Investigation of the role of prolactin in mammary gland development and carcinogenesis.
Samantha Richelle Oakes
A thesis submitted for the Degree of Doctor of Philosophy in the Faculty of Medicine at the University of New South Wales.
December 2006 ORIGINALITY STATEMENT
‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project's design and conception or in style, presentation and linguistic expression is acknowledged.’
Signed ……………………………………………......
Date ……………………………………………......
ii “I wish to record my unbounded admiration for the work of the experimenter in his struggle to wrest interpretable facts from an unyielding Nature who knows so well how to meet our theories with a decisive No-or with an inaudible Yes.”
Hermann Weyl (1885-1955)
Weyl H. (1931) Gruppentherie und Quantenmechanik p2 English translation: Robertson, HP. (1931) The Theory of Groups and Quantum Mechanics, p xx
I couldn’t agree more!
iii Acknowledgements
As I sit here and reflect on the past few years, I cannot help but think about all of the people who have helped me along the way. Firstly, I must thank my supervisor, Chris Ormandy. Thank you for taking the chance on a mad motorcyclist, for knowing that it was not experience, but the ability to acquire knowledge that was important as a scientist. Over the past six years, with your encouragement and support, I have grown confident in my abilities as a researcher, and most important of all I have come out of it loving science more than when I began. I feel proud to have been part of your laboratory, and I could not have asked for a more compatible supervisor. I thank you for the wonderful experiences I have had over this journey. Since beginning work at the Garvan Institute, I have been humbled by working with so many incredible and fascinating people. To Roger Daly, my co-supervisor, thank you for many stimulating conversations, and your continuous support. To Jess Harris and Fi Robertson, thank you for being so welcoming when I first started and for allowing me to share your windowsill. Jess, you inspired me with your intelligence and lateral thinking, and Fi, your friendship has meant a great deal to me. To my colleague Matt Naylor, I cannot begin to thank you for all the advice and training you have given me over the course of my thesis. There may be plenty of competition between us, but I respect you in so many ways, and you push me to be a better scientist. Heres to plenty more ‘friendly’ competitions at high altitude! I also thank you for our collaboration on the secretory activation project. I thank Margaret Gardiner-Garden for her collaboration involving the statistical analysis of the work in this thesis, and to Michael Kazlauskas for his wonderful work on the Elf5 transgenic. To Prue Stanford, I treasure your advice and friendship, thank you for always being an ear to cry to. I also thank you for your collaboration with the Elf5 retroviral experiment and the immunohistochemistry. To my dear friends Heidi Hilton and Katrina Blazek, I am so privileged to have worked with you. Thank you for many wonderful scientific and non-scientific discussions, and thank you for being my stunning bridesmaids. I thank Liz Caldon for her many words of wisdom and her valuable friendship, it’s an honour working in the same lab as you. To Rhian, thank you for all of your hard work and assistance in my final year of my PhD. To Maria, Simon, Karl, Daniel, Andrea and Chehani, thank you all for being part of this journey. I also thank Sarah iv Eggleton for her amazing expertise involving immunohistochemistry. To Jay, Alex, Darren, Tillman, Nick, Gillian, Eileen, Marcello, Luke, Amanda, Ruth and everyone else in the cancer program, thank you for your help and friendship. I also would like to thank Jeffrey Green for providing the SV40T mouse, and Lewis Chodosh for the tetracycline inducible MTB mouse. Finally I have to thank my family. To Mum and Dad, I will always be thankful for your gifts to me of love, support and education. I am so proud to be your daughter and I am blessed to have you both. For this, I dedicate this thesis to Mum and Dad, as your hard work and sacrifices have contributed to my achievements, this degree being the greatest of them all. To my brother Steven, ‘bro’, thank you for your love and support, it means the world to me. Thank you also to Nanna, Pa, Greg, Collette and Vicky, you all have made me who I am today. Finally I thank my husband, Brad. I know it hasn’t been easy, but you have helped me through one of the hardest experiences in my life. For that, I will be eternally grateful.
Sammy
v Abbreviations
-casein Casein alpha Acyl ATP citrate lyase Aldo3 Aldolase 3, C Angptl4 Angiopoietin-like 4 AP-1 Activator protein 1 -Casein casein beta bGH Bovine growth hormone BrdU 5-Bromo-2’-deoxyuridine Btn1a1 Butyrophilin, subfamily 1, member A1 Ccnd1 Cyclin D1 CD24a CD24a antigen Cdc6 Cell division cycle 6 homolog (S. cerevisiae) Cebp CCAAT/enhancer binding protein (C/EBP), beta Cebp CCAAT/enhancer binding protein (C/EBP), delta Cel Carboxyl ester lipase CFP Cleared fat pad Cidea Cell death-inducing DNA fragmentation factor, alpha subunit-like effector A Cldn3 Claudin 3 Copz1 Coatomer protein complex, subunit zeta 1 Cyp51 Cytochrome P450, family 51 -Casein Casein delta Dox Doxycycline dpc days post coitus dpp days post partum E Estrogen ECM Extracellular matrix EGF Epidermal growth factor EGFP Enhanced green fluorescent protein EGFR/ErbB/Her1 Epidermal growth factor receptor/v-erb-b2 erythroblastic leukemia viral oncogene homolog Ehf Ets homologous factor Elf3 (ESE-1) E74-like factor 3 Elf5 (ESE-2) E74-like factor 5 Elf5-/- E74-like factor 5 knockout Elf5+/- E74-like factor 5 heterozygote Elovl5 Elongation of very long chain fatty acids 5 Er Estrogen receptor ER Estrogen receptor alpha Erbb2/Her2 v-erb-b2 erythroblastic leukemia viral oncogene homolog 2/Hairy-related 2 ErbB3/Her3 v-erb-b2 erythroblastic leukemia viral oncogene homolog 3/Hairy-related 3
vi Erbb4/Her4 v-erb-b2 erythroblastic leukemia viral oncogene homolog 4/Hairy-related 4 Erg v-ets erythroblastosis virus E26 oncogene homolog Fak Focal adhesion kinase FCS Fetal calf serum Fdps Farnesyl diphosphate synthase Fos FBJ osteosarcoma oncogene Foxa1/HNF-3 Forkhead box A1 gene Fyn Fyn proto-oncogene G1p2/Isg15 ISG15 ubiquitin-like modifier Gal Galanin Gal-/- Galanin knockout Gal+/+ Galanin wildtype Gapdh Glyceraldehyde-3-phosphate dehydrogenase Gata3 GATA binding protein 3 GFP Green fluorescent protein GH Growth hormone GM-CSF Granulocyte-macrophage colony-stimulating factor Grb2 Growth factor bound protein 2 HandE Haematoxylin and Eosin HGMIN High-grade mammary intraepithelial neoplasia Id1 Inhibitor of DNA binding 1 Id2 Inhibitor of DNA binding 2 Igf2 Insulin-like growth factor 2 IKK (IKK ) Kinase IkB Ink4a (Cdkn2a) Cyclin-dependent kinase inhibitor 2A IRES Internal ribosome entry site Jak2 Janus kinase 2 Jun Jun oncogene -casein Casein kappa Kcna1 Potassium voltage-gated ion channel, shaker-related subfamily 1 Lalba Lactalbumin Ldlr Lipoprotein receptor Lef1 Lymphoid enhancer binding factor 1 LGMIN Low-grade mammary intraepithelial neoplasia MapK/Erk Mitogen activate protein kinase Mcmds 2-7 Mini chromosome maintenance deficient 2-7 homolog (S. cerevisiae) MEC Mammary epithelial cell MMTV Mouse mammary tumor virus MTB (MMTV-rtTA-pA) Mouse mammary tumor virus-tetracycline-dependent transactivator-polyadenylation NF1/CTF Nuclear factor 1/CCAAT-binding transcription factor NF- B Nuclear Factor-kappa B NMU Nitrosomethylurea
vii Nrg Neuregulin Oct1 Octamer-binding transcription factor-1 Orc6 Origin recognition complex, subunit 6-like Pcna Proliferating cell nuclear antigen PCR Polymerase chain reaction Pea3 Ets domain protein 3 Pg Progesterone Pgr/PR Progesterone receptor PgrA Progesterone receptor A PgrB Progesterone receptor B PI3K phosphatidylinositol 3-kinase PKB/Akt Protein kinase B PL Placental lactogen pMapK/pErk phosphorylated Mitogen activate protein kinase pPKB/pAkt phosphorylated Protein kinase B p-Prl phosphorylated Prolactin Prl Prolactin Prlr Prolactin receptor Prlr-/- Prolactin receptor knockout Prlr+/- Prolactin receptor heterozygote Prlr+/+ Prolactin receptor wildtype PTHrP Parathyroid hormone-related protein PTHrPr Parathyroid hormone-related protein receptor Raf v-raf-1 murine leukemia viral oncogene homolog 1 Rank Receptor activator of NF-kB RankL Receptor activator of NF-kB-ligand Ras Ras proto-oncogene Rfc3 Replicating factor C 3 Rfc4 (Rfc 1 4) Replicating factor C 4 Rfc5 (Rfc 1 5) Replicating factor C 5 Scd2 Stearoyl-CoA desaturase 2 Shc Src homology 2 domain containing transforming protein SHC src homology collagen siRNA small interfering RNA Slc34a2 Solute carrier family 34 (sodium phosphate), member 2 Slc39a8 Solute carrier family 39 (zinc transporter), member 8 Smad Mothers against decapentaplegic homolog Socs1 Suppressor of cytokine signalling 1 Socs2 Suppressor of cytokine signalling 2 Sox4 Sry-related gene a4 Sppr2a Small proline-rich protein 2A Sqle Squalene epoxidase Srebf1/Srebp1 Sterol regulatory element-binding protein 1 Stat3 Signal transducer and activator of transcription 3 Stat5a Signal transducer and activator of transcription 5a Stat5b Signal transducer and activator of transcription 5b
viii SV40T Simian virus 40 T-antigen Tcf1/Hnf1 Transcription factor 1 TEB Terminal end bud TetOn Tetracycline on TGF Transforming growth factor, alpha Tgf-1 Transforming growth factor, beta 1 Tm4sf1 Transmembrane 4 superfamily member 1 Tmprss2 Transmembrane protease, serine 2 Wap Whey acidic protein Wdnm1/Expi Extracellular proteinase inhibitor Wnt4 Wingless-related MMTV integration site 4 WT wildtype -casein Casein gamma YY 1 Y1 transcription factor Zo1 Tight junction protein 1
ix Abstract
The pituitary hormone prolactin (Prl) is essential for alveolar morphogenesis and plays a role in breast carcinogenesis, however the mechanism that underlies these actions remains to be defined. Alterations in serum Prl provide the primary endocrine signal regulating developmental events in the mammary gland in sexually mature mammals. Prl production and post-translational phosphorylation by the pituitary is regulated by the neuropeptide Galanin (Gal) in response to hypothalamic signals integrating neuronal and endocrine inputs. Prl exerts its effects on the mammary epithelium in two ways, indirectly by modulation of the systemic hormonal environment, for example the release of progesterone from the corpus luteum, and directly by binding to Prl receptors (Prlr) within the mammary epithelium. Prl binding to Prlr initiates signalling predominantly via activation of the Jak2/Stat5 pathway, leading to altered patterns of gene transcription. One of these target genes is the ets transcription factor Elf5, which is required by the epithelium for alveolar morphogenesis. This thesis aims to further our understanding of the mechanisms by which prolactin exerts its influence on the mammary gland during alveolar morphogenesis and carcinogenesis. Transcript profiling revealed a lactation signature of 35 genes in Prlr+/- mice, Gal-/- mice and mice treated with a Prl mutant (S179D) that mimics phosphorylated Prl. We discovered that the majority of changes in gene expression were produced by prolactin rather than by Gal. The action of Gal was predominantly via modulation of Prl phosphorylation and release, as its effects were very similar to that of S179D. Knockout of Elf5 phenocopied knockout of Prlr, resulting in failure of alveolar morphogenesis and reduced expression of milk and lipid synthesis genes. Forced Elf5 expression at puberty resulted in aberrant differentiation of the terminal end buds and milk protein synthesis during ductal morphogenesis. Re-expression of Elf5 in Prlr-/- mammary epithelial cells completely rescued alveolar morphogenesis. These observations indicate that Elf5 is a master regulator of alveolar morphogenesis downstream of the Prlr. Loss of mammary epithelial Prlr resulted in reduced proliferation of low-grade neoplastic lesions resulting in increased tumour latency in the C3(1)/SV40T model of mammary carcinogenesis. There was no change in the growth rate, proliferation nor the morphology of tumours in Prlr-/-/C3(1)/SV40T transplants, thus Prl acts early in x carcinogenesis to drive the proliferation of pre-invasive lesions resulting in faster progression to cancer.
xi Publications arising from this thesis
1. Robertson FG, J, Harris J, Naylor MJ, Oakes SR, Kindblom J, Dillner K, Wennbo H, Tornell J, Kelly PA, Green J and Ormandy CJ (2003). Prostate development and carcinogenesis in prolactin receptor knockout mice. Endocrinology 144: 3196- 3205. (Appendix I) 1. Davison EA, Lee CS, Naylor MJ, Oakes SR, Sutherland RL, Hennighausen L, Ormandy CJ and Musgrove EA (2003). The CDK inhibitor p27 (Kip1) regulates both DNA synthesis and apoptosis in mammary epithelium but is not required for its functional development during pregnancy. Mol. Endocrinol. 17:2436-2447. (Appendix II) 1. Harris J, Stanford PM, Oakes SR and Ormandy CJ (2004). Prolactin and the prolactin receptor - new targets of an old hormone. Ann. Med. 36:414-425. (Appendix III) 1. Naylor MJ*, Oakes SR*, Gardiner-Garden M, Harris J, Ho TWC, Li FC, Wynick D, Walker AM and Ormandy CJ (2005). Transcriptional changes underlying the secretory activation stage of mammary gland development. Mol. Endocrinol. 19:1868-1883. (* Joint first authors). (Appendix IV) 1. Harris J, Stanford PM, Sutherland K, Oakes SR, Naylor MJ, Robertson FG, Blazek KD, Kazlauskas M, Hilton HN, Wittlin S, Alexander WS, Lindeman GJ, Visvader JE and Ormandy CJ. (2006) Socs2 and elf5 mediate prolactin-induced mammary gland development. Mol. Endocrinol. 20:1177-1187. (Appendix V) 1. Oakes SR, Hilton HN and Ormandy CJ. (2006) Key stages in mammary gland development - The alveolar switch: coordinating the proliferative cues and cell fate decisions that drive the formation of lobuloalveoli from ductal epithelium. Breast Cancer Res. 8:207. (Appendix VI) 1. Oakes SR, Robertson FG, Kench JG, Gardiner-Garden M, Wand MP, Green JE and Ormandy CJ. (2006) Loss of mammary epithelial prolactin receptor delays tumour formation by reducing cell proliferation in low-grade preinvasive lesions. Oncogene. in press. (Appendix VII)
xii Table of Contents Declaration ii Acknowledgments iv Abbreviations vi Abstract x Publications arising from this thesis xii Table of Contents xiii
Chapter 1 Introduction...... 1
1.1 Preface...... 1
1.2 Cellular morphology...... 2
1.3 Embryonic development...... 2
1.4 Ductal branching morphogenesis...... 4
1.5 Tissue remodelling during pregnancy ...... 7
1.6 The pituitary hormone prolactin ...... 9 1.6.1. Splice Variants ...... 10 1.6.2. Proteolytic Variants ...... 10 1.6.3. Phosphorylated prolactin ...... 10 1.6.4. Other Post-translational modifications ...... 11
1.7 The prolactin receptor...... 12
1.8 Prolactin and progesterone initiation of alveolar morphogenesis...... 15
1.9 Molecular modulators of Prl induced alveolar morphogenesis ...... 17
1.10 Transcription factors involved in alveolar morphogenesis...... 22 xiii 1.11 Other factors involved in alveolar morphogenesis ...... 25
1.12 The ets transcription factor Elf5...... 27
1.13 Prolactin in mammary carcinogenesis...... 28
1.14 Aims of this thesis ...... 31
Chapter 2 Materials and Methods...... 33
2.1 Animal husbandry ...... 33
2.2 Tumor investigation ...... 34
2.3 Trypan blue supravital staining...... 34
2.4 Mammary epithelial transplantation...... 34 2.4.1. Prlr-/- and Elf5-/- study...... 34 2.4.2. C3(1)/SV40T transplantation study...... 34
2.5 Cryogenic preservation of mammary epithelium...... 37
2.6 Elf5 expression in Elf5/MTB transgenic mice ...... 37
2.7 Retroviral infection of mammary epithelial cells...... 37
2.8 Mammary whole-mounting...... 38
2.9 mRNA and protein isolation ...... 38
2.10 Cryomounting and sectioning ...... 38
2.11 HC11 cell differentiation protocol and siRNA ...... 39
2.12 Transcript profiling...... 39
2.13 Quantitative PCR ...... 40
xiv 2.14 Quantitative PCR analysis ...... 42
2.15 Western blotting ...... 42
2.16 Immunocytochemistry...... 43
2.17 Immunofluorescence...... 43
2.18 Statistics ...... 44
Chapter 3 Transcriptional regulation of mammary secretory activation ...... 45
3.1 Introduction...... 45
3.2 Results...... 48 3.2.1. Stat5 phosphorylation was decreased in Prlr+/- mice that are unable to lactate...... 48 3.2.2. Transcript profiling of three models of failed lactation identified a lactation signature comprised of 35 key genes...... 51 3.2.3. Quantitative PCR validated the transcript profiling results obtained from the three models of failed lactation...... 53 3.2.4. Failed secretory activation occurred at the same stage in all three models of failed lactation...... 55 3.2.5. The lactation signature contained a small number of genes important for milk and lipid synthesis...... 55 3.2.6. Rescued lactation in Prlr+/- mice was associated with increased expression of genes involved in proliferation...... 62 3.2.7. Hierarchical clustering demonstrated that gene expression changes as a result of loss of Gal were more closely related to treatment with S179D, than that of loss of an allele of Prlr...... 65 3.2.8. siRNA knockdown of Stat5a in the HC11 cell model of mammary cell differentiation modulated genes involved in the lactation signature...... 68 xv 3.3 Discussion...... 72
Chapter 4 The role of Elf5 during alveolar morphogenesis...... 77
4.1 Introduction...... 77
4.2 Results...... 81 4.2.1. Production of a renewable source of Elf5-/- mammary glands...... 81 4.2.2. Mammary epithelial Elf5 was not required for branching morphogenesis. .81 4.2.3. Mammary epithelial loss of Elf5 resulted in failed alveolar morphogenesis, which resembled loss of mammary epithelial Prlr...... 83 4.2.4. The Elf5 heterozygote mammary phenotype was partially penetrant...... 86 4.2.5. Alveolar proliferation was attenuated in both Elf5 and Prlr null epithelium...... 91 4.2.6. Elf5 was essential for milk protein synthesis...... 91 4.2.7. Polarisation of alveoli was normal in Elf5 and Prlr null epithelium...... 95 4.2.8. Transcript profiling of Elf5-/- epithelium at days 4 and 6 of pregnancy ...... 98 4.2.9. Forced expression of Elf5 in pubertal mammary glands resulted in differentiation of the terminal end bud...... 104 4.2.10. Forced expression of Elf5 in mature mammary epithelium resulted in alveolar differentiation and milk protein expression...... 107 4.2.11. Retroviral re-expression of Elf5 rescued alveolar morphogenesis in Prlr-/- primary mammary epithelial cells...... 107
4.3 Discussion...... 113
Chapter 5 Prolactin in mammary gland carcinogenesis...... 121
5.1 Introduction...... 121
xvi 5.2 Results...... 125 5.2.1. The latency to tumour formation was increased in C3(1)/SV40T without the Prlr...... 125 5.2.2. The morphology of tumours was not significantly different between Prlr-/- /C3(1)/SV40T and WTC3(1)SV40T mice...... 125 5.2.3. There was a trend towards reduced neoplasia in Prlr-/-/C3(1)/SV40T compared to WTC3(1)SV40T mice...... 129 5.2.4. The mRNA expression of SV40T was unaltered in C3(1)/SV40T mice that lack the Prlr...... 129 5.2.5. Body weight and mammary ductal side branching was reduced in mice that lack the Prlr...... 132 5.2.6. The latency to tumour formation was increased in C3(1)/SV40T transplants without the Prlr...... 135 5.2.7. The mRNA and protein expression of SV40T was unaltered in mammary epithelium without the Prlr...... 135 5.2.8. Loss of epithelial Prlr does not change the histology and morphology of SV40T induced tumours...... 139 5.2.9. SV40T induced neoplasia was delayed in Prlr-/- mammary epithelial cells...... 141 5.2.10. Cellular proliferation in SV40T-induced neoplasia was mediated by Prlr within mammary epithelium...... 144 5.2.11. Apoptosis via activation of Caspase 3 was unaltered by Prlr within mammary epithelium...... 144 5.2.12. Transcript profiling of 8 week Prlr-/-/C3(1)/SV40T transplants to identify genes that may be important for Prl modulated carcinogenesis...... 148
5.3 Discussion...... 154
Chapter 6 General discussion ...... 160
6.1 Overview ...... 160
xvii 6.2 The role of prolactin during the proliferative phase of alveolar morphogenesis ...... 160
6.3 The role of prolactin during lactogenesis...... 164
6.4 The role of prolactin in carcinogenesis ...... 168
6.5 Summary and future directions ...... 172 References ...... 174 Appendices...... 207 Appendix I ...... 208 Appendix II...... 219 Appendix III...... 232 Appendix IV...... 245 Appendix V ...... 262 Appendix VI...... 274 Appendix VII ...... 285 Appendix VIII...... 297 Appendix IX...... 300 Appendix X ...... 316 Appendix XI...... 319 Appendix XII ...... 321
xviii Chapter 1 Introduction
1.1 Preface Breast cancer remains the most commonly diagnosed malignancy in Australian women. The number of new cases of breast cancer per annum increased from 5318 cases in 1983, to 12027 cases in 2002, a real increase in incidence, however the mortality from breast cancer has fallen from 35 deaths/100000 females in 1943 to 23.4 deaths/100000 females in 2004 (The Australian Institute of Health and Welfare Breast cancer in Australia: an overview, 2006 http://www.aihw.gov.au/publications/can/bca06/bca06.pdf). The decline in mortality was first observed in the early 1990’s, and can be attributed, at least in part, to enhanced screening programs like BreastScreen introduced in 1991, and improved therapeutic technologies. In spite of this, there were still a total of 2641 deaths in women as a direct result of breast cancer in 2004, demonstrating the need for further improvements in treatment. One approach is to further our understanding of the mechanisms underlying breast cancer and to use this knowledge as the basis for these improvements. Carcinogenesis involves the dysregulation of genes resulting in uncontrolled cell division, aberrant cell survival, increased angiogenesis and acquired cell motility resulting in invasion and metastasis. These processes resemble normal cell behaviour at various stages of mammary gland development, but multiple epigenetic and genetic events enable or drive a cancer cell to escape normal developmental restrictions. Therefore an understanding of the mechanisms of breast carcinogenesis must begin with an understanding of the cues that drive the formation of a mammary gland. The mouse mammary gland is the primary model utilised for understanding human breast development and carcinogenesis. Mouse mammary gland morphology is comparable to that of the human breast (Cardiff et al., 1999), and with increasingly rapid developments in murine transgenic technologies (Capecchi, 2005), the mouse has proved a valuable system for the investigation of the genes involved in mammary gland development and carcinogenesis. In this thesis, we have utilised mutant mouse models, including knockout mouse models of the prolactin receptor (Prlr) (Ormandy et al., 1997b) and the Ets transcription factor Elf5 (Donnison et al., 2005), the MTB system for mammary specific
1 inducible gene expression (Gunther et al., 2002), and the C3(1)/SV40T model of mouse mammary carcinogenesis (Maroulakou et al., 1994), to further our understanding of prolactin mediated mammary cell proliferation, differentiation and carcinogenesis.
1.2 Cellular morphology The mammary gland consists of the complex network of ducts, formed from a variety of cell types that are, collectively termed the mammary epithelium, which sits inside a mammary fat pad comprised of adipocytes, fibroblasts, and cells of the vasculature and immune systems, collectively termed the stroma. The mammary epithelium can be broadly divided into two compartments, the luminal epithelium and the myo-epithelium (Figure 1.1). The luminal epithelium undergoes differentiation during pregnancy to form the milk secreting lobuloalveoli and the contractile myo-epithelium, which surrounds the ducts and contributes to the delivery of milk during lactation (Hennighausen et al., 2005). Surrounding these two epithelial cell layers is a sheath of basement membrane and this in turn is surrounded by a fibrous layer of extracellular matrix (ECM), sitting in a bed of fibrous connective tissue and adipocytes that contain the vasculature (Figure 1.1). Signalling from the ECM is essential for the development of the mammary epithelium (reviewed in (Fata et al., 2004)). The development of the mammary gland occurs in discrete stages dependent on the hormonal events surrounding reproduction, and achieves its fully functional state only during lactation following a successful pregnancy.
1.3 Embryonic development The mammary epithelium develops from a localised thickening of the ectoderm. A mammary bud emerges from the milk line that forms on both sides of the midventral line of the embryo at around embryonic day (E-) 10.5 (Hovey et al., 2002). At E11.5, five pairs of lens shaped placodes form at locations along the milk line. Between E11.5 and E12.5, bulbous shaped mammary buds emerge from the placodes and push into the underlying mesenchyme. These mammary buds remain quiescent until approximately E15.5-16.5, when they proliferate and push through the mesenchyme into the developing fat pad, where they form a small branched ductal tree immediately adjacent to the nipple. This process
2 Figure 1.1. The structure of a terminal end bud and the cellular compartments of the mammary gland. The mammary gland can be broadly divided into the luminal epithelium and the surrounding myoepithelium. Surrounding these two layers is a sheet of basement membrane, which in turn is surrounded by a fibrous layer of extracellular matrix in a bed of loose connective tissue. The extracellular matrix, fibroblasts, vasculature and adipocytes (blue) are components of the stroma. The terminal end bud is a specialised structure at the ends of extending ducts. It is comprised of an outer layer of undifferentiated cap cells (yellow), which contribute to the formation of the myoepithelium, and inner layers of highly proliferative luminal epithelial cells (purple). An apoptotic plate is present at the rear of luminal cells, which produces the canalisation of the ducts.
3 requires the expression of parathyroid hormone-related protein (PTHrP) in the mammary bud, and the expression of its receptor (PTH1R) in the mammary mesenchyme (Foley et al., 2001; Wysolmerski et al., 1998). PTHrP induces expression of androgen receptor in the basal epidermis (Dunbar et al., 1999). In males, the expression of androgens from the fetal testes, which bind to the androgen receptor in the mammary mesenchyme, results in the complete destruction of the mammary bud. Thus male mice do not develop nipples (Kratochwil et al., 1976). A subset of males show incomplete destruction of the underlying epithelium, which can be found as a small under developed mammary duct in some individuals (Kratochwil, 1977). Once in the mammary fat pad the epithelium undergoes ductal branching morphogenesis, resulting in the development of a rudimentary ductal tree with a small number of branches, which is present at birth (Hens et al., 2005).
1.4 Ductal branching morphogenesis After birth, the mammary anlage continues to grow isometrically, but the hormonal changes that occur during puberty induce allometric growth of the mammary epithelium. Bulbous, highly motile terminal end buds (TEBs) form at the ends of the ducts (Williams et al., 1983). The TEBs consist of an outer layer of undifferentiated cap cells, which contribute to the formation of the myoepithelium, and inner layers of highly proliferative luminal epithelial cells concentrated towards the cap layer (Figure 1.1). There is also a large amount of apoptosis in the body cells of the TEB, which is important in the formation of the hollow ducts enabling milk delivery during lactation (Humphreys et al., 1996). The close proximity of apoptotic cells and mitotic cells suggests that molecules that regulate growth may also induce apoptosis in nearby cells within the TEB (Humphreys, 1999). The spatial and temporal distribution of Transforming growth factor beta (Tgf ) and its inhibition of cellular growth during ductal morphogenesis via epithelial-stromal interactions (Silberstein et al., 1990), suggest that it maybe a key molecule regulating the growth and patterning of the TEB (Daniel et al., 1996). Also, variations in the expression patterns of the extracellular matrix proteins collagen I, IV and laminin are observed in the mammary gland during ductal elongation, and inhibition of 1-integrin and laminin results in impaired TEB formation (Klinowska et al., 1999). Therefore the extracellular matrix may play a part in the regulation of cell growth and apoptosis in the TEB. However, the 4 mechanism underlying the patterning of proliferation and apoptosis within the TEB is not understood. Estrogen (E) (Daniel et al., 1987; Silberstein et al., 1994), growth hormone (GH) (Walden et al., 1998) Egf (Coleman et al., 1990), Tgf (Daniel et al., 1996; Pierce et al., 1993; Silberstein et al., 1990) and Igf1 (Ruan et al., 1992) have been demonstrated to have either stimulatory or inhibitory effects on the TEB during ductal outgrowth. The onset of puberty triggers a rapid increase in proliferation within the TEBs resulting in the extension and bifurcation of the mammary epithelium throughout the mammary fat pad (Figure 1.2). Secondary side branches emerge laterally from the expanding epithelium and shorter tertiary branches form along the ducts, in response to cycling ovarian hormones including Progesterone (Pg). This process continues until the mammary epithelium reaches the periphery of the fat pad and forms a complex network of ducts (Sternlicht, 2006). The stromal environment is crucial for the expansion of the mammary epithelium and is dependent on stromal estrogen receptor (ER) (Cunha et al., 1997), and epidermal growth factor receptor (EGFR) (Wiesen et al., 1999). In addition, the extent of ductal side branching is dependent on the strain of the stroma, as demonstrated by mammary stromal- epithelial recombination. When epithelium from a strain such as C57BL/6, which shows no side branching, is recombined with the stroma of the highly side branched 129 strain, it results in a highly side branched morphology, further demonstrating the importance of the stroma during ductal morphogenesis. (Naylor et al., 2002). Tgf secreted by the epithelium inhibits lateral side branching, and is believed to account for the open mammary architecture of the mammary gland (Hinck et al., 2005). Once the TEBs reach the periphery of the mammary fat pad, the TEBs and lateral buds regress and differentiate into quiescent alveolar buds, which appears to be via an extra-epithelial mechanism requiring prolactin (Prl), Prlr and the signal transducer and activator of transcription 5a (Stat5a) (Brisken et al., 1999; Horseman et al., 1997; Liu et al., 1998).
5 Figure 1.2. Mammary gland remodelling. At birth the mammary epithelium exists a rudimentary ductal tree comprised of only a small number of ducts. The epithelium grows isometrically until puberty when the TEBs undergo a rapid increase in proliferation resulting in ductal elongation and bifurcation, extending the epithelium throughout the fat pad. Ductal side branches continue to sprout from the lateral surfaces of the ducts after maturity resulting in the epithelium completely filling the mammary fat pad. This process is collectively termed ductal morphogenesis. Once the epithelium reaches the periphery of the fat pad, the TEBs regress to form alveolar buds. The onset of pregnancy results in alveolar morphogenesis, and begins with a rapid increase in alveolar proliferation, which continues until mid-pregnancy. This is followed by alveolar differentiation, termed lactogenesis I (secretory initiation) which results in the expression of milk genes in a temporal order. The last phase of alveolar differentiation or Lactogenesis II (secretory activation), results in milk and lipid movement into the alveolar lumens ready for lactation. Milk ejection maintains lactation, and after weaning the mammary gland undergoes rapid programmed cell death (involution), returning the gland to a near-virgin state. 6 1.5 Tissue remodelling during pregnancy The majority of mammary development that occurs during pregnancy is characterised by massive tissue remodelling. During the alveolar morphogenesis phase (Richert et al., 2000), rapid and global proliferation of the epithelial cells occur within the ductal branches and developing alveoli (Figure 1.2). This increases both epithelial cell number and epithelial surface area, actions essential for sufficient milk production during lactation. Cell differentiation becomes dominant from mid pregnancy as the gland moves into the secretory initiation phase (Richert et al., 2000). The developing alveoli cleave and the alveolar cells become polarised and form a sphere-like single layer of epithelial cells that envelops a circular lumen, connected to the ductal network via a single small duct. Each individual alveolus is surrounded by a basket-like architecture of contractile myo- epithelial cells. The myoepithelium of the alveoli is discontinuous so that the luminal cells directly contact the underlying basement membrane, which forms part of the ECM. Some cells of the ductal network also contact the basement membrane. Contact is required for complete lobuloalveolar differentiation (Fata et al., 2004; Streuli et al., 1995), seen morphologically, by the appearance of lipid droplets (Neville et al., 2002) and by the initiation of gene expression in a defined order (Rudolph et al., 2003). Nearing parturition, alveolar tight junctions close and milk and colostrum proteins move into the alveolar lumen, in preparation for active milk secretion post partum, which marks the onset of the secretory activation phase (Nguyen et al., 2001) (Figure 1.3). The epithelial expansion is paralleled by equally dramatic changes in other tissue compartments. Adipocytes lose their lipid content and remain as long projections scattered throughout the alveolar epithelium (Neville et al., 1998). A huge expansion of the vasculature also occurs within the stroma, to provide the large quantities of energy, sugars, amino acids and solutes required for milk production (Djonov et al., 2001).
7 Figure 1.3 Alveolar morphogenesis. Mammary whole mounts (Carmine alum stain top row) and mammary cellular architecture (low power middle row, high power bottom row) in virgin, 12 days posts coitus (dpc), 18dpc and 1 day post partum (1dpp) murine mammary glands. Ductal epithelial cells (arrow) and myoepithelial cells (arrowhead) arise from a common mammary epithelial stem cell. Massive epithelial cell proliferation occurs at the onset of pregnancy, which is co-ordinated predominantly by prolactin and progesterone. At mid pregnancy (12dpc) developing alveoli continue to proliferate and polarise to form a sphere-like single layer of epithelial cells enveloping a circular lumen (X). This is followed by further cell proliferation and differentiation categorised by the expression of milk genes and the formation of cytoplasmic lipid droplets (*). At 18dpc, alveoli have large amounts of lipid and milk protein expression is increased. At parturition, tight junctions between alveolar cells close and milk proteins and lipid are secreted into the alveolar lumen (X). An expansion of the vasculature (open arrows) and reduction in adipocyte (A) area is also apparent in the stroma.
8 Developmental events are also elicited elsewhere in the animal. For example the gut and liver enlarge dramatically to cope with the energy needs of gestation and lactation. The brain is programmed for correct maternal behaviour by Prl (Lucas et al., 1998). Thus mammary alveolar development is part of a larger developmental program of adaptation to pregnancy and lactation. Another striking aspect of tissue remodelling during pregnancy is its cyclical nature. Following weaning nearly all of the development induced by pregnancy is removed by programmed cell death during the involution phase, only to redevelop with the next pregnancy. This observation first led researchers to hypothesise that mammary tissue must contain persistent self-renewing mammary stem cells (reviewed in (Smalley et al., 2003)). The ability of small epithelial transplants to recapitulate a complete and fully functional epithelial mammary gland reinforced this view (DeOme et al., 1959). The presence of single mammary stem cell was indicated by limiting dilution experiments and the existence of committed progenitor cells was demonstrated by transplants that showed limited developmental capacity (Smith et al., 2003). This cell was recently isolated and elegantly demonstrated to be capable of producing a renewable and complete mammary epithelium (Shackleton et al., 2006; Stingl et al., 2006). Thus it is hypothesised, based on a paradigm developed in the hematopoietic system, that a primary mammary epithelial stem cell gives rise to a hierarchy of epithelial progenitor cell lineages that ultimately produce the different cells found in the mammary epithelium (Shackleton et al., 2006; Smalley et al., 2005; Stingl et al., 2006). The flux of cells through these lineages is likely to be controlled by, and in turn control, the patterns of gene expression that comprise the alveolar switch- the mechanism by which prolactin initiates these events. Integrating our knowledge of gene expression patterns with the emerging knowledge regarding stem cell lineages and their interactions offers us an unprecedented opportunity to understand this phase of mammary development.
1.6 The pituitary hormone prolactin PRL is a polypeptide hormone synthesised and secreted primarily by the lactotrophic cells of the anterior pituitary (Stricker et al., 1928). PRL is also synthesized at extra-pituitary sites such as the mammary epithelium, placenta, uterus, brain and the 9 immune system. Evidence for PRL production in the lacrimal gland, adrenal gland, corpus luteum, prostate, testes and pancreas also exists (Reviewed in (Ben-Jonathan et al., 1996; Freeman et al., 2000)). PRL is part of a large family of related hormones, which includes growth hormone (GH) and placental lactogen (PL) (Shome et al., 1977). A proximal promoter region directs PRL expression in the pituitary (Berwaer et al., 1991), while a distal promoter controls extra-pituitary sites of expression (Berwaer et al., 1994). Although the majority of pituitary PRL exists as an unmodified 23kDa protein, a number of variants of PRL have been found in many species which are a result of alternative splicing, proteolytic cleavage, phosphorylation, glycosylation and other post translational modifications (Reviewed in (Sinha, 1995). 1.6.1. Splice Variants There is evidence for alternative splicing of the PRL transcript in the brain (Emanuele et al., 1992), and proteins smaller than 23kDa cross-react with a PRL antibody are present in the pituitary (Sinha et al., 1987b). One cross-reactive variant is 21kDa in size, has a tyrosine peptide map that resembles that of PRL (Sinha et al., 1988) and is present in non-lactating post-partum women (Liu et al., 1990). However, these proteins have not been linked to a transcript and the majority of PRL variants result from post-translational modification. 1.6.2. Proteolytic Variants Proteolytic cleavage of PRL occurs at sites of PRL production (Sinha, 1995), via Cathespsin D, resulting in two fragments of 16kDa and 6kDa (Baldocchi et al., 1993) or via kallikerin giving rise to a 22kD fragment (Anthony et al., 1993). The 16kD fragment has anti-angiogenic properties (Clapp et al., 1993) and inhibits prostate tumour growth possibly by suppressing blood vessel formation (Kim et al., 2003). 1.6.3. Phosphorylated prolactin Phosphorylation of PRL occurs within the secretory granules of the lactotroph and is mediated by p21-activated protein kinase (PAK2) (Tuazon et al., 2002). The function of PRL phosphorylation is widely debated, and it has been shown that phosphorylated PRL (p-PRL) can have both agonistic and antagonistic effects. PRL activity is commonly measured in Nb2 lymphoma cells that proliferate following treatment with lactogenic hormones (Gout et al., 1980; Tanaka et al., 1980). The action of phosphorylated prolactin in
10 this and other cell lines is controversial. p-PRL fails to stimulate proliferation of Nb2 lymphoma cells to the extent of unmodified PRL (Wicks et al., 1995). Inhibition of PRL stimulated proliferation occurs in a dose dependent, antagonistic manner (Wang et al., 1993). Treatment of Nb2 lymphoma cells with a molecular mimic of p-PRL (S179D) also antagonises PRL activity (Chen et al., 1998). Furthermore, S179D inhibits growth in the rat mammary gland (Kuo et al., 2002), and antagonises PRL induced proliferation in breast cancer cell lines (Schroeder et al., 2003). S179D may preferentially activate the mitogen activated protein kinase (Mapk) pathway in HC11 mouse mammary epithelial cells, resulting in increased -Casein expression (Wu et al., 2003). An increase in -Casein expression also occurs in the mammary glands of S179D treated rats (Kuo et al., 2002). However, S179D has also been reported to act as an agonist of Nb2 proliferation and can activate the Jak/Stat pathway, indicating that the action of this PRL mutant is not fully understood (Bernichtein et al., 2001). 1.6.4. Other Post-translational modifications PRL is also modified by glycosylation, deamination, sulfonation, polymeration and formation of complexes with other molecules, although the function of these post- translational modifications is not well understood (Reviewed in (Sinha, 1995)). Glycosylation of PRL has been reported in humans (Lewis et al., 1985), rats (Sinha et al., 1987a) and mice (Bollengier et al., 1989). The ratio of glycosylated PRL to unmodified prolactin is decreased during pregnancy (Markoff et al., 1988), and glycosylated PRL has been reported to down-regulate unmodified PRL (Hoffmann et al., 1993). Thus, like p- PRL, glycosylated PRL appears to play an important role in the regulation and maintenance of unmodified PRL. De-amination of PRL involves the loss of ammonia from asparagine or glutamine residues has been reported to occur in nearly all species including the rat, mouse and human (Graf et al., 1970; Kohmoto, 1975; Sinha, 1995; Sinha et al., 1981). Sufonation of tyrosine residues in Prl has been reported in sheep (Kohli et al., 1988) and buffalo (Chadha et al., 1991) and the function of these post-translationally modified forms of PRL is currently unknown. PRL can also form dimers, polymers and complexes with other molecules (Sinha, 1995; Soong et al., 1982). Importantly, 24% of patients with hyperprolactinemia can be accounted for by the presence of prolactin complexes (macroprolactin), which are believed to have limited bioactivity (Smith et al., 2002).
11 Therefore measurements of hyperprolactinemia are complicated by the presence of macroprolactin (Fahie-Wilson, 2003). PRL can also bind to immunoglobulin, which results in increased proliferation of cells from patients with chronic lymphocytic leukemia (Walker et al., 1995). Thus, careful consideration of the various post-translational modified forms of PRL should be undertaken, when predicting breast cancer risk based upon measurements of circulating PRL (Hankinson et al., 1998).
1.7 The prolactin receptor Prolactin receptor (PRLR) is a member of the class I receptor superfamily of membrane receptors and is closely related to the growth hormone receptor (Boutin et al., 1988; Kelly et al., 1991). In the human, mice and rats, the gene encoding the Prlr transcript is found on chromosome 5, 15 and 2 respectively. Alternative splicing of the PRLR transcript results in the formation of several isoforms, which are referred to as the short, intermediate and long, and these differ in the length of their cytoplasmic tail (Reviewed in (Goffin et al., 1996a; Goffin et al., 1997)). In contrast, the extracellular domain of PRLR isoforms is identical and is highly conserved among species (Bole-Feysot et al., 1998). The function of the various isoforms of PRLR has been hypothesised using various cytoplasmic domain mutants of the PRLR (Lebrun et al., 1995a; Lebrun et al., 1995b). Tyrosine residues present in only the long and intermediate forms of the receptor are required for recruitment and phosphorylation of the kinase Jak2, a major mediator of PRL signalling resulting in mammary epithelial cell differentiation in vivo. The intermediate and long forms of the PRLR have varying proliferative activity and signalling via Jak2 and Fyn in Cho-K1 cell lines, suggesting alternate roles for the various isoforms of PRLR (Kline et al., 1999). The functions of the shorter isoforms of PRLR remain unclear, however both the short and long isoforms are mitogenic to NIH 3T3 cells via stimulation of the Mapk pathway (Das et al., 1995). In addition, the expression of these isoforms varies at different time points of mammary development and during the estrous cycle (Buck et al., 1992; Clarke et al., 1997a). Thus it appears that levels of the various isoforms of the PRLR may play an important role during mammary development. Interestingly, a decreased ratio of the short form to the long form of PRLR has been observed in tumour tissue when compared to adjacent normal tissues, suggesting that modulation of the various isoforms of PRLR may 12 play an important role during carcinogenesis (Meng et al., 2004). The PRLR is activated by PRL, Placental lactogen and GH, which result in the dimerization of the receptor and the activation of various signalling pathways discussed below (Bole-Feysot et al., 1998; Freeman et al., 2000). Mice with a germ line null mutation of Prlr produced by replacing exon 5 with the NEO cassette, was used to investigate the role of the Prlr during mammary gland development (Ormandy et al., 1997b). Exon 5 encodes part of the conserved extracellular domain containing a pair of cysteine residues, which are essential for ligand binding and receptor activation (Rozakis-Adcock et al., 1991; Rozakis-Adcock et al., 1992). Homozygous female mice are infertile and have several reproductive defects including irregular mating, reduced fertilisation rates, degeneration of fertilised embryos within the oviduct, no pseudopregnancy and failure of embryonic implantation (Ormandy et al., 1997a). Ductal elongation and bifurcation was normal in Prlr knockout (Prlr-/-) females, however a failure of ductal side branching was observed. Also, the TEBs persist in the mammary glands of Prlr-/- mice, as they fail to differentiate into the quiescent alveolar buds, despite the mammary epithelium reaching the periphery of the mammary fat pad. During their first pregnancy, approximately two thirds of Prlr heterozygous females on a mixed 129SvPas/Ola or 129Sv/C57BL/6 fail to nurse pups due to reduced lobuloalveolar development. Mammary transplantation was performed to investigate whether Prlrs are required in the epithelium during alveolar morphogenesis (Brisken et al., 1999). Mature Prlr-/- transplants taken in nulliparous animals, displayed normal ductal morphogenesis and regression of TEBs. Pg is reduced in Prlr-/- animals, and Pg pellet administration rescues ductal side branching, suggesting that Prl regulates Pg secretion and this is responsible for the branching morphogenesis defects in Prlr-/- animals (Binart et al., 2000). During pregnancy however, alveolar morphogenesis fails in Prlr-/- mammary epithelium, resulting failed lobuloalveoli development milk secretion (Figure 1.4). Recombination experiments of Prlr-/- stroma and wildtype epithelium, resulted in normal lobuloalveolar development, suggesting that epithelial Prlrs are required for alveolar morphogenesis (Brisken et al., 1999).
