp., 47°C, and room temp., respectively). Slides were then rinsed in DIG wash buffer (1 min at room temp.) and equilibriated in DIG buffer 2(30 min at room temp ).

EM.4.4.4. Immunological detection Sections were incubated in a humidified chamber (60 min at room temp.) with lOOpl per section of an alkaline phosphatase conjugated anti-DIG sheep polyclonal antibody (diluted 1:500 in DIG buffer 2). Two washes in DIG buffer 1(15 min at room temp.) were followed by a rinse in DIG buffer 3 (2 min at room temp ). Colour detection reagent was prepared immediately prior to usage. Each section had lOOpl of colour detection reagent applied and the slides incubated in the dark in a humidified chamber (15 hours at room temp ). Slides were finally washed in DIG buffer 4(30 min at room temp.) and mounted in glycerin jelly.

EM.5. Imaging methods and quantitation A minimum of 100 fibres of each fibre type, per slide stained for MHC, were manually counted, and the result expressed as a percentage of type I fibres. The KS400 imaging system (Kontron, UK), interfaced to an Olympus Vanox AH-2 fluorescence microscope with a Photonic Science low light camera, was utilised. Images were stored of the same fibres stained with type I MHC antibody, IGF-I anti-sense probe, and IGF-I sense probe. Three fields from one section per slide were measured, in contrast to the fibre proportion analysis which incorporated a much larger field of analysis.

171 A program was written with Kontron KS400 software for the measurement of fibre area and stain density. Loops in this macro allowed the measurement of these parameters in the same field for both fibre types. The experimenter was blind regarding treatment stage of tissue or sense/anti-sense section. The fibre area data were corrected for fibres in a non-transverse orientation. ISH signal was expressed in terms of stain area in a given field. Fibre size and number data, for the same field, allowed the expression of ISH signal as stain area per fibre area, or stain area per fibre.

EM.6. Reverse transcriptase polymerase chain reaction methodology Total RNA was extracted and subjected to RT-PCR with two oligonucleotide primers (IGF-IEa and IGF-IEc). These primers enabled the selective amplification of either the IGF-IEa or IGF-IEc cDNA Both the IGF-IEa and IGF-IEc transcripts are expressed in stretched rabbit skeletal muscle, undergoing hypertrophy. The IGF-IEa transcript is also present in normal muscle Hence, the Ec-peptide is believed to be a local growth regulator which is switched on by mechanical rather than hormonal stimuli. The two types of cDNA differed by the presence (IGF-IEc) or absence (IGF-IEa) of a 52bp insert The 5' primer sequence (5 GCTTGCTCACCTTTACCAGC3 ) was common to both probes (complementary to exon 3 of the IGF-I gene), whilst the sequence of the 3' primer for the IGF-IEa probe was 5'AAATGT ACTTCCTTCTGGGTCT3 ' (complementary to parts of exons 4 and 6). For the IGF-IEc probe, the 3' primer sequence was 5 TTGGGCATGTCAGTGTGG3 ' (complementary to parts of exons 5 and 6), being different to the IGF-IEa sequence due to the 52bp insert. The RT-PCR reaction was performed for 35 cycles in a thermal block (Hybaid) with sequential dénaturation (94 °C, Imin), annealing (65 °C, Imin) and elongation (72 °C, Imin), before a final elongation step for 7min at 72 °C. The product of the RT-PCR reaction was separated on a 1.5% agarose gel. Results are presented for only three patients as the aim, of this experiment, was to establish the presence of IGF-I mRNA in the muscle by an alternative methodology to ISH. Hence, the data from ISH is also presented for each patient, together with the outcome from the RT-PCR.

172 Chapter 5 Results

ER.1. Muscle fibre proportions The percentage of type I fibres was initially examined for data dispersion and found not to follow a normal distribution (figure ER.1 a). Transformation of the data, by either taking logarithms or square roots (results not shown) failed to achieve a normal distribution. The data were therefore analysed with the non-parametric K-W ANOVA method. There were no significant changes in the proportion of type I and type II muscle fibres over the trial period (figure ERIb). Analysis of data, by expressing data from each patient as a percentage of the baseline values, also failed to show any difference between groups (data not shown). When the data were examined for gender differences, again, no significant differences were observed (figure ER.Ib). However, in females, a trend was observed for a decline (18.5%) in the percentage of type I fibres over the trial period. Examination of data for differences due to age of patients failed to reveal any influence of this parameter. Similarly, dividing data into those patients who agreed to continue with post-trial r-hGH therapy, and those who did not (six patients for, four against) did not reveal any changes. Finally, fibre proportions in euthyroid subjects were compared with those in patients receiving thyroxine replacement, and no differences were observed (data not shown).

Summary ► Fibre proportions are not affected by r-hGH therapy. ► Females appear to exhibit a reduction in type I fibres over the trial period, however, this was not significant.

173 7

S 5 (0 o .

20 30 4 0 5 0 60 7 0 % type I fibres Observed expected (normal)

Figure ER.la Graph showing data distribution for the proportion of type I fibres (samples from patients at the three trial stages). Data expressed as a percentage of type I fibres. Expected normal distribution also shown.

^ 2 0

all male female

GHOOm P-6m g GH-6/12m

Figure ER.lb; Histogram showing the proportion of type I fibres for all patients and for patients according to gender. Data expressed as a percentage of type I fibres. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m -patients after six or twelve months r-hGH. Error bars show SEM.

174 ER.2. Muscle fibre area The data were not normally distributed (figure ER.2a). Although logarithm transformation of the data failed to achieve a normal distribution, the square root transformation of data method, resulted in an approximately normal distribution (Shapiro-Wilk statistic=0.78,//gwrg ER.2h). The transformed data were analysed with the parametric ANOVA method. The size of individual muscle fibres, for both fibre types, was smaller (10-12%, p=0.045) in the placebo group relative to the baseline and r-hGH treated groups {figure ER2c). Thus, r-hGH treatment, in relation to the placebo group, showed an increase in muscle fibre area. These changes were not statistically obvious when data were examined in relation to individual fibre types though a similar trend existed {figure ER.2c). Gender differences were not observed, for the transformed data, in relation to r-hGH therapy for either fibre type (data not shown). However, males demonstrated larger fibre areas (13.6%, p=0.0004) than females for fibre types as a whole, and also for each fibre type (type I 13.9%, p=0.006; type II 13.3%, 022). figure ER.2d A trend existed for type I fibres to be greater in size {5Vo, figure ER.2d) than type II fibres, however these were not significant differences. Further analysis, after transformation of data to reflect changes from baseline, revealed a similar trend. Following two-way ANOVA analysis, there was no interaction between gender differences and those observed due to trial group or fibre type. No evidence was found to suggest variations in fibre size between patients who agreed to continue on r-hGH and those who did not, or between euthyroid subjects and those receiving thyroxine therapy (data not shown). Examination of data for differences due to age were only possible for the 35-44y and 45-53y age brackets, due to small patient numbers. Fibre area, for both fibre types, in the 35-44y bracket was greater (12.1%, ^<0.00\, figure ER.2e) than that in the 45-53y bracket. The latter effect was also observed for the fibre types when analysed separately (type I fibres 11.7%, p=0.005; type II fibres 12.6%, ^=0.0\9, figure ER.2e). Following two-way ANOVA analysis, no interaction was observed for the age related decline in fibre size and the trial group status of the patients. However, a significant interaction existed between age and gender of patients, for fibre size in both fibre types (age x gender, p=0.012). The latter effect was evaluated by analysing male and female data for

175 fibre size separately, with age as a variable. A more marked decrease existed in females as compared to males, for fibre size atrophy with age (males 7.9% decrease with age, p=0.028; females 25.1% decrease with age, p<0.001). The small numbers involved, however, meant that the latter result had to be treated with a degree of scepticism and has thus not been graphically presented. A positive correlation was found for serum IGF-I levels and fibre size (Spearman correlation =0.343, p=0.013). Conversely, a negative correlation was present for fibre area and serum thyroxine concentrations (Spearman correlation =-0.246, p=0.046).

Summary ► Fibre size is reduced in patients during the placebo arm of the study as compared with the baseline GH-deficient subjects. ► Fibre size exhibits an increase in the r-hGH treated subjects as compared to the placebo group. ► Males have larger fibre areas than females. ► Older subjects (45-53y) have smaller fibre areas than younger patients (35-44y). ► There is a positive correlation between serum IGF-I levels and fibre size, whilst serum thyroxine values are negatively correlated with fibre area.

176 16

S JD 6

0 2000 3000 4000 5000 7000 $000 9000 10000 fibre area (micrometers) observed expected (normal)

Figure ER.2a; Graph showing data distribution for type I and II muscle fibre area (samples from patients at all three trial stages). Expected normal distribution also shown.

corrected fibre area observed expected (normal)

Figure ER.2b; Graph showing data distribution for type I and II muscle fibre area. following square root correction (samples from patients at all three trial stages). Expected normal distribution also shown.

177 100

m 60

type l&ll type I type II

GHCMXn P-6m GH-8/12m

Figure ER.2c: Histogram showing muscle fibre area (following square root transformation). Results are shown for individual fibre types and for pooled data from both fibre types. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m -patients after six or twelve months r-hGH. Error bars show SEM.

100

1 8 (TJ 2

all male female

typel&ll type I type II

Figure ER.2d; Histogram showing fibre type and gender differences for fibre area (square root transformed). Error bars show SEM, ***p<0.001, **p<0.01, *p<0.05.

178 100

T3 g 60 8 g (0 40 £12

type type I type II

46-63y

Figure ER.2e: Histogram showing age differences for fibre area (square root transformed). Error bars show SEM, ***p<0.001, **p<0.01, *p<0.05.

179 ER.3. In situ hybridisation ER.3.1. In situ hybridisation signal localisation and distribution The ISH signal for IGF-I message displayed a cytoplasmic location for muscle fibres (see plates 7ii & iii\ similar to that seen for paralyzed or stretched rat skeletal muscle, processed with a similar methodology The staining pattern was consistent across muscle fibres in a section, though individual fibres did show differing levels of staining compared to neighbouring fibres. The signal was generally present as grains arranged along longitudinal stacks (see plate 7iii). Additional ISH signal was distributed in connective tissue, interstitial cells, blood vessels, and intramuscular nerves (see plate 7i). The expression of the IGF-I gene by the muscle fibres themselves was strongly indicated by the distribution of the ISH signal. The pattern of staining with the sense probe was similar to that with the anti-sense probe, though of a reduced density (see plates 7i, ii & iii). Fibre size was found to correlate with the density of ISH signal (for both fibre types; ISH signal and fibre area, Pearson correlation=0.34, p=0.012). A common observation was that smaller diameter fibres expressed more ISH signal than that present in larger fibres (see plate 7iii). Since individual fibre signal values were pooled (with the remainder of values from a section), quantification of this phenomenon was not possible.

ER.3.2. In situ hybridisation signal data expression, distribution and statistical analysis ER.3.2.1. Data expression Initial analysis of the raw data revealed a marked inter-experiment difference in stain density. Thus, for non-specific/background staining, there was a 4 to 52% difference in staining, with the sense probe, in between separate experimental runs (data not shown). Tissue from each patient was processed in one batch, to avoid problems in interpretation. In order to allow group comparisons, two complementary approaches were carried out: i) Inter-patient variation correction; data were expressed as a percentage of the baseline GH-deficient values. This approach results in the loss of data spread for the

180 baseline group and therefore may result in loss of information regarding contributory factors. ii) Inter-experiment correction; both sense and anti-sense data were expressed relative to the sense data for that experimental group with the highest level of sense staining. The scaling factor for this correction was obtained from the following calculation; 1- (a b) = scaling factor, where a is data from experiment with low sense staining, where b is data from experiment with highest sense staining.

Though the signal with the sense probe was generally weak, in order to obtain a true estimate of the IGF-I mRNA, for each patient, the signal with the anti-sense probe was subtracted from that present for the sense probe. Hence, two sets of data were obtained after this subtraction; those originating from the scale factor corrected data and those derived from the original sense/anti-sense data. These measurements were then expressed relative to the morphological parameters of muscle fibre count or muscle fibre area: (i) S ISH stain area/ S number of fibres, i.e. signal per fibre (S/F). (ii) S ISH stain area/ S fibre area, i.e. signal per fibre area, expressed as a percent (S/FA).

The experimental groups thus obtained were as follows: (i) S/F, from scale factor corrected data. (ii) S/F as percentage baseline, from original data. (iii) S/FA from scale factor corrected data. (iv) S/FA as percentage baseline, from original data.

The above procedures for data expression are summarised in Chart ER.3 a.

181 ISH SIGNAL ISH SIGNAL a-SENSE SENSE a-SENSE SENSE express relative to sense correction factor a-sense minus sense=signal (S) a-SENSE SENSE S per fibre area S per fibre 1 (S/FA) (S/F) a-sense minus sense=signal (S)

express relative to baseline group S per fibre area S per fibre (S/FA) (S/F)

Chart ER-3a: Flow chart demonstrating the two methods adopted for data expression. Left panel; Inter-experimental correction. Right panel: Inter-patient correction.

182 ER.3.2.2. Distribution of the data The data from either anti-sense or sense staining were shown not to follow a normal distribution either before or after scaling factor conversion (only anti-sense data for scale factor corrected data is figure ER.3a for S/F, figure ER.3b for S/FA). Transformation of the raw data by taking logarithms or square roots (results not shown) failed to achieve a totally normal distribution. However, the Shapiro-Wilk normality test did indicate a near normal distribution, following log transformation, for both sense and anti-sense data (Shapiro-Wilk statistic 0.32-0.8). The data were therefore analysed with both non-parametric and parametric (for log corrected data) methods of statistical analysis.

ER.3.2.3. Results of statistical analysis ER.3.2.3.1. Length of sample storage analysis All muscle biopsy samples had been stored for a period of sixteen to thirty months at -80 °C before ISH analysis making it important to consider the possible effects of storage on mRNA. It was not possible to investigate mRNA degradation in these samples for the first sixteen month period. However, it was possible to extrapolate the rate of an exponential decay in mRNA, if such a decay was indeed shown to be present, for samples stored between sixteen and thirty months. Storage at temperatures below -70 °C is known not to result in mRNA degradation for years In any case, even if some degradation has occurred, it should not be a problem as mRNA need not be completely intact in order to be detected by ISH The usage of a housekeeping gene, whose levels would remain stable over the trial stages, was not considered absolutely necessary since the design of the study, with an internal control group (placebo treated), should have ensured that anomalies, due to sample degradation effects, would have been identified. The length of storage was divided into three equal periods (16-20m, 21-25m and 26- 30m). Three separate evaluations of signal degradation were made: i) ANOVA; comparing the signal density for ISH staining revealed no difference in signal betweeen samples stored for between sixteen and thirty months at -80 °C (for S/FA, figure ER. 3 c).

183 ii) Regression analysis; since changes in ISH signal with r-hGH therapy may have obscured sample degradation effects, regression analysis, with length of storage as a covariant, was carried out on logarithm transformed patient groups (results shown for log S/FA, log S/F, log of non-scale factor corrected a-sense and sense data). The percentage baseline groups were not utilised as the same values for ISH signal at baseline were bound to skew the data. Data is presented for the regression significance and for as a percentage {table ER.3d). The latter indicates the contribution of the covariant to the main effect (in this case r-hGH treatment). There was no evidence from the regression analysis for an influence of length of storage on ISH signal. Though data presented is for r-hGH treatment, with six and twelve months treatment considered separately, analysis with treatment values as one group showed very similar results. iii) Regression analysis; this additional regression analysis was carried out with covariant being months of storage {table ER.3a). There was a possible contribution of length of storage to the log S/FA data, however, this was not significant (p=0.077).

184 22 1

20 -

ü) 0)C ro Q. o JDQ) 3E c

100 200 300 500 600 700 S/F Observed expected (normal)

Figure ER.3a: Graph showing data distribution for ISH signal in both type I and II fibres in muscle taken from all trial stages. ISH signal expressed as signal per fibre (S/F). Expected normal distribution also shown.

10 Q. 8 n0) E 6 C3

0 0 2 3 5 S 8 %S/FA observed expected (normal)

Figure ER3b: Graph showing ISH signal data distribution for both type I and II fibres in muscle taken from all trial stages. ISH signal expressed as signal per fibre area (S/FA). Expected normal distribution also shown.

185 Storage/3 Storage/3 Months Months

P= R' (%) = P= R '(% ) = log S/FA 0.133 4.8 0.077 6.6 log S/F 0.144 4.6 0.9 0.03 log «-sense S/FA 0.219 3.3 0.307 2.2 log sense S/FA 0.584 0.66 0.854 0.07

Table ER.3a; Results from an ordinary least squares regression for length of sample storage. The analysis was carried out with factor levels assigned for baseline GH- deficient patients, patients after six months placebo, patients after six months r-hGH therapy, and patients after twelve months r-hGH therapy. Dependent variables were logarithm and scale factor corrected data for S/FA and S/F, as well as logarithm corrected values for the non-scale factor corrected «-sense/sense S/FA data. Regression covariants were length of storage divided into three equal periods (Storage/3) and length of storage expressed as months for which samples were stored.

I 16-20ltl ^ 21-25m ^ 26-30m I

Figure ER.3c; Histogram showing differences in ISH signal with length of storage (months). ISH signal expressed, for both fibre types, as signal per fibre area (S/FA). Error bars show SEM.

186 ER.3.2.3.2. Trial group differences There was no statistical evidence, with parametric or non-parametric ANOVA methods, for a change in muscle fibre IGF-I message expression, with either placebo or r-hGH therapy in GH-deficient adults. This was true for all methods of data expression outlined earlier with the appropriate test. However, for S/FA (scale factor corrected or percentage baseline, ER.3d) and S/F (percentage baseline, ER.3e), there was a consistent trend, for the r-hGH treated groups, to display a greater ISH signal than the baseline/placebo groups (9.6-49%). This was found to be statistically significant for the twelve month r-hGH group (for both fibre types log S/FA, GH-12m 34.8%>GH-0m, t-test ^=0.0'^!, figure ER.3f plate 7ii & iii).

187 250

200

«150 c en (A I ^ 100

S/FA S/FA %b Œ GHD-Om P-6m

Figure ER.3d: Histogram showing ISH signal for both type I and II fibres. ISH signal displayed as signal per fibre area (S/FA), and as signal per fibre area and expressed as a percentage of baseline (S/FA %b). Note; Y-axis reflects, for S/FA, scale factor corrected values for ISH signal (100 added to all values for axis scaling), and for S/FA %b, a percentage scale derived from original ISH signal data. GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH. Error bars show SEM.

188 250

200

«150 c (Ag> X ^ 100

S/F S/F %b V 7 ^ GHD-Om P-6m GI-P6/12m □

Figure ER3e: Histogram showing ISH signal for both type I and II fibres. ISH signal expressed as signal per fibre (S/F), and as signal per fibre expressed as a percentage of baseline (S/F %b). Note: Y-axis reflects, for S/F, scale factor corrected values for ISH signal, and for S/F %b, a percentage scale derived from original ISH signal data GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH. Error bars show SEM.

189 2.5

I(Q 1.5 (OI i 1

0.5

log S/FA log S/FA %b [ GHOOm P-6m GH-6m GH-12m GH^12m ]

Figure ER.3f: Histogram showing ISH signal for type I&II fibre types following logarithm transformation. ISH signal demonstrated as signal per fibre area (S/FA), and as signal per fibre area expressed relative to the baseline group (S/FA %b). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH Note: for S/FA, logarithmic transformation resulted in some groups having negative values, hence, a value of one has been added to all groups in the log S/FA segment above Error bars show SEM.

190 ER.3.2.3.3. Fibre type differences For S/FA, expressed either as scale factor corrected or as a percentage of baseline, type I fibres demonstrated a greater (37%, K-W ANOVA p=0.037 and 43%, p=0.036 respectively, ER.3g, ER.3h and summarised in ER.3i, plate 7ii & iii) signal density than type II fibres. A similar difference was seen following logarithm transformation for S/FA data (ANOVA p=0.075 & p=0.039, for scale factor corrected and percentage baseline respectively), and for S/F data (ANOVA p=0.04 & 0.039, for scale factor corrected and percentage baseline respectively).

ER.3.2.3.4, Gender differences The most obvious feature for gender differences was a 20% greater level of ISH signal in males, as compared to females, for type II fibres (for log S/FA expressed as a percentage of baseline, ANOVA figure ER.3i). For both fibre types together, an elevated signal in males (12%) was also observed, though found not to be statistically significant with the ANOVA method (for log S/FA expressed as a percentage of baseline, ANOVA Oil, figure E3i\ for log S/F, t-test p=0.042).

ER.3.2.3.5. Interactions between variables For the logarithm converted data it was possible to examine, utilising two-way ANOVA, the interaction between gender and fibre type. There was no interaction between these two variables for any of the appropriate groups, suggesting that these were separate effects influencing the level of IGF-I message expression.

ER.3.2.3.6. Other parameters Examination of data for differences due to age of patients, failed to show any influence of this variable (data not shown). Dividing data into those patients who agreed to continue with post-trial r-hGH therapy, and those who did not, did not reveal any changes for IGF-I message (data not shown). Two-way ANOVA analysis for S/FA (type I&n fibres, logarithm transformed), revealed some interaction between thyroxine status and trial group (p=0.057). This was perhaps due to a 65.4% greater IGF-I message in normothyroid subjects as compared to patients on thyroxine replacement (p=0.069, data not shown).

191 There was no correlation between circulating IGF-I and muscle fibre IGF-I mRNA expression. Similarly, T3 and T4 levels were not correlated with ISH IGF-I expression.

ER.3.2.3.7. Results summary ► Length of sample storage does not affect the level of ISH signal. ► An increase in IGF-I mRNA signal is observed following twelve months r-hGH therapy as compared to the baseline levels. ► Type I fibres display a greater IGF-I ISH signal than type II fibres. ► Males tend to have higher levels of IGF-I mRNA than females; this is statistically significant for type II fibres. ► Patients on thyroxine replacement appear to have reduced levels of muscle IGF- I message than subjects who are euthyroid; however, this is not a statistically significant finding.

192 type I/ll type II

^ GHD-Om GH-6m l-12m g GH-6/12m |

Figure ER.3g; Histogram showing ISH signal for fibre types expressed as signal per fibre area (S/FA). GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/l2m =patients after six or twelve months r-hGH Error bars show SEM, *p<0.05.

500

400

300

I 200

100

type type type II GHD-Om P-6m GH-6m GH-12m GH-6/12m ]

Figure ER3h: Histogram showing ISH signal for fibre types. ISH signal expressed as a percentage of baseline for signal per fibre area (S/FA). GHD-Om=patients before r- hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH. Error bars show SEM, *p<0.05.

193 2.5

1.5 (A 1

0.5

all male female [ type I/ll typel typgii I

Figure ER3i; Histogram summarising overall fibre type and gender differences for ISH signal The ISH signal demonstrated as the logarithm transformed values for signal per fibre area (S/FA), expressed as a percentage of baseline. Error bars show SEM, *p<0.05.

194 Plate 7i Photomicrograph of human skeletal muscle demonstrating IGF-1 message expression byin situ hybridisation.

A: Serial muscle section demonstrating «-sense signal in the region of a blood vessel (arrow). B: Serial muscle section demonstrating sense signal in the region of a blood vessel (arrow; same area as in A). C: Serial muscle section demonstrating «-sense signal in the region of a nerve trunk (arrow). D: Serial muscle section demonstrating sense signal in the region of a nerve trunk (arrow; same area as C). E: Muscle section demonstrating «-sense signal at the edge of muscle fibres (arrowhead). F: Muscle section demonstrating «-sense signal at the edge of muscle fibres (arrowhead).

Note: Magnification for fields A,B,C and D =xlO, scale bar =100pm. Magnification for fields E and F =x40, scale bar =25pm. Tissue from baseline GH-deficient trial subjects. The tissue was processed as described in EM.4.4. except for a reduction in the pre-treatment incubation with 0.2N HCl (lOmin instead of 20min). A longer pre-treatment period was required to demonstrate muscle fibre cytoplasmic staining probably due to the presence of myofibrils inside skeletal muscle fibres

195 V - , ^

Iff,?' fk i- ■:. . . ■•« . ■- p v . ■ <*..V^*V - * ' ' '

i'- < m - %,

______r îflT " S'#* <

' - y # •'î?.>:vA'r:'':_ ,

• T Plate 7ii Photomicrograph of skeletal muscle from a baseline GH-deficient subject demonstrating myosin heavy chain and IGF-I message expression.

A: Serial muscle section demonstrating M HCI immunostaining (arrowhead). B: Serial muscle section demonstrating a-sense signal (arrowhead; same fibre as in A). C: Serial muscle section demonstrating sense signal (arrowhead; same fibre as in A).

Note: MHC type I fibres display greater ISH signal than type H fibres (not stained in A). Magnification in all fields =x20. Scale bar =50pm.

196

Plate 7iii Photomicrograph of skeletal muscle from trial subjects following twelve months r-hGH therapy (demonstrating myosin heavy chain and IGF-1 message expression).

A: Serial muscle section demonstrating MHC I immunostaining (arrowhead). B: Serial muscle section demonstrating «-sense signal (arrowhead; same fibre as in A). C: Muscle section demonstrating «-sense signal (arrow =small diameter fibre). D: Muscle section demonstrating «-sense signal (longitudinal section)

Note: For A and B, MHC type I fibres display greater ISH signal than type H fibres (not stained). A & B are mirror images. For C, small diameter fibres (arrow) display greater ISH staining than larger fibres. For D, note stacking arrangement of IGF-I message (arrow). Magnification in all fields =x20. Scale bar =50pm.

197

ER.4. Reverse transcriptase polymerase chain reaction results and comparison with in situ hybridisation signal ER.4.1. Patient A The subject was a 48y male (see table BR.L, patient 11), on placebo for six months, then on r-hGH for the next six months. In both baseline and placebo muscle samples, there was no demonstration of the appropriate mRNA transcripts, with either the IGF- lEa, or IGF-IEc probes. Following six months of r-hGH therapy, appropriate-sized products were observed for only the IGF-IEa probe (301 bp for IGFI-IEa, 353bp for IGF-IEc, figure ER.4a). The ISH signal included data from both fibre types and was expressed as the signal per fibre area. The baseline and placebo values were similar, whilst the six month r-hGH treated sample showed about a 40% increase in IGF-I signal over the former groups {table ER.4a).

ER.4.2. Patient B This was a 38y male subject (see table BR.L, patient 2) who received r-hGH for the full duration of the study. With the baseline sample, no RT-PCR product was visible with either the IGF-IEa, or IGF-IEc probes. For both the six month and twelve month specimens, PCR products, of the appropriate size were observed with the IGF-IEa probe (figure ER.4b). The ISH signal was low in the baseline group and increased after six months of r-hGH treatment. A further six months of r-hGH therapy resulted in a further 34.5% increase in signal per fibre area (table ER.4b)

ER.4.3. Patient C The subject was a 45y male (see table B R L , patient 1) who received r-hGH for the full duration of the trial period. In contrast to patients one and two, the baseline sample for patient three did demonstrate RT-PCR products with the IGF-IEa probe at this trial stage (figure ER.4b). Following six months of r-hGH therapy, IGF-IEa and IGF-IEc products were not detectable in samples from this patient (figure ER.4b). This patient showed the highest level of ISH signal at the baseline stage as compared with the other nine biopsy samples. Following six and twelve months of r-hGH therapy

198 there was over a 60% decrease in ISH signal as compared with the baseline sample {table ER. 4b).

199 301bp

Figure ER4a; Result of RT-PCR for patient A; on placebo for six months, then on r- hGH for the next six months. Patient A was a 48y male subject. Lane 1: Ikb DNA ladder Lane 2: GHD-Om (IGF-IEc) Lane 3: GHD-Om (IGF-IEa) Lane 4; P-6m (IGF-IEc) Lane 5 : P-6m (IGF-IEa) Lane 6; GH-6m (IGF-IEc) Lane 7: GH-6m (IGF-IEa) Lane 8: no cDNA (IGF-IEc) Lane 9: no cDNA (IGF-IEa)

Note. A 301 bp product in lane 7 (specific for IGF-IEa).

