Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. The role of the mammary fat pad during mammogenesis A thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Animal Science at Massey University Russell Charles Hovey 1996 ABSTRACT Development of the female mammary gland involves the proliferation and morphogenesis of epithelial cel ls within a matrix of adipose and connective tissue which constitutes the mammary fat pad. The objective of this research was to investigate the mechanisms by which this stromal environment locally regulates postnatal mammogenesis. Initial experiments showed that the mouse mammary fat pad liberates a diffusible activity in vitro which stimulates the growth of mouse mammary epithelial cells and enhances their proliferative response to insulin-like growth factor-I, epidermal growth factor and insulin. This effect was specific to these mitogens, and of a variety of cell lines tested was most pronounced for mouse mammary epithelial cells. Subsequent investigations indicated that these responses were likely induced by unsaturated fatty acids, particularly linoleic acid, from mammary adipocytes. Such responses may be effected by increased intracellular signalling via the actions of protein kinase C. The mitogenic capacity of the mouse mammary fat pad was also evaluated across several physiological states. Mammary fat pad-stimulated proliferation during the estrus cycle was increased at estrus concomitant with a phase of ductal elongation in vivo. In certain medium treatments there was evidence for epithelial upregulation of the mitogenic effect of the mammary fat pad, where intact mammary tissue was more stimulatory than mammary fat pad cleared of endogenous epithelium. Further experiments demonstrated that while the mitogenic effect of the mammary fat pad was unaltered by ovariectomy, ovarian function was required for this effect to be increased by exogenous progesterone. The effect of estrogen was independent of ovarian function but was altered by the local epithelial-stromal interaction, where it increased the mitogenic effect of epithelium-free mammary fat pad and decreased that of intact mammary tissue. Mitogenic stimulation by mammary tissues also declined during virginal development to be least in mature virgin and mid-pregnant states. Stimulation by intact mammary tissue increased during lactation, while that from epithelium-free mammary fat pad remained constant in the presence of steroid hormones and increased in the presence of growth factors. 11 Further experiments investigated the stromal regulation of epithelial growth within the ruminant mammary gland. Differences between the ruminant and rodent mammary fat pad were emphasised in vitro where ovine mammary fat pad stimulated the growth of mouse mammary epithelial cells but did not markedly potentiate their growth factor­ responsiveness. A subsequent study examined the expression of stroma-derived growth factors within the ruminant mammary gland during postnatal development, and their regulation by several physiological influences. The level of insulin-like growth factor (IGF)-I mRNA in the ovine mammary fat pad was elevated prior to puberty and during late gestation, while IGF-ll mRNA was upregulated in mammary parenchyma of virgin ewes in a transcript-specific manner. Abundance of IGF-I mRNA in mammary tissues of prepubertal ewe lambs tended to be increased by exogenous estrogen whereas IGF-ll mRNA levels were reduced. Messenger RNA for keratinocyte growth factor (KGF) was detected within the ovine mammary fat pad throughout development as 2.4 and 1 .5 kb mRNA transcripts which were expressed by stromal adipocytes and fibroblasts, respectively. The level of KGF mRNA in mammary tissues of prepubertal lambs was increased by ovariectomy and decreased by estrogen, while KGF mRNA expression in cultures of mammary fibroblasts was suppressed by dexamethasone. Messenger RNA for hepatocyte growth factor, a paracrine mitogen and morphogen for mammary epithelial cells, was expressed in the ovine mammary fat pad and by cultured mammary fibroblasts. The abundance of basic fibroblast growth factor (bFGF) mRNA was highest within the ovine mammary fat pad, while in vitro results suggest bFGF may be a paracrine/autocrine mitogen for multiple cel l types within the mammary gland. Basic FGF gene expression in mammary tissues of prepubertal ewes was reduced by estrogen treatment. For each of these growth factors there was evidence suggesting that their expreSSIOn within the mammary fat pad was upregulated by the adjacent mammary epithelium. In conclusion, these findings indicate that the mammary fat pad may stimulate the proliferation of mammary epithelial cells during postnatal mammogenesis by a variety of influences. Such mechanisms may involve the direct stimulation of epithelial growth or the modulation of epithelial responsiveness to other mitogens. These effects may function to mediate the actions of certain mammogenic hormones. Furthermore, strong iii evidence indicates that mammary growth may be locally regulated by the interaction between epithelial and stromal cells . IV ACKNOWLEDGEMENTS There is a large number of people who have assisted in the realisation of this thesis - to each of you I express my sincerest gratitude. To the members of my committee: Tom McFadden for introducing me to the area of mammary biology, for assisting with aspects of this research, and for his friendship. To Duncan Mackenzie for his ongoing, quality support and supervision, and to Stuart McCutcheon for his unselfish assistance at all times. Thanks to Steve Davis for initiating and supporting this research programme. A great deal of thanks are also extended to Helen Davey for her friendship and advice, for col laborating on aspects of this work, and a will ingness to talk science to the oddest of hours! The research presented herein has also been ably assisted by a number of people to whom I am particularly grateful: Megan Callaghan for her preparedness to lend a hand at every turn along with her smiles and cheer, Rob McLaren for advice and those jokes, and to Danielle Auldist, Steve Ellis, Stephen Eichler, Bid Clark, Rebecca Grieve, Christa Van Loon for graciously providing valuable assistance when it was needed. Provision of samples used in these experiments by Danielle and Steve is also acknowledged. The technical advice and assistance so generously provided by Alan Nixon, Christine Ford, Tom Wheeler, Colin Prosser, Angela Hodges, Adrian Molenaar and Jim Peterson are also greatly appreciated. Specific thanks to Glenda and Bobby Smith, Warwick Donaldson, and Cowley Harris who tended the animals used in these studies. Thanks also to Ric Broadhurst, John Parr and Denny Laboyrie for their expert assistance in such matters. I would also like to extend my gratitude to others for their valuable discussion on a variety of matters, people such as Dick Wilkins, Kerst Stelwagen, Vicki Farr, Mike Akers, Miriam Weber, Rita Lee, John Smith, Phil L'Huillier, Ian Garthwaite, and Pierre Lacasse. Often a small input makes a large contribution. There are numerous others who in one way or another have assisted my efforts, whether by their interest and support, or by their advice and assistance; to name but a few: Brett Whyte, Mahmoud Kiaei, Brydon Bennett, Gina Nicholas, Angela Beaton, Karyn Dupreez, Brett Langley, Raewyn Towers, Mark Thomas, Sally-Anne Turner, Pavla Misica, Dave Nation, Martin Auldist. Particular thanks to Brett Whyte and Danielle v Auldist for their assistance in compiling this thesis. I also thank all those around the Ruakura campus and the Dairying Research Corporation who, over a beer, a coffee, or while passing in the corridor, have afforded me their assistance and support. Thanks also to my friends and family in Australia for their ongoing interest and support. I would also like to thank those who provided me with cDNA probes or antibodies: Dr Eric Wong, Virginia Polytech; Dr Jeff Parrott, UCSF; Dr Judith Abraham, Scios Nova; Dr Jane MitcheIl , Moredun Research Institute; Dr Steven Wilson, The Cleveland Clinic Foundation, and the NIDDK Hormone and Pituitary Program. The provision of cell l ines by Drs Daniel Medina and Jeff Turner is also appreciated. This research was funded by the New Zealand Foundation for Research, Science, and Technology. Personal assistance was provided in the form of a Massey University postgraduate scholarship, and a Helen E. Akers Scholarship. TABLE OF CONTENTS ABSTRACT ACKNOWLEDGEMENTS LIST OF TABLES LIST OF FIGURES ABBRE VIATIONS CHA PTER 1 INTRODUCTION 1.1 On tog eny of mammary gland d evelopm en t 1 . 1 . 1 Embryonic development 1 . 1 .2 Prepubertal development 1 . 1 .3 Peripubertal and postpubertal development 1 . 1 .4 Oestrous cycle growth 1 . 1 .5 Gestational development 1 . 1 .6 Lactational growth 1.2 Mammary morphog en esis and his tog en esis 1 .2. 1 Cell heterogeneity 1 .2.2 The ductal end bud 1 .2.3 The mammary duct 1 .2.4 Alveolar and lobuloalveolar development 1.3 Th e s tromal en vironm en t 1.3 . 1 The embryonic mesenchyme 1 .3 .2 The mammary fat pad 1 .3 .3 Mammary transplantation studies 1 .3 .4 The epithelial-stromal reaction 1 .3 .5 Modell ing epithelial-stromal associations 1 .3 .6 The extracellular matrix 1 .4 Fac tors in fluencing mammary gland d evelopm en t 1 .4. 1 The ovarian steroids 1 .4. 1 . 1 Oestrogen 1 .4. 1 .2 Local oestrogen biosynthesis vi Pag e IV XV XVI XXI 1 2 2 4 5 6 7 1 0 1 1 1 1 1 2 1 3 1 4 1 6 1 6 1 6 1 8 22 24 25 27 28 28 30 1 .4. 1 .3 The oestrogen receptor (ER) 1 .4. 1 .4 Progesterone 1 .4. 1 .5 The progesterone receptor (PgR) 1 .4.2 The pituitary hormones 1 .4.2. 1 Prolactin (Prl) 1 .4.2.2 The prolactin receptor 1 .4.2.3 Growth hormone (GH) 1 .4.2.4 The growth hormone receptor 1 .4.3 The placental lactogens (PLs) 1 .4.4 Plane of nutrition 1 .4.4. 1 Rodents 1 .4.4.2 Ruminants 1 .4.5 Dietary fat 1 .4.5 . 1 In vivo studies 1 .4.5.2 Mechanisms of dietary fat influence 1 .4.6 Peptide growth factors 1 .4.6. 1 The insulin-like growth factors (IGFs) 1 .4.6. 1 . 1 IGF-I 1 .4.6. 1 .2 The IGF-I receptor 1 .4.6. 1 .3 IGF-II 1 .4.6. 1 .4 The IGF-II receptor 1 .4.6. 1 .5 The IGF binding proteins (IGF-BPs) 1 .4.6.2 Epidermal growth factor (EGF)-like polypeptides 1 .4.6.2. 1 In vitro studies 1 .4.6.2.2 In vivo studies 1 .4.6.2.3 The EGF receptor (EGFR) 1 .4.6.3 Hepatocyte growth factor (HGF) 1 .4.6.4 The fibroblast growth factors (FGFs) 1 .4.6.4. 1 acidic fibroblast growth factor (aFGF) 1 .4.6.4.2 basic fibroblast growth factor (bFGF) 1 .4.6.4.3 Keratinocyte growth factor (KGF) 1.5 Purpos e and scop e of th e in ves tiga tion Vll 3 1 32 34 36 36 38 40 42 42 44 44 45 47 47 48 50 50 50 52 53 54 54 55 56 57 59 6 1 63 63 64 66 68 CHAPTER 2 2.1 Abs trac t 2.2 In trod uc tion MAMMAR Y FAT PROLIFERATI VE PAD MODULATES THE RESPONSE OF MOUSE MAMMAR Y EPITHELIAL CELLS TO SPECIFIC MITOGENS IN VITRO 2.3 Ma terials and M ethods 2.3 . 1 Reagents and chemicals 2.3.2 Cell culture 2.3 .3 Co-cultures 2.3 .4 Statistical analyses 2.4 R es ul ts 2.5 Disc ussion CHAPTER 3 MOUSE MAMMAR Y FAT PAD SPECIFICALL Y REGULATES GROWTH FACTOR -INDUCED PROLIFERATION OF MAMMAR Y EPITHELIAL viii 70 7 1 7 1 73 73 73 74 75 75 84 CELLS IN VITRO 88 3.1 Abs trac t 89 3.2 In trod uc tion 89 3.3 Ma terials a nd M ethods 9 1 3 .3 . 1 Materials 9 1 3 .3 .2 Cel l culture 92 3 .3 .3 Co-cultures and conditioned medium 92 3.3.4 Ligand binding assays 93 3 .3 .5 Statistical analyses 93 3.4 R es ul ts 93 3 .4. 1 Co-cultured mammary fat pad potentiates mitogenic stimulation 93 3 .4.2 In vitro effects of the mammary fat pad are due to a diffusible �� � 3 .4.3 Mitogenic potentiation by the mammary fat pad is growth factor -specific 97 3.4.4 Effect of mammary fat pad on IGF-I and EGF receptors 3 .4.5 Responses to mammary fat pad are cell type-specific 3.5 Discussion CHAPTER 4 4.1 Abstract 4.2 Introduction ROLE FOR MAMMAR Y FAT PAD-DERI VED UNSATURATED FATT Y ACIDS IN POTENTIATING GROWTH FACTOR -STIMULATED GROWTH O F MAMMAR Y EPITHELIAL CELLS IN VITRO 4.3 Materials and Methods 4.3 . 