13 Figure 1.4. Mammary epithelial transplants without Prlr resulted in failed alveolar morphogenesis. Mammary whole mounts of WT (A-D) and Prlr-/- mammary epithelial transplants (E-H), Carmine stain in virgin animals (A and E) and at 12.5dpc (B and F), 18.5dpc (C and G) and 1dpp (D and H).
14 1.8 Prolactin and progesterone initiation of alveolar morphogenesis. The formation of the milk secreting structures during pregnancy is dependent on a synergy between Prl and Pg signalling (Neville et al., 2002). These hormones trigger an initial wave of cell proliferation during days 2-6 of pregnancy (Traurig, 1967). The progesterone receptor (Pgr) knockout mouse demonstrated that Pg is required for alveolar morphogenesis, and epithelial recombination experiments demonstrated that Pgr in the mammary epithelium, not the stroma, was essential for epithelial cell proliferation (Brisken et al., 1998). Not all mammary epithelial cells express Pgr, and therefore are unable to respond to Pg directly. Mammary gland chimeras made from Pgr+/+ and Pgr-/- mammary epithelial cells (MECs), demonstrated that Pgr-/- epithelial cells proliferate in response to Pg and therefore must respond to a paracrine factor from Pgr+/+ cells (Brisken et al., 2002). Indeed, in the epithelium, proliferating cells segregate with Pgr positive cells (Conneely et al., 2003). This is also true for estrogen receptor (Er) positive cells (Clarke et al., 1997b)}. Further, steroid receptor positive cells are in close proximity to proliferating cells, indicating that proliferation is mediated, at least in part, by a paracrine mechanism. This heterogeneous receptor patterning observed in the luminal epithelium is required for complete lobuloalveolar development (Grimm et al., 2002). Wingless-related MMTV integration site 4 (Wnt4) and Receptor activator of NF- B-ligand (RankL) are targets of the Pgr signalling pathway and maybe the paracrine factors responsible for cellular proliferation in steroid receptor negative cells. Over- expression of the proto-oncogene Wnt1 can rescue pregnancy-induced ductal side branching in Pgr knockout mice, indicating that a Wnt factor may be an important paracrine mediator of Pg-induced ductal side branching during early pregnancy (Brisken et al., 2000). Mammary transplants of Wnt4-/- epithelium have demonstrated that Wnt4 acts in a paracrine fashion to stimulate epithelial ductal side branching during early pregnancy. In these experiments, normal lobuloalveolar proliferation was observed during the later half of pregnancy, indicating that other factors mediating proliferation in late pregnancy may be involved. The RankL target, Nuclear Factor-kappa B (NF- B), is required for Cyclin D1 (Ccnd1) activation via the kinase IkB (IKK ) in neighbouring proliferating cells. Germ 15 line deletion of both RankL and its receptor (Rank) in mice resulted in failed alveolar morphogenesis due to reduced proliferation and increased apoptosis of alveolar epithelial cells. These effects were mediated by Protein kinase B (PKB/Akt), demonstrating that this pathway is essential for the formation of lobuloalveolar structures (Fata et al., 2000). The RankL/NF- B/Ccnd1 pathway is now known to be crucial for the formation of alveolar structures during pregnancy (Cao et al., 2003), and NF- B is essential for Pg driven proliferation within alveoli (Conneely et al., 2003). RankL also co-localises with Pgr’s in response to pregnancy levels of estrogen (E) and Pg, indicating this is an important part of the response. In primary mammary epithelial cell (MECs) cultures, Pg acts in synergy with E to increase Cyclin D1 transcription resulting in increased proliferation (Said et al., 1997). Together these data indicate that Pg may drive the proliferation of neighbouring cells via RankL/NF- B resulting in Cyclin D1 transcription (Figure 1.7 A and B). Pgr consists of 2 isoforms, PgrA and PgrB, which are expressed from a single gene. The PgrB isoform is essential and sufficient for alveolar morphogenesis during pregnancy. Alveoli in PgrB knockout mice fail to develop due to impaired proliferation of the ductal and alveolar compartment, which is possibly mediated via activation of RankL (Cao et al., 2003). Pituitary Prl stimulation of ovarian Pg assists in maintaining the required levels of Pg during early pregnancy (Binart et al., 2000). In addition, up-regulation of Pgr expression by Prl, and Prl receptor (Prlr) expression by Pg, suggests that these hormones may interact in a synergistic manner to control alveolar development. As mentioned earlier, Prlr-/- mice have demonstrated the importance of this receptor during mammary development (Ormandy et al., 1997b). The cooperation of the Pg and Prl signalling pathways is further demonstrated by the similarities of the mammary gland phenotypes observed in Pgr-/- epithelium and Prlr-/- epithelium during alveolar morphogenesis, however the point at which the two signalling pathways interact, is not yet understood. The neuronal peptide galanin (Gal) regulates Prl secretion from the pituitary lactotrophs (Wynick et al., 1993). In addition, the mammary epithelium is responsive to Gal, as it augments alveolar morphogenesis in mammary explants in the presence of Prl (Naylor et al., 2003a). Gal-/- mice show increased levels of the inhibitory phosphorylated form of Prl (Naylor et al., 2005b), and are unable to nurse pups due to failed secretory activation (Wynick et al., 1998). Therefore Gal has dual actions: firstly an indirect role by
16 modulating pituitary Prl and phosphorylated Prl release; and secondly a direct cell autonomous role in the formation of lobuloalveoli during pregnancy. Hormone regulation of alveolar morphogenesis is schematically represented in Figure 1.5. Other hormones can influence alveolar morphogenesis. Growth hormone (Gh) may act in combination with Prl to mediate alveolar proliferation. Gh treatment restores alveolar morphogenesis but inhibits lactation in Prlr+/- mammary glands (Allan et al., 2002). Placental lactogen is released from the placenta during pregnancy and can fully compensate for Prl, allowing alveolar morphogenesis in Prl-/- mice (Horseman, 1999).
1.9 Molecular modulators of Prl induced alveolar morphogenesis Members of the prolactin-signalling pathway are essential for normal alveolar morphogenesis (Hennighausen et al., 2005). Prlr dimerization occurs after Prl binding and leads to the phosphorylation of the associated Janus kinase (Jak2) (Han et al., 1997; Pezet et al., 1997), which in turn phosphorylates specific residues on the Prlr (Lebrun et al., 1995a). Stat5 is then recruited to the receptor and is phosphorylated by Jak2 (DaSilva et al., 1996). Phosphorylated Stat5 is then translocated to the nucleus where it can activate transcription of multiple genes (Wakao et al., 1995) involved a variety of processes during alveolar morphogenesis including establishment of epithelial polarity and cell-cell interactions, stromal epithelial interactions and milk protein expression during lactation (Figure 1.7 C). Both isoforms of Stat5, Stat5a and Stat5b, when knocked out in mice, result in lobuloalveolar defects (Liu et al., 1995; Liu et al., 1997; Teglund et al., 1998). The phenotype is more severe in combined Stat5a/Stat5b knockout animals. One class of genes activated by the prolactin-signalling pathway are the suppressor of cytokine signalling (Socs) members, which act to shut down the prolactin-signalling pathway. Socs1 knockout mice show precocious development during pregnancy, and Socs1+/- mice can restore the lobuloalveolar defects present in Prlr+/- mice due to Prlr haplo-insufficiency (Lindeman et al., 2001). Similarly, loss of Socs2 can also rescue lactation in Prlr+/- females (Harris et al., 2006). The canonical Prl signalling pathways are schematically represented in Figure 1.6. Prl binding to the Prlr, can result in docking of Src homology 2 domain containing transforming protein (Shc) and Growth factor receptor bound protein 2 (Grb2), leading to
17 . Figure 1.5. Hormone control of mammary gland remodelling. At puberty, the terminal end buds respond to ovarian E and the epithelium extends throughout the fat pad. Pg directly influences ductal side branching, and its secretion from the corpus luteum in the ovaries is modulated by pituitary prolactin (Prl). During pregnancy, increases in Pg and P influence alveolar proliferation. Placental lactogen (PL) may take over the functions of Prl during early lactogenesis. Pg withdrawal is required for secretory activation, and further increase in Prl is essential for this stage. Galanin (G) augments secretory activation in the mammary gland. Pituitary Prl secretion is stimulated by Gal, which also regulates its phosphorylation (p-Prl).
18 Figure 1.6. Prl Signalling pathways Prlr is dimerised by Prl binding, resulting phosphorylation of Jak2 which is constitutively associated with the cytoplasmic domain Prlr. Jak2, then recruits Stat5 to the Prlr receptor, where in turn it is phosphorylated. Stat5 then form dimers with other activated Stat5 molecules, and is then translocated to the nucleus, to bind to DNA and activate the transcription of a number of target genes involved in proliferation and differentiation of the mammary epithelium. Our transcript profiling identified a number of potential targets of this process including the transcription factor Elf5. Stat5 initiates transcription of Socs2, which acts as a negative feedback loop to restrict Prl signalling. Prl has also been shown to signal via the PI3K and the MapK pathways.
19 the activation of the Map kinase pathway via the activation of Ras and Raf pathway (Das et al., 1996b; Erwin et al., 1995; Piccoletti et al., 1997). More recently, Prl and TGF over- expression in bi-transgenic mice was demonstrated to synergistically increase Mapk phosphorylation associated with significantly reduced latency of mammary macrocyst formation (Arendt et al., 2006). This study suggests that Prl may interact with other signalling pathways such as the TGF /EGFR pathways via Mapk, which may play an important role during tumour development. In addition, the Src family of kinases can associate with the Prlr, and may result the activation of protein kinase B (PKB) and in turn phosphatidylinositol 3-kinase (PI3K), which may be important in the proliferative and anti- apoptotic effects of Prl signalling during mammary development (Fresno Vara et al., 2001; Tessier et al., 2001). A schematic representation of Prl/Prlr signalling is illustrated in Figure 1.6. Transcript profiling of prolactin receptor knockout mammary glands identified a panel of genes that require Prlr-mediated signalling for increased expression during early pregnancy (Harris et al., 2004; Ormandy et al., 2003) (Figure 1.7 C). Two members of the collagen family and laminin were identified. These molecules are cell adhesion components of the ECM and play an important part of the epithelial-stromal signalling required for full lobuloalveolar differentiation and gene expression (Rudolph et al., 2003; Streuli et al., 1995). Alveolar morphogenesis induced by Prl involves the establishment of polarity and cell-cell communication. The maintenance of cellular polarity is regulated by the closure of tight junctions, and the expression of tight junction proteins Claudin-3 and Claudin-7 was reduced in Prlr-/- mammary transplants. The gap junction protein Connexin 26 was also identified and is involved in the exchange of small ions and metabolites. Recently, Connexin-26 was shown to be important in full lobuloalveolar development and in the prevention of alveolar cellular apoptosis (Bry et al., 2004). Wnt4 was also down regulated in Prlr-/- transplants, indicating that it is potentially a target of Prlr signalling. The downstream target of Wnt, -catenin, has specific actions in both the luminal and myoepithelial compartments of the epithelium, and as a component of cell-cell junctions appears to have a role in signalling to luminal epithelial cells
20 Figure 1.7. Molecular control of alveolar morphogenesis. Signalling from the Pgr and Prlr is essential for alveolar morphogenesis. Increases in serum Pg and Prl result in luminal cell proliferation during early pregnancy, which continues throughout gestation. Heterogenous receptor patterning is essential for complete alveolar morphogenesis (A and B). Tgf- 1 signalling via phosphorylation of Smad results in the transcription of target genes, which act to control proliferation in steroid receptor positive cells (A). Wnt4 and RankL are transcribed in response to Pgr signalling, probably in cooperation with Prl signalling, and appear to stimulate proliferation of neighbouring cells via paracrine mechanisms. RankL binds to its receptor Rank in a neighbouring cell and activates the RankL/NF-kB pathway resulting in Cyclin-D1 transcription and proliferation (B). Wnt4 binds and activates its target -catenin, which has specific roles for both luminal and myo-epithelium for cell fate decisions involving both proliferation and differentiation (B). Prl binds to Prlr and activates the Jak2/Stat5 cascade resulting in the transcription of genes including various transcription factors (TFs) involved in epithelial morphogenesis and branching (Wnt4, A), establishment of epithelial polarity and cell-cell interactions (claudins and connexins), stromal epithelial interactions (collagen and laminin), proteins that regulate it’s own pathway (Socs1/2) and lactation (serotonin and milk proteins) (C). Prl signalling also results in the transcription of Cyclin D1 via an Igf2 dependent mechanism. The ets transcription factor Elf5 is transcribed in response to Prl and can completely compensate for the loss of Prlr signalling. Laminin in the extracellular matrix binds to 1- integrin when contact between the basement membrane and the luminal epithelium is established, and is essential for the maintenance of alveolar cell polarity and differentiation. ErbB4 and its ligands complement Prlr signalling as activation of ErbB4 results in Stat5 phosphorylation and translocation to the nucleus.
21 (Imbert et al., 2001; Teuliere et al., 2005). Indeed, activation of -catenin within the basal epithelial cells results in premature differentiation of the luminal epithelium during pregnancy and persistent proliferation resulting in tumours. These tumours consisted predominantly of undifferentiated basal cells, which were amplified in response to - catenin activation, thereby implicating this molecule in cell fate decisions in the mammary gland (Teuliere et al., 2005). RankL was also identified as a potentially Prl regulated gene (Ormandy et al., 2003; Srivastava et al., 2003). Ccnd1 null mutants exhibit significantly delayed alveolar cell proliferation and impaired lactation, which was shown to be epithelial cell autonomous (Fantl et al., 1999). Interestingly, Prl can induce Ccnd1 expression via induction of Igf2, independent of RankL induction (Brisken et al., 2002). The similarities between Prl and Pg mediated effects on both RankL and Wnt signalling is further evidence of the co-operation of these pathways for alveolar cell proliferation during early pregnancy (Figure 1.7 A). Gene expression profiling of Prl-/- mice has also identified unique targets of mammary development. Expression of tryptophan hydroxylase, the rate-limiting enzyme in serotonin biosynthesis, is increased by prolactin during pregnancy and lactation. Accumulation of serotonin due to milk engorgement experienced during weaning or experimentally via teat sealing, inhibits milk gene expression and can induce involution, providing a mechanism that is put in place by prolactin to stop lactation at weaning. (Matsuda et al., 2004).
1.10 Transcription factors involved in alveolar morphogenesis Prl and Pg and other factors induce the transcription of genes via activation of target transcription factors. These include Stat5 and the steroid hormone receptors as discussed previously, which bind to DNA and result in the transcription of genes involved in many aspects of alveolar morphogenesis. Further, some of these target genes are transcription factors also, which act induce the expression of genes or groups of genes involved in lobuloalveolar development. Some transcription factors which appear to be involved in alveolar morphogenesis include the Homeobox genes, Helix-loop-helix genes, Stats, Tcf/Lef family, NF- B, Cebp family, Nuclear factor family and the Ets transcription factors. The regulation of cellular proliferation during mammary development by the 22 homeobox genes, helix-loop-helix genes, stats and ets transcription factors has been reviewed previously (Coletta et al., 2004). Pg and Prl are hypothesised to influence the expression of -catenin, via induction of the Wnt pathway as discussed earlier. -catenin regulates the activity of the Tcf/Lef family of transcription factors which appear to mediate -catenin signalling and therefore may play a role during alveolar morphogenesis (Hatsell et al., 2003). Inhibition of -catenin results in alveolar apoptosis and greatly reduced milk production capacity. Mice lacking Lef-1 demonstrate a failure to form the alveolar bud at embryonic day 13 (van Genderen et al., 1994). Lef-1 is co-expressed with -catenin, and shows a similar expression pattern in response to PTHrP (Foley et al., 2001). Thus Lef-1 may act to mediate the actions of - catenin, however its effects during alveolar morphogenesis are still unclear. The nuclear factor (NF1) family of transcriptions factors also play a role in functional differentiation as they regulate the transcription of milk protein genes such as WAP, -lactalbumin and -Lactoglobulin (Murtagh et al., 2003). The NF1-C2 isoform member of this family induces the expression of the milk genes carboxyl ester lipase and WAP. Prl regulates the protein expression of NF1-C2 in NmuMG cells, and its expression is reduced in the nucleus of Prlr-/- luminal cells at mid-pregnancy, indicating that NF1-C2 may be regulated by Prl signalling during pregnancy and involved in expression of milk genes in preparation for lactation (Johansson et al., 2005). The helix-loop-helix transcription factors Id1 and Id2 have varying expression in the mammary gland. Id1 expression is increased during early pregnancy, remains low during lactation and rises again at involution. Unlike Id1, Id2 remains high during lactation, indicating that these isoforms have specific functional roles during alveolar morphogenesis (Desprez et al., 2003). Id1 is specifically expressed by the expanding epithelium during the alveolar proliferative phase and is inversely correlated with the expression of -Casein, therefore it appears to be an important factor during early alveolar proliferation. Id1 also regulates Clusterin, a gene involved in the regulation of cell-cell interactions. Additionally, lobuloalveolar development is severely impaired in Id2 knockout mice (Miyoshi et al., 2002; Mori et al., 2000). Reduced proliferation and increased apoptosis has been observed in mammary epithelium lacking Id2, resulting in the failure to form alveolar structures and consequently failure of lactation. Id2 also promotes differentiation in MEC cultures, 23 indicating Id2 is essential for the differentiation of the mammary epithelium (Desprez et al., 2003). The transcription factor NF- B discussed earlier is essential for Pg induced alveolar cell proliferation resulting in Cyclin D1 transcription (Cao et al., 2003; Conneely et al., 2003). NF- B can also induce the transcription many genes involved in the regulation of apoptosis. NF- B levels are induced during pregnancy, decline during lactation and are re- induced during lactation implying a role in mammary gland remodelling. It is also hypothesised that NF- B is an essential “checkpoint” of apoptosis, whose actions are dependent on association with specific transcriptional regulators. Thus NF- Bisan important transcription factor controlling both proliferation and apoptosis in the epithelium during pregnancy (Clarkson et al., 1999). The Cebp family of proteins appear to be important regulators of alveolar morphogenesis. (For review see (Grimm et al., 2003)). Cebp and Cebp isoforms are increased during pregnancy and decline during lactation, indicating that they play a critical role in alveolar morphogenesis, and early milk gene expression. Transplantation experiments have revealed that Cebp is required in epithelial cells for normal lobuloalveolar development during pregnancy, and Cebp knockout mice display phenotypes similar to Pgr, Prlr, Stat5a/b, Ccnd1, Id2 and RankL knockouts. Interestingly Pgr expression was dramatically increased in the mammary glands of C/ebp null mice, and in addition, the expression of Pgr was unusually uniform within the epithelium. These effects were associated with a 10-fold decrease in the rate of proliferation. There was however no change in the expression of Cebp in the mammary glands of Pgr knockout mice, indicating that Cebp is upstream of Pgr, and possibly controls the spatial distribution of epithelial cells, which influence proliferation in alveolar progenitors (Seagroves et al., 2000). Cebp null epithelium significantly increased Tgf and Smad2 signalling, and this pathway is known to inhibit cellular proliferation (Barcellos-Hoff et al., 2000). Cell cycle progression in Cebp null MECs was blocked at the G1/S transition preventing these cells from proliferating in response to early pregnancy levels P and E (Grimm et al., 2005). Therefore, Cebp is essential for controlling cell fate decisions within the mammary gland, including attenuating Pgr expression resulting in mammary epithelial cell differentiation during pregnancy. 24 The expression of the Ets transcription factor sub family Pea3, are elevated at the onset of pregnancy but decline during mid pregnancy to low levels at lactation and involution suggesting a role in early pregnancy induced ductal outgrowth. There are 3 members of the Pea3 subfamily, which are expressed by both the myoepithelium and the luminal cells, however their expression varies during pregnancy suggesting multiple signalling roles during alveolar morphogenesis. The expression of all members of the family remains in the myoepithelium during pregnancy, however the expression of the ER81 member declines in the luminal epithelium 7 days after impregnation. Increased numbers of dividing cells was observed in the terminal end buds of Pea3 knockout mice, and mammary gland transplants of Pea3 knockout epithelium displayed reduced mammary branching during pregnancy suggesting a role for Pea3 in progenitor cell differentiation (Kurpios et al., 2003).
1.11 Other factors involved in alveolar morphogenesis The receptor tyrosine kinase, ErbB (epidermal growth factor) family and their ligands, are important mediators of all aspects of mammary development. There are 4 receptors: Epidermal growth factor receptor (Egfr)/ErbB/Her1, ErbB2/Her2/neu, ErbB3/Her3 and ErbB4/Her4, which are activated by a variety of ligands inducing activation via diamerization and cross phosphorylation. ErbB ligands share a 50 amino acid domain, which is homologous to Egf. Mice expressing a truncated dominant negative allele of ErbB2 did not exhibit a phenotype until late pregnancy where alveoli failed to expand and distend, indicating that ErbB2 is critical for secretory activation (Jones et al., 1999). Conditional deletion of ErbB4 within the mammary gland at pregnancy demonstrated a critical role for this receptor during alveolar morphogenesis (Long et al., 2003). Alveolar expansion was reduced from 13.5 days post coitus in mammary epithelium lacking ErbB4 resulting in incomplete alveolar development and failure to nurse pups due to reduced milk gene expression. Alveolar proliferation was attenuated and Stat5 phosphorylation was abolished. The ErbB4 ligand Neuregulin/Heregulin-1 (Nrg) promotes lobuloalveolar development and the expression of milk genes when used in mammary gland explants (Yang et al., 1995), indicating a role for this ligand in lobuloalveolar development. In addition, mice that lack the alpha form of Nrg show a similar phenotype to ErbB4 knockout 25 with reduced alveolar proliferation and differentiation, demonstrated by reduced -casein expression in reduced alveoli expansion (Li et al., 2002). Other ErbB ligands also appear to have overlapping functions for mammary gland development. Amphiregulin null animals have reduced alveolar development, however the phenotype was much more severe in a triple mutant including knockouts of Tgf and Egf (all ligands of the ErbB family), indicating overlapping and compensatory roles for these ligands during alveolar morphogenesis (Luetteke et al., 1999). Triple mutants developed poorly organised and differentiated alveoli, had reduced milk protein expression and often pups born to these mice did not survive. Amphiregulin loss was also associated with reduced Stat5 phosphorylation. Our transcript profiling experiments demonstrated that amphiregulin was down-regulated in Prlr-/- epithelium (Ormandy et al., 2003), indicating that amphiregulin may be modulated by Prlr signalling. These data together indicate important roles for the ErbB receptors and ligands during alveolar morphogenesis. The overlapping phenotypes observed in Prlr, Pgr and ErbB knockout mice suggest there may be some cross talk between these receptors, which is yet to be fully understood. The cell surface receptor, 1 integrin, which are present on luminal epithelial cells, are essential mediators of ECM signalling via its ligands collagen and laminin (Zutter et al., 1998). Mammary epithelium in mice lacking -integrin in the luminal cells, displayed reduced proliferation and alveolar disorganisation (Li et al., 2005). The focal adhesion kinase (Fak), which is important in protein complexes that connect the ECM to the actin cytoskeleton was also reduced in these mice. Conditional deletion of 1 integrin during early pregnancy and late pregnancy demonstrates that this molecule was important for both the formation of lobuloalveolar structures and for functional differentiation (Naylor et al., 2005a). In these mammary glands, luminal epithelium becomes dissociated from the basement membrane, and cellular polarity is compromised as luminal epithelial cells protrude into the alveolar luminal space. In addition, Prl stimulated milk protein expression via phosphorylation of Stat5 was largely absent in primary mammary epithelial cells lacking 1-integrin, indicating that it is essential for Prl-induced activation of Stat5 (Naylor et al., 2006). The cytokine Tgf 1 is an important regulator of mammary cell proliferation during pregnancy (Barcellos-Hoff et al., 2000). Tgf 1 is restricted to the luminal epithelial cells 26 and can controls cell proliferation via phosphorylation of Smad following Tgf- receptor activation (Massague et al., 2000). Tgf- 1 heterozygote mice display accelerated lobuloalveolar development due to increased proliferation, indicating that the expression of Tgf- 1 restricts alveolar cell proliferation. Epithelial cell proliferation was increased more than 15-fold in Tgf- 1 null ovariectomised animals treated with E and Pg compared to wildtype mice (Ewan et al., 2002). In animals treated with E and Pg, Tgf- 1 expression was restricted to the steroid receptor positive epithelial cells, indicating that Tgf- 1 may play an important role in restricting epithelial cell proliferation in these cells (Ewan et al., 2005).
1.12 The ets transcription factor Elf5 Our transcript profiling experiments identified a number of transcription factors that showed reduced expression in response to a loss of the prolactin receptor, and in combination with profiling of a cell based model of positive prolactin action we identified the ets transcription factor Elf5 (Harris et al., 2006). Ets transcription factors are identified by a highly conserved DNA binding domain (the ets domain), which binds to sites containing a central GGA motif (Sharrocks et al., 1997). Ets transcription factors regulate gene expression during the differentiation of multiple tissues including vascular, lymphoid, muscle and bone (reviewed in (Oikawa et al., 2003). Elf5 (e74-like factor 5 or ESE-2) is an epithelial specific member of the Elf subfamily of Ets transcription factors, and is closely related to the epithelial specific Elf3 (ESE-1) and Ehf (ESE-3) (Oettgen et al., 1999; Zhou et al., 1998). The predicted protein products of mouse Elf5 and human ESE-2 are 95% identical and are expressed as 2 isoforms produced by alternative start sites. Such high conservation of sequence implies similar conservation of function (Zhou et al., 1998). Elf5 is expressed specifically in the luminal cells of mammary tissue (Harris et al., 2006), and its expression is increased dramatically during pregnancy, to levels that far exceed those seen in other tissues. Elf5 can also bind to an ets like domain in the proximal promoter of Whey acidic protein (Wap) and induce its expression independently of lactogenic hormones, indicating that Elf5 may be an important mediator of alveolar differentiation during mid-pregnancy (Thomas et al., 2000). Elf5-/- mice die in utero due to a placentation defect (Donnison et al., 2005). Elf5+/- mice did not lactate due to failed alveolar development, and in some mice where alveoli had formed, differentiation into 27 functional secretory units was severely impaired (Zhou et al., 2005). Mammary epithelial cell proliferation was reduced throughout alveolar morphogenesis and secretory activation, and mammary epithelial transplants demonstrated that this effect was cell autonomous. The levels of Elf5 are reduced in Prlr+/- glands and there is no similar reduction in the expression of Prlr in Elf5+/-, implying that Elf5 is downstream of the Prlr (Zhou et al., 2005). Elf5 therefore appears to be a master-regulator of the alveolar switch required for alveolar morphogenesis. What remains to be elucidated, is how much Prl signalling is controlled by the transcription factor Elf5. One of the specific aims of this thesis is to further understand the role of Elf5 in Prlr signalling.
1.13 Prolactin in mammary carcinogenesis The Nurses Health Study I (http://www.channing.harvard.edu/nhs/) is a large prospective study initiated in 1976. A case-control study conducted using this cohort examined the risk of breast cancer conferred by elevated serum PRL levels. Blood samples were collected between 1989 and 1990, and 306 post-menopausal women were subsequently diagnosed with breast cancer before 1994. These women were matched to 448 control subjects. Measurement of serum PRL demonstrated that top quartile serum PRL conferred a higher relative risk (2.03 fold 95%CI 1.24-3.31 p=0.01) of developing breast cancer compared to women with bottom quartile serum prolactin. (Hankinson et al., 1999). The effect was independent of plasma sex steroid hormones and exclusion of cases diagnosed within 2 years of blood collection resulted in the same conclusion. The cohort was updated with 851 cases diagnosed by 2000, matched to 1275 controls (Tworoger et al., 2004). Overall the same positive correlation between breast cancer risk and serum PRL levels was seen (1.35 fold 95%CI 1.02-1.76 p=0.01), and this association varied by sex hormone receptor status, with ER+ PR+ tumours having an increased relative risk of 1.78 (95% CI, 1.28, 2.50; P-trend < 0.001) compared to ER- tumours (0.76 95% CI, 0.43, 1.32; P-trend=0.28). Rodent cancer models recapitulate the sensitivity of the human breast to PRL (Welsch et al., 1977; Wennbo et al., 2000). Pituitary grafts or transgenic strategies that increase serum Prl levels result in mammary cancer. For example, transgenic mice that over-express human growth hormone, which binds both Prlr and GH receptors, develop 28 mammary carcinoma while mice over expressing the growth hormone receptor-restricted ligand bGH do not (Wennbo et al., 1997). Over-expression of rat Prl using the lipocalin promoter to drive expression predominantly in mammary epithelium produces ER positive tumours at a higher rate than other mouse mammary cancer models (Rose-Hellekant et al., 2003). Prl as a modulator of pre-initiated cancer has been examined in chemical carcinogenesis (Welsch et al., 1975) and transgenic oncogene models. Prl and Prlr mRNA were detected in nitrosomethylurea (NMU) carcinogen-induced tumours and a rat Prl antiserum inhibited NMU-induced tumour cell proliferation by up to 70%, compared to normal rabbit serum and GH antiserum (Mershon et al., 1995). Mice expressing the polyoma middle-T antigen oncogene develop tumours in the first weeks of life, but when crossed with Prl knockout mice they developed tumours significantly later (Vomachka et al., 2000). These, and other experiments, have demonstrated that Prl alone at high levels is sufficient to produce mammary cancer, and that its loss can retard tumour formation in response to an oncogenic stimulus. The mechanisms behind these important observations are not clear. It has generally been assumed that prolactin exerts its effects via direct modulation of the mammary epithelial cell (reviewed in (Clevenger et al., 2003; Vonderhaar, 1999) and there is evidence consistent with this. Thus prolactin receptors (PRLR) are expressed at high levels predominantly by steroid hormone receptor positive breast cancer cells and tumours (Bonneterre et al., 1987; Ormandy et al., 1997c), but at low levels by most tumours (Reynolds et al., 1997). Prolactin causes an increase proliferation (Das et al., 1997; Ginsburg et al., 1995; Llovera et al., 2000) and cyclin D1 expression (Schroeder et al., 2003) in breast cancer cell lines selected for prolactin sensitivity (Schroeder et al., 2002). Likewise the use of PRLR antagonists can reduce proliferation (Chen et al., 1999; Goffin et al., 1996b; Llovera et al., 2000) in some breast cancer cell lines. PRL is also produced by mammary epithelial cells and has been hypothesised to act via an autocrine mechanism (Clevenger et al., 1995; Ginsburg et al., 1995; Reynolds et al., 1997). Transplants of mouse mammary epithelium lacking the Prl gene show normal development during pregnancy, but show a three-fold reduction in cell proliferation at parturition, the time at which PRL production by the epithelium becomes apparent (Naylor et al., 2003b). PRL may also exert a direct effect via the modulation of the sensitivity of the epithelial cell to the action of
29 other hormones. For example exogenous Prl modulates the expression of progesterone receptors by human breast cancer cells (Ormandy et al., 1997c), while endogenous PRL can influence ER alpha levels (Gutzman et al., 2004a). Recent work in mice, however, demonstrates that prolactin also exerts potent indirect effects on the mammary gland via modulation of the systemic endocrine system. Null mutation of Prl or Prlr in mice (Horseman et al., 1997; Ormandy et al., 1997b) results in disruption of ovarian, pituitary and other endocrine systems (Clement-Lacroix et al., 1999). Thus failed mammary ductal side branching during ductal morphogenesis is restored by transplanting Prlr-/- epithelium into the cleared fat pads of hosts with normal endocrine function (Brisken et al., 1999) or by Pg pellet administration (Binart et al., 2000), as discussed earlier. Modulation of the endocrine system by prolactin provides a potential mechanism underlying the results to date in mouse models and the Nurses Health Study. We have scant knowledge regarding the point in the carcinogenic process where prolactin exerts its effect. It may exert an effect as a continuation of its normal developmental role, or acquire a novel role due to dysregulation in cancer. There is evidence for the latter. Prlr expression can be increased in cancer compared to adjacent normal tissue (Touraine et al., 1998). Alternatively altered ratios of a number of different splicing variants and isoforms could modulate signalling from this receptor in cancer (Clevenger et al., 1995).
30 1.14 Aims of this thesis Prolactin controls mammary gland development and influences carcinogenesis but the mechanisms that underlie these actions remain to be defined. This thesis has the overall aim of increasing our understanding of the mechanisms by which prolactin exerts its profound influence over the mammary gland. Alveolar morphogenesis occurs during pregnancy, resulting epithelial cell proliferation and differentiation to form the lobuloalveolar units capable of milk secretion during lactation. The onset of alveolar morphogenesis is triggered by increases in Pg and Prl. Prl has two actions; firstly it modulates the secretion of Pg via Prlrs within the corpus luteum of the ovaries, resulting in mammary proliferation (Binart et al., 2000), and secondly it acts directly on the mammary epithelium to drive the formation of lobuloalveoli and enable milk secretion (Brisken et al., 1999). Prl binding to its receptor within mammary epithelium activates the Jak/Stat5 signalling cascade, resulting in the transcription of genes involved in alveolar morphogenesis and secretory activation (Bole-Feysot et al., 1998). Although the initial signalling events are now well defined, the transcriptional program that executes alveolar morphogenesis and secretory activation is mostly unknown. Elucidating this program was a major aim of this thesis. In chapter 3, we sought to understand the transcriptional program of secretory activation during alveolar morphogenesis using three mouse models of failed lactation. These included loss of a single allele of Prlr, loss of the neuropeptide Gal and treatment with a molecular mimic of phosphorylated Prl. We used Affymetrix GeneChip transcript profiling of all three models to identify a common lactation signature important for secretory activation. We also investigated the unique and shared actions of Prl, Gal and phosphorylated Prl. We have identified the ets transcription factor Elf5 as a Prl regulated gene during pregnancy. Elf5 is required by the epithelium for full lobuloalveolar development (Zhou et al., 2005), however its contribution to Prl mediated alveolar morphogenesis has not been fully investigated. In Chapter 4 a more complete view of Elf5 action is presented using a unique Elf knockout model. We investigated the role of the transcription factor Elf5 during alveolar morphogenesis. We used mammary transplantation and Elf5-/- and Prlr-/- epithelium to understand the contribution of Elf5 in Prl mediated alveolar morphogenesis. We used 31 Affymetrix GeneChip transcript profiling to investigate the transcriptional program modulated by Elf5. By using transgenic mouse technology, we tested whether Elf5 expression in the absence of pregnancy was sufficient for mammary cell differentiation. Finally, we examined whether re-expression of Elf5 in Prlr null epithelium was sufficient to rescue alveolar morphogenesis. The large US Nurses Health study demonstrated that patients with the highest levels of serum PRL, confer a two fold increased relative risk of developing breast cancer (Hankinson et al., 1999). In addition elevated Prl is a complete carcinogen in a number of mouse models (Rose-Hellekant et al., 2003; Welsch et al., 1977; Wennbo et al., 2000). The mechanistic basis for these observations is unknown. Prl is a weak mitogen for some breast cancer cells (Das et al., 1997), and thus a mitogenic role for Prl during mammary carcinogenesis has long been hypothesised as the mechanism of prolactin action, but this hypothesis has not been well tested. Nothing was known regarding the stage of carcinogenesis during which prolactin exerts its effect, and a direct role for prolactin on the mammary epithelium was assumed, but its equally possible indirect effects had not been examined. Thus in Chapter 5 we examine these aspects of prolactin action during carcinogenesis. Firstly we examined whether systemic loss of Prl modulated the formation of tumours in the C3(1)/SV40T model of mouse mammary carcinogenesis. We then used mammary transplantation of Prlr-/-/C3(1)/SV40T to investigate whether Prlr within the mammary epithelium, in the absence of systemic endocrine disruption, was important for Prl modulated carcinogenesis. We also sought to understand the stage at which Prl at its greatest effects during carcinogenesis. And finally, we used Affymetrix GeneChip transcript profiling to identify Prl regulated genes, which may be important in modulating SV40T induced carcinogenesis. The aims of these chapters enabled the overall aim of this thesis to be achieved. We now have an improved understanding of the transcriptional control of secretory activation modulated by Prl, Gal and phosphorylated Prl, as well as the contribution of Elf5 to Prl mediated alveolar morphogenesis. Finally, this thesis makes a significant contribution to the understanding of the role of Prl during mammary carcinogenesis.
32 Chapter 2 Materials and Methods 2.1 Animal husbandry All experiments involving mice were performed under the supervision of and in accordance with the regulations of the Garvan/St. Vincent’s Animal Experimentation Committee. The Prlr-/- mouse was generated as previously described (Ormandy et al., 1997b). The germ line Elf5-/- mouse was generated by insertion of the puromycin cassette into the second exon of the Elf5 gene (Donnison et al., 2005). Placental defects were overcome in the Elf5-/- by injecting Elf5-/- embryonic stem cells into a wildtype tetraploid blastocyst, which was then implanted into the uterus of a pseudo-pregnant foster mother on a 129 background (Unpublished, Peter Pfeffer, Ag Research, NZ). Elf5-/- epithelium was maintained by serial transplantation using techniques described below. The MTB mice were a gift from Lewis Chodosh (Abramson Cancer Centre, PA, USA). Transgenic mice carrying the TetOn-Elf5-IRES-EGFP construct were generated by oocyte injection techniques (Ozgene, WA, Australia) and were on an FVB/N background. The C3(1)/SV40T animals (Maroulakou et al., 1994) were on an inbred FVB/N background, and the core colony was maintained by homozygous matings. Mice heterozygous for both Prlr-/- and C3(1)/SV40T were produced by mating homozygous Prlr-/- males and homozygous C3(1)/SV40T females. Female heterozygous progeny were then back crossed to homozygous Prlr-/- males to produce mice heterozygous or wildtype (WT) for C3(1)/SV40T, and Prlr-/-. Control WT mice were produced by using WT males in an identical but separate scheme to ensure similar genetic diversity between groups (Robertson et al., 2003). Rag1-/- mice (Mombaerts et al., 1992) of the C57BL/6J strain were purchased from Animal Resource Centre, Perth, Australia. All animals were housed with food and water ad libitum with a 12 hr day/night cycle at 22°C and 80% relative humidity. Rag1-/- mice were administered Resprim (Alphapharm, Carole Park, Australia), containing Sulfamethoxazole/Trimethoprim via drinking water (5mg/1mg per 5ml drinking water) in alternate weeks. Mice wildtype for C3(1)/SV40T were weighed weekly and aged to 50 weeks.
33 2.2 Tumor investigation 20 WT/C3(1)/SV40T and 25 Prlr-/-/C3(1)/SV40T mice were observed twice weekly for tumour formation (whole animal study). The date of the first palpable tumour was recorded (age of detection) and tumour size monitored using vernier callipers. The volume of each tumour was estimated using the major and minor axes of palpable tumours (Attia et al., 1966). Mice were sacrificed when the tumour burden reached 10% of the animal’s body weight (ethical end point) or earlier if the animal became unhealthy. At sacrifice tumours were collected for routine haematoxylin and eosin (H&E) histochemistry, and remaining mammary glands were collected for whole mount histology as described below. 2.3 Trypan blue supravital staining Trypan blue injection was performed as previously described (Cappell, 1929). Briefly a 0.5% w/v solution of Trypan Blue powder (Sigma Aldrich, Castle Hill, Aus) in μ distilled H2O was boiled for 1min. The solution was then filtered through a 0.22 m filter (Millipore MA, USA) and stored at -20°C. 500μL of filtered trypan blue solution was injected IP into mice 24hrs prior to sacrifice. 2.4 Mammary epithelial transplantation Mammary epithelium transplants were performed as previously described (DeOme et al., 1959). 2.4.1. Prlr-/- and Elf5-/- study Approximately 1cm3 section of the 4th mammary gland was excised from mature (>10 week) from either WT or knockout donor mice or transplants. Trypan blue injection 24hrs prior to sacrifice was used to identify the location of the mammary epithelium in serial transplants. WT and knockout epithelium was then transplanted in to the alternate cleared 4th mammary fat pads of 4-week-old Rag1-/- mice (Figure 2.1). Host mice were then housed to 12 weeks post-transplant, a time when the mammary epithelium completely fills the mammary fat pad. Host mice were then sacrificed or time mated and mammary transplants harvested at various time points of pregnancy. 2.4.2. C3(1)/SV40T transplantation study Approximately 1cm3 section of the 4th mammary gland was excised from 5-8 week old WT/C3(1)/SV40T and Prlr-/-/C3(1)SV40T donors (prior to the onset of neoplasia) and transplanted into the cleared 4th mammary fat pad of 4-week-old Rag1-/- mice. 2 cohorts 34 Figure 2.1. Mammary transplantation technique 1. (A) Wildtype epithelium (Donor 1) and knockout epithelium (Donor 2) were transplanted into the alternate cleared fat pads of the same 4-week-old Rag1-/- immune-compromised host. Host mice were aged to maturity at 12 weeks of age and time mated (Prlr-/- and Elf5-/- study), or aged to investigate the development of neoplasia and tumor (Prlr-/-/C3(1)/SV40T study). (B-D) An illustration of the cleared mammary fat pad technique. The lymph node, nipple and bridge between the 4th and 5th mammary gland of a 4-week-old Rag1-/- host is cauterised (black arrows). The mammary gland is bisected at approximately 1mm past the superior border of the cauterised (purple ovals) lymph node (C). The caudal portion of the mammary gland is removed and a 1mm3 piece of mammary tissue (blue circle) is inserted into a small pocket inside the cleared mammary fat pad (D). 35 Figure 2.2. Mammary transplantation technique 2. Wildtype epithelium (Donor 1) and knockout epithelium (Donor 2) were transplanted into the cleared fat pads (see Figure 2.1 B-D) of 4-week-old Rag1-/- immune-compromised hosts. Host mice were aged and the development of palpable tumors was investigated (Prlr-/- /C3(1)/SV40T study).