GHD-Om=patient before r-hGH treatment, P-6m=patient after six months placebo, GH- 6m =patient after six months r-hGH.

200 Type I&n fibres GHD-Om 2.43 P-6m 2.52

GH-6m 4.35

Table ER4a: ISH signal for patient A; on placebo for six months, then on r-hGH for the next six months. ISH signal expressed as signal per fibre area.

GHD-Om=patient before r-hGH treatment, P-6m=patient after six months placebo, GH- 6m =patient after six months r-hGH.

Type I&H fibres Type I&H fibres (patient B) (patient C) GHD-Om 0.32 7.12 GH-6m 2.73 2.71 GH-12m 4.17 1.9

Table ER.4b: ISH signal for patients B and C; on r-hGH for six months, then on r- hGH for a further six months. ISH signal expressed as signal per fibre area.

GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH.

201 1 2 3 4 5 6 7 8 9 10 11 12

Figure ER4b; Result of RT-PCR for patients B and C; both patients were on r-hGH for twelve months. Patient B was a 3By male subject and patient C was a 45y male subject. Lane 1: Ikb DNA ladder Lane 2: GHD-Om (IGF-IEa) Lane 3: GHD-Om (IGF-IEc) Lane 4: GH-6m (IGF-IEa) Lane 5: GH-6m (IGF-IEc) Lane 6: GH-12m (IGF-IEa) Lane 7: GH-12m (IGF-IEc) Lane 8: Ikb DNA ladder Lane 9: GHD-Om (IGF-IEa) Lane 10: GHD-Om (IGF-IEc) Lane 11 : GH-6m (IGF-IEa) Lane 12: GH-6m (IGF-IEc)

Note: A 301bp product in lanes 4, 6 and 9 (specific for IGF-IEa) A 353bp product in lanes 5 and 7 (specific for IGF-IEc)

GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m =patients after six or twelve months r-hGH.

202 Chapter 5 Discussion

ED I. Muscle fibre proportions The general descriptions, for vastus lateralis muscle fibre proportions, imply that the ratio of type I to II fibres is 1:2, or 1:1.1 for type I IIA IIB However, a broad range of adult human studies have indicated that the percentage of type I fibres, for this muscle, may be over 40%. Hence, values of -50% (78 subjects of varying physical activity, aged 27y) 39.1-51% (12 physically active but not resistance trained subjects, aged 20.65±1.5y) and 47.5-52.3% (21 sedentary subjects, aged 20.1- 21.3y) have been observed. In addition, control subjects matched to GH-deficient patients, also displayed type I fibre proportions of 45% and 46.3% (39y and 42.5y, respectively, see table ED.l) Hence, the fibre proportions in the present study fall within the normal range, as well as being consistent with values from other studies on GHD (table ED.l). Preece et al ^ and Rutherford et al reported no change in fibre proportions, after cessation of r-hGH therapy, whilst Whitehead et al and Cuneo et al found no effect on fibre proportions, following r-hGH substitution. The present study has replicated data from the aforementioned reports in terms of lack of change in fibre

proportions following r-hGH treatment. Is there a role for GH in adult GH-deficient subjects in relation to muscle fibre proportions? Treatment with r-hGH therapy results in an increase in MHC mRNA levels in nutritionally depleted subjects Further, the demonstration by Rutherford et al of an increase in contractile properties in GH- deficient adults is consistent with an elevation in type II fibres. The lack of change in this and other biopsy studies therefore suggests either a methodological problem due to small tissue sampling or lack of fibre type subclassification. Alternatively, effects of GH may have been masked by other influences. A non-significant increase in type I fibres was observed by Whitehead et al in the placebo arm of their study (table ED.l). Hence, with ten patients, a trend was observed for an increase in type I fibres in GHD, whilst no change was seen with r-hGH treatment (table ED.l). Conversely, Doerga et al in a preliminary report, observed far lower type I fibre levels in GH- deficient adults, as compared with control subjects (table ED.l).

203 The most likely candidate for an alternative influence to GH on muscle fibre type expression is thyroid hormone. The present study has attempted to infer the influence of thyroid hormone given the well-established effect of this hormone on MHCs and fibre types, as well as the known alteration in thyroid hormone levels with r-hGH treatment in GH-deficient adults 222,528-530 It has been shown in hypophysectomised rats that thyroxine stimulation of skeletal growth is independent of GH and IGF-I Further, thyroxine can alter MHC gene expression in both cardiac and skeletal rat muscle Hence, hyperthyroidism in rats is associated with a reduction in type I fibres and an increase in intermediate and fast isoforms ^25,529,530 mechanism by which thyroid hormones influence MHC gene expression is believed to be via T3 activation of the muscle-specific myogenin family and also by transcriptional effects of T3 on MHC promoters ^^ 2,533 gaudies in long-term hypophysectomised rats (500-900 days), and mutant dwarf mice, indicate an increased percentage of type I fibres in soleus and biceps branchii muscles, respectively Conversely, short-term hypophysectomised rats (14 days) were shown to have a reduced proportion of type I fibres in soleus and extensor digitorum longus muscles Similarly, type I MHC mRNA was reduced in the gastrocnemius muscles of short-term hypophysectomised rats (14 days) Shorey et al have explained some of these differences as due to thyroxine being present at near normal levels in the short-term hypophysectomised rats. Hence, lack of GH in the short-term hypophysectomised rats would result in a reduction of type I fibres. The long-term effect is described as due to thyroxine deficiency resulting in an increase of type I fibres Further, thyroid stimulating hormone injection failed to alter changes in short-term hypophysectomised rats, as compared with r-hGH injections which resulted in an increase of type I fibres Interestingly, Loughana and Bates reported that their results following hypophysectomy were far more pronounced in three week, compared to two week, hypophysectomised rats. The aforementioned studies suggest that the levels of thyroid hormones may influence or mask effects due to GHD or r-hGH therapy. This is also apparent from the finding in rats that MHC adaptation occurs in a hierarchial manner, with thyroid hormone status predominating over increased functional demand The direction of such an effect would depend on the physiological levels of the thyroid hormones. Hence, in the

204 present study, despite significant reductions in serum thyroid hormone levels, none of the patients approached subnormal levels at any stage. Also, analysis of fibre proportion data, to distinguish between patients who received thyroxine and euthyroid subjects, did not reveal any difference between these groups. This would suggest that the 'normal' effect of thyroid hormones, that of decreasing the proportion of type I fibres, would predominate. In summary, both GH and thyroid hormones are believed to influence fibre proportions with the effect of GH more prominent in situations of relative thyroid hormone insufficiency and vice versa. It is possible that thyroid hormones were responsible for a decrease in type I fibres, particularly in female GH-deficient patients, and that six months therapy with r-hGH failed to reverse these changes due to an inherent hierarchial system of control dominated by thyroid hormones. Any attempt to clarify this phenomenon would involve, for example, treatment of long-term hypophysectomised rats with either r-hGH or thyroxine. It is interesting to note that testosterone treatment increases the proportion of type II fibres in rat and human studies Since all male patients from whom muscle tissue was processed, bar one, were on testosterone replacement, any effect of this hormone should have been observed in the gender analysis. However, no such change was observed in male subjects during the clinical trial and therefore testosterone did not have an obvious influence on muscle fibre proportions in this study. An additional consideration relates to a possible influence of r-hGH therapy on muscle pathways of aerobic and anaerobic metabolism. Future investigations should incorporate histochemical reactions for metabolic enzymes (eg. succinate dehydrogenase, nicotinamide adenine dinucleotide) to clarify this issue.

205 Study N GHD-Om P-6m GH-6/12m Con

Whitehead et al 1989 quad.fern. 5 44.7 48.7

Whitehead et al 1989 quad.fem. 5 52.3 52.2

Cuneo et al 1992 vas.lat. 21 47.1 42.3 35.6 45.0

Doerga et al 1994 vas.lat. 13 22.8 46.3

Hasan,W. (all patients) 10 42.9 42.8 40.7

Hasan,W. (male patients) 7 40.8 41.3 40.5

Hasan,W. (female patients) 3 50.6 44.7 41.2 Table ED.l; Summary table for investigations into the percentage of type I fibres in GH-deficient subjects. N=number of GH-deficient subjects in study, GHD-Om=patients before r-hGH treatment, P-6m=patients after six months placebo, GH-6/12m=patients after six or twelve months r-hGH therapy, Con=control subjects. Apart from N values, all others refer to the percentage of type I fibres.

206 ED.2. Muscle fibre size An influence of r-hGH on muscle fibre size has been noted in the present study. A number of other investigations have described similar effects. Hence, following r-hGH therapy, an increase in diameter of both type I and II fibres has been noted in the extensor digitorum longus and soleus muscles of normal rats In mutant dwarf mice, muscle fibres generally display smaller fibres than in control mice Though Rutherford et al reported a decrease in quadricep cross-sectional area, following discontinuation of r-hGH therapy for a year in GH-deficient adults, they did not observe any statistical changes in muscle fibre area over this period probably due to small numbers (number of patients biopsied was five). The same group did report a decrease, in quadriceps cross-sectional area and fibre area, in a patient with multiple hormone deficiencies, whilst a subject with isolated GHD did not display such changes, following discontinuation of r-hGH therapy Though there were no statistical changes in fibre area observed in the aforementioned studies, four out of five patients showed fibre atrophy in the range 10-40% (85.6% of baseline values, p=0.066) Treatment with r-hGH of GH-deficient adults, however, failed to increase fibre area in tissue taken from vastus lateralis and quadriceps femoris muscles The present study has demonstrated a decrease in fibre area in the placebo arm of the investigation. Hence, the trend observed by Rutherford et al ^9.483,484 decrease after twelve months discontinuation of r-hGH therapy) has been confirmed with a similar reduction of fibre area in the present investigation (10-12% decrease after six months of placebo in GH-deficient adults). In this context, the length of GHD at the start of the present study was one to two years for eight of the patients, whilst the other two patients had been GH-deficient for three and five years. The reduction in fibre size may therefore be understood as a consequence of the prolongation of GHD during the placebo stage of the trial. Confirmation of such an effect requires a larger patient population with a range of periods of GHD. Since fibre area was observed to increase, as compared to the placebo group, following r-hGH therapy, the present investigation is the first to demonstrate an increase in fibre area, for adult GH-deficient patients on r-hGH therapy. Cuneo et al explained their failure to demonstrate changes in fibre area in r-hGH treated GH-deficient subjects, as a consequence of lack of statistical power, given a 9% coefficient of variation for mean

207 fibre areas for biopsies taken at two separate occasions. However, the lack of change in the latter study can also be explained by the relatively short period of therapy with r-hGH (only six months), or due to the small numbers in the investigation (only seven patients treated with r-hGH were biopsied). Cuneo et al were able to show, however, a 5% increase in quadriceps femoris cross-sectional area, following six months r-hGH therapy. Other studies have also documented reductions in thigh muscle cross-sectional area in GH-deficient patients, and increases following r-hGH therapy 23,189,190. however, this methodology does not distinguish between increases in muscle bulk and that due to connective tissue growth. Further, in some cases results have not been corrected for changes due to bone growth, water accumulation, and fat redistribution. Other observations in the present study relate to gender and age differences. It is well established that the diameters of both type I and II fibres are larger in men than in women for the same muscle Also, type II fibres in the adult male are usually larger than type I fibres; in women, on the other hand, type I fibres are slightly larger Hence, in the course of the present investigation, the observation of a greater fibre size in males replicated the known situation. However, this study did not reproduce the subtler gender differences in fibre size, peculiar to each fibre type, perhaps because of the small number of patients in each group. The muscle weakness associated with ageing is known to reflect muscle fibre atrophy, which occurs concurrently with bone loss and muscle wasting Hence, the demonstration of smaller fibres in older subjects in this study was a confirmation of these age-changes. Of interest, however, was the lack of influence of r-hGH therapy on this parameter, given the widespread interest in GH as an ergogenic aid. The inadequacy of GH in preventing age-related changes in muscle has, however, been noted in a number of studies in ageing subjects 216,489,538 Type I fibre hyperplasia was also indicated by the data of Cuneo et al (29.5% increase in fibre number following therapy, p>0.05). Hence, increases in fibre size, following r-hGH therapy, may be accompanied by hyperplasia, particularly of type I fibres. The latter proposal is supported by the demonstration of atrophy and hypoplasia in soleus muscles from dwarf mice by the increased size and DNA:protein ratio in r-hGH-treated rats by the stimulation of satellite cells by IGF-I in regenerating rat

208 muscle and by the promotion of muscle protein anabolism by local infusion of GH or IGF-I in normal and postoperative subjects If limited hyperplasia was occurring, then variability in size of muscle fibres would be expected. It was a common observation in the present investigation, though not quantified, that small diameter fibres, particularly type I fibres, expressed a greater level of IGF-I mRNA (see plate 7iii). This would tend to support the suggestion that r-hGH therapy in these patients was accompanied by an increase in fibre number.

ED.3. Insulin-like growth factor I mRNA expression in muscle ED.3.1. In situ hybridisation A recent study on adult human vastus lateralis muscle has reported that IGF-I immunoreactivity was increased, following exercise, in muscle interstitial cells, satellite cells, vascular wall cells, and some diffuse staining of the cytoplasm was occasionally observed for muscle fibres Taaffe et al carried out a study on the effects of exercise, and r-hGH therapy during exercise, in elderly men. They observed, with quantitative PCR methodology, no change in vastus lateralis muscle mRNA for IGF-I, IGF-I receptor, GH receptor, IGF-II, or IGF-II receptor after exercise, or following r- hGH administration during resistance exercise training. Though the lack of effect in the Taaffe et al investigation may be ascribed to a reduced responsiveness in ageing the role of the GH/IGF-I axis in adult human muscle remains an open question. The function of IGF-I in animal muscle has been extensively examined. Aside from its role in development, the importance of IGF-I in models of nerve/muscle regeneration have been explored. Hence, immunoreactivity for IGF-I has been demonstrated in regenerating rat and mice skeletal muscle The latter studies were able to show IGF-I immunoreactivity in satellite cells, intramuscular nerves, blood vessels, myoblasts and in immature muscle cells. Additionally, Jennische and Hansson observed that immunoreactivity in myoblasts and myotubes was strictly cytoplasmic and associated with polyribosomes. They suggested that local synthesis of the peptide was occurring in the myogenic cells. A similar conclusion was propounded by Caroni and Schneider who demonstrated IGF-I mRNA upregulation in paralyzed and denervated rat skeletal muscle. The production of IGF-I message in the latter study was initially by muscle fibres, suggesting that the subsequent interstitial cell proliferation was induced

209 by the fibre-derived IGF-I. Caroni and Schneider also suggested that the IGF-I gene in muscle may be coded for jointly with other genes, including those for myogenin and acetylcholine receptors. In addition to an influence on interstitial cell proliferation, IGF- I produced by the muscle fibres is believed to have an autocrine function (on myogenic differentiation), and a paracrine influence on intramuscular nerves, other muscle fibres, fibroblasts and on glia 494,496,503,504,540

The evidence from human and animal studies indicate that the level of IGF-I protein expression by the muscle fibres themselves, in the normal non-pathologic state, may be limited. Hence, the presence of an appreciable IGF-I ISH signal in all groups, in the present investigation, may reflect a paracrine role for fibre-derived IGF-I in humans. In this context, a vastus lateralis muscle biopsy from an aged female subject revealed a level of ISH signal comparable to that seen with the trial patients (results not shown). Han et al observed a differential distribution pattern of IGFs message and protein in a wide range of tissues from human foetuses. They ascribed these differences to the accumulation of peptide in specific cells which contained IGF receptors and the ability to internalise the protein In the present study, the prominent expression of IGF-I message in the cytoplasm of muscle fibres may not represent the site of action of the protein. A more detailed evaluation of IGF-I action in skeletal muscle is found in the general discussion (F.2.2). A consistent trend was observed for IGF-I ISH signal to be greater in the r-hGH treated subjects as compared to the baseline and placebo groups. Additionally, twelve months r-hGH treated patients exhibited greater levels of ISH staining than baseline subjects. Hence, a modest increase in IGF-I mRNA expression may indicate a local function for this growth factor. A GH-dependency for muscle-derived IGF-I is thus indicated with a possible function in relation to muscle fibre growth. The lack of statistical evidence for these apparent changes probably reflects the small patient numbers involved. The possibility that circulating levels of IGF-I may have limited the influence of autocrine/paracrine -derived IGF-I cannot be dismissed. Further discussion on this issue will be continued later (see ED. 3.3.)

210 ED.3.2. Reverse transcriptase polymerase chain reaction Though extremely limited, the RT-PCR data suggest a confirmation of the ISH results. The general trend of ISH signal, for each of the three patients, was reproduced by the products obtained with RT-PCR Of particular consequence was the lack of RT-PCR products in the placebo samples, as deterioration of IGF-I signal in the baseline groups as a result of prolonged storage of samples, may have accounted for the absence of reaction product (the presence of a band in the six month r-hGH treated muscle signifies presence of IGF-I signal at the same approximate time point as in the placebo sample). The presence, in one subject (patient B), of both IGF-IEa and IGF-IEc transcripts following r-hGH therapy, is indicative of the possible GH-dependency for both E-peptide isoforms. The Ec-peptide is believed to be a local growth regulator influenced by mechanical rather than hormonal stimuli Hence, an increased expression following r-hGH therapy may indicate a novel control over this transcript. However, these results are preliminary and the main conclusion to be derived is the broad compatability of results obtained with the ISH and RT-PCR methodologies.

ED.3.3. Function of muscle insulin-like growth factor I It has been demonstrated in rats that transcription of IGF-I in skeletal muscle, following r-hGH treatment, precedes elevation of circulating IGF-I levels Since GH is known to regulate IGF-I at the transcription level a more detailed understanding of IGF-I mRNA production is required before the role of GH in this process can be elucidated. Several studies have demonstrated that, even in the absence of GH (following hypophysectomy), resistance training in rats results in an increased production of muscle IGF-I mRNA Also, the suppression of GH release in rats (with a- GHRH antibodies) causes a much greater reduction of IGF-I in the liver than in the muscle In addition, following exercise, GH suppressed rats display a significant increase in muscle exon 1 mRNA and no changes in liver IGF-I mRNA The conclusion reached by Zanconato et al ^ is that GH is a less important regulator of IGF-I production in muscle compared with liver. However, Isgaard et al were able to demonstrate a decrease in skeletal IGF-I mRNA levels following hypophysectomy and a return to normal levels after r-hGH therapy. Similarly, Turner et al showed an

211 increased expression of IGF-I mRNA in both the liver and muscle of rats implanted with GH secreting cells. The evidence above would tend to suggest that IGF-I levels in the liver would be more sensitive to GHD than those in muscle. Following r-hGH therapy, however, increased levels of IGF-I would be expected in both tissues. Since Taaffe et al were unable to show either of these changes in ageing, a state of relative GH insufficiency, they suggested that one confounding factor may be the time at which the biopsy was taken, as the message half-lives for IGF-I and its receptors are unknown. Studies in rats indicate a rapid increase in IGF-I mRNA within an hour of r-hGH injection, and a maximal response at 6-12h Interestingly, Laursen et al were able to obtain similar short-term metabolic effects (including increases in serum IGF-I) by either continuous infusion, or h interval injections, for 36h. It is not possible, without further investigation, to exclude the influence of sampling time point on the IGF-I mRNA data from the present study. A complementary approach to quantify IGF-I protein expression was not possible due to lack of sufficient time. Changes in IGF-I message expression with r-hGH therapy, in the present study, were not as dramatic as the elevations in serum IGF-I levels (see BR.2.). Interestingly, whereas twelve months of r-hGH therapy had no further increase, compared to six months of r-hGH treatment on serum IGF-I, IGF-I mRNA in muscle was observed to be higher at twelve compared to six months of r-hGH treatment. The differential regulation of liver- and muscle-derived IGF-I may account for these findings. In addition, improvements in muscle parameters, following r-hGH therapy in GH-deficient adults, may occur more gradually than can be appreciated from a general observation on serum IGF-I levels. In support of the latter proposal, one of the reasons why some studies ^*2,486 failed to note increases in muscle parameters, following r-hGH therapy in GH-deficient adults, may be because duration of treatment has been restricted to six months. The chief asset of the ISH approach, in the present study, was in clarifying the responses of different fibre types for IGF-I mRNA. There was no clear cut distinction between fibre types, as far as changes in IGF-I message with r-hGH treatment was concerned. However, a distinct elevation of IGF-I mRNA in type I fibres, compared to type II fibres, was observed at all trial stages. This novel observation suggests the

212 possibility of a GH-independent role for IGF-I. This may involve paracrine actions on muscle innervation and interstitial cells, as well as influencing specific metabolic characteristics of the type I fibres. It is recognised that modulation by local factors such as IGF-I may increase biochemical variability between fibres within a motor unit An aspect of the latter action may be the modulation by IGF-I of muscle specific regulatory genes such as myogenin The finding in the present study of greater levels of IGF-I mRNA in male muscle fibres (particularly in type II fibres), as compared with female samples, suggests a molecular basis for the gender differences observed in muscle fibre size. Interestingly, Czerwinski et al have demonstrated a role for IGF-I mRNA in rat muscle hypertrophy but not in muscle atrophy. An inference could also be attempted for the known greater strength in men, however, the observation by Rutherford et al who noted increased strength in GH-deficient men following r-hGH therapy, would imply a GH-dependence of this process, which is not suggested by the data from the present study.

ED.4. Conclusion The results fi*om the ISH experiments suggest a small GH-dependent increase in IGF-I message, and the likelihood of a non-GH dependent role for IGF-I, in type I fibres, in particular. The RT-PCR data has shown a possible GH-dependent increase for both IGF-IE transcripts on r-hGH therapy. Hence, a natural extension of the present study would be to focus on differential regulation of GH-dependent and independent IGF-I transcripts with ISH methodology. An interesting outcome may be the demonstration of fibre type-specific alterations in, for example, exon 1- and 2-derived IGF-I transcripts. The physiological consequence of alterations in fibre proportions, size and IGF-I levels, may relate to the reversal of fatigue and increased strength observed in patients treated with r-hGH. Since the isometric specific force of type II fibres is greater than that of type I fibres it is proposed that increasing fibre size, together with a preferential increase in type II fibre number, contributes to the overall greater strength seen in hypopituitary adults treated with r-hGH. A contribution to the latter effect may be the elevated IGF-I message in type II fibres in males, as this would complement the changes due to r-hGH therapy. The increase observed in fibre size was enough to reverse the

213 atrophy in the placebo period, however, fibre size did not rise above the baseline levels. Similarly, increases in IGF-I message in the muscle on r-hGH therapy were not dramatic. Hence, a maintenance role of GH on muscle parameters can be implied from this study.

214 CHAPTER 6 GENERAL DISCUSSION

215 F.l. Introduction The initial aim of this discussion will be to attempt to define the mechanism of GH action in various target tissues. The dual effector theory proposes that direct actions of GH, through GH receptors, result in local production of IGF-I, which in turn augments the functions of GH. Instead of being the sole mediator, IGF-I is thus regarded as part of a cascade initiated by GH. The theory suggests that targets may also be primed for the synthesis of other factors by direct actions of GH. For example, GH may be an initiator of target-derived neurotrophic activity, as suggested by the data from the present study (see chapter 3). It is hypothesised that GH 'priming' of target cells governs the neurotrophic support provided to innervating neurones. Further, it is proposed that the deficiency of GH in various situations (ie GHD, ageing etc) is responsible for fimctional alterations of target tissues and hence, their innervation. The defective thermoregulation characteristic of GHD and ageing can be regarded as a consequence of failure of GH to provide trophic support to the ESGs, resulting in a reduction in neurotrophic support for the innervating nerve fibres with consequent impairments in neuronal activation of sweating. The interactions of thyroid and sex hormones with GH allow for a further broadening of the suggested hypothesis. Evidence is provided (see chapters 3 & 5) that gender differences influence the responsiveness of target organs to GH. Focussing on the sudomotor apparatus, it is suspected that greater changes in the sweating profile following r-hGH therapy in females may be linked to similar changes in the functional innervation. Further, thyroid hormone interactions with GH may also be influenced by gender and direct influences of thyroid hormones on the sudomotor apparatus cannot be ruled out. The above suggestions will be expanded in the following sections:

F.2. Direct or indirect? Growth hormone and insulin-like growth factor I There is increasing evidence to suggest that though many functions of GH and IGF-I are common, there remain activities which cannot be duplicated by each other, or require the additive effect of both factors 37,108,197,199 basis for these varied responses may be tissue specific. In order to illustrate the complexity of interactions which exist in the GH/IGF-I axis, the mechanism for GH/IGF-I action in one well

216 characterised target will be examined so that avenues for understanding other systems may be unveiled.

F.2.1. Direct or indirect? Growth hormone, insulin-like growth factor I and bone Bone has long been recognised as a target of the GH/IGF-I axis, particularly as GH is the only recognised hormone to stimulate longitudinal bone growth in a dose-dependent manner (for reviews see Isaksson et al 209,542,543^ Initial observations in this field demonstrated decreases of epiphyseal cartilage width in hypophysectomised animals whilst GH-treatment resulted in increased thickness of the growth plate through proliferation of chondrocytes. The subsequent somatomedin hypothesis (seeAI.6.) was based on the lack of GH effect on cartilage metabolism in vitro whilst GH- dependent plasma peptides, in similar explant cartilage cultures, stimulated a number of cellular functions associated with cell multiplication and growth. Evidence from in vivo experiments also suggested the duplication of GH effects by infusion of IGF-I. However, the doses of IGF-I used were in some instances 50-fold higher than the dose of GH required to stimulate longitudinal bone growth. Further, no dose-response relationship was seen in IGF-I infused animals, nor was there a correlation between serum IGF-I and the magnitude of bone growth in animals treated with IGF-I or GH In order to reconcile these differences between in vivo and in vitro studies, attention focussed on the characteristics of the end organ response. Local administration of GH or IGF-I to hypophysectomised rats was shown to stimulate epiphyseal chondrocytes; these effects on bone growth were not augmented by simultaneous administration of GH and IGF-I. The site of action was further elucidated by the finding that GH was primarily acting on progenitor cells (prechondrocytes or young differentiating chondrocytes) whilst IGF-I was stimulating proliferating chondrocytes. A crucial role for locally released IGF-I was suggested by the demonstration of IGF-I mRNA production, independently of serum IGF-I levels, in growth plates of GH-treated hypophysectomised rats. This local IGF-I production, and the stimulatory effect of GH on the growth plate, could be abolished by infusion of antibodies to IGF-I with concurrent local GH administration. These and other observations led to the proposal of a dual effector theory of GH action. This hypothesis suggests a direct stimulation by GH of the differentiation of mesenchymal precursor cells (including prechondrocytes

217 and satellite cells in muscle), and the subsequent promotion by IGF-I of the clonal expansion of these cells, previously primed by the direct action of GH. Isaksson et al propose a further level of synergy where GH-stimulated cells become more responsive to IGF-I. In addition, GH may stimulate or enhance local production oflGF- I, whilst IGF-I may downregulate GHRs thus strengthening the synergistic interactions of these growth factors. Though a paracrine role of IGF-I is appealing, human studies in particular (including the present investigation) have argued for the importance of liver-produced IGF-I in the maintenance of bone mass Hence, the relative contributions of systemic and locally produced IGF-I remain a matter of dispute (see 199,516,544-546^ Care should also be exerted in extrapolating from observations in bone to other tissues since, for example, the kidney, spleen and thymus of hypophysectomised rats are more sensitive to treatment with IGF-I, as compared to GH These cautionary aspects should not detract from the main conclusion; namely, for bone and perhaps other tissues, GH and IGF-I cooperate for a maximal growth response, this synergy involving local generation of IGF-I and direct actions of GH on target cells.