1 Cell cultures 4.3.2 Conditioned medium 4.3.3 Conditioned medium analyses 4.3 .4 Cell growth experiments 4.3.5 Western blotting 4.3.6 Fatty acid analyses 4.3.7 Statistical analyses 4.4 Results 4.4. 1 Analysis of the mitogenic activity in CM 4.4.2 Fatty acid effects on COMMA- ID cell growth 4.4.3 Effects of CM on intracellular signalling 4.5 Discussion CHAPTER S MODULATION O F STEROID- AND GROWTH FACTOR-STIMULATED GROWTH O F MAMMAR Y EPITHELIAL CELLS BY THE MAMMAR Y FAT PAD AND EPITHELIAL -STROMAL IX 99 99 102 105 106 106 1 08 108 1 08 1 08 109 109 109 1 1 0 1 1 0 1 10 1 12 1 1 5 1 1 8 INTERACTIONS DURING THE OESTROUS C YCLE 1 22 5.1 Abstract 5.2 Introduction 5.3 Materials and Methods 1 23 1 24 1 25 5 .3 . 1 Animals 5 .3 .2 Cell cultures 5 .3 .3 Assessment of mammary development 5 .3 .4 Oestrogen receptor (ER) immunoblotting 5 .3 .5 Statistical analyses 5.4 Results 5 .4. 1 Mammary gland development 5 .4.2 Co-cultures 5 .4.3 CM modulation of steroid action 5.5 Discussion CHAPTER 6 IN FLUENCE O F ONTOGENIC STATE , O VARIECTOM Y, AND O VARIAN STEROID HORMONES ON THE IN VITRO PROLI FERATI VE x 1 25 1 26 1 27 1 27 1 27 1 28 1 28 1 28 1 35 1 36 E FFECT O F THE MAMMAR Y FAT PAD 1 4 1 6.1 Abstract 6.2 Introduction 6.3 Materials and Methods 6.3 . 1 Animals 6.3.2 Co-cultures 6.3.3 Mammary gland whole mounts 6.3 .4 Statistical analyses 6.4 Results 6.4. 1 Effect of ontogenic state 6.4. 1 . 1 Mammary gland development 6.4. 1 .2 Co-cultures 6.4.2 Effects of ovariectomy and ovarian steroids 6.4.2. 1 Mammary development 6.4.2.2 Co-cultures 6.5 Discussion 1 42 143 145 145 146 146 146 147 147 1 47 149 1 54 1 54 1 56 1 59 CHAPTER 7 7.1 Abstract 7.2 Introduction DI FFERENTIAL MODULATOR Y E FFECTS O F MURINE AND O VINE MAMMAR Y FAT PAD ON THE PROLI FERATION O F MAMMAR Y EPITHELIAL CELLS IN VITRO 7.3 Materials and Methods 7.3 . 1 Cell cultures 7.3.2 Animals and tissues 7 .3 .3 Co-cultures and conditioned medium 7.3 .4 Statistical analyses 7.4 Results 7.5 Discussion CHAPTER 8 PREPARATION O F A MAMMAR Y FAT PAD PARENCH YMA ·FREE AND SUBSE QUENT Xl 1 65 1 66 1 67 1 68 1 68 1 69 1 69 1 70 1 7 1 1 82 MAMMAR Y GLAND DE VELOPMENT IN S HEEP 1 86 8.1 Abstract 1 87 8.2 Introduction 1 88 8.3 Materials and Methods 1 89 8 .3 . 1 Animals 1 89 8 .3 .2 Surgical procedure 1 89 8 .3 .3 Experimental design 1 90 8 .3 .4 Mammary sampling 1 9 1 8 .3 .5 Hormonal induction of lactation and milking 1 9 1 8 .3 .6 Chemical analyses 1 9 1 8 .3.7 Histology and whole mounts 1 92 8 .3 .8 Statistical analyses 1 92 8.4 Results 1 92 8 .4. 1 Mammary CFP technique 1 92 8 .4.2 Mammary gland development 1 97 8.5 Discussion 200 CHAPTER 9 9.1 Abstract 9.2 Introduction DE VELOPMENTAL E XPRESSION SUGGESTS ROLES FOR LOCALL Y-DERI VED INSULIN-LIKE GROWTH FACTORS-I AND -11 DURING O VINE MAMMOGENESIS 9.3 Materials and Methods 9.3 . 1 Animals and tissues 9.3.2 Probes and labelling 9.3 .3 Northern analysis 9.3 .4 Statistical analyses 9.4 Results 9.S Discussion CHAPTER 10 MULTIPLE FACTORS REGULATE THE PARACRINE E XPRESSION O F KERATINOC YTE GROWTH FACTOR (KG F) IN THE DE VELOPING xii 206 207 207 209 209 209 209 2 10 2 10 2 1 5 RUMINANT MAMMAR Y GLAND 2 1 9 10.1 Abstract 220 10.2 Introduction 220 10.3 Materials and Methods 222 1 0.3 . 1 Cell culture 222 1 0.3 .2 Proliferation studies 223 1 0.3 .3 Animals and tissues 223 1 0.3 .4 Probes and labelling 224 1 0.3.5 RNA isolation and Northern analysis 224 1 0.3.6 Statistical analyses 225 10.4 Results 226 1 0.4. 1 KGF is specifically mitogenic for mammary epithelium 226 1 0.4.2 Mammary stroma expresses KGF mRNA in vitro 226 10.4.3 The KGF gene is differentially transcribed in the ovine mammary �md 2� 1 0.4.4 Mammary KGF expression is physiologically regulated 10.5 Discussion CHAPTER 1 1 E XPRESSION AND ROLE O F PARACRINE GROWTH FACTORS DURING DE VELOPMENT O F xiii 23 1 233 THE RUMINANT MAMMAR Y GLAND 238 11.1 Abstract 239 1 1.2 Introduction 240 1 1.3 Materials and Methods 241 1 1 .3 . 1 Tissues 241 1 1 .3 .2 Cell cultures 242 1 1 .3 .3 Probes and Northern analysis 243 1 1 .3 .4 Conditioned medium 243 1 1 .3 .5 Fractionation of CM 243 1 1 .3 .6 Statistical analyses 244 11.4 Results 244 1 1 .4. 1 Expression of growth factor mRNA in the ovine mammary gland 244 1 1 .4.2 In vitro expression of growth factor mRNA by mammary cells 249 1 1 .4.3 Growth stimulation by mammary fibroblasts 249 11.5 Discussion 255 CHAPTER 12 GENERAL DISCUSSION AND CONCLUSIONS 12.1 General Discussion 1 2. 1 . 1 Mitogenic effects of the mammary fat pad 1 2. 1 .2 Influence of epithelial-stromal interactions on the mitogenic capacity of the mammary fat pad 1 2. 1 .3 The hormonal regulation of mitogenic stimulation by the 259 260 260 265 mammary fat pad 267 1 2. 1 .4 Effect of ontogenic state on the mitogenic capacity of the mammary fat pad 270 1 2. 1 .5 A combined model for fat pad stimulation of mammogenesis 272 12.2 Conclusions 274 xiv APPENDICES 275 Appendix 1 Time course of DNA synthesis by COMMA- I D cells Appendix 2 Appendix 3 Appendix 4 Appendix 5 Appendix 6 Appendix 7 Appendix 8 Appendix 9 Appendix 1 0 Appendix 1 1 Appendix 1 2 Appendix 1 3 RE FERENCES cultured in hormone-free BM. Growth of COMMA- I D cells in response to medium conditioned with various amounts of mouse mammary fat pad tissue. IGF-I l igand binding assay parameters. EGF ligand binding assay parameters. Growth of COMMA- ID cells in the presence of co-cultured mouse mammary fat pad and various combinations of IGF-I and insulin. DNA synthesis by COMMA- I D cells in response to acidic- and basic FGF after 24 h culture. Effect of trypsin on the mitogenic effect of IGF-I + EGF. Effect of treating CM with heparin sepharose. Growth of COMMA- ID cel ls in medium supplemented with IGF-I + EGF and various concentrations of BSA, or in the presence of co-cultured mammary fat pad. Growth of COMMA- ID cells in the presence of co-cultured mammary or ovarian fat pad tissue. Growth of COMMA- ID cells in various concentrations of CM stored for 6 months at -80°C. Northern analysis of KGF mRNA expression in various cell lines. List of publications. 276 277 278 279 280 28 1 282 283 284 285 286 287 288 289 xv LIST OF TABLES Table Page 3.1 Growth of various cell lines in response to mitogens in the absence and presence of co-cultured mammary fat pad. 1 0 1 4.1 Effects of treatments applied to CM on the subsequent growth of COMMA- I D cells. 1 1 1 4.2 Concentrations of major fatty acids in CM. 1 1 2 4.3 Effect of unsaturated fatty acids and prostaglandin E2 on the growth of COMMA- I D cells. 1 1 3 5.1 Assessment of mammary gland development in virgin female BALB/C mice during the oestrous cycle. 1 3 1 5.2 Least Squares mean COMMA- I D growth responses to co-cultured mammary tissue in medium supplemented with ovarian steroid hormones. 6.1 Mean body and mammary tissue weights of sham-operated and ovariectomised female mice treated for 5 days with saline, progesterone, 1 33 oestrogen, or progesterone + oestrogen. 1 56 6.2 Least squares means of COMMA- I D cell growth responses to co- cultured CFP and MFP tissue from sham-operated or ovariectomised female mice treated for 5 days with saline, progesterone, oestrogen, or progesterone + oestrogen. 1 58 6.3 Change in the weight of co-cultured CFP and MFP explants from sham­ operated and ovariectomised female mice treated for 5 days with saline, progesterone, oestrogen, or progesterone + oestrogen. 1 59 8.1 Lactogenic and mammogenic responses in mammary glands of surgically modified ewes. 201 9.1 Effects of ovariectomy and oestrogen on IGF-I and -IT mRNA levels in prepubertal ovine mammary tissues. 2 1 5 xvi LIST OF FIGURES Figure Page 1.1 Composite drawing of the ductal terminal end bud made from light and 1.2 2.1 2.2 electron microscope analyses. Schematic representation of breast development. Effect of medium changing routine on the growth response of COMMA- I D cells to co-cultured mammary fat pad. Growth of COMMA- I D cells in response to co-cultured mammary fat pad and FCS. 2.3 Effect of BSA and co-cultured mammary fat pad on COMMA- I D cell proliferation. 2.4 Effect of PGE2 and indomethacin on the growth of COMMA- I D cells . 2.5 Response of COMMA- I D cel ls to mammogenic hormones and co­ cultured mammary fat pad. 2.6 2.7 2.8 Effect of rbGH on COMMA- I D cel l growth in the absence or presence of co-cultured mammary fat pad. Response of COMMA- I D cells to growth factors and co-cultured mammary fat pad. Phase contrast micrographs of COMMA- I D cells after 5 days culture in various treatments. 3.1 Concentration-response curves for rnitogens added to cultures of COMMA- I D cells in the absence or presence of co-cultured mammary 1 3 1 5 76 77 78 79 80 8 1 82 83 �p�. � 3.2 Response of COMMA- I D cells to increasing concentrations of CM supplemented with mitogens. 3.3 Growth of COMMA- I D cells in response to growth factors in CM or in co-cultures. 3.4 Specific binding of IGF-I and EGF to COMMA- I D cells cultured in BM or CM supplemented with various treatments. 96 98 1 00 4.1 Growth of COMMA- I D cells in response to various concentrations of CM and linoleic acid-BSA added to cultures either alone or in the xvii presence of IGF-I + EGF. 1 14 4.2 Comparison of the mitogenic effects of CM and linoleic acid-BSA alone and in the presence of IGF-I and/or EGF. 4.3 Western blotting of protein tyrosine phosphorylation in COMMA- I D cells fol lowing their incubation for 30 mins or 6 h in various medium treatments. 4.4 Effect of various concentrations of the PKC inhibitor, OHMG, on the growth of COMMA- ID cells in BM or CM either alone or in the presence of IGF-I + EGF. 5.1 Mammary gland morphology in virgin female BALB/c mice during the oestrous cycle. 5.2 Histomorphology of mammary parenchyma at oestrus and dioestrus. 5.3 In vitro mitogenic stimulation by virgin mouse CFP and MFP during the oestrous cycle in hormone-free BM, or BM supplemented with 1 7P­ oestradiol, progesterone, or 1 7p-oestradiol + progesterone. 5.4 In vitro mitogenic stimulation by virgin mouse CFP and MFP during the oestrous cycle in BM supplemented with IGF-I, EGF, or IGF-I + 1 1 5 1 1 6 1 17 1 29 1 30 1 32 EGF. 1 34 5.5 In vitro proliferation of COMMA- I D cells in BM or CM supplemented with ovarian steroid hormones. 5.6 Western blotting of ER proteins in COMMA- I D cells after their culture in various medium treatments. 6.1 Ontogeny of postnatal mammary gland development and 1 35 1 36 morphogenesis in female BALB/c mice. 1 48 6.2 Body and mammary tissue weights of female BALB/c Inlce during postnatal development. 6 .3 Ontogeny of mitogenic stimulation by mouse CFP and MFP in hormone 149 free BM, or BM supplemented with IGF-I, EGF, or IGF-I + EGF. 1 50 6.4 Ontogeny of mitogenic stimulation by mouse CFP and MFP in BM supplemented with 1 7�-oestradiol , progesterone, or 1 7�-oestradiol + XVlll progesterone. 1 52 6.5 Change in the weight of co-cultured MFP and CFP explants from mice at various stages of development. 6.6 Mammary gland development in sham-operated and ovariectomised female mice treated for 5 days with excipient, 1 7�-oestradiol, progesterone, or 1 7�-oestradiol + progesterone. 6.7 Growth of COMMA- I D cells in response to various medium supplements either alone, or when co-cultured with explants of CFP or 1 53 1 55 MFP. 1 57 7.1 Growth response of COMMA- I D cells to co-cultured ovine FP and 1 0% FCS. 1 7 1 7.2 Growth of COMMA- I D cells in response to co-cultured ovine FP, IGF- I, and EGF. 1 72 7.3 Response of COMMA- I D cells to a range of concentrations of CM prepared using ovine FP, both alone and in the presence of various mitogenic supplements. 1 73 7.4 Effect of co-cultured ovine FP, IGF-I and rbGH on the growth of COMMA- I D mouse mammary epithelial cells. 1 74 7.5 Growth of COMMA- ID and MAC-T cells in response to co-cultured mammary tissues and 10% FCS. 1 75 7.6 Effect of IGF-I, EGF, and co-cultured mammary tissues on the growth of COMMA- I D and MAC-T cells. 