36 were generated. The first consisted of 32 Rag1-/- mice with Prlr-/-/C3(1)/SV40T donor epithelium and 32 mice with WT/C3(1)/SV40T donor epithelium (Figure 2.2). This cohort was investigated for the development of palpable tumours as described above. The second cohort comprised 21 mice with Prlr-/-/C3(1)/SV40T and WT/C3(1)/SV40T epithelium in alternate inguinal fat pads (Figure 2.1). Mammary epithelial transplants were collected at 8, 22 and 32 weeks for whole-mount analysis of early neoplastic lesions. Mice were injected with 5-Bromo-2’-deoxyuridine (Sigma, Germany) dissolved in distilled H2O (100μg/g body weight) two hours prior to sacrifice by CO2 ashphyxiation. 2.5 Cryogenic preservation of mammary epithelium Mammary transplants were collected from Rag1-/- hosts carrying Elf5-/- mammary epithelium in one 4th inguinal fat pad and WT mammary epithelial transplants in the contralateral 4th inguinal fat. Transplants were placed quickly in 65% minimal essential medium (Gibco), 25% Fetal calf serum and 10% DMSO in 1.5ml Cryovials. The cryovials were then placed in a polystyrene container and slow frozen at -80°C overnight. The following day, the cryovials were transferred to liquid nitrogen for long-term storage. 2.6 Elf5 expression in Elf5/MTB transgenic mice Elf5/MTB mice were produced by breeding heterozygous TetOn-Elf5-IRES-GFP mice with MTB heterozygote transgenic mice (Gunther et al., 2002), to produce mice carrying a single allele of MTB and a single allele of TetOn-Elf5-IRES-GFP. Elf5/MTB mice were given 2:50mg/ml doxycycline:sucrose in their drinking water for a period of 3-4 weeks. After a period of 3-4 weeks, mice were sacrificed and the glands were collected and whole-mounted or snap frozen in liquid nitrogen as outlined below. 2.7 Retroviral infection of mammary epithelial cells. Mouse Elf5 cDNA was isolated from mammary gland cDNA by PCR and cloned into the retroviral vector polyPOZ (a gift from Dr. T. Dale (Naylor et al., 2000)). Elf5- IRES-LacZ-polyPOZ and LacZ-polyPOZ ecotropic retroviruses were packaged in Phoenix- Eco cells (a gift of Philip Achacoso and Garry Nolan, Stanford University Medical Center, Stanford, CA) by transient transfection using FuGENE-6 Reagent (Roche). Viral supernatant was harvested by filtration through a 0.45μlm filter. Primary MECS were harvested from mammary glands of 11-13 week-old virgin Prlr-/- mice. Briefly, the 4th inguinal mammary glands were dissected out under sterile conditions, finely chopped, and 37 subjected to three or four rounds of collagenase (10mg/ml) digestion in 2.5% fetal bovine serum/HEPES-buffered RPMI 1640. The purified epithelial cells were plated in DMEM: Ham’s F12 (Gibco) supplemented with 10% fetal bovine serum, 5μg/ml insulin, 10ng/ml epidermal growth factor, 5μg/ml hydrocortisone, and 10ng/ml cholera toxin (all additives from Sigma). Primary MECs were subjected to four rounds of retroviral infection by addition of viral supernatant plus 8μg/ml polybrene. 0.5-1x106 MECs were injected into the cleared 4th inguinal mammary fat pad of a 3-week-old Rag1-/- female mouse prepared as described above. 2.8 Mammary whole-mounting Mammary whole-mounts were made using the Carmine alum technique as described before (Bradbury et al., 1995). Mammary whole mounts were visualised using a Leica MZ12 stereomicroscope and imaged using a Leica DC200 digital camera. Quantitative analysis of neoplasia and tumour was performed using the public domain NIH Image (developed at the U.S. National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih-image/). Briefly, the area of neoplasia and tumour (distinguished using whole-mount gross morphology) was estimated by tracing the perimeter of each lesion manually, using photomicrographs of Carmine alum stained whole-mounts imported into NIH Image software. These areas were then converted into a percentage of total mammary gland area. Mammary whole-mounts were then peeled off the slides and paraffin embedded. 4μm sections were cut for routine haematoxylin and eosin cytochemistry and immune-histochemistry. 2.9 mRNA and protein isolation The fourth inguinal mammary glands was frozen in liquid nitrogen before storage at -80°C prior to use. Total RNA was extracted using TRIZOL Reagent (Gibco/Invitrogen Vic, Australia) according to the manufacturers instructions. 2.10 Cryomounting and sectioning Mammary transplants were harvested and placed on sterile gauze soaked in sterile saline. The transplant was then bisected along the lateral plane. Half of the transplant was mounted on a glass slide and Carmine alum stained as detailed above. The remainder was placed in cold Optimal Cutting Temperature (OCT) mounting medium in a labelled Cryomould and frozen over dry ice. The blocks were then stored at -80°C. 10μm sections 38 were cut using the Leica CM 3050 S cryostat cooled to a temperature of -30°C, and placed on plain glass slides. Sections were stored at -80°C prior to HandE and immune- fluorescence staining. 2.11 HC11 cell differentiation protocol and siRNA HC11 cells were maintained in 10% fetal calf serum (FCS) RPMI 1640 (Gibco/Invitrogen Vic, Australia) containing 10ng/ml EGF. HC11 cells at a concentration of 1x106/2mls 10% FCS/RPMI/EGF were plated in 6 well plates (Day 0) and allowed to grow to 70-80% confluence (Day 1). On Day 1, a 1:8 mix of Lipofectamine 2000 (Invitrogen Vic. Australia):Opti-MEM (Gibco/Invitrogen Vic, Australia) was allowed to sit for 5-10mins at room temperature. Meanwhile 1nmol Stat5 or control EGFP small interfering RNA (siRNA) was added to 200μ l serum free medium. The Lipofectamine/Opti-MEM was then added to the siRNA/serum free medium mix and allowed to form complexes for 15-20mins at room temperature. HC11 cells were then washed twice in serum free medium and 800μl of serum free medium was added to the cells. 200μl of the siRNA complexes were then added to each well. The cells were ° μ incubated for 4hrs in a 37 C5%CO2 incubator after which, 500 l 30% FCS was added to each well. The medium was then replaced by 10% FCS/RPMI/EGF maintenance medium on day 2. On day 3, the medium was changed and replaced with EGF free maintenance medium was added on day 3. On day 4 1 of 1μM dexamethasone and 5μg/ml prolactin was added to each well. This was repeated on each day until day 8 of the protocol. Cells were harvested at days 4, 6 and 8 of cell culture. 2.12 Transcript profiling Total RNA was extracted using Trizol reagent described above. RNA was purified using RNeasy Mini Spin columns (Qiagen, Vic, Australia) and resuspended in RNase free
H2O. RNA quantification and quality control was performed by gel electrophoresis or on the 2100 Bioanalyzer (Agilent, Vic, Australia). For the transcript profiling of failed secretory activation, arrays were formed in duplicate using 4-6 glands per treatment group in two separate replicate experiments. cDNA synthesis was performed using the Superscript II (Invitrogen), and synthesis of biotin-labelled cRNA was performed using the BioArray High Yield RNA Transcript labelling kit (Enzo Life Sciences, Farmingdale, NY). cRNA was then hybridised to Affymetrix MGU74Av2 GeneChips (CA, USA) overnight 39 according to the manufacturers protocol. For the transcript profiling of Elf5-/- and Prlr-/- C3(1)/SV40T transplants, arrays were performed in triplicate using single separate samples per chip. Synthesis of biotin-labelled cRNA was performed using the One-Cycle Target Labelling and Control Reagent kit (Affymetrix), and cRNA was hybridised to Mouse Genome 430 2.0 GeneChip array (Affymetrix). Analysis of microarrays was performed in routines in the language R http://www.R-project.org/ or using the Affymetrix GeneChip version 5 software (MAS 5.0). Data was quantile normalised using the RMA function in the “affy” package in R. ANOVA was carried out using limma package in R (Smyth, 2004) and principal components analysis was carried out on scaled data using the prcomp function in the “stats” package in R. 2.13 Quantitative PCR Single stranded cDNA was produced by reverse transcription using 1ug RNA in a 20uL reaction (Promega WI, USA). Quantitative PCR was performed using LightCycler technology (Roche Diagnostics, Basel, Switzerland). PCR reactions were performed in a 10μl volume with 1μL cDNA, 5pmoles of each primer and FastStart DNA Master SYBR Green I enzyme mix (Roche) as per manufacturers instructions. Name Primer Sequence Ampl Exten Anne Name icon sion aling lengt time Tm h (bp) (s) (°C) Glyceraldehyde- Gapd F TGACATCAAGAAGGTGGTGAAGC 117 5 60 3-phosphate Gapd R AAGGTGGAAGAGTGGGAGTTGCTG dehydrogenase Actin, beta, Actin F GTGGGCCGCTCTAGGCACCA 374 15 60 cytoplasmic ActinR CGGTTGGCCTTAGGGTTCAGGGGGG UDP-galactose Ugalt2 F GGTGGTTGGAATAGAAGAGCACAC 449 18 61 translocator 2 Ugalt2 R CAAGACCGAGACCCAGGAAAAC Folate receptor 1 Folr1 F TGGAGTTGGCGATTAGAGTCTGAC 336 14 61 (adult) Folr1 R GAGGCAGGTGTCTTGGATAAAGTG Sialyltransferase Siat1 F TGTAAAATGGGGGTGACAATCC 422 17 61 1 (beta- Siat1 R CTCTTGCTGACCTCTTGAAGGAAC galactoside alpha- 2,6- sialytransferase) Cytochrome Cyp51 F AAAGGTAATGGGGTCGTGTAGTTG 438 18 60 P450, family 51 Cyp51 R GCACAGAATACGGGCAATGATAC CutA divalent Cuta F TGTCCCAACGAAAAAGTCGC 403 16 62 cation tolerance Cuta R AAAGGCATCAGGAGCAGGAGAG homolog Coatomer protein Copz1 F CAGCACAAGTGGGTTTGGAGTG 379 15 63 complex, subunit Copz1 R TGAGGAGAAGGAACACGGCAAG zeta 1 40 Casein delta Csnd F TATTACCCATCTACCCCCAGCC 210 9 61 Csnd R GAAACCCACAAGCAGACCTAACAC Casein beta Csnb F TTCACCTCCTCTCTTGTCCTCCAC 186 8 63 Csnb R GGGGCATCTGTTTGTGCTTG Extracellular Wdnm1 F TGACAATGACTACTGCCTGGGC 84 4 68 protease inhibitor Wdnm1 R TTCCAAAACTGCGTGGGGGC (Expi) Whey acidic Wap F TGCCTCATCAGCCTCGTTCTTG 308 12 62 protein Wap R CTGGAGCATTCTATCTTCATTGGG Cell death- Cidea F GACTTCCTCGGCTGTCTCAATG 448 18 61 inducing DNA Cidea R GAAACTGATTCGTATCCACGCAG fragmentation factor, alpha subunit-like effector A V-erb-b2 Erbb3 F TCTACCAAGTGGAACAGGAGAGGC 402 16 63 erythroblastic Erbb3 R CACCAACAAACGGAGTCTGGAAG leukemia viral oncogene homolog 3 Aldolase 3, C Aldo3 F TGCCAGTATGTTACAGAGAAGGTCC 390 16 62 isoform Aldo3 R CCGCTTGATAAACTCCTCAGTAGC Stearoyl- Scd2 F GCTGGGGCGAGACTTTTGTAAAC 405 16 64 Coenzyme A Scd2 R TGGCTTCTGGAACAGGAACTGC desaturase 2 Signal transducer Stat5a F CACAGGTGGAAGATTGGGGTTC 219 9 68 and activator of Stat5a R CGAATGGAGAAAAGGGATGGTG transcription 5a Signal transducer Stat5b F GTTCCTCTGCCAGGTAGTCCATAG 197 9 62 and activator of Stat5b R CCACTCCCCATCCAAAAACC transcription 5b Keratin complex Keratin-18 F TGTTCATAGTGGGCACGGATGTC 384 16 69 1, acidic, gene 18 Keratin-18 R CAAGATCATCGAAGACCTGAGGGC Simian virus 40 SV40T F CCTGGAATAGTCACCATG 424 15 55 SV40T R GAGAAAAACACCTATGAATGT Prolactin receptor Prlr F CATGGATACTGGAGTAGATGGGGC 350 14 49 Prlr R TTGCACAGCCACTTCTTCCTCTCC E74-like factor 5 Elf5 F TGGACTCCGTAACCCATAGCACCT 231 10 65 Elf5 R AGGAGATGCAGTTGGCATCAAGCT Estrogen receptor ERa F GACCAGATGGTCAGTGCCTT 185 8 59 alpha ERa R ACTCGAGAAGGTGGACCTGA Transmembrane Tmem56 F GAGTGGGAGGAGATAGCAAGACTG 244 8 61 protein 56 Tmem56 R TTGTGTGCGTGAGGAAACCTG Mitogen activated Map3k5 F GAGACAGAGTTGTGTTAGGGAAGGG 264 11 61 protein kinase Map3k5 R TTGAACGAAGGAGAGCAGAGAGGC kinase kinase 5 CD24a antigen CD24a F AGGAGCCAAAACTGTAAACCCAG 219 9 61 CD24a R TAGAACCAAGCCCCCTTTCAGG Forkhead box A1 Foxa1 F CAGACCTGTAAACTCGTGTGGTGG 485 20 60 Foxa1 R TTCCAGACCCGTGCTAAATACTTC Transmembrane Tmprss2 F CCAGGGAGCACAGTCAAACAAG 476 19 62 protease, serine 2 Tmprss2 R TGAGAAAGGGAAGACCTCGGAC Deleted in liver Dlc1 F TCAGTGAGAAAACCTGTCCGAG 323 13 61 cancer 1 Dlc1 R TTGGAAGAAGAGTTCATCTGGGTC 41 Table 2.1 Quantitative PCR primers Gene name and sequence of forward and reverse primers used for quantitative PCR. The length of the amplicon, extension times (secs) and the annealing temperature (°C) is also tabulated.
2.14 Quantitative PCR analysis Each cycle in the linear phase of the reaction corresponds to a two-fold difference in transcript levels between samples. Reactions were performed in duplicate or triplicate. Each reaction was normalised to a house keeping control, Actin, Gapdh or Keratin 18. Relative quantitation was performed by comparing the crossing points of individual samples. The absolute copy number in each sample was determined by using a standard curve of known concentration. The amplicon for each individual primer set was purified and quantitated. A standard curve was then produced ranging in concentration from 1x101 to 1x107 copies/μl. 1μl of cDNA from unknown samples was then amplified in the same PCR reaction as the standard curve. 2.15 Western blotting Following RNA extraction using TRIZOL reagent, protein was extracted according to the manufacturers instructions. Protein (20μg per lane) was separated using SDS-PAGE (Bio-Rad Laboratories, CA), transferred to polyvinylidine difluoride (Millipore Corp., MA), and blocked overnight with 1% skim milk powder, 50nM sodium phosphate, 50mM NaCl, and 0.1% Tween 20. Membranes were incubated with -Milk (Accurate Chemical and Scientific Corp., Westbury, NY) -Stat5a (Upstate, Westbury, NY) -p-Stat5a (Cell Signalling Technology, Danvers, MA), -MapK (New England Biolabs, Beverley, MA), - p44/42-MapK (Cell Signalling Technology, Danvers, MA), -Akt (New England Biolabs, Beverley, MA), -p-Akt (Cell Signalling Technology, Danvers, MA), -Elf5-N20 (Santa Cruz, CA) -SV40T (Santa Cruz Biotechnology, Santa Cruz, CA) and - -actin (Sigma- Aldrich, Castle Hill, Australia). Specific binding was detected using horseradish peroxidase-conjugated secondary antibodies (Amersham Briosciences, IL) with Chemiluminescence Reagent (PerkinElmer, CT) and Fuji Medical X-ray Film (Fujifilm, Tokyo).
42 2.16 Immunocytochemistry 4μm mammary gland sections were baked at 80°C for 5 mins and placed in Xylene for de-parraffinisation. Antigen retrieval was performed using target retrieval solution low pH (S1699) 20min water bath ( -5-Bromo-2’-dexyuridine (BrdU) clone Bu20a), high pH (S2367) 20min water bath with -Milk (Accurate Chemical and Scientific Corp., Westbury, NY) and high pH (S2367) 30sec pressure cooker ( -cleaved Caspase 3 Asp175, (Cell Signalling Technology, Danvers, MA), -Elf5 (Santa Cruz Biotechnology, Santa Cruz, CA). Slides were blocked in 3% H2O2 and a protein block ( -Cleaved Caspase 3 only) prior to application of 1:12000 -Milk, 1:600 -Elf5, 1:200 anti-BrdU or 1:100 anti- cleaved Caspase 3 primary antibody. Secondary antibody was Envision mouse ( -BrdU), Envision rabbit ( -cleaved Caspase 3 and -Milk) or Link-Label goat ( -Elf5) secondary applied for 30mins. Visualisation was via diaminobenzidine (DAB+). All immune- cytochemistry reagents were purchased from Dako Cytomation (Botany, Australia) unless otherwise stated. All sections were visualised on a Leica DMRB light microscope and imaged using a Leica DC200 camera. 2.17 Immunofluorescence 10μm mammary gland sections were fixed for 2 minutes in 1:1 Acetone:Methanol at -20°C. Sections were then air dried, washed for 2mins in 1X phosphate buffered saline (PBS) and then blocked in 1X PBS containing 1% normal goat serum (PBS/NGS). 50μL primary antibody was applied containing 1:100 -Zo1 (Zymed Laboratories, San Francisco, CA) or 1:200 - -Catenin (BD Biosciences, San Jose, CA) in PBS/NGS under a coverslip at 4°C overnight. The sections were then washed with PBS/NGS for 5 mins followed by 4 washes of PBS only. Secondary antibody 1:500 -mouse Cy2 (Jackson Immuno, West Grove, PA) and -rabbit Cy3 (Jackson Immuno, West Grove, PA) and 1:200 ToPro3 (Molecular Probes, Invitrogen, Mount Waverley, Australia) in PBS was then applied at room temperature for 2 hours. Sections were then washed 4 times with PBS and a final wash of water before cover slipping in Confocal-Matrix (Micro-Tech-Lab, Graz, Austria). Sections were visualized and photographed on a Leica DM RBE TCS (Upright Confocal) SP1.
43 2.18 Statistics The difference between Prlr genotype, Gal genotype and treatment with S179D on the quantitative expression of genes was examine by an unpaired t-test (Statview). The effect of Prlr genotype on the rate of weight gain in whole animals was assessed by the coefficients corresponding to the interactions between genotype and time ( 5, 6, and 7)in the following mixed effects model (Laird et al., 1982) using the nlme package in R (Pinheiro et al., 2005): weightij = 0 +Ui +( 1 +Vi)timeij + 2genoBi + 3genoCi+ 4genoDi + 5time ij genoBi + 6time ij genoCi + 7timeij genoBi + errorij where i indexes animal and j indexes measurement. yij is weight measured in grams, genoA (reference), genoB, genoC, and genoD represent indicator variables for the four genotypes. Time is measured as weeks since birth and 2 2 Ui~N(0, U ) and V i~N(0, V ). The effect of genotype on the rate of increase in tumour volume was assessed using a similar model where yij is the cube root of volume, time is days since the detection of a palpable tumour. The difference between Prlr genotypes in terms of mean SV40T mRNA levels, wholemount analyses, BrdU and cleaved caspase 3 immuno-cytochemistry and tumour morphology was examined by an analysis of variance and the Fisher’s PLSD p-value was used to determined significance (Statview, SAS Institute, NC, USA). The difference between Prlr geneotype in tumour multiplicity was examine by an unpaired t-test (Statview). The effect of Prlr genotype on time until detection of a tumour of a defined size, or attainment of ethical end point, was determined by Kaplan Meier survival analysis (logrank statistic) (Statview). The effect of genotype on cellular structure in tumours was determined by a Chi Square statistic (Statview). In all analyses, a p<0.05 corresponded to statistical significance.
44 Chapter 3 Transcriptional regulation of mammary secretory activation 3.1 Introduction Alveolar morphogenesis results in the formation of the basic architecture of the secretory mammary gland. During the first half of pregnancy, rapid epithelial cell proliferation occurs predominantly in response to Prl and Pg (Brisken, 2002). This process results in an increase in the epithelial surface area, and the formation of the alveolar secretory units, required for milk production during parturition (Richert et al., 2000). Mammary differentiation or lactogenesis begins around mid-pregnancy and the now formed alveoli become predominantly composed of milk-producing cells. Differentiation of these cells is marked by the presence of cytoplasmic lipid droplets within the cytoplasm (Neville et al., 2002). Lactogenesis occurs in 2 distinct stages (Hartmann, 1973), described as the secretory initiation and activation phases of mammary development. Secretory initiation, the first phase of lactogenesis results in the temporal expression of a number of milk genes, the timing of which is important for terminal differentiation (Robinson et al., 1995). The milk genes extra-cellular proteinase inhibitor (Wdnm-1) and - casein are expressed earliest, with marked increase in their expression first observed at 9.5 and 11.5dpc respectively. This is followed by expression of Whey acidic protein (Wap) and alpha-lactalbumin (Lalba) at 16.5dpc. Transcript profiling of various stages of mammary gland development recapitulated these observations (Rudolph et al., 2003). Furthermore, the expression of milk genes measured by Affymetrix MAS5 software increases up to 20 fold over that present in virgin animals, reaching maximal levels during lactation. However, these changes may be a gross underestimate as the effects of probe saturation by highly expressed genes is significant. We have also demonstrated that expression changes measured by Affymetrix transcript profiling often are conservative measurements of the real biological differences in gene expression (Ormandy et al., 2003; Robertson et al., 2003). The second phase of lactogenesis, or secretory activation occurs around parturition and is characterised by the closure of alveolar tight junctions and the movement of lipid and milk into the alveolar lumina (Neville et al., 2002). This process is dependent on 45 progesterone withdrawal and elevated levels of Prl (Deis et al., 1983; Nishikawa et al., 1994). As suckling begins, further expansion of the alveolar epithelium occurs, and milk protein expression increases. Copious milk secretion during lactation is dependent on Prl mediated milk secretion and oxytocin production, enabling milk ejection (Neville et al., 2002). Although the hormonal regulation of secretory activation has been well described, the global gene expression changes that drive these processes are not well understood. In this chapter we have used three models of failed lactation, in order to identify novel genes that are essential for secretory activation. These three models include mice that lack an allele of the Prlr, mice that lack the neuropeptide Galanin (Gal), and mice treated with the a synthetic form of phosphorylated Prl. Prlr heterozygote (Prlr+/-) mice on a mixed 129SvPas/Ola or 129Sv/C57BL6 background failed to nurse pups on their first pregnancy (Brisken et al., 1999; Ormandy et al., 1997a). The development of lobuloalveoli in these animals was impaired, resulting in insufficient milk production, due reduced expression of the Prlr (Binart et al., 2003). Therefore this model was selected to identify genes that were changing in failed lactation due to haplo-insufficiency of Prlr. Galanin is an autocrine/paracrine growth factor for the Prl-secreting pituitary lactotrophic cells (Kaplan et al., 1988; Wynick et al., 1993), and as a result Gal-/- mice display a failure of estrogen-induced lactotroph proliferation and reduced serum Prl levels. Thus, Gal-/- mice failed to lactate and were unable to nurse pups (Wynick et al., 1998). Gal also modulated the action of Prl during pregnancy via a mammary epithelial cell autonomous mechanism (Naylor et al., 2003a), as mammary explants co-cultured with Gal and Prl produce more and larger lobuloalveoli, compared to Prl treatment alone. This model was selected to identify the genes regulated by Gal and it’s actions in synergy with Prl, which were responsible for failed secretory activation in the absence of Gal. Several post-translational modifications of Prl exist, which may be involved in the regulation of Prl function (Sinha, 1995). Phosphorylation of Prl appears to be the most important (Ho et al., 1993; Oetting et al., 1986). Treatment of rats with a molecular mimic of phosphorylated Prl (S179D) inhibits lobuloalveolar development during pregnancy, but results in increased -casein expression (Kuo et al., 2002), a counter-intuitive finding
46 interpreted as evidence for a mechanism involving forced precocious differentiation preventing full lobuloalveolar development. In an in vitro model S179D can stimulate proliferation and activate the Jak2/Stat5 and Mapk pathways (Bernichtein et al., 2001). In contrast, S179D can antagonise the actions unmodified prolactin in MCF-7 breast cancer cells, resulting in reduced proliferation, Cyclin D1 levels and phosphorylation of Stat5 and Mapk (Schroeder et al., 2003). Thus the actions of S179D are controversial. Work in our laboratory has shown that mice treated with S179D via intra-scapular mini-osmotic pump, resulted in failed lobuloalveolar development and reduced -casein expression (Naylor et al., 2005b), thus in our hands S179D acted predominantly as an antagonist. Seventy five percent of mice treated with S179D failed to nurse pups associated with reduced lobuloalveolar development and milk protein expression (Naylor et al., 2005b). Gal also regulates p-Prl, as the ratio of p-Prl to unmodified Prl is increased in Gal-/- mice (Naylor et al., 2005b). Thus, transcript profiling of mice treated with S179D, would enable us to investigate the genes differentially expressed as a result of antagonism of the Prl pathway, and also investigate the transcriptional regulation of p-Prl in Gal mediated mammary gland development. The aim of this chapter was to investigate the changes in gene expression among these three mouse models of failed lactation, to in order to understand the transcriptional regulation of secretory activation.
47 3.2 Results 3.2.1. Stat5 phosphorylation was decreased in Prlr+/- mice that are unable to lactate. Two thirds of Prlr+/- mice on a mixed 129SvPas/Ola or 129Sv/C57BL/6 background fail to lactate on their first pregnancy (Brisken et al., 1999), a phenotype, that can be rescued by increasing the expression of the short form of Prlr (Binart et al., 2003). Lobuloalveolar development in Prlr+/- mice is variously retarded in individual animals (Figure 3.1). Some mammary glands taken at parturition from Prlr+/- mice, display a phenotype that is representative of days 12-14 of pregnancy (Figure 3.1 C) associated with lactational failure, while others develop mammary glands with more advanced lobuloalveoli (Figure 3.1 D) and show variable degrees of lactation. Thus, the phenotype is partially penetrant and is presumably dependent on allelic variants or haplotypes that segregate in a mixed genetic background, as breeding onto a pure C57BL/6 results in complete penetrance (Hennighausen L, Lindeman GJ personal communication). We assessed the activation of Stat5, the major mediator of Prl signalling during alveolar morphogenesis (Bole-Feysot et al., 1998), using an antibody against phosphorylated Stat5 (pStat5). Stat5 phosphorylation was decreased in Prlr+/ mice that were unable to lactate (Figure 3.2 C No lact.) compared to WT animals (Figure 3.2 C Full lact.), with no similar decrease in the levels of total Stat5. The levels of pStat5 were variable in Prlr+/- mice capable of lactation (Figure 3.2 C Partial Lact.), however these levels were above those in mice that exhibited lactation failure. A similar reduction in Stat5 phosphorylation was also observed in mice treated with S179D (Figure 3.2 A). Gal can also directly regulate the Stat5 pathway, as it can induce phosphorylation of Stat5 in mammary explants treated with Gal (Naylor et al., 2003a). We found no alterations in the MapK and PI3K signalling pathways with loss of a single copy of Prlr (Figure 3.2 C) and treatment with S179D (S179D westerns performed by Dr Matthew Naylor) (Figure 3.2 B). These results indicated that failed lactation observed in Prlr+/ mice and mice treated with S179D, was probably via a mechanism requiring phosphorylation of Stat5.
48 Figure 3.1. Lobuloalveolar development in Prlr+/- mice. Mammary whole mounts of WT (A and B) and Prlr+/- mice (C and D) at 1dpp, carmine stain. The phenotype of lobuloalveoli is variable in Prlr+/- mice, and displays lobuloalveolar development from that which resembles mid-pregnancy (C) to more advanced lobules, which are characteristic of late pregnancy (D).
49 Figure 3.2. Phosphorylation of Stat5 was reduced in Prlr+/- that failed to lactate. Western analysis of Stat5 phosphorylation and Mapk and PI3K signalling pathways in Gal+/+ mice (WT), Gal+/+ mice treated with S179D (WT + S179D, A and B), Prlr+/+ mice and Prlr+/- mice that lactated successfully (Partial lact) and failed to lactate (No lact, C).
50 3.2.2. Transcript profiling of three models of failed lactation identified a lactation signature comprised of 35 key genes. To understand the transcriptional regulation as a result of failed secretory activation, we compared the expression profiles between Gal-/- mice and Gal wildtype (Gal+/+) controls (denoted – Gal), Gal+/+ mice treated with S179D and saline treated controls (denoted + S179D), and Prlr+/- mice compared to Prlr WT (Prlr+/-) controls (denoted – Prlr). Pooled RNA from 4-6 replicates from each experimental group was hybridised to MGU74Av2Affymetrix GeneChips. (Array hybridisation of Gal and S179D experimental groups was perfomed by Dr Matthew Naylor). The experiment was performed in duplicate. To identify differentially regulated genes within and between groups, we undertook a Venn analysis of gene expression (Figure 3.3 A). Of the 12488 total probe sets on each GeneChip, 7278 genes were absent in the mammary glands of all three models of failed secretory activation (Figure 3.3 A dark orange). Of the remaining 5210 detectable genes, 939 genes were differentially expressed, in at least 1 of the models of failed secretory activation. Figure 3.3 A illustrates the number of increasing genes (I) and decreasing genes (D) for each subset of – Prlr (pink), - Gal (yellow) and + S179D (blue). The number of genes that increased and decreased in the intersections between these models is also shown for –Prlr and - Gal (Figure 3.3 B), - Prlr and + S179D (Figure 3.3 C) and + S179D and – Gal (Figure 3.3 D). The false discovery rate for each subset, determined by permutation analysis is denoted in brackets (Figure 3.3 A). The false discovery rate of the – Gal set and the + S179D set was 80% and the intersection of these two experiments was 30%. Thus genes differentially expressed in the – Gal and + S179D sets, independent of the – Prlr set could be considered biologically irrelevant. In contrast, the false discovery rate in the – Prlr set was estimated at only 21%. In addition the false discovery rates of the intersection between the – Prlr set and the – Gal or + S179D was 14% and 19% respectively. Therefore any gene expression changes observed – Prlr sets could be considered biologically relevant. This analysis indicated that the function of Gal in this model was predominantly via modulation of Prl action, and also, that S179D has virtually no off target effects.
51 Figure 3.3. Venn analysis identified differentially regulated genes in three models of failed lactation. Venn analysis was performed on the gene expression changes in mammary glands at 1dpp as a result of loss of an allele of Prlr (- Prlr), loss of Galanin (- Gal) and treatment with S179D (+ S179D, A). The number of genes increasing (I) and decreasing (D) is represented, and the false discovery rates are indicated in the brackets. The number of increasing and decreasing genes for the intersections between – Prlr and – Gal (B), - Prlr and + S179D (C) and + S179D and – Gal (D) is represented.
52 It was clear from this analysis that of these three models of failed lactation, loss of Prlr resulted in the greatest gene expression changes during secretory activation. 654 out of a total of 939 differentially expressed genes were changing with loss of an allele of Prlr. The majority of gene expression changes (~ 70%) were in the negative direction, suggesting that the bulk of Prlr signalling results in activation of transcription. The expression of most of these genes (602) was not influenced by S179D, thus the overall effects of S179D or phosphorylated Prl were not via modulation of Prl action. 109 genes were found at the intersection of the – Gal and –Prlr set (Figure 3.3 orange and green), which were indicative of Gal control of pituitary Prl secretion (Wynick et al., 1998), and Gal modulation of Prl action within the mammary gland (Naylor et al., 2003a). Treatment with S179D resulted in changes in gene expression that were common with both – Gal and – Prlr (35 genes, Figure 3.3 white) and with the – Prlr independent of – Gal (52 genes, Figure 3.3 purple). Interestingly, of these sets that were common to both + S179D and – Prl (87 genes comprised of a set containing 52 and 35 genes), 84 of the 87 genes changes in the same direction. This indicated that in respect to these genes, S179D acted predominantly as an antagonist of Prl action during secretory activation. Importantly, we identified 35 genes that were differentially expressed and common to all three models of failed lactation. We denoted this set as the lactation signature, and used these genes in further analysis.
3.2.3. Quantitative PCR validated the transcript profiling results obtained from the three models of failed lactation. We then selected nine genes within the lactation signature, and used quantitative PCR to validate our transcript profiling analysis (Figure 3.4). These experiments were performed in triplicate. In all cases, the Affymetrix increasing or decreasing calls (Figure 3.4 grey bars) were confirmed by quantitative PCR (Figure 3.4 black bars). In addition, the estimates of fold change using Affymetrix analysis, was a conservative estimate of what we observed using quantitative PCR. Interestingly, of the 9 genes we selected for quantitative PCR
53 Figure 3.4. Quantitative PCR confirmed differential expression of the genes identified in the lactation signature. Bar graphs of the estimates of fold change as a result of loss of a single allele of Prlr (- Prlr), loss of Gal (- Gal), and treatment with S179D (+S179D) from Affymetrix transcript profiling (grey bars) and quantitative PCR (black bars). Quantitative PCR of 9 genes within the lactation signature confirmed the gene expression changes observed with Affymetrix transcript profiling.
54 validation, the magnitude and fold change of these genes was comparable between all three models of failed lactation. Therefore similarities in the magnitude of gene expression suggested that lactation failure as a result of loss of a single allele of Prlr, loss of Gal and treatment with S179D, occurred at a similar stage of secretory activation.
3.2.4. Failed secretory activation occurred at the same stage in all three models of failed lactation. Trajectory clustering of gene expression throughout mammary gland development, identified groups of genes, which were temporally regulated during alveolar morphogenesis (Rudolph et al., 2003). The milk genes Wdnm1 and the caseins are expressed from the earliest time points of pregnancy, followed by Wap from mid-pregnancy and subsequently -Casein and Lalba during the later time points of pregnancy. Milk protein expression reaches maximal levels in the first few days of lactation. In contrast, the genes encoding the lipid biosynthetic machinery, remain low during pregnancy, and rise sharply during early lactation (Rudolph et al., 2003). We investigated the behaviour of the milk protein genes (Figure 3.5) and the lipid synthesis genes (Figure 3.6) identified in this study, using our data sets of failed secretory activation. The first observation we made was that all three models have proceeded through secretory activation, as the all milk proteins (Figure 3.5) and the lipid synthesis genes (Figure 3.6) were expressed. Thus, failure to nurse pups in the three models of failed lactation, was due to insufficient gene expression resulting in reduced lactation. These findings are consistent with the morphology observed in these models, as in all cases lobuloalveolar development occured, but stalled at a premature stage of lactogenesis. Interestingly, the magnitude of decreased milk protein and lipid synthesis gene expression, with the exception of -Casein, was highly similar among all of the models. Therefore loss of an allele of Prlr, knockout of Gal and treatment with S179D results in developmental arrest at the same stage of secretory activation.
3.2.5. The lactation signature contained a small number of genes important for milk and lipid synthesis. Using stringent analysis we identified a lactation signature comprised of 35 genes differentially expressed in all three models of failed secretory activation. Hierarchical
55 Figure 3.5. Milk protein gene expression was similarly decreased in all three models of failed secretory activation. The expression of the milk protein genes identified in (Rudolph et al., 2003), expressed a fraction of control levels in mammary glands at 1dpp with loss of a single allele of Prlr, loss of Gal and treatment with S179D. All milk proteins are expressed in each of the failed models of lactation, and with the exception of -Casein are reduced to similar levels.
56 Figure 3.6. Lipid synthesis genes were all expressed and were similarly decreased in all three models of failed secretory activation. The expression of the lipid synthesis genes identified in (Rudolph et al., 2003), expressed a fraction of control levels in mammary glands at 1dpp with loss of a single allele of Prlr, loss of Gal and treatment with S179D. All milk proteins are expressed in each of the failed models of lactation, and are reduced to similar levels.
57 clustering clearly divided the mice that failed to nurse pups (Figure 3.7 failed lactation) from those that were able to lactate sufficiently (Figure 3.7 lactation). Using comprehensive literature searches, we discovered that the functions of most of these genes were metabolic processes involved in milk production, such as synthesis of triacylglycerols and cholesterol from glucose, transport of fatty acids fro the circulation, and lactose synthesis. This discovery was proof of principal that we had had identified genes involved in milk delivery, however many of the genes identified in the lactation signature had not previously been associated with lactogenesis. The formation of cytoplasmic lipid droplets is characteristic of lactogenesis (Neville et al., 2002), thus it is no surprise that genes involved in the synthesis of lipid were identified in the lactation signature. An example of such a gene is Cidea, a gene involved in energy balance and adiposity (Zhou et al., 2003). Cidea knockout mice have an increased metabolic rate, and lipolysis of brown adipose tissue, and are resistant to diet induced obesity and diabetes. Therefore, Cidea is a key gene involved in metabolic processes, and our transcript profiling places Cidea as a potential regulator of secretory activation in the mammary gland. Two enzymes involved in the lipid synthesis pathways were also identified, and these were Stearoyl-CoA desaturase 2 (Scd2) and Aldolase 3, C isoform (Aldo3). Scd2 is a key enzyme involved in the production of mono-unsaturated fatty acids (Enoch et al., 1976). Scd-2 knockout mice have reduced triglyceride levels in the plasma, skin and liver during embryonic development, and die within 24hrs of birth due to skin permeability barrier function as a result of impaired lipid synthesis (Miyazaki et al., 2005). Aldo3 is a key enzyme involved in glucose metabolism and is involved in the conversion glucose to pyruvate (Villar-Palasi et al., 1970). In addition to Scd2 and Aldo3, we identified the ATP citrate lyase (Acly) (Beigneux et al., 2004), the lipoprotein receptor (Ldlr), farnesyl diphosphate synthase (Fdps) (Welch et al., 1996), Squalene epoxidase (Sqle) and elongation of very long chain fatty acids 5 (Elovl5) (Leonard et al., 2002), all genes involved in the synthesis of fatty acids. Therefore, down regulation of the genes in our three models of failed lactation, suggests a role in lipid metabolism during secretory activation. Interestingly, we also identified several key transcription factors in the lactational signature. The Sterol regulatory element-binding protein 1 (Srebp1/Srebf1) is a member of
58 a large family of transcription factors involved in the synthesis and uptake of cholesterol, fatty acids, triglycerides and phospholipids (Horton et al., 2002). Srebf1 regulates the transcription of a large number of genes involved in the synthesis of monounsaturated and saturated fatty acids, tryacylglycerides and phospholipids. Our transcript profiling suggests that is also a key enzyme for the production of lipids during lactogenesis. The transcription factors CCAAT enhancer-binding protein delta (Cebp) and the Sry-related gene a4 (Sox4) were up regulated in the three models of failed lactation. Cebp has a role in controlling the growth arrest and apoptosis in the mammary gland and is activated by Stat3 (Gigliotti et al., 2003; Hutt et al., 2000). The role of Sox4 is not clear in the breast, but its transcription can be induced by progesterone (Graham et al., 1999; McGowan et al., 1999). The up regulation of these genes in our models of failed lactation, suggest that their loss is required for secretory activation. In addition, the identification of transcription factors in the lactation signature, indicates that our transcript profiling has been successful in identifying novel processes driven by transcription factors, necessary for secretory activation. Table 3.1 summarises the 35 genes identified in the lactation signature, and sample gene ontology functions are given for each. Probe Set ID Genebank Title Gene Gene Ontology accession # Symbol 100091_at D87990 UDP-galactose translocator 2 Ugalt2 UDP-galactose transporter activity 101019_at U74683 Cathepsin C Ctsc Cysteine-type endopeptidase activity] IEA 102114_f_at AI326963 Angiopoietin-like 4 Angptl4 Positive regulation of lipid metabolism 102335_at AF033017 Potassium channel, subfamily Kcnk1 Potassium ion transport K, member 1 103202_at AW047476 Guanylate nucleotide binding Gbp3 GTPase activity protein 3 104477_at AW047643 Mus musculus transcribed EST Unknown sequences 160109_at X70298 SRY-box containing gene 4 Sox4 Transcription factor 160207_at AW121639 ATP citrate lyase Acly Acetyl-CoA biosynthesis 160424_f_at AI846851 Farnesyl diphosphate synthetase Fdps Cholesterol biosynthesis 160469_at M62470 Thrombospondin 1 Thbs1 Carboxypeptidase A activity 160546_at AW121134 Aldolase 3, C isoform Aldo3 Glycolysis 160549_at M23568 TPA regulated locus Tparl Unknown 160832_at Z19521 Low density lipoprotein Ldlr Cholesterol metabolism receptor 160894_at X61800 CCAAT/enhancer binding Cebpd Transcription factor 59 protein (C/EBP), delta 92642_at M25944 Carbonic anhydrase 2 Car2 Lyase activity 93085_at D44456 Proteosome (prosome, Psmb9 Ubiquitin-dependent macropain) subunit, beta type 9 protein catabolism (large multifunctional protease 2) 93264_at AI843895 Sterol regulatory element Srebf1 Lipid metabolism binding factor 1 93294_at M70642 Connective tissue growth factor Ctgf Regulation of cell growth 93496_at AI852098 ELOVL family member 5, Elovl5 Lipid metabolism elongation of long chain fatty acids (yeast) 93785_at M64782 Folate receptor 1 (adult) Folr1 Folic acid transport 94322_at D42048 Squalene epoxidase Sqle Sterol biosynthesis 94432_at AI117157 Sialyltransferase 1 (beta- Siat1 Protein amino acid galactoside alpha-2,6- glycosylation sialyltransferase) 94916_at AW122260 Cytochrome P450, 51 Cyp51 Cholesterol biosynthesis 95114_s_at AI844999 Divalent cation tollerence Cuta Unknown 95149_at AW121088 Coatomer protein complex, Copz1 Protein transporter subunit zeta 1 activity 95758_at M26270 Stearoyl-Coenzyme A Scd2 Lipid metabolism desaturase 2 96119_s_at AA797604 Angiopoietin-like 4 Angptl4 Positive regulation of lipid metabolism 96771_at AI006228 v-erb-b2 erythroblastic Erbb3 Signal transduction leukemia viral oncogene homolog 3 (avian) 97442_at AW124340 Solute carrier 39 8 Slc39a8 Ion transport 97531_at X95280 G0/G1 switch gene 2 G0s2 Signal transduction 98814_at V00740 Casein delta Csnd Milk protein 98822_at X56602 Interferon, alpha-inducible G1p2 Iron ion transporter protein activity 98994_at AF081499 Solute carrier family 34 (sodium Slc34a2 Ion transport phosphate), member 2 99098_at AW045533 Farnesyl diphosphate synthetase Fdps Cholesterol biosynthesis 99994_at AF041376 Cell death-inducing DNA Cidea Lipid metabolism fragmentation factor, alpha subunit-like effector A Table 3.1: The 35 genes identified in the lactation signature. The 35 genes commonly changing in Prlr+/- mice, Gal-/- mice and mice treated with S179D, which failed to lactate at 1dpp. Affymetrix ID (Probe set ID), Genebank accession number, Title, Gene Symbol and Gene Ontology are tabulated.
60 Figure 3.7. Hierarchical clustering of the 35 genes in the key lactation signature. Heat map and dendrogram of the 35 genes commonly changed in all three models of failed secretory activation. Low to high expression is represented in greyscale from white to black. Hierarchical clustering of these 35 genes clearly identified the models of failed secretory activation (Prlr+/- no lactation (NL), Gal-/- and S179D from those models that successfully nursed pups (Prlr+/- lactation (L), Gal+/+ and Prlr+/+. Treatment of the Gal-/- mouse resulted in a successfully rescue of lactation in one of the experimental replicates.
61 3.2.6. Rescued lactation in Prlr+/- mice was associated with increased expression of genes involved in proliferation. In order to gain insight into the partial penetrance of the Prlr+/- phenotype, we investigated the changes in gene expression that occurred between Prlr+/- that were unable to lactate and those that lactated sufficiently for pup survival. We wanted to investigate the ability of the segregating genetic elements to rescue lactation, which is observed in 25% of the mice in a mixed 129 background. Of the top 100 genes ranked by their p-value, 25% of these genes represented genes involved DNA replication. An example is Orc6, a gene involved in the origin recognition complex, which selects sites in the genome for the initiation of replication (DePamphilis, 2003). At this location, a pre-replicative complex is formed from the ordered assembly of genes including Cdc6, and Mcmds 2-7. This complex is an important step in co-ordinating DNA replication with the cell cycle (Bell et al., 2002). Proliferating cell nuclear antigen (Pcna) and the Replication factor Cs (Rfc 1 4 and 5) are also involved in the DNA replication machinery (Johnson et al., 2005). Pcna is also widely used as a marker of proliferation (Landberg et al., 1997). In addition Cyclin D1 (Ccnd1) is essential for G1-S transition of the cell cycle (Caldon et al., 2006). A list of these genes is shown in Table 3.2 and their involvement in the DNA replication machinery and cell cycle progression is tabulated. Interestingly, the expression of these genes was not only above that of the Prlr+/ mice that could not lactate, but they were also higher than the WT animals (Figure 3.8). Therefore, we believe that rescued lactation in Prlr+/- mice, was due to abnormally high expression of genes involved in cellular proliferation and mitosis. This is an interesting finding, and the first indication of a potential mechanism of the partial penetrance of the Prlr+/- phenotype on a mixed 129 background.
62 Probe Set ID Genebank ID Title Gene name Function 160159_at AK088816 Ccnb1 Cyclin B1, related sequence 1 Cell cycle 103418_at AW122092 Rfc1 4 Replication factor C (activator 1) 4 DNA replication 96772_at J04620 Prim 1 DNA primase, p49 subunit DNA replication Cdc6 Cell division cycle 6 homolog (S. Cell cycle 103821_at AJ223087 cerevisiae) 103207_at D13543 Pola 1 DNA polymerase alpha 1, 180 kDa DNA replication Mad2l1 MAD2 (mitotic arrest deficient, Cell cycle 99632_at U83902 homolog)-like 1 (yeast) Smc2 SMC2 structural maintenance of Cell cycle 97421_at U42385 chromosomes 2-like 1 104001_at AA959940 Rfc1 5 Replication factor C (activator 1) 5 DNA replication Mki67 Antigen identified by monoclonal Cell proliferation 99457_at X82786 antibody Ki 67 102001_at M14223 Rrm 2 Ribonucleotide reductase M2 DNA replication Mcmd3 Minichromosome maintenance DNA replication 100062_at X62154 deficient 3 99578_at U01915 Top2a Topoisomerase (DNA) II alpha DNA metabolism Orc6 Origin recognition complex, DNA replication 95712_at AW045261 subunit 6-like Mcmd4 Mini chromosome maintenance DNA replication 93041_at D26089 deficient 4 homolog (S. cerevisiae) 99662_at AI194767 Pcnt Pericentrin Centrrosome 100612_at K02927 Rrm1 Ribonucleotide reductase M1 DNA replication Cdc20 Cell division cycle 20 homolog (S. Cell cycle 96319_at AW061324 cerevisiae) 98478_at U95826 Ccng2 Cyclin G2 Cell cycle Mda5 Interferon induced with helicase C DNA helicase 103446_at AA959954 domain 1 Mcmd7 Mini chromosome maintenance DNA replication 93356_at D26091 deficient 7 (S. cerevisiae) Cdc2a Cell division cycle 2 homolog A Cell cycle 100128_at M38724 (S. pombe) Mcmd2 Mini chromosome maintenance DNA replication 93112_at D86725 deficient 2 (S. cerevisiae) 101065_at X57800 Pcna Proliferating cell nuclear antigen Cell proliferation Mcmd5 Mini chromosome maintenance DNA replication 100156_at D26090 deficient 5 (S. cerevisiae) 94232_at AI849928 Ccnd1 Cyclin D1 Cell cycle 92809_r_at X17069 Fkbp4 FK506 binding protein 4 (59 kDa) DNA replication Ppicap Peptidylprolyl isomerase C- DNA replication 97507_at X67809 associated protein Table 3.2: 27 genes involved in DNA replication and cell proliferation up regulated in Prlr+/-, which lactated sufficiently to nurse pups. The 27 DNA replication and proliferation genes differentially regulated between Prlr+/- that failed to lactate compared to Prlr+/- that lactated sufficiently to nurse pups. Affymetrix ID (Probe set ID), Genebank accession number, Title, Gene Symbol and Gene Ontology are given for each.