F2.2. Direct or indirect? Growth hormone, insulin-like growth factor I and muscle The liver is the main organ contributing to the circulating pool of IGF-I; it has limited numbers of type I IGF-I receptors and has traditionally been regarded as the principal initiator of IGF-I effects in the body An important implication of the dual effector theory, discussed above, is that systemic IGF-I may have specific functions in particular tissues. However, liver-derived IGF-I can no longer be considered the main source of IGF-I action in the body The shift of emphasis towards autocrine and/or paracrine actions of IGF-I brings with it some controversy as to the role of GH in these actions of IGF-I. It is evident that particular tissues will have to be examined for evidence indicating the influence of either factor, and the route of action. Skeletal muscle has been extensively studied with regard to regenerative and growth responses following stimuli such as denervation, exercise, stretch, electrical stimulation and GH or IGF-I treatment (see EL). There is substantial evidence from animal studies which suggests that IGF-I mRNA and/or protein is increased in skeletal muscle following these stimuli 99,499-501,504-508,517,539,547,548 process of muscle regeneration and

218 growth involves the proliferation and fixsion of satellite cells with adjacent fibres, resulting in an increase in myonuclei numbers for fibre growth and repair The source of muscle IGF-I in rats, in response to most stimuli, appears to involve satellite cells,

fibroblasts and immature and mature muscle cells 500,506,507,517,539,547 gi^^gig^al muscle fibroblasts and myofibre cells in culture also secrete IGFBPs in response to stretch In situations of muscle denervation, aside from interstitial cell proliferation, sprouting of intramuscular nerves also occurs The signal for these changes has been identified as a rapid but transient elevation in muscle fibre IGF-I mRNA production Evidence fi-om studies on motoneuron regeneration suggest that IGF-I and its binding proteins are additionally produced by Schwann cells IGF-I derived from these sources is transported in nerve by retrograde transport presumably to exert a neurotrophic function The differential distribution of IGFBPs in muscle has been demonstrated by Ma et al in the adult mouse Hence, IGFBP-2 was immunohistochemically localised to mouse skeletal muscle fibre basement membranes, whilst IGFBP-1 was present at the neuromuscular junction, as well as within intramuscular nerves suggesting its axonal transport It is clear from the above evidence that locally-derived IGF-I has a dual role in skeletal muscle, acting in an autocrine manner on muscle fibre growth as well as fulfilling a paracrine function on motoneurones. The presence of differing IGFBPs in the local environment is indicative of tight regulation of IGF-I in this tissue system. However, a further level of control for IGF-I may be by circulating GH Skeletal muscle IGF-I mRNA expression is decreased in hypophysectomised rats and replacement with r-hGH normalises levels of IGF-I expression The possible direct nature of GH action is reflected by an early increase in GHR/BP mRNA in regenerating normal rat myoblasts and myotubes The localisation of the GHR/BP mRNA to muscle fibres has also been shown in adult rat gastrocnemius muscle; the ISH signal was found to be stronger around the nuclei of skeletal muscle cells In addition, the tyrosine kinase signal transduction pathway has been shown to be activated by GH through its receptor in cultured skeletal muscle cells from normal rats Though studies in humans are limited, there is some evidence indicating that similar mechanisms to the above exist in humans. GHR mRNA is present in the cytoplasm of adult human skeletal muscle fibres with ISH signal greatest close to muscle cell nuclei

219 After various regimes of physical exercise, circulating GH levels increase in both trained and untrained individuals Following seven days of strenuous exercise, Hellsten et al observed IGF-I immunoreactive cells in skeletal muscle biopsy samples. These cells included endothelial and smooth muscle cells of blood vessels, as well as interstitial cells tentatively identified as satellite cells; occasional muscle cells with diffuse staining of the cytoplasm were also observed displaying IGF-I immunoreactivity Due to elevation in plasma of biochemical markers for muscle injury (creatine kinase and hydroxyproline), Hellsten et al suggest that the increased IGF-I immunoreactivity was probably due to muscle regenerative responses. Hence, GH may either directly influence muscle growth and regenerative responses, or, indirectly stimulate muscle IGF-I. That local rather than systemic sources of IGF-I are important in muscle plasticity is indicated by improvements in muscle growth, following r-hGH therapy in hypophysectomised rats or dwarf mice, without appreciable changes in serum IGF-I levels i9^’552,553 indeed. Pell and Bates observed greater increases in muscle IGF-I immunostaining following GH treatment than after IGF-I administration in dwarf mice The evidence thus far presented has shown the importance, in both bone and muscle, for a thorough evaluation of the contributions of systemic and autocrine and/or paracrine modes of IGF-I action. Further, the direct role of GH has been examined, particularly in regard to its control over local tissue IGF-I production. The finding from the present study, of a lack of correlation between serum IGF-I levels and the muscle IGF-I mRNA signal also suggests a direct influence of GH on the production of IGF-I by the muscle fibres. Further, the positive correlation of muscle fibre size with both serum and muscle IGF-I indicate that both sources of IGF-I are likely to have a function in muscle growth. An important finding from this investigation is the demonstration of roughly comparable myofibre IGF-I mRNA signal in GHD and also following r-hGH treatment. If muscle IGF-I production is a normal feature of adult muscle then this suggests an important role of IGF-I in normal muscle maintenance and growth responses. The observed influence of gender and fibre type on the level of IGF-I ISH signal also suggests an involvement of muscle IGF-I in the normal maintenance of muscle fibres. The demonstration of IGF-I mRNA in interstitial cells, blood vessels, and intramuscular

220 nerves suggests a role for locally produced IGF-I in the maintenance and plasticity of a range of peripheral tissues. In summary, observations from this investigation, and from the literature, suggest that skeletal muscle is a target of GH acting directly, and indirectly via both hepatic and muscle generation of IGF-I. A logical inference from these findings is that GH is involved in both muscle growth in response to physiological stimuli, and via its generation of local IGF-I, in the normal plasticity of the neuromuscular junction and motor innervation. Other targets of the GH/IGF-I axis may include the sudomotor apparatus which has been examined in this study in both humans and rats.

F.3. Growth hormone and the nervous system; a re-examination There is increasing evidence to suggest that the endocrine and nervous systems are much more closely integrated than was previously thought. For example, classical neurotrophins not only influence the neuroendocrine axis, but can also be regarded as having endocrine fiinctions Conversely, many hormones have effects which may be autocrine and/or paracrine and are thus not typically endocrine. Further, the interactions of GH, IGF-I, and thyroid hormones (amongst others) with the nervous system suggests that they may fulfill a 'neurotrophic' function Much of the evidence in support of these proposals has been outlined in Chapters 1, 3 and 4, however, some aspects related to GH in particular will be explored below.

F.3.1. Neurotrophins and growth hormone Aside from NGF expression in endocrine organs (eg. thyroid, parathyroid, adrenal, pineal and reproductive glands) the role of this neurotrophin in the hypothalamus- pituitary axis is of considerable interest. The presence of NGF and its receptors in the hypothalamus (pTS^*^^) and pituitary {trk A) has been described in rats Patterson and Childs have observed NGF immunoreactivity in 51% of cells containing GH, whilst trk A immunoreactivity was present in 23% of GH-secreting cells. Borson et.at., were able to demonstrate, in the primate pituitary, immunohistochemical localisation of and neurotrophins (with a neurotrophin antibody binding to NGF, brain- derived neurotropic factor and neurotrophin-3). The novel finding, in the latter study.

221 was the lack of co-localisation of neurotrophin-containing cells with hormone-secreting cells, suggesting the presence of a distinct population of neurotrophin cells in the anterior pituitary. An endocrine role for neurotrophins was postulated but remains debatable until the presence of neurotrophins in secretory granules has been shown Subsequent studies suggest a paracrine action of pituitary-derived NGF on the proliferation and differentiation of hormonal cells, particularly mammotrophs Treatment of pancreatic cell lines with r-hGH results in an increase of both p75^°^ and trk A receptor binding sites and of p75^^^ mRNA levels These actions are believed to be direct, since GHR sites exist on the p-cells and there appears to be a putative GH- responsive element in the p75^°^ gene; this GH-sensitive motif is also present in the genes for IGF-I, SS etc NGF in the anterior pituitary is therefore possibly under control from GH, acting in a paracrine manner. Conversely, NGF or other neurotrophins may influence GH production from somatotropic cells. Further, the presence of GHR sites on tissues which produce or are responsive to neurotrophins, could be indicative of a similar control of neurotrophin levels by GH. The latter proposal has further implications which will be discussed in relation to the sudomotor apparatus.

F.3.2. The influence of growth hormone on its regulatory neurons The positive effect of GH administration on peripheral nerve regeneration in normal and adult hypophysectomised rats indicates a role for this hormone in nerve-target interations in maturity. GH actions in the nervous system may occur through classical endocrine actions where circulating IGF-I acts as an intermediary. GH may also have a direct action on the target itself, with a subsequent indirect effect on nerves either through IGF-I, or through some other neurotrophin. There exists the intriguing third possibility that GH may also influence nerves directly. The feedback of GH on its hypophysiotropic neurons (see AJ. 7.1.) calls for debate on the precise route of GH action on these neurons. Aside from feedback control in maturity, GH affects the survival and maturation of the GHRH and SS neurons in the developing hypothalamus Development and phenotypic differentiation of GHRH and SS neurons are normal in dwarf mice, however, lack of feedback signals from GH result in regression of SS neurons and increased numbers of GHRH neurons The

222 mechanism for GH feedback control appears to involve activation of the c-fos gene, in SS neurons of the periventricular nucleus, and also in NPY neurons in the arcuate nucleus NPY-containing axons have synaptic connections with SS neurons in the periventricular nucleus, however, the location of their cell bodies is unknown; if these cell bodies were located in the arcuate then this would indicate that NPY neurons can indirectly influence GH secretion GHR mRNA has been colocalised with both SS mRNA in neurons of the periventricular nucleus and with NPY mRNA in neurons of the arcuate nucleus The levels of NPY in the neurons of the arcuate, but not in other regions of the hypothalamus, are dependent on the GH status of the animal suggesting a direct action of GH on these neurons Chan et al suggest that NPY neurons in the arcuate probably have connections with GHRH neurons in this region given the close proximity of these neurons. In addition, an inhibition of GH secretion following intracerebroventricular administration of NPY has been observed What is apparent from these studies is the direct nature of GH action, through its receptors in the hypothalamus, on its regulatory (and other) neurons.

F.4. Growth hormone, insulin-like growth factor I and the sudomotor apparatus The above discussion has provided evidence for an action of GH on its hypophysiotropic neurons which is quite divorced from classical understandings of GH feedback inhibition on its secretion. Evidence has also been provided (see F.3.) to indicate a direct interaction of GH with neurotrophins. Aside from clarifying the various modes of GH action, the sections on GH and bone/muscle interactions (see F.2 .) have also shown the importance of receptors and binding proteins in modulating growth factor action. The above inferences will be discussed in relation to the actions of GH in the skin. The present study has shown that GH exerts a trophic effect on ESG secretory coils. Further, sudomotor nerve growth has been shown to be a feature of acromegaly, and r-hGH therapy to GH-dehcient patients results in alterations in the neuropeptide content of ESG innervation. ISH on skin samples from GH-treated trial subjects, indicated the presence of IGF-I mRNA signal on ESG secretory coils (see plate 8). There was no specific staining in the periacinar connective tissue, suggesting that sudomotor nerve fibres situated in the connective tissue, did not express IGF-I signal.

223 These preliminary observations demonstrate that human ESGs can express IGF-I with the potential for consequent trophic actions on periacinar nerves. In the absence of data from the baseline patients it is not possible to conclude whether r-hGH treatment induced IGF-I expression. Some evidence is presented below, particularly in regard to the distribution of GH and IGF-I in the skin, in support of this suggestion. The presence of GHR/BP (protein and message) in human adult and foetal skin has been demonstrated 398-400,S6i q h r /b p protein expression is seen in the epidermis, dermal fibroblasts, sebaceous glands, ESGs, and dermal blood vessels. However, GHR mRNA has only been localised to the epidermis, dermal fibroblasts and sebaceous glands. In rat skin, GHR/BP protein and message are found in similar locations to that in humans with the additional presence of GHR/BP immunoreactivity in peripheral nerves and skeletal muscle fibres 397,550,552,563 ^GF-I mRNA in human and rat skin is found primarily in the epidermis and in dermal fibroblasts, whilst immunohistochemical staining demonstrates the protein in hair follicles and sebaceous glands ^^ 2,400,541,553 Cultured human kératinocytes, but not fibroblasts, express mRNA for the IGF-I receptor Of particular import is the differential distribution of IGFBPs in adult human skin Batch et al localised IGFBP-3 (the major circulating IGFBP) mRNA to the epidermis whilst IGFBP-2, IGFBP-4, and IGFBP-5 mRNA was present in ESGs. Each IGFBP can both inhibit or potentiate IGF-I action However, the half-life of the IGF-I-IGFBP-3 complex is much greater than that of IGF-I associated with the less abundant and smaller-sized 'minor' IGFBPs (-1,-2,-4,-5 and -6) Also, the minor IGFBPs are not restricted to the circulation, hence, their involvement in local modulation can be inferred IGFBPs derived locally may prevent the diffusion of IGF-I and hence restrict its actions to the site of production This may occur by the association of IGFBPs with proteins in the cell surface, or in the extracellular matrix, resulting in an increased IGF-I avilability to its receptor Taken together with the preliminary demonstration of IGF-I mRNA in ESGs in the present study, this result indicates that IGF-I produced by ESGs may have an autocrine/paracrine function. Hence, GH-induced IGF-I production would be accompanied by the production of IGFBPs, also from the ESGs. The IGFBPs, by restricting IGF-I action to the vicinity of the ESG, would serve a key function in control over IGF-I action. Given the well

224 established influence of IGF-I on peripheral nerves, an obvious target of IGF-I action would be the sudomotor fibres. The mechanism thus far presented is of an initial direct effect of GH, through its receptor on ESGs, inducing IGF-I production and glandular growth. The growth response may also be a consequence of IGF-I acting in an autocrine manner, or through circulating IGF-I. Until distribution of IGF-I receptors in skin has been evaluated, these possibilities cannot be dismissed. It is also proposed that the generation by ESGs of specific IGFBPs, whose role is linked to local control over IGF-I function, is indicative of a neurotrophic role of ESG-derived IGF-I on sudomotor nerves. Neurotrophic factors for ESG innervation in maturity have not been extensively investigated. The target-derived cholinergic differentiation factor (s) responsible for the transmitter switch in sudomotor neurons during development, remains to be identified (see CL4.) Though leukemia inhibitory factor and ciliary neurotrophic factor can induce cholinergic differentiation in culture, transgenic mice deficient in either factor have a normal neurotransmitter phenotype 289,443,565 ygiigygj to be permissive for developing sudomotor neurons, allowing the neurons to survive and grow but not influencing their phenotype The demonstration of p75^°^ staining on rat sudomotor nerves in this study is indicative of a role in adulthood for the neurotrophin family (i.e. NGF, brain-derived neurotrophic factor, neurotrophin-3 etc). This is also believed to be true in humans since trk A immunoreactivity has been observed on ESG secretory coil and duct cells These observations allow the proposal of an interaction between NGF and IGF-I as neurotrophic factors acting in synergy for the sudomotor neurons in maturity. The similar biochemical pathways and mechanisms of NGF and IGF-I action, the potentiation by IGF-I of NGF receptors, and the co-localisation of IGF-I and p75^*^^ in the same neuronal population has been previously commented on (see DL3.4.). A further aspect of cooperation will be commented on below. The NGF gene encodes for a protein (78 NGF) which is synthesised in a pre-pro form 78 NGF is a dimer made up of two a-, two P-, and two y-subunits, with the y- subunits responsible for the protease activity and also involved in intracellular processing of the pre-pro form of the NGF molecule^^^’^^^. The p-subunits constitute the biologically active, secreted form of the peptide 78 NGF can modulate IGF-I action by cleaving of IGFBP-3, -4, -5, and -6, hence suggesting an additional level of

225 interaction between the two growth factors Though the latter study employed a protease assay methodology, the inhibition, by 7S NGF of IGF-I binding to neurons in culture, also suggests the validity of this finding Is there an additional role for GH in this scenario? Without the influence of epigenetic factors in determining neuronal connectivity, the complexity of the nervous system would be considerably curtailed Thus, sympathetic neurons induced to innervate large targets exhibit greater cell body diameters and more complex dendritic arborisations than do neurons that innervate smaller targets The demonstration of hypertrophic glands in acromegaly, and a correspondingly increased innervation, is compatible with excess GH levels inducing nerve-target changes at the level of the ESGs. An alternative explanation would involve circulating GH affecting the thermoregulatory centre in the hypothalamus, resulting in an increased innervation to the ESGs, which would direct the subsequent elevation in the sweating profile. However, this interpretation suggests that the other changes which occur in the skin of acromegalic subjects (eg. increased epidermal thickness, hyperpigmentation, dermal connective tissue changes, increased sebum secretion etc; see CI.8.) may also be explained by alterations in central control of growth and metabolic processes. What has become apparent from the discussion on the mechanism of GH action in relation to bone, muscle and ESGs, is the importance of direct effects of GH on target organs. It is suggested that the GH/IGF-I axis serves as a bridge allowing central modulation of peripheral effectors. An additional role of the GH/IGF-I axis could be in its interaction with the spinal and/or axon reflex sweat responses (see CL 3., ^62-264^ The local nature of these alternate sweating pathways would allow for modulation by paracrine factors. For such a function to exist, however, a physiological significance for these reflex arcs needs to be proven. Localised hyperhidrosis associated with cutaneous diseases has already been commented on (see CL 3. & CL 8.) and may be compensatory if there are large areas of anhidrosis In addition, it has recently been shown, in normal subjects undergoing physical exercise in a temperate environment or under heat stress, that while an exercise-induced rise in core temperature is critical to acclimation (including cutaneous blood flow and sweating responses), heat stress induces a much more complete acclimation Regan et al propose that not only is the cutaneous input important

226 in acute regulation of core temperature, but it may be fundamental to the acclimation process through which the body adapts to restrict changes in core temperature. Does the latter investigation have consequences for a local modulation of sweating? Kruger et al suggests that ESGs may provide an example of a peripheral feedback loop whereby cutaneous thermoreceptors stimulated by an elevation in temperature might then activate ESGs in their immediate vicinity to dissipate heat and maintain body temperature. Such a mechanism would presumably also include vasodilator responses, hence, the presence of a factor with an influence on both sensory and sympathetic populations can be envisaged to play a crucial role in peripheral thermoregulatory responses. The presence of both NGF and IGF-I in the local environment, taken together with the increased secretion of GH in situations of elevated temperature, allow the possibility of modulation by these factors of the sweating apparatus. The role of GH in thermoregulation has been one focus in the present study. A link has been suggested between GH status, the sweating profile and changes in ESG morphology and innervation. Further, a direct role of GH on the effector organ has been established from a study in rats. A mechanism of direct actions of GH on the sudomotor apparatus has been proposed taking into consideration better understood pathways in bone and muscle. It has also been suggested that the skin serves not just as a passive conveyor or recipient of nerve impulses, but that it generates local thermoregulatory responses. It is hypothesised that GH, through interactions with IGF-I and NGF, may play a crucial role in modulating the morphology and plasticity of the sudomotor apparatus. In consequence, in situations of GHD, including ageing, alterations in thermoregulation may be a consequence of lack of trophic support to the sudomotor apparatus.

F.5. Gender, thyroid hormones and growth hormone interactions Gender differences were observed for various experimental markers. Some of these confirmed previous observations. Thus, males exhibited greater levels than females for serum IGF-I (28%), sweat rate (37.5%), and muscle fibre size (13.6%). Other findings were more novel, such as increased levels of VIP and PGP staining in sudomotor nerves (13%), ESG size (8%), and IGF-I mRNA in muscle (12%) in males compared to females . Further, T3 concentrations were found to vary according to the thyroxine

227 status of the patient. Hence, euthyroid subjects demonstrated greater levels of T3 in males (19%), and patients on thyroxine replacement exhibited increased T3 concentrations in females (13%) as compared to males. Patients on thyroxine replacement also showed greater levels of circulating thyroxine in females (22%) as compared to males. Two-way ANOVA analysis indicated that, apart from thyroid hormones, these results were independent of the GH status of the patients. Given the small study group, extrapolations of the novel data to the general population may not be appropriate. However, the results suggest that patients displayed gender- specific dififerences for key physiological parameters. As an aid to interpreting the value of experimental findings, the incorporation into clinical trials of parameters known to display a sex difference, may be used as a tool in indicating 'normality' of a heterogeneous study group. However, these results also illustrate the dangers of under estimating gender differences in clinical trial design, particularly with regard to placebo- controlled studies. The skewed distribution of females (7 in placebo group, 3 treated with r-hGH for 12 months) in the present investigation probably contributed to the lack of statistical power for certain analyses. This is exemplified by the greater responsiveness of females to r-hGH therapy for some parameters. Hence, increases in females were greater than those in males for sweat rate (males 31%, females 52.5%) and AChE staining around ESGs (males 19.5%, females 41%). A gender-specific response to r-hGH therapy has previously been reported by Johannsson et al with males responding to treatment better than females for serum IGF-I levels and total body weight. In addition, Johansson et al have observed differences between clinical trial groups at the baseline stage for certain parameters. It is apparent that human studies with small numbers of subjects need to be carefully evaluated for gender differences, and that designation of placebo administration should not be random in relation to gender. A further aspect to the role of gender in determining GH-responsivity relates to the mechanisms of GH action in the skin. The finding of greater sweating rates in male as compared to female trial subjects has already been elaborated upon (see CD. 3.). The increased responsiveness of females to r-hGH therapy may reflect the observation that peak GH levels in response to most pharmacological stimuli are greater in normal premenopausal females than males The link which has been shown in this study

228 between changes in the sweating profile and cholinergic innervation suggest that the effect of GH on the sudomotor apparatus may be influenced by interactions with sex hormones. The presence of both androgen and oestradiol receptor immunoreactivity (see CD. 3.) on ESGs is consistent with this scenario. What could the mechanism of interaction between GH and sex hormones involve? Sex hormones are known to modify the metabolic and growth-promoting effects of GH in peripheral tissues It has been suggested that androgens are involved in long-term remodelling of secretory coil structure and in the maintenance of sex differences a possible mechanism may be via androgen-dependent tissue kallikreins. The presence in human sweat of kallikrein and kininase suggests that a kallikrein-kinin system is present in ESGs Glandular kallikrein is may be involved in cleavage of bioactive precursor peptides of NGF and/or epidermal growth factor Interestingly, the glandular kallikrein, prostate specific antigen, can act as an IGFBP-3 protease releasing IGF-I suggesting an influence over GH action in the skin for these proteases and hence for androgens. The sex differences observed in this study for sweating, acinar size, and innervation may therefore be explained by the greater activity of kallikreins modulating NGF and IGF-I activity in males. The mechanism of oestradiol action is unclear, however, future studies on the interaction of sex hormones with the GH/IGF-I axis are warranted particularly with regard to possible influences on neurotrophic function. In relation to gender differences, a further contributory factor may be the influence of thyroid hormones. Thyroxine and triiodothyronine levels were reduced in the placebo group, relative to baseline values, in females only (11% and 24%, respectively). Thus, for those experimental parameters demonstrating a correlation with thyroid hormones, a gender-specific influence, particularly for females in the placebo group, may be postulated. However, the fluctuating relationship between thyroid hormones and GH status (see BD.3.) does not allow for clear inferences to emerge in relation to the gender data from the present study. Interactions were also observed between thyroxine and parameters such as urinary total pyridinoline (negative correlation), muscle fibre size (negative correlation) and neuropeptides in sudomotor nerves (negative correlation with AChE and VIP; positive correlation with PGP epidermal/dermal staining). A role of thyroid hormones has also been suggested for muscle fibre proportions (see ED. 1.). These results have consequences for the mechanism of GH action and it is also possible

229 that some beneficial effects of r-hGH therapy may be mediated through thyroid hormones. Hence, it has been shown that r-hGH treatment, without the concurrent administration of thyroid hormones, may not be as effective in reversing growth impairments in hypopituitary humans Further, IGF-I bioactivity is reduced in both hypothyroidism and hyperthyroidism (despite normalised IGF-I levels in hyperthyroidism) In relation to the nervous system, it has been demonstrated in culture that a synergy between thyroxine and IGF-I determines cortical neuron development and differentiation. Addition of T3 to hypothalamic cultures results in maturation of neurones coincident with increased IGF-I levels Thyroxine also increases neurite outgrowth of peripheral regenerating nerves An interaction between NGF and thyroid hormones exists with developmental expression of NGF possibly modulated by thyroxine The influence of thyroid hormones on the function of the GH/IGF-I axis is indicated from the above observations. It is apparent that a more thorough evaluation of the role of thyroid hormones, particularly on the neuronal aspects, is required.

F.6. Conclusion Adult growth hormone deficiency syndrome involves changes in the musculo-skeletal system as well as thermoregulatory disturbances. The present study has investigated these problems before and after growth hormone therapy. Prior to treatment, GH- deficient subjects display decreased sweating rates, alterations in sweat gland innervation, and reductions in muscle fibre size and IGF-I mRNA expression. Therapy with r-hGH reverses some of these deficits. The mechanism of GH interaction with these systems has been investigated and is believed to involve direct actions of GH on targets leading to local IGF-I generation. In addition, GH effects on sweat gland innervation suggest an indirect neurotrophic function for this hormone, possibly via generation of local, target-derived neurotrophins. This study has also shown an influence of sex and thyroid hormones on patient responses to r-hGH therapy. It is apparent that beneficial effects of r-hGH substitution involve interplay with several body systems, including muscle, skin appendages and the peripheral nervous system.

230 m

Plate 8 Photomicrograph of skin from a trial subject, following twelve months r-hGH therapy, demonstrating IGF-I message expression by in situ hybridisation.

A: ESGs demonstrating a-sense signal (xIO magnification). B: ESGs demonstrating sense signal (xlO magnification). C: ESGs demonstrating a-sense signal (x40 magnification).

Note: The ESG periluminal cell staining (arrowhead). Scale bar for A and B =1000pm. Scale bar for C =25pm.

231 Appendix G.I. Data presentation and statistical methods

The variability in response to r-hGH therapy between patients and the confounding influences of gender, age, and thyroid hormone status are aspects which have had to be carefiilly evaluated as part of this study. Due to these influences, the data generated were considerable and thus only results which indicated changes have been presented. Patients received either placebo or r-hGH for the first six months after which all subjects received r-hGH for the subsequent six months (see BM. L). Subgroups for data analysis could have included: la) baseline patients to receive r-hGH, six months r-hGH, twelve months r-hGH. lb) baseline patients to receive placebo, six months placebo, six months r-hGH. or, 2) all baseline patients, six months placebo, six months r-hGH, twelve months r-hGH.

The latter method of subdividing trial subjects was chosen in order to allow maximum numbers to be included in analysis. The major disadvantage of this approach is in the comparison of the six months placebo and twelve months r-hGH groups since they are composed of separate patient populations. A complementary approach, involving expressing each patient's six and twelve month data as a percentage of the baseline values, was also carried out. Though baseline data spread is lost, this method minimises differences due to inter-patient variability. Before statistical analysis, data were checked for normality of distribution. This consisted of a visual assessment of observed and expected normal frequency distributions (Unistat statistical package 3.0a). In addition, the Shapiro-Wilk goodness of fit test was carried out on the data. The latter test provides a test statistic (from 0 to 1) which is indicative of the normality of the data (see for experimental usage see

574)

The parametric test of choice was the classic experimental analysis of variance (ANOVA). However, student's t-test was also utilised where appropriate. The ANOVA method allowed the influence of more than one factor to be evaluated (two-way ANOVA, regression analysis). Post hoc comparisons of data groups were carried out

232 with the Tukey-highest standard deviation method If data were not normally distributed transformation of the data was carried out. Logarithmic and square root transformation are established methods of normalising data spread (see for experimental usage see Non-parametric methods included the Kruskal-Wallis ANOVA (K-W ANOVA) and the Mann-Whitney U test (M-W U test). A 95% confidence interval was used for all statistical analysis. For age and gender analysis, in order to increase numbers for analysis, patients at all trial stages were included in the statistical tests. However, individual trial stages were also examined for age and gender differences and where trial status influenced these analyses, then this has been stated. In addition, two-way ANOVA was carried out to identify GH influences in relation to gender and age.