1 77 7.7 Proliferative response of COMMA- I D and MAC-T cells to aFGF, bFGF, and co-cultured mammary tissues. 1 78 7.8 Growth of COMMA- I D and MAC-T cells in response to mammogenic hormones and co-cultured mammary tissues. 1 80 7.9 Effect of ovarian steroids and co-cultured mammary tissues on the growth of COMMA- I D and MAC-T cells. 1 8 1 8.1 Histomorphogenesis of parenchymal regrowth within the mammary fat pad of virgin ewe lambs. 1 93 8.2 Mammary tissues from the CFP procedure. 8.3 Photographs of the udder of a ewe in which one gland had been xix 1 95 prepared as a CFP or a sham-operated CFP. 1 96 8.4 Liveweights of virgin CFP and control ewe lambs. 1 97 8.5 Weight of CFP and intact mammary glands from ewes sacrificed at various stages of postnatal development. 1 98 8.6 Development of parenchyma within the intact mammary gland of CFP ewes sacrificed at various stages of postnatal development. 8.7 Histology of mammary parenchyma and fat pad tissue from a 3-month old virgin ewe lamb. 9.1 Northern analysis of mammary IGF-I mRNA in postnatal ovme 1 99 205 mammary tissues. 2 1 1 9.2 Northern analysis of IGF-II mRNA expression in ovine mammary tissues during postnatal development. 2 1 3 10.1 Effect of various concentrations of recombinant human KGF on the proliferation of MAC-T bovine mammary epithelial cells and ovine mammary fibroblasts. 227 10.2 Northern analysis of KGF gene expression in primary cultures of bovine mammary epithelial and fibroblast cells. 228 10.3 Northern analysis of KGF mRNA in ovine mammary tissues throughout development. 229 10.4 KGF gene expression by ovine mammary cells in vitro. 230 10.5 Expression of KGF mRNA in various ovine tissues. 23 1 10.6 Quantification of KGF mRNA in ovine mammary tissues throughout postnatal development. 232 10.7 Effect of ovariectomy and oestrogen on KGF gene expression in mammary parenchyma and extra-parenchymal MFP of prepubertal ewe lambs. 10.8 Histology of mammary and adipose tissues from a prepubertal ewe 233 lamb. 237 11.1 Northern analysis of bFGF mRNA in postnatal ovine mammary tissues. 245 1 1.2 Northern analysis of HGF mRNA expression in ovine mammary tissues during postnatal development. 11.3 Effect of ovariectomy and oestrogen on bFGF mRNA levels in mammary parenchyma and extra-parenchymal MFP of prepubertal ewe xx 247 lambs. 248 11.4 In vitro expression of mRNA for HGF and bFGF by cultures of bovine mammary epithelial and fibroblast cells. 250 11.5 Dose-dependent response of COMMA- I D cell proliferation to ovine mammary fibroblast CM. 1 1.6 DNA synthesis by COMMA- I D cells in response to size-fractionated CM from cultures of ovine mammary fibroblasts. 11.7 Morphological characteristics of ovine mammary epithelial organoids cultured within collagen gels for 7 days in the presence of either 1 0% 25 1 252 FCS or ovine mammary fibroblast CM. 253 1 1.8 Proliferation of MAC-T bovine mammary epithelial cells and ovine mammary fibroblasts in response to bFGF and IGF-I. 254 12.1 Diagrammatic representation of potential growth regulatory mechanisms within the mammary gland. 273 °c aFGF ANOVA BCA b bFGF BM bp BP BSA BW cAMP cDNA CFP CIDR CM cpm CTP Da DMBA DMEM DNA ECL ECM EDso EDTA EFA EGF EGFR ER FCS FGF FP FPLC pg ng Ilg mg kg g GH GLM h HBSS HEPES HGF HPLC IGF Ig LIST OF ABBREVIATIONS degrees celsius acidic fibroblast growth factor analysis of variance bicinchoninic acid bovine basic fibroblast growth factor basal medium base pairs binding protein bovine serum albumin bodyweight 3 ' ,5' -cyclic AMP complementary deoxyribonucleic acid surgically-cleared mammary fat pad controlled intra-vaginal drug release conditioned medium counts per minute cytidine triphosphate daltons dimethylbenz [a] anthracene Dulbecco's modified Eagle's medium deoxyribonucleic acid enhanced chemiluminesence extracellular matrix median effective dose ethy lenediamine tetra-acetate essential fatty acids epidermal growth factor epidermal growth factor receptor oestrogen receptor foetal calf serum fibroblast growth factor fat pad fast protein l iquid chromatography pico-, nano-, micro-, milli-, kilo-, gram growth hormone general linear model human Hank's balanced salt solution N-2-hydroxyethyl- l -piperazine-N' -2-ethane sulfonic acid hepatocyte growth factor high-performance l iquid chromatography insulin-like growth factor immunoglobulin XXI l.m. l.p. l.v. KGF III ml l LSD nmllm mm cm m MFP MMTV pM nM IlM mM M Mr mRNA NEFA NMU n.s. OHMG ovex PAR PBS PDGF PG PgR PKC PL PMA Prl r REML RIA RNA RT-PCR s.c. SDS SEM SSC TBS TGF Tris UV intramuscular intraperi toneal intravenous keratinocyte growth factor micro-, milli-, litre least significant difference nano-, micro-, milli-, centi-, metre mammary fat pad mouse mammary tumour virus pico-, nano-, micro-, milli-, mole molecular weight messenger ribonucleic acid non-esterified fatty acids N-nitroso-N-methylurea not significant I -O-hexadecy 1-2-0-methy I glycerol ovariectomised parenchyma phosphate-buffered saline platelet-derived growth factor prostaglandin progesterone receptor protein kinase C placental lactogen I 2-myristate I 3-acetate prolactin recombinant residual maximum likelihood radioimmunoassay ribonucleic acid reverse transcriptase polymerase chain reaction subcutaneous sodium dodecyl sulfate standard error of the mean O. I 5M NaCl, O.0 1 5M trisodium citrate tris-buffered saline transforming growth factor Tris(hydroxymethyl)aminomethane ultraviolet XXll 1 CHAPTER! INTRODUCTION 2 1.1 ONTOGEN Y O F MAMMAR Y GLAND DE VELOPMENT The mammary gland of the female mammal undergoes extensive development during embryonic and postnatal growth, its ultimate function realised at parturition when it synthesises and secretes milk to nourish the offspring. During this development, mammary epithelial cells undergo extensive proliferation and morphogenesis within the confines of a stromal matrix, the mammary fat pad. This unique pattern of organogenesis is regulated by an array of systemic and local influences, many of which are specific to certain reproductive states. An appreciation of this process is fundamental to more in depth investigation into specific areas of mammary developmental biology. 1.1.1 Embryonic de velopment The first structures to arise in the mammary gland are the mammary bands; bilateral zones of ectodermal thickening on the ventrolateral body wall which become apparent in the rodent around day 1 1 of embryonic life (reviewed by Topper and Freeman, 1 980). This structural formation is associated with the local morphogenetic movement of cells rather than cell proliferation (Balinsky, 1 950) . The mammary bands subsequently divide into paired individual buds, the number and position of which correspond to the ultimate glands in the mature female (Knight and Peaker, 1 982). Underlying these mammary buds are two distinguishable mesenchymal compartments. A dense mesenchyme comprised of several fibroblastic layers directly associates with the epithelial anlage while a second compartment develops separately from the mammary anlage and is the precursor tissue for the mammary fat pad (Kimata et al. , 1 985). Lipid accumulation occurs in this latter tissue at day 1 6 of embryonic development (Sakakura et al. , 1 982). The rodent mammary bud grows very slowly to day 1 5 .5 of embryonic development (resting phase), although recent evidence indicates that this period is associated with extensive DNA synthesis in the primary duct (Cunha and Horn, 1 996). Rapid proliferation between days 1 6 and 2 1 of embryogenesis results in the formation of a mammary rudiment consisting of several branched, canalised cords that have penetrated the dense mesenchyme and which lie positioned within the fat pad precursor tissue at 3 birth (Balinsky, 1 950). Such development is associated with the formation of one or more primary sprouts, the number of which dictates the number of galactophores per teat or nipple. Cows, goats, sheep, mice and rats have one opening while humans have 1 5-25 (Anderson, 1 978). A generally similar pattern of development is observed in the embryonic ovine and bovine mammary gland (Knight and Peaker, 1 982; Raynaud, 1 96 1 ) . The mammary gland of male sheep grows at a constant rate of 2.8 times that of body weight while the female gland grows at 5 times that of body weight between days 44 and 70 of embryonic development (Martinet, 1 962). By day 70, secondary ducts have developed, the teat cistern is evident, and parenchymal tissue is well developed. Mammary growth then declines to 1 .7 times the rate of body growth until birth (Martinet, 1 962) . The four mammary buds are apparent in the bovine foetus at the 4 to 8 cm stage; at the 1 9 cm stage the mammary cord becomes canalised to form the streak canal and cistern, and at the 1 6 to 23 cm stage secondary branches arise from the dilated cistern (Raynaud, 1 96 1 ). The mammary rudiment of the embryonic female bovine undergoes most rapid growth from the 7th month of gestation. Within the mammary gland of the foetal human, a mass of epithelium l ies embedded in mesenchyme at 5 weeks of gestational age, and by 1 3 weeks the mammary bud forms cords surrounded by a fib rob lastic stroma (reviewed by Raynaud, 1 96 1 ; Kellokumpu­ Lehrinen et al., 1 987). Canalisation of ducts occurs at 20 weeks and epithelial cells develop a secretory appearance by the last trimester of pregnancy (Tobon and Salazar, 1 974) . Sexual dimorphism in the embryonic mammary gland involves hormonal action via unique epithelial-mesenchymal interaction (Sakakura, 1 99 1 ). Development of the mammary rudiment in male and female rodents is similar up until day 1 4 of embryonic development. The production of androgens by the foetal testis and the presence of epithel ial-induced androgen receptors in the mammary mesenchyme at this time (Heuberger et al. , 1 982) causes mesenchymal condensation around the mammary bud and rupture of the epithelial stalk (Kratochwil and Schwartz, 1 976). The mammary rudiment in male mice undergoes extensive necrosis whereas in the male rat it remains within the mammary fat pad but is disconnected from the epidermis (Imagawa et al., 4 1 994a). Induction of the opposite sexual phenotype can be induced by male gonad­ irradiation, or by treatment of females with testosterone (reviewed by Imagawa, 1 994a). 