63 Figure 3.8. Hierarchical clustering of the top 27 genes with differential expression between Prlr+/- with failed lactation and those that were able to lactate. Heat map and dendrogram of 27 genes up regulated in Prlr+/-, which lactated sufficiently to nurse pups. Low to high expression is represented in greyscale from white to black. Hierarchical clustering of these 27 genes separated Prlr+/- that could lactate (Prlr+/- lactation (L) with those that exhibited failed lactation (Prlr+/- no lactation (NL). Interestingly the expression of cell proliferation genes and genes involved in the replication machinery was not only greater than Prlr+/- that failed to lactate, but also above that of the Prlr+/+ replicates.
64 3.2.7. Hierarchical clustering demonstrated that gene expression changes as a result of loss of Gal were more closely related to treatment with S179D, than that of loss of an allele of Prlr. We used hierarchical clustering in an attempt to understand the similarities between the three models of failed secretory activation. We first clustered the entire 939 genes that were differentially expressed in all three models of failed lactation. This analysis grouped the samples according to the genetic background of the sample. The experiments involving Gal and S179D were on the 129OlaHsd background, whereas the Prlr experiments were on the mixed 129SvPas/129OlaHsd background. Thus of the 939 differentially expressed genes, the overwhelming gene expression changes were due to the genetic background of the strain and not the changes in genotype. Our laboratory has previously demonstrated that ductal morphogenesis is dependent in the strain of the stroma (Naylor et al., 2002). Thus, these observations suggest that care must be taken in the interpretation of biology from mouse models, and the researcher must take into account the strain from which the model is derived. To overcome these issues, we compared the genes that were differentially expressed between Gal+/+ and Prlr+/+, and removed these from the analysis. This left a set of 316 genes, which were differentially expressed in all three models of failed lactation, but were not due the genetic background of the animal. Hierarchical clustering of these 316 genes resulted in the identification of 4 clusters (Figure 3.9 A). The Gal+/+ and Prlr+/+ experiments formed a cluster together, which was closely related to a second cluster containing Prlr+/- that were capable of lactation. Unrelated to this cluster, was a cluster containing the Prlr+/- that failed to lactate, and a fourth cluster with the Gal-/- mice and the mice treated with S179D. Gal-/- mice treated with unmodified Prl (U-Prl) to rescue alveolar morphogenesis, separated into 2 clusters, indicating that rescue of lobuloalveolar development was less effective in one of the experimental replicates. Interestingly, the pattern of differential gene expression, more closely resembles that of Gal knockout rather than loss of a single allele of Prlr. We then investigated the similarities of the models based upon the gene expression changes identified in Venn analysis. When we clustered the experiments based on the 297 genes that changes in the – Prlr Venn set, we produced the same 4 clusters, but the differences between - Gal and + S179D were less obvious (Figure 3.9 B). Clustering the
65 genes that changed in the + S179D model resulting in the separation of the Gal-/- mice and the S179D treated samples from the other experiments (Figure 3.9 C). Using the genes that were changing as a result of loss of Gal, hierarchical clustering separated the WTs and the Prlr+/- that could lactate from those experiments resulting in failed secretory activation (Figure 3.9 D). Therefore, these results suggested that the transcriptional changes resulting from loss of the neuropeptide Gal was more closely related to treatment via S179D, the phosphorylated Prl mimic. Interestingly, the ratio of phosphorylated Prl to unmodified Prl was increased in the pituitaries of Gal-/- mice, suggesting that Gal acts to inhibit the phosphorylation of Prl in the mouse pituitary (Naylor et al., 2005b). Thus, Gal may modulate the ratio of pituitary Prl and its phosphorylated form, which in turn may regulate secretory activation during pregnancy.
66 Figure 3.9. Hierarchical clustering of genes within the Venn sets demonstrated that Gal-/- mice and mice treated with S179D were the most similar. Hierarchical clustering dendrograms were used to cluster the models of failed secretory activation based upon similarities in gene expression. Clustering of the set of 326 genes that changed in at least 1 of the three models of failed secretory activation, and were not different between Gal+/+ and Prlr+/+ (A). Clustering using genes that changed as a result of loss of an allele of Prlr (297 genes, B), treatment with S179D (84 genes, C) and loss of Gal (71 genes, D). 67 3.2.8. siRNA knockdown of Stat5a in the HC11 cell model of mammary cell differentiation modulated genes involved in the lactation signature. To test whether Stat5a, the major mediator of Prl signalling (Bole-Feysot et al., 1998), could directly regulate genes identified in the lactation signature, we used the HC11 normal mouse mammary epithelial cell line (Ball et al., 1988). HC11 cells proliferate in response to EGF and become confluent after three days of culture. Treatment with Prl from day 4 of culture results in dome formation and -Casein expression. We used short interfering RNA (siRNA) targeted against Stat5a, to reduced Stat5 expression and examine the expression of 2 of the genes in the lactation signature. HC11 cells were transfected with either Stat5a siRNA or control EGFP siRNA at day 2 during the growth phase of the differentiation protocol. siRNA against Stat5a inhibited dome formation (Figure 3.9 B) compared to siRNA against EGFP (Figure 3.9 A asterixis). Western analysis demonstrated that siRNA against Stat5a significantly lowered Stat5a protein expression at day 4 compared to siRNA against EGFP (Figure 3.10 A, p=0.014). Expression of Stat5a protein remained lower in cells treated with siRNA against Stat5a at day 6, compared to those treated with siRNA against EGFP, but recovered by day 8 (Figure 3.10 A). We investigated the mRNA expression of Stat5a and Stat5b at day 4, 6 and 8 of the differentiation protocol (Figure 3.10 B). The expression of the Stat5a transcript was consistent with Stat5a protein expression. In contrast, the expression of the highly homologous Stat5b transcript remained unchanged between cells treated with siRNA against Stat5a and EGFP, indicating the specificity of our approach. Reducing Stat5a levels in HC11 cells resulted in a decrease in the levels of the milk protein genes (Figure 3.10 B) Wdmn1 (p=0.018 at day 4 and p=0.06 for combined day 6 and day 8 data) and -Casein (not detected at day 4 and p=0.05 for combined day 6 and day 8 data). Thus reducing Stat5a with a siRNA approach resulted in attenuated Prl-mediated expression of milk protein genes in the HC11 differentiation protocol. We also examined the expression of the Ets transcription factor Elf5, which we have previously identified as a Prl target in transcript profiling experiments of Prlr-/- mammary epithelium at days 2, 4 and 6 of pregnancy (Harris et al., 2006). siRNA knockdown of Stat5a resulted in reduced expression of Elf5 at day 4 (Figure 3.10 B p=0.06), the time when Stat5a expression was lowest. Thus we confirmed that by altering
68 Figure 3.10. siRNA treatment of HC11 cells results in reduced dome formation. Dome formation in HC11 cells grown in the 8-day differentiation protocol. Treatment with Prl results in an increase in thickness of HC11 cells at multiple sites in a 2D assay (A asterixis). Treatment with siRNA against Stat5 inhibited dome formation in HC11 cells.
69 Figure 3.11. siRNA against Stat5 resulted in reduced expression of Aldo3 and Scd2. Transfection with Stat5 siNRA resulted in significantly reduced Stat5a protein at day 4 compared to siRNA against EGFP, detected by western blot (B). Quantification of Stat5 protein is shown by densitometry is shown in A. siRNA against Stat5 significantly lowered the expression of the milk genes -Casein and Wdnm-1, and the Prl target Elf5 at day 4. Scd2 and Aldo3, two genes out of the 35 identified in the lactation signature, were down regulated in response to Stat5 siRNA compared to EGFP siNRA at day 4.
70 the total levels of Stat5, we could modulate the expression of Prl/Stat5 target genes. We then tested whether Stat5a reduction would modulate the levels of 2 of the genes in the lactation signature (Figure 3.11 B). The lipogenic enzymes, Aldo3, the key step in the conversion of glucose to pyruvate, and Scd2, essential for the formation of monounsaturated fatty acids (Enoch et al., 1976), were both down regulated in all three models of failed secretory activation. Knockdown of Stat5a resulted in decreased expression of both Aldo3 and Scd2 at day 4 (p=0.06 and p<0.001 respectively). By day 6 and day 8, the mRNA expression of these genes had recovered. Therefore, these experiments give additional insight into the regulation of lipid synthesis during secretory activation, and place Aldo3 and Scd2 downstream of Stat5, possibly within the Prl- signalling cascade.
71 3.3 Discussion Secretory activation is the final step in the formation of a functional secretory unit in the mammary gland during pregnancy. It is characterised by the closure of alveolar tight junctions, and the movement of milk and lipid into the alveolar lumens (Neville et al., 2002). Pg withdrawal is essential for this stage (Nguyen et al., 2001), as ovariectomy in pregnant mice initiates tight junction closure. Prl is also required for secretory activation, as inhibition of pituitary PRL secretion by Bromocriptine suppresses milk production (Kulski et al., 1978). Work in our laboratory has also demonstrated a mammary cell autonomous role for the neuropeptide Galanin during lactogenesis. Thus, the hormonal control of secretory activation has been well characterised (Reviewed in (Neville et al., 2002)). Although our knowledge of the transcriptional program that regulates alveolar morphogenesis is expanding (Brisken et al., 2006), much is still to be understood. Advancements in genomic sequencing (Nadeau et al., 2001; Venter et al., 2001) combined with the introduction of micro array technology (Fodor et al., 1991) within the last 15 years, has rapidly expanded our understanding of gene expression in biological processes. In this chapter, we have used Affymetrix GeneChip technology combined with transgenic mouse models, to further the understanding of the transcriptional regulation of secretory activation. Transcript profiling using micro array technology represents a snap shot of the processes operating within a tissue. Thus it is important to expand our model systems and data sets and combine these results in order to understand the potential mechanisms in operation. The mouse models of loss of a single allele of Prlr, loss of Gal and treatment with S179D, result in failed lactation at a stage that has proceeded through secretory activation, as all lipid synthesis genes were expressed, albeit reduced compared to control mice (Figure 3.6). Thus, these models represent a unique time point for the identification of genes required for full secretory activation. In addition, they allow us to understand the transcriptional regulation of secretory activation, as a result of failed hormone signalling, namely that of Prl and Gal. Comparison with other transcript profiling studies, such the comprehensive experiments by Rudolph and colleagues (Rudolph et al., 2003), examining global gene expression changes throughout mammary gland development, enable further characterisation of the genetic program that co-ordinates alveolar morphogenesis. 72 Statistical algorithms can be used to predict false discovery rates as a result of large numbers of observations, an inherent problem with transcript profiling data (Yang et al., 2006). By combining three different models of failed secretory activation performed in duplicate, we were able to identify 35 genes representative of a lactation signature (Figure 3.3). Using multiple testing we estimated the false discovery rate of these genes at 11%. Therefore, we can be confident that the majority of the genes identified within the lactation signature, are biological relevant during secretory activation. In addition, a large proportion of these genes were involved in lipid synthesis, a process essential for the formation of milk lipids during lactogenesis (McManaman et al., 2004). This was further proof that we had identified genes critical in regulating milk production during secretory activation. Many of these genes had not previously been placed in the secretory activation pathway, and thus are interesting candidates for further study in the mammary gland. We identified several transcription factors within the lactation signature. These include Srebf1, a key gene involved in the synthesis of lipids (Horton et al., 2002), Cebp, which has a role controlling mammary gland development and lipogenesis (Gigliotti et al., 2003; Mason et al., 2003) and Sox4, whose function in the mammary gland is yet to be determined. As transcription factors, these genes represent potential programmers of gene expression essential for secretory activation downstream of the transcription factor Stat5, which is essential for secretory activation and -Casein transcription (Schmitt-Ney et al., 1992). In this chapter, we have also shown, that reduced Stat5 levels in HC11 cells resulted in attenuated expression of two lipid synthesis genes (Scd2 and Aldo3), identified in the lactation signature (Figure 3.11). It is clear from our studies in this chapter that Prl mediated signalling results in a transcriptional hierarchy that we do not fully understand, and more investigation is required to dissect the pathways involved in lactogenesis. We used hierarchical clustering, to group the models of failed lactation using the genes identified by Venn analysis (Figure 3.9). We used this approach to gain insight into the hormonal regulation of mammary gland gene expression during secretory activation. Curiously, we could never separate the experiments between loss of Gal and treatment with S179D, using genes that changed in the loss of Prlr, loss of Gal or treatment with S179D. Therefore the pattern of gene expression as a result of loss of Gal was very similar to that observed with treatment with S179D. Work in our laboratory has demonstrated that Gal
73 modulates pituitary Prl phosphorylation (Naylor et al., 2005b). Thus this appears to be an important mechanism in the transcriptional control of secretory activation. The majority of gene expression changes (654/939 genes) were detected with loss of a single allele of Prlr. Thus Prl was clearly having the greatest effects on gene expression during secretory activation. Our transcript profiling data also provided an insight into the partial penetrance of the Prlr heterozygote phenotype. Prlr+/- mice were on a mixed 129 background, and two thirds of these mice fail to lactate following their first pregnancy. An analysis of gene expression resulted in the discovery of down regulation of a large number of genes involved in DNA replication (Bell et al., 2002) and proliferation in Prlr+/- mice that failed to lactate compared to those that were able to lactate (Figure 3.8). Two of these genes were proliferating cell nuclear antigen (Pcna) and Ki67, widely used markers of proliferation (Landberg et al., 1997). Our results in this chapter suggest, that in the set of mice not capable of nursing pups, lactation fails as a result of decreased mammary epithelial cell proliferation, with resulting decreased alveolar surface area insufficient for pup survival. These findings are consistent with the histology of Prlr+/- mammary glands, which show varying degrees of lobuloalveolar development ((Brisken et al., 1999), Figure 3.1). In addition, when crossed to a poorly epithelial side-branched mouse strain such as the C57Bl/6, all Prlr+/- mice fail to lactate on their first and subsequent pregnancies (Hennighausen L, Lindeman GJ personal communication). Thus, partial penetrance of the Prlr+/- phenotype in a mixed genetic background was due to decreased epithelial proliferation, resulting in decreased surface area for sufficient lactation. In summary, we transcript profiled three mouse models of failed lactation, including loss of a single allele of Prlr, loss of Gal and treatment with S179D, to identify a lactation signature comprised of 35 genes (A schematic diagram for secretory activation is illustrated in Figure 3.12). Many of these genes were placed in the secretory activation pathway for the first time, and the majority of these genes were involved in the synthesis of lipids. We used gene expression profiles to demonstrate, that all three models had proceeded through secretory activation, but failed to nurse pups due to insufficient milk and lipid gene expression. We also discovered that the majority of gene expression changes crucial for secretory activation, were via the Prlr. Treatment with the molecular mimic of
74 phosphorylated Prl, S179D, resulted in similar gene expression changes to that of loss of Gal. Therefore by comparing gene expression profiles between these models, we gained insight into the regulation of secretory activation by Prl, Gal and phosphorylated Prl. A complete understanding of hormonal regulation and gene expression during secretory activation, may assist in understanding why early full-term pregnancy is protective against breast cancer (MacMahon et al., 1970).
75 Figure 3.12. Model of secretory activation. The majority of signalling during secretory activation occurs by Prl binding to its receptor in mammary epithelium. This results activation of Jak2/Stat5 and results in Stat5 nuclear translocation to the nucleus where it transcribes a number of genes involved in secretory activation including other transcription factors such as Elf5 and Sox4 and the inhibition of Cebp and Srebf1. Srebf is involved in transcribing genes involved in the synthesis of lipid. Other lipid synthesis genes transcribed by this process are Fdps, Cyp51, Sque, Elovl5, Scd2, Aldo3 and Acyl. Copz1 and Ugalt are involved in the synthesis of lactose. It is currently unclear what molecular mechanisms are initiated in the secretory cell by Gal binding to its receptor.
76 Chapter 4 The role of Elf5 during alveolar morphogenesis. 4.1 Introduction Alveolar morphogenesis during pregnancy involves massive tissue remodelling resulting in the formation of lobuloalveoli capable of milk secretion (Oakes et al., 2006). This process requires binding of Prl to its receptor within the mammary epithelium (Brisken et al., 1999), triggering a cascade of events resulting in the activation and dimerisation of Stat5 (DaSilva et al., 1996; Han et al., 1997). Stat5 is then translocated to the nucleus, where it induces the transcription of a number of target genes involved in alveolar morphogenesis (Wakao et al., 1995). As we have reported in Chapter 3, some of these target genes are themselves transcription factors, which regulate various processes involved in the formation of a functional mammary gland. The helix-loop-helix transcription factor Id1 is highly expressed by the expanding epithelium during early pregnancy and inversely correlated with the expression -Casein (Parrinello et al., 2001). Id1 appears to play an important role in promoting epithelial cell proliferation and inhibiting differentiation in the mammary epithelium during pregnancy (Desprez et al., 1995). In contrast, its closely related family member Id2 is important for differentiation of the lobuloalveolar compartment and secretory activation (Miyoshi et al., 2002; Mori et al., 2000). The RankL/Rank target NF-B is essential for mediating proliferation signals induced by Pg (Cao et al., 2003; Conneely et al., 2003). We have also shown that both RankL and NF-B are potential targets of Prlr signalling (Ormandy et al., 2003). Epithelial Cebp regulates the proliferation and differentiation of the alveolar epithelium during pregnancy, and controls the expression of milk genes -Casein, Wap and Wdnm1 (Robinson et al., 1998; Seagroves et al., 1998). In addition, Cebp controls the distribution of progesterone receptor positive cells in the mammary epithelium, suggesting that it is an important factor in determining cell fate (Seagroves et al., 2000). In chapter 3 we identified, Srebf and Cebp, which were consistently altered in all three models of failed lactation. Srebf controls the expression of lipid metabolism genes (Horton et al., 2002), while Cebp is involved in controlling fatty acid synthesis (Mason et al., 2003), processes essential for secretory activation during pregnancy. Recently Gata3 has emerged 77 as an important factor in alveolar morphogenesis and appears to play an important role in luminal cell fate specification (Kouros-Mehr et al., 2006a; Kouros-Mehr et al., 2006b). Gata3 was down regulated in Prlr-/- transplants at days 2, 4 and 6 of pregnancy (Ormandy et al., 2003). Conditional deletion of Gata3 during ductal morphogenesis results in failed TEB formation and ductal outgrowth (Kouros-Mehr et al., 2006a). Chronic loss of Gata3 resulted in undifferentiated luminal cell expansion and dissociation with the basement membrane, which led to apoptosis in the long term. Lobuloalveolar development was impaired and pup weight was reduced after long term conditional deletion of Gata3, however this may have been a consequence of the luminal epithelial defects prior to pregnancy, and not an effect on the differentiated epithelium per se. Thus it is now clear that the developmental events during pregnancy are executed by a cascade of transcription factors that are initiated by hormonal signals, principally prolactin and progesterone. Defining this genomic regulatory network and placing it in the context of stem and progenitor cell hierarchy and tissue morphology offers an unprecedented understanding of the processes that build a mammary gland, and which may go awry in cancer. Although, our understanding of the transcriptional control of alveolar morphogenesis is expanding, much is still to be understood. To identify the transcriptional targets of the Prlr, essential for alveolar morphogenesis, our laboratory used transcript profiling of Prlr-/- mammary transplants at days 2, 4 and 6 of pregnancy (Ormandy et al., 2003), a time where the epithelial cell number differences between WT and Prlr-/- epithelium were negligible. To identify genes specific to the epithelium, only differentially expressed genes, which were absent in cleared mammary fat pads were used in further analysis. This analysis generated a large list of genes that may be important for Prl mediated alveolar morphogenesis. To reduce this list to small number of tightly regulated genes, these profiles were compared to a cell model of Prl action. SCp2 cells will differentiate and produce milk proteins when treated with insulin, hydrocortisone and Prl. Transcript profiles were generated from SCp2 cells grown on Matrigel for 48 hours in the presence of insulin, hydrocortisone either with or without Prl. Comparison with the transcript profiles of Prlr-/- mammary epithelium at days 2, 4 and 6 of pregnancy, generated a small list of epithelial specific Prl regulated genes. One of these genes was Elf5, an epithelial specific member of the Ets transcription factor family.
78 The large Ets transcription factor family are characterised by a highly conserved DNA binding domain (the ets domain), which binds to sites containing a central GGA motif (Sharrocks et al., 1997). Ets transcription factors are further classified by their animo acid similarities and conservation of protein domains. Elf5 (e74-like factor 5 or ESE-2) was isolated from a mouse lung cDNA library using a probe containing the Ets domain of Elf3 (e74-like factor 3 or ESE-1) (Zhou et al., 1998). Thus it is most closely related to Elf3. Mouse Elf5 and human ESE-2 are 95% identical and are expressed as 2 isoforms produced by alternative start sites. Such high conservation of sequence implies similar conservation of function (Zhou et al., 1998). Work in our laboratory has demonstrated that Elf5 expression is restricted to luminal mammary cells (Harris et al., 2006), and its expression is increased dramatically during pregnancy, to levels that far exceed those seen in other tissues. Elf5 can also bind to an ets like domain in the proximal promoter of whey acidic protein (Wap) and induce its expression independently of lactogenic hormones, indicating that Elf5 may be an important mediator of alveolar differentiation during mid-pregnancy (Thomas et al., 2000). Elf5-/- mice die in utero due to a placentation defect (Donnison et al., 2005). Elf5+/- mice fail to lactate on the first pregnancy, and appear to be a phenocopy of the Prlr+/- (Ormandy et al., 1997b). Lobuloalveolar development was severely impaired in mammary transplants of Elf5+/- epithelium, demonstrating that Elf5 has a cell autonomous role in the mammary epithelium during alveolar morphogenesis (Zhou et al., 2005). The levels of Elf5 are reduced in Prlr+/- glands and there is no similar reduction in the expression of Prlr in Elf5+/-, implying that Elf5 is downstream of the Prlr (Zhou et al., 2005). We have also demonstrated in Chapter 3, that by reducing the expression of Stat5 in HC11 normal mouse mammary epithelial cells, we can attenuate the expression of Elf5, implying that Elf5 is also downstream of Stat5 (Figure 3.11). Therefore, the aim of this chapter was to understand the extent to which Elf5 controls Prl mediated alveolar morphogenesis. To achieve this aim, we used several experimental approaches. Firstly, we obtained an Elf5 knockout (Elf5-/-) mouse generated using tetraploid rescue to overcome its embryonic lethality, generated at AgResearch Crown Research Institute, New Zealand. We took mammary epithelial biopsies from the Elf5-/- mouse, and performed mammary transplantation in order to compare the mammary phenotype of Elf5-/- epithelium to Prlr-/- epithelium during pregnancy. Secondly we used the
79 MTB system for mammary specific inducible Elf5 expression, to understand the effects of forced expression of Elf5 in the absence of pregnancy. Thirdly, we used a retroviral delivery system to over-express Elf5 in Prlr-/- primary mammary epithelial cells to determine whether Elf5 could completely compensate for the Prlr during pregnancy.
80 4.2 Results 4.2.1. Production of a renewable source of Elf5-/- mammary glands. A null mutation of the Elf5 gene was generated by homologous recombination resulting in insertion of a puromycin cassette into exon 2, 76bp downstream of the ATG (Donnison et al., 2005). Elf5-/- mice fail to form a chorion resulting in mid-gestational lethality at approximately E8.5. To overcome embryonic lethality, diploid embryo- tetraploid embryo aggregation chimeras were produced (Tanaka et al., 2001), and a single female was successfully carried to sexual maturity. This female was fertile but unable to lactate port partum, and was sent to us as a gift from Peter Pfeffer (AgResearch Crown Research Institute NZ). Mammary biopsies were taken from this mouse and transplanted into the cleared 4th mammary fat pads of Rag1-/- hosts. Twenty primary mammary transplants of Elf5-/- epithelium in one mammary fat pad, and 20 transplants of WT mammary epithelium in the alternate 4th mammary fat pad were performed. All 40 primary mammary epithelial transplants successfully engrafted indicating that epithelial loss of Elf5 does not dramatically affect mammary regenerative capacity, thus loss of Elf5 does not result in the loss of the primary mammary stem cell population required for ductal morphogenesis. All secondary serial transplants of Elf5-/- epithelium generated successful mammary outgrowths, suggesting that the self-renewal capacity of mammary stem cells via asymmetrical cell division (Smalley et al., 2003) was not dramatically altered by loss of Elf5. Mammary epithelial cell dilution and transplantation experiments are underway to determine if a more subtle defect is present, with outcomes to be reported subsequent to the submission of this thesis. A set of Elf5-/- and WT primary mammary epithelial transplants has been cryogenically preserved to provide material for future experiments.
4.2.2. Mammary epithelial Elf5 was not required for branching morphogenesis. Mature 12-week-old transplants generated from either WT (Figure 4.1 A and C) or Elf5-/- (Figure 4.1 B and D) epithelium displayed complete ductal morphogenesis. At 12 weeks, both WT and Elf5-/- epithelium completely filled the mammary fat pad. Ductal bifurcation and elongation appeared normal (Figure 4.1 E and F black lines). In addition,
81 Figure 4.1. Ductal morphogenesis was normal in Elf5-/- mammary transplants. Mammary whole mounts of WT (A and C) and Elf5-/- (B and D) mammary epithelial transplants, Carmine Stain. At 12 weeks WT and Elf5-/- epithelium developed a full ductal outgrowth, and terminal end buds had regressed into alveolar buds (arrows). Schematic representation of ductal branching within WT (E) and Elf5-/- (F) epithelium. Ductal elongation and bifurcation (black lines) as well as secondary (blue lines) and tertiary (red lines) braches was normal in epithelium that lacked Elf5.
82 secondary (Figure 4.1 E and F blue lines) and tertiary (Figure 4.1 E and F red lines) side branches were present, indicating that homozygous loss of Elf5 is not required for ductal branching morphogenesis. Once the mammary epithelium reaches the periphery of the fat pad, TEBs regress to quiescent alveolar buds (Sternlicht, 2006). We observed no persistence of TEBs in Elf5-/- epithelium (Figure 4.1 C and D arrows), suggesting that Elf5 is not required in the epithelium for this TEB regression. Ductal morphogenesis and TEB regression was rescued in mammary epithelial transplants lacking Prlr (Brisken et al., 1999), thus Elf5 like the Prlr does not have a cell autonomous role in the mammary epithelium for these processes.
4.2.3. Mammary epithelial loss of Elf5 resulted in failed alveolar morphogenesis, which resembled loss of mammary epithelial Prlr. Elf+/- mammary epithelial transplants exhibited a failure of complete lobuloalveolar development, thus Elf5 is required in the epithelium during alveolar morphogenesis (Zhou et al., 2005). Elf5+/- mammary epithelium has significantly reduced milk protein expression and mice fail to nurse pups on their first, and in many cases subsequent pregnancies. This phenotype is a copy of that seen in Prlr+/- animals. The original Elf5-/- female generated by tetraploid embryonic stem cell rescue exhibited lactational failure, and we hypothesised that Elf5-/- mammary glands would phenocopy Prlr-/- mammary glands, To directly compare the Elf5-/- and the Prlr-/- phenotype during pregnancy, we performed mammary transplantation of Elf5-/- and Prlr-/- epithelium with WT epithelium in the contralateral glands of the same Rag1-/- host animal. We then time mated the host mice, and collected mammary epithelium at days 6.5, 12.5, 18.5 days post coitus (dpc) and 1 day post partum (dpp) and compared development (Figure 4.2). The first phase of alveolar morphogenesis involves rapid proliferation of the alveolar buds to increase the surface of the mammary epithelium (Figure 4.2 A and B). During the second half of pregnancy, differentiation takes over as the dominant force resulting in the formation of polarised lobuloalveoli (Figure 4.2 C), capable of milk secretion at parturition (Figure 4.2 D). Prlrs in the mammary epithelium are required for alveolar morphogenesis (Brisken et al., 1999). Thus, transplants of Prlr null epithelium exhibited failed alveolar proliferation and development of lobuloalveoli (Figure 4.2 E-H). Mammary transplants of Elf5 null
83 Figure 4.2. Alveolar morphogenesis failed in both Prlr-/- and Elf5-/- mammary epithelial transplants. Mammary whole mounts of WT (A-D), Prlr-/- (E-H) and Elf5-/- (I-L) mammary epithelial transplants at 6.5dpc (A, E and I), 12.5dpc (B, F and J), 18.5dpc (C, G and K) and 1dpp (D, H and L), Carmine stain.
84 Figure 4.3. Lobuloalveoli failed to develop in both Prlr-/- and Elf5-/- mammary epithelial transplants. Haematoxylin and eosin histochemistry of WT (A-D), Prlr-/- (E-H) and Elf5-/- (I-L) mammary epithelial transplants at 6.5dpc (A, E and I), 12.5dpc (B, F and J), 18.5dpc (C, G and K) and 1dpp (D, H and L).
85 epithelium also displayed failed alveolar proliferation and differentiation (Figure 4.2 I-L) that was directly comparable to that observed in Prlr-/- transplants. Lipid droplets are present in the alveolar cells of WT epithelium during late pregnancy, which marks the onset of secretory activation (Figure 4.3 A-D). In contrast, epithelium null for both the Prlr (Figure 4.3 E-H) and Elf5 (Figure 4.3 I-L) contain cells that never produce cytoplasmic lipid droplets, indicating that secretory activation has failed in both models. At parturition, milk and lipid move into the alveolar lumens, resulting in alveolar distension enabling the delivery of milk (Figure 4.3 D). Prlr-/- (Figure 4.3 H) and Elf5-/- (Figure 4.3 L) epithelium at parturition never take on this appearance, and although some expansion of the epithelium has occurred, alveoli do not appear distended with secretions. Thus, the failure of lobuloalveolar development is morphologically similar in Elf5-/- epithelium compared to Prlr-/- epithelium.
4.2.4. The Elf5 heterozygote mammary phenotype was partially penetrant Eighty four percent of Elf5+/- mice on a mixed 129SvJ/C57Bl/6J background fail to lactate on the first pregnancy (Zhou et al., 2005). The authors of this study suggested that the partially penetrant phenotype was probably due to a genetic modifier present in one of the genetic backgrounds. The degree of ductal side branching in individual animals could be a potential explanation, as the 129 strain has a highly ductal side branched morphology compared to the poorly side branched C57Bl/6J, a phenotype which is dependent on stromal factors (Naylor et al., 2002). Increased ductal side branching in individual mice may facilitate sufficient lobuloalveolar development enabling pup survival. The phenotype observed in mammary transplants of Elf5+/- in this study (Figure 4.4), offer an additional or alternative explanation for this partially penetrant phenotype. Mammary transplants of WT epithelium displayed fully differentiated lobuloalveoli that completely filled the mammary fat pad (Figure 4.4 A and B). In comparison, Elf5+/- transplants taken at 1dpp displayed varying degrees of lobuloalveolar development, from well-differentiated alveolar lobules (Figure 4.4 C and D) and intermediate alveolar lobules (Figure 4.4 E and F) to the least developed with a morphology, which resembles mid-pregnancy (Figure 4.4 G and H). Interestingly, on close examination, Elf5+/- mammary transplants demonstrated varying
86 Figure 4.4. The Elf5+/- phenotype was partially penetrant. Mammary whole mounts of WT (A and B), and Elf5+/- epithelium (C-I) at 1dpp, Carmine stain. Elf5+/- transplants displayed varying degrees of lobuloalveolar development, from well-differentiated lobules (C and D), intermediate lobules (E and F) and poorly differentiated lobules (G and H). Areas of the same ductal tree displayed chimerism, as some areas displayed poorly differentiated lobules (H), in close proximity to areas, which had formed well-developed lobules (I).
87 Figure 4.5. Chimerism in Elf5+/- transplants correlated with lipid formation in the epithelium, immune reactivity to milk staining and heterogeneous Elf5 staining. Haematoxylin and eosin histochemistry (A-C) and immune-histochemistry to milk (D-F) and Elf5 (G-H) of WT (A, D and G), well-differentiated Elf5+/- epithelium (B, E and H), poorly developed Elf5+/- epithelium (C, F and I) at 1dpp and WT epithelium at 2dpc. Well- differentiated areas of Elf5+/- transplants contained cytoplasmic lipid droplets (B) and strong immune-reactivity to milk (E) and Elf5 (H). In contrast, poorly-differentiated areas of the same ductal tree formed intermediate lobules with reduced cytoplasmic lipid droplets (C) and attenuated immune-reactivity to milk (F). Poorly differentiated areas of Elf5+/- epithelium showed heterogenous staining of Elf5 with the luminal cells of some alveoli with strong Elf5 staining and others with little or no immune-reactivity to Elf5 antibody (I, black arrows). Heterogenous Elf5 staining was also observed in WT epithelium at 2dpc, negative cells indicated by black arrows (J). Elf5 peptide block negative control (K). 88 degrees of development within the same mammary transplant. In some areas the mammary epithelium failed to develop lobuloalveoli (Figure 4.4 H), whereas other areas of the same mammary tree contained well-differentiated lobuloalveoli (Figure 4.4 I). In areas that displayed well-differentiated alveoli, large lipid droplets were observed within the luminal epithelial cells (Figure 4.5 B). In addition strong milk staining was apparent within the alveolar lumens (Figure 4.5 E). In contrast, in areas of the same gland that displayed immature lobules, the presence of lipid within the epithelium (Figure 4.5 C) and immunoreactivity to the milk antibody was reduced (Figure 4.5 F). The well-developed areas of Elf5+/- mammary epithelium displayed strong immunoreactivity to Elf5 (Figure 4.5 H) in most of the luminal epithelial cells, and was similar to what was observed in WT mammary transplants (Figure 4.5 G). In contrast the areas of poorly developed alveolar lobules displayed a heterogenous pattern of Elf5 staining, and contained luminal epithelial cells with no immunoreactivity to Elf5 (Figure 4.5 I arrows). Therefore, the pattern of Elf5 expression is more variable in transplants lacking a single copy of Elf5, resulting in differing degrees of lobuloalveolar development. Heterogeneity of Elf5 in luminal epithelial cells was also seen in WT mammary transplants at 2dpc (Figure 4.5 J), and suggests that during pregnancy Elf5 must be induced in the remainder of Elf5 negative luminal epithelial cells for complete lobuloalveolar development. Thus, homogeneous expression within the luminal epithelium is required for the full development of lobuloalveoli, presumably to push the lobular progenitor (Kordon et al., 1998; Shackleton et al., 2006; Stingl et al., 2006) into a differentiated alveolar cell. This observation is intriguing and will be discussed in more detail later in this chapter. Variation in lobuloalveolar development in Elf5+/- transplants is similar to what is observed in Prlr+/- mice. Two thirds of Prlr+/- mice on a mixed 129 SvPas/Ola or 129Sv/C57Bl/6 fail to lactate on their first pregnancy due to lobuloalveolar defects. Prlr+/- mice, which can nurse pups, displayed increased numbers of well-developed lobules, in comparison to Prlr+/- that failed to lactate and displayed less well-developed alveolar morphology (Ormandy et al., 2003). Western analysis demonstrated that milk protein expression at 1dpp was decreased in Prlr+/- mice that were unable to nurse pups, compared to Prlr+/- mice that were able to lactate and WT mice (Figure 4.6 A). A similar decrease in milk protein expressed was observed in Elf5+/- transplants at parturition (Figure 4.6 B),
89 Figure 4.6. Milk protein was decreased in Prlr+/- mice that could not nurse pups and Elf5+/- transplants. Western analysis of milk protein expression in Prlr+/- mice (A) and Elf5+/- transplants (B) at parturition. Milk protein expression was decreased in Prlr+/- mice that failed to lactate sufficiently to nurse pups (No lact.) compared to wildtype and Prlr+/- mice that lactated sufficiently to nurse pups (Partial lact.). Milk protein expression was also decreased in Elf5+/- transplants at 1dpp (B) and was directly correlated with the extent of lobuloalveolar development (lane number in B is matched to the mammary gland number in C). (C) Mammary whole mounts of wildtype transplants (1, 3, 5 and 7) and in Elf5+/- transplants (2, 4, 6 and 8) at 1dpp, Carmine stain. An antibody to -Actin was used as a loading control (A and B). 90 which was directly correlated to the extent of lobuloalveolar development observed in these glands (Figure 4.6 C).
4.2.5. Alveolar proliferation was attenuated in both Elf5 and Prlr null epithelium. Proliferation of the alveolar epithelium is critical for increasing the surface area of the mammary epithelium to facilitate lactation. For example, mice carrying a null mutation of Cyclin-D1, a key modulator of G1-S cell cycle progression (Sicinski et al., 1997), exhibited failed alveolar proliferation and were unable to nurse pups (Fantl et al., 1999). To test whether proliferation was perturbed in epithelium lacking Prlr or Elf5, we used immune- histochemistry against BrdU. Mice were injected with 5-Bromo-2’-deoxyuridine (Sigma,
Germany) dissolved in distilled H2O two hours prior to sacrifice by CO2 ashphyxiation. BrdU incorporates in the DNA of synthesising cells, and can be used to estimate the rate of proliferation in a cell population. Proliferation was reduced in mammary epithelium lacking both the Prlr (Figure 4.7 F-J) and Elf5 (Figure 4.7 K-O). There was a trend towards reduced proliferation as measured by BrdU immunohistochemistry in mammary epithelium lacking Prlr or Elf5 (Figure 4.8). Proliferation was also measured by the total number of epithelial cells per field sampled within each gland. We observed a large decrease in the total number of epithelial cells per field of view sampled (Figure 4.8 B). These observations indicate that both Prlr and Elf5 modulate proliferation to a similar extent in alveolar epithelium during pregnancy.
4.2.6. Elf5 was essential for milk protein synthesis. The expression of milk proteins in the mammary epithelium occurs in a programmed manner. Trajectory clustering of gene expression during alveolar morphogenesis resulted in identification of groups of genes, which were expressed at all time points of pregnancy (Rudolph et al., 2003). The earliest milk genes expressed in the mammary gland are the extracellular protease inhibitor (Expi/Wdnm1) and the caseins,
91 which are detectable in virgin mammary glands. The expression of Wap is increased during mid-pregnancy, and subsequently, Casein- and alpha lactalbumin are increased during late
Figure 4.7. Proliferation determined by BrdU immunohistochemistry was decreased in Prlr-/- and Elf5-/- transplants.
92 Immune-histochemistry to BrdU in WT (A-E), Prlr-/- (F-J) and Elf5-/- transplants (K-O) at 2.5dpc (A, F and K), 6.5dpc (B, G and L), 12.5dpc (C, H and M), 18.5dpc (D, I ad N) and 1dpp (E, J and O).
93 Figure 4.8. Proliferation in Prlr-/- and Elf5-/- transplants using BrdU immunhistochemistry. (A) Bar graph depicting the percentage of BrdU positive cells in WT transplants (yellow bars), Prlr-/- transplants (pink bars) and Elf5-/- transplants (blue bars) at 2.5dpc, 4.5dpc, 6.5dpc, 12.5dpc, 16.5dpc, 18.5dpc and 1dpp. (B) The absolute number of cells/sampling field at 6.5dpc, 12.5dpc, 16.5dpc, 18.5dpc and 1dpp in WT transplants (yellow line), Prlr-/- transplants (pink line) and Elf5-/- transplants (blue line).
94 pregnancy. An antibody targeted to the milk proteins Casein- , Casein- and WAP was used to investigate whether Prlr and Elf5 epithelium had differentiated to produce milk (Figure 4.9). Milk protein expression can be detected as early as 2.5dpc in WT mammary epithelium (Figure 4.9 A). The intensity of milk protein expression increased throughout pregnancy, culminating in intense brown staining in the WT mammary glands at parturition (Figure 4.9 E). Immunoreactivity to the milk antibody was detected in mammary epithelium from Prlr-/- transplants, with a low level amount of expression detected within the ducts and alveoli from 12dpc (Figure 4.9 H). Interestingly, there was almost no detectable milk protein expression in the mammary glands from Elf5-/- transplants at all time points of pregnancy. Elf5 expression can be induced by synthetic progesterone treatment in T47D cells (Heidi Hilton; unpublished data), thus it is possible that in the absence of Prlr signalling, Elf5 expression can be induced by progesterone. This is clearly not sufficient for full lobuloalveolar development, but enables some milk protein expression in mammary epithelium without the Prlr. This may be a point of convergence with between prolactin and progesterone action during pregnancy. These data indicate that Elf5 is a transcription factor required for alveolar differentiation during alveolar morphogenesis, and that it is possibly downstream of the Prlr. Also, that Elf5 appears to regulate the expression of milk proteins from the earliest stages of pregnancy.
4.2.7. Polarisation of alveoli was normal in Elf5 and Prlr null epithelium. Polarisation of the mammary epithelium during pregnancy is essential for directing milk and lipid secretion into the alveolar lumen (Barcellos-Hoff et al., 1989). Polarisation requires signalling from the extra-cellular matrix (Fata et al., 2004), and the maintenance of cell adhesion (Gumbiner, 1996). Using immune-fluorescence, we tested whether mammary epithelium that lacks Prlr and Elf5 can polarise during pregnancy. Zo1 is component of tight junctions and is a marker of the apical membrane (Stevenson et al., 1986). -catenin is a component of cell adherens junctions (Huber et al., 1996), and is found on the lateral surfaces of luminal and alveolar cells. During alveolar morphogenesis the alveolar cells cleave and form a single layer of epithelial cells surrounding a circular lumen. Zo1 and - catenin expression was tightly restricted to the apical and lateral membranes respectively in
95 Figure 4.9. Immune-reactivity to milk was decreased in Prlr-/- and almost absent in Elf5-/- transplants. Immune-histochemistry to Milk in WT (A-E), Prlr-/- (F-J) and Elf5-/- transplants (K-O) at 2.5dpc (A, F and K), 6.5dpc (B, G and L), 12.5dpc (C, H and M), 18.5dpc (D, I ad N) and 1dpp (E, J and O).
96 Figure 4.10. Polarisation was normal in Prlr-/- and Elf5-/- transplants. Immune-fluorescence to Zo-1 (Red) and -Catenin (Green) in WT (A-C), Prlr-/- (D-F) and Elf5-/- transplants (G-I) at parturition. Topro was used to identify nuclei. Zo-1 staining was restricted to the apical membrane, and -Catenin to the lateral surfaces of epithelial cells. Epithelium without Prlr or Elf5 resulted in no change in the location of Zo-1 and -Catenin.
97 WT alveolar cells throughout pregnancy (Figure 4.10 A-C). This expression remains at parturition when the alveoli are distended and secreting milk (Figure 4.10 C). Although, the formation of alveoli is severely impaired in mammary epithelium without Prlr or Elf5, polarisation of the alveolar cells was normal (Figure 4.10 D-I). Thus Prlr and Elf5 are required for alveolar morphogenesis, but not required for the polarisation of the alveolar epithelium.