233 Appendix G.2. Assays for serum IGF-I, free T3, free T4, and urinary total pyridinoline crosslinks

G.2.1.Serum IGF-I assay The details of the competitive radioimmunoassay protocol are as follows: A polyclonal rabbit anti-IGF-I antiserum and a radioactively labelled IGF-I tracer were added to a serum sample. The non-labelled IGF-I in serum competes with the labelled IGF-I tracer for the limited amount of binding sites in the antiserum (anti- IGF-I antibodies are added in a substoichiometric amount as compared to the total amount of labelled and non-labelled IGF-I). After an overnight incubation at room temperature, the immune complex was precipitated with a second antibody (polyclonal anti-rabbit IgG antiserum) in a polyethylene glycol environment. The radioactivity of the precipitated immune complex was then counted in a gamma counter. The amount of precipitated radioactivity in a sample is reciprocally proportional to the concentration of IGF-I. IGFBPs may interfere with the assay. Thus, IGFBPs were precipitated with acidified ethanol prior to the assay. To further reduce the influence of the IGFBP activity, a truncated IGF-I analogue was utilised as a tracer. In the analogue, des (1-3) rhIGF-I, three amino acids are missing from the amino terminal resulting in a lower affinity for IGFBPs as compared to native IGF-I. The assay has a cross-reactivity for IGF-II of about 1 %. For insulin, proinsulin and smaller synthesized IGF-I fragments the cross-reactivity is negligible. (Protocol information courtesy of Pharmacia & Upjohn Ltd, UK; document 94 96 681)

G.2.2. Serum free T3 and T4 assay The test principal is an ELISA method using streptavidin technology (Enzymun-test FT-3/FT-4). The plasma is treated with EDTA buffer and incubated with streptavidin incubation solution. An incubation solution, containing biotinylated anti-T3/T4 antibodies, is added allowing plate wall binding. A washing stage separates the unbound

234 constituents and a substrate-chromagen solution solution is added. Absorbance is plotted allowing an assessment of free T3/T4 concentration. (Protocol information courtesy of Boehringer Mannheim, UK; document 1092.309.1411730)

G.2.3. Urinary total pyridinoline crosslinks assay Extracted freeze-dried acid hydrolysates were re-suspended in 20mM pentafluoropropionic acid (PFPA). Separations were carried out using an Exsil 100 5- microns 0DS2 column (100mm x 2.1mm I.D.) eluted with lOmM PFPA in water at 0.15 ml/min and detected using a Jasco 821-FP detector (xenon lamp: excitation 290nm, emission 400nm). Fluorescent response was linear from 269 to 8620fmol for pyridinoline. The limits of detection were 28fmol.^^^

235 Appendix G.3. Gelatin coating of slides and acetylcholinesterase methodology

G.3.1. Gelatin-chrom alum coating of slides Slides are washed in 75% alcohol and air-dried. The coating solution consists of 1% gelatin and 0.3% chromic potassium sulphate; for 300ml add (i) 3g gelatin powder to 250ml H2O, warm to 50°C, (ii) add 0.3g chromic potassium sulphate to 60ml H 2O and warm to dissolve. Mix both solutions, filter, allow to cool to 37°C. Dip slides for Imin, drain of excess and dry at 37°C. Coated slides may be stored for several weeks.

G.3.2. Acetylcholinesterase methodology Incubation medium constituents; Phosphate buffer (pH 6.02); mix KH 2PO4 (O.IM, 13.61g/l) and Na 2HP04 (O.IM, 14.2g/l). Sodium citrate (O.IM, 2.94g/100ml), copper sulphate (anhydrous, 0.03M, 479mg/100ml), potassium ferricyanide (0.003 8 M, 126mg/100ml). Store stock solutions at4°C. For 100ml add 50mg acetylthiocholine iodide to 65ml phosphate buffer (pH 6.02). Then add sodium citrate (5ml), copper sulphate (10ml), H 2O (10ml), potassium ferricyanide ( 10ml), and 100 pi of inhibitor. Three inhibitors were utilised: Iso-OMPA (tetraisopropylpyrophosphoramine; Mwt =342.4, working solution =10'^ M); an irreversible inhibitor for abolishing non-specific staining. Eserine (Mwt= 649, working solution =0.42x10'^^M); a competitive inhibitor for abolishing both specific AChE and non-specific staining. B.W. 284C51 (Mwt =560, working solution =10'^M); a competitive inhibitor for abolishing specific AChE staining.

236 Appendix G.4.

I n s i t u hybridisation procedures

G.4.1. Silane coating of slides Wash slides with in 10% hydrochloric acid/70% ethanol (30s), DEPC-H 2O (30s), 95% ethanol (30s). Bake slides (180°C) overnight after wrapping in aluminium film. Dip slides in 2% 3-aminopropyltriethoxsilane (30s; 2% in acetone), wash in acetone (2xlmin), DEPC-H 2O (2xlmin). Dry at 37°C overnight.

G.4.2. Probe labelling for ISH G.4.2.I. Linearize DNA by digestion with restriction enzymes To two eppendorf tubes on ice were added 5 pi plasmid DNA (IGF-I rabbit cDNA in PBS, 0.62pg/pl), 2pl buffer B and 1 Ipl DEPC-H 2O. Finally 2pl of the appropriate restriction enzyme was added; Bam HI (anti-sense) and Hind III (sense). After mixing and a quick spin the tubes were incubated at 37°C for 120min.

G.4.2.2. Check digestion with agarose gel A 1% agarose gel was prepared; To 0.3g agarose, 30ml IxTAE buffer was added. After heating in a microwave to dissove (Imin) the solution was cooled to -1 5 °C, then 2pl ethidium bromide was added (care carcinogenic). The agarose solution was poured into the electrophoresis apparatus and the gel left to set for 30min. 1 xTAE buffer was poured till the gel block was covered. One sample per lane was applied. The gel was run for 30min (lOOV, 75A): Lane one, 20pl Bam HI cut DNA + Ipl load buffer Lane two, 20pl Hind III cut DNA + Ipl load buffer Lane three, 20pl uncut plasmid DNA + Ipl load buffer (control) Lane four Ipl Ikb DNA ladder + Ipl load buffer + 19pl H 2O (ladder for size of DNA) The gel was viewed under UV light; linearized DNA moves slower than uncut DNA thus the uncut DNA band is higher than the linearized DNA bands.

237 G.4.2.3. Purify linearized DNA with DNA purification matrix The Prep-A-Gene DNA Purification Matrix (Bio-Rad Laboratories, USA) method was utilised. Under UV light a scalpel blade was used to cut the gel around the band area in order to excise the two linearized DNA bands. The gel slices were placed in eppendorf tubes and centrifuged to bring the gel slice down. The volume of this slice was estimated to within 20% by placing a similar tube next to it and dropping water into the empty tube until the equivalent volume was achieved to that of the gel slice. The amount of Prep-A-Gene matrix required to bind all of the DNA present was calculated as follows; The capacity of matrix for supercoiled DNA is 0.2pg DNA per microliter of completely resuspended matrix, i.e. 5pi matrix is needed for each pg DNA to be adsorbed. The volume for both Bam HI and Hind III slices was 200pi, the DNA concentration was 0.62pg/pl for 5pi plasmid DNA, hence the DNA concentration was 3.1pg per tube. Since for each tube 5pi matrix per pg DNA is required, thus 15.5pi matrix per tube was needed. Based on the volume of the gel slice plus the amount of Prep-A-Gene required for total DNA binding, 3 volumes of Prep-A-Gene binding buffer were added to the gel slice and agitated gently to dissolve. The tube was heated at 37-55°C for several minutes to assist in dissolving the agarose, followed by the addition of the predetermined amount of Prep-A-Gene matrix. After gentle mixing by flicking the tube, or vortexing briefly, the eppendorf tubes were incubated for 5-lOmin at room temperature. End-over-end rocking during this binding step was carried out to assist DNA to Prep-A-Gene binding. Centrifuging for 30s in a micro-centrifuge resulted in pelleting of the DNA-containing Prep-A-Gene matrix. The supernatant was removed, either with a pipett tip attached to a vacuum aspirator or with a pipettor, and the pellet containing the bound DNA was rinsed. This was carried out by resuspending the pellet gently in an amount of binding buffer equivalent to 50 times the amount of added matrix, using brief vortexing or by flicking the centrifuge tube. After centrifugation as above the rinse step was repeated. The supernatant was disposed of and the Prep-A-Gene pellet washed three times with a 5Ox pellet volume of prepared wash buffer. After the last wash and centrifugation to pellet the matrix, all traces of liquid in the tube were carefully removed by removing most of the supernatant. Re-centrifugation of the tube to pellet the matrix firmly was

238 followed by removal of the remaining supernatant. Ethanol and high concentrations of salt can inhibit enzymes thus it is important to remove the last traces of wash buffer before eluting DNA from the Prep-A-Gene pellet, hence, the tubes were placed in a speed-vac for several minutes prior to the addition of the elution buffer. The bound DNA was eluted by resuspending the Prep-A-Gene matrix pellet in one volume of elution buffer (15 pi sterile double distilled water), followed by incubation at 50°C for 5mins. A spin for 5mins to make a solid pellet was then carried out. DNA dissolves in the elution buffer and at least 75% is transferred to the supernatant. Hence the supernatant was removed to a fresh tube and another 15 pi elution buffer added to the original tube. Resuspension followed by a quick spin and removal of the supernatant to the fresh tube was carried out to yield an additional 10-15% recovery. The concentration od DNA was verified, once centrifugation of the fresh tube to check for the presence of unwanted matrix had shown no trace of pellet.

G.4.2.4. Determine concentration of DNA A 1% agarose gel was run as in to check the two DNA samples. The presence of a single sharp band of the predicted size for each sample indicated the presence of only linear DNA: Lane 1; Ipl DNA sample cut by BamHI + Ipl north buffer + 8pl H^O Lane 2; 1 pi DNA sample cut by Hindlll + 1 pi north buffer + 8pl H 2O Lane 3; Ipl Ikb DNA ladder + 1 pi load buffer + 8pl H 2O

G.4.2.5. In vitro transcription to generate single-stranded riboprobe A nucleic acid (RNA) labelling kit (no. 1175 025; Boehringer Mannheim, UK, with SP6,T7 and T3 RNA polymerases) was utilised. One digoxigenin-11-UTP residue is incorporated every 20-25 nucleotides (from Ipg template DNA can transcribe -lOpg of full length DIG-labelled RNA). To two eppendorf tubes on ice were added 9pl linear DNA (~1 pg; cut by either BamHI or Hind IB), 2pl NTP labelling mix (lOx cone ), 2pl lOx transcription buffer, 2pl RNA polymerase (T3 or T7), Ipl RNase inhibitor and 4pl DEPC-H 2O. After brief mixing and centrifugation the tubes were incubated at 37° C for 2h. This was followed by the addition of 2pl DNase I to remove any remaining template DNA with

239 incubation at 37°C for 15min. The reaction was stopped by 2pl 0.2M EDTA solution (pH8.0) per tube. Precipitation of the labelled RNA, with 2.5pl 4M LiCl (0.1 vol.) and 75pl pre-chilled (-20°C) 100% ethanol (3 vol.), was accomplished by leaving the tubes at -70°C for 30min. Pelleting of the RNA was by centrifugation at 14 OOOg for 30min (4°C). The supernatant was removed and the pellet washed with 80% ethanol, 50pi per eppendorf tube (5min, cool spin). A brown pellet was obtained with any remaining ethanol removed with a pipette. The pellet was then dried in a speed-vac (10-15min, the more transparent the pellet the less the salt concentration, the better). Finally,the pellet was dissolved in lOOpl of TE buffer + 2pl RNase inhibitor, at 37°C for 30min.The riboprobe was stored at -20°C.

G.4.2.6. Run minigel to check size and concentration of templates To check the cRNA probe a 1% agarose gel was run with control RNA (or DNA) of known size and concentration. Lane one: 2pl anti-sense riboprobe + 8pl RNA load buffer Lane two: 2pl sense riboprobe + 8pl RNA load buffer Lane three: Ipl control DNA +lpl load buffer + 8pl DEPC-H 2O Lane four: Ipl Ikb DNA ladder + Ipl load buffer + 8pl DEPC-H 2O The gel was run for 30min, lOOV, 50A. The concentration of probe was determined from the band sizes. For the control DNA (4935.4), lpg/4935.4 in Ipl. For cRNA main band was taken as that of the transcript (anti-sense 1061.7, sense 942.34). Concentration of anti-sense riboprobe (Ipg x 530.85/4935.4 in Ipl = 107.5ng/pl). Concentration of sense riboprobe (Ipg x 471.2/4935.4 in Ipl = 95.47ng/pl).

G.4.2.7. Check DIG incorporation into probe using a dot blot To a small piece of nylon membrane (positively charged, Boehringer Mannheim, UK) was applied l-2pl anti-sense or sense riboprobe. The membrane was cross-linked to the cRNA with 2-3 min UV light exposure. After equilibriating the membrane in DIG buffer 1 for Imin, the membrane was blocked by gently agitating it in DIG buffer 2 for 30min

240 in a petri dish. The membrane was then incubated for 30 min with the anti-DIG antibody (anti-DIG-alkaline phosphatase in DIG buffer 2, 1:5000, working concentration 150 mU/ml). Two washes in DIG buffer 1, 15min each, to remove unbound antibody was followed by equilibriation of the membrane in DIG buffer 3 for 2min. Incubation in fresh colour solution was carried out in the dark for 12h (colour solution, 45pi NBT solution and 35pi X-phosphate in 10ml DIG buffer 3). The reaction was stopped by a 5 min wash of the membrane in DIG buffer 1.

G.4. 3. Reagents and chemicals All reagents and chemicals from Boehringer Mannheim, UK, unless specified otherwise.

Anti-digoxigenin-AP,Fab fragments; Anti-digoxigenin (Fab) conjugated to alkaline phosphatase. Blocking reagent, stock solution; for 100ml add to 85 ml buffer 1, lOg blocking reagent. Heat (microwave; 3-5mins) and shake to dissolve. Make up with buffer 1 to volume. Autoclave and store at 4°C or -20°C. Color Solution;for 10ml add to 9ml buffer 3, 45 pi NBT, 35 pi X-phosphate and 2.4 mg levamisole. Make up with buffer 3 to volume, use fresh and keep in the dark. DEPC-HjO; Add 1ml diethylpyrocarbonate (DEPC) to 1000ml double-distilled H 2O. Stir at room temperature for 30min. Autoclave and store at room temperature (Sigma, UK). Digoxigenin buffer 1;for 1000ml add to 850ml DEPC-H 2O, 11.61g maleic acid (O.IM), 8.77g NaCl (0.15M). Adjust to pH 7.5 with conc. NaOH. Make up with DEPC-H2O to volume. Autoclave and store at room temperature (note: alkaline conditions needed for DIG). Digoxigenin buffer 2;for 100ml add to 85ml buffer 1, 10ml block stock solution. Make up with buffer 1 to volume. Store at room temperature. Digoxigenin buffer 3;for 500ml add to 400ml DEPC-H 2O, 15.76g Tris-HCl (200mM, Trizma hydrochloride), 5.84g NaCl (200mM). Adjust pH to 9.5 with lOM NaOH. Make up with DEPC-H 2O to volume. Separately prepare 10.16g MgCl 2 (lOOmM) in 400ml DEPC-H2O. Make up with DEPC-H 2O to volume. Add one volume of the first solution to one volume of the MgCl 2 solution. Store at room temperature.

241 Digoxigenin buffer 4;for 500ml add to 400ml DEPC-H2O, 0.186g EDTA (1 mM), dissolve then add 0.788g Tris-HCl (lOmM). Adjust to pH 8.0 with ION NaOH. Make up with DEPC-H 2O to volume. Store at room temperature. Digoxigenin wash Buffer;for 100ml add to 85ml buffer 1, 0.3ml Tween 20 (0.3%, Tween 20 is polyoxyethylenesorbitanmonolaurate). Make up with buffer 1 to volume. Store at room temperature. DNA, labelled control; digoxigenin-labelled pBR328 DNA, random primed labelled. DNase I, RNase-free; 10 units/pl.

EDTA, 0.5M; for 1000ml add to 800ml DEPC-H 2O, 186g EDTA 2 H2 O m.wt.=372.2. Adjust to pH 8.0 with ION NaOH, solution will now dissolve. Make up with DEPC- H2O to volume. Autoclave and store at -20°C. Ethanol; absolute ethanol, chilled at -20°C, diluted in DEPC-H 2O (Sigma, UK). Glycine 2mg/ml; for 400ml add to 350ml IxPBS, glycine 800mg. Make up with IxPBS to volume. Use fresh (Sigma, UK). HCl, 0.2N ; for 500ml add to 400ml DEPC-H 2O, 20ml 5N HCl. Make up with DEPC- H2O to volume. Store at room temperature (Sigma, UK). Hybridisation Buffer; 50% de-ionised formamide, 5xSSC, 5xDenhardt’s soln, yeast t-RNA (lOOpg/ml, sterile-filtered), heat denatured and sheared salmon-sperm DNA (250pg/ml), 4mM ethylenediaminetetracetic acid (EDTA), 0.05M Na phosphate buffer pH6.5, 4xSSC, 5% dextran sulphate, 5xDenhardt’s soln (0.02% BSA, 0.02% PVP, 0.02% Ficoll 400). Store at -70°C or -20°C. LiCl, 4M; lithium chloride solution made up with DEPC-H 2O. Autoclave and store at -20°C (Sigma, UK). MOPS, lOx; morpholinopropansulfonic acid (200mM), sodium acetate (50mM), EDTA (lOmM), mix, pH7.0, autoclave. MOPS buffer; 0.01% (w/v) bromophenol blue, 50pl glycerol, make up to 500pl with DEPC-H2O. NBT solution;75 mg/ml nitroblue tétrazolium salt in 70% (v/v) dimethylformamide, DIG-RNA labelling kit, store at -20°C. NTP labelling mixture, xlO; lOmM ATP, lOmM CTP, lOmM GTP, 6.5mM UTP, 3.5mMDIG-ll-UTP; in Tris-HCl, pH7.5 (+20°C). Nylon membranes; positively charged, DIG-function tested.

242 Paraformaldehyde (4%) ; for 400ml add to 350 ml DEPC-H 2O, paraformaldehyde 16g. Heat to 60°C, add a few drops (4-5) IM NaOH. Cool in ice, add PBS tablets = 2 tabs. Filter and adjust pH to 7.2, make up with DEPC H 2O to volume. Use fresh (Sigma, UK). PBS, lOx; for 500ml add to 400ml DEPC-H 2O, 25 tablets Sigma PBS (Sigma, UK). Make up with DEPC-H 2O to volume. Autoclave and store at room temperature. Prep-A-Gene DNA Purification Matrix; Bio-Rad Laboratories, USA. Pronase 10 pg/ml, RNA free, lyophylised; for 400ml add to 350ml IxPBS, 4mg pronase. Make up with IxPBS to volume, store at -20°C. RNA dilution buffer; mix DEPC-H2O, 20xSSC and formaldehyde 5:3:2. RNA loading buffer; 250pl deionized formamide, 83 pi formaldehyde 37% (w/v), 50pl lOx. RNase 100 pg /ml, DNase free, lyophylised;for 100ml add to 85ml 2xSSC, lOmg RNase. Make up with 2xSSC to volume, store at -20°C. RNA polymerase, T7/T3/SP6; 20 units/pl. RNase inhibitor; 40000 units/ml, Pharmacia Biotech, UK. SSC, 20x; for 500ml add to 400ml DEPC-H 2O, 87.65g NaCl (3M), 44. Ig sodium citrate (0.3M). Adjust pH to 7.0 with ION NaOH at 20°C. Make up with DEPC-H 2O to volume. Autoclave and store at room temperature. TEA, 0.1 M; for 1000ml add to to 850ml DEPC-H 2O, triethanolamine.Cl 18.57g. Adjust to pH 8.0 with ION NaOH, make up with DEPC H 2O to volume. Use fresh. TE buffer; lOmM Tris, ImM EDTA (pH8.0, +20°C). Transcription buffer, xlO; 400mM Tris-HCl, pH8.0 (+20°C), 60mM MgCl 2, lOOmM DTT, lOOmM NaCl, 20mM spermidine, RNase inhibitor lU/pl. X-pbospbate solution; 50mg/ml 5 -bromo-4-chloro-3 -indolyl phosphate (X- phosphate), toluidinium salt in 100% dimethylformamide, store at -20°C.

243 Bibliography

1. Marie P. Sur deux cas d'acromegalie; hypertrophie singulière non congénitale des extrémités supérieures, inférieures et cephaliques. Rev Med Paris 1886; 6:297-333.

2. Lewis DD. Hyperplasia of the chromophil cells of the hypophysis as the cause of acromegaly, with report of a case. Johns Hopkins Med J 1905; 16:157-164.

3. Cushing H. Partial hypophysectomy for acromegaly. Ann Surg 1909; 50:1002-1007.

4. Baumann G. Growth hormone heterogeneity: genes, isohormones, variants, and binding proteins. [Review]. Endocr Rev 1991; 12:424-449.

5. Rappaport EB, Graham DJ. Pituitary growth hormone from human cadavers: neurologic disease in ten recipients. Neurology 1987; 37:1211-1213.

6. Dean HJ, Friesen HG. Growth hormone therapy in Canada: end of one era and beginning of another. Can Med Assoc J 1986; 135:297-301.

7. Buchanan CR, Preece MA, Milner RD. Mortality, neoplasia, and Creutzfeldt-Jakob disease in patients treated with human pituitary growth hormone in the United Kingdom. B M J 1991; 302:824-828.

8. Powell-Jackson J, Weller RO, Kennedy P. Creutzfeld-Jakob disease after administration of human growth hormone. Lancet 1985; 2:244-246.

9. Eshkol A. Introduction to GH therapy. Horm Res 1992; 37(suppl2):l-3.

10. Ho KY, Weissberger A, Stuart MC, Day RO, Lazarus L. The pharmacokinetics, safety and endocrine effects of authentic biosynthetic human growth hormone in normal subject. Clin Endocrinol Oxf 1989; 30:335-345.

11. Wallstrom A, Trulsson L, Guilbaud O, Gunnarsson R. The Kabi International Growth Study: concept and organization. Acta Paediatr Scand Suppl 1990; 370:105-106.

244 12. Christiansen JS, Jorgensen JO, Pedersen SA, Muller J, Jorgensen J, Moller J, et al. GH-replacement therapy in adults. Horm Res 1991; 36 Suppl 1:66-72.

13. Kaplan SL. The newer uses of growth hormone in adults. [Review]. Adv Intern Med 1993; 38:287-301.

14. Christiansen JS, Jorgensen JO. Beneficial effects of GH replacement therapy in adults. Acta Endocrinol Copenh 1991; 125:7-13.

15. Shetty KR, Duthie EH. Anterior pituitary fimction and growth hormone use in the elderly. Endocrinol Metab Clin North Am 1995; 24:213-231.

16. Bengtsson BA. The consequences of growth hormone deficiency in adults. [Review]. Acta Endocrinol Copenh 1993; 128 Suppl 2:2-5.

17. de Boer H, Blok GJ, van der Veen EA. Clinical aspects of growth hormone deficiency in adults. [Review]. Endocrine Reviews 1995; 16:63-86.

18. Jorgensen JO, Muller J, Moller J, Wolthers T, Vahl N, Juul A, et al. Adult growth hormone deficiency. [Review]. Horm Res 1994; 42:235-241.

19. Cuneo RC, Salomon F, McGauley GA, Sonksen PH. The growth hormone deficiency syndrome in adults. [Review]. Clin Endocrinol Oxf 1992; 37:387-397.

20. Bjork S, Jonsson B, Westphal O, Levin JE. Quality of life of adults with growth hormone deficiency: a controlled study. Acta Paediatr Scand Suppl 1989; 356:55-59.

21. Rosen T, Bosaeus I, Tolli J, Lindstedt G, Bengtsson BA. Increased body fat mass and decreased extracellular fluid volume in adults with growth hormone deficiency. Clin Endocrinol Oxf 1993; 38:63-71.

22. Salomon F, Cuneo RC, Hesp R, Sonksen PH. The effects of treatment with recombinant human growth hormone on body composition and metabolism in adults with growth hormone deficiency. N Engl J Med 1989; 321:1797-1803.

245 23. Jorgensen JO, Pedersen SA, Thuesen L, Jorgensen J, Ingemann-Hansen T, Christiansen JS. Beneficial effects of growth hormone treatment in GH-deficient adults. Lancet 1989; 1221-1225.

24. McGauley GA. Quality of life assessment before and afl;er growth hormone treatment in adults with growth hormone deficiency. Acta Paediatr Scand Suppl 1989; 356:70-72.

25. Whitehead HM, Boreham C, Mcllrath EM, Sheridan B, Kennedy L, Atkinson AB, et al. Growth hormone treatment of adults with growth hormone deficiency: results of a 13-month placebo controlled cross-over study. Clin Endocrinol Oxf 1992; 36:45-52.

26. Sonksen PH, Cuneo RC, Salomon F, McGauley G, Wiles CM, Wilmshurst P, et al. Growth hormone therapy in adults with growth hormone deficiency. Acta Paediatr Scand Suppl 1991; 379:139-146.

27. Degerblad M, Almkvist 0, Grunditz R, Hall K, Kaijser L, Knutsson E, et al. Physical and psychological capabilities during substitution therapy with recombinant growth hormone in adults with growth hormone deficiency. Acta Endocrinol Copenh 1990; 123:185-193.

28. Rosen T, Johannsson G, Hallgren P, Caidahl K, Bosaeus I, Bengtsson BA. Beneficial effects of 12 months replacement therapy with recombinant human growth hormone to growth hormone deficient adults. Endocrinol Metab 1994; 1:55-56.

29. Preece MA, Round JM, Jones DA. Growth hormone deficiency in adults-an indication for therapy? Acta Paediatr Scand Suppl 1987; 331:76-79.

30. Angelin B, Olivecrona H, Ericsson S, Rudling M. Growth hormone and low-density lipoproteins. [Review]. Acta Endocrinol Copenh 1993; 128 Suppl 2:26-28.

31. Haymond MW, Horber FF, Mauras N. Human growth hormone but not insulin-like growth factor I positively affects whole-body estimates of protein metabolism. Horm Res 1992; 38(suppll):73-75.

32. de Boer H, Blok GK, Voerman HJ, De Vries PMJ, van der Veen EA. Body composition in adult growth hormone-deficient men, assessed by anthropometry and bioimpedance analysis. J Clin Endocrinol Metab 1992; 75:833-837.

246 33. Binnerts A, Deurenberg P, Swart GR, Wilson JH, Lamberts SW. Body composition in growth hormone-deficient adults. Am J ClinNutr 1992; 55:918-923.

34. Keller U, Miles JM. Growth hormone and lipids. Horm Res 1991; 36 Suppl 1:36-40.

35. Sonksen PH, Salomon F, Cuneo R. Metabolic effects of hypopituitarism and acromegaly. Horm Res 1991; 36 Suppl 1:27-31.

36. Salomon F, Cuneo R, Sonksen PH. Growth hormone and protein metabolism. Horm Res 1991; 36(suppll):41-43.

37. Clemmons DR, Underwood LE. Clinical review 59: Uses of human insulin-like growth factor-I in clinical conditions. [Review]. J Clin Endocrinol Metab 1994; 79:4-6.

38. Jadresic A, Banks LM, Child DF, Diamant L, Doyle FH, Fraser TR, et al. The acromegaly syndrome. Relation between clinical features, growth hormone values and radiological characteristics of the pituitary tumours. Q J Med 1982; 51:189-204.

39. Nabarro JD. Acromegaly. [Review]. Clin Endocrinol Oxf 1987; 26:481-512.

40. Mohtch ME. Clinical manifestations of acromegaly. [Review]. Endocrinol Metab Clin North Am 1992; 21:597-614.

41. Roach P, Eastman RC. Acromegaly. Curr Opin Endocrinol Diabet 1994; 100-105.

42. Rudd BT. Growth, growth hormone and the somatomedins: a historical perspective and current concepts. [Review]. Ann Clin Biochem 1991; 28:542-555.