1.1.2 Prepubertal de velopment A substantial degree of mammary growth occurs in rodents, and indeed in other species, prior to the onset of puberty. Ductal end buds on the mammary rudiment of neonatal rodents are probably formed in response to maternal hormones. These are transient and do not reappear until around 3-4 weeks of age (Imagawa et al., 1 994a). Correspondingly, the rate of mammary growth in rats and mice is isometric until the onset of positive allometric growth at around 3-4 weeks of age. This phase of growth involves the ramification of epithelium into the mammary fat pad to establish a highly­ branched network of ducts lined with a single layer of luminal epithelial cells (Flux, 1 954). The perinatal human breast demonstrates epithelial differentiation and synthesis of "witches milk" in response to both maternal and endogenous hormones (Mayer and Klein, 1 96 1 ). Morphological development within the prepubertal human breast is variable, where mammary parenchyma may range from being a collection of simple elongate ducts to a well developed set of branched ducts with terminal lobules (Anbazhagan et al., 1 99 1 ) . The human breast then remains relatively dormant until puberty. Rat mammary tissue closely resembles that in the human gland due to the presence of terminal lobular units and an associated intralobular connective tissue (Sakakura, 1 99 1 ). Mammary parenchyma in the prepubertal heifer consists of a gland cistern and a duct system lined with double-layered epithelium; associated with these ducts are terminal alveolar structures (Mayer and Klein, 1 96 1 ). This parenchymal tissue as a whole expands and becomes less elongate during this development (Swett et al. , 1 955). Mammary gland area in rats (Sinha and Tucker, 1 966) and mice (Flux, 1 954) increases at 3 .5 and 5 times that of metabolic body weight gain to 40 and 56 days of age, respectively. Positive allometric mammary growth commences in heifers at 2-3 months of age; thereafter the rate of increase in mammary DNA content is 3 .5 times faster than that for metabolic liveweight until 9 months of age (Sinha and Tucker, 1 969b). In ewe lambs, a period of rapid mammary parenchymal growth has been recorded between 8 5 and 1 2 (Wallace, 1 953), and 1 2 and 1 6 weeks (Anderson, 1 975) of age. Similarly, lohnsson and Hart ( 1 985) recorded positive allometric mammary growth in ewe lambs between 4 and 20 weeks. This period of allometric mammary growth is particularly critical for subsequent mammogenesis and the milk yield potential of heifers and ewes (Sejrsen, 1 994). Attempts to predict the milk yield potential of dairy heifers based on the extent of mammary development in the prepubertal mammary gland" have proven unreliable (Elliott, 1 957). It is well established that this prepubertal phase of mammogenesis requires ovarian function in mice (Flux, 1 954; Bresciani, 1 968) and heifers (Wall ace, 1 953 ; Purup et aI. , 1 993b). This likely reflects a requirement for oestrogen, given its ability to restore ovariectomy-abrogated growth (Silver, 1 953). Several reports similarly indicate that mammogenesis in prepubertal rats requires ovarian function (Cowie, 1 949; Silver, 1 953; Paape and Sinha, 1 97 1 ) . However, the reviews of Imagawa et al. ( 1 990; 1 994), quoting such studies as Astwood et al. ( 1 937) and Reece and Leonard ( 1 94 1 ), state that mammogenesis in prepubertal rats is ovary-independent. This discrepancy likely reflects the type of measurements made, and that development of the rat gland is probably suppressed, although not completely abrogated, by ovariectomy. Three separate studies (Wallace, 1 953; lohnsson, 1 984; Ellis et al., 1 996a) have indicated that prepubertal allometric mammary growth in sheep is ovary-independent. Using hypophysectomised, ovariectomised and adrenalectomised female mice it has been established that completely normal prepubertal mammary development requires oestrogen + adrenal corticoid + growth hormone, where progesterone can substitute for adrenal corticoids (reviewed by Imagawa et al. , 1 990). 1.1.3 Peripubertal and postpubertal de velopment Allometric ductal growth in mice continues after the onset of puberty at 28-42 days of age, slows between 56 and 84 days, and plateaus by around 1 00 days (Flux, 1 954). Likewise, there is only a small increase in total mammary DNA between 50-60 and 1 10 days of age in female rats (Tucker, 1 969). Peripubertal mammogenesis in rodents involves the elongation and branching of mammary ducts until the ductal tree has extended to the bounds of the mammary fat pad. Ductal elongation is directed by highly mitotic terminal end bud structures which are able to penetrate the surrounding fatty 6 stroma (Daniel and Silberstein, 1 987; Section 1 .2.2). Some mice and rats may also display tertiary alveolar budding within the ductal network; the extent to which this occurs depends upon the strain (Imagawa et al. , 1 994) and their stage of oestrous (Dulbecco et al. , 1 982). The onset of puberty not only initiates stromal and epithelial proliferation within the human breast, but also results in a conspicuous increase in its size due to the deposition of substantial amounts of fat in mammary adipocytes (Russo and Russo, 1 987). Parenchymal growth is associated with the formation of terminal end buds and alveolar structures, where alveolar buds are clustered around a ductal termination to form terminal duct lobular units (Moffat and Going, 1 996). Formation of lobules commences 1 -2 years after the first menses; these lobules first appear at the periphery of the breast and extend centrally (Monaghan et al., 1 990) . Mammary gland growth in heifers continues to be allometric to around 9 months of age, thereafter slowing so that total mammary DNA content of 1 2- and 1 6-month old heifers is not different (Sinha and Tucker, 1 969b). This development corresponds to the growth of mammary parenchyma toward the bounds of the mammary fat pad (Swett et al., 1 955). A similar pattern of parenchymal development has been observed in sheep (Wallace, 1 953). 1.1.4 Oestrous cycle growth The mammary epithelium undergoes cyclical patterns of cell proliferation and morphogenesis during early puberty in response to the changing hormonal profiles of the oestrous cycle as the ductal tree extends toward the bounds of the mammary fat pad. This leads to net cumulative increases in mammary gland size (Sinha and Tucker, 1 969a; Vonderhaar, 1 988). Total mammary DNA content in rats undergoes the greatest increases during the first and second cycles (Sinha and Tucker, 1 969a). Within the oestrous cycle of the mature rat, the mitotic index of mammary epithelium is greatest at metoestrus and dioestrus while the duration of DNA synthesis is longest at metoestrus (Grahame and Bertalanffy, 1 972; Purnell and Kopen, 1 976). Using immature and pubertal rats, Dulbecco et al. ( 1 982) examined DNA synthesis in epithelial cells of ducts and terminal end buds during the oestrous cycle. End bud labelling showed peaks at early oestrus and late oestrus-metoestrus, where the distribution of this label was 7 dependent upon the epithelial cel l type (based on nuclear appearance) examined. Duct and ductule labelling was highest in late oestrus-metoestrus. The mammary gland is described as being most morphologically developed at oestrus and least developed at dioestrus (Lotz and Krause, 1 978). It has not been determined whether morphology of the mammary gland and the type of epithelial cells that proliferate within it during the oestrous cycle at the onset of puberty (when the fat pad is not fully occupied) differs from that in a fully mature gland (when ductal elongation has ceased) . It is conceivable that differences do exist; such temporal alteration may be present in the results of Sinha and Tucker ( 1 969a), where elevated mammary DNA content at pro-oestrus and oestrus was maintained into metoestrus in the first and second cycles, while peak DNA content was only measured at oestrus in cycles 3-5 . The significance of these changes within the mammary gland is emphasised by the fact that the stage of the oestrous cycle at which chemical carcinogen is administered to rats influences the extent and latency period of tumorigenesis (Lindsey et al. , 1 98 1 ; Ratko and Beattie, 1 985). Epithelial DNA synthesis within the mammary gland of parous women is greater in the luteal than the fol licular phase of the oestrous cycle (Masters et aI., 1 977; reviewed by Laidlaw et al., 1 995). Again, it is possible that the human breast may proliferate more in the follicular phase during early puberty. The DNA content of the heifer mammary gland is greatest at oestrus relative to other stages of the cycle (Sinha and Tucker, 1 969b). This cyclical variation is accompanied by increased secretory activity at oestrus when epithelial cell s assume a cuboidal appearance (Hammond, 1 927) . Levels of hydroxyproline, a measure of collagen, are also elevated in the heifer mammary gland at oestrus (Sinha and Tucker, 1 969b). 1.1.5 Gestational de velopment The majority of postnatal mammogenesis occurs during pregnancy, although the relative contribution of such growth is species-dependent (reviewed by Cowie et al., 1 980) . Likewise, there are substantial species differences in the relative contribution of maternal and foetal influences during this development (Thordarson and Tal amantes , 1 987). Gestational mammary development for a given species is� DNA doubling time remains relatively constant and is primarily a function of gestation length (Sheffield and Anderson, 1 985). 8 Gestational mammogenesis in rodents accounts for between 60 and 80% of mammary gland growth that occurs during pregnancy and early lactation (Munford, 1 964). One exception is the hamster, in which the mammary gland is essentially fully developed at term (Sinha et al., 1 970) . In the mouse, approximately 30% of total development occurs between days 6 and 1 2 of pregnancy, and approximately 50% occurs between day 1 2 and parturition (Brookreson and Turner, 1 959). The first phase of gestational mammogenesis in rodents involves extensive side-branching of ducts and the budding of alveoli into the interductal spaces of the mammary fat pad. The autoradiographic studies of Traurig ( 1 967) show peaks in DNA synthesis at days 4 and 1 2 of gestation, with similar findings for mitotic index reported by Grahame and Bertalanffy ( 1 972). It is probable that the first peak is stimulated by maternal influences such as progesterone and prolactin, and that the second is due to the onset of placental stimulation. Studies with pseudopregnant rodents indicate that maternal factors are indeed the primary impetus for mammogenesis until days 1 0- 1 2 of pregnancy, after which a placental influence is required to achieve the full development typical of a normal pregnancy (Wrenn et al. , 1 966; Desjardins et al., 1 968). Associated with epithelial growth during pregnancy is an increased accumulation of collagen within the mammary parenchyma, while the number of cells within the mammary fat pad remains unchanged (Paape and Sinha, 1 97 1 ) . Another factor that may account for up to 50% of pregnancy-associated mammary growth in rats is self-licking (Roth and Rosenblatt, 1 968), although the mechanism responsible for this dramatic effect remains unknown. The changing hormonal environment during pregnancy also induces mammary epithelial cel ls to differentiate and assume their capacity to synthesise milk constituents. A small rise in mRNA levels for certain milk proteins can be detected in the mouse mammary gland as early as day 5 (reviewed by Rosen, 1 987) although more substantial increases in the level of milk constituents such as a-lactalbumin occur around day 1 5 . This corresponds to a time when epithelial cells also demonstrate a marked increase in cell volume (Foster, 1 977). It is during the periparturient and early lactation periods, however, that maximal milk synthesis is initiated. As for rodents, epithelial growth during pregnancy in the ruminant mammary gland is exponential, and in cows equates to a cell doubling time of 87 days (Sheffield and Anderson, 1 985). However, development of the ruminant mammary gland differs from 9 that of rodents in that the ruminant gland is almost completely developed at term. It has been consistently reported that under normal conditions total mammary DNA content does not change after parturition in goats (Anderson et al., 1 98 1 ), heifers (Swanson and Poffenbarger, 1 979; Baldwin, 1 966), and sheep (Anderson, 1 975b). The mammary parenchyma of ruminants gradually replaces the interspersed adipose tissue as it undergoes extensive alveolar development during gestation (Tucker, 1 969). This process involves extensive tissue remodelling and reduces the stromal connective tissue to narrow bands (Cowie et al., 1 980; Smith et al., 1 989a). As emphasised by Swanson and Poffenbarger ( 1 979), these changes are unlikely to be reflected in total mammary weight. The total amount of mammary parenchyma inflects around 70-80 days in goats, 80- 1 1 5 days in sheep, and 1 1 0- 140 days in cows; a period when there is noticeable formation of alveoli-filled lobules. The highest percentage of epithelial tissue and eH]-thymidine-Iabelled epithelial cells in the mammary gland of ewes is around day 1 1 5 of gestation (Smith et al., 1 989a). This phase is also associated with an increase in epithelial cell size (Feldman, 1 96 1 ), a general appearance of secretory activity within the mammary gland, and an elevation in lactose synthesis and the ratio of RNA to DNA (Swanson and Poffenbarger, 1 979). As in the rodent, development of the ruminant mammary gland is influenced by a variety of maternal and foetal factors which are further discussed in respective sections of this review. Davis et al. ( 1 993) made the interesting observation that udder growth in hemimastectomised ewes remained at 50% of controls until day 1 44 of pregnancy, after which the single gland underwent compensatory growth to contain 70% of the DNA in control udders at term. Although the mechanisms which promote this compensatory growth are unknown, such a response illustrates the substantial growth potential of the mammary gland in the periparturient period. Development of the human mammary gland during pregnancy has been summarised by Russo and Russo ( 1 987). Early gestational mammogenesis involves both the proliferation of peripheral ducts and the formation of new lobules, where a range of morphologies is generally evident. By the end of the first half of pregnancy the ductal tree is essentially established. Thereafter, differentiated acini become increasingly apparent within some lobules and demonstrate an accumulation of secretion within the lumen. However, the development of individual lobules is quite heterogeneous, as some 1 0 may be fully differentiated whilst others may undergo extensive proliferation, even during lactation. 