4.2.8. Transcript profiling of Elf5-/- epithelium at days 4 and 6 of pregnancy Our experiments indicated that Elf5 was the transcription factor regulating Prlr mediated alveolar differentiation. To understand the transcriptional program mediated by Elf5, we investigated the differential expression of genes in Elf5-/- transplants at days 4 and 6dpc. We also transcript profiled cleared mammary fat (CFP) pads at the same time points of pregnancy, to isolate an epithelial specific population of genes. All experiments were performed in triplicate. After filtering for probe sets that did not reach an intensity criterion of a mean over 10 in at least one of the experimental conditions, we reduced the number of genes to 34409/45101 total probe sets on the MOE430v2 GeneChip. We clustered the 34409 genes using principal components analysis to visualise successful separation of the experimental groups (Figure 4.11). Expression changes between WT and either Elf5-/- were determined using an analysis of variance. Using an unadjusted p<0.001, we identified 69 (Appendix XIII) and 429 differentiated genes (Appendix IX) in Elf5-/- epithelium at day 4 and 6 respectively, and 24 genes that were commonly changing at day 4 and 6 combined (Table 4.1). Using 500 random permutations we determined the false discovery rate as 10/69 (14.5%) at day 4 and 12/429 (2.8%) at day 6, and 0/24 (0%) for the differentiated regulated genes in common at day 4 and 6 combined. Thus we can be confident that the changes in gene expression as a result of loss of epithelial Elf5 were biologically relevant. Interestingly, we identified a much larger gene set with epithelial loss of Elf5 at day 6 associated with a small false discovery rate. This suggests the transcriptional program of gene expression modulated by Elf5 was greatest at day 6 of pregnancy compared to day 4. We did however detect a small number of genes (24) that were regulated by Elf5 at days 4 and 6 combined (Table 4.1). One of these genes was the carboxyl ester lipase (Cel), and
98 Figure 4.11. Principal components analysis of Elf5-/- transplants. 34409 genes met an intensity criterion of a mean over 10 in at least one of the experimental conditions. Principal components analysis using these 34409 genes successfully separated the WT and Elf5-/- mammary epithelium at days 4 and 6 of pregnancy and the cleared fat pads.
99 was found to be decreasing genes in Elf5-/- mammary glands at days 4 and 6 of pregnancy. It was also enriched in the epithelium (p<0.001 WT vs CFP at Day 4 and 6, Appendix VIII and IX). Reduced expression of Cel was previously observed in Elf5+/- mammary glands at 18dpc and 1dpp (Zhou et al., 2005), thus the presence of this gene in this set, is confirmation that our approach was reliable. In addition, the identification of this gene in the tightly regulated set of 24 commonly regulated genes implies that Cel is maybe under direct transcriptional control by Elf5, although more investigation is needed to confirm this hypothesis. To understand the functional groups over-represented by the differentiated genes as a result of loss of Elf5, we analysed the genes changing with a p<0.001 at days 4 and 6 of pregnancy using Ingenuity Pathways Analysis (http://www.ingenuity.com/). The functional groups with the highest significance are illustrated in Figure 4.12. A list of genes identified in the top 4 over-represented functional categories that were decreasing or increasing is given in Appendix X and Appendix XI respectively. Interestingly, many of the decreasing genes were associated with processes involved in lipid synthesis. We also investigated the expression of the panel of milk genes identified by Rudolph and colleagues (Rudolph et al., 2003). -Casein did not make the cut off intensity criterion of a mean over 10 in at least of the experimental replicates, thus was not detected. This is not surprising as it is the last casein to be expressed in the mammary gland during pregnancy. We detected decreased expression of most of the caseins, Btn1a1 and Cel as a result of Elf5. Lalba and Wap, genes expressed from mid-pregnancy, did not show significant differential expression in Elf5-/- mammary transplants. Thus, using transcript profiling we confirmed that Elf5 was a critical mediator of milk protein gene transcription and lipid synthesis.
100 FC Probe Set ID Gene Title Gene Symbol FC day4 day6 Mammary gland RCB-0526 Jyg-MC(A) cDNA, RIKEN full-length enriched library, clone:G830022P11 1425583_at product:unclassifiable, full insert sequence --- -5.59 -4.81 1425500_x_at Hypothetical protein LOC625794 LOC625794 -4.42 -2.99 1417257_at Carboxyl ester lipase Cel -3.70 -3.51 1426851_a_at Nephroblastoma overexpressed gene Nov -2.55 -1.78 1460372_at Dual oxidase maturation factor 1 Duoxa1 -1.67 -1.52 1418762_at CD55 antigen Cd55 -1.44 -1.57 1426153_a_at Desmoglein 2 Dsg2 1.49 1.49 1417569_at Neurocalcin delta Ncald 1.62 1.58 ST6 (alpha-N-acetyl-neuraminyl-2,3-beta-galactosyl-1,3)-N- 1417616_at acetylgalactosaminide alpha-2,6-sialyltransferase 2 St6galnac2 1.73 2.53 1455301_at Expressed sequence BQ952480 BQ952480 1.76 1.71 1417089_a_at Creatine kinase, mitochondrial 1, ubiquitous Ckmt1 1.79 1.88 1418091_at Transcription factor CP2-like 1 Tcfcp2l1 1.90 2.79 1450197_at Retinal pigment epithelium 65 Rpe65 1.91 2.08 1452514_a_at Kit oncogene Kit 2.23 2.91 1419373_at ATPase, H+ transporting, lysosomal V1 subunit B1 Atp6v1b1 2.24 3.37 1435998_at Gene model 288, (NCBI) Gm288 2.36 2.14 1459253_at ------2.58 3.47 1422178_a_at RAB17, member RAS oncogene family Rab17 2.64 3.15 1424890_at Basonuclin 1 Bnc1 2.91 1.95 1449203_at Solute carrier organic anion transporter family, member 1a5 Slco1a5 3.02 3.61 1426904_s_at DnaJ (Hsp40) homolog, subfamily C, member 10 Dnajc10 3.70 4.42 1452230_at DnaJ (Hsp40) homolog, subfamily C, member 10 Dnajc10 4.17 4.05 1426905_a_at DnaJ (Hsp40) homolog, subfamily C, member 10 Dnajc10 4.21 4.88 3110079O15Ri 1453089_at RIKEN cDNA 3110079O15 gene k 5.31 4.85 Table 4.1. Differentially regulated genes at day 4 and 6 of pregnancy in Elf5-/- mammary transplants. Genes with an unadjusted p<0.001 were compared between Elf5-/- and WT mammary transplants at day 4 and 6 of pregnancy. 24 genes were commonly differentially regulated at both days 4 and 6. The Affymetrix ID, gene title, symbol and the fold changes at day 4 and 6 of pregnancy are tabulated. Green indicated decreasing genes and orange indicates increasing genes.
101 Figure 4.12. Functional groups significantly over-represented using differentially expressed genes identified by loss of epithelial Elf5 at days 4 and 6 of pregnancy. Bar graphs of the significant functional categories over-represented by differentially expressed genes in Elf5-/- mammary transplants at days 4 and 6 of pregnancy. Ingenuity Pathways Analysis.
102 Figure 4.13. The expression of milk genes in mammary epithelium without Elf5 at days 4 and 6 of pregnancy. Bar graph of the differential expression of the milk genes identified by Rudolph and colleagues (Rudolph et al., 2003). The milk genes -Casein, Lalba and Wap are expressed at later time points of pregnancy, and do not show significant changes. Decreased expression of most of the caseins, Btn1a1 and Cel all showed reduced expression as a result of Elf5.
103 4.2.9. Forced expression of Elf5 in pubertal mammary glands resulted in differentiation of the terminal end bud. Our results indicated that Elf5 was a critical modulator of alveolar differentiation pregnancy. To determine if elevation of mammary epithelial cell Elf5 was sufficient to induce differentiation and milk protein expression in virgin mammary epithelium, we used the MTB system for mammary-specific and tetracycline inducible gene expression (Gunther et al., 2002). MTB mice carry the reverse tetracycline-dependent transactivator (rtTA) under promotional control of MMTV, conferring mammary specific expression to the Tet system. These mice were obtained as a gift from Prof. Lewis Chodosh and were crossed with transgenic mice constructed using a transgenic cassette assembled in our lab by Mr Michael Kazluaskas, designed to express Elf5 from a Tet-operator and incorporating an IRES-EGFP as a marker for expression. The animals were produced by Ozgene at the Animal Resource Center Perth, Western Australia by traditional oocyte injection techniques. Elf5 expression was induced in bi-transgenic mice with oral administration of Doxycycline (Dox) for 3 weeks. UV excitation of mammary whole mounts resulted in robust green fluorescence throughout the mammary epithelium (Figure 4.14 C and D). Elf5 mRNA expression was induced up to 29 fold in mice treated with Dox compared to untreated control mice (data not shown). At the onset of puberty TEBs form at the ends of the ducts, which proliferate in response to cycling ovarian hormones resulting in extension of the mammary epithelium throughout the fat pad (Hinck et al., 2005). We forced the expression of Elf5 in bi- transgenic mice at 5 weeks, a time where TEBs are present and are facilitating epithelial expansion. At 8 weeks, the mammary epithelium in untreated control bi-transgenic mice had almost extended throughout the fat pad (Figure 4.15 A). In contrast, the extension of mammary epithelium from Dox-treated bi-transgenic mice was restricted (Figure 4.15 D). TEBs were visible at the ends of elongating ducts in untreated control mice, as expected, as the epithelium had not yet reached the periphery of the fat pad (Figure 4.15 B). TEBs were also present in Dox-treated bi-transgenic mice, however they displayed an abnormal morphology (Figure 4.15 E). Histology revealed that these abnormal terminal end buds had lost their unidirectional morphology (Figure 4.15 F), which is evident in normal TEBs (Figure 4.15 C). In addition both the cap and body cell layers of the TEB were absent.
104 Figure 4.14. Doxycycline treatment of bi-transgenic mice carrying the MTB and the TetON-Elf5-IRES-EGFP construct resulted in robust green fluorescence of the mammary epithelium. Mammary whole mounts of MTB-Elf5 transgenic mice with no Dox (A and B) and treated with Dox (C and D) at 8 weeks of age. Bright field images of Carmine stained epithelium (A and C) and fluorescence under UV excitation (B and D). There was no green fluorescence observed in mice without Dox treatment, in comparison to mice that were treated with Dox, which showed robust green fluorescence throughout the epithelium.
105 Figure 4.15. Doxycycline treatment of MTB-Elf5 bi-transgenic mice resulted in stalled ductal elongation and differentiation of the terminal end bud. Mammary whole mounts of MTB-Elf5 transgenic mice with no Dox (A and B) and treated with Dox (C and D) at 8 weeks of age. Dox treatment of bi-transgenic mice resulted in stalled epithelial extension throughout the fat pad (D) and irregular terminal end buds (E). Haematoxylin and eosin histochemistry of terminal end buds (C and F) demonstrated that Dox treatment in bi-transgenic mice resulted in abnormal TEB morphology (F) and alveolar-like differentiation (F asterixis).
106 Instead, single cell, circular, alveolar-like projections were visible where the body cells should be (Figure 4.15 F asterisk). Thus, forced expression of Elf5 during ductal morphogenesis, resulted in stalled epithelial extension and forced the apparent differentiation of the TEBs.
4.2.10. Forced expression of Elf5 in mature mammary epithelium resulted in alveolar differentiation and milk protein expression. In order to determine the functional significance of Elf5 expression in the virgin mammary gland, we forced the expression of Elf5 from 7 weeks of age, a time where the mammary epithelium has almost reached the periphery of the fat pad. At 11 weeks, the mammary ducts around the nipple were grossly distended (Figure 4.16 A). Distended ducts displayed high EGFP expression, indicative of Elf5 expression (Figure 4.16 B). We believed that the ductal distension was due to the engorgement of milk around the nipple, a process that marks the onset of the second stage of lactogenesis (Neville et al., 2002). Thus we used western analysis to investigate whether the expression of milk proteins was increased in Dox-treated bi-transgenic mice (Figure 4.16 B). Milk protein expression was robustly induced in bi-transgenic mice treated with Dox (Figure 4.16 C). Quantitative PCR demonstrated that mRNA expression of Wap and -Casein was induced ~1500 and ~1000 fold respectively (data not shown). Thus, expression of Elf5, in the absence of pregnancy is sufficient for epithelial differentiation and secretory activation.
4.2.11. Retroviral re-expression of Elf5 rescued alveolar morphogenesis in Prlr-/- primary mammary epithelial cells. Our findings suggested that Elf5 might be one of the primary transcription factor regulating alveolar morphogenesis. Also, the similarities between the phenotypes observed in Elf5-/- and Prlr-/- transplants, implied that Elf5 was a critical modulator of Prlr signalling during alveolar morphogenesis. Thus, to test whether Elf5 could compensate for Prlr during alveolar morphogenesis, we used genetic complementation. The technical aspects of this work was a collaboration among Dr. Prue Stanford, virus production and Elf5 immunohistochemistry, Jessica Harris, vector construction and Samantha Oakes, MEC
107 Figure 4.16. Doxycycline treatment of 7-week-old MTB-Elf5 bi-transgenic mice resulted ductal distension and milk production. Mammary whole mounts of MTB-Elf5 transgenic mice treated with Dox, bright field image, Carmine stain (A) and dark field image, green fluorescence (B). Dox treatment of 7- week-old bi-transgenic mice resulted in distension of the ducts around the nipple (A), which correlated with strong green fluorescence (B). Western analysis of milk protein expression demonstrated that Dox treatment in bi-transgenic mice resulted in massive over- expression of milk protein compared to untreated bi-transgenic mice.
108 transplantation. All three were also involved in the analysis of results. Isolated primary MECs injected into a cleared fat pad will form a normal mammary outgrowth and can undergo alveolar morphogenesis (Smith, 1996). MECs isolated from Prlr-/- mice formed a normal ductal outgrowth but could not undergo alveolar morphogenesis during pregnancy, as evident in mammary whole mounts at 1dpp (Figure 4.17 D). Alveoli failed to expand and distend, and epithelial cells did not contain cytoplasmic lipid droplets (Figure 4.17 E). Milk expression was also severely impaired, observed using immune-histochemistry against milk protein (Figure 4.17 F). To test whether Elf5 could compensate for loss of Prlr in primary mammary epithelial cells, we re-expressed Elf5 in Prlr-/- MECs using the PolyPOZ retrovirus (Naylor et al., 2000) encoding Elf5 (Harris et al., 2006). In 4 out of 28 successful Prlr-/- MEC engraftments, lobuloalveolar development was completely rescued by re-expression of Elf5 (Figure 4.15 G and H). Immunohistochemistry against milk proteins demonstrated that re- expression of Elf5 in Prlr-/- MECs resulted in full secretory activation that was similar to that of the endogenous epithelium (Figure 4.17 I). Interestingly in a further 4 MEC transplants, we also observed a chimeric pattern within the mammary epithelium (Figure 4.17 J). Advanced alveolar lobules were observed in some areas of the epithelium, as demonstrated by histology (Figure 4.17 K) and immunohistochemistry to milk protein antibody (Figure 4.17 L). However, in other areas of the same ductal tree there was areas of failed alveolar morphogenesis, demonstrated by failed expansion of the alveolar epithelium and reduced milk protein expression (Figure 4.17 K and L arrow). Areas of complete rescue were associated with the greatest immunoreactivity to Elf5 (Figure 4.18 E and F), which was similar to endogenous epithelium (Figure 4.18 G and H). This was in contrast to areas of incomplete rescue (Figure 4.18 C and D), which displayed Elf5 staining intermediate between Prlr-/- MECs (Figure 4.18 A and B) and endogenous epithelium (Figure 4.18 G and H). Interestingly, the pattern of Elf5 staining was heterogenous in the luminal epithelial cells of Prlr-/- MECs and Prlr-/- MECs with partial rescue, with an absence of Elf5 staining in some of the luminal epithelial cells (Figure 4.18 B and D arrows). This is similar phenotype to that observed in areas of incomplete lobuloalveolar development in Elf5+/- transplants (Figure 4.5 I arrows). The infected MECs used in these mammary transplants were from a heterogenous population of Elf5 expressing and Elf5 non- expressing cells.
109 Therefore we hypothesise that the chimerism we observed in 4 of these transplants was due to both Elf5-negative and Elf5-positive stem cells contributing to the same mammary epithelial outgrowth. Our findings here and those demonstrated in the Elf5+/- transplants (Figure 4.5) suggest that full lobuloalveolar development requires homogeneous luminal cell expression of Elf5. In summary we have shown that Elf5 can completely compensate for Prlr during alveolar morphogenesis. Therefore Elf5 is a master regulator of lobuloalveolar development during pregnancy, and possibly a lobular cell fate determinant in the mammary gland.
110 Figure 4.17. Retroviral re-expression of Elf5 in Prlr-/- mammary epithelial cells rescues alveolar morphogenesis. Mammary whole mounts (A, D, G and J), haematoxylin and eosin histochemistry (B, E, H and K) and immune-histochemistry to milk (C, F, I and L) of endogenous epithelium (A- C), Prlr-/- mammary epithelial cell (MEC) transplants (D-L) and at 1dpp. Re-expression of Elf5 in Prlr-/- MECs results in rescued alveolar morphogenesis (G-I). In some MEC transplants, chimeric rescue was observed resulting from heterogenous infection of the Elf5 retrovirus in Prlr-/- MECs. Arrows (J-L) indicate areas of incomplete rescue.
111 Figure 4.18. Retroviral rescue of alveolar morphogenesis in Prlr-/- mammary epithelial by Elf5 was associated with increased Elf5 staining. Immune-histochemistry to Elf5 in Prlr-/- mammary epithelial cell (MEC) transplants (A and B), partial rescue (C and D), full rescue (E and F) and endogenous epithelium (G and H) at 1dpp. Prlr-/- mammary epithelial cell (MEC) transplants without Elf5 displayed low levels of Elf5 staining (A and B), and had a punctate pattern in the luminal epithelium. Prlr-/- mammary epithelial cell (MEC) transplants without Elf5 and areas of partial rescue in Prlr-/- MECs with Elf5 displayed heterogeneous staining of Elf5, and contained cells with no expression of Elf5 (B and D arrows). In contrast, areas of complete rescue in Prlr-/- MECs with Elf5 (E and F), displayed strong and homogeneous Elf5 staining similar to what is observed in endogenous epithelium (G and H).
112 4.3 Discussion In this chapter we have investigated the role of the Ets transcription factor Elf5 in mammary gland development and compared this to mammary epithelial cell loss of Prlr. Prl binding to its receptor in the mammary epithelium triggers a cascade of events, which leads to the activation of the Jak/Stat5 pathway resulting in the transcription of a number of genes important in alveolar morphogenesis during pregnancy (Hennighausen et al., 2005). The Elf5 transcript was down regulated in Prlr-/- mammary glands during early pregnancy, and up regulated with Prl treatment in SCp2 cells (Harris et al., 2006; Naylor et al., 2003a). In addition, quantitative PCR demonstrated reduced mRNA expression of Elf5 in Prlr+/- mammary glands, with no similar decrease in expression of Prlr mRNA in Elf5+/- (Zhou et al., 2005). Thus, Elf5 appears to be downstream of the Prlr. A cell autonomous role for Elf5 during alveolar morphogenesis was previously demonstrated using mammary transplants of Elf5+/- epithelium (Zhou et al., 2005). Lobuloalveolar development was impaired in Elf5+/- mice and they failed to nurse pups on their first and often subsequent pregnancies. In addition, proliferation and the expression milk protein genes Wdnm1, -casein, Cel and Wap were reduced in Elf5+/- glands. Thus Elf5 is required for both alveolar proliferation and differentiation. These experiments were performed using Elf5 heterozygous mice, as homozygous mice are embryonic lethal. To investigate the effects of homozygous loss of Elf5 within the mammary gland, we used an Elf5-/- model generated independently of the above studies, in which embryonic lethality was rescued by tetraploid rescue (Donnison et al., 2005). We compared the phenotypes of Elf5-/- and Prlr-/- epithelium using mammary transplantation to investigate the contribution of Elf5 to Prl mediated alveolar morphogenesis. Mammary transplantation involves taking a small biopsy of mammary epithelium from a donor mouse and transplanting this into the cleared fat pad of an immune- compromised host (DeOme et al., 1959). A mammary outgrowth will form from this biopsy and undergo full ductal morphogenesis to fill the mammary fat pad. This ability is dependent on mammary epithelial stem cells within the biopsy (Kordon et al., 1998; Shackleton et al., 2006; Stingl et al., 2006). Mammary transplants of Elf5-/- epithelium were able to generate full mammary outgrowths in both primary and secondary generations, thus
113 loss of epithelial Elf5 does not dramatically alter the regenerative capacity of the primary mammary stem cells, at least during ductal morphogenesis. Ductal morphogenesis is dependent on estrogen E and Pg (Sternlicht, 2006). Prl indirectly modulates ductal side branching, as it modulates the release of Pg via Prlr receptors within the corpus luteum of the ovaries (Binart et al., 2000). Therefore ductal side branching of Prlr-/- mammary transplants was normal ((Brisken et al., 1999), Figure 1.4), but side branching was absent in Prlr-/- animals unless treated with exogenous progesterone (Binart et al., 2000). In this chapter, we report that epithelial Elf5 was also not required in the epithelium for ductal morphogenesis (Figure 4.1), as ductal bifurcation during elongation, and ductal side branching and TEB end bud regression in Elf5-/- epithelial transplants was normal. In our laboratory we have demonstrated that Elf5 is regulated by Progestin (a synthetic Pg) in T47D human breast cancer cells, thus our results demonstrate that Elf5 in the mammary epithelium is not a mediator of Pg induced ductal morphogenesis in the normal mammary gland in the non-pregnant state. In contrast, mammary epithelium lacking Elf5 results in failed alveolar morphogenesis that resembles loss of mammary epithelial Prlr. Proliferation of the alveolar epithelium is dependent on signals from the Pgr in synergy with Prl signalling. Pg may signal via paracrine intermediaries such as Wnt4 (Brisken et al., 2000) or RankL (Conneely et al., 2003), resulting in the proliferation of adjacent steroid receptor negative epithelial cells. Transcript profiling experiments performed in our laboratory have indicated that Prlr may also signal via RankL (Ormandy et al., 2003), potentially resulting in Cyclin D1 transcription. Prl can also signal to Cyclin D1 via Igf2, resulting in the proliferation of primary mammary epithelial cells (Brisken et al., 2002). In addition, Prlr signalling through MapK may be important for the mitogenic action of Prl in mammary cells (Das et al., 1996a; Das et al., 1996b). Work from our laboratory has demonstrated a role for extra- pituitary Prl in the proliferation of the alveolar epithelium at the end of pregnancy (Naylor et al., 2003b). In this chapter, we have demonstrated that proliferation of the alveolar epithelium was similarly reduced in Prlr-/- and Elf5-/- mammary transplants during early pregnancy (4.6 and 4.7). This work strongly confirms the observations that proliferation was reduced in Elf5+/- mice (Zhou et al., 2005). In addition, the similarities of failed alveolar proliferation between epithelium lacking Elf5 and that without Prlr, suggest that
114 Elf5 may involved in the modulation of Prl mediated proliferation of the alveolar epithelium. Lobuloalveolar development was severely impaired and milk protein expression was significantly reduced in epithelium lacking Prlr and Elf5. Furthermore, we discovered a greater reduction in the immunoreactivity to a milk antibody in Elf5-/- transplants compared to Prlr-/- transplants. The milk antibody used in this chapter (Figure 4.6 and 4.9), recognises the milk proteins -Casein, -Casein and Wap (Naylor et al., 2003a). Thus Elf5 is a critical regulator of milk protein synthesis during pregnancy, as the absence of Elf5 in the epithelium results in almost no expression of these milk proteins during pregnancy. In Chapter 3, we demonstrated that Stat5 in HC11 cells regulated Elf5 transcription, thus Elf5 appears to be downstream of Stat5. There are also several Stat5 binding sites (GAS elements) in the promoter region of Elf5 (Figure 4.19), suggesting that Stat5 may bind to the promoter of Elf5. Stat5 binds to the promoter region of target genes such as -Casein (Wakao et al., 1995) and WAP (Pittius et al., 1988), which in turn drives their expression. Interestingly, members of the Ets transcription factor family (Ets1 and Ets2) can form protein interactions with Stat5 in T cells (Rameil et al., 2000). The ability of Ets transcription factors to form complexes with other structurally unrelated proteins is a characteristic of this family (Verger et al., 2002). It is also quite clear that transcription of target genes is not simply a mechanism involving a single transcription factor binding to its gene target (Rosenfeld et al., 2006). Therefore, we propose that although Elf5 appears to be regulated by the Stat5 pathway, it is possible that a post-translational transcriptional complex forms between Stat5a and Elf5, which is responsible for the transcription of the genes such as Wap and -Casein. The transcription factor Cepb is also required by the epithelium for mammary differentiation and milk gene transcription (Robinson et al., 1998; Seagroves et al., 1998), thus may also be involved in similar transcription complex. To understand the transcriptional program mediated by Elf5 during alveolar morphogenesis, we performed transcript profiling of Elf5-/- epithelium at days 4 and 6 of pregnancy. An analysis of the gene expression patterns revealed that Elf5 regulates genes involved in both milk protein and lipid synthesis. We confirmed the regulation of the important milk gene Cel by Elf5 observed in the studies by Zhou and colleagues (Zhou et al., 2005). Thus, Cel appears to be an important target of Elf5 mediated transcription and
115 Figure 4.19. The genomic sequence of Elf5 and the location of GAS elements (Stat5 binding sites). The genomic sequence of the Ets transcription factor Elf5 (NC_000068). Alternating introns and exons are shown in blue and pink respectively. The location of the ATG (black arrow) and the coding sequence is indicated. 5 putative Stat5 binding GAS elements (sequence TTC nnn GAA) were detected, and indicated by the orange stars. The distances between the axes marks are 2000bp.
116 requires further investigation. The lipolytic enzyme Cel is abundant in milk and hypothesised to substitute for pancreatic enzyme for nutrient digestion in the infant prior to the maturation of the pancreas (Hui et al., 2002). The promoter region of Cel contains 2 Stat5 GAS binding sites, as well as binding sites for CTF/NF-1, Oct-1, Ets 1, YY 1, Cebp and the glucocorticoid receptor (Kannius-Janson et al., 1998), suggesting possible cis- regulation by various transcription factors. The CCAAT-binding transcription factor (CTF)/nuclear factor 1 (NF1) (Kannius-Janson et al., 1998) binds to the promoter of Cel to drive its expression. NF1-C2 may also be regulated by Prlr signalling during pregnancy (Johansson et al., 2005). Thus the transcription of Cel may occur as a result of a transcriptional complex such as the one described above and may include NF-1 and Elf5. Alternatively or in addition, Elf5 may drive the transcription of other transcription factors involved in lobuloalveolar differentiation resulting in the expression of milk genes such as Cel, and act close to origin of the transcriptional hierarchy mediating alveolar morphogenesis. We therefore propose a model for Elf5 action during alveolar morphogeneis in Figure 4.20. We hope to continue these studies and directly compare the transcript profiles of Elf5-/- and Prlr-/- epithelium. These studies will be important in understanding the contribution of the Elf5 transcriptional regulation to Prlr signalling. The difference in the immune reactivity to milk we observed in Prlr-/- transplants is possibly due to low but not absent expression of Elf5 in epithelium lacking the Prlr. Although Elf5 expression was significantly reduced in Prlr-/- mammary glands at 2, 4 and 6 of pregnancy, it was still detected above levels observed in the cleared fat pads (Harris et al., 2006). This reduced level of expression of Elf5 in Prlr-/- mammary glands, was clearly not sufficient for alveolar morphogenesis, but perhaps enabling for some milk protein expression observed in Figure 4.8. Further evidence to support this, is the observation of chimeric lobuloalveolar development within individual Elf5+/- transplants (Figure 4.4) and chimeric rescue within Prlr-/- mammary epithelial cells re-expressing Elf5 (Figure 4.17). We observed a heterogenous pattern of Elf5 immunoreactivity in Elf5+/- transplants (Figure 4.5 I) and partially rescued Prlr-/- MECs at 1dpp (Figure 4.18 B), compared to WT (Figure 4.5 G) and endogenous glands (Figure 4.18 F and H) at 1dpp. thus homogeneous expression of Elf5 is required for full lobuloalveolar development and milk protein expression.
117 Figure 4.20. A model of transcriptional control by Elf5 Elf5 appears to be downstream of Prlr and transcribed by Stat5. Various transcriptional mechanisms may lead to complete alveolar morphogenesis. (A) Stat5 transcription of Elf5, other transcription factors (TFs) eg NF1 and the milk (MGs) and lipid genes (LGs) resulting in alveolar morphogenesis. (B) Upon transcription by Stat5, Elf5 transcribes other transcription factors, milk genes, lipid genes, Cel and NF-1 involved in alveolar morphogenesis. (C) Elf5 is part of a greater transcription factor complex resulting in the transcription of genes involved in alveolar morphogenesis including Cel and the milk and lipid genes. (D) Various transcription factors transcribe genes involved in specific processes during alveolar morphogenesis. Specific transcription factor binding sites are indicated by coloured circles.
118 These findings suggest that alveolar morphogenesis is dependent on robust and uniform expression of Elf5 throughout the luminal epithelium. The observation of heterogeneous Elf5 staining in WT mammary epithelium at early pregnancy and uniform expression at late pregnancy, suggests that at some point during pregnancy, Elf5 is switched on in most of the alveolar epithelial cells. The mammary stem cell hierarchy is only beginning to be defined (Smalley et al., 2003), but it is believed that commited lobular and luminal progenitors exist in the mammary epithelium (Kordon et al., 1998), downstream of a multipotent and self renewing single mammary stem cell (Shackleton et al., 2006; Stingl et al., 2006). Our findings here may suggest that Elf5 is a critical cell fate determinant driving the differentiation of an Elf5 negative lobular progenitor into an Elf5 positive differentiated alveolar cell. To test this hypothesis, we are now investigating the effects of homozygous loss of Elf5 on the mammary stem cell populations. Understanding how Elf5 modulates the various stem cell populations in the mammary epithelium is an important step as mammary stem cells that maintain self-renewal capacity, appear to be important targets for carcinogenesis (Smalley et al., 2003). It is clear that the heterogeneous patterning of Elf5 is more severe in Elf5+/- mammary transplants, which is associated with failed lobuloalveolar development. Mammary cell patterning in the luminal epithelium is a critical part of lobuloalveolar development (Grimm et al., 2002). Pg and Prlr expression is restricted in a subset of epithelial cells, and is controlled by Cebp (Robinson et al., 1998; Seagroves et al., 2000). Elf5 expression in the luminal epithelium may also be restricted to the steroid receptor positive cells during the earliest phases of pregnancy. Elf5 haploinsufficiency may result in further reduction in the number of Elf5 positive cells, and thus upon transplantation of a tissue fragment with an uncontrolled number of epithelial cells, this may result in the variable phenotype observed in Elf5+/- transplants. It also possible that gene silencing initiated by epigenetic events in mammary glands without sufficient Elf5, contribute this phenotype. In addition, heterogeneity in Elf5+/- transplants may be the result of alterations in the number of luminal and/or alveolar progenitors due to loss of a single copy of Elf5, within the mammary donor fragment in the previous paragraph.
119 The findings in this chapter indicated that Elf5 appears to be the limiting factor during alveolar morphogenesis. To test whether the expression of Elf5 alone, in the absence of the pregnancy induced increases in Prl and Pg, we forced the expression of Elf5 in virgin mammary glands (Figure 4.14-4.16). These experiments resulted in apparent pre-mature differentiation of the TEBs in expanding epithelium and robust expression of the milk proteins Wap and -Casein in mature epithelium. We also demonstrated that the expression of Elf5 could completely rescue alveolar morphogenesis in epithelium lacking the Prlr (Figure 4.17). Thus we demonstrated that a single transcription factor, Elf5, was sufficient for alveolar differentiation. These findings are quite extraordinary, and suggest that Elf5 is the transcription factor downstream of Prl signalling, which controls the hierarchy of gene expression during alveolar morphogenesis, and possibly also a cell fate determinant for alveolar epithelium. In summary we have demonstrated that mammary epithelial loss of Elf5 is a phenocopy of mammary epithelial loss of Prlr. Elf5 is required by the epithelium for full lobuloalveolar development but not ductal morphogenesis. Elf5 controls the expression of genes involved in alveolar morphogenesis and we hypothesise that Elf5 interacts with Stat5 signalling to control the expression of milk genes. Elf5 is the primary transcription factor mediating Prl signalling within the mammary epithelium, driving the development of lobuloalveoli. In addition, robust and uniform expression of Elf5 in the luminal epithelium is required to switch the epithelium to an alveolar cell fate, and Elf5 expression alone, in the absence of pregnancy, forces epithelial cell differentiation. Thus we believe that Elf5 is a master regulator of alveolar morphogenesis.
120 Chapter 5 Prolactin in mammary gland carcinogenesis 5.1 Introduction In the previous chapters we have investigated the genes essential for alveolar morphogenesis and secretory activation. We have demonstrated that the Prl target Elf5 is a master regulator of alveolar morphogenesis. We have also identified genes involved in mammary secretory activation during pregnancy. It is important to understand the processes involved in the growth and differentiation of the normal mammary gland, as the dysregulation of these mechanisms may lead to the development of cancer. Examples include over-expression of the epidermal growth factor receptor (Erbb2) (Slamon et al., 1987), disregulation of the steroid receptors (Anderson, 2002), and disruption of the cell cycle in breast cancer (Zafonte et al., 2000). Therapies aimed at restricting the activities of ER (Tamoxifen) and Erbb2 (Trastuzumab), are effective in the treatment of breast cancers characterized by increased expression of ER and Erbb2 respectively (Navolanic et al., 2005). Thus understanding the mechanisms involved in mammary development is a critical step in developing agents useful in the treatment of disease. Our research and that of others, has demonstrated the role of Prl and its molecular targets during alveolar morphogenesis, and a large body of literature exists implicating the role of Prl in breast cancer (Reviewed in (Clevenger et al., 2003)). A role of Prl during carcinogenesis has been recognised since the 1970’s (Welsch et al., 1977). In addition, elevated serum Prl levels were associated with breast cancer in several studies (Bird et al., 1981; Hill et al., 1976; Kwa et al., 1981). However, it was not until results of the large US Nurses Health Study (http://www.channing.harvard.edu/nhs/index.html) was published, that the association between serum PRL levels and breast cancer risk was seriously considered (Hankinson et al., 1999). Plasma PRL levels were analysed in 306 post-menopausal patients diagnosed with breast cancer after blood collection compared to 448 control subjects. Patients with the highest quartile serum PRL levels compared to those with the lowest quartile serum PRL levels had a two-fold higher relative risk of developing breast cancer (Hankinson et al., 1999). This study was updated with an additional 545 postmenopausal patients with breast cancer and 827 matched controls in 2004 (Tworoger et al., 2004). The same positive correlation between breast cancer risk and serum PRL levels
121 was observed with a significant correlation found in patients that developed ER+ PGR+ tumours. In patients who developed ER- tumours there was a reduction in relative risk to 0.75, indicating that raised PRL levels are not a risk for ER- cancers, or could be slightly protective. Prl is also a mitogen for breast cancer cells (Clevenger et al., 2003), and Prl treatment results in an increase proliferation (Das et al., 1997; Ginsburg et al., 1995; Llovera et al., 2000) and cyclin D1 expression (Schroeder et al., 2003) in breast cancer cell lines selected for prolactin sensitivity (Schroeder et al., 2002). Likewise the use of PRLR antagonists can reduce proliferation (Chen et al., 1999; Goffin et al., 1996b; Llovera et al., 2000) in some breast cancer cell lines. In comparison to the mitogenic effects of estrogen and growth factors on breast cancer cell lines the mitogenic effects of PRL are weak. In rodent models the converse is true, raised Prl is a potent and complete carcinogen in these models (Welsch et al., 1977; Wennbo et al., 2000). For example, Prl and Prlr mRNA were detected in nitrosomethylurea (NMU) carcinogen-induced tumours and a rat Prl antiserum inhibited NMU-induced tumour cell proliferation by up to 70%, compared to normal rabbit serum and GH antiserum (Mershon et al., 1995). Over-expression of rat Prl using the lipocalin promoter to drive expression predominantly in mammary epithelium produces estrogen receptor (ER) positive tumours at a higher rate than other mouse mammary cancer models (Rose-Hellekant et al., 2003). Mice expressing the polyoma middle-T antigen oncogene develop tumours in the first weeks of life, but when crossed with Prl knockout mice they developed tumours significantly later (Vomachka et al., 2000). Thus several in- vitro and in-vivo models using rodents have demonstrated a clear and unequivocal role for raised Prl as a complete carcinogen, however it is still unclear how prolactin exerts these this effects. It is not known whether Prl acts via the mammary epithelial cell or via modulation of the systemic hormone environment, and we do not know at what point during carcinogenesis Prl has its greatest effects. The aims of this chapter were to determine whether Prl acts directly via the epithelial cell to modulate the growth of cancer, and at what point during carcinogenesis Prl has its greatest effects. To achieve these aims, we have used a mouse model of mammary carcinogenesis. The C3(1)SV40T model of mammary cancer provides a reproducible series of defined neoplastic lesions that progress to invasive carcinoma (Figure 5.1) and resemble
122 the human disease (Green et al., 2000; Maroulakou et al., 1994; Shibata et al., 1998). Female mice homozygous for C3(1)/SV40T, develop mammary neoplasia in all mammary glands by 3 months of age (Figure 5.1 B, C, F G), which progresses to invasive carcinoma (Figure 5.1 D and H) by 6 months of age. We crossed the C3(1)/SV40T transgenic mouse with the Prlr-/- mouse, which lacks Prl function, and investigated tumor formation. We also used mammary gland transplantation of Prlr-/-/C3(1)/SV40T epithelium to understand whether Prl acts indirectly via the pituitary-ovarian axis or whether it acts directly by binding to Prlr within the mammary epithelium to modulate SV40T induced carcinogenesis.
123 Figure 5.1 Tumor formation in the C3(1)/SV40T model of mouse mammary carcinogenesis. Carmine stained whole mounts of the 4th mammary gland from wildtype (A) and C3(1)/SV40T mice (B-D), Carmine Stain. Haematoxylin and eosin histochemistry of mammary epithelium from wildtype (E) and C3(1)/SV40T mice (F-H). Low-grade neoplasia appears as aberrant proliferation of the mammary epithelium resulting in abnormal budding and branching of the epithelium (B) and an increase in the number and layers of luminal epithelial cells. Abnormal luminal cell proliferation continues resulting in filling of the ductal epithelium characteristic of high-grade neoplasia (Figure C and G). Eventually, cancer cells escape the confines of the basement membrane and invade the surrounding stroma, thus forming an invasive tumour (D and H).
124 5.2 Results 5.2.1. The latency to tumour formation was increased in C3(1)/SV40T without the Prlr. The development of palpable tumours was monitored in Prlr-/- and WT mice that carried the C3(1)/SV40T construct. C3(1)/SV40T mice which lacked Prlr had significantly increased latency to palpable tumour formation (200 ± 9 days) compared to control C3(1)/SV40T mice (175 ± 7 days, logrank p=0.033; Figure 5.2 A). Prlr-/-/C3(1)/SV40T mice also reached a tumour burden of 10% body weight significantly later (243 ± 15 days) compared to WT/C3(1)/SV40T mice (217 ± 15 days, logrank p=0.032; Figure 5.2 B). To determine whether Prlr signalling affected tumour growth rate, a mixed effects linear model was applied to the cubed root of tumour volume for each experimental animal group. Tumors that were filled with fluid at the ethical end point were excluded from the analysis. Results are plotted with day 0 as the day the tumour was detected (Figure 5.3). No significant difference in the rate of change in tumour volume was detected in Prlr-/- /C3(1)/SV40T mice compared to control C3(1)/SV40T mice (p=0.45).
5.2.2. The morphology of tumours was not significantly different between Prlr-/-/C3(1)/SV40T and WTC3(1)SV40T mice. Histological examination of tumour tissues collected at the ethical end point was undertaken (Table 5.1). Some of the smaller tumours exhibited areas that had not yet invaded through the basement membrane and represent a stage similar to human ductal carcinoma in situ. The invasive tumours demonstrated a high-grade morphology, high mitotic index, coarse chromatin structure, pleomorphic nuclei and foci of necrosis. Tumors often displayed more than one low power architectural pattern including acinar and/or glandular, papillary and solid areas. Acinar and glandular patterns were grouped together for analysis as they often tended to merge into each other rendering reliable distinction problematic. These pathologies have been described in detail previously (Cardiff et al., 2000). Cellular morphology demonstrated 2 main variants: firstly tumour cells with a high nucleocytoplasmic ratio, hyperchromatic nuclei, coarse chromatin and mild to moderate pleomorphism designated (Type A); and secondly cells with a lower nucleocytoplasmic ratio, a smaller amount of eosinophilic cytoplasm, more vesicular chromatin and more 125 Figure 5.2. The latency to tumor formation and the time to reach ethical end point was significantly increased in Prlr-/-/C3(1)/SV40T mice. (A) Tumor free survival curve (Kaplan-Meier). Time is represented in days after birth. Prlr- /-/C3(1)/SV40T mice (circles) develop palpable tumors significantly later compared to WT/C3(1)/SV40T mice (squares, p=0.033). (B) Survival curve. Time is represented in days after birth. Prlr-/-/C3(1)/SV40T mice (circles) reach the ethical end point of 10% tumor burden significantly later than WT/C3(1)/SV40T mice (squares, p=0.032).
126 Figure 5.3. There was no difference in rate of growth between tumours that formed in Prlr-/-/C3(1)/SV40T mice and those that formed in WT/C3(1)/SV40T mice. Tumor volume trellis plot. The cube root of tumor volume (volume1/3) is plotted with respect to time after detection (days). Time 0 is the day of initial detection. There was no significant difference in the rate of change in tumor volume from WT/C3(1)/SV40T mice (left box) and Prlr-/-/C3(1)/SV40T mice (right box) (p=0.45).
127 % area of tumor Cellular Structure
Acinar/ Epithelium Genotype n Glandular Papillary Solid Type A Type B WT/C3(1)/SV40T 30 17±3 11±3 72±5 67 33 Reached 10% -/- tumor burden Prlr /C3(1)/SV40T 29 18±3 19±5 63±6 74 26 p-value 0.86 0.19 0.25 Chi Square p=0.28
Did not reach WT/C3(1)/SV40T 12 23±4 16±7 61±8 77 23 10% tumor Prlr-/-/C3(1)/SV40T 20 15±3 10±4 75±6 67 33 burden p-value 0.14 0.46 0.17 Chi Square p=0.12 WT/C3(1)/SV40T 42 19±3 12±3 69±4 70 30 All tumors -/- collected Prlr /C3(1)/SV40T 49 17±2 15±4 68±4 71 29 p-value 0.49 0.47 0.84 Chi Square p=0.88 Table 5.1. Tumour morphology and cellular structure in tumors from WT/C3(1)/SV40T and Prlr-/-/C3(1)/SV40T mice. n; number of tumors. The average percent of tumor area classified as acinar and/or glandular, papillary and solid is given ± standard error. The percentage of tumors classified as having type A or type B cellular structure is also shown.
128 marked pleomorphism (Type B). Some tumours demonstrated intermediate forms and were classified according to the predominant features. Areas of necrosis were observed in 98% of tumours at collection. Statistical analysis of tumour classification in a blinded fashion revealed no significant difference in tumour architecture or cellular morphology between tumours derived from Prlr-/-/C3(1)/SV40T mice or control C3(1)/SV40T mice.
5.2.3. There was a trend towards reduced neoplasia in Prlr-/- /C3(1)/SV40T compared to WTC3(1)SV40T mice. Mammary whole mounts were analysed for the development of neoplastic lesions. Analysis of mammary whole mounts collected at the ethical end point (Figure 5.4) demonstrated that C3(1)/SV40T mice that lacked Prlr displayed a trend toward reduced lesion-area measured as a percentage of total mammary fat pad area (8.9 ± 1.3%) compared to WT/C3(1)/SV40T mice at ethical end point (13.6 ± 2.2%, p=0.08).
5.2.4. The mRNA expression of SV40T was unaltered in C3(1)/SV40T mice that lack the Prlr. Some mammary specific promoters such as MMTV and WAP are sensitive to pregnancy and/or hormone stimuli (Hutchinson et al., 2000). This complicates investigation of endocrine-mediated carcinogenesis as observed effects may simply be due to changes in transgene expression. The C3(1) rat prostatic steroid binding protein (PSBP) promoter used here is not steroid hormone responsive (Green et al., 2000; Shibata et al., 1998). To confirm that SV40T expression was not altered by Prlr genotype, quantitative real time PCR was used to examine the relative expression of SV40T mRNA. Expression of SV40T mRNA was detected in all mammary glands from 12 week C3(1)/SV40T inguinal mammary glands (Figure 5.5). There was no significant difference in the relative expression of SV40T between mammary glands from 12 week old WT/C3(1)/SV40T (18 ± 0.5 units) and Prlr-/- /C3(1)/SV40T mice (18 ± 1.2 units, t-test p=0.94).
129 Figure 5.4. There was a trend towards reduced neoplasia in the mammary glands of Prlr-/-/C3(1)/SV40T mice. Bar graph summary of mammary whole-mount analysis. Prlr-/-/C3(1)/SV40T mice (white bars) had significantly reduced area of lesions as a percentage of total mammary fat pad area compared to WT/C3(1)/SV40T mice (black bars, p=0.046). The area of lesions when classified into neoplasia and tumor was reduced in Prlr-/-/C3(1)/SV40T mice compared to WT/C3(1)/SV40T mice although this was not statistically significant (p=0.13 and p=0.12 respectively).