43. Baumann G. Growth hormone-binding proteins. [Review]. Proc Soc Exp Biol Med 1993; 202:392-400.

44. Leung DW, Spencer SA, Cachianes G, Hammonds RG, Collins C, Henzel WJ, et al. Growth hormone receptor and serum binding protein: purification, cloning and expression. Nature 1987; 330:537-543.

247 45. Sotiropoulos A, Goujon L, Simonin G, Kelly PA, Postel-Vinay MC. Evidence for generation of the grovvth hormone-binding protein through proteolysis of the growth hormone membrane receptor. Endocrinology 1993; 132:1863-1865.

46. Kelly PA, Goujon L, Sotiropoulos A, Dinerstein H, Esposito N, Edery, et al. The GH receptor and signal transduction. [Review]. Horm Res 1994; 42:133-139.

47. Rosenfeld RG. Circulating growth hormone binding proteins. [Review]. Horm Res 1994; 42:129-132.

48. Spencer SA, Leung DW, Godowski PJ, Hammonds RG, Waters MJ, Wood WI. Growth hormone receptor and binding protein. [Review]. Rec Prog Horm Res 1990; 46:165-81; discus.

49. Daughaday WH, Trivedi B. Absence of serum growth hormone binding protein in patients with growth hormone receptor deficiency (Laron dwarfism). Proc Natl Acad Sci USA 1987; 84:4636-4640.

50. Rotwein P, Gronowski AM, Thomas MJ. Rapid nuclear actions of growth hormone. [Review]. Horm Res 1994; 42:170-175.

51. Kelly PA, Ali S, Rozakis M, Goujon L, Nagano M, Pellegrini I, et al. The growth hormone/prolactin receptor family. [Review]. Rec Prog Horm Res 1993; 48:123-164.

52. Silbergeld A, Klinger B, Keret R, Eshet R, Almonte AS, Laron Z. Serum growth hormone-binding protein (GHBP) activity is decreased by administration of insulin-like growth factor I in three Laron syndrome siblings with normal GHBP. Proc Soc Exp Biol Med 1994; 206:324-327.

53. Gluckman PD, Butler AA. Endocrine/paracrine interactions following administration of insulin-like growth factor I. [Review]. Acta Paediatrica 1994; Supplement. 399:180-182.

54. Wells JA, Cunningham BC, Fuh G, Lowman HB, Bass SH, Mulkerrin MG, et al. The molecular basis for growth hormone-receptor interactions. [Review]. Rec Prog Horm Res 1993; 48:253-275.

248 55. Frank SJ, Gilliland G, Kraft AS, Arnold CS. Interaction of the growth hormone receptor cytoplasmic domain with the JAK2 tyrosine kinase. Endocrinology 1994; 135:2228-2239.

56. Chow JC, Ling PR, Qu Z, Laviola L, Ciccarone A, Bistrian BR, et al. Growth hormone stimulates tyrosine phosphorylation of JAK2 and STAT5, but not insulin receptor substrate-1 or SHC proteins in liver and skeletal muscle of normal rats in vivo. Endocrinology 1996; 137:2880-2886.

57. Sotiropoulos A, Perrot-Applanat M, Dinerstein H, Pallier A, Postel-Vinay MC, Finidori J, et al. Distinct cytoplasmic regions of the growth hormone receptor are required for activation of JAK2, mitogen-activated protein kinase, and transcription. Endocrinology 1994; 135:1292-1298.

58. Waters MJ, Rowlinson SW, Clarkson RW, Chen CM, Lobie PE, Norstedt G, et al. Signal transduction by the growth hormone receptor. [Review]. Proc Soc Exp Biol Med 1994; 206:216-220.

59. Lobie PE, Barnard R, Waters MJ. The nuclear growth hormone receptor binding protein. Antigenic and physicochemical characterization. J Biol Chem 1991; 266:22645-22652.

60. Lobie PE, Mertani H, Morel G, Morales-Bustos O, Norstedt G, Waters, et al. Receptor-mediated nuclear translocation of growth hormone. J Biol Chem 1994; 269:21330-21339.

61. Corpas E, Harman SM, Blackman MR Human growth hormone and human aging. [Review]. Endocr Rev 1993; 14:20-39.

62. Ho KY, Weissberger A. Secretory patterns of growth hormone according to sex and age. Horm Res 1990; 33 Suppl 4:7-11.

63. Van Den Berg G, Veldhuis JD, Frolich M, Roelfsema F. An amplitude-specific divergence in the pulsatile mode of growth hormone (GH) secretion underlies the gender difference in mean GH concentrations in men and premenopausal women. Endocrinology 1996; 81:2460-2467.

64. Cuneo RC, Wallace JD. Growth hormone, insulin-like growth factors and sport. Endocrinol Metab 1994; 3-13.

249 65. Horvath S, Palkovits M, Gores T, Arimura A. Electron microscopic immimocytochemical evidence for the existence of bidirectional synaptic connections between growth hormone-releasing hormone- and somatostatin-containing neurons in the hypothalamus of the rat. Brain Res 1989; 481:8-15.

66. Hindmarsh PC. Growth and growth hormone secretion. Curr Opin Endocrinol Diabet 1995; 2:91-96.

67. Brook CG, Hindmarsh PC, Stanhope R. Growth and growth hormone secretion. [Review]. J Endocrinol 1988; 119:179-184.

68. Mayo KE, Godfrey PA, Suhr ST, Kulik DJ, Rahal JO. Growth hormone-releasing hormone: synthesis and signaling. [Review]. Rec Prog Horm Res 1995; 50:35-73.

69. Kamegai J, Minami S, Sugihara H, Higuchi H, Wakabayashi 1. Growth hormone induces expression of the c-fos gene on hypothalamic neuropeptide-Y and somatostatin neurons in hypophysectomized rats. Endocrinology 1994; 135:2765-2771.

70. Frohman LA, Jansson JO. Growth hormone-releasing hormone. [Review]. Endocr Rev 1986; 7:223-253.

71. Sugihara H, Minami S, Okada K, Kamegai J, Hasegawa O, Wakabayashi 1. Somatostatin reduces transcription of the growth hormone gene in rats. Endocrinology 1993; 132:1225-1229.

72. Tannenbaum GS, McCarthy GF, Zeitler P, Beaudet A. Cysteamine-induced enhancement of growth hormone-releasing factor (GRF) immunoreactivity in arcuate neurons: morphological evidence for putative somatostatin/GRF interactions within hypothalamus. Endocrinology 1990; 127:2551-2560.

73. Phelps CJ, Dalcik H, Endo H, Talamantes F, Hurley DL. Growth hormone-releasing hormone peptide and mRNA are overexpressed in GH-deficient Ames dwarf mice. Endocrinology 1993; 133:3034-3037.

74. Bertherat J, Timsit J, Bluet-Pajot M l, Mercadier JJ, Gourdji D, Kordon C, et al. Chronic growth hormone (GH) hypersecretion induces reciprocal and reversible changes in mRNA levels

250 from hypothalamic GH-reieasing hormone and somatostatin neurons in the rat. J Clin Invest 1993;91:1783-1791.

75. Sato M, Frohman LA. Differential effects of central and peripheral administration of growth hormone (GH) and insulin-like growth factor on hypothalamic GH-releasing hormone and somatostatin gene expression in GH-deficient dwarf rats. Endocrinology 1993; 133:793-799.

76. Minami S, Kamegai J, Sugihara H, Hasegawa O, Wakabayashi 1. Systemic administration of recombinant human growth hormone induces expression of the c-fos gene in the hypothalamic arcuate and periventricular nuclei in hypophysectomized rats. Endocrinology 1992; 131:247-253.

77. Jaffe CA, Barkan AL. Treatment of acromegaly with dopamine agonists. [Review]. Endocrinol Metab Clin North Am 1992; 21:713-735.

78. Fairhall KM, Mynett A, Robinson 1C. Central effects of growth hormone-releasing hexapeptide (GHRP-6) on growth hormone release are inhibited by central somatostatin action. J Endocrinol 1995; 144:555-560.

79. Chihara K, Iwasaki J, Minamitani N, Kaji H, Matsukura S, Tamaki N, et al. Effect of vasoactive intestinal polypeptide on growth hormone secretion in perifused acromegalic pituitary adenoma tissues. J Clin Endocrinol Metab 1982; 54:773-779.

80. Epelbaum J, Tapia-Arancibia L, Besson J, Rotsztejn WH, Kordon C. Vasoactive intestinal peptide inhibits release of somatostatin from hypothalamus in vitro. Eur J Pharmacol 1979; 58:493-495.

81. Dalcik H, Phelps CJ. Median eminence-afferent vasoactive intestinal peptide (VIP) neurons in the hypothalamus: localization by simultaneous tract tracing and immunocytochemistry. Peptides 1993; 14:1059-1066.

82. Fahim A, Rettori V, McCann SM. The role of calcitonin gene-related peptide in the control of growth hormone and prolactin release. Neuroendocrinology 1990; 51:688-693.

251 83. Netti C, Guidobono F, Sibilia V, Pagani F, Braga PC, Pecile A. Evidence of a central inhibition of growth hormone secretion by calcitonin gene-related peptide. Neuroendocrinology 1989; 49:242-247.

84. Muller EE, Locatelli V, Ghigo E, Celia SG, Loche S, Pintor C. Involvement of brain catecholamines and acetylcholine in growth hormone deficiency states. Pathophysiological, diagnostic and therapeutic implications. [Review]. Drugs 1991; 41:161-177.

85. Conway S, Richardson L, Spéciale S, Moherek R, Mauceri H, Krulich L. Interaction between norepinephrine and serotonin in the neuroendocrine control of growth hormone release in the rat. Endocrinology 1990; 126:1022-1030.

86. Salmon WD, Daughaday WH. A hormonally controlled serum factor which stimulates sulfate incorporation by cartilage in vitro. J Lab Clin Med 1957; 49:825-836.

87. Quin JD. The insulin-like growth factors. [Review]. Q J Med 1992; 82:81-90.

88. Leroith D, Adamo M, Werner H, Roberts CT, Jr. Insulin-like growth factors and their receptors as growth regulators in normal physiology and pathological states. Trends Endocrinol Metab 1991;2:134-139.

89. Daughaday WH, Rotwein P. Insulin-like growth factors I and II. Peptide, messenger ribonucleic acid and gene structures, serum, and tissue concentrations. [Review]. Endocr Rev 1989; 10:68-91.

90. Malozowski S, Parmer TG, Trojan S, Merriam GR, Gibori G, Roberts CT, et al. Growth hormone (GH) modulates insulin-like growth factor I (IGF-I) and type I IGF receptor mRNA levels in the ovary of prepubertal GH-deficient rats. Eur J Endocrinol 1995; 132:497-501.

91. Holly JM, Wass JAH. Insulin-like growth factors; autocrine, paracrine or endocrine? New perspectives of the somatomedin hypothesis in the light of recent developments. [Review]. J Endocrinol 1989; 122:611-618.

92. Leroith D, Roberts CT, Jr. Insulin-like growth factor I (IGF-I): a molecular basis for endocrine versus local action?. [Review]. Mol Cell Endocrinol 1991; 77:C57-C6I.

252 93. Jones JI, Clemmons DR. Insulin-like growth factors and their binding proteins: biological actions. [Review]. Endocr Rev 1995; 16:3-34.

94. Hintz RL. Role of growth-hormone and insulin-like growth-factor-binding proteins. [Review]. Horm Res 1990;33:105-110.

95. Moxham CP, Jacobs S. lnsulin/lGF-1 receptor hybrids: a mechanism for increasing receptor diversity. [Review]. J Cell Biochem 1992; 48:136-140.

96. de Meyts P, Wallach B, Christofifersen CT, Urso B, Gronskov K, Latus, et al. The insulin-like growth factor-1 receptor. Structure, hgand-binding mechanism and signal transduction. [Review]. Horm Res 1994; 42:152-169.

97. Du J, Meng XP, Delafontaine P. Transcriptional regulation of the insulin-like growth factor-1 receptor gene: evidence for protein kinase C-dependent and -independent pathways. Endocrinology 1996; 137:1378-1384.

98. Thomas MJ, Kikuchi K, Bichell DP, Rotwein P. Rapid activation of rat insulin-like growth factor-1 gene transcription by growth hormone reveals no alterations in deoxyribonucleic acid-protein interactions within the major promoter. Endocrinology 1994; 135:1584-1592.

99. Zanconato S, Moromisato DY, Moromisato MY, Woods J, Brasel JA, Leroith D, et al. Effect of training and growth hormone suppression on insulin-like growth factor 1 mRNA in young rats. J Appl Physiol 1994; 76:2204-2209.

100. Mathews LS, Norstedt G, Palmiter RD. Regulation of insulin-like growth factor 1 gene expression by growth hormone. Proc Natl Acad Sci USA 1986; 83:9343-9347.

101. Trainer PJ, Holly J, Medbak S, Rees LH, Besser GM. The effect of recombinant lGF-1 on anterior pituitary fimction in healthy volunteers. Clin Endocrinol Oxf 1994; 41:801-807.

102. Harel Z, Tannenbaum GS. Synergistic interaction between insulin-like growth factors-1 and -n in central regulation of pulsatile growth hormone secretion. Endocrinology 1992; 131:75 8-764.

253 103. Henderson TA, Johnson EM, Jr., Osborne PA, Jacquin MF. Fetal NGF augmentation preserves excess trigeminal ganglion cells and interrupts whisker-related pattern formation. J Neurosci 1994; 14:3389-3403.

104. Fagin JA, Fernandez Mejia C, Melmed S. Pituitary insulin-like growth factor-I gene expression: regulation by triiodothyronine and growth hormone. Endocrinology 1989; 125:2385-2391.

105. Leung K, Rajkovic LA, Peters E, Markus I, Van Wyk JJ, Ho KY. Insulin-like growth factor 1 and insulin down-regulate growth hormone (GH) receptors in rat osteoblasts: Evidence for a peripheral feedback loop regulating GH action. Endocrinology 1996; 137:2694-2702.

106. Guler H, Zapf J, Froesh ER. Short-term metabolic effects of recombinant human insulin-like growth factor 1 in healthy adults. N Engl J Med 1987; 317:137-140.

107. Gluckmann PD, Ambler GR What is the function of circulating insulin-like growth factor-2 in postnatal life?. [Review]. Mol Cell Endocrinol 1993; 92:C1-C3.

108. Laron Z. Somatomedin-1 (insulin-like growth factor-I) in clinical use. Facts and potential. [Review]. Drugs 1993; 45:1-8.

109. Cotterill AM, Camacho-Hubner C, Holly JM, Savage MO. The effect of recombinant human insulin-like growth factor-I treatment on growth hormone secretion in two subjects with growth hormone insensitivity (Laron syndrome). Clin Endocrinol Oxf 1993; 39:119-122.

110. Laron Z, Klinger B. IGF-I treatment of adult patients with Laron syndrome: preliminary results. Clin Endocrinol Oxf 1994; 41:631-638.

111. LaiZ, Roos P, Zhai O, Olsson Y, Fholenhag K, Larsson C, et al. Age-related reduction of human growth hormone-binding sites in the human brain. Brain Res 1993; 621:260-266.

112. Zhai Q, Lai Z, Roos P, Nyberg F. Characterization of growth hormone binding sites in rat brain. Acta Paediatr.Suppl. 1994; 406:92-95.

254 113. Mertani HC, Waters MJ, Jambou R, Gossard F, Morel G. Growth hormone receptor binding protein in rat anterior pituitary. Neuroendocrinology 1994; 59:483-494.

114. Lobie PE, Garcia-Aragon J, Lincoln DT, Barnard R, Wilcox JN, WatersMJ. Localization and ontogeny of growth hormone receptor gene expression in the central nervous system. Dev Brain Res 1993;74:225-233.

115. Noguchi T, Sugisaki T, NishikawaN, Tsukada Y. Restoration of microcephalic cerebrum with hypomyelination in the growth hormone-deficient mouse (lit): stimulatory effects of GH restricted to the first 20 days of postnatal life. Neurochem Res 1988; 13:249-252.

116. Johansson JO, Larson G, Andersson M, Elmgren A, Hynsjo L, Lindahl A, et al. Treatment of growth hormone-deficient adults with recombinant human growth hormone increases the concentration of growth hormone in the cerebrospinal fluid and affects neurotransmitters. Neuroendocrinology 1995; 61:57-66.

117. Dinn JJ, Dinn El. Natural history of acromegalic peripheral neuropathy. Q J Med 1985; 57:833-842.

118. Jamal GA, Kerr DJ, McLellan AR, Weir AI, Davies DL. Generalised peripheral nerve dysfiinction in acromegaly: a study by conventional and novel neurophysiological techniques. J Neurol Neurosurg Psychiat 1987; 50:886-894.

119. Laron Z, Anin S, Klipper Aurbach Y, Klinger B. Effects of insulin-like growth factor on linear growth, head circumference, and body fat in patients with Laron-type dwarfism. Lancet 1992;339:1258-1261.

120. Laron Z, Klinger B. Laron syndrome: clinical features, molecular pathology and treatment. Horm Res 1994; 42:198-202.

121. Rosen T, Wiren L, WiUiehnsen L, Wiklund I, Bengtsson BA. Decreased psychological well-being in adult patients with growth hormone deficiency. Clin Endocrinol Oxf 1994; 40:111-116.

255 122. McGauley GA, Cuneo RC, Salomon F, Sonksen PH. Psychological well-being before and after growth hormone treatment in adults with growth hormone deficiency. Horm Res 1990; 33 Suppl 4:52-54.

123. Bengtsson BA, Eden S, Lonn L, Kvist H, Stokland A, Lindstedt G, et al. Treatment of adults with growth hormone (GH) deficiency with recombinant human GH. J Clin Endocrinol Metab 1993;76:309-317.

124. Le Guellec D, Trembleau A, Pechoux C, Gossard F, Morel G. Ultrastructural non-radioactive in situ hybridization of GH mRNA in rat pituitary gland: pre-embedding vs ultra-thin fi-ozen sections vs post-embedding. J Histochem Cytochem 1992; 40:979-986.

125. Phelps CJ, HofiBnan GE. Isolated deficiency of immunocytochemically detected somatostatin in Snell dwarf, but not in "little," mice. Peptides 1987; 8:1127-1133.

126. Hurley DL, Bartke A, Wagner TE, Wee BE, Phelps CJ. Increased hypothalamic somatostatin expression in mice transgenic for bovine or human GH. J Neuroendocrinol 1994; 6:539-548.

127. Phelps CJ, Hurley DL. Development of hypophysiotropic neuron abnormalities in GH- and PRL-deficient dwarf mice. Rec Prog Horm Res 1993; 48:489-496.

128. Hurley DL, Phelps CJ. Hypothalamic preprosomatostatin messenger ribonucleic acid expression in mice transgenic for excess or deficient endogenous growth hormone. Endocrinology 1992; 130:1809-1815.

129. Phelps CJ, Carlson S, Hurley DL. Hypothalamic dopaminergic neurons in transgenic dwarf mice: histofluorescence, immunocytochemical, and in situ hybridization studies. Anat Rec 1991; 231:446-456.

130. Phelps CJ. Pituitary hormones as neurotrophic signals: anomalous hypophysiotrophic neuron differentiation in hypopituitary dwarf mice. [Review]. Proc Soc Exp Biol Med 1994; 206:6-23.

131. Kanje M, Skottner A, Lundborg G. Effects of growth hormone treatment on the regeneration of rat sciatic nerve. Brain Res 1988; 475:254-258.

256 132. UUman M, Oldfors A. Effects of growth hormone on skeletal muscle. 1. Studies on normal adult rats. Acta Physiol Scand 1989; 135:531-536.

133. Ullman M, Alameddine H, Skottner A, Oldfors A. Effects of growth hormone on skeletal muscle. II. Studies on regeneration and denervation in adult rats. Acta Physiol Scand 1989; 135:537-543.

134. Recio-Pinto E, Ishii DN. Effects of insulin, insulin-like growth factor-ll and nerve growth factor on neurite outgrowth in cultured human neuroblastoma cells. Brain Res 1984; 302:323-334.

135. Recio-Pinto E, Lang FF, Ishii DN. Insulin and insulin-like growth factor 11 permit nerve growth factor binding and the neurite formation response in cultured human neuroblastoma cells. Proc Natl Acad Sci USA 1984; 81:2562-2566.

136. Recio-Pinto E, Rechler MM, Ishii DN. Effects of insulin, insulin-like growth factor-11, and nerve growth factor on neurite formation and survival in cultured sympathetic and sensory neurons. J Neurosci 1986; 6:1211-1219.

137. NeffNT, Prevette D, Houenou LJ, Lewis ME, Glicksman MA, Yin QW, et al. Insulin-like growth factors: putative muscle-derived trophic agents that promote motoneuron survival. J Neurobiol 1993; 24:1578-1588.

138. Li L, Oppenheim RW, Lei M, Houenou LJ. Neurotrophic agents prevent motoneuron death following sciatic nerve section in the neonatal mouse. J Neurobiol 1994; 25:759-766.

139. Ishii DN. Neurobiology of insulin and insulin-like growth factors. In: Loughlin SE, Fallon JH, editors. Neurotrophic factors. London: Academic Press, Inc. 1993:415-442.

140. Caroni P, Becker M. The downregulation of growth-associated proteins in motoneurons at the onset of synapse elimination is controlled by muscle activity and IGFl. J Neurosci 1992; 12:3849-3861.

141. Burgess SK, Jacobs S, Cuatrecasas P, SahyounN. Characterization of a neuronal subtype of msulin-like growth factor 1 receptor. J Biol Chem 1987; 262:1618-1622.

257 142. Lesniak MA, Hill JM, Kiess W, Rojeski M, Pert CB, Roth J. Receptors for insulin-like growth factors I and H: autoradiographic localization in rat brain and comparison to receptors for insulin. Endocrinology 1988; 123:2089-2099.

143. Shemer J, Raizada MK, Masters BA, Ota A, Leroith D. Insulin-like growth factor I receptors in neuronal and ghal cells. Characterization and biological effects in primary culture. J Biol Chem 1987; 262:7693-7699.

144. Rotwein P, Burgess SK, Milbrandt JD, Krause JE. Differential expression of insulin-like growth factor genes in rat central nervous system. Proc Natl Acad Sci USA 1988; 85:265-269.

145. Schechter R, Whitmire J, Beju D, Jackson KW, Harlow R, Gavin JR, 3rd. An immunohistochemical and in situ hybridization study of insulin-like growth factor 1 within fetal neuron cell cultures. Brain Res 1995; 670:1-13.

146. Masters BA, Werner H, Roberts CT, Jr., Leroith D, Raizada MK. Insulin-like growth factor I (lGF-1) receptors and lGF-1 action in oligodendrocytes from rat brains. Regul Pept 1991; 33:117-131.

147. Matsuo K, Niwa M, Kurihara M, Shigematsu K, Yamashita S, Ozaki M. Receptor autoradiographic analysis of insulin-like growth factor-1 (lGF-1) binding sites in rat forebrain and pituitary gland. Cell Mol Neurobiol 1989; 9:357-367.

148. Torres-Aleman 1, Naftolin F, Robbins RJ. Trophic effects of insulin-like growth factor-1 on fetal rat hypothalamic cells in culture. Neuroscience 1990; 35:601-608.

149. McMorris FA, Smith TM, DeSalvo S, Furlanetto RW. Insulin-like growth factor I/somatomedin C: a potent inducer of oligodendrocyte development. Proc Natl Acad Sci USA 1986; 83:822-826.

150. Leroith D, Roberts CT, Werner H, Bondy C, Raizada M, Adamo ML. Insulin-like growth factors in the brain. In: Loughlin SE, Fallon JH, editors. Neurotrophic factors. London: Academic Press, Inc. 1993:391-414.

258 151. Aizenman Y, de Vellis J. Brain neurons develop in a serum and glial free environment: effects of transferrin, insulin, insulin-like growth factor-I and thyroid hormone on neuronal survival, growth and differentiation. Brain Res 1987; 406:32-42.

152. Xue ZG, Le Douarin NM, Smith J. Insulin and insulin-like growth factor-1 can trigger the differentiation of catecholaminergic precursors in cultures of dorsal root ganglia. Cell Different Dev 1988;25:1-10.

153. Pahlman S, Meyerson G, Lindgren E, Schalling M, Johansson 1. Insulin-like growth factor 1 shifts from promoting cell division to potentiating maturation during neuronal differentiation. Proc Natl Acad Sci USA 1991; 88:9994-9998.

154. Martin DM, YEE D, Carlson RO, Feldman EL. Gene expression of the insulin-like growth factors and their receptors in human neuroblastoma cell lines. Mol.Brain Res. 1992; 15:241-246.

155. Zackenfels K, Oppenheim RW, Rohrer H. Evidence for an important role of IGF-1 and IGF-II for the early development of chick sympathetic neurons. Neuron 1995; 14:731-741.

156. Beck KD, Knusel B, Hefti F. The nature of the trophic action of brain-derived neurotrophic factor, des( 1 -3)-insulin-like growth factor-1, and basic fibroblast growth factor on mesencephalic dopaminergic neurons developing in culture. Neuroscience 1993; 52:855-866.

157. Gutierrez-Ospina G, Calikoglu AS, Ye P, D'Ercole AJ. In vivo effects of insulin-like growth factor-I on the development of sensory pathways: Analysis of the primary somatic sensory cortex (SI) of transgenic mice. Endocrinology 1996; 137:5484-5492.

158. Carson MJ, Behringer RR, Brinster RL, McMorris FA. Insulin-like growth factor I increases brain growth and central nervous system myelination in transgenic mice. Neuron 1993; 10:729-740.

159. Rosenfeld RG, Bengtsson BA. Effects of growth hormone and insulin-like growth factors on the central nervous system. [Review]. Acta Paediatrica 1994; Supplement. 406:89-91.

259 160. Caruso-Nicoletti M, Mancuso M, Mandara G, Scuderi A, di Pietro M. Occurrence of myelinated nerve fibres in a patient with Laron syndrome treated with insulin-like growth factor I. Acta Paediatr.Suppl. 1995; 411:108 Abstract.

161. Andersson I, Billig H, Fryklund L, Hansson HA, Isaksson O, Isgaard J, et al. Localization of lGF-1 in adult rats. Immunohistochemical studies. Acta Physiol Scand 1986; 126:311-312.

162. Hansson HA, Nilsson A, Isgaard J, Billig H, Isaksson O, Skottner A, et al. Immunohistochemical localization of insulin-like growth factor I in the adult rat: Histochemistry 1988; 89:403-410.

163. Marks JL, Porte D, Jr., Baskin DG. Localization of type I insulm-like growth factor receptor messenger RNA in the adult rat brain by in situ hybridization. Mol Endocrinol 1991; 5:1158-1168.

164. Aguado F, Fernandez T, Martinez-Murillo R, Rodrigo J, Cacicedo L, Sanchez-Franco F. Immunocytochemical localization of insulin-like growth factor I in the hypothalamo-hypophyseal system of the adult rat. Neuroendocrinology 1992; 56:856-863.

165. Sara VR, Carlsson-Skwirut C, Drakenberg K, Giacobini MB, Hakansson L, Nordberg A, et al. The biological role of truncated insulin-like growth factor-1 and the tripeptide GPE in the central nervous system. [Review]. Ann NY Acad Sci 1993; 692:183-191.

166. Ishii DN, Marsh DJ. On the therapeutic potential for insulin-like growth factor use in motor neuron disease. [Review]. Exp Neurol 1993; 124:96-99.

167. Kanje M, Skottner A, Sjoberg J, Lundborg G. Insulin-like growth factor I (lGF-1) stimulates regeneration of the rat sciatic nerve. Brain Res 1989; 486:396-398.

168. Sjoberg J, Kanje M. Insulin-like growth factor (IGF-1) as a stimulator of regeneration in the freeze-injured rat sciatic nerve. Brain Res 1989; 485:102-108.