1.1.6 Lactational growth As indicated earlier, the extent to which the mammary gland develops during lactation is largely species-dependent. The mechanisms which promote such growth are essentially unknown. A wide range of studies have investigated aspects of mammary growth during lactation (reviewed by Knight and Wilde, 1 987). A major proportion of mammary development in rats, mice and guinea pigs occurs during early lactation in order to yield maximum DNA content by about day 10 of lactation (reviewed by Munford, 1 964). This growth corresponds to a peak of epithelial DNA synthesis on days 2 and 3 (Traurig, 1 96 1 ) in association with increased cell proliferation in the surrounding connective tissue. It is generally assumed that the normal ruminant gland does not undergo such development during early lactation (Knight and Wilde, 1 987), although one report (Knight and Peaker, 1 984) suggests that goats display a small increment of growth in this time. � Growth of the mammary gland during lactation may be induced by various means. Several studies have demonstrated that increased lactation demand due to hemimastectomy (Knight, 1 987) and increased milking frequency can stimulate mammary development (Wilde et al. , 1 987). Likewise, increased suckling intensity stimulates mammary growth in a variety of species (reviewed by Tucker, 1 987). It is not known by what mechanisms lactational growth is regulated. Tucker ( 1 987) points out that the results of various experiments indicate an unlikely involvement of ovarian hormones or� It also appears that the effect of growth hormone during lactation is to promote galactopoiesis more so than mammary development (Gluckman et aI. , 1 987). One possibility is that the lactating mammary epithelium requires adequate energy before it can proliferate (Goodwill et al., 1 996) . The potential for lactational growth and increased milk yield make the underlying mechanisms for such a response particularly intriguing. Embryonic and postnatal development of the mammary gland is a process which has fascinated researchers from a number of fields within the area of developmental biology. 1 1 The dramatic changes in epithelial proliferation and morphogenesis across several reproductive states in themselves provide valuable insight into the factors controlling mammary development. Growth within each of these stages is integral to the overall development of the gland before it can assume its ultimate role in milk synthesis and secretion. These changes also afford an excellent model for investigating the physiological mechanisms which regulate specific aspects of mammary gland function. Furthermore, an opportunity to study essentially the entire course of organogenesis in the postnatal animal is of particular utility to the biologist. 1.2 MAMMAR Y MORPHOGENESIS AND HISTOGENESIS The mammary parenchyma displays a range of morphologies as it ramifies into the stromal tissues of the mammary fat pad. These are generally characteristic of different reproductive states and frequently represent responses to specific local and systemic influences. Such changes result in the mammary gland becoming an elaborate structure capable of providing copious quantities of milk to the offspring on demand. Epithelium with these various morphologies differs in its risk of progressing to a tumorous phenotype. Furthermore, the nature of epithelial histogenesis and morphogenesis differs substantially across a range of species. An appreciation of the histology and morphology of the mammary gland may enable a better understanding of the mechanisms which serve to regulate normal and neoplastic mammogenesis. 1.2.1 Cell heterogeneity Mammary parenchyma comprises a heterogeneous population of epithelial cell types; 10 different types have been identified in the adult rat mammary gland (Dulbecco e t al. 1 983). Particular attention has focussed on the pluripotency of transplanted mammary epithelium and the possible existence of a stem cell popUlation able to undergo complete mammary regeneration. Williams and Daniel ( 1 983) proposed that the cap cells of the ductal end bud were one such population. A cytologically distinct, pale-staining cell type present throughout the developing mouse mammary gland may also represent a form of stem cell (Smith and Medina, 1988). The abil ity of the COMMA- I D mammary cell line to repopulate the mammary fat pad (Daniel son et al. , 1 984) led to its 1 2 subcloning to identify specific lineages that may represent stem cell types (Danielson et aI. , 1 989). The recent findings of Smith ( 1 996) indicate that separate stem cell populations do exist within the mammary gland and that they serve as separate progenitors for ductal and lobular growth. 1.2.2 The ductal end bud The end bud is a bulbous termination of the mammary duct that is found in the pre- and pubertal mammary gland of rodents and humans. Intriguingly, such a structure has not been identifiable in mammary tissues of ewe lambs and heifers by histological examination (Akers, 1 990; Ellis et aI. , 1 995), although a ful l assessment based on several other parameters remains to be conducted. The ductal end bud serves to provide a population of differentiated ductal and myoepithelial cells for ductal elongation, and to direct the path of ductal progression (Daniel and Silberstein, 1 987). These structures range from 0. 1 to 0.5 mm in diameter and are largest at the periphery of the ductal tree. The distal, basal layer of the rodent end bud is ensheathed by cap cells - cells which lack polarity and an organised cytoskeleton, and which are only loosely adherent with one another (Williams and Daniel, 1 983; Figure 1 . 1 ) . As cap cells progress along the periphery of the end bud they acquire structural and ultrastructural characteristics of myoepithelial cells . Some cap cel ls may also migrate towards the lumen of the end bud to become ductal cells . Hence, the cap cells may in fact represent a pluripotent cell type within the mammary gland (Dulbecco et al., 1 983). The apparent absence of end buds within the ruminant mammary gland raises the question as to how the mammary ducts progress into the mammary fat pad and the origin of the myoepithelial cell population in such species. 1 3 bl Figure 1.1 Composite drawing of the ductal terminal end bud made from light and electron microscope analyses. Adipocytes (a) abut against cap cells at the tip (left) . Fibrous components and fibrocytes (f) comprise the connective tissue tunic around the neck region. The basal lamina (bl) is represented as a cutaway to expose the underlying cap cells (cp). Cap cells are cuboidal but become progressively flattened toward the midregion of the end bud, then differentiate into and are continuous with myoepithelial cells (mc) in the neck region. The basal lamina overlying myoepithelial cells in the midregion is 1 4 times thicker than that at the tip. Mitosis is seen in the cap and body cells (bd). Reproduced from Williams and Daniel ( 1 983) with permission. 1.2.3 The mammary duct The outer layer of the mammary duct subtending the terminal end bud consists of myoepithelial cells positioned on the basal lamina. These cells are arranged longitudinally along the ductal axis and form a collar of cells around the inner ductal epithelium (Warburton et aI., 1 982; Emerman and Vogl, 1 986). These inner layers of ductal cells may comprise several populations which are responsive to different stimuli (Sapino et al., 1 990). While primary and secondary ducts are l ined with several layers of epithelium, terminal ducts are generally lined with a single layer of epithelial cells (Williams and Daniel, 1 983). Luminal epithelial cells possess well developed intercellular junctions and have short, blunted microvil l i . 1 4 1.2.4 Al veolar and lobuloal veolar de velopment Pregestational development of the mouse mammary gland involves the establishment of a sparse ductal tree within the confines of the mammary fat pad. In some strains, small amounts of alveolar budding may arise from ducts in response to hormonal changes during the oestrous cycle; these alveoli comprise a lumen surrounded by a single layer of epithelial cells (Vonderhaar, 1 984) and may contain evidence of secretory activity. Extensive alveolar budding commences with the onset of pregnancy, and by 6 to 8 days post coitus these alveoli have begun to cluster to form true lobuloalveolar structures surrounding small lumina (Vonderhaar, 1 988). By the end of pregnancy, the gland consists of lobules composed of many acini. As alveolar formation proceeds, myoepithelial cells alter their orientation to form a basket arrangement around each acinus (Emerman and Vogl , 1 986). The extent of alveolar development within the pubertal mammary gland varies across species. Mayer and Klein ( 1 96 1 ) attributed this variation to the relative length of the luteal phase during the oestrous cycle. This association is not surprising given the well established influence of progesterone on alveolar growth (Haslam, 1 988a; Lydon et al., 1 996). Consequently, in species such as rats and mice where the fol licular phase predominates, only ductal morphogenesis is typically evident prior to the onset of pregnancy. In contrast, the mammary gland of species that have a long luteal phase (such as the bitch) may display lobuloalveolar development during puberty which is similar to that seen in pregnancy (Mayer and Klein, 1 96 1 ) . The pubertal human breast displays several parenchymal morphologies which are distinct from those in the widely studied mouse mammary gland. These features have been previously classified in detail (Russo and Russo, 1 987). As the ductal tree continues to elongate, the terminal end buds undergo lateral and dichotomous branching after the first menses, leading to the formation of small ductules or alveolar buds. These buds are arranged around a terminal duct to form a type 1 lobule composed of approximately 1 1 alveolar buds (Figure 1 .2). With recurrent oestrous cycles, these type 1 lobules progress to type 2 lobules, and then to type 3 . This progression is heterogeneous within the gland and involves ongoing alveolar budding such that type 2 and type 3 lobules contain approximately 47 and 80 alveolar buds, respectively. 