130 Figure 5.5. The mRNA expression of SV40T was not altered in the mammary glands of Prlr-/-/C3(1)/SV40T mice. Bar graph of relative SV40T mRNA expression. There was no significant difference in the relative expression of SV40T mRNA in the inguinal mammary glands from WT/C3(1)/SV40T mice and Prlr-/-/C3(1)/SV40T (p=0.94)
131 5.2.5. Body weight and mammary ductal side branching was reduced in mice that lack the Prlr. To determine if Prlr-/- mice differ in body weight compared to control mice, both WT and C3(1)/SV40T animals were aged and weighed weekly (Figure 5.6). A mixed effects linear model demonstrated that both Prlr-/- and Prlr-/-/C3(1)/SV40T animals gained weight at a reduced rate compared to WT and WT/C3(1)/SV40T mice (p=0.006 and p=0.005 respectively; Figure 5.6). Control Prlr-/- mice were approximately 17% lighter (average 27.8 ± 0.8g) at 50 weeks of age compared to control WT mice (average 33.6 ± 1.3g). Reduced body weight in female Prlr-/- mice is due to reduced abdominal fat stores via a mechanism that includes altered endocrine environment (Freemark et al., 2001) and possibly Prlr expression by adipocytes (Ling et al., 2000). Thus the changes in body weight we observed. We collected mammary whole mounts from 12 week old mice without the C3(1)/SV40T transgene. Ductal side branching was significantly reduced in the mammary glands of Prlr-/- animals (Figure 5.7 B) compared to WT (Figure 5.7 A), as reported previously in prolactin and prolactin receptor knockout (Horseman et al., 1997; Ormandy et al., 1997b). These results are due to Prl modulation of progesterone levels via the pituitary- ovarian axis (Binart et al., 2000). These changes in body weight and mammary epithelial cell content potentially confound our results regarding altered tumour latency. These problems also potentially confound the results obtained in many other rodent models of prolactin action. Increased prolactin levels produced via pituitary graft, prolactin injection or transgenic methods, or loss of prolactin produced by knockout or pituitary ablation, may have altered the systemic hormonal environment causing undetected changes in body weight and mammary epithelial cell number. To investigate this problem we utilised mammary epithelial transplantation. This procedure rescues the defect in ductal side branching and negates the body weight issue by placing the test glands in a normal endocrine environment (Brisken et al., 1999).
132 Figure 5.6. Body weight is significantly reduced in mice that lack the Prlr. Body weight trellis plot for WT (top left) and Prlr-/- (top right) mice and WT/C3(1)/SV40T (bottom left) and Prlr-/-/C3(1)/SV40T (bottom right) mice. Age (days) is represented on the horizontal axis and body weight (grams) on the vertical axis. The increase in body weight is significantly more gradual in Prlr-/- and Prlr-/-/C3(1)/SV40T mice compared to control WT and WT/C3(1)/SV40T mice (p=0.006 and p=0.005 respectively).
133 Figure 5.7. Mammary ductal side branching is reduced in mice that lack the Prlr. (A) Prlr-/-/C3(1)/SV40T inguinal mammary gland at 12 weeks, Carmine stain. (B) WT/C3(1)/SV40T inguinal mammary gland at 12 weeks, Carmine stain. Ductal elongation and bifurcation was normal in Prlr-/- mice, however secondary lateral side branches were significantly reduced. This is due to the pituitary-ovarian action of Prl during ductal morphogenesis (Brisken et al., 1999).
134 5.2.6. The latency to tumour formation was increased in C3(1)/SV40T transplants without the Prlr. Mammary glands made from Prlr-/- epithelium developed palpable tumours significantly later than mammary glands with WT/C3(1)/SV40T epithelium (289 ± 22 days verses 236 ± 24 days, logrank p<0.001; Figure 5.8). We looked directly for an effect of Prlr genotype on tumour growth rate using a mixed effects linear model described above (Figure 5.9). Tumors that were filled with fluid at ethical end point were excluded from the analysis. Overall there was no significant difference in the rate of tumour growth in tumours derived from Prlr-/-/C3(1)/SV40T epithelium compared to tumours from WT/C3(1)/SV40T epithelium (p=0.33). WT/C3(1)/SV40T transplants produced a total of 29 tumours while Prlr-/-/C3(1)/SV40T transplants produced 27, with mean tumours per transplant of 1.6 ± 0.2 and 1.2 ± 0.3 respectively, revealing no detectable significant (p=0.23) difference in tumour frequency between genotypes.
5.2.7. The mRNA and protein expression of SV40T was unaltered in mammary epithelium without the Prlr. We examined SV40T levels in 8 (56 days), 22 (154 days) and 32 week (224 days) old C3(1)/SV40T mammary glands formed by transplantation (Figure 5.10 A). No significant difference was observed at 8, 22 and 32 weeks post surgery between WT/C3(1)/SV40T (16 ± 0.8, 18 ± 1.0 and 19 ± 0.2 units) and Prlr-/-/C3(1)/SV40T epithelial transplants (17 ± 0.5, 18 ± 0.3 and 19 ± 0.3 units, p=0.10, p=0.64, p=0.78 respectively). Western blotting using an antibody against SV40T protein was used to determine the protein expression of SV40T in 8 week (56 day) old transplants (Figure 5.10 B). Detection of b-actin protein was used as a loading control. The average volume of SV40T protein in WT/C3(1)/SV40T transplants was 8.6 ± 0.7% which was not significantly different to Prlr-/- /C3(1)/SV40T transplants (10.6 ± 2.4%, p=0.45), indicating that like the mRNA expression of SV40T, the protein expression of the transgene is not altered by the presence of the Prlr (Figure 5.10 C). Univariate regression analysis demonstrated that SV40T mRNA expression was not a predictor for age of detection (p=0.52), tumour latency (p=0.95) and days with tumour (p=0.52).
135 Figure 5.8. Tumour latency is significantly increased in mammary epithelium without the Prlr. Tumor free survival curve. Time is represented in days after transplantation. Palpable tumors were detected in Prlr-/-/C3(1)/SV40T transplants (circles) significantly later compared to WT/C3(1)/SV40T transplants (p<0.001).
136 Figure 5.9. There was no significant difference in rate of growth of tumours that formed in Prlr-/-/C3(1)/SV40T transplants compared to those that formed in WT/C3(1)/SV40T transplants. Tumor volume trellis plot. The cube root of tumor volume (volume1/3) is plotted with respect to time (days). Time 0 is the day of initial detection. There was no significant difference in the rate of change in tumor volume from WT/C3(1)/SV40T transplants (left box) and Prlr-/-/C3(1)/SV40T transplants (right box) (p=0.33).
137 Figure 5.10. SV40T mRNA and protein expression is not altered in epithelium without the Prlr. Relative expression of SV40T mRNA in 55 day (8 week), 154 (22 week) and 224 (32 week) old transplants. There was no significant difference in the expression of SV40T at 8, 22 and 32 weeks between Prlr-/-/C3(1)/SV40T (red bars) and WT/C3(1)/SV40T (blue bars) transplants (p=0.07, p=0.63, p=0.78 respectively). (B) Western blot of SV40T protein expression in 55 day (8 week) old transplants. -Actin is shown as a loading control. 5 donor animals are indicated by the numbers above the blot. (C) Average volume of SV40T protein normalised to -Actin. There was no significant difference the expression of SV40T protein between WT/C3(1)/SV40T (blue bar) and Prlr-/-/C3(1)/SV40T (red bar) 8 week old transplants.
138 5.2.8. Loss of epithelial Prlr does not change the histology and morphology of SV40T induced tumours. We then investigated whether loss of Prlr in the epithelium changed the histological appearance of SV40T induced tumours. Tumor tissues were collected and HandE histology was undertaken in a similar manner to C3(1)/SV40T mice described above. There was similar diversity in the microscopic features of lesions observed in tumours taken from mammary gland transplants, the histopathology was comparable to that observed in C3(1)/SV40T mice described above. Areas of necrosis were found in 100% of tumours taken from mice that had reached 10% tumour burden. Only 16/25 and 2/14 palpable tumours collected for histological investigation reached the ethical end point from control C3(1)/SV40T epithelium and Prlr-/-/C3(1)/SV40T epithelium respectively. The large latencies observed in the formation of palpable tumours from these transplants resulted in the lengthening of the experiment beyond the normal healthy life span of a Rag1-/- immune- compromised host. Therefore, a large proportion of tumours were collected prior to reaching the pre-determined end point size. There was little variation in tumour architecture or cellular morphology between tumours derived from mammary glands made from Prlr-/- /C3(1)/SV40T epithelium and control C3(1)/SV40T epithelium (Table 5.2). There was no significant difference in percentage areas of papillary and solid or cellular structure. A small increase was detected in the percentage of tumours that displayed acinar/glandular characteristics in tumours from Prlr-/-/C3(1)/SV40T epithelium, but we detected no difference in tumour type (acinar/glandular, papillary and solid) as a function of high and low Prlr expression level in tumours from WT/C3(1)/SV40T epithelium (p=0.22, p=0.15 and p=0.54 respectively; data not shown). This suggests that the small increase in acinar/glandular tumours may simply reflect the longer latency in Prlr-/-/C3(1)/SV40T transplants rather than an effect of Prlr. Overall, Prlr null epithelium does not appear to change the mechanism determining the morphology of SV40T-induced tumours.
139 Tumor morphology Cellular Structure (%) (%)
Acinar/ Epithelium Genotype n Glandular Papillary Solid Type A Type B WT/C3(1)/SV40T 25 21±4 17±6 61±7 60 40 All tumors -/- collected Prlr /C3(1)/SV40T 14 35±6 20±6 45±7 57 43 p-value 0.042 0.76 0.10 Chi Square p=0.67 Table 5.2. Tumor morphology and cellular structure in tumors from WT/C3(1)/SV40T and Prlr-/-/C3(1)/SV40T transplants. n; number of tumors. The average percent of tumor area classified as acinar and/or glandular, papillary and solid is given ± standard error. The percentage of tumors classified as having type A or type B cellular structure is also shown.
140 5.2.9. SV40T induced neoplasia was delayed in Prlr-/- mammary epithelial cells. In order to determine whether the presence of Prlr in mammary epithelium can modulate the development of SV40T-induced neoplasia, we collected mammary glands made from WT/C3(1)/SV40T and Prlr-/-/C3(1)/SV40T epithelial transplants at 22 (154 days) (Figure 5.11) and 32 weeks (224 days) (Figure 5.12). The Rag1-/- C57BL/6J mouse strain used as our transplant host develops very few mammary ductal side branches, a feature of this mouse strain that is dependent on factors from the stroma and not the epithelial donor (Naylor et al., 2002). Thus donor tissue from a mixed FVB/N and 129Ola/Pas strain develops a mammary tree that shows a predominantly primary ductal branching pattern, formed by bifurcation during ductal elongation at pregnancy (Y-shaped junctions), with sparse side branches (T-shaped junctions), when transplanted into C57BL/6J Rag1-/- hosts. Abnormal development in WT/C3(1)/SV40T transplants first appears as an increased number of short side branches at abnormally close spacing (Figure 5.11 A), a feature that is not seen in control transplants without the C3(1)/SV40T construct. In contrast, Prlr-/-/C3(1)/SV40T epithelial transplants at the same age exhibit the same developmental abnormality, but at a greatly reduced frequency (Figure 5.11 B). To quantify neoplastic area, we assessed the area occupied by SV40T-induced lesions in carmine alum stained Prlr-/-/C3(1)/SV40T mammary whole mounts. C3(1)/SV40T mammary epithelium lacking Prlr had a significantly smaller area of total lesions at 22 (7.6 ± 1.7%; Figure 5.11 C) and 32 weeks (11.7 ± 2.4%; Figure 5.12 C) compared to control C3(1)/SV40T epithelium (18.4 ± 1.2 and 24.1 ± 2.8%, p<0.001 and p=0.005 respectively). We divided the lesions into neoplasia and tumour. Prlr-/- /C3(1)/SV40T transplants also had less neoplastic or tumour area at 22 weeks (7.5 ± 1.6 and 0.2 ± 0.1% respectively) and 32 weeks (7.3 ± 1.1 and 4.4 ± 1.6% respectively) than control C3(1)/SV40T transplants at 22 weeks (17.4 ± 1.2 and 1.0 ± 0.4%; p<0.001 and p=0.06 respectively) and 32 weeks (14.4 ± 2.0 and 9.8 ± 2.6%; p=0.009 and p=0.11 respectively). The ratio of neoplasia to tumour in Prlr-/-/C3(1)/SV40T transplants at 22 weeks was greater than in WT/C3(1)/SV40T transplants, however, this ratio equalled control levels by 32 weeks.
141 Figure 5.11. Mammary neoplasia was reduced in Prlr-/-/C3(1)/SV40T transplants at 22 weeks Mammary whole mounts of WT/C3(1)/SV40T transplants (A) and Prlr-/-/C3(1)/SV40T transplants (B ) at 22 weeks post transplantation. Scale bars represent 500μm. (C) Bar graph summary of mammary whole-mount analysis at 22 weeks. Prlr-/-/C3(1)/SV40T transplants (red bars) had less area of total lesions, neoplasia and tumor as a percentage of total mammary gland area compared to WT/C3(1)/SV40T transplants (blue bars) (p<0.001, p<0.001 and p=0.061 respectively).
142 Figure 5.12 Mammary neoplasia was reduced in Prlr-/-/C3(1)/SV40T transplants at 22 weeks Mammary whole mounts of WT/C3(1)/SV40T transplants (A) and Prlr-/-/C3(1)/SV40T transplants (B ) at 32 weeks post transplantation. Scale bars represent 500μm. (C) Bar graph summary of mammary whole-mount analysis at 32 weeks. Prlr-/-/C3(1)/SV40T transplants (red bars) had less area of total lesions, neoplasia and tumor as a percentage of total mammary gland area compared to WT/C3(1)/SV40T transplants (blue bars) (p=0.005, p=0.009 and p=0.11).
143 5.2.10. Cellular proliferation in SV40T-induced neoplasia was mediated by Prlr within mammary epithelium. H&E histology allows the division of neoplasia into low-grade mammary intraepithelial neoplasia (LGMIN) and high-grade mammary intraepithelial neoplasia (HGMIN). LGMIN displayed the presence of stratified atypical ductal epithelial cells with elongated, hyperchromatic and pleomorphic nuclei. HGMIN were present at multiple foci, and showed greater cellular crowding, more stratification, loss of polarity and increased pleomorphism and hyperchromatism. Often neoplastic cells completely filled the ductal lumen. Invasive lesions are distinguished from HGMIN by breaching the basement membrane and stromal invasion. We used BrdU immune-histochemistry to investigate the effect of Prlr on SV40T-induced cellular proliferation within these lesion types (Figure 5.13). Mice were injected with 5-Bromo-2’-deoxyuridine (BrdU, Sigma, Germany) dissolved in distilled H2O (100μg/g body weight), two hours prior to euthanasia by CO2 asphyxiation. A significant increase was detected in the proliferation rate of cells from WT/C3(1)/SV40T pre-invasive lesions (17.0 ± 1.2%) compared to ‘typical’ WT/C3(1)/SV40T ductal epithelium displaying a normal epithelial morphology (7.8 ± 1.8%, Figure 5.13 E; t-test p=0.019). We detected significantly less proliferation in low grade and high grade MIN lesions from Prlr-/-/C3(1)/SV40T transplants (11.3 ± 1.2 and 13.0 ± 1.4%) compared to control C3(1)/SV40T transplants (17.0 ± 1.2 and 17.5 ± 2.0% Figure 5.13 E; t-test p=0.003 and p=0.067), demonstrating that loss of Prlr results in reduced SV40T induced proliferation in early stage lesions. There was no significant difference in the number of proliferating cells in invasive lesions from Prlr-/-/C3(1)/SV40T transplants compared to control C3(1)/SV40T transplants.
5.2.11. Apoptosis via activation of Caspase 3 was unaltered by Prlr within mammary epithelium. We investigated the effect of a loss of Prlr signalling on apoptosis using an antibody raised against the cleaved and active form of Caspase 3, a marker of cellular apoptosis (Figure 5.14). We detected no cleaved Caspase 3 positive cells in typical epithelium from both Prlr-/-/C3(1)/SV40T and control C3(1)/SV40T transplants. A low level of apoptosis was detected in LGMIN, which increased in HGMIN lesions and was maintained at the
144 same level in invasive lesions. There was no significant difference in the rates of apoptosis in LGMIN, HGMIN and invasive lesions from Prlr-/-/C3(1)/SV40T (0.2 ± 0.1, 1.4 ± 0.3 and
145 Figure 5.13. Proliferation was reduced in low-grade neoplastic lesions formed in Prlr-/- /C3(1)/SV40T transplants. Immunohistochemistry using an antibody against 5-Bromo-2’-dexyuridine (BrdU) in low- grade (A and C) and high-grade lesions (B and D) from WT/C3(1)/SV40T transplants (A and B) and Prlr-/-/C3(1)/SV40T transplants (C and D). (E) The percentage of BrdU positive cells in areas of typical ductal epithelium, LGMIN, HGMIN and invasion was reduced in Prlr-/-/C3(1)/SV40T transplants (red bars) compared to WT/C3(1)/SV40T transplants (blue bars). Prlr-/-/C3(1)/SV40T transplants had significantly reduced proliferation via detection of BrdU staining in lesions classified as LGMIN and HGMIN (p=0.003 and p=0.067 respectively).
146 Figure 5.14. There was no difference in the rate of apoptosis in lesions from Prlr-/- /C3(1)/SV40T and WT/C3(1)/SV40T transplants. Immunohistochemistry using an antibody against cleaved Caspase-3 in WT/C3(1)/SV40T transplants (A) and Prlr-/-/C3(1)/SV40T transplants (B). There was an undetectable level (ND) of cleaved Caspase 3 staining in typical ductal epithelium from WT/C3(1)/SV40T and Prlr-/-/C3(1)/SV40T transplants. There was no significant difference in the rate of apoptosis via measurement of cleaved Caspase 3 in LGMIN, HGMIN and invasive lesions from Prlr-/-/C3(1)/SV40T transplants (red bars) compared to WT/C3(1)/SV40T transplants (blue bars; p=0.42, p=0.59 and p=0.92, respectively).
147 1.2 ± 0.5%) and control C3(1)/SV40T transplants (0.3 ± 0.1, 1.2 ± 0.2 and 1.3 ± 0.2%, Figure 5.14 C; t-tests p=0.42, p=0.59 and p=0.92 respectively). These results indicate that Prlr signalling has no effect on cellular survival via apoptotic mechanisms involving cleavage of Caspase 3 in SV40T induced lesions.
5.2.12. Transcript profiling of 8 week Prlr-/-/C3(1)/SV40T transplants to identify genes that may be important for Prl modulated carcinogenesis. To identify the genes regulated by Prl that may be important in modulating proliferation in this model, we performed transcript profiling of Prlr-/-/C3(1)/SV40T and WT/C3(1)/SV40T mammary glands at 8 weeks post-transplantation. We chose 8 weeks, as this time point was prior to the development of neoplasia and invasive tumor. This would overcome the problems foreseen by profiling whole mammary glands at later time points containing tissue from all stages of carcinogenesis. Thus, we would be able to investigate the genes regulated by Prl, that predispose the epithelium to carcinogenesis under the influence of SV40T. We collected 4 Prlr-/-/C3(1)/SV40T, 4 WT/C3(1)/SV40T transplants and 2 cleared fat-pads at 8 weeks post-transplantation. We collected the cleared fat-pads in order to identify an epithelial enriched population of genes and we hybridised RNA to Mouse Genome 430 2.0 GeneChip arrays. Quality control analysis revealed that one of the Prlr-/-/C3(1)/SV40T arrays was an outlier (Figure 5.15). Therefore we excluded this array and its matched control from further analysis, and analysed the remaining chips. An intensity filter was applied on each group such that the average intensity of either the WT/C3(1)/SV40T or Prlr-/-/C3(1)/SV40T samples was greater than the 20th percentile of overall intensity on the GeneChip. Genes that differed significantly between either the WT/C3(1)/SV40T or Prlr-/-/C3(1)/SV40T and the cleared fat pads were denoted the epithelial enriched population. This set comprised 2508 epithelial enriched genes and the top 425 genes (p<0.05) that were differentially regulated between WT/C3(1)/SV40T or Prlr-/-/C3(1)/SV40T are listed in Appendix XII. Principal components analysis was then used to visualise the successful separation of each experimental group (Figure 5.16). After multiple testing, 51 epithelial enriched genes (p<0.005) were identified as significantly changing between 8 week old WT/C3(1)/SV40T and Prlr-/-/C3(1)/SV40T samples (Table 5.3). The false discovery rate
148 Figure 5.15. Quality control plot of 8 week C3(1)/SV40T arrays. Array data was quantile normalised using RMA function in R. The logged intensities of each array were then plotted in a histogram. All arrays passed the quality control analysis except for the 4th Prlr-/-/C3(1)/SV40T array (yellow dashed). This array and its matched contralateral control gland were therefore excluded from further analysis.
149 Figure 5.16 Principal components analysis of 8 week C3(1)/SV40T arrays. An analysis of variance was used to identify genes that were significantly different between cleared fat pad and either WT/C3(1)/SV40T and Prlr-/-/C3(1)/SV40T. 2508 genes differed significantly between Prlr-/-/C3(1)/SV40T and WT/C3(1)/SV40T 8 week old mammary transplants. Principal components analysis demonstrated the set of 2508 epithelial enriched genes successfully clustered the Prlr-/-/C3(1)/SV40T and WT/C3(1)/SV40T experimental groups
150 Fold Q-PCR Probe Set ID Gene Title Gene Symbol P-Value Change FC 1448556_at Prolactin receptor Prlr 0.00001 -4.34405 1437397_at Prolactin receptor Prlr 0.00002 -3.85888 Potassium voltage-gated channel, shaker-related subfamily, member 1437230_at 1 (Kcna1), mRNA Kcna1 0.00009 2.20202 DNA segment, Chr 17, human 1417822_at D6S56E 5 D17H6S56E-5 0.00020 -2.46929 1458347_s_at Transmembrane protease, serine 2 Tmprss2 0.00023 -2.17936 1417156_at Keratin complex 1, acidic, gene 19 Krt1-19 0.00027 -1.99045 1425853_s_at Prolactin receptor Prlr 0.00029 -9.43127 Potassium voltage-gated channel, shaker-related subfamily, member 1417416_at 1 Kcna1 0.00043 2.41945 Solute carrier family 7 (cationic amino acid transporter, y+ 1436555_at system), member 2 Slc7a2 0.00044 -1.86263 1419154_at Transmembrane protease, serine 2 Tmprss2 0.00052 -1.75441 -2.714 1441102_at Prolactin receptor Prlr 0.00053 -2.97489 Mitogen activated protein kinase 1421340_at kinase kinase 5 Map3k5 0.00063 1.77449 -1.003 1418496_at Forkhead box A1 Foxa1 0.00076 -1.75833 -2.232 1437502_x_at CD24a antigen Cd24a 0.00093 -1.96892 -1.001 1437019_at RIKEN cDNA 2200001I15 gene 2200001I15Rik 0.00120 -1.78120 1450206_at Deleted in liver cancer 1 Dlc1 0.00125 1.63074 -1.073 1448182_a_at CD24a antigen Cd24a 0.00131 -1.70128 -1.001 Ral guanine nucleotide dissociation stimulator,-like 1, mRNA (cDNA clone MGC:18430 1440662_at IMAGE:4241244) Rgl1 0.00143 1.58337 1448169_at Keratin complex 1, acidic, gene 18 Krt1-18 0.00162 -1.57223 Transmembrane 4 superfamily 1445008_at member 1 Tm4sf1 0.00168 1.60314 1436910_at RAS protein activator like 2 Rasal2 0.00177 1.56831 0 day neonate eyeball cDNA, RIKEN full-length enriched library, clone:E130307M10 product:unclassifiable, full insert 1440305_at sequence --- 0.00190 1.57799 Myxovirus (influenza virus) 1451905_a_at resistance 1 Mx1 0.00195 1.66491 1460569_x_at Claudin 3 Cldn3 0.00213 -1.55385 DNA segment, Chr 17, human 1417821_at D6S56E 5 D17H6S56E-5 0.00215 -3.44588 Replication factor C (activator 1) 1432538_a_at 3 Rfc3 0.00215 -1.51772 1416034_at CD24a antigen Cd24a 0.00220 -1.93706 -1.001 1452534_a_at High mobility group box 2 Hmgb2 0.00221 -1.52648 151 1434553_at Transmembrane protein 56 Tmem56 0.00238 -1.69910 -2.330 1457349_at Gene model 69, (NCBI) Gm69 0.00258 -1.74478 1424351_at WAP four-disulfide core domain 2 Wfdc2 0.00265 -1.79208 Spinocerebellar ataxia 2 homolog 1443516_at (human) (Sca2), mRNA Sca2 0.00274 1.90710 1429906_at RIKEN cDNA A930035E12 gene A930035E12Rik 0.00279 -1.48464 1435742_at RIKEN cDNA 1110034C04 gene 1110034C04Rik 0.00294 1.48527 Autophagy-related 7 (yeast) 1447147_at (Atg7), mRNA Apg7l 0.00296 1.57912 Transmembrane 4 superfamily 1439925_at member 1 Tm4sf1 0.00304 1.50040 1451701_x_at Claudin 3 Cldn3 0.00305 -1.64741 1419700_a_at Prominin 1 Prom1 0.00317 -1.84661 1427095_at CUB domain containing protein 1 Cdcp1 0.00345 -1.50406 RIKEN cDNA 5730403B10 gene, mRNA (cDNA clone MGC:8188 1447951_at IMAGE:3590511) 5730403B10Rik 0.00352 1.57338 1417373_a_at Tubulin, alpha 4 Tuba4 0.00359 -1.73973 1439079_a_at Erbb2 interacting protein Erbb2ip 0.00395 1.56864 1425469_a_at RIKEN cDNA 9030208C03 gene 9030208C03Ri 0.00429 -2.08281 1458601_at RIKEN cDNA 8030447M02 gene 8030447M02Rik 0.00431 1.57284 PolyA specific ribonuclease 1437883_s_at subunit homolog (S. cerevisiae) Pan3 0.00440 1.44913 1417896_at Tight junction protein 3 Tjp3 0.00443 -1.47392 ATPase, Na+/K+ transporting, alpha 2 polypeptide, mRNA (cDNA clone MGC:36347 1434893_at IMAGE:4955003) Atp1a2 0.00464 1.69004 1449369_at Transmembrane protease, serine 2 Tmprss2 0.00466 -2.60076 -2.714 Glycine amidinotransferase (L- arginine:glycine 1423569_at amidinotransferase) Gatm 0.00470 1.78041
1453070_at Protocadherin 17 Pcdh17 0.00472 1.46844 1460711_at RIKEN cDNA 4930461P20 gene 4930461P20Rik 0.00479 1.54287 Table 5.3: Top 51 epithelial enriched genes differentially regulated between Prlr-/- /C3(1)/SV40T and WT/C3(1)/SV40T transplants at 8 weeks. The top 51 epithelial enriched genes differentially regulated between Prlr-/-/C3(1)/SV40T and WT/C3(1)/SV40T transplants with an unadjusted p<0.005. The Affymetrix ID, gene title, symbol, unadjusted p value, Affymetrix fold change and quantitative PCR fold changes are tabulated. Increasing or decreasing genes are indicated by positive and negative values respectively.
152 was determined by permutation analysis and was as 7/51 (13.7%). Using quantitative PCR we confirmed the differntial expression of the Tmprss2, Tmem56 and Foxa1 (Table 5.3). 30 of the top 51 (59%) epithelial enriched genes were down-regulated in response to loss of epithelial Prlr in C3(1)/SV40T transplants, suggesting that the majority of Prlr signalling functions were stimulatory. We have observed this previously, in transcript profiling experiments of Prlr-/- mammary glands at parturition reported in Chapter 3 of this thesis. Several of these genes observed in the set of 51 differentially regulated genes, were genes that we had previously observed in transcript profiling experiments of Prlr-/- mammary epithelium at early time points of pregnancy (Ormandy et al., 2003). These genes include Keratin 18, Keratin 19 and Claudin-3. We also identified the prolactin receptor, which was down regulated in 4 out of the 51 probe sets, consistent with the reverse transcriptional direction of the NEO cassette employed as a targeting strategy in this knockout model. Therefore we could be confident that our statistical approach was a reliable method of detecting Prl regulated genes. Interestingly several of the differentially regulated genes, including Cldn3, CD24a, Tm4sf1, Kcna1 and Tmprss2 were represented by multiple probes sets within the top 51, suggesting that they are not in this list simply by chance.
153 5.3 Discussion Large-scale prospective studies of breast cancer have demonstrated that prolactin (Hankinson et al., 1999; Tworoger et al., 2004), estrogen (Missmer et al., 2004), and the androgenic precursors of estrogen (Hankinson et al., 1998) are hormones that increase the risk of breast cancer in women who experience serum levels within the top quartile of the population range. In this chapter, we have examined how prolactin acts to modulate carcinogenesis, using a model in which mammary cancer is initiated by the SV40T oncogene in the absence or presence of an intact prolactin-signalling pathway. Using whole animals or transplanted glands we have been able to contrast the direct action of prolactin on the mammary epithelial cell with its indirect actions. By a combination of longitudinal survival analysis and cross sectional histological studies we have defined the influence of prolactin over latency, numbers and types of tumours produced by SV40T, and we have identified the stage in this carcinogenic process where prolactin acts. Tumor latency increased as a consequence of the loss of Prlr by 26 days in whole animals (12.0%) and by 53 days (22.5%) in transplants. This comparison shows that the direct effect of prolactin via the mammary epithelial prolactin receptor is the predominant mechanism by which prolactin modulates mammary carcinogenesis. This is also the first demonstration of a mammary cell autonomous effect of prolactin outside of pregnancy. The indirect effects of prolactin are complex. Loss of the prolactin receptor caused reduced estrogen and progesterone levels but increased the level of parathyroid hormone (Clement- Lacroix et al., 1999). Prlr-/- mice also had reduced insulin levels and sensitivity (Freemark et al., 2001) and decreased body weight. Given this degree of endocrine disruption other endocrine defects probably remain to be discovered. Despite these changes in the endocrine environment, comparison of the difference in relative tumour latency between tumour genotypes in the whole animal (26 days) and transplant experiments (53 days) indicates that the combined effects of these indirect actions on carcinogenesis are negligible or operate to reduce the effect of prolactin. The stage of the carcinogenic process that is influenced by prolactin has not previously been defined. Transplants lacking Prlr showed greatly reduced areas of neoplasia and longer latency to the first palpable tumour. An analysis of cell proliferation showed a reduction in cell proliferation in the neoplasias as a result of a loss of the Prlr. 154 Apoptosis was unaffected by genotype at every stage examined. Thus prolactin acts at the very earliest stages of carcinogenesis to increase cell proliferation in neoplastic lesions, resulting in a greater area of neoplasia and a more rapid emergence of invasive tumours. We also found that loss of Prlr did not influence the growth of invasive lesions in this model of carcinogenesis. Overall there was no difference in the proliferation rate or the growth rate of invasive lesions between prolactin genotypes in either the whole animals or the transplants. An analysis of WT tumour transplants showed no relationship between Prlr expression level and tumour growth rate despite the detection of the expected difference between genotypes of latency to palpable tumour. Close inspection of Figure 5.9 shows a dichotomy of tumour growth rates in Prlr-/- transplants compared to WT. Two distinct types of growth rate are seen in Prlr-/- tumours, a majority with slow growth and 3 tumours that showed initial slow growth followed by a dramatic increase in growth rate. This dichotomy is reflected in the size of the BrdU error bar for Prlr-/- tumours (Figure 5.13 E). WT tumours show a broad spectrum of growth rates. Although, it is tempting to speculate that this difference in growth pattern may reflect a fundamental difference in tumour biology between genotypes, this effect was not seen in whole animals or in relation to Prlr expression level. A few advanced tumours may remain sensitive to prolactin, but in our experiments their frequency was not sufficient to influence the analysis. Loss of sensitivity to Prl should be seen in the wider context of a loss of hormone sensitivity in general by these tumours; this model becomes estrogen insensitive early in progression (Green et al., 2004). Mouse models do not currently offer a way to produce hormone sensitive tumours, with the only possible exception being over expression of prolactin, which results in about 20-30% ER positivity with less sensitivity to ER ablation (Rose-Hellekant et al., 2003). This limitation prevents us from drawing conclusions regarding the ability of prolactin to drive the proliferation of hormone sensitive tumours, but demonstrates prolactin action during the earliest stages of carcinogenesis. Thus prolactin may facilitate tumour formation in two ways: by increasing the number of neoplastic cells prolactin increases the chance of a tumourigenic event; and by driving the proliferation of neoplastic cells prolactin forces cell divisions that may replicate tumourigenic genetic or epigenetic events. Once these events occur the resulting invasive lesions generally become independent of prolactin. This observation offers a explanation, in
155 addition to autocrine Prl action, of why bromocriptine treatment of patients with advanced breast cancer was not successful (Bonneterre et al., 1988; Manni et al., 1989; Peyrat et al., 1984), and why prolactin treatment of breast cancer cell lines does not have a generalised and potent effect on proliferation. Our finding that prolactin acts on very early neoplastic lesions challenges the prevailing assumption that prolactin acts primarily during late-stage disease to drive invasive tumour growth. In an attempt to understand the underlying mechanisms of Prl mediated proliferation in pre-invasive lesions, we performed transcript profiling of 8 week old C3(1)/SV40T transplants prior to the development of neoplasia. This enabled us to investigate gene expression changes as a result of loss of Prlr, under the influence of SV40T, prior to the development of invasive tumours. We identified 51 genes, enriched in the mammary epithelium that showed significant change in expression between WT/C3(1)/SV40T and Prlr-/-/C3(1)/SV40T 8 week old mammary transplants. Several proof of principal genes were discovered in the set of 51 differentially regulated genes. The first represented by four probes in the top 51, was the prolactin receptor, which indicated that our approach was reliable. The two genes Keratin 18 and 19 were significantly down- regulated in C3(1)/SV40T transplants that lacked Prlr. We have previously identified these genes in transcript profiling experiments of Prlr-/- mammary epithelium at early time points of pregnancy (Ormandy et al., 2003). Keratin 18 is directly regulated by the Ets transcription factor Elf5 (Yaniw et al., 2005). In addition, we have shown in Chapter 4, that Elf5 is a key mediator of Prlr signalling during mammary gland development, thus the presence of Keratin-18 in the current experiment, is further confirmation that we have detected Prl regulated genes. Our transcript profiling of Prlr-/- mammary epithelium during early pregnancy also identified the integral membrane protein Claudin-3 (Ormandy et al., 2003). Here again, we detected 2 probes representing Claudin-3 as down-regulated in C3(1)/SV40T mammary epithelium lacking Prlr. Claudins are major structural components of tight junctions, and are important in maintaining cellular polarity (Morita et al., 1999). However, their down-regulation in Prlr-/-/C3(1)/SV40T epithelium, may suggest a role in modulating SV40T induced carcinogenesis. These observations add validity to our transcript profiling technique, but more intriguingly, suggest that the mechanisms driving
156 proliferation in the mammary gland during pregnancy may also be important for driving proliferation during carcinogenesis. The replicating factor C 3 (Rfc 3) was down regulated in C3(1)/SV40T transplants which lacked Prlr. Replicating factor 3 is a gene involved in the DNA replication machinery (Johnson et al., 2005). Interestingly, we also detected down regulation of the closely related family members Rfc 1 and Rfc 4 in Prlr+/- mice not capable of lactation in chapter 3. Thus discovery of these related probe sets in two separate transcript profiling studies, is suggestive that Prl may have a biological role in regulating proliferation via the DNA replication machinery, both in the normal mammary gland and in cancer cells. Until now, these are the first reports implicating Prl in the control of the DNA machinery, and clearly more work is required to fully understand this mechanism. Another important observation is the repetition of some of the probe sets in the top 51, such as Cldn3, Kcna1 and Tmprss2, suggesting that they are not in this list simply by chance. Kcna1 is the potassium voltage-gated ion channel, shaker-related subfamily 1. Cellular proliferation and apoptosis are regulated by the exchange of ion solutes across the membrane mediated by the voltage-gated channels (Lang et al., 2005). There is also increasing evidence that ion channels have a role in the development of cancer (Kunzelmann, 2005). However their diversity and involvement in either apoptosis or proliferation may depend on the cell of origin and regulation by various signalling pathways (Lang et al., 2005). C3(1)/SV40T transplants which lack Prlr resulted in increased expression of Kcna1, and this may suggest that it is playing a regulatory role in the proliferation in SV40T induced epithelial cells. The transmembrane protease, serine 2 is androgen regulated and can form gene fusions with ets transcription factors, which appear to play a role in the development of prostate cancer (Rubin et al., 2006). Tmprss2 can form 14 different gene fusions with the ets transcription factor Erg, which may be dependent on the location from where a cancer arises in the prostate (Clark et al., 2006). Also, in all cases except one, the pointed domain was retained suggesting that this domain was important for these gene fusions. The pointed domain is hypothesised to be involved in homo-oligomerisation, heterodiamerisation and transcriptional repression (Sharrocks, 2001). In chapter 3 we demonstrated that Elf5 is downstream of the Prlr. Co-incidentally we find that Tmprss2 is down-regulated in Prlr-/-
157 /C3(1)/SV40T mammary epithelium. There is at this stage, no evidence that Tmprss2 has a role in breast cancer progression. Our results here suggest however, that Prl may regulate Tmprss2 in the breast potentially by binding to an Ets factor like Elf5, and hypothesise that this may be important in modulation of SV40T induced carcinogenesis. Another interesting gene to come out of these arrays is the Forkhead box A1 gene, Foxa1 (HNF-3a). Foxa1 was down regulated in C3(1)/SV40T epithelium lacking Prlr. Recently, it has been shown that chromatin binding of ER requires the presence of Foxa1 resulting in the transcription of estrogen target genes in human breast cancer cells (Carroll et al., 2005). Prlr and ER are co-expressed in human breast cancer cells, and Prlr is up regulated by long-term exposure to estrogen, suggesting that cross-regulation of these receptors occurs in breast cancer (Ormandy et al., 1997c). Chromatin immune-precipitation demonstrated that the transcription factor Gata3 binds to the promoter of Foxa1 (Kouros- Mehr et al., 2006a) and thus might be regulated by Gata3. Interestingly work in our laboratory discovered that Gata3 was decreased in Prlr-/- transplants at days 2, 4 and 6 of pregnancy (Ormandy et al., 2003). Thus it appears that Gata3 and Foxa1 are regulated at least in part by the Prlr. These findings together suggest a point of cross talk between Prlr regulation of ER activities, although much more research is needed to confirm these hypotheses. We hope to continue to investigate the function of these genes, in an attempt to define the mechanisms underlying Prl modulated proliferation of SV40T induced pre- invasive lesions. In summary, we demonstrated that Prl acts directly via the epithelial cell, to modulate the proliferation of pre-invasive lesions in the SV40T model of mouse mammary carcinogenesis. Prl does not affect tumor growth rate, morphology, or the proliferation of invasive lesions. Thus Prl acts at the earliest stages of carcinogenesis, to modulate the proliferation of neoplasia, resulting in faster progression to tumor formation. These findings have important implications for the treatment of human disease. Agents antagonising prolactin action, such as S179D prolactin and G129R prolactin (Goffin et al., 2005) may prove to be useful in preventing the progression of hyperplastic and neoplastic lesions to invasive cancer. Improvements in imaging and diagnostic techniques are currently under development to allow the identification of these early lesions. Prolactin receptor antagonists
158 should be considered as agents for their treatment, both as an adjuvant to surgery and hormonal therapy, or as a component of preventative therapy.
159 Chapter 6 General discussion 6.1 Overview The primary function of the mammary gland is to provide nutrition to the infant. This function is made possible by extensive tissue remodelling during pregnancy, collectively termed alveolar morphogenesis (Neville et al., 2002), that results in the formation of the lobuloalveolar units.. Together Prl and Pg activate various signalling pathways crucial for the regulation of alveolar morphogenesis and subsequently secretory initiation and activation. A number of downstream effectors important for proliferation and differentiation of the mammary gland have been identified, however their arrangement in a complete transcriptional program of alveolar morphogenesis is not yet defined. The aims of this thesis were increase our understanding of the mechanisms by which prolactin exerts its profound influence over the mammary gland. In this thesis, we have expanded our knowledge of the regulatory networks that control alveolar morphogenesis and secretory activation, and gained insight into how prolactin exerts its proliferative influence during carcinogenesis. A schematic diagram of the regulatory networks underlying these processes is illustrated in Figure 6.1.
6.2 The role of prolactin during the proliferative phase of alveolar morphogenesis The first phase of alveolar morphogenesis involves massive luminal cell proliferation to increase the surface area of the alveolar epithelium, and is initiated by Prl in the mouse. Although ER is required for pubertal ductal elongation and bifurcation, it is dispensable for alveolar morphogenesis (Bocchinfuso et al., 1997). Coitus in the mouse triggers rapid increases in the circulating levels of Prl, which acts to maintain and expand the ovarian corpus luteum, resulting in an increase in circulating Pg. Pg binding to the Pgr is essential for mammary cell proliferation, as mammary epithelial transplants, which lack Pgr result in failed alveolar morphogenesis (Brisken et al., 1998). The longer isoform of Pgr (Pgrb) is required for this process, through induction of its paracrine mediator RankL (Mulac-Jericevic et al., 2003). Pgrb induces expression of RankL, which in turn mediates proliferation of neighbouring cells through transcription of Ccnd1. Ccnd1 is required in the
160 epithelium for the proliferative phase of alveolar morphogenesis (Fantl et al., 1999). RankL is activated by its ligand Rank, which triggers a cascade of events leading to activation of Ikk and the transcription factor NF-B, resulting in Ccnd1 transcription and cellular proliferation (Cao et al., 2001). Null mutation of RankL and its receptor resulted in failed lobuloalveolar development (Fata et al., 2000). Pg also co-localises with Wnt4, which is important for mediating Pg proliferative signals within the mammary epithelium via paracrine mechanisms (Brisken et al., 2000). Thus Pg interaction with its Pgr in a subset of mammary epithelial cells is required for the proliferative phase of alveolar morphogenesis. Heterogeneous steroid receptor patterning is required for alveolar morphogenesis (Grimm et al., 2002; Seagroves et al., 1998; Seagroves et al., 2000), and steroid receptor positive cells segregate with proliferating cells. The transcription factor Cepb is an essential cell fate determinant in the mammary gland, which regulates epithelial Pgr as well as Prlr patterning, as Pgr and Prlr expression was increased and became more uniform in Cepb null animals (Grimm et al., 2002). These studies also demonstrate the close relationship in the regulation of Prl and Pg, and therefore suggest similar mechanisms of action. These observations are further supported by the co-expression and cross-regulation of Prlr and Pgr in human breast cancer cells (Ormandy et al., 1992; Ormandy et al., 1997c), suggesting that they act synergistically during carcinogenesis. Pituitary Prl stimulation of Prlrs in the ovaries assists in maintaining the required levels of Pg during early pregnancy (Binart et al., 2000). Thus Prl regulation of Pgr can account for some of the proliferative actions of the Prlr during early pregnancy. The neuropeptide Gal is important for maintaining the levels of Prl during pregnancy. Gal is an important mediator for estrogen stimulated pituitary lactotroph proliferation (Wynick et al., 1993), and regulates the secretion of Prl (Wynick et al., 1998). Gal is also a local growth factor for the mammary epithelium during pregnancy and augments lobuloalveolar development in co-operation with Prl (Naylor et al., 2003a). In addition Gal modulates the ratio of unmodified Prl and its inhibitory form phosphorylated Prl (Naylor et al., 2005b). Prl is essential in the mammary epithelium during the proliferative phase of alveolar morphogenesis. Prlr+/- mice fail to lactate on their first pregnancy and have lobuloalveolar defects (Ormandy et al., 1997b). Prlr-/- mice are infertile due to several reproductive defects, therefore mammary transplantation was used to
161 investigate loss of Prlr within mammary epithelium, Mammary transplants of Prlr-/- epithelium result in failed alveolar proliferation and differentiation (Brisken et al., 1999). The best-characterised signalling pathway activated by Prl is the Jak2/Stat5 pathway (Bole- Feysot et al., 1998). Prlr dimerization occurs after Prl binding and leads to the phosphorylation of Jak2 (Han et al., 1997; Pezet et al., 1997), which in turn phosphorylates specific residues on the Prlr (Lebrun et al., 1995a). Stat5 is then recruited to the receptor and is phosphorylated by Jak2 (DaSilva et al., 1996). Phosphorylated Stat5 is then translocated to the nucleus where it can activate transcription of multiple genes (Wakao et al., 1995). Inactivation of Stat5 and Jak2 in mice result in severe lobuloalveolar defects (Liu et al., 1997; Wagner et al., 2004). Control of this signalling pathway occurs through the Socs1/2 (Harris et al., 2006; Lindeman et al., 2001) and Caveolin 1 (Park et al., 2002) proteins via interaction with Prlr/Jak2/Stat5 (Kubo et al., 2003). Interestingly, loss of Caveolin 1 resulted in the up regulation the Er expression, suggesting that Prlr and Er are similarly regulated in mammary cells (Sotgia et al., 2006). Prlr activation also leads to the activation of the Ras/Mek/Mapk pathway (Das et al., 1996a), which involves the kinases Grb2, Shc, Sos, Ras and Raf (Das et al., 1996b), and results in the proliferation of NOG-8 normal mammary and T47D breast cancer cells. Activation of the Mapk pathway is important for Prl mediated proliferation of epithelial cells reviewed in (Bole-Feysot et al., 1998). However there is evidence to suggest that the Mapk and the Stat pathways may converge at the level of Mapk phosphorylation (Ihle, 1996). Also, Prl activation of the Prlr can result in the association of the Src family of kinases, resulting in the activation of PI3K and in turn AKT/PKB, which is important in the proliferative and anti-apoptotic effects of Prl signalling during alveolar morphogenesis (Berlanga et al., 1997; Fresno Vara et al., 2001; Tessier et al., 2001). This was demonstrated in HC11 cells where Prl prevented Tgf induced inactivation of Akt via phosphorylation of PI3K, and reduced Tgf induced apoptosis (Bailey et al., 2004). Crossing the Prl-/- with the dominant negative mutant of Tgf prevented Tgf induced hyperplastic alveolar development, associated with phosphorylation of Akt. Tgf binding to its receptor in mammary epithelium results in phosphorylation of Smad, which in turn act acts to restrict cell proliferation during alveolar morphogenesis (Reviewed in (Barcellos-
162 Hoff et al., 2000)). Thus, Prl activation of the PI3K pathway is important for the survival of mammary cells, and a point in convergence with Tgf signalling. Our transcript profiling experiments have identified several other candidates that may be important for Prl mediated proliferation of the alveolar epithelium (Ormandy et al., 2003). NF-B and RankL were down regulated in response to loss of mammary epithelial Prlr, thus this may be another point of convergence with Pg receptor signalling resulting in the proliferation of Pg negative cells, possibly via Ccnd1. In addition, Prl induces the transcription of Igf2, which stimulates mammary epithelial cell proliferation via increasing the expression of Ccnd1 (Brisken et al., 2002). The Erbb ligand, amphiregulin was also down regulated in Prlr-/- epithelium, suggesting co-operation between Prlr and the Erbb receptors. Amphiregulin restores proliferation in ovariectomised animals suggesting a key role for this pathway in epithelial cell proliferation (Kenney et al., 1996). Our transcript profiling also identified the transcription factor Gata3, which has been recently identified as essential for maintaining differentiated luminal cell epithelium during pregnancy (Kouros- Mehr et al., 2006a). In Chapter 3 we have identified several other potential targets of Prl mediated proliferative responses. Prlr+/- mammary mice that failed to lactate and had reduced lobuloalveolar development had reduced expression of the proliferation coupled proteins Pcna and Ki67 (Landberg et al., 1997). In addition we identified several other key components of the DNA replication machinery, the Rfcs 1 and 4 and the Mcmds. These proteins are important components for the ordered assembly of the origin of replication, which occurs during cell division (Bell et al., 2002). In addition to these genes, we also identified several other mediators of the cell cycle (Table 3.2). We once again identified Ccnd1 in this set, further confirmation that Prlr is important for its regulation. Prl modulation of the cell cycle in mammary cells has been demonstrated in several studies (Kaneko et al., 1998; Schroeder et al., 2003; Wu et al., 2006), and Stat5, after activation by Prl can bind and activate the promoter of Ccnd1 (Brockman et al., 2002). However, the findings in this thesis are the first reports of Prl mediated regulation of the DNA replication machinery, and therefore needs to be further examined. The transcription factor Elf5, is a target of Prlr signalling, and is required for Prl mediated mammary epithelial cell proliferation, as Elf5-/- mammary epithelial transplants
163 demonstrated similar defects in mammary cell proliferation to Prlr-/- mammary epithelial transplants during early pregnancy. Therefore Elf5 must play an important role in the regulation of Prl mediated proliferative responses in the mammary epithelium, presumably by regulating the transcription of many of genes mentioned above. Thus the control of the proliferative phase of alveolar morphogenesis requires both Prl and Pg signalling, resulting the regulation of a vast array of genes, involved in mammary cell proliferation (Figure 6.1 blue symbols).