169. Contreras PC, Steffler C, Yu E, Callison K, Stong D, Vaught JL. Systemic administration of rhIGF-1 enhanced regeneration after sciatic nerve crush in mice. J Pharmacol Exp Therap 1995 ; 274:1443-1449.

260 170. Hansson HA, Dahlin LB, Danielsen N, Fryklund L, Nachemson AK, Rozell B, et al. Evidence indicating trophic importance of IGF-I in regenerating peripheral nerves. Acta Physiol Scand 1986; 126:609-614.

171. Nachemson AK, Hansson HA, Lundborg G. Neurotropism in nerve regeneration, an immunohistochemical study. Acta Physiol Scand 1988; 133:139-148.

172. Gehrmann J, Yao DL, Bonetti B, Bondy CA, Breimer M, Zhou J, et al. Expression of insulin-like growth factor-I and related peptides during motoneuron regeneration. Exp Neurol 1994; 128:202-210.

173. Hansson HA, Rozell B, Skottner A. Rapid axoplasmic transport of insulin-like growth factor I in the sciatic nerve of adult rats. Cell Tissue Res 1987; 247:241-247.

174. Jung-Testas I, Schumacher M, Robel P, Baulieu EE. Actions of steroid hormones- and growth factors on glial cells of the central and peripheral nervous system. J Steroid Biochem Molec Biol 1994; 48:145-154.

175. Cheng HL, Randolph A, YEE D, Delafontaine P, Tennekoon G, Feldman EL. Characterization of insulin-like growth factor-I and its receptor and binding proteins in transected nerves and cultured Schwann cells. J Neurochem 1996; 66:525-536.

176. Rabinovsky ED, Le WD, McManaman JL. Differential effects of neurotrophic factors on neurotransmitter development in the IMR-32 human neuroblastoma cell line. J Neurosci 1992; 12:171-179.

177. Aguado F, Sanchez-Franco F, Rodrigo J, Cacicedo L, Martinez-Murillo R. Insulin-like growth factor I-immunoreactive peptide in adult human cerebellar Purkinje cells: co-localization with low-affinity nerve growth factor receptor. Neuroscience 1994; 59:641-650.

178. Patterson PH. Cytokines and the function of the mature nervous system. [Review]. Comptes Rendus de 1 Académie des Sciences - Serie lii. Sciences de la Vie 1993; 316:1141-1157.

179. Raben MS. Growth hormone. 2. Clinical use of human growth hormone. N Engl J Med 1962; 266:82-86.

261 180. Degerblad M, Grunditz R, Hall K, Sjoberg HE, Saaf M, Thoren M. Substitution therapy with recombinant growth hormone (somatrem) in adults with growth hormone deficiency. Acta Paediatr Scand Suppl 1987; 337:170-171.

181. Hager A. Should adults with growth hormone deficiency be maintained on growth hormone substitution therapy? Acta Paediatr Scand Suppl 1989; 362:72-75.

182. Salomon F, Cuneo RC, Hesp R, Sonksen PH. Growth hormone deficiency in adults. Acta Paediatr Scand Suppl 1989; 356:69 Abstract.

183. Laursen T, Jorgensen JO, Christiansen JS. Metabolic effects of growth hormone administered subcutaneously once or twice daily to growth hormone deficient adults. Clin Endocrinol Oxf 1994; 41:337-343.

184. Laursen T, Jorgensen JOL, Christiansen JS. Metabolic responses to growth hormone (GH) administered in a pulsatile, continuous or combined pattern. Endocrinol Metab 1994; 1:33-40.

185. Johannsson G, Bjamason R, Bramnert M, Carlsson LMS, Degerblad M, Manhem P, et al. The individual responsiveness to growth hormone (GH) treatment in GH-deficient adults is dependent on the level of GH-binding protein, body mass index, age, and gender. J Clin Endocrinol Metab 1996; 81:1575-1581.

186. Cittadini A, Cuocolo A, Merola B, Fazio S, Sabatini D, Nicolai E, et al. Impaired cardiac performance in GH-deficient adults and its improvement after GH replacement. Am J Physiol 1994; 267:E219-25.

187. Rutherford OM, Beshyah SA, Schott J, Watkins Y, Johnston DG. Contractile properties of the quadriceps muscle in growth hormone-deficient hypopituitary adults. Clin Sci Oxf 1995; 88:67-71.

188. Rutherford OM, Beshyah SA, Johnston DG. Quadriceps strength before and after growth hormone replacement in hypopituitary adults: relationship to changes in lean body mass and IGF-I. Endocrinol Metab 1994; 1:41-47.

262 189. Sartorio A, Narici M, Conti A, Monzani M, Faglia G. Quadriceps and hand-grip strength in adults with childhood-onset growth hormone deficiency. Eur J Endocrinol 1995; 132:37-41.

190. Sartorio A, Narici MV. Growth hormone (GH) treatment in GH-deficient adults: effects on muscle size, strength and neural activation. Clin Physiol 1994; 14:527-537.

191. Gluckmann PD, Ambler GR. Therapeutic use of insulin-like growth factor I: lessons from in vivo animal studies. [Review]. Acta Paediatr.Suppl. 1992; 383:134-136.

192. Lin S, Lin CR, Gurovsl^ 1, Lusis AJ, Sawchenko PE, Rosenfeld MG. Molecular basis of the httle mouse phenotype and implications for cell type-specific growth. Nature 1993; 364:208-213.

193. Brown-Borg HM, Borg KE, Mellska CJ, Bartke A. Dwarf mice and the ageing process (scientific correspondence). Nature 1996; 384:33

194. Noguchi T, Sugisaki T, Kanamatsu T, Satoh 1, Nishikawa N. Reduced vasoactive intestinal polypeptide immunoreactivity in the pituitaries of hormone-deficient mutant mice. J Endocrinol 1988; 118:179-185.

195. Charlton HM, Clark RG, Robinson 1C, Goff AE, Cox BS, Bugnon C, et al. Growth hormone-deficient dwarfism in the rat: a new mutation. J Endocrinol 1988; 119:51-58.

196. Wanke R, Wolf E, Hermanns W, Folger S, Buchmuller T, Brem G. The GH-transgenic mouse as an experimental model for growth research: clinical and pathological studies. [Review]. Horm Res 1992; 37 Suppl 3:74-87.

197. Behringer RR, Lewin TM, Quaife CJ, Palmiter RD, Brinster RL, D'Ercole AJ. Expression of insulin-like growth factor I stimulates normal somatic growth in growth hormone-deficient transgenic mice. Endocrinology 1990; 127:1033-1040.

198. Ayling CM, Moreland BH, Zanelli JM, Schulster D. Human growth hormone treatment of hypophysectomized rats increases the proportion of type-1 fibres in skeletal muscle. J Endocrinol 1989; 123:429-435.

263 199. Clark RG, Carlsson LMS, Mortensen D, Cronin MJ. Additive effects on body growth of insulin-like growth factor-I and growth hormone in hypophysectomized rats. Endocrinol Metab 1994; 1:49-54.

200. Ho KY, Veldhuis JD. Diagnosis of growth hormone deficiency in adults (workshop report). Endocrinol Metab 1994; 1:61-63.

201. Thomer MO, Bengtsson BA, Ho KY, Albertsson-Wikland K, Christiansen JS, Faglia G, et al. The diagnosis of growth hormone deficiency (GHD) in adults (letter). J Clin Endocrinol Metab 1995; 80:3097-3098.

202. Jorgensen JO, Pedersen SA, Laurberg P, Weeke J, Skakkebaek NE, Christiansen JS. Effects of growth hormone therapy on thyroid function of growth hormone-deficient adults with and without concomitant thyroxine-substituted central hypothyroidism. J Clin Endocrinol Metab 1989; 69:1127-1132.

203. Thomas MR, Miell JP, Taylor AM, Ross RJ, Amao JR, Jewitt DE, et al. Endocrine and cardiac paracrine actions of insulin-like growth factor-1 (lGF-1) during thyroid dysfimction in the mechanism of heart weight/body weight change during abnormal thyroid fimction?. J Mol Endocrinol 1993; 10:313-323.

204. Miell JP, Taylor AM, Zini M, Maheshwari HG, Ross RJ, Valcavi R. Effects of hypothyroidism and hyperthyroidism on insulin-like growth factors (IGFs) and growth hormone- and IGF-binding proteins. J Clin Endocrinol Metab 1993; 76:950-955.

205. De Gennaro Colonna V, Bertola G, Coco CB, Bifano M, Cocchi D, Maggi A, et al. Hypothalamic-pituitary somatotropic function in prepubertal hypothyroid rats: effect of growth hormone replacement therapy. Proc Soc Exp Biol Med 1991; 196:432-437.

206. Fisher DA, Hoath S, Lakshmanan J. The thyroid hormone effects on growth and development may be mediated by growth factors. [Review]. Endocrinol Exp 1982; 16:259-271.

207. Ismail IS, Scanlon MF. Endocrine and metabolic effects of growth hormone treatment. Hosp Update 1993; 19:443-448.

264 208. Jorgensen JO, Moller J, Laursen T, Orskov H, Christiansen JS, Weeke, et al. Growth hormone administration^timulates energy expenditure and extrathyroidal conversion of thyroxine to triiodothyronine in a dose-dependent manner and suppresses circadian thyrotrophin levels: studies in GH-deficient adults. Clin Endocrinol Oxf 1994; 41:609-614.

209. Isaksson OG, Lindahl A, Nilsson A, Isgaard J. Mechanism of the stimulatory effect of growth hormone on longitudinal bone growth. [Review]. Endocr Rev 1987; 8:426-438.

210. Hayden JM, Mohan S, Baylink DJ. The insulin-like growth factor system and the coupling of formation to resorption. [Review]. Bone 1995; 17:93s-98s.

211. Ljunghall S, Lindh E, Johansson AG. Endocrine interactions of insulin-like growth factor 1 on bone. [Review]. Acta Paediatr.Suppl. 1994; 399:178-179.

212. Spencer EM, Liu CC, Si EC, Howard GA. In vivo actions of insulin-like growth factor-I (IGF-1) on bone formation and resorption in rats. Bone 1991; 12:21-26.

213. Bang P, Brandt J, Degerblad M, Enberg G, Kaijser L, Thoren M, et al. Exercise-induced changes in insulin-like growth factors and their low molecular weight binding protein in healthy subjects and patients with growth hormone deficiency. Eur J Clin Invest 1990; 20:285-292.

214. Bang P, Eriksson U, Sara V, Wivall 1-L, Hall K. Comparison of acid ethanol extraction and acid gel filtration prior to IGF-1 and IGF-11 radioimmunoassays : Improvement of determinations in acid ethanol extracts by the use of truncated IGF-1 as radioligand. Acta Endocrinol 1991; 124:620-629.

215. James IT, Walne AJ, Perrett D. Measurement of pyridinium crosslinks.[Review]. Ann Clin Biochem 1996; 33:397-420.

216. Taaffe DR, Pruitt L, Reim J, Hintz RL, Butterfield G, Hoffinan AR, et al. Effect of recombinant human growth hormone on the muscle strength response to resistance exercise in elderly men. J Clin Endocrinol Metab 1994; 79:1361-1366.

265 217. Sandler LM, Burrin JM, Williams G, Joplin OF, Carr DH, Bloom SR. Effective long-term treatment of acromegaly with a long-acting somatostatin analogue (SMS 201-995). Clin Endocrinol Oxf 1987; 26:85-95.

218. White MC, James RA, Chattel]ee S, Mother M, Hall K, Dunne MJ, et al. Administration of high-dose octreotide (Sandostatin) by continuous subcutaneous infiision for the treatment of acromegaly. Horm Res 1990; 33 Suppl 1:13-8; discuss.

219. Zapf J, Walter H, Froesch ER. Radioimmunological determination of insulinlike growth factors I and H in normal subjects and in patients with growth disorders and extrapancreatic tumor hypoglycemia. J Clin Invest 1981; 68:1321-1330.

220. Hammerman MR. Insulin-like growth factors and aging. [Review]. Endocrinol Metab Clin North Am 1987; 16:995-1011.

221. Juul A, Bang P, Hertel NT, Main K, Dalgaard P, Jorgensen K, et al. Serum insulin-like growth factor-I in 1030 healthy children, adolescents, and adults: relation to age, sex, stage of puberty, testicular size, and body mass index. J Clin Endocrinol Metab 1994; 78:744-752.

222. Monson JP, Weaver JU, Noonan K, Burrin JM. Effect of growth hormone on serum thyroid hormone concentrations and desmopressin requirements in adults with hypopituitarism and central diabetes insipidus: a preliminary report. Endocrinol Metab 1994; 1:51-55.

223. Lippe BM, Van Herle AJ, LaFranchi SH, Uller RP, Lavin N, Kaplan SA. Reversible hypothyroidism in growth hormone-deficient children treated with human growth hormone. J Clin Endocrinol Metab 1975; 40:612-618.

224. Binnerts A, Swart OR, Wilson JHP, Hoogerbrugge N, Pols HA, Birkenhager JC, et al. The effect of growth hormone administration in growth hormone deficient adults on bone, protein, carbohydrate and hpid homeostasis, as well as on body composition. Clin Endocrinol Oxf 1992; 37:79-87.

225. Wuster C. Growth hormone and bone metabohsm. [Review]. Acta Endocrinol Copenh 1993; 128 Suppl 2:14-18.

266 226. Thoren M, Soop M, Degerblad M, Saaf M. Preliminary study of the effects of growth hormone substitution therapy on bone mineral density and serum osteocalcin levels in adults with growth hormone deficiency. Acta Endocrinol Copenh 1993; 128 Suppl 2:41-43.

227. Kaufinan JM, Taelman P, Vermeulen A, Vandeweghe M. Bone mineral status in growth hormone-deficient males with isolated and multiple pituitary deficiencies of childhood onset. J Clin Endocrinol Metab 1992; 74:118-123.

228. Vandeweghe M. Long-term consequences of growth hormone deficiency acquired in childhood. [Review]. Acta Endocrinol Copenh 1993; 128 Suppl 2:6-8.

229. Degerblad M, Elgindy N, Hall K, Sjoberg HE, Thoren M. Potent effect of recombinant growth hormone on bone mineral density and body composition in adults with panhypopituitarism. Acta Endocrinol Copenh 1992; 126:387-393.

230. Gambert SR, Schultz BM, Hamdy RC. Osteoporosis. [Review]. Endocrinol Metab Clin North Am 1995;24:317-371.

231. Collins KJ. The autonomic nervous system and the regulation of body temperature. In: Bannister R, Mathias CJ, editors. Autonomic failure: A textbook of clinical disorders of the autonomic nervous system. Third ed. Oxford: Oxford University Press, 1992:212-230.

232. MacKenzie MA, Schonbaum E, Hermus AR, Wollersheim HC, Thien T, Smals, et al. Sudomotor function in human poikilothermia. Neurology 1995; 45:1602-1607.

233. Appenzeller 0. Temperature regulation. In: Appenzeller O, editor. The autonomic nervous system - An introduction to basic and clinical concepts. 4th ed. Oxford: Elsevier, 1990:35-140.

234. Sato K, Kang WH, Saga K, Sato KT. Biology of sweat glands and their disorders. I. Normal sweat gland fimction. [Review]. J Am Acad Dermatol 1989; 20:537-563.

235. al-Tamer YY, Hadi EA. Age dependent reference intervals of glucose, urea, protein, lactate and electrolytes in thermally induced sweat. Eur J Clin Chem Clin Biochem 1994; 32:71-77.

267 236. Robertshaw D. Hyperhidrosis and the sympatho-adrenal system. Med Hypotheses 1979; 5:317-322.

237. Collins KJ. Sweat Glands: Eccrine and Apocrine. In: Greaves MW, Schuster S, editors. Handbook of Experimental Pharmacology, Vol. 87/1 Pharmacology of the Skin. 2nd ed. Berlin Heidelberg: Springer-Verlag, 1989:193-212.

238. Rechardt L, Waris T, Rintala A. Innervation of human axillary sweat glands. Histochemical and electron microscopic study of hyperhidrotic and normal subjects. Scand J Plast Reconstruct Hand Surg 1976; 10:107-112.

239. Tainio H, Vaalasti A. Electron microscopic study on the innervation of the human axillary sweat glands. Acta Histochem 1988; 83:167-171.

240. Ebling FJ. Apocrine glands in health and disorder. [Review]. Int J Dermatol 1989; 28:508-511.

241. Wells TR, Landing BH, Sandhu M, Lipsey AI. Microdissection study of the incidence of branched eccrine sweat glands and the number of eccrine glands per unit area of infants' and childrens' skin. Pediatr Pathol 1986; 6:301-307.

242. Sato K, Sato F. Individual variations in structure and function of human eccrine sweat gland. Am J Physiol 1983; 245:R203-R208.

243. Sato K, Nishiyama A, Kobayashi M. Mechanical properties and functions of the myoepithelium in the eccrine sweat gland. Am J Physiol 1979; 237:C177-C184.

244. Jones CJ, Kealey T. Electrophysiological and dye-coupling studies on secretory, myoepithehal and duct cells in human eccrine sweat glands. J Physiol Lond 1987; 389:461-481.

245. Wilson SM, Pediani JD, Cockbum F, Bovell DL, Jenkinson DM, Paton JY, et al. The cholinergic regulation of potassium (86Rb+) permeability in sweat glands isolated from patients with cystic fibrosis. Exp Physiol 1991; 76:573-578.

268 246. Sato K, Sato F. Defective beta adrenergic response of cystic fibrosis sweat glands in vivo and in vitro. J Clin Invest 1984; 73:1763-1771.

247. Pedersen PS. Human reabsorptive sweat duct in primary cell culture. Cellular mechanisms of ion transport and their regulation with special reference to cystic fibrosis. [Review]. Dan Med Bull 1993;40:208-223.

248. McDougall AJ, McLeod JG. Autonomic neuropathy, 1. Clinical features, investigation, pathophysiology, and treatment. J Neurol Sci 1996; 137:79-88.

249. Navarro X, Verdu E, Wendelscafer-Crabb G, Kennedy WR. Innervation of cutaneous structures in the mouse hind paw: a confocal microscopy immunohistochemical study. J Neurosci Res 1995;41:111-120.

250. Kennedy WR, Wendelschafer-Crabb G, Brelje TC. Innervation and vasculature of human sweat glands: an immunohistochemistry-laser scanning confocal fluorescence microscopy study. J Neurosci 1994; 14:6825-6833.

251. Dale HH, Feldberg W. The chemical transmission of secretory impulses to the sweat glands of the cat. J Physiol Lond 1934; 82:121-128.

252. Aavik OR. Cholinesterases in human skin. J Invest Dermatol 1955; 24:103-106.

253. Shelley WB, Hurley HJ. The physiology of the human axillary apocrine sweat gland. J Invest Dermatol 1953; 285-290.

254. Bumstock G. The changing face of autonomic neurotransmission. Acta Physiol Scand 1986; 126:67-91.

255. Leblanc G, Landis S. Development of choline acetyltransferase (CAT) in the sympathetic innervation of rat sweat glands. J Neurosci 1986; 6:260-265.

256. Sato K. Sweat induction from an isolated eccrine sweat gland. Am J Physiol 1973; 225:1147-1152.

269 257. Symes AJ, Rao MS, Lewis SE, Landis SC, Hyman SE, Fink JS. Ciliary neurotrophic factor coordinately activates transcription of neuropeptide genes in a neuroblastoma cell line. Proc Natl Acad Sci USA 1993; 90:572-576.

258. Grant MP, Landis SC. Developmental expression of muscarinic cholinergic receptors and coupling to phospholipase C in rat sweat glands are independent of iimervation. J Neurosci 1991 ; 11:3772-3782.

259. Ryer HI, Katz SE, Serby M. Muscarinic receptors on human eccrine sweat gland in aging and Alzheimer's disease. Biol Psychiat 1995; 37:259-264.

260. Grant MP, Landis SC, Siegel RE. The molecular and pharmacological properties of muscarinic cholinergic receptors expressed by rat sweat glands are unaltered by denervation. J Neurosci 1991; 11:3763-3771.

261. Tanaka E, Uchiyama S, Nakano S. Effects of calcitonin gene-related peptide and vasoactive intestinal peptide on nicotine-induced sweating in man. J Auton Nerv Syst 1990; 30:265-268.

262. Arunodaya GR, Taly AB. Sympathetic skin response: a decade later. [Review]. J Neurol Sci 1995; 129:81-89.

263. Wiesenfeld Hallin Z. Intrathecal vasoactive intestinal polypeptide modulates spinal reflex excitability primarily to cutaneous thermal stimuli in rats. Neurosci Lett 1987; 80:293-297.

264. Collins KJ, Weiner JS. Axon reflex sweating. Clin Sci Oxf 1961; 21 No.3:333-344.

265. Sato K, Kang WH, Saga K, Sato KT. Biology of sweat glands and their disorders. II. Disorders of sweat gland function [see comments]. [Review]. J Am Acad Dermatol 1989; 20:713-726.

266. Ryder RE, Marshall R, Johnson K, Ryder AP, Owens DR, Hayes TM. Acetylcholine sweatspot test for autonomic denervation. Lancet 1988; 1:1303-1305.

267. Robinson T, Fotherby M, Potter J. Changes in the sweatspot test with ageing and relation to cardiovascular autonomic function. Clin Auton Res 1995; 5:135-138.

270 268. Uno H. Sympathetic innervation of the sweat glands and piloarrector muscles of macaques and human beings. J Invest Dermatol 1977; 69; 112-120.

269. Sato K, Sato F. Role of calcium in cholinergic and adrenergic mechanisms of eccrine sweat secretion. Am J Physiol 1981; 241:C113-C120.

270. Sato K, Sato F. Cychc AMP accumulation in the beta adrenergic mechanism of eccrine sweat secretion. Pflugers Arch 1981; 390:49-53.

271. Sato K, Sato F. Cholinergic potentiation of isoproterenol-induced cAMP level in sweat gland. Am J Physiol 1983; 245:C189-C195.

272. Lundberg JM, Hokfelt T, Schultzberg M, Uvnas-Wallensten K, Kohler C. Occurrence of vasoactive intestinal polypeptide (VlP)-like immunoreactivity in certain cholinergic neurons of the cat: evidence from combined immunohistochemistry and acetylcholinesterase staining. Neuroscience 1979; 4:1539-1559.

273. Lundberg JM, Anggard A, Fahrenkrug J, Hokfelt T, Mutt V. Vasoactive intestinal polypeptide in cholinergic neurons of exocrine glands: functional significance of coexisting transmitters for vasodilation and secretion. Proc Natl Acad Sci USA 1980; 77:1651-1655.

274. Vaalasti A, Tainio H, Rechardt L. Vasoactive intestinal polypeptide (VlP)-like immunoreactivity in the nerves of human axillary sweat glands. J Invest Dermatol 1985; 85:246-248.

275. Landis SC, Fredieu JR. Coexistence of calcitonin gene-related peptide and vasoactive intestinal peptide in cholinergic sympathetic innervation of rat sweat glands. Brain Res 1986; 377:177-181.

276. Tainio H, Vaalasti A, Rechardt L. The distribution of substance P-, CGRP-, galanin- and ANP-like immunoreactive nerves in human sweat glands. Histochem J 1987; 19:375-380.

277. Dalsgaard CJ, Jembeck J, Stains W, Kjartansson J, Haegerstrand A, Hokfelt T, et al. Calcitonin gene-related peptide-like immunoreactivity in nerve fibers in the human skin. Relation

271 to fibers containing substance P-, somatostatin- and vasocactive intcstinalpolypcptidc-likc immunoreactivity. Histochemistry 1989; 91:35-38.

278. Hartschuh W, Weihe E, Reinecke M. Peptidergic (neurotensin, VIP, substance P) nerve fibres in the skin. Immunohistochemical evidence of an involvement of neuropeptides in nociception, pruritus and inflammation. Br J Dermatol 1983; 109 Suppl 25:14-17.

279. Johansson O. Evidence for PHl-immunoreactive nerve fibres in the human skin; coexistence with VIP? Med Biol 1986; 64:67-73.

280. Eedy DJ, Shaw C, Armstrong EP, Johnston CF, Buchanan KD. Vasoactive intestinal peptide (VIP) and peptide histidine methionine (PHM) in human eccrine sweat glands: demonstration of iimervation, specific binding sites and presence in secretions. Br J Dermatol 1990; 123:65-76.

281. Lindh B, Haegerstrand A, Lundberg JM, Hokfelt T, Fahrenkrug J, Cuello AC, et al. Substance P-, VIP- and CGRP-like immunoreactivities coexist in a population of cholinergic postganghonic sympathetic nerves innervating sweat glands in the cat. Acta Physiol Scand 1988; 134:569-570.

282. Kumazawa K, Sobue G, Mitsuma T, Ogawa T. Modulatory effects of calcitonin gene-related peptide and substance P on human cholinergic sweat secretion. Clin Auton Res 1994; 4:319-322.

283. Habecker BA, Landis SC. Noradrenergic regulation of cholinergic differentiation. Science 1994; 264:1602-1604.

284. Landis SC, Keefe D. Evidence for neurotransmitter plasticity in vivo: Developmental changes in properties of cholinergic sympathetic neurons. Dev Biol 1983; 98:349-372.

285. Stevens LM, Landis SC. Development and properties of the secretory response in rat sweat glands: relationship to the induction of cholinergic function in sweat gland innervation. Dev Biol 1987; 123:179-190.

286. Stevens LM, Landis SC. Developmental interactions between sweat glands and the sympathetic neurons which innervate them: effects of delayed iimervation on neurotransmitter plasticity and gland maturation. Dev Biol 1988; 130:703-720.

272 287. Braslis KG, Shulkes A, Fletcher DR, Hardy KJ. Pharmacokinetics and organ-specific metabolism of calcitonin gene-related peptide in sheep. J Endocrinol 1988; 118:25-31.

288. Grant MP, Francis NJ, Landis SC. The role of acetylcholine in regulating secretory responsiveness in rat sweat glands. Mol Cell Neurosci 1995; 6:32-42.

289. Habecker BA, Tresser SJ, Rao MS, Landis SC. Production of sweat gland cholinergic differentiation factor depends on innervation. Dev Biol 1995; 167:307-316.

290. Thoenen H. The changing scene of neurotrophic factors. [Review]. Trends Neurosci 1991; 14:165-170.

291. Korsching S. The neurotrophic factor concept: a reexamination. [Review]. J Neurosci 1993; 13:2739-2748.

292. Cowen T. Ageing in the autonomic nervous system: a result of nerve-target interactions? A review. [Review]. Mech Ageing Dev 1993; 68:163-173.

293. Hefti F. Neurotrophic factor therapy for nervous system degenerative diseases. [Review]. J Neurobiol 1994; 25:1418-1435.

294. Ip NY, Yancopoulos GD. Neurotrophic factors and their receptors. [Review]. Ann Neurol 1994; 35 Suppl:S13-6.

295. Tuszynski MH, Gage FH. Neurotrophic factors and diseases of the nervous system. [Review]. Ann Neurol 1994; 35 Suppl:S9-12.

296. Bothwell M. Functional interactions of neurotrophins and neurotrophin receptors. [Review]. Annu Rev Neurosci 1995; 18:223-253.

297. Main K, Nilsson KO, Skakkebaek NE. Influence of sex and growth hormone deficiency on sweating. Scand J Clin Lab Invest 1991; 51:475-480.

298. Terenghi G, Sundaresan M, Moscoso G, Polak JM. Neuropeptides and a neuronal marker in cutaneous innervation during human foetal development. J Comp Neurol 1993; 328:595-603.

273 299. Dobson RL. The human eccrine sweat gland. Structural and functional interrelationships. Arch Environ Health 1965; 11:423-429.

300. Robertshaw D. Neuroendocrine control of sweat glands. [Review]. J Invest Dermatol 1977; 69:121-129.

301. Johansson O, Vaalasti A. Immunohistochemical evidence for the presence of somatostatin-containing sensoiy nerve fibres in the human skin. Neurosci Lett 1987; 73:225-230.