1 5 Associated with this increase in number i s a concomitant decrease i n the size of each individual unit. A · B n c l o b 3 Figure 1.2 Schematic representation of breast development. (A) At puberty or during its onset, the ducts grow and divide in a dichotomous and sympodial basis ending in terminal end buds. (B) After the first menstruation, the first lobular structures appear (lobules type 1 ) ; they are composed of alveolar buds (AB). Some branches end in terminal end buds or terminal ducts. (C) The number of lobules increases with age and in the adult nulliparous female breast, three types of lobule may be found (lobules types 1 , 2, and 3); n, nipple; lob, lobule. Reproduced from Russo and Russo ( 1 987) with permission. In studying the local regulation of mammary gland growth, it is important that species differences in parenchymal morphogenesis is acknowledged. This recognition may be of particular relevance when evaluating the regulation of processes such as branching morphogenesis, where it is conceivable that different factors influence the virgin human breast and the mouse mammary gland. There is essentially no information regarding parenchymal morphogenesis and histogenesis within the ruminant mammary gland, 1 6 although a realisation i s slowly emerging that i t differs substantially from that i n the widely studied rodent gland (Sheffield, 1 988b; Akers, 1 990) 1.3 THE STROMAL EN VIRONMENT Mammary epithelial cells within the embryonic and postnatal mammary gland demonstrate an extensive interaction with the constituents of the surrounding stroma. While early investigations considered this environment to be a relatively inert matrix in which epithelial cells grow, numerous lines of evidence now indicate that this is far from the case. There are also substantial differences in the relative proportions of various stromal tissues within the mammary gland across species, a fact which may warrant particular consideration in studies of human breast disease and ruminant mammogenesls. 1.3.1 The embryonic mesenchyme Several studies have utilised tissue recombination strategies to demonstrate that the mesenchyme of the embryonic mammary gland exerts strong influence on the adjacent epithelium. Cunha et al. ( 1995) showed that mouse mammary mesenchyme induces the epidermis to adopt a mammary epithelial phenotype. Studies by Sakakura et al. ( 1 976; 1 982) also showed that the development of mammary epithelium depends on the type of mammary mesenchyme in which it grows. Various mesenchymes can induce mammary epithelium to assume a morphogenesis which is dependent upon the origin of the mesenchyme, while mammary epithelium stil l retains its milk synthetic potential . The precursor mammary fat pad best supports organotypic development of several foetal epithelial tissues at day 14 of embryogenesis (Sakakura et al. , 1 987). Mammary fat pad precursor tissue transplanted into the adult mouse mammary gland or co-transplanted with epithelium under the renal capsule facil itates normal morphogenesis while the fibroblastic mesenchyme induces a nodular hyperplasia. 1.3.2 The mammary fat pad The mammary fat pad is a vernacular term that refers to the subcutaneous depot of adipose tissue in which the mammary gland develops. This fat pad develops from a 1 7 specific population of precursor mesenchyme which i s evident in the embryonic mouse mammary gland from day 1 4 (Sakakura, 1 987) . Transplantation of this precursor tissue to the kidney capsule of mature hosts results in its differentiation into adipose tissue after 2 weeks (Sakakura et al., 1 982). This tissue becomes less compact on days 1 5- 1 6 of embryonic development, and on days 1 6- 1 7 the preadipocytes proliferate, form lobular structures with a capillary network, and begin to accumulate lipid (Sakakura, 1 987). The mammary fat pad in neonatal rodents is readily evident as a depot of white adipose tissue. The first sign of a definitive mammary fat pad in the bovine foetus is around day 80 of gestation (Sheffield, 1 988b) . The mature mammary fat pad consists of several stromal elements including adipocytes, connective tissues, blood vessels, nerves and the lymphatic components. There are, however, species differences in the relative proportions of these tissues within the mammary fat pad, the pattern of development that they undergo, and the extent to which the stroma interacts with the developing epithelium. As is evidenced histologically, the major proportion of the rodent mammary fat pad is comprised of adipocytes, with a thin sheath of connective tissue enveloping the established ducts. The mouse mammary fat pad is also serviced by a defined vascular network (Soemarwoto and Bern, 1 958), and it is drained by lymphatics to one or more lymph nodes. The rat mammary gland demonstrates a net increase in the collagen content of both the fat pad and parenchymal fractions during puberty, while ovariectomy induces an increase in collagen and the total DNA content of the mammary fat pad (Paape and Sinha, 1 97 1 ) . The protein content of the mouse mammary fat pad remains relatively constant across mature virgin, pregnant and lactation stages, although it is unclear as to why the mammary fat pad demonstrates an increase in its total DNA content into lactation (Bandyopadhyay et al., 1 995). The mammary fat pad of the human breast is somewhat different. At birth there is an already well developed network of connective and vascular tissue, and by prepuberty the developing glandular epithelium l ies in veins of connective tissue that form conductive pathways for pubertal growth of the eventual duct system (Mayer and Klein , 1 96 1 ; Anbazhagan et al., 1 99 1 ). After the onset of puberty, the human breast demonstrates a considerable increase in size due to deposition of substantial amounts of fat and the initiation of stromal proliferation (Knight and Peaker, 1 982). Volume of the breast is 1 8 greatest around the time of menses (Cowie et al., 1 980) and can change by as much as 20% during the oestrous cycle due to changes in stromal volume (R�nnov-Jessen et al. , 1 996). Epithelial growth is preceded by stromal proliferation and the deposition of inter- and intralobular fibroblastic connective tissue (R�nnov-Jessen et al. , 1 996). As a result, and in contrast to the mouse mammary gland, human mammary epithelium develops while surrounded by a collagenous intralobular stroma (R�nnov-Jessen et al. , 1 996) . A similar although less pronounced tissue architecture is seen in the rat (Sakakura, 199 1 ) . There also exists a substantial proportion of interlobular collagenous stroma which consists of fibroblasts dispersed at a low density. The ratio of stroma to parenchyma during the course of development of the human mammary gland was reported by Russo and Russo ( 1 987). During puberty only 1 0% of the breast is parenchyma; the remaining 90% is comprised of stroma, of which 1 7% is intralobular stroma. With age, the breast of nulliparous women comprises 30% parenchyma, and 28% intralobular stroma. By late pregnancy, parenchyma occupies 73% of gland area with concomitant declines in area of both the inter- and intralobular stroma. Only subjective observations have been made regarding the nature of the mammary fat pad in ruminants. Mayer and KJein ( 1 96 1 ) state that in the neonatal calf, "the non­ glandular structures are almost mature in form, with already-established vascular and lymphatic systems. The adipose and connective tissues are also well organised. The early partitioning of the adipose tissue by the connective tissue system is remarkable, and the connective septa serve as paths for the future extension of the epithelial structures". Fat is deposited within the mammary fat pad during prepuberty and puberty, although the extent to which this occurs may be influenced by the animal ' s plane of nutrition. As the parenchymal elements progress into the mammary fat pad during prepuberty and puberty they apparently replace the adipose tissue and subsequently become enveloped by a zone of fibroblastic connective tissue similar to that seen in the human breast (Sheffield, 1 988b). 1.3.3 Mammary transplantation studies The ability to transplant normal and neoplastic mammary epithelium to various sites has facilitated investigations into several aspects of mammary biology (reviewed by Hoshino, 1 978; Sheffield, 1 988b; Medina, 1 996), especially the interrelationships 1 9 between different tissue types within the gland. A large proportion of these studies in mice and rats have been aided by the cleared fat pad technique, a procedure described by DeOme et al. ( 1 959) that involves ablating the mammary epithelial rudiment at around 3 weeks of age. This manipulation yields a mammary fat pad devoid of epithelium which can subsequently serve as a transplantation site. Several modifications of this original procedure have also been described (Hoshino, 1 962). A number of important observations have been made from transplantation studies in rodents. Firstly, repopulation experiments indicate that the amount of parenchyma which ultimately develops within the mammary gland reflects the volume of the mammary fat pad in which it grows rather than a characteristic of the epithelium itself (Hoshino, 1 978). Results from several transplantation studies also show that the mammary epithelial cel l population possesses a great deal of pluripotency. Grafts from various sites of the mammary ductal tree, including the terminal end buds, can re-establish a normal ductal network when transplanted to a virgin host (Hoshino, 1 964; Sakakura et al., 1 979; Ormerod and Rudland, 1 986). The only region incapable of such growth is the nipple. Similar potential for regrowth is possessed by foetal (Sakakura et aI., 1 979), neonatal (Currie et al., 1 977) and male (Blair and Moretti, 1 970) mammary tissue. These recombinants undergo morphogenetic and functional differentiation when hosts are mated or hormonally treated (Sekhri et al. , 1 967; Hoshino, 1 983). Immediately after its transplantation, mammary tissue undergoes extensive remodelling and proliferation within the new stromal environment (Hoshino, 1 978). Other recombination strategies have also been developed. One approach involves the isolation and in vitro genetic manipulation of mammary epithelial cells which are then inoculated into a cleared mammary fat pad or other sites, enabling the resultant phenotype to be examined (Edwards et al., 1 995; Edwards et aI. , 1 996) . Other studies have used the mammary fat pad to demonstrate that mammary epithelial cell lines such as COMMA- I D (Daniel son et al., 1 984) can undergo a complete and normal morphogenesis in vivo. Transplantation studies have also revealed that normal parenchymal elements locally regulate their growth and spacing within the mammary gland (Faulkin and DeOme, 1 960). Other experiments have shown that hyperplastic alveolar nodules are similar to normal epithelium in that they only grow within a depot of adipose tissue and, 20 furthermore, that their growth is inhibited by the presence of endogenous mammary epithelium relative to their growth in a cleared mammary fat pad (Miller et al. , 1 98 1 ) . Transplantation approaches have also enabled investigation into the lifespan of mammary epithel ium. The work of Daniel and co-workers (reviewed by Daniel and Silberstein, 1 987) demonstrated that serially transplanted epithelium had a finite lifespan, and that growth rate declined by approximately 1 5% with each generation. This lifespan was a function of the number of cell divisions rather than actual cell age. Senescent epithelium did, however, retain an ability to respond to cholera toxin, leading to the suggestion that epithelial cells become refractory to in vivo mitogenic stimulation. On the other hand, cells from hyperplastic alveolar nodules were immortalised and did not demonstrate senescent tendencies (Daniel et aI., 1 968). In contrast, the reports of Hoshino (reviewed by Hoshino, 1 978) suggested that serially transplanted mammary epithelium does not suffer any reduction in its survival potential. A likely reason for these conflicting results is that the measurements made by Hoshino related to the success of transplant recovery, whereas