6.3 The role of prolactin during lactogenesis From midpregnancy, differentiation takes over as the dominant process occurring in the mammary epithelium. This phase is characterised by the presence of cytoplasmic lipid droplets within the alveolar epithelium (Neville et al., 2002). This process involves further expansion of the alveolar epithelium, polarisation, alveolar differentiation and the synthesis of milk and lipid proteins in a temporal order (Rudolph et al., 2003). Luminal cell contact with the extracellular matrix is also essential for this process (Fata et al., 2004). Prlr signalling through Jak2/Stat5 is indispensable for alveolar differentiation, and is demonstrated by the failure of lobuloalveolar development and milk production in mouse models with null mutations of these genes (Brisken et al., 1999; Cui et al., 2004; Wagner et al., 2004). Stat5 has been shown to directly promote the transcription of -Casein (Schmitt- Ney et al., 1992) and Wap (Pittius et al., 1988). Stat5 is also activated by members of the EGF, growth hormone, interleukins, colony stimulating factor GM-CSF, thrombopoietin (reviewed in (Groner et al., 2000)) and recently identified integrin signalling (Naylor et al., 2005a). Contact with the extracellular matrix is required for lobuloalveolar development (Fata et al., 2004), and ablation of 1-intergrin results in lobuloalveolar defects and reduced pup weight (Naylor et al., 2005a). 1-intergrin provides an important link between the extracellular matrix and the alveolar epithelium via the extracellular matrix protein laminin (Klinowska et al., 1999). The epidermal growth factor receptor (Erbb4) is an essential mediator of Stat5 signalling during alveolar morphogenesis, as only partial lobuloalveolar
164 Figure 6.1. The regulatory networks involved in Prlr mediated alveolar morphogenesis and carcinogenesis. Prl binding to its receptor results in the activation of various signalling pathways involved in proliferation (blue) and differentiation of the mammary gland (pink). The final stage of alveolar morphogenesis or secretory activation (purple) results in the secretion of the various components of milk into the alveolar lumens. These processes are tightly regulated by a number of regulatory molecules involved in regulation of proliferation (green) and cell signalling (dark blue). Many of these processes are dysregulated during breast cancer, and the molecules that have been shown to have a role in the aetiology of breast cancer are highlighted in red.
165 development is observed in Erbb4 knockout mice associated with reduced Stat5 phosphorylation (Long et al., 2003). Activation of Stat5 by Erbb4 occurs via unique serine phosphorylation sites (Clark et al., 2005), different to that mediated by Jak2 on tyrosine 694 (Gouilleux et al., 1994). The Erbb4 ligand heregulin contributes to mammary cell proliferation and differentiation in the presence of E and Pg, although this pathway is not sufficient for full lobuloalveolar development (Jones et al., 1996). Also, although Erbb4 knockout mice have lobuloalveolar defects, they do not resemble the complete failure of lobuloalveolar development observed in Prlr-/- mammary epithelial transplants (Brisken et al., 1999). The same is true for other mouse knockout models (Luetteke et al., 1999; Naylor et al., 2005a). Thus, Prl is the major mediator of Stat5 phosphorylation contributing to the development of lobuloalveoli during alveolar morphogenesis, however it is clear that there are many signal transduction pathways, which contribute to Stat5 phosphorylation and in turn lobuloalveolar development. Cross talk between the cytokine receptors, growth factor receptors and receptor tyrosine kinases is at this stage not well defined and needs further investigation. In Figure 6.1 we illustrate the various signalling molecules that contribute to the differentiation of the mammary epithelium during pregnancy. In chapter 4, we demonstrated that Elf5 is a master regulator of Prl mediated alveolar morphogenesis; a study, which unequivocally confirms and expands on previous studies demonstrating an essential and cell autonomous role for Elf5 during lobuloalveolar development (Zhou et al., 2005). Elf5 also appears to be the “alveolar switch” driving differentiation of the lobular progenitor into the differentiated alveolar cell. Elf5 controls the expression of a wide variety of genes involved in milk and lipid synthesis during the differentiation of the mammary gland during pregnancy. In addition, our studies place Elf5 at the origin of the transcriptional hierarchy responsible for the formation of a differentiated mammary gland. However the design of this transcriptional hierarchy is still unclear (See Figure 4.20), and more research is needed to fully elucidate these mechanisms. It is likely that the transcription of genes involved in lobuloalveolar differentiation is controlled by a cascade of transcription factors (Visvader et al., 2003), initiated by Prl and Pg and showing increased specificity of action as the cascade develops. For example Stat5 and Elf5 are initiators of this cascade while Srebf1 appears later in the process and has effects restricted to lipogenesis. A number of transcription factors have important roles in alveolar
166 morphogenesis including Stat5 (Liu et al., 1997; Wakao et al., 1995), Id2 (Desprez et al., 2003), Ets (Coletta et al., 2004), Cebp (Grimm et al., 2003), NF-B (Cao et al., 2003), NF-1 (Murtagh et al., 2003) and Tcf/Lef (Hatsell et al., 2003). An example is the transcription factor Id2. Id2 is essential for lobuloalveolar development as Id2 knockout mice displayed failed alveolar morphogenesis. Id2 also promotes differentiation in MEC cultures, indicating Id2 is essential for the differentiation of the mammary epithelium (Desprez et al., 2003). Interestingly Id2 was demonstrated to be regulated by the paracrine mediator RankL (Kim et al., 2006), thus providing a link between Pg (Mulac-Jericevic et al., 2003) and possibly the Prlr, as we have shown that both RankL and its target NF-B are regulated by the Prlr (Ormandy et al., 2003). Prl also activates the transcription factor NF-1 resulting in expression of the milk gene Cel during alveolar morphogenesis through a post-translational mechanism requiring Jak2 (Nilsson et al., 2006). Therefore complex transcriptional networks exist to control alveolar morphogenesis, which are only beginning to be discovered (Figure 6.1). Alveolar morphogenesis also involves closure of tight junctions to maintain cellular polarity (Nguyen et al., 2001), and requires the withdrawal of Pg. We have demonstrated that Claudin 3 is regulated by Prlr signalling (Ormandy et al., 2003), and this protein is an important component of tight junctions (Morita et al., 1999). Prl also regulates Connexin 26, a component of gap junctions that allow the passage of ions and metabolites across cell membranes. Connexin 26 is also required in the mammary epithelium during alveolar morphogenesis (Bry et al., 2004). Withdrawal of Pg, triggers secretory activation phase of lactogenesis (Deis et al., 1983). This stage is characterised by the movement of milk and lipid proteins into the alveolar lumen in preparation for lactation. It is at the end of this stage that the synthesis of the various components of milk reaches maximal levels (Rudolph et al., 2003). In Chapter 3 we demonstrated that this process is primarily under control by Prl, and the actions of Gal were predominantly indirect, presumably via regulation of pituitary Prl secretion (Wynick et al., 1998), and regulation of Prl phosphorylation (Naylor et al., 2005b). Thus during secretory activation, the effects of Gal are not necessarily via direct effects on the epithelial cell. We also discovered that the actions of phosphorylated Prl during this phase of alveolar development were via antagonism of Prlr action. These studies demonstrate the dependence
167 on Prlr signalling during secretory activation, and add further information about the regulatory processes governing its regulation. A key transcription factor important this process if Srebf1, a key gene involved in the synthesis of lipids (Horton et al., 2002). In Chapter 3, we demonstrated that the expression of Srebf1 was under the control of Prlr signalling. Thus we can now place Srebf1 downstream of the Prlr. How Srebf fits in with the transcription hierarchy, and whether it is downstream of Elf5 requires further investigation. We also demonstrated that Aldo3 and Scd2, two key genes involved in de novo synthesis of lipids from glucose, were downstream of the Prlr/Stat5 cascade. Once again, whether these genes are direct targets of Stat5, or also require the presence of Elf5 needs to be determined. In addition, we discovered several other genes involved in regulation of transcription (Sox4 and Cebp), lipid synthesis (Angptl4, Acyl, Fdps, Elovl5, and Cyp51), transport (Copz1, Slc39a8, Slc34a2 and G1p2) and cell signalling (Erbb3), which are important for the secretory activation phase of mammary development. It would now be interesting to investigate how these genes fit into the signalling network driving secretory activation.
6.4 The role of prolactin in carcinogenesis The large Nurses Health Study demonstrated that patients with the highest levels of serum PRL confer a two-fold higher increase relative risk in developing breast cancer, independent of circulating estrogens (Hankinson et al., 1999; Tworoger et al., 2004). Prl has been consistently demonstrated as a weak mitogen for breast cancer cells (reviewed in (Clevenger et al., 2003)) and a potent mitogen and a complete carcinogen in animal models (Welsch et al., 1977). It has been hypothesised that Prl acts via its receptor within mammary epithelial cells to exert its effects during carcinogenesis, as both mammary tissue and breast cancer tissue express Prlr (Clevenger et al., 1995), but until now it has not been proven in vivo. Our studies in Chapter 5 of this thesis demonstrate a potent mitogenic action of epithelial Prl receptor in the growth and progression of pre-invasive lesions in the C3(1)/SV40T model of mouse mammary carcinogenesis. Given the role of Prlr signalling in the proliferation of the alveolar epithelium during normal alveolar morphogenesis, the effects of Prlr described in Chapter 5 are not surprising but this is the first demonstration of the normal developmental effects of prolactin cooperating with a potent oncogene to 168 enhance carcinogenesis. Little is known about the molecular mechanisms that are responsible for the mitogenic action of Prl in breast cancer and much remains to be discovered here. Estrogen is a potent mitogen and carcinogen for breast cancer (Reviewed in (Yager et al., 2006)), and the estrogen receptor antagonist tamoxifen is effective in preventing the recurrence of estrogen receptor and progesterone receptor positive breast cancer both at the site of the primary lesion and in the contralateral breast (Houssami et al., 2006; Martino et al., 2004). Prlr is associated with the PGR+ and ER+ breast cancers (Ormandy et al., 1997c), thus is it likely that they regulate and are regulated by similar mechanisms. A very good example of this is the inhibition of both ER and Prlr signalling by Caveolin 1 (Park et al., 2002; Sotgia et al., 2006). Caveolin 1 is an important regulator of cell signalling and proliferation mammalian cells and has a tumor suppressor role in the mammary gland (Williams et al., 2006). Loss of Caveolin-1 and Ink4a (Cdkn2a) results in proliferation of primary mammary epithelial fibroblasts (Williams et al., 2004), and Caveolin-1 is now emerging as an important gene in malignant transformation (Williams et al., 2005). Therefore up regulation of ER and Prlr in loss of function mutations of Caveolin-1 may play an important role in mediating breast cancer cell proliferation. The transcription factor, Foxa1 has been recently demonstrated as an important transcription factor in mediating the estrogen response in breast cancer cells (Carroll et al., 2005), and it is also regulated by the transcription factor Gata3 (Kouros-Mehr et al., 2006a). Our work in Chapter 5 and in other studies (Ormandy et al., 2003), places both these two transcription factors downstream of the Prlr, and thus indicates another point of convergence of PRLR and ER in breast cancer cells. These findings are fascinating and the mechanisms that lead to Prlr regulation of Foxa1 and other Er modulators should be investigated, as the modulation of the Prlr could be used alongside treatments aimed at inhibiting the action of estrogen receptor. Exogenous PRL is a weak mitogen for MCF7 breast cancer cells (Vonderhaar, 1989), however the effects of PRL are complicated by endogenous PRL production by these cells (Shaw-Bruha et al., 1997). To overcome these effects and study the mitogenic action of PRL, Schroeder and colleagues generated MCF7 breast cancer cells devoid of endogenous PRL production (Schroeder et al., 2002). PRL treatment in these cells resulted in increased cell proliferation and expression of Cyclin D1. Thus PRL appears to mediate
169 proliferation of breast cancer cells via Cyclin D1. The role of Cyclin D1 in breast cancer cell proliferation is well characterised (reviewed in (Sutherland et al., 2004)), and Cyclin D1 is one of the most commonly over expressed genes in breast cancer (Buckley et al., 1993). Cyclin D1 is potently induced by E and Pg (Sutherland et al., 1998), as well as the epidermal growth factor receptors (Richer et al., 1998), and therefore Prl modulation of the levels of the steroid and growth factor receptors (Ormandy et al., 1997c) has to be considered for the regulation of Cyclin D1, although this may be insignificant compared to the potent effects of estrogen. In addition, we have shown that the transcription factor NF- B can be regulated by Prl. Cyclin D1 is a target of NF-B transcription (Cao et al., 2001), and this pathway has been shown to have a role during carcinogenesis (Cao et al., 2003). In this thesis, we have demonstrated that Prl also regulated Cyclin D1 in the mammary epithelium during secretory activation. However our transcript profiling did not reveal Cyclin D1 as a differentially expressed gene in C3(1)/SV40T mammary epithelium prior to the development of neoplasia. Therefore Prl does not appear to modulate cell proliferation via Cyclin D1 in C3(1)/SV40T epithelium prior to the development of neoplasia. Given the findings in human breast cancer cells and the role of Prl in modulating Cyclin D1 in the normal mammary gland, Prl modulation of Cyclin D1 must be considered as a potential mechanism of Prl action in early breast cancer development, and may be revealed in other models of carcinogenesis. Prl can interact with the growth factor or receptor tyrosine kinases during carcinogenesis. The role of epidermal growth factor signaling in the pathogenesis in breast cancer is well recognized (Reviewed in (Gschwind et al., 2004)). Recently, Prl was shown to induce phosphorylation of the epidermal growth factor receptors including Erbb2 in T47D breast cancer cells (Huang et al., 2006). In this study, both PRL and EGF synergistically activated the MAPK/ERK pathway via the src homology collagen (SHC), an SH2-containing adaptor protein upstream of MAPK. Prl also interacts with a ligand of the epidermal growth factor receptor, TGF (Schroeder et al., 2001). In a mouse model where both Prl and TGF were co-expressed in the mammary gland, the latency to mammary macrocyst formation was significantly reduced, associated with increased Erk1/2 phosphorylation compared to TGF over-expression alone (Arendt et al., 2006). Thus Prl
170 in co-operation with the epidermal growth factor receptors and interaction with the MapK pathways may be a critical mechanism for proliferation during carcinogenesis. Tgf is a potent inhibitor of mammary cell proliferation and has a role as a tumor suppressor (Reviewed in (Wakefield et al., 2001)). In T47D breast cancer cells PRL inhibited the expression of TGF resulting in increased proliferation at PRL concentrations above physiological levels (Philips et al., 2004). In addition, Prl inhibits TGF induced apoptosis via activation of the PI3K/Akt pathway in normal mammary epithelial cells (Bailey et al., 2004). The role of Akt in cell survival (Schwertfeger et al., 2001), and the role of TGF in regulating cell proliferation via the Smad pathway (Wakefield et al., 2001) has been demonstrated, suggesting that this may be an important mechanism of Prl mediated cell survival in breast cancer. This is evidenced by the fact that PRL is a survival factor for breast cancer cells (Perks et al., 2004), however we did not demonstrate a survival role for Prlr in Chapter 5. This and the considerable lack of evidence suggesting that Prl acts to increase the survival of breast cancer cells, suggests that is more likely that PRL primarily acts as a mitogen and not a survival factor during carcinogenesis. The Mapk pathway has been demonstrated in mediating the proliferative response of Prlr in breast cancer cells (Das et al., 1996b), and inhibition of the Prlr receptor with a Prlr antagonist results in reduced proliferation and Mapk phosphorylation in T47D breast cancer cells (Llovera et al., 2000). Activation of the Mapk pathway is important for cell proliferation but not milk protein synthesis in Nb2 lymphoma cells (Yu et al., 1998). In addition Prl stimulated the proliferation of T47D and MCF7 breast cancer cells via activation of the Mapk and PI3K pathways resulting in Cyclin D1 transcription (Acosta et al., 2003). These effects were mediated by phosphorylation of the kinase Src. In addition PRL activation of the Mapk pathway in MCF7 cells without endogenous PRL, results in the activation of AP1 (Gutzman et al., 2004b). The AP1 transcription factor complex can be important in mediating proliferation, survival and invasion in breast cancer (Eferl et al., 2003). However, activation of the various components of AP1 such as Jun and Fos can result in both oncogenic and tumour suppressor functions, and requires further investigation. Therefore the activation of the Mapk and PI3K pathways may be important in conferring Prl induced proliferation during carcinogenesis.
171 The studies presented in this thesis have identified another potential mechanism of Prl induced proliferation. The mammary glands Prlr+/- mice which displayed reduced proliferation of lobuloalveoli, had reduced expression of genes involved in the DNA replication machinery (Bell et al., 2002) such as the Rfcs and the Mcmds. Interestingly, we also identified another member of the DNA machinery (Rfc3) as down regulated in C3(1)/SV40T transplants without the Prlr prior to the development of neoplasia. That we have identified these genes in two separate experiments implies that we have identified a novel mechanism of Prl action. The priming of cells by the Prlr in preparation for cell division represents an early event leading to the proliferation of cells under influence of other oncogenes, and is an intriguing mechanism for further investigation. Thus the mechanism of Prl mediated proliferation of breast cancer cells may be mediated through a variety mechanisms involving signaling through Stat5, Mapk, PI3K and modulation of the epidermal growth factor, Er and Pgr signaling. Many of these pathways represent normal mammary gland developmental pathways, which can be dysregulated during cancer. Some of these mechanisms are highlighted in red in Figure 6.1, although the precise mechanism by which Prl exerts its proliferative effects during carcinogenesis is, as yet, not defined. It is important to note that many of the studies aimed at identifying a mechanism for PRL mediated proliferation of breast cancer in vitro, are performed in immortalized breast cancer cells such as MCF7 and T47D breast cancer cells. They are complicated by their endogenous expression of Prl and the Prlr, which makes investigation of the exogenous effects of the Prl difficult (Clevenger et al., 1995). It is also unclear what stage of carcinogenesis these breast cancer cells represent. We have clearly shown in Chapter 5, that Prl acts to modulate the proliferation of pre-invasive lesions and its effects are lost in invasive and estrogen receptor negative breast cancer. Therefore we must attempt to design improved in vitro and in vitro models that better represent the earliest phases of carcinogenesis, in order to understand the mechanisms by which Prl exerts its effects in breast cancer.
6.5 Summary and future directions In summary, we have made significant contributions to the understanding of how Prl exerts its profound influence on the mammary epithelium during alveolar 172 morphogenesis and carcinogenesis. It is clear that the transcription factor Elf5 is a master regulator of Prl induced alveolar morphogenesis of the mammary gland during pregnancy, but how Elf5 fits in to the transcriptional hierarchy during alveolar morphogenesis remains to be defined. We have also identified several other regulatory molecules, which mediate the action of Prlr signalling during lactogenesis. More research is required to characterise the mechanisms by which these molecules contribute to milk secretion in the mammary gland. We have also shown that mammary epithelial Prlr can mediate the proliferation of pre-invasive lesions in the C3(1)/SV40T model of carcinogenesis. This is the first time that a cell autonomous role for Prlr during carcinogenesis in vivo has been shown. Importantly we demonstrated that these actions occurred specifically during the earliest phase of carcinogenesis, prior to the development of invasive and estrogen receptor negative carcinoma. The mechanisms modulated by Prlr signalling that are important for the proliferation of pre-invasive lesions must now be investigated, in order to design improved therapeutic or preventive strategies to breast cancer.
173 References
Acosta JJ, Munoz RM, Gonzalez L, Subtil-Rodriguez A, Dominguez-Caceres MA, Garcia-Martinez JM, Calcabrini A, Lazaro-Trueba I & Martin-Perez J. (2003). Src mediates prolactin-dependent proliferation of T47D and MCF7 cells via the activation of focal adhesion kinase/Erk1/2 and phosphatidylinositol 3-kinase pathways. Mol Endocrinol, 17(11), 2268-82. Allan GJ, Tonner E, Barber MC, Travers MT, Shand JH, Vernon RG, Kelly PA, Binart N & Flint DJ. (2002). Growth hormone, acting in part through the insulin-like growth factor axis, rescues developmental, but not metabolic, activity in the mammary gland of mice expressing a single allele of the prolactin receptor. Endocrinology, 143(11), 4310-9. Anderson E. (2002). The role of oestrogen and progesterone receptors in human mammary development and tumorigenesis. Breast Cancer Res., 4(5), 197-201. Anthony PK & Powers CA. (1993). Characterization of cysteamine induction of the 22K prolactin variant in the rat pituitary. Neuroendocrinology., 57(1), 167-76. Arendt LM, Rose-Hellekant TA, Sandgren EP & Schuler LA. (2006). Prolactin potentiates transforming growth factor alpha induction of mammary neoplasia in transgenic mice. Am J Pathol., 168(4), 1365-74. Attia MA & Weiss DW. (1966). Immunology of spontaneous mammary carcinomas in mice. V. Acquired tumor resistance and enhancement in strain A mice infected with mammary tumor virus. Cancer Research., 26(8), 1787-800. Bailey JP, Nieport KM, Herbst MP, Srivastava S, Serra RA & Horseman ND. (2004). Prolactin and transforming growth factor-beta signaling exert opposing effects on mammary gland morphogenesis, involution, and the Akt-forkhead pathway. Mol Endocrinol., 18(5), 1171-84. Baldocchi RA, Tan L, King DS & Nicoll CS. (1993). Mass spectrometric analysis of the fragments produced by cleavage and reduction of rat prolactin: evidence that the cleaving enzyme is cathepsin D. Endocrinology., 133(2), 935-8.
174 Ball RK, Friis RR, Schoenenberger CA, Doppler W & Groner B. (1988). Prolactin regulation of beta-casein gene expression and of a cytosolic 120-kd protein in a cloned mouse mammary epithelial cell line. Embo J., 7(7), 2089-95. Barcellos-Hoff MH, Aggeler J, Ram TG & Bissell MJ. (1989). Functional differentiation and alveolar morphogenesis of primary mammary cultures on reconstituted basement membrane. Development., 105(2), 223-35. Barcellos-Hoff MH & Ewan KB. (2000). Transforming growth factor-beta and breast cancer: Mammary gland development. Breast Cancer Res., 2(2), 92-9. Beigneux AP, Kosinski C, Gavino B, Horton JD, Skarnes WC & Young SG. (2004). ATP-citrate lyase deficiency in the mouse. J Biol Chem., 279(10), 9557-64. Bell SP & Dutta A. (2002). DNA replication in eukaryotic cells. Annu Rev Biochem., 71 333-74. Ben-Jonathan N, Mershon JL, Allen DL & Steinmetz RW. (1996). Extrapituitary prolactin: distribution, regulation, functions, and clinical aspects. Endocrine Reviews., 17(6), 639-69. Berlanga JJ, Gualillo O, Buteau H, Applanat M, Kelly PA & Edery M. (1997). Prolactin activates tyrosyl phosphorylation of insulin receptor substrate 1 and phosphatidylinositol-3-OH kinase. J Biol Chem., 272(4), 2050-2. Bernichtein S, Kinet S, Jeay S, Llovera M, Madern D, Martial JA, Kelly PA & Goffin V. (2001). S179D-human PRL, a pseudophosphorylated human PRL analog, is an agonist and not an antagonist. Endocrinology., 142(9), 3950-63. Berwaer M, Martial JA & Davis JR. (1994). Characterization of an up-stream promoter directing extrapituitary expression of the human prolactin gene. Mol Endocrinol., 8(5), 635-42. Berwaer M, Monget P, Peers B, Mathy-Hartert M, Bellefroid E, Davis JR, Belayew A & Martial JA. (1991). Multihormonal regulation of the human prolactin gene expression from 5000 bp of its upstream sequence. Mol Cell Endocrinol., 80(1-3), 53-64. Binart N, Helloco C, Ormandy CJ, Barra J, Clement-Lacroix P, Baran N & Kelly PA. (2000). Rescue of preimplantatory egg development and embryo implantation in prolactin receptor-deficient mice after progesterone administration. Endocrinology., 141(7), 2691-7.
175 Binart N, Imbert-Bollore P, Baran N, Viglietta C & Kelly PA. (2003). A short form of the prolactin (PRL) receptor is able to rescue mammopoiesis in heterozygous PRL receptor mice. Molecular Endocrinology., 17(6), 1066-74. Bird CE, Cook S, Owen S, Sterns EE & Clark AF. (1981). Plasma concentrations of C-19 steroids, estrogens, FSH, LH and prolactin in post-menopausal women with an without breast cancer. Oncology., 38(6), 365-8. Bocchinfuso WP & Korach KS. (1997). Mammary gland development and tumorigenesis in estrogen receptor knockout mice. J Mammary Gland Biol Neoplasia., 2(4), 323-34. Bole-Feysot C, Goffin V, Edery M, Binart N & Kelly PA. (1998). Prolactin (PRL) and its receptor: actions, signal transduction pathways and phenotypes observed in PRL receptor knockout mice. Endocrine Reviews., 19(3), 225-68. Bollengier F, Velkeniers B, Hooghe-Peters E, Mahler A & Vanhaelst L. (1989). Multiple forms of rat prolactin and growth hormone in pituitary cell subpopulations separated using a Percoll gradient system: disulphide-bridged dimers and glycosylated variants. J Endocrinol., 120(2), 201-6. Bonneterre J, Mauriac L, Weber B, Roche H, Fargeot P, Tubiana-Hulin M, Sevin M, Chollet P & Cappelaere P. (1988). Tamoxifen plus bromocriptine versus tamoxifen plus placebo in advanced breast cancer: results of a double blind multicentre clinical trial. European Journal of Cancer & Clinical Oncology., 24(12), 1851-3. Bonneterre J, Peyrat JP, Beuscart R, Lefebvre J & Demaille A. (1987). Prognostic significance of prolactin receptors in human breast cancer. Cancer Research., 47(17), 4724- 8. Boutin JM, Jolicoeur C, Okamura H, Gagnon J, Edery M, Shirota M, et al. (1988). Cloning and expression of the rat prolactin receptor, a member of the growth hormone/prolactin receptor gene family. Cell., 53(1), 69-77. Bradbury JM, Edwards PA, Niemeyer CC & Dale TC. (1995). Wnt-4 expression induces a pregnancy-like growth pattern in reconstituted mammary glands in virgin mice. Developmental Biology., 170(2), 553-63. Brisken C. (2002). Hormonal control of alveolar development and its implications for breast carcinogenesis. J Mammary Gland Biol Neoplasia, 7(1), 39-48.
176 Brisken C, Ayyannan A, Nguyen C, Heineman A, Reinhardt F, Jian T, et al. (2002). IGF-2 is a mediator of prolactin-induced morphogenesis in the breast. Developmental Cell, 3(6), 877-87. Brisken C, Heineman A, Chavarria T, Elenbaas B, Tan J, Dey SK, McMahon JA, McMahon AP & Weinberg RA. (2000). Essential function of Wnt-4 in mammary gland development downstream of progesterone signaling. Genes Dev, 14(6), 650-4. Brisken C, Kaur S, Chavarria TE, Binart N, Sutherland RL, Weinberg RA, Kelly PA & Ormandy CJ. (1999). Prolactin controls mammary gland development via direct and indirect mechanisms. Developmental Biology., 210(1), 96-106. Brisken C, Park S, Vass T, Lydon JP, O'Malley BW & Weinberg RA. (1998). A paracrine role for the epithelial progesterone receptor in mammary gland development. Proc Natl Acad Sci U S A, 95(9), 5076-81. Brisken C & Rajaram RD. (2006). Alveolar and Lactogenic Differentiation. J Mammary Gland Biol Neoplasia, 17 17. Brockman JL, Schroeder MD & Schuler LA. (2002). PRL activates the cyclin D1 promoter via the Jak2/Stat pathway. Molecular Endocrinology., 16(4), 774-84. Bry C, Maass K, Miyoshi K, Willecke K, Ott T, Robinson GW & Hennighausen L. (2004). Loss of connexin 26 in mammary epithelium during early but not during late pregnancy results in unscheduled apoptosis and impaired development. Dev Biol, 267(2), 418-29. Buck K, Vanek M, Groner B & Ball RK. (1992). Multiple forms of prolactin receptor messenger ribonucleic acid are specifically expressed and regulated in murine tissues and the mammary cell line HC11. Endocrinology., 130(3), 1108-14. Buckley MF, Sweeney KJ, Hamilton JA, Sini RL, Manning DL, Nicholson RI, et al. (1993). Expression and amplification of cyclin genes in human breast cancer. Oncogene., 8(8), 2127-33. Caldon CE, Daly RJ, Sutherland RL & Musgrove EA. (2006). Cell cycle control in breast cancer cells. J Cell Biochem., 97(2), 261-74. Cao Y, Bonizzi G, Seagroves TN, Greten FR, Johnson R, Schmidt EV & Karin M. (2001). IKKalpha provides an essential link between RANK signaling and cyclin D1 expression during mammary gland development. Cell., 107(6), 763-75.
177 Cao Y & Karin M. (2003). NF-kappaB in mammary gland development and breast cancer. J Mammary Gland Biol Neoplasia., 8(2), 215-23. Capecchi MR. (2005). Gene targeting in mice: functional analysis of the mammalian genome for the twenty-first century. Nat Rev Genet., 6(6), 507-12. Cappell DF. (1929). Intravitam and supravital staining. I. The principles and general results. The Journal of Pathology and Bacteriology, 32(3), 595-628. Cardiff RD, Anver MR, Gusterson BA, Hennighausen L, Jensen RA, Merino MJ, et al. (2000). The mammary pathology of genetically engineered mice: the consensus report and recommendations from the Annapolis meeting. Oncogene, 19(8), 968-88. Cardiff RD & Wellings SR. (1999). The comparative pathology of human and mouse mammary glands. J Mammary Gland Biol Neoplasia, 4(1), 105-22. Carroll JS, Liu XS, Brodsky AS, Li W, Meyer CA, Szary AJ, et al. (2005). Chromosome-wide mapping of estrogen receptor binding reveals long-range regulation requiring the forkhead protein FoxA1. Cell., 122(1), 33-43. Chadha N, Kohli R, Kumari GL & Muralidhar K. (1991). Physico-chemical and immunological characteristics of pituitary prolactin from water buffaloes (Bubalus bubalis). Mol Cell Biochem., 105(1), 61-71. Chen TJ, Kuo CB, Tsai KF, Liu JW, Chen DY & Walker AM. (1998). Development of recombinant human prolactin receptor antagonists by molecular mimicry of the phosphorylated hormone. Endocrinology., 139(2), 609-16. Chen WY, Ramamoorthy P, Chen N, Sticca R & Wagner TE. (1999). A human prolactin antagonist, hPRL-G129R, inhibits breast cancer cell proliferation through induction of apoptosis. Clinical Cancer Research, 5(11), 3583-93. Clapp C, Martial JA, Guzman RC, Rentier-Delure F & Weiner RI. (1993). The 16- kilodalton N-terminal fragment of human prolactin is a potent inhibitor of angiogenesis. Endocrinology., 133(3), 1292-9. Clark DE, Williams CC, Duplessis TT, Moring KL, Notwick AR, Long W, et al. (2005). ERBB4/HER4 potentiates STAT5A transcriptional activity by regulating novel STAT5A serine phosphorylation events. J Biol Chem., 280(25), 24175-80. Clark J, Merson S, Jhavar S, Flohr P, Edwards S, Foster CS, et al. (2006). Diversity of TMPRSS2-ERG fusion transcripts in the human prostate. Oncogene, 16 16.
178 Clarke LA, Wathes DC & Jabbour HN. (1997a). Expression and localization of prolactin receptor messenger ribonucleic acid in red deer ovary during the estrous cycle and pregnancy. Biol Reprod., 57(4), 865-72. Clarke RB, Howell A, Potten CS & Anderson E. (1997b). Dissociation between steroid receptor expression and cell proliferation in the human breast. Cancer Res., 57(22), 4987-91. Clarkson RW & Watson CJ. (1999). NF-kappaB and apoptosis in mammary epithelial cells. J Mammary Gland Biol Neoplasia., 4(2), 165-75. Clement-Lacroix P, Ormandy C, Lepescheux L, Ammann P, Damotte D, Goffin V, et al. (1999). Osteoblasts are a new target for prolactin: analysis of bone formation in prolactin receptor knockout mice. Endocrinology., 140(1), 96-105. Clevenger CV, Chang WP, Ngo W, Pasha TL, Montone KT & Tomaszewski JE. (1995). Expression of prolactin and prolactin receptor in human breast carcinoma. Evidence for an autocrine/paracrine loop. American Journal of Pathology., 146(3), 695-705. Clevenger CV, Furth PA, Hankinson SE & Schuler LA. (2003). The role of prolactin in mammary carcinoma. Endocrine Reviews., 24(1), 1-27. Coleman S & Daniel CW. (1990). Inhibition of mouse mammary ductal morphogenesis and down-regulation of the EGF receptor by epidermal growth factor. Dev Biol., 137(2), 425-33. Coletta RD, Jedlicka P, Gutierrez-Hartmann A & Ford HL. (2004). Transcriptional control of the cell cycle in mammary gland development and tumorigenesis. J Mammary Gland Biol Neoplasia, 9(1), 39-53. Conneely OM, Jericevic BM & Lydon JP. (2003). Progesterone receptors in mammary gland development and tumorigenesis. J Mammary Gland Biol Neoplasia., 8(2), 205-14. Cui Y, Riedlinger G, Miyoshi K, Tang W, Li C, Deng CX, Robinson GW & Hennighausen L. (2004). Inactivation of Stat5 in mouse mammary epithelium during pregnancy reveals distinct functions in cell proliferation, survival, and differentiation. Mol Cell Biol., 24(18), 8037-47.
179 Cunha GR, Young P, Hom YK, Cooke PS, Taylor JA & Lubahn DB. (1997). Elucidation of a role for stromal steroid hormone receptors in mammary gland growth and development using tissue recombinants. J Mammary Gland Biol Neoplasia., 2(4), 393-402. Daniel CW, Robinson S & Silberstein GB. (1996). The role of TGF-beta in patterning and growth of the mammary ductal tree. J Mammary Gland Biol Neoplasia., 1(4), 331-41. Daniel CW, Silberstein GB & Strickland P. (1987). Direct action of 17 beta- estradiol on mouse mammary ducts analyzed by sustained release implants and steroid autoradiography. Cancer Res., 47(22), 6052-7. Das R & Vonderhaar BK. (1995). Transduction of prolactin's (PRL) growth signal through both long and short forms of the PRL receptor. Mol Endocrinol., 9(12), 1750-9. Das R & Vonderhaar BK. (1996a). Activation of raf-1, MEK, and MAP kinase in prolactin responsive mammary cells. Breast Cancer Research & Treatment., 40(2), 141-9. Das R & Vonderhaar BK. (1996b). Involvement of SHC, GRB2, SOS and RAS in prolactin signal transduction in mammary epithelial cells. Oncogene., 13(6), 1139-45. Das R & Vonderhaar BK. (1997). Prolactin as a mitogen in mammary cells. Journal of Mammary Gland Biology & Neoplasia., 2(1), 29-39. DaSilva L, Rui H, Erwin RA, Howard OM, Kirken RA, Malabarba MG, Hackett RH, Larner AC & Farrar WL. (1996). Prolactin recruits STAT1, STAT3 and STAT5 independent of conserved receptor tyrosines TYR402, TYR479, TYR515 and TYR580. Mol Cell Endocrinol., 117(2), 131-40. Deis RP & Delouis C. (1983). Lactogenesis induced by ovariectomy in pregnant rats and its regulation by oestrogen and progesterone. J Steroid Biochem., 18(6), 687-90. DeOme KB, Faulkin LJJ, Bern HA & Blair PB. (1959). Development of mammary tumors from hyperplastic alveolar nodules transplanted into gland-free mammary fat pads of female C3H mice. Cancer Research., 19(5), 515-520. DePamphilis ML. (2003). The 'ORC cycle': a novel pathway for regulating eukaryotic DNA replication. Gene., 310 1-15. Desprez PY, Hara E, Bissell MJ & Campisi J. (1995). Suppression of mammary epithelial cell differentiation by the helix-loop-helix protein Id-1. Mol Cell Biol., 15(6), 3398-404.
180 Desprez PY, Sumida T & Coppe JP. (2003). Helix-loop-helix proteins in mammary gland development and breast cancer. J Mammary Gland Biol Neoplasia, 8(2), 225-39. Djonov V, Andres AC & Ziemiecki A. (2001). Vascular remodelling during the normal and malignant life cycle of the mammary gland. Microsc Res Tech., 52(2), 182-9. Donnison M, Beaton A, Davey HW, Broadhurst R, L'Huillier P & Pfeffer PL. (2005). Loss of the extraembryonic ectoderm in Elf5 mutants leads to defects in embryonic patterning. Development, 132(10), 2299-308. Dunbar ME, Dann PR, Robinson GW, Hennighausen L, Zhang JP & Wysolmerski JJ. (1999). Parathyroid hormone-related protein signaling is necessary for sexual dimorphism during embryonic mammary development. Development., 126(16), 3485-93. Eferl R & Wagner EF. (2003). AP-1: a double-edged sword in tumorigenesis. Nat Rev Cancer, 3(11), 859-68. Emanuele NV, Jurgens JK, Halloran MM, Tentler JJ, Lawrence AM & Kelley MR. (1992). The rat prolactin gene is expressed in brain tissue: detection of normal and alternatively spliced prolactin messenger RNA. Mol Endocrinol., 6(1), 35-42. Enoch HG, Catala A & Strittmatter P. (1976). Mechanism of rat liver microsomal stearyl-CoA desaturase. Studies of the substrate specificity, enzyme-substrate interactions, and the function of lipid. J Biol Chem., 251(16), 5095-103. Erwin RA, Kirken RA, Malabarba MG, Farrar WL & Rui H. (1995). Prolactin activates Ras via signaling proteins SHC, growth factor receptor bound 2, and son of sevenless. Endocrinology., 136(8), 3512-8. Ewan KB, Oketch-Rabah HA, Ravani SA, Shyamala G, Moses HL & Barcellos- Hoff MH. (2005). Proliferation of estrogen receptor-alpha-positive mammary epithelial cells is restrained by transforming growth factor-beta1 in adult mice. Am J Pathol., 167(2), 409-17. Ewan KB, Shyamala G, Ravani SA, Tang Y, Akhurst R, Wakefield L & Barcellos- Hoff MH. (2002). Latent transforming growth factor-beta activation in mammary gland: regulation by ovarian hormones affects ductal and alveolar proliferation. Am J Pathol., 160(6), 2081-93. Fahie-Wilson M. (2003). In hyperprolactinemia, testing for macroprolactin is essential. Clin Chem., 49(9), 1434-6.
181 Fantl V, Edwards PA, Steel JH, Vonderhaar BK & Dickson C. (1999). Impaired mammary gland development in Cyl-1(-/-) mice during pregnancy and lactation is epithelial cell autonomous. Developmental Biology., 212(1), 1-11. Fata JE, Kong YY, Li J, Sasaki T, Irie-Sasaki J, Moorehead RA, et al. (2000). The osteoclast differentiation factor osteoprotegerin-ligand is essential for mammary gland development. Cell, 103(1), 41-50. Fata JE, Werb Z & Bissell MJ. (2004). Regulation of mammary gland branching morphogenesis by the extracellular matrix and its remodeling enzymes. Breast Cancer Res, 6(1), 1-11. Fodor SP, Read JL, Pirrung MC, Stryer L, Lu AT & Solas D. (1991). Light- directed, spatially addressable parallel chemical synthesis. Science., 251(4995), 767-73. Foley J, Dann P, Hong J, Cosgrove J, Dreyer B, Rimm D, Dunbar M, Philbrick W & Wysolmerski J. (2001). Parathyroid hormone-related protein maintains mammary epithelial fate and triggers nipple skin differentiation during embryonic breast development. Development., 128(4), 513-25. Freeman ME, Kanyicska B, Lerant A & Nagy G. (2000). Prolactin: structure, function, and regulation of secretion. Physiol Rev, 80(4), 1523-631. Freemark M, Fleenor D, Driscoll P, Binart N & Kelly P. (2001). Body weight and fat deposition in prolactin receptor-deficient mice. Endocrinology., 142(2), 532-7. Fresno Vara JA, Caceres MA, Silva A & Martin-Perez J. (2001). Src family kinases are required for prolactin induction of cell proliferation. Mol Biol Cell., 12(7), 2171-83. Gigliotti AP, Johnson PF, Sterneck E & DeWille JW. (2003). Nulliparous CCAAT/enhancer binding proteindelta (C/EBPdelta) knockout mice exhibit mammary gland ductal hyperlasia. Exp Biol Med (Maywood). 228(3), 278-85. Ginsburg E & Vonderhaar BK. (1995). Prolactin synthesis and secretion by human breast cancer cells. Cancer Research., 55(12), 2591-5. Goffin V, Bernichtein S, Touraine P & Kelly PA. (2005). Development and potential clinical uses of human prolactin receptor antagonists. Endocr Rev, 6 6. Goffin V & Kelly PA. (1996a). Prolactin and growth hormone receptors. Clinical Endocrinology., 45(3), 247-55.