302. Berg TJ, Levy DM, Reid G, Abraham RR. The effects of vasoactive intestinal polypeptide and substance P on methacholine-induced sweating and vascular flare in diabetic neuropathy. Clin Auton Res 1995; 5:159-164.

303. Fahrenkrug J. Transmitter role of vasoactive intestinal peptide. [Review]. Pharmacol Toxicol 1993; 72:354-363.

304. Said S. Vasoactive intestinal polypeptide (VIP): current status. [Review]. Peptides 1984; 5:143-150.

305. Luis J, Martin JM, el Battari A, Marvaldi J, Pichon J. The vasoactive intestinal peptide (VIP) receptor: recent data and hypothesis. [Review]. Biochimie 1988; 70:1311-1322.

306. Ganz P, Sandrock AW, Landis SC, Leopold J, Gimbrone MA, Jr., Alexander RW. Vasoactive intestinal peptide: vasodilatation and cyclic AMP generation. Am J Physiol 1986; 250:H755-H760.

307. Hartschuh W, Reinecke M, Weihe E, Yanaihara N. VlP-immunoreactivity in the skin of various mammals: immunohistochemical, radioimmunological and experimental evidence for a dual localization in cutaneous nerves and merkel cells. Peptides 1984; 5:239-245.

308. Heinz-Erian P, Sudhir P, Said S. Receptors for vasoactive intestinal peptide on isolated human sweat glands. Peptides 1986; 7:S151-S154.

309. Kummer W, Herbst WM, Heym C. Vasoactive intestinal polypeptide receptor-like immunoreactivity in human sweat glands. Neurosci Lett 1990; 110:239-243.

274 310. Kaji A, Shigematsu H, Fujita K, Maeda T, Watanabe S. Parasympathetic iimervation of cutaneous blood vessels by vasoactive intestinal polypeptide-immunoreactive and acetylcholinesterase-positive nerves: histochemical and experimental study on rat lower lip. Neuroscience 1988; 25:353-362.

311. Sugenoya J, Ogawa T, Jmai K, Ohnishi N, Natsume K. Cutaneous vasodilatation responses synchronize with sweat expulsions. Eur J Appl Physiol Occ Physiol 1995; 71:33-40.

312. Schwarzschild MA, Zigmond RE. Secretin and vasoactive intestinal peptide activate tyrosine hydroxylase in sympathetic nerve endings. J Neurosci 1989; 9:160-166.

313. Lundberg JM, Hokfelt T. Multiple co-existence of peptides and classical transmitters in peripheral autonomic and sensory neurones-functional and pharmacological implications. In: Hokfelt T, Fuke K, Pemow B, editors. Coexistence of neuronal messengers: a new principle in chemical transmission. Oxford: Elsevier, 1986:241-262.

314. Lundberg JM, Anggard A, Emson P, Fahrenkrug J, Hokfelt T. Vasoactive intestinal polypeptide and cholinergic mechanisms in cat nasal mucosa: studies on choline acetyltransferase and release of vasoactive intestinal polypeptide. Proc Natl Acad Sci USA 1981; 78:5255-5259.

315. Lundberg JM. Vasoactive intestinal polypeptide enhances muscarinic ligand binding in cat submandibular salivary gland. Nature 1982; 295:147-149.

316. Sato K, Sato F. Effect of VIP on sweat secretion and cAMP accumulation in isolated simian eccrine glands. Am J Physiol 1987; 253:R935-R941.

317. Tainio H. Cytochemical localization of VIP-stimulated adenylate cyclase activity in human sweat glands. Br J Dermatol 1987; 116:323-328.

318. Kruger L, Silverman JD, Mantyh PW, Stemini C, Brecha NC. Peripheral patterns of calcitonin-gene-related peptide general somatic sensory innervation: cutaneous and deep terminations. J Comp Neurol 1989; 280:291-302.

275 319. Lindh B, Lundberg JM, Hokfelt T. NPY-, galanin-, VIP/PHI-, CGRP- and substance P-immunoreactive neuronal subpopulations in cat autonomic and sensory ganglia and their projections. Cell Tissue Res 1989; 256:259-273.

320. Nyberg F, Le Greves P, Terenius L. Modulation of endopeptidase activity by calcitonin gene related peptide: a mechanism affecting substance P action? Biochimie 1988; 70:65-68.

321. Johansson O, Hilhges M, Wang L. Somatostatin-like immunoreactivity is found in dendritic guard cells of human sweat ducts. Peptides 1993; 14:401-403.

322. Ishida Yamamoto A, Tohyama M. Calcitonin gene-related peptide-like immunoreactive cells in the secretory portions of rat sweat glands. Cell Tissue Res 1990; 260:621-623.

323. Savage MV, Brengelmann GL, Buchan AM, Freund PR. Cystic fibrosis, vasoactive intestinal polypeptide, and active cutaneous vasodilation. J Appl Physiol 1990; 69:2149-2154.

324. Heinz-Erian P, Dey RD, Flux M, Said S. Deficient vasoactive intestinal peptide iimervation in the sweat glands of cystic fibrosis patients. Science 1985; 229:1407-1408.

325. Karanth SS, Springall DR, Lucas S, Levy D, Ashby P, Levene MM, et al. Changes in nerves and neuropeptides in skin from 100 leprosy patients investigated by immunocytochemistry. J Pathol 1989; 157:15-26.

326. Rao KS, Balakrishnan S, Oommen PK, Swamy MK, Durai V. Restoration of plantar sweat secretion in the feet of leprosy patients. Indian J Lep 1987; 59:442-449.

327. Terenghi G, Bunker CB, Liu YF, Springall DR, Cowen T, Dowd PM, et al. Image analysis quantification of peptide-immunoreactive nerves in the skin of patients with Raynaud's phenomenon and systemic sclerosis. J Pathol 1991; 164:245-252.

328. Anand P, Springall DR, Blank MA, Sellu D, Polak JM, Bloom SR. Neuropeptides in skin disease: increased VIP in eczema and psoriasis but not axillary hyperhidrosis. Br J Dermatol 1991; 124:547-549.

276 329. Haustein UF, Schliebs R, Schaller J. Changes in muscarinic acetylcholine receptor binding in skin slices of cholinergic urticaria. Acta Derm Venereol Stockh 1990; 70:208-211.

330. Kennedy WR, Sakuta M, Sutherland D, Goetz FC. The sweating deficiency in diabetes mellitus: methods of quantitation and clinical correlation. Neurology 1984; 34:758-763.

331. Faerman I, Faccio E, Calb I, Razumny J, Franco N, Dominguez A. Autonomic neuropathy in the skin: a histological study of the sympathetic nerve fibres in diabetic anhidrosis. Diabetologia 1982; 22:96-99.

332. Levy DM, Reid G, Rowley DA, Abraham RR. Quantitative measures of sympathetic skin response in diabetes: relation to sudomotor and neurological function. J Neurol Neurosurg Psychiat 1992; 55:902-908.

333. Lindberger M, Schroder HD, Schultzberg M, Kristensson K, Persson A, Ostman J, et al. Nerve fibre studies in skin biopsies in peripheral neuropathies. 1. Immunohistochemical analysis of neuropeptides in diabetes mellitus. J Neurol Sci 1989; 93:289-296.

334. Levy DM, Karanth SS, Springall DR, Polak JM. Depletion of cutaneous nerves and neuropeptides in diabetes mellitus: an immunocytochemical study. Diabetologia 1989; 32:427-433.

335. Pedersen SA, Welling K, Michaelsen KF, Jorgensen JO, Christiansen JS, Skakkebaek NE. Reduced sweating in adults with growth hormone deficiency. Lancet 1989; 2:681-682.

336. Juul A, Skakkebaek NE. Growth hormone deficiency and hyperthermia. Lancet 1991; 338:887

337. Main K, Price DA, Savage MO, Skakkebaek NE. Decreased sweating in seven patients with Laron syndrome. J Clin Endocrinol Metab 1993; 77:821-823.

338. Juul A, Behrenscheer A, Tims T, Nielsen B, Halkjaer-Kristensen J. Impaired thermoregulation in adults with growth hormone deficiency during heat exposure and exercise. Clin Endocrinol Oxf 1993; 38:237-244.

277 339. Juul A, Hjortskov N, Jepsen LT, Nielsen B, Halkj aer-Kristensen J, Vahl N, et al. Growth hormone deficiency and hyperthermia during exercise: A case controlled study of sixteen GH-deficient patients. J Clin Endocrinol Metab 1995; 80:3335-3340.

340. Thomer MO, Chait A, Aitken M, Benker G, Bloom SR, Mortimer CH, et al. Bromocriptine treatment of acromegaly. B M J 1975; 1:299-303.

341. Weeke J, Gundersen HJ. The effect of heating and central cooling on serum TSH, GH, and norepinephrine in resting normal man. Acta Physiol Scand 1983; 117:33-39.

342. Okada Y, Matsuoka T, Kumahara Y. Human growth hormone secretion during exposure to hot air in normal adult male subjects. J Clin Endocrinol Metab 1972; 34:759-763.

343. Buckler JM. The relationship between changes in plasma growth hormone levels and body temperature occurring with exercise in man. Biomedicine 1973; 19:193-197.

344. Kukkonen Haijula K, Kauppinen K. How the sauna affects the endocrine system. Ann Clin Res 1988; 20:262-266.

345. Galbo H, Houston ME, Christensen NJ, Holst JJ, Nielsen B, Nygaard E. The effect of water temperature on the hormonal response to prolonged swimming. Acta Physiol Scand 1979; 105:326-337.

346. Berg GR, Utiger RD, Schalch DS, Reichlin S. Effect of central cooling in man on pituitary-thyroid fimction and growth hormone secretion. J Appl Physiol 1966; 21:1791-1794.

347. Okada Y, Miyai K, Iwatsubo H, Kumahara Y. Human growth hormone secretion in normal adult subjects during and after exposure to cold. J Clin Endocrinol Metab 1970; 30:393-395.

348. Bauer M, Caroff S, Winokur A, Koenig R. Neuroendocrine responses to cold stress in normal subjects and depressives. Psychoneuroendocrinology 1987; 12:483-490.

349. Ruch W, Mixter RC, Russell RM, Garcia JF, Gale CC. Aminergic and thermoregulatory mechanims in hypothalamic regulation of growth hormone in cats. Am J Physiol 1977; 233:E61-E69.

278 350. Abramovici A, Josefsberg Z, Mimoimi M, Liban E, Laron Z. Histopathological features of the skin in hypopituitarism and Laron-type dwarfism. Isr J Med Sci 1983; 19:515-519.

351. Aloia JF, Grover RW. Dermal changes in osteoporosis following prolonged treatment with human growth hormone. J Cutan Pathol 1976; 3:222-231.

352. Jorgensen PH, Andreassen TT, Jorgensen KD. Growth hormone influences collagen deposition and mechanical strength of intact rat skin. A dose-response study. Acta Endocrinol Copenh 1989; 120:767-772.

353. Oxlund H, Jorgensen PH, Ortoft G, Andreassen TT. Alterations in the cross-links of skin collagen of rats treated with biosynthetic growth hormone. Connect Tissue Res 1991; 26:65-75.

354. Murphy LJ, Vrhovsek E, Lazarus L. Identification and characterization of specific growth hormone receptors in cultured human fibroblasts. J Clin Endocrinol Metab 1983; 57:1117-1124.

355. Granot I, Halevy 0, Hurwitz S, Pines M. Growth hormone and insulin-like growth factor I regulate collagen gene expression and extracellular collagen in cultures of avian skin fibroblasts. Mol Cell Endocrinol 1991; 80:1-9.

356. Cook JJ, Haynes KM, Werther GA. Mitogenic effects of growth hormone in cultured human fibroblasts. Evidence for action via local insulin-like growth factor I production. J Clin Invest 1988;81:206-212.

357. Bhora FY, Dunkin BJ, Batzri S, Aly HM, Bass BL, Sidawy AN, et al. Effect of growth factors on cell proliferation and epithelialization in human skin. J Surg Res 1995; 59:236-244.

358. Jensen UB, Jensen TG, Jensen PK, Rygaard J, Hansen BS, Fogh J, et al. Gene transfer into cultured human epidermis and its transplantation onto immunodeficient mice: an experimental model for somatic gene therapy. J Invest Dermatol 1994; 103:391-394.

359. Neely EK, Morhenn VB, Hintz RL, Wilson DM, Rosenfeld RG. Insulin-like growth factors are mitogenic for human kératinocytes and a squamous cell carcinoma. J Invest Dermatol 1991; 96:104-110.

279 360. Herndon DN, Barrow RE, Kunkel KR, Broemeling L, Rutan RL. Effects of recombinant human growth hormone on donor-site healing in severely burned children. Ann Surg 1990; 212:424-9; discuss.

361. Belcher HJ, Mercer D, Judkins KC, Shalaby S, Wise S, Marks V, et al. Biosynthetic human growth hormone in burned patients: a pilot study. Bums 1989; 15:99-107.

362. Thiboutot DM. Dermatological manifestations of endocrine disorders. J Clin Endocrinol Metab 1995; 80:3082-3087.

363. Feingold KR, Ehas PM. Endocrine-skin interactions. Cutaneous manifestations of pituitary disease, thyroid disease, calcium disorders, and diabetes. [Review]. J Am Acad Dermatol 1987; 17:921-940.

364. Burton JL, Libman LJ, Cunliffe WJ, Wilkinson R, Hall R, Shuster S. Sebum excretion in acromegaly. B M J 1972; 1:406-408.

365. Holt PJ, Marks R. Epidermal architecture, growth, and metabolism in acromegaly. B M J 1976; 1:496-497.

366. Matsuoka LY, Wortsman J, Kupchella CE, Eng A, Dietrich JE. Histochemical characterization of the cutaneous involvement of acromegaly. Arch Intern Med 1982; 142:1820-1823.

367. Thompson RJ, Doran JF, Jackson P, Dhillon AP, Rode J. PGP 9.5-a new marker for vertebrate neurons and neuroendocrine cells. Brain Res 1983; 278:224-228.

368. Wilkinson KD, Lee KM, Deshpande S, Duerksen-Hughes P, Boss JM, Pohl, et al. The neuron-specific protein PGP 9.5 is a ubiquitin carboxyl-terminal hydrolase. Science 1989; 246:670-673.

369. Webster HL. Laboratory diagnosis of cystic fibrosis. Clin Lab Sci 1983; 18:313-338.

370. Polak JM, Van Noorden SV. An introduction to immunocytochemistry: current techniques and problems. 1988; 2nd ed., 11, 1-58. Oxford: Oxford University Press.

280 371. Kamovsky MJ, Roots LA. A 'direct-colouring' thiocholine method for cholinesterases. J Histochem Cytochem 1964; 12:219-222.

372. Abdel-Rahman TA, Cowen I. Neurodegeneration in sweat glands and skin of aged rats. J Auton Nerv Syst 1994; 46:55-63.

373. Cowen I. Image analysis of FITC-immunofluorescence histochemistry in perivascular substance P-positive nerves. Histochemistry 1984; 81:609-610.

374. Cowen T, Haven AJ, Bumstock G. Pontamine sky blue: a counterstain for background autofluorescence in fluorescence and immunofluorescence histochemistry. Histochemistry 1985; 82:205-208.

375. Thrasivoulou C, Cowen T. Regulation of rat sympathetic nerve density by target tissues and NGF in maturity and old age. Eur J Neurosci 1995; 7:381-387.

376. Andrews T, Lincoln J, Milner P, Bumstock G, Cowen T. Differential regulation of tyrosine hydroxylase protein and activity in rabbit sympathetic neurones after long-term cold exposure: altered responses in ageing. Brain Res 1993; 624:69-74.

377. al-Shawaf AA, Kendall MD, Cowen T. Identification of neural profiles containing vasoactive intestinal polypeptide, acetylcholinesterase and catecholamines in the rat thymus. J Anat 1991; 174:131-143.

378. Cowen T, Thrasivoulou C, Shaw SA, Abdel-Rahman TA. Transplanted sweat glands from mature and aged donors determine cholinergic phenotype and altered density of host sympathetic nerves. J Auton Nerv Syst 1996; 58:153-162.

379. Abdel-Rahman TA, Collins KJ, Cowen T, Rustin M. Immunohistochemical, morphological and functional changes in the peripheral sudomotor neuro-effector system in elderly people. J Auton Nerv Syst 1992; 37:187-197.

380. Cowen T, Thrasivoulou C. A microscopical assay using a densitometric application of image analysis to quantify neurotransmitter dynamics. J Neurosci Methods 1992; 45:107-116.

281 381. Hjortskov N, Jepsen LT, Nielsen B, Juul A, Skakkebaek NE. Pilocarpine iontophoresis test; an index of physiological sweat secretion? Clin Physiol 1995; 15:409-414.

382. Juul A, Main K, Nielsen B, Skakkebaek NE. Decreased sweating in growth hormone deficiency: does it play a role in thermoregulation?. [Review]. J Pediatr Endocrinol 1993; 6:39-44.

383. Yamashita Y, Ogawa T, Ohnishi N, Imamura R, Sugenoya J. Local effect of vasoactive intestinal polypeptide on human sweat-gland function. Jpn J Physiol 1987; 37:929-936.

384. Rudman D, Feller AG, Nagraj HS, Gergans GA, Lalitha PY, Goldberg AF, et al. Effects of human growth hormone in men over 60 years old [see comments]. N Engl J Med 1990; 323:1-6.

385. Rudman D, Feller AG, Cohn L, Shetty KR, Rudman IW, Draper MW. Effects of human growth hormone on body composition in elderly men. [Review]. Horm Res 1991; 36 Suppl 1:73-81.

386. Gibson TE, Shelley WB. Sexual and racial differences in the response of sweat glands to acetylcholine and pilocarpine. J Invest Dermatol 1948; 137-142.

387. Lobeck CC, Huebner D. Effect of age, sex, and cystic fibrosis on the sodium and potassium content of human sweat. Pediatrics 1962; 172-173.

388. Kenney WL. A review of comparative responses of men and women to heat stress. [Review]. Environ Res 1985; 37:1-11.

389. De Berker D, Rees JL. Normal sweat secretion rate in patients with alopecia areata. Br J Dermatol 1995; 132:402-404.

390. Yousef MK, Dill DB, Vitez TS, Hillyard SD, Goldman AS. Thermoregulatory responses to desert heat: age, race and sex. J Gerontol 1984; 39:406-414.

391. Shikata N, Kurokawa I, Andachi H, Tsubura A. Expression of androgen receptors in skin appendage tumors: an immunohistochemical study. J Cutan Pathol 1995; 22:149-153.

282 392. Ruizeveld de Winter JA, Trapman J, Vermey M, Mulder E, Zegers ND, van der Kwast TH. Androgen receptor expression in human tissues: an immunohistochemical study. J Histochem Cytochem 1991; 39:927-936.

393. Fraser D, Padwick ML, Whitehead M, Coffer A, King RJ. Presence of an oestradiol receptor-related protein in the skin: changes during the normal menstrual cycle. Br J Obstet Gynaecol 1991; 98:1277-1282.

394. Choudhry R, Hodgins MB, van der Kwast TH, Brinkmann AO, Boersma WJ. Localization of androgen receptors in human skin by immunohistochemistry : implications for the hormonal regulation of hair growth, sebaceous glands and sweat glands. J Endocrinol 1992; 133:467-475.

395. Carpenter AJ, Nunneley SA. Endogenous hormones subtly alter women's response to heat stress. J Appl Physiol 1988; 65:2313-2317.

396. Quirion R, Araujo D, Nair NP, Chabot JG. Visualization of growth factor receptor sites in rat forebrain. Synapse 1988; 2:212-218.

397. Lobie PE, Breipohl W, Lincoln DT, Garcia-Aragon J, Waters MJ. Localization of the growth hormone receptor/binding protein in skin. J Endocrinol 1990; 126:467-471.

398. Oakes SR, Haynes KM, Waters MJ, Herington AC, Werther GA. Demonstration and localization of growth hormone receptor in human skin and skin fibroblasts. J Clin Endocrinol Metab 1992; 75:1368-1373.

399. Mertani HC, Delehaye-Zervas MC, Martini JF, Postel-Vinay MC, Morel G. Localization of growth hormone receptor messenger RNA in human tissues. Endocrine 1995; 3:135-142.

400. Tavakkol A, Elder JT, Griffiths CE, Cooper KD, Talwar H, Fisher GJ, et al. Expression of growth hormone receptor, insulin-like growth factor 1 (IGF-1) and IGF-1 receptor mRNA and proteins in human skin. J Invest Dermatol 1992; 99:343-349.

401. Batch JA, Mercuri FA, Edmondson SR, Werther GA. Localization of messenger ribonucleic acid for insulin-like growth factor-binding proteins in human skin by in situ hybridization. J Clin Endocrinol Metab 1994; 79:1444-1449.

283 402. Foster KG, Ellis FP, Dore C, Exton-Smith AN, Weiner JS. Sweat responses in the aged. Age & Ageing 1976; 5:91-101.

403. Collins KJ, Easton JC, Belfield-Smith H, Exton-Smith AN, Pluck RA. Effects of age on body temperature and blood pressure in cold environments. Clin Sci Oxf 1985; 69:465-470.

404. Collins KJ. Effects of cold on old people. [Review]. Br J Hosp Med 1987; 38:506-508.

405. Kenney WL, Fowler SR Methylcholine-activated eccrine sweat gland density and output as a function of age. J Appl Physiol 1988; 65:1082-1086.

406. Inoue Y, Nakao M, Araki T, Murakami H. Regional differences in the sweating responses of older and younger men. J Appl Physiol 1991; 71:2453-2459.

407. Ferrer T, Ramos MJ, Perez-Sales P, Perez-Jimenez A, Alvarez E. Sympathetic sudomotor function and aging. Muscle & Nerve 1995; 18:395-401.

408. Garwood M, Engel BT, Kusterer JP. Skin potential level: age and epidermal hydration effects. J Gerontol 1981; 36:7-13.

409. Montagna W, Carhsle K. Structural changes in ageing skin. [Review]. Br J Dermatol 1990; 122 Suppl 35:61-70.

410. Smith L. Histopathologic characteristics and ultrastructure of aging skin. Cutis 1989; 43:414-424.

411. Navarro X, Kennedy WR. Changes in sudomotor nerve territories with aging in the mouse. J Auton Nerv Syst 1990; 31:101-107.

412. Navarro X, Kamei H, Kennedy WR. Effect of age and maturation on sudomotor nerve regeneration in mice. Brain Res 1988; 447:133-140.

413. Navarro X, Kennedy WR. Effect of age on collateral reinnervation of sweat glands in the mouse. Brain Res 1988; 463:174-181.

284 414. Zadik Z, Chalew SA, McCarter RJ, Meistas M, Kowarski AA. The influence of age on the 24-hour integrated concentration of growth hormone in normal individuals. J Clin Endocrinol Metab 1985;60:513-516.

415. Hill DJ. What is the role of growth hormone and related peptides in implantation and the development of the embryo and fetus. [Review]. Horm Res 1992; 38 Suppl 1:28-34.

416. Borst SE, Millard WJ, Lowenthal DT. Growth hormone, exercise, and aging: the fiiture of therapy for the frail elderly. [Review]. J Am Geriatr Soc 1994; 42:528-535.

417. Xu X, Bennett SA, Ingram RL, Sonntag WE. Decreases in growth hormone receptor signal transduction contribute to the decline in insulin-like growth factor I gene expression with age. Endocrinology 1995; 136:4551-4557.

418. Toogood AA, O'Neil PA, Shalet SM. Beyond the somatopause: Growth hormone deficiency in adults over the age of 60 years. J Clin Endocrinol Metab 1996; 81:460-465.

419. Bothwell M. Tissue localization of nerve growth factor and nerve growth factor receptors. [Review]. Curr Topic Microbiol Immunol 1991; 165:55-70.

420. Levi-Montalcini R, Dal Toso R, della Valle F, Skaper SD, Leon A. Update of the NGF saga. J Neurol Sci 1995; 130:119-127.

421. Walicke PA. Novel neurotrophic factors, receptors, and oncogenes. [Review]. Annu Rev Neurosci 1989; 12:103-126.

422. Strand FL, Rose KJ, Zuccarelli LA, Kume J, Alves SE, Antonovich FJ, et al. Neuropeptide hormones as neurotrophic factors. Physiol Rev 1991; 71,4:1017-1046.

423. Bothwell M. Keeping track of neurotrophin receptors. [Review]. Cell 1991; 65:915-918.

424. Bothwell M. p75^: A receptor for all. Science 1996; 272:506-507.

425. Gavazzi I, Cowen T. Can the neurotrophic hypothesis explain degeneration and loss of plasticity in mature and ageing autonomic nerves? J Auton Nerv Syst 1996; 58:1-10.

285 426. Alberch J, Perez-Navarro E, Arenas E, Marsal J. Involvement of nerve growth factor and its receptor in the regulation of the cholinergic function in aged rats. J Neurochem 1991; 57:1483-1487.

427. Mearow KM, Kril Y, Diamond J. Increased NGF mRNA expression in denervated rat skin. Neuroreport 1993; 4:351-354.

428. Diamond J. Nerve-skin interactions in adult and aged animals. [Review]. Prog Clin Biol Res 1994; 390:21-44.

429. Vega JA, Vazquez E, Naves FJ, del Valle ME, Calzada B, Represa JJ. Immunohistochemical localization of the high-affmity NGF receptor (gpl40-trkA) in the adult human dorsal root and sympathetic ganglia and in the nerves and sensory corpuscles supplying digital skin. Anat Rec 1994; 240:579-588.

430. Landis SC, Fredieu JR, Yodlowski M. Neonatal treatment with nerve growth factor anti serum eliminates cholinergic sympathetic innervation of rat sweat glands. Dev Biol 1985; 112:222-229.

431. Rao MS, Landis SC. Characterization of a target-derived neuronal cholinergic differentiation factor. Neuron 1990; 5:899-910.

432. Lee KF, Bachman K, Landis S, Jaenisch R. Dependence on p75 for innervation of some sympathetic targets. Science 1994; 263:1447-1449.

433. Scharfinann R, AtoufF, Tazi A, Czemichow P. Growth hormone and prolactin regulate the expression of nerve growth factor receptors in lNS-1 cells [see comments]. Endocrinology 1994; 134:2321-2328.

434. Polak M, Scharfinann R, SeiUieimer B, Eisenbarth G, Dressier D, Verma IM, et al. Nerve growth factor induces neuron-like differentiation of an insulin-secreting pancreatic beta cell line. Proc Natl Acad Sci USA 1993; 90:5785

435. Aigner L, Caroni P. Absence of persistent spreading, branching, and adhesion in GAP-43-depleted growth cones. J Cell Biol 1995; 128:647-660.

286 436. Lewis ME, Neff NT, Contreras PC, Stong DB, Oppenheim RW, Grebow PE. Insulin-like growth factor-I: potential for treatment of motor neuronal disorders. [Review]. Exp Neurol 1993; 124:73-88.

437. Gavazzi I, Cowen T. Axonal regeneration from transplanted sympathetic ganglia is not impaired by age. Exp Neurol 1993; 122:57-64.

438. Spiteri Grech J, Bartlett JM, Nieschlag E. Regulation of testicular insulin-like growth factor-I in pubertal growth hormone-deficient male rats. J Endocrinol 1991; 131:279-285.

439. Nieves-Rivera F, Kerrigan JR, Krieg RJ, Jr., Egan J, Hwang LJ, TruumeesE., et al. Altered growth hormone (GH) secretion in vivo and in vitro in the diabetes-prone BB/Worcester rat. Growth Reg 1993; 3:235-244.

440. Young WG, Zhang CZ, Li H, Lobie PE, Waters MJ. A bromodeoxyuridine immunocytochemical and morphometric study of the influence of growth hormone on cell proliferation in odontogenic mesenchyme of the Lewis dwarf rat. Arch Oral Biol 1993; 38:207-214.

441. Johnson RJR, Santer RM, Cowen T. Neuron numbers in the myenteric plexus of aged rats: effects of dietary restriction. Society for Neuroscience Abstracts 1996; E-11:307.16 Abstract.