a more valid parameter would be the outgrowth potential of epithelial transplants as measured by Daniel and co-workers. The stromal requirements for normal mammary gland development have been investigated in the extensive studies of Hoshino (reviewed by Hoshino, 1 978). When transplanted to subcutaneous sites, mammary grafts were maintained, but did not demonstrate any outgrowth beyond the adipose tissue associated with the original transplant. This limited outgrowth also occurred in transplants to the peritoneal cavity and the anterior chamber of the eye. In contrast, grafts to the cleared mammary fat pad, the perirenal fat pad or the mesometrial fat pad demonstrated extensive ductal outgrowth into the transplanted and host stroma. The extent of outgrowth was generally greatest in the cleared mammary fat pad, although substantial amounts of growth were also observed in the perirenal fat pad. A high rate of mammary graft recovery was also achieved in transplants to the interscapular depot of brown adipose tissue (Hoshino, 1 967), although no measure was given as to the extent of the outgrowth. Nevertheless, the results of these studies indicate that mammary epithelium has an inherent requirement for a depot of adipose tissue in order for it to undergo normal growth and morphogenesis. The locality of the recipient fat pad also influences where metastatic tumours will arise; metastases from the mammary fat pad primarily arise in the lung 2 1 whilst metastases from ovarian and mesenteric adipose tissues generally occur i n the liver, spleen and diaphragm (Elliott et al. , 1 992) . The mammary fat pad of the immunologic ally-deficient athymic nude mouse and rat has been widely utilised as a site for heterologous mammary transplants (reviewed by Sheffield, 1 988b). Interestingly, Welsch et al. ( 1 987) showed that normal, but not carcinomatous, rat mammary epithelium can survive and grow within the mouse mammary fat pad. Attempts to transplant human mammary tissue have met with l imited success. Although Outzen and Custer ( 1 975) reported outgrowths from transplanted human mammary heterografts, the epithelium which they used was from abnormal mammary tissue. Transplantations conducted by Jensen and Wellings ( 1 976) did not show any outgrowth of human epithelium into the host mammary fat pad; instead transplants appeared as rounded grey structures encapsulated by a glistening sheath of connective tissue. Similar results were reported by Sheffield and Welsch ( 1 988) and Yang et al. ( 1 995). DNA synthesis in these epithelial spheroids could be stimulated by hormones, yet their morphology remained unaltered. Such studies have been extended to the transplantation of bovine mammary tissue into the mammary fat pads of nude mice (Sheffield and Welsch, 1 986). Similar to transplanted human cells, bovine cells do not outgrow but instead form hollow spheroid structures ensheathed by connective tissue. The same observations were reported by Ellis and Akers ( 1 995) using the MAC­ T bovine mammary cell line. Both studies reported increased DNA synthesis of spheroids in response to exogenous hormones without morphological change. The reason for species differences in the ability of transplanted epithelium to outgrow into the mouse mammary fat pad is intriguing. It is unlikely that it reflects different hormonal conditions in host mice, as tissue slices of human and bovine mammary tissue are maintained, grow and differentiate when transplanted subcutaneously in nude mice (Welsch et aI., 1 979; Sheffield and Welsch, 1 986). One likely explanation is that the mouse mammary fat pad does not provide a stromal environment suitable for the outgrowth of human and bovine mammary epithelium. Sheffield ( 1 988b) reported that the composition of the extracellular matrix that surrounds spheroids of heterografted bovine epithelium differs to that in normal bovine mammary tissue. Human and bovine epithelium normally grows within a matrix of connective tissue enriched with matrix components such as collagen and glycosaminoglycans. Furthermore, the mammary fat 22 pad in both of these species has an extensive, pre-established mesh of connective tissue. In contrast, the mouse mammary fat pad comprises predominantly adipocytes, and proliferating ductal epithelium abuts onto adipocytes rather than being enveloped by a fibroblastic connective tissue. Transplantation studies have also shown that both intrinsic epithelial properties and the host environment can influence mammogenesis and tumorigenesis. For example, Alston-Mills and Rivera ( 1 985) and Ethier and Cundiff ( 1 987) have shown that the growth of transplanted DMBA-induced tumours is affected by properties such as growth factor independence. Others indicate that host factors such as age (Daniel et ai. , 1 968) and dietary fat (Jp and Sinha, 1 98 1 ) can influence the growth of transplanted epithelium. 1.3.4 The epithelial-stromal reaction The encroachment of parenchyma into the adjacent mammary fat pad leads to the formation of an intimate association between the mammary epithelium and the surrounding stromal elements. Within the mouse mammary gland, cap cells and the basal lamina of the advancing terminal end bud are in direct contact with adipocytes of the mammary fat pad (Daniel and Silberstein, 1 987). This association induces DNA synthesis in stromal cells within 250 �m of the end bud, where eH]-thymidine labelling of stromal cells is greatest adjacent to the epithelium and decreases with distance (Berger and Daniel, 1 983; Dulbecco et ai., 1 982). Interestingly, this response distance of 250 �m is the same as that which separates ducts of the virgin mammary gland (Faulkin and DeOme, 1 960). This DNA synthetic response by stromal cells can only be induced by epithelium that is undergoing proliferation, for senescent ductal transplants do not evoke this effect (Daniel and Silberstein, 1 987). Proliferating epithelium in the ductal end bud also induces a local upregulation of stromal epidermal growth factor (EGF) receptors (Daniel and Si lberstein, 1 987). Furthermore, mammary parenchyma can locally influence lipolysis and lipogenesis in nearby adipocytes of the mammary fat pad (Elias et al., 1 973; Kidwell et al. , 1 982; Bartley et al., 1 98 1 ) by liberating diffusible, soluble factors (Lucas et al., 1 976). This epithelial influence on adipocyte metabol ism might also be mediated by local immune mast cells (Kidwell and Shaffer, 1 984; Haslam, 1 988b). It is not known to what extent this local influence on adipocyte metabolism subsequently regulates the growth of the mammary epithelium. 23 The constricted flank of the ductal end bud also demonstrates extensive stromal reaction. This region, encased by proliferating fibrocytes synthesising interstitial collagen (Williams and Daniel, 1 983), is enriched with sulfated glycosaminoglycans such as chondroiton sulfate (Silberstein and Daniel, 1 982). Fibrosis within this zone may be induced by the newly differentiated myoepithelium and/or the synthesis of the basal lamina (Wicha et al. , 1 980). One question that does arise is whether this fibroblastic stroma originates from cells of the mammary fat pad, or from mesenchymal cells that have co-migrated with the advancing ductal epithelium. In the case of mammary epithelial heterografts which overexpress the Wnt- l on co gene and induce extensive stromal fibrosis (Cunha and Horn, 1 996), the resultant fibroblastic stroma has been shown to originate from constituents of the host' s mammary fat pad (G.R. Cunha, personal communication). It remains to be determined whether the extensive fibrosis that occurs around parenchyma in the human and ruminant mammary gland similarly arises from the stromal cells of the mammary fat pad proper. The myoepithelial cell population is closely associated with epithelial cells and forms a collar of longitudinally-oriented cells around the ductal epithelium in the nonpregnant female (Haslam, 1 988b). The myoepithelium may therefore act to regulate exposure of epithelial cells to the underlying basement membrane. The most pronounced example of stromal reaction is seen in the vicinity of breast tumours, where desmoplasia may account for greater than 90% of cells in tumours such as infiltrating ductal carcinomas (R�nnov-Jessen et aI . , 1 996). This response may involve the upregulated expression of genes for certain mitogens such as IGF-II (Singer et al., 1 995). The importance of this stromal reaction may be emphasised by the finding that human adenocarcinoma cells transplanted to athymic nude mice have an increased tumour take and growth rate when they are co-inoculated with human fibroblasts (Noel et al., 1 994). In addition to markedly altering the composition of the extracellular matrix, desmoplasia may encourage the appearance of a modified fibroblast cell type, the myofibroblast, which is positive for a-smooth muscle actin. The stromal revelation of this cell type during tumour progression may further alter the composition of the extracellular matrix, or may allow contractile forces to be exerted during the growth of tumorous epithelium (R�nnov-Jessen et al. , 1 996) . 24 1.3.5 Modelling epithelial -stromal associations With an increasing recognition that local influences play an important role during mammary development and tumorigenesis, numerous attempts have been made to model the interactions that take place between the epithelium and the adjacent stroma (reviewed by Ip and Darcy, 1996). One such approach has been to "condition" cel l culture medium by' incubating it with different constituents of the mammary gland stroma. This conditioned medium can then be added to cultures of mammary epithelial cells to determine their response to soluble factors liberated by the stromal tissue or cells . Numerous studies have prepared conditioned medium from cultured monolayers of foetal mammary stroma (Grey et al. , 1 989), and stroma of adult rodent (Sasaki et aI., 1 994), human (Van Roozendaal et al. , 1 996) and bovine (Woodward et al . , 1 992) mammary tissue, as well as mammary tumour cel ls (Cappelletti et aI. , 1 993). Other studies have conditioned medium using differentiated 3T3-L l preadipocytes (Levine and Stockdale, 1 984; Rahimi et al., 1 994), isolated mammary adipocytes, or explants of mammary fat pad tissue (Beck and Hosick, 1 988). Almost al l of these studies have demonstrated that the stromal constituents produce some form of biological activity (particularly polypeptide growth factors) capable of modifying such characteristics as anchorage-dependence, morphogenesis, proliferation, and milk protein synthesis. Other studies have investigated physical and physico-chemical influences on epithelial growth and morphogenesis by co-culturing epithelial and stromal cells . These co­ cultures may incorporate live cell populations (Wang and Haslam, 1 994), stromal cells inhibited by glutaraldehyde or killed by irradiation (Levine and Stockdale, 1 985; Haslam, 1 986), or the matrix synthesised by stromal cel ls. Such conditions have been shown to stimulate epithelial DNA synthesis; likewise, stromal DNA synthesis is stimulated in live mixed cultures. Other characteristics such as epithel ial morphology, milk protein expression (Wiens et al., 1 987) and progesterone receptor levels may also be altered. A major disadvantage of such an approach is that in many cases it is difficult to segregate the epithelial and stromal cell types for routine assay. A third, similar approach has been to culture epithelial and stromal populations that are physically separated from each other while stil l allowing bidirectional diffusion of soluble factors. This approach has included the use of culture well inserts and the co-casting of 25 mammary fat pad explants in collagen along with epithelial organoids (Carrington and Hosick, 1 985). 