442. D'Costa AP, Ingram RL, Lenham JE, Sonntag WE. The regulation and mechanisms of action of growth hormone and insulin-like growth factor 1 during normal ageing. [Review]. J Reprod Fertil Supp 1993; 46:87-98.

443. Rao MS, Patterson PH, Landis SC. Multiple cholinergic differentiation factors are present in footpad extracts: comparison with known cholinergic factors. Development 1992; 116:731-744.

444. Rohrer H. Cholinergic neuronal differentiation factors: evidence for the presence of both CNTF-like and non-CNTF-like factors in developing rat footpad. Development 1992; 114:689-698.

287 445. Ure DR, Campenot RB, Acheson A. Cholinergic differentiation of rat sympathetic neurons in culture; effects of factors applied to distal neurites. Dev Biol 1992; 154:388-395.

446. Rao MS, Landis SC. Cell interactions that determine sympathetic neuron transmitter phenotype and the neurokines that mediate them. [Review]. J Neurobiol 1993; 24:215-232.

447. Berczi I, Nagy E, Asa SL, Kovacs K. The influence of pituitary hormones on adjuvant arthritis. Arth Rheum 1984; 27:682-688.

448. Neidhart M, Fluckiger EW. Hyperprolactinaemia in hypophysectomized or intact male rats and the development of adjuvant arthritis. Immunology 1992; 77:449-455.

449. Robertson JG, Pickering KJ, Belford DA. Insulin-like growth factor I (IGF-I) and IGF-binding proteins in rat wound fluid. Endocrinology 1996; 137:2774-2781.

450. Williams RM, Singh J, Sharkey KA. Innervation and mast cells of the rat exorbital lacrimal gland: the effects of age. J Auton Nerv Syst 1994; 47:95-108.

451. Kimura M, Masuda T, Hiwatashi N, Toyota T, Nagura H. Changes in neuropeptide-containing nerves in human colonic mucosa with inflammatory bowel disease. Pathology International 1994; 44:624-634.

452. PmceUi C, Fantini F, Romualdi P. Skin levels of vasoactive intestinal polypeptide in atopic dermatitis. Arch Dermatol Res 1991; 283:230-232.

453. Senapati A, Anand P, McGregor GP, Ghatei MA, Thompson RP, Bloom SR. Depletion of neuropeptides during wound healing in rat skin. Neurosci Lett 1986; 71:101-105.

454. Konttinen YT, Hukkanen M, Kemppinen P, Segerberg M, Sorsa T, Malmstrom M, et al. Peptide-containing nerves in labial salivary glands in Sjogren's syndrome. Arth Rheum 1992; 35:815-820.

455. Donnerer J, Schuligoi R, Stein C. Increased content and transport of substance P and calcitonin gene-related peptide in sensory nerves mnervatmg inflamed tissue: evidence for a regulatory function of nerve growth factor in vivo. Neuroscience 1992; 49:693-698.

288 456. Pincelli C, Fantmi F, Giaimetti A. Neuropeptides and skin inflammation. Dermatology 1993; 187:153-158.

457. Westerman RA, Carr RW, Delaney CA, Morris MJ, Roberts RG. The role of skin nociceptive afferent nerves in blister healing. Clin Exp Neurol 1993; 30:39-60.

458. Keith IM, Jin J, Saban R. Nerve-mast cell interaction in normal guinea pig urinary bladder. J Comp Neurol 1995; 363:28-36.

459. Hill CE, Jelinek H, Hendry LA, McLennan IS, Rush RA. Destruction by anti-NGF of autonomic, sudomotor neurones and subsequent hyperinnervation of the foot pad by sensory fibres. J Neurosci Res 1988; 19:474-482.

460. Gergely J, Dubowitz V. Disorders of voluntary muscle. 4th ed. Edinburgh: Churchill Livingstone, 1981.

461. Schiaffmo S, Reggiani C. Myosin isoforms in mammalian skeletal muscle. J Appl Physiol 1994; 77:493-501.

462. Hoh JFY, Fitzsimons RB. Myosin isoforms in normal and dystrophic human and murine species. In: Strohman RC, Wolf S, editors. Gene expression in muscle. New York: Plenum Press, 1984:43-62.

463. Landon DN. Skeletal muscle pathology. Edinburgh: Churchill Livingstone, 1982.

464. Delp MD, Duan C. Composition and size of type I, IIA, IID/X, and IIB fibres and citrate synthase activity of rat muscle. J Appl Physiol 1996; 80:261-270.

465. Pette D, Staron RS. Cellular and molecular diversities of mammalian skeletal muscle fibres. Rev Physiol Biochem Pharmacol 1990; 116:1-76.

466. Caiozzo VJ, Haddad F, Baker MJ, Baldwin KM. Influence of mechanical loading on myosin heavy-chain protein and mRNA isoform expression. J Appl Physiol 1996; 80:1503-1512.

289 467. Kraemer WJ, Patton JF, Gordon SE, Harman EA, Deschenes MR, Reynolds K, et al. Compatibility of high-intensity strength and endurance training on hormonal and skeletal muscle adaptations. J Appl Physiol 1995; 78:976-989.

468. Staron RS, Karapondo DL, Kraemer WJ, Fry AC, Gordon SE, Falkel JE, et al. Skeletal muscle adaptations during early phase of heavy-resistance training in men and women. J Appl Physiol 1994; 76:1247-1255.

469. Hortobagyi T, Hill JP, Houmard JA, Fraser DD, Lambert NJ, Israel RG. Adaptive responses to muscle lengthening and shortening in humans. J Appl Physiol 1996; 80:765-772.

470. Larsson L, Sjodin B, Karlsson J. Histochemical and biochemical changes in human skeletal muscle with age in sedentary males, age 22-65 years. Acta Physiol Scand 1978; 103:31-39.

471. Essen-Gustavsson B, Borges O. Histochemical and metabolic characteristics of human skeletal muscle in relation to age. Acta Physiol Scand 1986; 126:107-114.

472. Lexell J, Taylor CC, Sjostrom M. What is the cause of the ageing atrophy? Total number, size and proportion of different fibre types studied in whole vastus lateralis muscle from 15- to 83-years-old men. J Neurol Sci 1988; 84:275-294.

473. Glenmark B. Skeletal muscle fibre types, physical performance, physical activity and attitude to physical activity in women and men. A follow-up from age 16 to 27. Acta Physiol Scand Suppl 1994; 151:1-47.

474. Trappe SW, Costill DL, Fink WJ, Pearson DR. Skeletal muscle characteristics among distance runners: a 20-yr follow-up study. J Appl Physiol 1995; 78:823-829.

475. Devor ST, White TP. Myosin heavy chain of immature soleus muscle grafts adapts to hyperthyroidism more than to physical activity. J Appl Physiol 1996; 80:789-794.

476. Engel AG. Metabohc and endocrine myopathies. In: Walton J, editor. Disorders of Voluntary Muscle. Edinburgh: Churchill Livingstone, 1981:664-711.

477. Mastaglia FL, Barwich DD, Hall R. Myopathy in acromegaly. Lancet 1970; 2:907-909.

290 478. Mastaglia FL. Pathological changes in skeletal muscle in acromegaly. Acta Neuropathol 1973; 24:273-286.

479. Khaleeli AA, Levy RD, Edwards RH, McPhail G, Mills KR, Round JM, et al. The neuromuscular features of acromegaly: a clinical and pathological study. J Neurol Neurosurg Psychiat 1984; 47:1009-1015.

480. Nagulesparen M, Trickey R, Davies MJ, Jenkins JS. Muscle changes in acromegaly. B M J 1976; 2:914-915.

481. Ruff RL, Weissmann J. Endocrine myopathies. Neurol Clin 1988; 6:575-592.

482. Cuneo RC, Salomon F, Wiles CM, Hesp R, Sonksen PH. Growth hormone treatment in growth hormone-deficient adults. I. Effects on muscle mass and strength. J Appl Physiol 1991; 70:688-694.

483. Rutherford OM, Jones DA, Round JM, Preece MA. Changes in skeletal muscle after discontinuation of growth hormone treatment in young adults with hypopituitarism. Acta Paediatrica Scandinavica - Supplement 1989; 356:61-3; discussion 64.

484. Rutherford OM, Jones DA, Round JM, Buchanan CR, Preece MA. Changes in skeletal muscle and body composition after discontinuation of growth hormone treatment in growth hormone deficient young adults. Clin Endocrinol Oxf 1991; 34:469-475.

485. Whitehead HM, Gilliland JS, Allen IV, Hadden DR. Growth hormone treatment in adults with growth hormone deficiency: effect on muscle fibre size and proportions. Acta Paediatr Scand Suppl 1989; 356:65-67.

486. Cuneo RC, Salomon F, Wiles CM, Round JM, Jones D, Hesp R, et al. Histology of skeletal muscle in adults with GH deficiency: comparison with normal muscle and response to GH treatment. HormRes 1992; 37:23-28.

487. Carmeh E, Hochberg Z, Livne E, Lichtenstein I, Kestelboim C, Silbermann M, et al. Effect of growth hormone on gastrocnemius muscle of aged rats after immobilization: biochemistry and morphology. J Appl Physiol 1993; 75:1529-1535.

291 488. Linderman JK, Gosselink KL, Booth FW, Mukku VR, Grindeland RE. Resistance exercise and growth hormone as countermeasures for skeletal muscle atrophy in hindlimb-suspended rats. Am J Physiol 1994; 267:R365-R371.

489. Taaffe DR, Jin IH, Vu TH, Hoffman AR, Marcus R. Lack of effect of recombinant human growth hormone (GH) on muscle morphology and GH-Insulin-like growth factor expression in resistance-trained elderly men. J Clin Endocrinol Metab 1996; 81:421-425.

490. Wemerman J. Effect of growth hormone on muscle and protein in critically ill patients. [Review]. Acta Endocrinol Copenh 1993; 128 Suppl 2:19-22.

491. Fong Y, Rosenbaum M, Tracey KJ, Raman G, Hesse DG, Matthews DE, et al. Recombinant growth hormone enhances muscle myosin heavy-chain mRNA accumulation and amino acid accrual in humans. Proc Natl Acad Sci USA 1989; 86:3371-3374.

492. Fryburg DA. Insulin-like growth factor 1 exerts growth hormone- and insulin-like actions on human muscle protein metabolism. Am J Physiol 1994; 267:E331-6.

493. Eustache 1, Seyfritz N, Gueritaud JP. Effects of insulin-like growth factors on organotypic cocultures of embryonic rat brainstem slices and skeletal muscle fibers. Brain Res 1994; Developmental Brain:284-292.

494. Caroni P, Schneider C, Kiefer MC, Zapf J. Role of muscle insulin-like growth factors in nerve sprouting: suppression of terminal sprouting in paralyzed muscle by IGF-binding protein 4. J Cell Biol 1994; 125:893-902.

495. Baker LP, Peng HB. Induction of acetylcholine receptor cluster formation by local apphcation of growth factors in cultured Xenopus muscle cells. Neurosci Lett 1995; 185:135-138.

496. Ma J, Yang SX, Ho GJ, Festoff BW. Insulin-like growth factor binding protein-1 at mouse neuromuscular synapses. Synapse 1994; 17:225-229.

497. Ma J, Yang SX, Ho GJ, Festoff BW. Insulin-like growth factor binding protein-1 is pre-synaptic at mouse neuromuscular synapses and is transported in nerve. Neurochem Res 1994; 19:1363-1368.

292 498. Dardevet D, Somet C, Attaix D, Baracos VE, Grizard J. Insulin-like growth factor-1 and insulin resistance in skeletal muscles of adult and old rats. Endocrinology 1994; 134:1475-1484.

499. Goldspink DF, Cox VM, Smith SK, Eaves LA, Osbaldeston NJ, Lee DM, et al. Muscle growth in response to mechanical stimuli. Am J Physiol 1995; 268:E288-97.

500. Perrone CE, Fenwick-Smith D, Vandenburgh HH. Collagen and stretch modulate autocrine secretion of insulin-like growth factor-1 and insulin-like growth factor binding proteins from differentiated skeletal muscle cells. J Biol Chem 1995; 270:2099-2106.

501. Czerwinski SM, Martin JM, Bechtel PJ. Modulation of IGF mRNA abundance during stretch-induced skeletal muscle hypertrophy and regression. J Appl Physiol 1994; 76:2026-2030.

502. Yeh JK, Aloia JF, Chen M, Ling N, Koo HC, Millard WJ. Effect of growth hormone administration and treadmill exercise on serum and skeletal IGF-I in rats. Am J Physiol 1994; 266:E129-35.

503. Allen RE, Rankin LL. Regulation of satellite cells during skeletal muscle growth and development. [Review]. Proc Soc Exp Biol Med 1990; 194:81-86.

504. Caroni P, Schneider C. Signaling by insulin-like growth factors in paralyzed skeletal muscle: rapid induction of IGF 1 expression in muscle fibers and prevention of interstitial cell proliferation by IGF-BP5 and IGF-BP4. J Neurosci 1994; 14:3378-3388.

505. Turner JD, Rotwein P, Novakofski J, Bechtel PJ. Induction of mRNA for IGF-I and -II during growth hormone-stimulated muscle hypertrophy. Am J Physiol 1988; 255:E513-E517.

506. Jennische E, Skottner A, Hansson HA. Satellite cells express the trophic factor IGF-I in regenerating skeletal muscle. Acta Physiol Scand 1987; 129:9-15.

507. Lefaucheur JP, Sebille A. Muscle regeneration following injury can be modified in vivo by immune neutralization of basic fibroblast growth factor, transforming growth factor beta 1 or insulin-like growth factor I. J Neuroimmunol 1995; 57:85-91.

293 508. Isgaard J, Nilsson A, Vikman K, Isaksson OG. Growth hormone regulates the level of insulin-like growth factor-I mRNA in rat skeletal muscle. J Endocrinol 1989; 120:107-112.

509. Ullman M, Ullman A, Sommerland H, Skottner A, Oldfors A. Effects of growth hormone on muscle regeneration and IGF-I concentration in old rats. Acta Physiol Scand 1990; 140:521-525.

510. FelsingNE, Brasel JA, Cooper DM. Effect of low and high intensity exercise on circulating growth hormone in men. J Clin Endocrinol Metab 1992; 75:157-162.

511. Hellsten Y, Hansson HA, Johnson L, Frandsen U, Sjodin B. Increased expression of xanthine oxidase and insulin-like growth factor I (IGF-I) immunoreactivity in skeletal muscle after strenuous exercise in humans. Acta Physiol Scand 1996; 157:191-197.

512. Loughna PT, Bates PC. Interactions between growth hormone and nutrition in hypophysectomised rats: skeletal muscle myosin heavy chain mRNA levels. Biochem Biophys Res Commun 1994; 198:97-102.

513. Ayling CM, Moreland B, Schulster D. Effects of growth hormone on hypophysectomized rat skeletal muscles: increase in weight and creatine kinase activity. Biochem Soc Trans 1986; Vol 14 6I9th meeting:I0I3-I0I4.

514. Russell-Jones DL, Weissberger A, Bowes SB, Kelly JM, Thomason M, Jones RH, et al. The effects of growth hormone on protein metabohsm in adult growth hormone deficient patients. Clin Endocrinol Oxf 1993; 38:427-431.

515. Florini JR, Ewton DZ. Skeletal muscle fiber types and myosin ATPase activity do not change with age or growth hormone administration. J Gerontol 1989; 44:B110-Bl 17.

516. Palmer RM, Loveridge N, Thomson BM, Mackie SC, Tonner E, Flint DJ. Effects of a polyclonal antiserum to rat growth hormone on circulating insulin-like growth factor (IGF)-I and IGF-binding protein concentrations and the growth of muscle and bone. J Endocrinol 1994; 142:85-91.

294 517. Yang S, Alnaqeeb M, Simpson H, Goldspink G. Cloning and characterization of an IGF-1 isoform expressed in skeletal muscle subjected to stretch. J Muse Res Cell Motil 1996; 17:487-495.

518. Edwards RHT, Young A, Wiles CM. Needle biopsy of skeletal muscle in the diagnosis of myopathy and the clinical study of muscle fonction and repair. N Engl J Med 1980; 302:261-271.

519. Ecob-Prince M, Hill M, Brown W. Immunocytochemical demonstration of myosin heavy chain expression in human muscle. J Neurol Sci 1989; 91:71-78.

520. Ecob-Prince M, Hill M, Brown W. Myosin heavy chain expression in human muscle cocultured with mouse spinal cord. J Neurol Sci 1989; 90:167-177.

521. Franchi LL, Murdoch A, Brown WE, Mayne CN, Elliott L, Salmons S. Subcellular localization of newly incorporated myosin in rabbit fast skeletal muscle undergoing stimulation-induced type transformation. J Muse Res Cell Motil 1990; 11:227-239.

522. Augood SJ, McGowan EM, Finsen BR, Heppelman B, Emson PC. Non-radioactive in situ hybridisation using alkaline phosphatase-labelled oligonucleotides. In: Wisden W, Morris BJ, editors. In situ hybridization protocols for the brain. London: Academic Press, 1994:81-97.

523. The DIG system user's guide for filter hybridisation. 1994; 1st ed. Germany. Boehringer Mannheim GmbH.

524. Nonradioactive in situ hybridization - Application manual. 1992; 1st ed. Germany. Boehringer Mannheim GmbH.

525. Chan YY, Steiner RA, Clifton DK. Regulation of hypothalamic neuropeptide-Y neurons by growth hormone in the rat. Endocrinology 1996; 137:1319-1325.

526. Banker BQ, Engel AG. Basic reactions of muscle. In: Engel AG, editor. Myology. New York: McGraw-Hill, 1986:845-907.

295 527. Doerga ME, Keizer HA, Geurten PGH, van Kranenburg GPJ, Snel YEM, Zelissen PMJ, et al. Growth hormone deficiency may induce a decrease in type I fibres of the vastus lateralis muscle. Endocrinol Metab 1994; 1:P5 Abstract.

528. d'Albis A, Tobin C, Janmot C, Couteaux R. Effect of testosterone and thyroid hormone on the expression of myosin in the sexually dimorphic levator ani muscle of rat. J Biol Chem 1992; 267:10052-10054.

529. Izumo S, Nadal-Ginard B, Mahdavi V. All members of the MHC multigene family respond to thyroid hormone in a highly tissue-specific manner. Science 1986; 231:597-600.

530. Caiozzo VJ, Herrick RE, Baldwin KM. Influence of hyperthyroidism on maximal shortening velocity and myosin isoform distribution in skeletal muscles. Am J Physiol 1991; 261:C285-C295.

531. Glasscock GF, Hein AN, Miller JA, Hintz RL, Rosenfeld RG. Effects of continuous infusion of insulin-like growth factor 1 and II, alone and in combination with thyroxine or growth hormone, on the neonatal hypophysectomized rat. Endocrinology 1992; 130:203-210.

532. Downes M, Carozzi AJ, Muscat GE. Constitutive expression of the orphan receptor, Rev-erbA alpha, inhibits muscle differentiation and abrogates the expression of the myoD gene family. Mol Endocrinol 1995; 9:1666-1678.

533. Sachs L, de Luze A, Lebrun JJ, Kelly PA, Demeneix BA. Use of heterologous DNA-based gene transfer to follow physiological, T3-dependent regulation of myosin heavy chain genes in Xenopus tadpoles. Endocrinology 1996; 137:2191-2194.

534. Shorey CD, Everitt AV, Armstrong RA, Manning LA. Morphometric analysis of the muscle fibres of the soleus muscle of the ageing rat: long-term effect of hypophysectomy and food restriction. Gerontology 1993; 39:80-92.

535. Stickland NC, Crook AR, Sutton CM. Effects of pituitary dwarfism in the mouse on fast and slow skeletal muscles. Acta Anatomica 1994; 151:245-249.

296 536. Rutherford OM, Jones DA. The relationship of muscle and bone loss and activity levels with age in women. Age & Ageing 1992; 21:286-293.

537. Phillips SK, Rook KM, Siddle NC, Bruce SA, Woledge RC. Muscle weakness in women occurs at an earlier age than in men, but strength is preserved by hormone replacement therapy. Clin Sci Oxf 1993;84:95-98.

538. Capuano Pucci D, RheaultW, Rudman D. Relationship between plasma somatomedin C and muscle performance in a geriatric male population. Am J Phys Med 1987; 66:364-370.

539. Jennische E, Hansson HA. Regenerating skeletal muscle cells express insulin-like growth factor I. Acta Physiol Scand 1987; 130:327-332.

540. Montarras D, Pinset C, Perez MC, Dan J, Gros F. Muscle differentiation: insulin-like growth factors as positive modulators of myogenic regulatory genes? Comptes Rendus de 1 Académie des Sciences - Serielii, Sciences de la Vie 1993; 316:1025-1031.

541. Han VKM, D'Ercole AJ, Lund PK. Cellular localization of somatomedin (insulin-like growth factor) messenger RNA in the human fetus. Science 1987; 236:193-197.

542. Lindahl A, Isgaard J, Isaksson O. Growth and differentiation. [Review]. Baillieres Clin Endocrinol Metab 1991; 5:671-687.

543. Isaksson OG, Ohlsson C, Nilsson A, Isgaard J, Lindahl A. Regulation of cartilage growth by growth hormone and insulin-like growth factor I. [Review]. Pediatr Nephrol 1991; 5:451-453.

544. Gosteli-Peter MA, Winterhalter KH, Schmid C, Froesch ER, Zapf J. Expression and regulation of insulin-like growth factor-I (IGF-I) and IGF-binding protein messenger ribonucleic acid levels in tissues of hypophysectomized rats infused with IGF-I and growth hormone. Endocrinology 1994; 135:2558-2567.

545. Gilhar A, Ish-Shalom S, PiUar T, Etzioni A, Silbermann M. Effect of antiinsulin-like growth factor I on epidermal proliferation of human skin transplanted onto nude mice treated with growth hormone. Endocrinology 1994; 134:229-232.

297 546. Hodgkinson SC, Spencer GS, Bass JJ, Davis SR, Gluckman PD. Distribution of circulating insulin-like growth factor-I (IGF-I) into tissues. Endocrinology 1991; 129:2085-2093.

547. Edwall D, Schalling M, Jennische E, Norstedt G. Induction of insulin-like growth factor I messenger ribonucleic acid during regeneration of rat skeletal muscle. Endocrinology 1989; 124:820-825.

548. Coleman ME, DeMayo F, Yin KC, Lee HM, Geske R, Montgomery C, et al. Myogenic vector expression of insulin-like growth factor I stimulates muscle cell differentiation and myofiber hypertrophy in transgenic mice. J Biol Chem 1995; 270:12109-12116.

549. Jennische E, Andersson GL. Expression of GH receptor mRNA in regenerating skeletal muscle of normal and hypophysectomized rats. An in situ hybridization study. Acta Endocrinol Copenh 1991; 125:595-602.

550. Mertani HC, Morel G. In situ gene expression of growth hormone (GH) receptor and GH binding protein in adult male rat tissues. Mol Cell Endocrinol 1995; 109:47-61.

551. Koistinen H, Koistinen R, Selenius L, Ylikorkala 0, Seppala M. Effect of marathon run on serum IGF-I and IGF-binding protein 1 and 3 levels. J Appl Physiol 1996; 80:760-764.

552. Guler HP, Zapf J, Scheiwiller E, Froesch ER. Recombinant human insulin-like growth factor I stimulates growth and has distinct effects on organ size in hypophysectomized rats. Proc Natl Acad Sci USA 1988; 85:4889-4893.

553. Pell JM, Bates PC. Differential actions of growth hormone and insulin-like growth factor-I on tissue protein metabolism in dwarf mice. Endocrinology 1992; 130:1942-1950.

554. Borson S, Schatteman G, Claude P, Bothwell M. Neurotrophins in the developing and adult primate adenohypophysis: a new pituitary hormone system? Neuroendocrinology 1994; 59:466-476.

555. Patterson JC, Childs GV. Nerve growth factor and its receptor in the anterior pituitary. Endocrinology 1994; 135:1689-1696.

298 556. Patterson JC, Childs GV. Nerve growth factor in the anterior pituitary: regulation of secretion. Endocrinology 1994; 135:1697-1704.

557. Missale C, Boroni F, Frassine M, Caruso A, Spano P. Nerve growth factor promotes the differentiation of pituitary mammotroph cells in vitro. Endocrinology 1995; 136:1205-1213.

558. Burton KA, Kabigting EB, Clifton DK, Steiner RA. Growth hormone receptor messenger ribonucleic acid distribution in the adult male rat brain and its colocalization in hypothalamic somatostatin neurons. Endocrinology 1992; 131:958-963.

559. Rettori V, Milenkovic L, Riedel M, McCann SM. Physiological role of neuropeptide Y (NPY) in control of anterior pituitary hormone release in the rat. Endocrinol Exp 1990; 24:37-45.

560. Rettori V, Milenkovic L, Aguila MC, McCann SM. Physiologically significant effect of neuropeptide Y to suppress growth hormone release by stimulating somatostatin discharge. Endocrinology 1990; 126:2296-2301.

561. Werther GA, Haynes K, Waters MJ. Growth hormone (GH) receptors are expressed on human fetal mesenchymal tissues-identification of messenger ribonucleic acid and GH-binding protein. J Clin Endocrinol Metab 1993; 76:1638-1646.

562. Lobie PE, Garcia-Aragon J, Wang BS, Baumbach WR, Waters MJ. Cellular localization of the growth hormone binding protein in the rat. Endocrinology 1992; 130:3057-3065.

563. Edmondson SR, Werther GA, Russell A, Leroith D, Roberts CT, Jr., Beck, et al. Localization of growth hormone receptor/binding protein messenger ribonucleic acid (mRNA) during rat fetal development: relationship to insulin-like growth factor-I mRNA. Endocrinology 1995; 136:4602-4609.

564. Conover CA, Clarkson JT, Bale LK. Factors regulating insulin-like growth factor-binding protein-3 binding, processing, and potentiation of insulin-like growth factor action. Endocrine 1996; 137:2286-2292.

565. Schotzinger R, Yin X, Landis S. Target determination of neurotransmitter phenotype in sympathetic neurons. [Review]. J Neurobiol 1994; 25:620-639.

299 566. Patterson PH. Instructive neuronal differentiation factors. In; Loughlin SE, Fallon JH, editors. Neurotrophic factors. London: Academic Press, Inc. 1993:527-563.

567. Rajah R, Bhala A, Nunn SE, Peehl DM, Cohen P. 7S nerve growth factor is an insulin-like growth factor-binding protein protease. Endocrinology 1996; 137:2676-2682.

568. Purves D, Snider W, Voyvodic JT. Trophic regulation of nerve cell morphology and innervation in the autonomic nervous system. Nature 1988; 336:123-128.

569. Regan JM, Macfarlane DJ, Taylor NAS. An evaluation of the role of skin temperature during heat adaptation. Acta Physiol Scand 1997; 158:365-375.

570. Zachmann M. Interrelations between growth hormone and sex hormones: physiology and therapeutic consequences. [Review]. HormRes 1992; 38 Suppl 1:1-8.

571. Kunkeler L, Hodgins MB. Androgens and sweating [letter]. Br J Dermatol 1994; 131:739

572. Binoux M, Faivre-Bauman A, Lassarre C, Barret A, Tixier-Vidal A. Triiodothyronine stimulates the production of insuhn-like growth factor (IGF) by fetal hypothalamus cells cultured in serum-free medium. Brain Res 1985; 353:319-321.

573. Altman DG. Practical statistics for medical research. 2nd ed. London: Chapman and Hall, 1992.

574. Snyder DK, Clemmons D, Underwood LE. Treatment of obese, diet-restricted subjects with growth hormone for 11 weeks: effects on anabolism, lipolysis, and body composition. J Clin Endocrinol Metab 1988; 67:54-61.

575. Hasan NAE, Neumann MM, De Souky MA, So K-F, Bedi KS. The influence of predegenerated nerve grafts on axonal regeneration from prelesioned peripheral nerves. J Anat 1996; 189:293-302.

300 1',‘t

ROYAL FREE HOSPITAL HAMPSTEAD