1.3.6 The extracellular matrix Literature concerning the role of the extracellular matrix in mammary development, differentiation, and lactogenesis is extensive and will not be reviewed in detail here. For comprehensive reviews of this subject the reader is referred to the reviews of Bissell and Hall ( 1 987); Aggeler et al. ( 1 988); Barcellos-Hoff and Bissell ( 1 989) and Blaschke et al. ( 1 994). It is pertinent, however, to briefly consider the involvement of the extracellular matrix in mammary gland development and morphogenesis, and the likely importance of the mammary fat pad in providing such a substratum. The extracellular matrix on which the mammary epithelium is positioned consists of three different strata (Sakakura, 1 99 1 ) . Basal to the epithelial cells i s a thin sheet referred to as the lamina densa that is separated from the epithelium by a narrow translucent space known as the lamina lucida. The lamina densa and the lamina lucida are collectively referred to as the basal lamina. Adjacent to the basal lamina and associated with the stromal components is a layer of variable thickness known as the reticular lamina. Epithelial cells in the virgin rodent mammary gland are actually separated from the basal lamina by a continuous collar of myoepithelial cells, whereas during pregnancy, epithelial cells become directly exposed to the basal lamina as the myoepithelial cells assume a basket-like arrangement (Haslam, 1 988b). The basal lamina that ensheaths the distal region of the end bud and the subtending ducts in the virgin mammary gland is approximately 1 00 nm thick, whilst that along the constricted neck region of the end bud may be 1 3-20 times this thickness (Daniel and Silberstein, 1 987). This latter region is a site of considerable sulfated glycosaminoglycan synthesis (Silberstein and Daniel, 1 982). Until recently it has been accepted that the epithelium produces components of the basal lamina such as laminin, type IV collagen and heparan sulfate proteoglycans (including nidogen and entactin), while cells of the stroma synthesise constituents of the reticular lamina such as type I and III collagens, fibronectin, and tenascin (Sakakura, 1 99 1 ) . These assumptions have been based on the results of immunolocalisation studies and in vitro experiments showing the presence of various matrix components in epithelial 26 cultures. However, more recent in situ hybridisation results indicate that only the stromal connective tissue cells and adipocytes, and not the epithelium, express mRNA for collagen I, collagen IV, and larninin in the mouse mammary gland (Keely et al., 1 995). Furthermore, this expression is developmentally regulated; collagen I mRNA levels are highest in prepuberty and early pregnancy and precede an upregulation of collagen IV and laminin expression. These findings emphasise the critical requirement for the mammary fat pad stroma during development and the importance of the epithelial-stromal reaction in regulating epithelial progression and morphogenesis. It remains unclear as to whether the detection of these matrix proteins in epithelial cultures in vitro is a function of cell adaptation to an artificial environment, or if it simply reflects contamination of cultures with stromal cells or their remnants. While the extracellular matrix plays an integral role in regulating the development and function of mammary epithelium, the extent of its involvement and the specific roles of its numerous constituents remain to be fully characterised. In vitro studies frequently adopt the technique of embedding primary epithelial cells within collagen gels to achieve sustained growth and hormonal responsiveness (Yang et al., 1 980). However, collagen I alone does not fulfil the matrix requirements for normal morphogenesis as shown in an experiment conducted by Daniel and co-workers (Daniel et aI., 1 984), where epithelium transplanted into collagen gel inside the mammary fat pad grew as radial spikes within the gel and only formed end buds when it reached the stroma of the mammary fat pad. It is possible that specific components of the extracellular matrix are prerequisite for certain stages of mammogenesis. For example, tenascin is only expressed by the foetal mesenchyme in response to epithelial induction (Sakakura, 1 99 1 ) while the stroma of the postnatal gland differentially expresses collagen I, collagen IV and larninin during virgin and gestational development (Keely et al. , 1 995). Other studies have shown that the synthesis of extracellular matrix components is regulated by various hormones and growth factors (Blum et al., 1 989a), further supporting suggestions for their stage­ specific roles. Furthermore, cell-surface receptors for these extracellular matrix proteins, the integrin subunits, are expressed by epithelial cells during specific stages of development (Anbazhagan et al. , 1 995) and are markedly repressed during lactation (Keely et aI., 1 995). Likewise, inhibition of type IV collagen synthesis in vivo using cis- 27 hydroxyproline leads to a reduction in epithelial growth and encourages the regression of epithelium to an involuted-like state (Wicha et al., 1 980) . The most pronounced requirement of epithelium for an extracellular matrix is seen in studies of milk protein synthesis where epithelial differentiation and hormonal responsiveness is only achieved when cells are cultured on substratum such as col lagen or a reconstituted membrane matrix such as Matrigel. Aside from having marked effects on epithelial characteristics such as morphogenesis and differentiation, the extracellular matrix can also regulate other cellular responses including the expression of hormone and growth factor receptors (Mohanam et al., 1 988; Haslam, 1 986) and cell proliferation (Salomon et al., 1 98 1 ; Wicha et al., 1 982). Taken together, this information strongly implicates the stroma as fulfilling a major role in regulating development of the mammary gland. This regulation may be effected both chemically and physically, and also likely depends upon the local reaction between epithelial cells and the surrounding stroma. Such information is of fundamental importance to our understanding of mammary gland biology. However, a great deal remains to be understood about the full extent of this regulation and its role during the course of normal and neoplastic mammogenesis. Furthermore, species differences remain to be explored, where it appears that there are several major distinctions between the stromal environment of the rodent mammary gland and that in other species such as humans and ruminants. 1.4 FACTORS IN FLUENCING MAMMAR Y DE VELOPMENT The postnatal proliferation of epithelial cells within the mammary fat pad is influenced by an array of systemically and locally derived factors. While it is well established which hormones primari ly control mammogenesis, in several cases their mechanism of action on the mammary gland is unknown. Furthermore, an increasing body of information suggests that the local synthesis of other factors within the mammary gland may be critical to the growth and morphogenesis of mammary epithelium. A role for these factors has been studied in the normal mammary gland of various species as well as in the tumorous mammary gland of rodents and humans. 1.4.1 The ovarian steroids 1.4.1.1 Oestrogen 28 Numerous studies have demonstrated the important role of oestrogen during postnatal mammogenesis (reviewed by Haslam, 1 987). Oestrogen is a potent stimulator of DNA synthesis in the mouse mammary gland (Traurig and Morgan, 1 964), particularly within the terminal end buds (Bresciani, 1 968). A requirement for the mammogenic effect of oestrogen is exemplified in prepubertal mice where ovariectomy-abrogated ductal growth can be restored by exogenous oestrogen (FlUX, 1 954) . While development of the rat mammary gland is unaffected by prepubertal ovariectomy, administration of oestrogen to ovariectomised female rats promotes both ductal and lobuloalveolar development (reviewed by Nandi et al., 1 995). Oestrogen also stimulates DNA synthesis in human breast tissue transplanted to athymic nude mice (Laidlaw et al. , 1 995). Short-term administration of supraphysiological doses of oestrogen to prepubertal ewe lambs (Ellis et al., 1 996a) and heifers (Woodward et al., 1 993) increases DNA synthesis in the mammary epithelium. This oestrogen-induced proliferation in the mammary gland of heifers was followed by a phase of stromal DNA synthesis. It is also well established that heifers and ewes grazing pastures with a high phyto-oestrogen content frequently demonstrate precocious mammary gland development (Adams, 1 995). Although ovariectomy abrogates prepubertal mammary development in heifers (Purup et al.� 1 993b), the involvement of oestrogen in this effect is somewhat unclear as the serum oestrogen concentration was only reduced by 30% (0. 1 pg/ml) fol lowing ovariectomy. In contrast, prepubertal mammary growth of ewe lambs is unaffected by ovariectomy (Wall ace, 1 953 ; Ellis et al., 1 996a). Others have shown that oestrogen, in combination with progesterone, can facilitate normal growth in ovariectomised heifers (Sud et al., 1 968); the combination of these two steroids is an essential requirement to adequately develop the mammary gland during hormonal induction of lactation (Sawyer et al., 1 986). Responsiveness of the mouse mammary gland to oestrogen-induced proliferation also varies during postnatal development (Haslam, 1 989). Specifically, oestrogen does not induce DNA synthesis in the mammary gland of 3- 14 day old female mice, while by 3-4 weeks of age both stroma and epithelium proliferate in response to oestrogen. Along 29 these l ines, it is unlikely that oestrogen serves a role during normal embryonic and perinatal mammary development given that the growth of the mammary rudiment is unaltered in female mice lacking a functional oestrogen receptor (Korach, 1 994). On the other hand, administration of exogenous oestrogen to embryonic mice does result in malformed mammary gland (Raynaud, 1 96 1 ). The finding that ductal elongation in hypophysectomised mice was not promoted by exogenous oestrogen indicated that pituitary hormones were also required to elicit the mammogenic effect of oestrogen (Lyons et ai., 1 958; Nandi, 1 958). Subsequent studies in triply-operated mice and rats showed that the combination of oestrogen + corticoid + growth hormone could fully restore mammary development (reviewed by Imagawa, 1 990). A synergistic response by ductal epithelium to oestrogen + prolactin has also been reported (Stoudemire et al., 1 975). However, it stil l remained unclear as to whether oestrogen initiated its effect by a direct action on the mammary gland, or via an indirect action through the pituitary. Using slow-release implants positioned within the immature mammary gland, Daniel et al. ( 1987) and Haslam ( 1 988d) demonstrated that oestrogen acts directly on the mammary gland to stimulate ductal growth. This mode of action is further supported by the fact that ductal growth is suppressed in the presence of implants containing antiestrogens (Silberstein et al., 1 994). In contrast, oestrogen­ stimulated proliferation in the mature gland is systemically mediated, while local oestrogen release can sti l l induce an upregulation of epithelial progesterone receptors (Haslam, 1 988d). Even though a local action of oestrogen had been shown, there remained the paradox that cultured epithelial cells from various species were not stimulated to proliferate by oestrogen in vitro (Yang et al. , 1 980; Richards et ai., 1 988). Only occasional reports have demonstrated oestrogen responsiveness of cultured mammary epithelium (Gompel et ai., 1 986). The establishment of an oestrogen-responsive breast cancer cell line (MCF-7) has led to its widespread adoption (Lippman et al., 1 976) as a model for hormone-dependent breast cancer. Yet while cultures of mammary epithelial cells do not proliferate in response to oestrogen, they