182 Goffin V & Kelly PA. (1997). The prolactin/growth hormone receptor family: structure/function relationships. Journal of Mammary Gland Biology & Neoplasia., 2(1), 7- 17. Goffin V, Kinet S, Ferrag F, Binart N, Martial JA & Kelly PA. (1996b). Antagonistic properties of human prolactin analogs that show paradoxical agonistic activity in the Nb2 bioassay. Journal of Biological Chemistry., 271(28), 16573-9. Gouilleux F, Wakao H, Mundt M & Groner B. (1994). Prolactin induces phosphorylation of Tyr694 of Stat5 (MGF), a prerequisite for DNA binding and induction of transcription. Embo J., 13(18), 4361-9. Gout PW, Beer CT & Noble RL. (1980). Prolactin-stimulated growth of cell cultures established from malignant Nb rat lymphomas. Cancer Res., 40(7), 2433-6. Graf L, Cseh G, Nagy I & Kurcz M. (1970). An evidence for deamidation of prolactin monomer. Acta Biochim Biophys Acad Sci Hung., 5(3), 299-303. Graham JD, Hunt SM, Tran N & Clarke CL. (1999). Regulation of the expression and activity by progestins of a member of the SOX gene family of transcriptional modulators. J Mol Endocrinol., 22(3), 295-304. Green JE, Desai KV, Ye Y, Kavanaugh C, Calvo A & Huh JI. (2004). Application of Gene Expression Profiling for Validating Models of Human Breast Cancer. Toxicol Pathol, 32(Supplement 1), 84-89. Green JE, Shibata MA, Yoshidome K, Liu ML, Jorcyk C, Anver MR, et al. (2000). The C3(1)/SV40 T-antigen transgenic mouse model of mammary cancer: ductal epithelial cell targeting with multistage progression to carcinoma. Oncogene., 19(8), 1020-7. Grimm SL, Contreras A, Barcellos-Hoff MH & Rosen JM. (2005). Cell cycle defects contribute to a block in hormone-induced mammary gland proliferation in CCAAT/enhancer-binding protein (C/EBPbeta)-null mice. J Biol Chem., 280(43), 36301-9. Grimm SL & Rosen JM. (2003). The role of C/EBPbeta in mammary gland development and breast cancer. Journal of Mammary Gland Biology & Neoplasia., 8(2), 191-204. Grimm SL, Seagroves TN, Kabotyanski EB, Hovey RC, Vonderhaar BK, Lydon JP, et al. (2002). Disruption of steroid and prolactin receptor patterning in the mammary gland
183 correlates with a block in lobuloalveolar development. Molecular Endocrinology., 16(12), 2675-91. Groner B, Fritsche M, Stocklin E, Berchtold S, Merkle C, Moriggl R & Pfitzner E. (2000). Regulation of the trans-activation potential of STAT5 through its DNA-binding activity and interactions with heterologous transcription factors. Growth Horm IGF Res., 10(Suppl B), S15-20. Gschwind A, Fischer OM & Ullrich A. (2004). The discovery of receptor tyrosine kinases: targets for cancer therapy. Nat Rev Cancer., 4(5), 361-70. Gumbiner BM. (1996). Cell adhesion: the molecular basis of tissue architecture and morphogenesis. Cell., 84(3), 345-57. Gunther EJ, Belka GK, Wertheim GB, Wang J, Hartman JL, Boxer RB & Chodosh LA. (2002). A novel doxycycline-inducible system for the transgenic analysis of mammary gland biology. Faseb J, 16(3), 283-92. Gutzman JH, Miller KK & Schuler LA. (2004a). Endogenous human prolactin and not exogenous human prolactin induces estrogen receptor alpha and prolactin receptor expression and increases estrogen responsiveness in breast cancer cells. Journal of Steroid Biochemistry & Molecular Biology., 88(1), 69-77. Gutzman JH, Rugowski DE, Schroeder MD, Watters JJ & Schuler LA. (2004b). Multiple Kinase Cascades Mediate Prolactin Signals to Activating Protein-1 in Breast Cancer Cells. Mol Endocrinol, 19 19. Han Y, Watling D, Rogers NC & Stark GR. (1997). JAK2 and STAT5, but not JAK1 and STAT1, are required for prolactin-induced beta-lactoglobulin transcription. Molecular Endocrinology., 11(8), 1180-8. Hankinson SE, Willett WC, Manson JE, Colditz GA, Hunter DJ, Spiegelman D, Barbieri RL & Speizer FE. (1998). Plasma sex steroid hormone levels and risk of breast cancer in postmenopausal women. Journal of the National Cancer Institute, 90(17), 1292- 9. Hankinson SE, Willett WC, Michaud DS, Manson JE, Colditz GA, Longcope C, Rosner B & Speizer FE. (1999). Plasma prolactin levels and subsequent risk of breast cancer in postmenopausal women. Journal of the National Cancer Institute., 91(7), 629-34.
184 Harris J, Stanford PM, Oakes SR & Ormandy CJ. (2004). Prolactin and the prolactin receptor: new targets of an old hormone. Ann Med, 36(6), 414-25. Harris J, Stanford PM, Sutherland K, Oakes SR, Naylor MJ, Robertson FG, et al. (2006). Socs2 and Elf5 mediate prolactin-induced mammary gland development. Mol Endocrinol, 9. Hartmann PE. (1973). Changes in the composition and yield of the mammary secretion of cows during the initiation of lactation. J Endocrinol., 59(2), 231-47. Hatsell S, Rowlands T, Hiremath M & Cowin P. (2003). Beta-catenin and Tcfs in mammary development and cancer. J Mammary Gland Biol Neoplasia., 8(2), 145-58. Hennighausen L & Robinson GW. (2005). Information networks in the mammary gland. Nature Reviews: Molecular Cell Biology, 6(9), 715-25. Hens JR & Wysolmerski JJ. (2005). Key stages of mammary gland development: molecular mechanisms involved in the formation of the embryonic mammary gland. Breast Cancer Res., 7(5), 220-4. Hill P, Wynder EL, Kumar H, Helman P, Rona G & Kuno K. (1976). Prolactin levels in populations at risk for breast cancer. Cancer Research., 36(11 Pt 1), 4102-6. Hinck L & Silberstein GB. (2005). Key stages in mammary gland development: the mammary end bud as a motile organ. Breast Cancer Res., 7(6), 245-51. Ho TW, Leong FS, Olaso CH & Walker AM. (1993). Secretion of specific nonphosphorylated and phosphorylated rat prolactin isoforms at different stages of the estrous cycle. Neuroendocrinology, 58(2), 160-5. Hoffmann T, Penel C & Ronin C. (1993). Glycosylation of human prolactin regulates hormone bioactivity and metabolic clearance. J Endocrinol Invest., 16(10), 807- 16. Horseman ND. (1999). Prolactin and mammary gland development. Journal of Mammary Gland Biology & Neoplasia., 4(1), 79-88. Horseman ND, Zhao W, Montecino-Rodriguez E, Tanaka M, Nakashima K, Engle SJ, Smith F, Markoff E & Dorshkind K. (1997). Defective mammopoiesis, but normal hematopoiesis, in mice with a targeted disruption of the prolactin gene. EMBO Journal., 16(23), 6926-35.
185 Horton JD, Goldstein JL & Brown MS. (2002). SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J Clin Invest., 109(9), 1125-31. Houssami N, Cuzick J & Dixon JM. (2006). The prevention, detection, and management of breast cancer. Med J Aust., 184(5), 230-4. Hovey RC, Trott JF & Vonderhaar BK. (2002). Establishing a framework for the functional mammary gland: from endocrinology to morphology. J Mammary Gland Biol Neoplasia, 7(1), 17-38. Huang Y, Li X, Jiang J & Frank SJ. (2006). Prolactin modulates phosphorylation, signaling and trafficking of epidermal growth factor receptor in human T47D breast cancer cells. Oncogene, 19 19. Huber O, Korn R, McLaughlin J, Ohsugi M, Herrmann BG & Kemler R. (1996). Nuclear localization of beta-catenin by interaction with transcription factor LEF-1. Mech Dev., 59(1), 3-10. Hui DY & Howles PN. (2002). Carboxyl ester lipase: structure-function relationship and physiological role in lipoprotein metabolism and atherosclerosis. J Lipid Res., 43(12), 2017-30. Humphreys RC. (1999). Programmed cell death in the terminal endbud. J Mammary Gland Biol Neoplasia., 4(2), 213-20. Humphreys RC, Krajewska M, Krnacik S, Jaeger R, Weiher H, Krajewski S, Reed JC & Rosen JM. (1996). Apoptosis in the terminal endbud of the murine mammary gland: a mechanism of ductal morphogenesis. Development., 122(12), 4013-22. Hutchinson JN & Muller WJ. (2000). Transgenic mouse models of human breast cancer. Oncogene, 19(53), 6130-7. Hutt JA, O'Rourke JP & DeWille J. (2000). Signal transducer and activator of transcription 3 activates CCAAT enhancer-binding protein delta gene transcription in G0 growth-arrested mouse mammary epithelial cells and in involuting mouse mammary gland. J Biol Chem., 275(37), 29123-31. Ihle JN. (1996). STATs and MAPKs: obligate or opportunistic partners in signaling. Bioessays., 18(2), 95-8.
186 Imbert A, Eelkema R, Jordan S, Feiner H & Cowin P. (2001). Delta N89 beta- catenin induces precocious development, differentiation, and neoplasia in mammary gland. J Cell Biol., 153(3), 555-68. Johansson EM, Kannius-Janson M, Gritli-Linde A, Bjursell G & Nilsson J. (2005). Nuclear factor 1-C2 is regulated by prolactin and shows a distinct expression pattern in the mouse mammary epithelial cells during development. Mol Endocrinol., 19(4), 992-1003. Epub 2005 Jan 6. Johnson A & O'Donnell M. (2005). Cellular DNA replicases: components and dynamics at the replication fork. Annu Rev Biochem., 74 283-315. Jones FE, Jerry DJ, Guarino BC, Andrews GC & Stern DF. (1996). Heregulin induces in vivo proliferation and differentiation of mammary epithelium into secretory lobuloalveoli. Cell Growth Differ., 7(8), 1031-8. Jones FE & Stern DF. (1999). Expression of dominant-negative ErbB2 in the mammary gland of transgenic mice reveals a role in lobuloalveolar development and lactation. Oncogene., 18(23), 3481-90. Kaneko H, Fujikawa T, Alam KS, Kaneko S, Tanaka M, Hibasami H & Nakashima K. (1998). Cooperative and differential effects of estrogen, prolactin, 22K and 20K human growth hormones on cyclin D1/PRAD1 gene expression in T-47D human breast cancer cells. Biochemistry & Molecular Biology International., 46(2), 411-4. Kannius-Janson M, Lidberg U, Hulten K, Gritli-Linde A, Bjursell G & Nilsson J. (1998). Studies of the regulation of the mouse carboxyl ester lipase gene in mammary gland. Biochem J., 336(Pt 3), 577-85. Kaplan LM, Gabriel SM, Koenig JI, Sunday ME, Spindel ER, Martin JB & Chin WW. (1988). Galanin is an estrogen-inducible, secretory product of the rat anterior pituitary. Proc Natl Acad Sci U S A., 85(19), 7408-12. Kelly PA, Djiane J, Banville D, Ali S, Edery M & Rozakis M. (1991). The growth hormone/prolactin receptor gene family. Oxf Surv Eukaryot Genes., 7 29-50. Kenney NJ, Smith GH, Rosenberg K, Cutler ML & Dickson RB. (1996). Induction of ductal morphogenesis and lobular hyperplasia by amphiregulin in the mouse mammary gland. Cell Growth Differ, 7(12), 1769-81.
187 Kim J, Luo W, Chen DT, Earley K, Tunstead J, Yu-Lee LY & Lin SH. (2003). Antitumor activity of the 16-kDa prolactin fragment in prostate cancer. Cancer Res., 63(2), 386-93. Kim NS, Kim HJ, Koo BK, Kwon MC, Kim YW, Cho Y, Yokota Y, Penninger JM & Kong YY. (2006). Receptor activator of NF-kappaB ligand regulates the proliferation of mammary epithelial cells via Id2. Mol Cell Biol., 26(3), 1002-13. Kline JB, Roehrs H & Clevenger CV. (1999). Functional characterization of the intermediate isoform of the human prolactin receptor. Journal of Biological Chemistry., 274(50), 35461-8. Klinowska TC, Soriano JV, Edwards GM, Oliver JM, Valentijn AJ, Montesano R & Streuli CH. (1999). Laminin and beta1 integrins are crucial for normal mammary gland development in the mouse. Dev Biol., 215(1), 13-32. Kohli R, Chadha N & Muralidhar K. (1988). Presence of tyrosine-O-sulfate in sheep pituitary prolactin. FEBS Lett., 242(1), 139-43. Kohmoto K. (1975). Mouse prolactin obtained by pituitary organ culture in a serum-free medium. Endocrinol Jpn., 22(5), 465-9. Kordon EC & Smith GH. (1998). An entire functional mammary gland may comprise the progeny from a single cell. Development., 125(10), 1921-30. Kouros-Mehr H, Slorach EM, Sternlicht MD & Werb Z. (2006a). GATA-3 Maintains the Differentiation of the Luminal Cell Fate in the Mammary Gland. Cell., 127(5), 1041-55. Kouros-Mehr H & Werb Z. (2006b). Candidate regulators of mammary branching morphogenesis identified by genome-wide transcript analysis. Dev Dyn., 235(12), 3404-12. Kratochwil K. (1977). Development and loss of androgen responsiveness in the embryonic rudiment of the mouse mammary gland. Dev Biol., 61(2), 358-65. Kratochwil K & Schwartz P. (1976). Tissue interaction in androgen response of embryonic mammary rudiment of mouse: identification of target tissue for testosterone. Proc Natl Acad Sci U S A., 73(11), 4041-4. Kubo M, Hanada T & Yoshimura A. (2003). Suppressors of cytokine signaling and immunity. Nat Immunol., 4(12), 1169-76.
188 Kulski JK, Hartmann PE, Martin JD & Smith M. (1978). Effects of bromocriptine mesylate on the composition of the mammary secretion in non-breast-feeding women. Obstet Gynecol., 52(1), 38-42. Kunzelmann K. (2005). Ion channels and cancer. J Membr Biol., 205(3), 159-73. Kuo CB, Wu W, Xu X, Yang L, Chen C, Coss D, Birdsall B, Nasseri D & Walker AM. (2002). Pseudophosphorylated prolactin (S179D PRL) inhibits growth and promotes beta-casein gene expression in the rat mammary gland. Cell Tissue Res, 309(3), 429-37. Epub 2002 Jul 12. Kurpios NA, Sabolic NA, Shepherd TG, Fidalgo GM & Hassell JA. (2003). Function of PEA3 Ets transcription factors in mammary gland development and oncogenesis. J Mammary Gland Biol Neoplasia., 8(2), 177-90. Kwa HG, Cleton F, Wang DY, Bulbrook RD, Bulstrode JC, Hayward JL, Millis RR & Cuzick J. (1981). A prospective study of plasma prolactin levels and subsequent risk of breast cancer. International Journal of Cancer., 28(6), 673-6. Laird NM & Ware JH. (1982). Random-effects models for longitudinal data. Biometrics, 38(4), 963-74. Landberg G & Roos G. (1997). The cell cycle in breast cancer. Apmis., 105(8), 575- 89. Lang F, Foller M, Lang KS, Lang PA, Ritter M, Gulbins E, Vereninov A & Huber SM. (2005). Ion channels in cell proliferation and apoptotic cell death. J Membr Biol., 205(3), 147-57. Lebrun JJ, Ali S, Goffin V, Ullrich A & Kelly PA. (1995a). A single phosphotyrosine residue of the prolactin receptor is responsible for activation of gene transcription. Proceedings of the National Academy of Sciences of the United States of America., 92(9), 4031-5. Lebrun JJ, Ali S, Ullrich A & Kelly PA. (1995b). Proline-rich sequence-mediated Jak2 association to the prolactin receptor is required but not sufficient for signal transduction. J Biol Chem., 270(18), 10664-70. Leonard AE, Kelder B, Bobik EG, Chuang LT, Lewis CJ, Kopchick JJ, Mukerji P & Huang YS. (2002). Identification and expression of mammalian long-chain PUFA elongation enzymes. Lipids., 37(8), 733-40.
189 Lewis UJ, Singh RN, Sinha YN & Vanderlaan WP. (1985). Glycosylated human prolactin. Endocrinology., 116(1), 359-63. Li L, Cleary S, Mandarano MA, Long W, Birchmeier C & Jones FE. (2002). The breast proto-oncogene, HRGalpha regulates epithelial proliferation and lobuloalveolar development in the mouse mammary gland. Oncogene., 21(32), 4900-7. Li N, Zhang Y, Naylor MJ, Schatzmann F, Maurer F, Wintermantel T, et al. (2005). Beta1 integrins regulate mammary gland proliferation and maintain the integrity of mammary alveoli. Embo J., 24(11), 1942-53. Lindeman GJ, Wittlin S, Lada H, Naylor MJ, Santamaria M, Zhang JG, et al. (2001). SOCS1 deficiency results in accelerated mammary gland development and rescues lactation in prolactin receptor-deficient mice. Genes & Development., 15(13), 1631-6. Ling C, Hellgren G, Gebre-Medhin M, Dillner K, Wennbo H, Carlsson B & Billig H. (2000). Prolactin (PRL) receptor gene expression in mouse adipose tissue: increases during lactation and in PRL-transgenic mice. Endocrinology, 141(10), 3564-72. Liu JH, Lee DW & Markoff E. (1990). Differential release of prolactin variants in postpartum and early follicular phase women. J Clin Endocrinol Metab., 71(3), 605-10. Liu X, Gallego MI, Smith GH, Robinson GW & Hennighausen L. (1998). Functional rescue of Stat5a-null mammary tissue through the activation of compensating signals including Stat5b. Cell Growth Differ., 9(9), 795-803. Liu X, Robinson GW, Gouilleux F, Groner B & Hennighausen L. (1995). Cloning and expression of Stat5 and an additional homologue (Stat5b) involved in prolactin signal transduction in mouse mammary tissue. Proc Natl Acad Sci U S A, 92(19), 8831-5. Liu X, Robinson GW, Wagner KU, Garrett L, Wynshaw-Boris A & Hennighausen L. (1997). Stat5a is mandatory for adult mammary gland development and lactogenesis. Genes & Development., 11(2), 179-86. Llovera M, Pichard C, Bernichtein S, Jeay S, Touraine P, Kelly PA & Goffin V. (2000). Human prolactin (hPRL) antagonists inhibit hPRL-activated signaling pathways involved in breast cancer cell proliferation. Oncogene., 19(41), 4695-705. Long W, Wagner KU, Lloyd KC, Binart N, Shillingford JM, Hennighausen L & Jones FE. (2003). Impaired differentiation and lactational failure of Erbb4-deficient
190 mammary glands identify ERBB4 as an obligate mediator of STAT5. Development., 130(21), 5257-68. Lucas BK, Ormandy CJ, Binart N, Bridges RS & Kelly PA. (1998). Null mutation of the prolactin receptor gene produces a defect in maternal behavior. Endocrinology., 139(10), 4102-7. Luetteke NC, Qiu TH, Fenton SE, Troyer KL, Riedel RF, Chang A & Lee DC. (1999). Targeted inactivation of the EGF and amphiregulin genes reveals distinct roles for EGF receptor ligands in mouse mammary gland development. Development., 126(12), 2739-50. MacMahon B, Cole P, Lin TM, Lowe CR, Mirra AP, Ravnihar B, Salber EJ, Valaoras VG & Yuasa S. (1970). Age at first birth and breast cancer risk. Bull World Health Organ., 43(2), 209-21. Manni A, Boucher AE, Demers LM, Harvey HA, Lipton A, Simmonds MA & Bartholomew M. (1989). Endocrine effects of combined somatostatin analog and bromocriptine therapy in women with advanced breast cancer. Breast Cancer Research & Treatment., 14(3), 289-98. Markoff E, Lee DW & Hollingsworth DR. (1988). Glycosylated and nonglycosylated prolactin in serum during pregnancy. J Clin Endocrinol Metab., 67(3), 519-23. Maroulakou IG, Anver M, Garrett L & Green JE. (1994). Prostate and mammary adenocarcinoma in transgenic mice carrying a rat C3(1) simian virus 40 large tumor antigen fusion gene. Proceedings of the National Academy of Sciences of the United States of America., 91(23), 11236-40. Martino S, Cauley JA, Barrett-Connor E, Powles TJ, Mershon J, Disch D, Secrest RJ & Cummings SR. (2004). Continuing outcomes relevant to Evista: breast cancer incidence in postmenopausal osteoporotic women in a randomized trial of raloxifene. J Natl Cancer Inst., 96(23), 1751-61. Mason RJ, Pan T, Edeen KE, Nielsen LD, Zhang F, Longphre M, Eckart MR & Neben S. (2003). Keratinocyte growth factor and the transcription factors C/EBP alpha, C/EBP delta, and SREBP-1c regulate fatty acid synthesis in alveolar type II cells. J Clin Invest., 112(2), 244-55.
191 Massague J & Chen YG. (2000). Controlling TGF-beta signaling. Genes Dev., 14(6), 627-44. Matsuda M, Imaoka T, Vomachka AJ, Gudelsky GA, Hou Z, Mistry M, et al. (2004). Serotonin regulates mammary gland development via an autocrine-paracrine loop. Dev Cell, 6(2), 193-203. McGowan EM & Clarke CL. (1999). Effect of overexpression of progesterone receptor A on endogenous progestin-sensitive endpoints in breast cancer cells. Mol Endocrinol., 13(10), 1657-71. McManaman JL, Palmer CA, Anderson S, Schwertfeger K & Neville MC. (2004). Regulation of milk lipid formation and secretion in the mouse mammary gland. Adv Exp Med Biol., 554 263-79. Meng J, Tsai-Morris CH & Dufau ML. (2004). Human prolactin receptor variants in breast cancer: low ratio of short forms to the long-form human prolactin receptor associated with mammary carcinoma. Cancer Res, 64(16), 5677-82. Mershon J, Sall W, Mitchner N & Ben-Jonathan N. (1995). Prolactin is a local growth factor in rat mammary tumors. Endocrinology., 136(8), 3619-23. Missmer SA, Eliassen AH, Barbieri RL & Hankinson SE. (2004). Endogenous estrogen, androgen, and progesterone concentrations and breast cancer risk among postmenopausal women. Journal of the National Cancer Institute, 96(24), 1856-65. Miyazaki M, Dobrzyn A, Elias PM & Ntambi JM. (2005). Stearoyl-CoA desaturase-2 gene expression is required for lipid synthesis during early skin and liver development. Proc Natl Acad Sci U S A., 102(35), 12501-6. Miyoshi K, Meyer B, Gruss P, Cui Y, Renou JP, Morgan FV, et al. (2002). Mammary epithelial cells are not able to undergo pregnancy-dependent differentiation in the absence of the helix-loop-helix inhibitor Id2. Molecular Endocrinology., 16(12), 2892- 901. Mombaerts P, Iacomini J, Johnson RS, Herrup K, Tonegawa S & Papaioannou VE. (1992). RAG-1-deficient mice have no mature B and T lymphocytes. Cell., 68(5), 869-77. Mori S, Nishikawa SI & Yokota Y. (2000). Lactation defect in mice lacking the helix-loop-helix inhibitor Id2. Embo J., 19(21), 5772-81.
192 Morita K, Furuse M, Fujimoto K & Tsukita S. (1999). Claudin multigene family encoding four-transmembrane domain protein components of tight junction strands. Proc Natl Acad Sci U S A., 96(2), 511-6. Mulac-Jericevic B, Lydon JP, DeMayo FJ & Conneely OM. (2003). Defective mammary gland morphogenesis in mice lacking the progesterone receptor B isoform. Proc Natl Acad Sci U S A., 100(17), 9744-9. Murtagh J, Martin F & Gronostajski RM. (2003). The Nuclear Factor I (NFI) gene family in mammary gland development and function. J Mammary Gland Biol Neoplasia., 8(2), 241-54. Nadeau JH, Balling R, Barsh G, Beier D, Brown SD, Bucan M, et al. (2001). Sequence interpretation. Functional annotation of mouse genome sequences. Science., 291(5507), 1251-5. Navolanic PM & McCubrey JA. (2005). Pharmacological breast cancer therapy (review). Int J Oncol., 27(5), 1341-4. Naylor M & Streuli C. (2006). Integrins and Development. Danin, E. (ed.). Landes Bioscience: Georgetown, Tx., pp 176-185. Naylor MJ, Ginsburg E, Iismaa TP, Vonderhaar BK, Wynick D & Ormandy CJ. (2003a). The neuropeptide galanin augments lobuloalveolar development. J Biol Chem, 278(31), 29145-52. Naylor MJ, Li N, Cheung J, Lowe ET, Lambert E, Marlow R, et al. (2005a). Ablation of {beta}1 integrin in mammary epithelium reveals a key role for integrin in glandular morphogenesis and differentiation. J Cell Biol, 171(4), 717-28. Naylor MJ, Lockefeer JA, Horseman ND & Ormandy CJ. (2003b). Prolactin regulates mammary epithelial cell proliferation via autocrine/paracrine mechanism. Endocrine., 20(1-2), 111-4. Naylor MJ, Oakes SR, Gardiner-Garden M, Harris J, Blazek K, Ho TW, et al. (2005b). Transcriptional changes underlying the secretory activation phase of mammary gland development. Mol Endocrinol, 19(7), 1868-83. Naylor MJ & Ormandy CJ. (2002). Mouse strain-specific patterns of mammary epithelial ductal side branching are elicited by stromal factors. Dev Dyn, 225(1), 100-5.
193 Naylor S, Smalley MJ, Robertson D, Gusterson BA, Edwards PA & Dale TC. (2000). Retroviral expression of Wnt-1 and Wnt-7b produces different effects in mouse mammary epithelium. J Cell Sci., 113(Pt 12), 2129-38. Neville MC, McFadden TB & Forsyth I. (2002). Hormonal regulation of mammary differentiation and milk secretion. Journal of Mammary Gland Biology & Neoplasia., 7(1), 49-66. Neville MC, Medina D, Monks J & Hovey RC. (1998). The mammary fat pad. J Mammary Gland Biol Neoplasia., 3(2), 109-16. Nguyen DA, Parlow AF & Neville MC. (2001). Hormonal regulation of tight junction closure in the mouse mammary epithelium during the transition from pregnancy to lactation. J Endocrinol, 170(2), 347-56. Nilsson J, Bjursell G & Kannius-Janson M. (2006). Nuclear Jak2 and transcription factor NF1-C2: a novel mechanism of prolactin signaling in mammary epithelial cells. Mol Cell Biol., 26(15), 5663-74. Nishikawa S, Moore RC, Nonomura N & Oka T. (1994). Progesterone and EGF inhibit mouse mammary gland prolactin receptor and beta-casein gene expression. Am J Physiol., 267(5 Pt 1), C1467-72. Oakes SR, Hilton HN & Ormandy CJ. (2006). The alveolar switch: coordinating the proliferative cues and cell fate decisions that drive the formation of lobuloalveoli from ductal epithelium. Breast Cancer Res., 8(2), 207. Oettgen P, Kas K, Dube A, Gu X, Grall F, Thamrongsak U, et al. (1999). Characterization of ESE-2, a novel ESE-1-related Ets transcription factor that is restricted to glandular epithelium and differentiated keratinocytes. J Biol Chem, 274(41), 29439-52. Oetting WS, Tuazon PT, Traugh JA & Walker AM. (1986). Phosphorylation of prolactin. J Biol Chem., 261(4), 1649-52. Oikawa T & Yamada T. (2003). Molecular biology of the Ets family of transcription factors. Gene, 303 11-34. Ormandy CJ, Binart N & Kelly PA. (1997a). Mammary gland development in prolactin receptor knockout mice. Journal of Mammary Gland Biology & Neoplasia., 2(4), 355-64.
194 Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, Buteau H, et al. (1997b). Null mutation of the prolactin receptor gene produces multiple reproductive defects in the mouse. Genes & Development., 11(2), 167-78. Ormandy CJ, de Fazio A, Kelly PA & Sutherland RL. (1992). Coordinate regulation of oestrogen and prolactin receptor expression by sodium butyrate in human breast cancer cells. Biochemical & Biophysical Research Communications., 182(2), 740-5. Ormandy CJ, Hall RE, Manning DL, Robertson JF, Blamey RW, Kelly PA, Nicholson RI & Sutherland RL. (1997c). Coexpression and cross-regulation of the prolactin receptor and sex steroid hormone receptors in breast cancer. Journal of Clinical Endocrinology & Metabolism., 82(11), 3692-9. Ormandy CJ, Naylor M, Harris J, Robertson F, Horseman ND, Lindeman GJ, Visvader J & Kelly PA. (2003). Investigation of the transcriptional changes underlying functional defects in the mammary glands of prolactin receptor knockout mice. Recent Progress in Hormone Research., 58 297-323. Park DS, Lee H, Frank PG, Razani B, Nguyen AV, Parlow AF, et al. (2002). Caveolin-1-deficient mice show accelerated mammary gland development during pregnancy, premature lactation, and hyperactivation of the Jak-2/STAT5a signaling cascade. Mol Biol Cell, 13(10), 3416-30. Parrinello S, Lin CQ, Murata K, Itahana Y, Singh J, Krtolica A, Campisi J & Desprez PY. (2001). Id-1, ITF-2, and Id-2 comprise a network of helix-loop-helix proteins that regulate mammary epithelial cell proliferation, differentiation, and apoptosis. J Biol Chem., 276(42), 39213-9. Perks CM, Keith AJ, Goodhew KL, Savage PB, Winters ZE & Holly JM. (2004). Prolactin acts as a potent survival factor for human breast cancer cell lines. Br J Cancer, 22 22. Peyrat JP, Vennin P, Bonneterre J, Hecquet B, Vandewalle B, Kelly PA & Djiane J. (1984). Effect of bromocriptin treatment on prolactin and steroid receptor levels in human breast cancer. European Journal of Cancer & Clinical Oncology., 20(11), 1363-7. Pezet A, Buteau H, Kelly PA & Edery M. (1997). The last proline of Box 1 is essential for association with JAK2 and functional activation of the prolactin receptor. Mol Cell Endocrinol., 129(2), 199-208.
195 Philips N & McFadden K. (2004). Inhibition of transforming growth factor-beta and matrix metalloproteinases by estrogen and prolactin in breast cancer cells. Cancer Letters., 206(1), 63-8. Piccoletti R, Bendinelli P & Maroni P. (1997). Signal transduction pathway of prolactin in rat liver. Molecular & Cellular Endocrinology., 135(2), 169-77. Pierce DF, Jr., Johnson MD, Matsui Y, Robinson SD, Gold LI, Purchio AF, Daniel CW, Hogan BL & Moses HL. (1993). Inhibition of mammary duct development but not alveolar outgrowth during pregnancy in transgenic mice expressing active TGF-beta 1. Genes Dev., 7(12A), 2308-17. Pinheiro J, Bates D, DebRoy S & Sarkar D. (2005). nlme: Linear and nonlinear mixed effects models. R package version 3.1-65. Pittius CW, Sankaran L, Topper YJ & Hennighausen L. (1988). Comparison of the regulation of the whey acidic protein gene with that of a hybrid gene containing the whey acidic protein gene promoter in transgenic mice. Mol Endocrinol., 2(11), 1027-32. Rameil P, Lecine P, Ghysdael J, Gouilleux F, Kahn-Perles B & Imbert J. (2000). IL-2 and long-term T cell activation induce physical and functional interaction between STAT5 and ETS transcription factors in human T cells. Oncogene., 19(17), 2086-97. Reynolds C, Montone KT, Powell CM, Tomaszewski JE & Clevenger CV. (1997). Expression of prolactin and its receptor in human breast carcinoma. Endocrinology., 138(12), 5555-60. Richer JK, Lange CA, Manning NG, Owen G, Powell R & Horwitz KB. (1998). Convergence of progesterone with growth factor and cytokine signaling in breast cancer. Progesterone receptors regulate signal transducers and activators of transcription expression and activity. Journal of Biological Chemistry., 273(47), 31317-26. Richert MM, Schwertfeger KL, Ryder JW & Anderson SM. (2000). An atlas of mouse mammary gland development. J Mammary Gland Biol Neoplasia, 5(2), 227-41. Robertson FG, Harris J, Naylor MJ, Oakes SR, Kindblom J, Dillner K, et al. (2003). Prostate development and carcinogenesis in prolactin receptor knockout mice. Endocrinology., 144(7), 3196-205.
196 Robinson GW, Johnson PF, Hennighausen L & Sterneck E. (1998). The C/EBPbeta transcription factor regulates epithelial cell proliferation and differentiation in the mammary gland. Genes Dev., 12(12), 1907-16. Robinson GW, McKnight RA, Smith GH & Hennighausen L. (1995). Mammary epithelial cells undergo secretory differentiation in cycling virgins but require pregnancy for the establishment of terminal differentiation. Development, 121(7), 2079-90. Rose-Hellekant TA, Arendt LM, Schroeder MD, Gilchrist K, Sandgren EP & Schuler LA. (2003). Prolactin induces ERalpha-positive and ERalpha-negative mammary cancer in transgenic mice. Oncogene., 22(30), 4664-74. Rosenfeld MG, Lunyak VV & Glass CK. (2006). Sensors and signals: a coactivator/corepressor/epigenetic code for integrating signal-dependent programs of transcriptional response. Genes Dev., 20(11), 1405-28. Rozakis-Adcock M & Kelly PA. (1991). Mutational analysis of the ligand-binding domain of the prolactin receptor. J Biol Chem., 266(25), 16472-7. Rozakis-Adcock M & Kelly PA. (1992). Identification of ligand binding determinants of the prolactin receptor. J Biol Chem., 267(11), 7428-33. Ruan W, Newman CB & Kleinberg DL. (1992). Intact and amino-terminally shortened forms of insulin-like growth factor I induce mammary gland differentiation and development. Proc Natl Acad Sci U S A., 89(22), 10872-6. Rubin MA & Chinnaiyan AM. (2006). Bioinformatics approach leads to the discovery of the TMPRSS2:ETS gene fusion in prostate cancer. Lab Invest., 86(11), 1099- 102. Rudolph MC, McManaman JL, Hunter L, Phang T & Neville MC. (2003). Functional development of the mammary gland: use of expression profiling and trajectory clustering to reveal changes in gene expression during pregnancy, lactation, and involution. J Mammary Gland Biol Neoplasia, 8(3), 287-307. Said TK, Conneely OM, Medina D, O'Malley BW & Lydon JP. (1997). Progesterone, in addition to estrogen, induces cyclin D1 expression in the murine mammary epithelial cell, in vivo. Endocrinology., 138(9), 3933-9.
197 Schmitt-Ney M, Happ B, Ball RK & Groner B. (1992). Developmental and environmental regulation of a mammary gland-specific nuclear factor essential for transcription of the gene encoding beta-casein. Proc Natl Acad Sci U S A., 89(7), 3130-4. Schroeder MD, Brockman JL, Walker AM & Schuler LA. (2003). Inhibition of prolactin (PRL)-induced proliferative signals in breast cancer cells by a molecular mimic of phosphorylated PRL, S179D-PRL. Endocrinology., 144(12), 5300-7. Schroeder MD, Rose-Hellekant TA, Sandgren EP & Schuler LA. (2001). Dysregulation of mammary Stats 1,3 and 5 and PRL receptors by overexpression of TGFalpha. Molecular & Cellular Endocrinology., 175(1-2), 173-83. Schroeder MD, Symowicz J & Schuler LA. (2002). PRL modulates cell cycle regulators in mammary tumor epithelial cells. Molecular Endocrinology., 16(1), 45-57. Schwertfeger KL, Richert MM, Anderson SM & Ryder JW. (2001). Mammary gland involution is delayed by activated Akt in transgenic mice. Mol Endocrinol., 15(6), 867-81. Seagroves TN, Krnacik S, Raught B, Gay J, Burgess-Beusse B, Darlington GJ & Rosen JM. (1998). C/EBPbeta, but not C/EBPalpha, is essential for ductal morphogenesis, lobuloalveolar proliferation, and functional differentiation in the mouse mammary gland. Genes Dev., 12(12), 1917-28. Seagroves TN, Lydon JP, Hovey RC, Vonderhaar BK & Rosen JM. (2000). C/EBPbeta (CCAAT/enhancer binding protein) controls cell fate determination during mammary gland development. Mol Endocrinol., 14(3), 359-68. Shackleton M, F. V, Simpson KJ, Stingl J, Smythe GK, Asselin-Labat M-L, Wu L, Lindeman GJ & Visvader JE. (2006). Generation of a functional mammary gland from a single stem cell. Nature, 439 84-88. Sharrocks AD. (2001). The ETS-domain transcription factor family. Nat Rev Mol Cell Biol., 2(11), 827-37. Sharrocks AD, Brown AL, Ling Y & Yates PR. (1997). The ETS-domain transcription factor family. Int J Biochem Cell Biol, 29(12), 1371-87. Shaw-Bruha CM, Pirrucello SJ & Shull JD. (1997). Expression of the prolactin gene in normal and neoplastic human breast tissues and human mammary cell lines: promoter
198 usage and alternative mRNA splicing. Breast Cancer Research & Treatment., 44(3), 243- 53. Shibata MA, Jorcyk CL, Liu ML, Yoshidome K, Gold LG & Green JE. (1998). The C3(1)/SV40 T antigen transgenic mouse model of prostate and mammary cancer. Toxicologic Pathology., 26(1), 177-82. Shome B & Parlow AF. (1977). Human pituitary prolactin (hPRL): the entire linear amino acid sequence. J Clin Endocrinol Metab., 45(5), 1112-5. Sicinski P & Weinberg RA. (1997). A specific role for cyclin D1 in mammary gland development. J Mammary Gland Biol Neoplasia., 2(4), 335-42. Silberstein GB, Strickland P, Coleman S & Daniel CW. (1990). Epithelium- dependent extracellular matrix synthesis in transforming growth factor-beta 1-growth- inhibited mouse mammary gland. J Cell Biol., 110(6), 2209-19. Silberstein GB, Van Horn K, Shyamala G & Daniel CW. (1994). Essential role of endogenous estrogen in directly stimulating mammary growth demonstrated by implants containing pure antiestrogens. Endocrinology., 134(1), 84-90. Sinha YN. (1995). Structural variants of prolactin: occurrence and physiological significance. Endocrine Reviews., 16(3), 354-69. Sinha YN & Gilligan TA. (1981). Identification of a less immunoreactive form of prolactin in the rat pituitary. Endocrinology., 108(3), 1091-4. Sinha YN & Jacobsen BP. (1987a). Glycosylated prolactin in the murine pituitary: detection by a novel assay and alteration of concentrations by physiological and pharmacological stimuli. Biochemical & Biophysical Research Communications., 148(1), 505-14. Sinha YN & Jacobsen BP. (1988). Structural and immunologic evidence for a small molecular weight ("21K") variant of prolactin. Endocrinology., 123(3), 1364-70. Sinha YN, Ott K & Vanderlaan WP. (1987b). Detection of multiple PRL- and GH- like proteins in human pituitary by Western blot analysis. American Journal of the Medical Sciences., 294(1), 15-25. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A & McGuire WL. (1987). Human breast cancer: correlation of relapse and survival with amplification of the HER- 2/neu oncogene. Science., 235(4785), 177-82.
199 Smalley M & Ashworth A. (2003). Stem cells and breast cancer: A field in transit. Nat Rev Cancer., 3(11), 832-44. Smalley MJ & Clarke RB. (2005). The mammary gland "side population": a putative stem/progenitor cell marker? J Mammary Gland Biol Neoplasia., 10(1), 37-47. Smith GH. (1996). Experimental mammary epithelial morphogenesis in an in vivo model: evidence for distinct cellular progenitors of the ductal and lobular phenotype. Breast Cancer Res Treat., 39(1), 21-31. Smith GH & Boulanger CA. (2003). Mammary epithelial stem cells: transplantation and self-renewal analysis. Cell Prolif., 36(Suppl 1), 3-15. Smith TP, Suliman AM, Fahie-Wilson MN & McKenna TJ. (2002). Gross variability in the detection of prolactin in sera containing big big prolactin (macroprolactin) by commercial immunoassays. J Clin Endocrinol Metab., 87(12), 5410-5. Smyth GK. (2004). Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol., 3(1), Article3. Soong YK, Ferguson KM, McGarrick G & Jeffcoate SL. (1982). Size heterogeneity of immunoreactive prolactin in hyperprolactinaemic serum. Clin Endocrinol (Oxf). 16(3), 259-65. Sotgia F, Rui H, Bonuccelli G, Mercier I, Pestell RG & Lisanti MP. (2006). Caveolin-1, mammary stem cells, and estrogen-dependent breast cancers. Cancer Res., 66(22), 10647-51. Srivastava S, Matsuda M, Hou Z, Bailey JP, Kitazawa R, Herbst MP & Horseman ND. (2003). Receptor activator of NF-kappaB ligand induction via Jak2 and Stat5a in mammary epithelial cells. J Biol Chem, 278(46), 46171-8. Sternlicht MD. (2006). Key stages in mammary gland development: the cues that regulate ductal branching morphogenesis. Breast Cancer Res., 8(1), 201. Stevenson BR, Siliciano JD, Mooseker MS & Goodenough DA. (1986). Identification of ZO-1: a high molecular weight polypeptide associated with the tight junction (zonula occludens) in a variety of epithelia. J Cell Biol., 103(3), 755-66. Stingl J, Eirew P, Ricketson I, Shackleton M, Vaillant F, Choi D, Li HI & Eaves CJ. (2006). Purification and unique properties of mammary epithelial stem cells. Nature., 439(7079), 993-7.
200 Streuli CH, Edwards GM, Delcommenne M, Whitelaw CB, Burdon TG, Schindler C & Watson CJ. (1995). Stat5 as a target for regulation by extracellular matrix. J Biol Chem, 270(37), 21639-44. Stricker P & Grueter R. (1928). Action du lobe anterieur de l'hypophyse sur la montee laiteuse. C. R. Soc. Biol., 99 1978-1980. Sutherland RL & Musgrove EA. (2004). Cyclins and breast cancer. J Mammary Gland Biol Neoplasia, 9(1), 95-104. Sutherland RL, Prall OW, Watts CK & Musgrove EA. (1998). Estrogen and progestin regulation of cell cycle progression. J Mammary Gland Biol Neoplasia, 3(1), 63- 72. Tanaka M, Hadjantonakis AK & Nagy A. (2001). Aggregation chimeras. Combining ES cells, diploid and tetraploid embryos. Methods Mol Biol., 158 135-54. Tanaka T, Shiu RP, Gout PW, Beer CT, Noble RL & Friesen HG. (1980). A new sensitive and specific bioassay for lactogenic hormones: measurement of prolactin and growth hormone in human serum. J Clin Endocrinol Metab., 51(5), 1058-63. Teglund S, McKay C, Schuetz E, van Deursen JM, Stravopodis D, Wang D, et al. (1998). Stat5a and Stat5b proteins have essential and nonessential, or redundant, roles in cytokine responses. Cell, 93(5), 841-50. Tessier C, Prigent-Tessier A, Ferguson-Gottschall S, Gu Y & Gibori G. (2001). PRL antiapoptotic effect in the rat decidua involves the PI3K/protein kinase B-mediated inhibition of caspase-3 activity. Endocrinology., 142(9), 4086-94. Teuliere J, Faraldo MM, Deugnier MA, Shtutman M, Ben-Ze'ev A, Thiery JP & Glukhova MA. (2005). Targeted activation of beta-catenin signaling in basal mammary epithelial cells affects mammary development and leads to hyperplasia. Development, 132(2), 267-77. Thomas RS, Ng AN, Zhou J, Tymms MJ, Doppler W & Kola I. (2000). The Elf group of Ets-related transcription factors. ELF3 and ELF5. Adv Exp Med Biol, 480 123-8. Touraine P, Martini JF, Zafrani B, Durand JC, Labaille F, Malet C, et al. (1998). Increased expression of prolactin receptor gene assessed by quantitative polymerase chain reaction in human breast tumors versus normal breast tissues. Journal of Clinical Endocrinology & Metabolism., 83(2), 667-74.
201 Traurig H. (1967). Cell Proliferation in the mammary gland during late pregnancy and lactation. Anatomical Record, 157 489-503. Tuazon PT, Lorenson MY, Walker AM & Traugh JA. (2002). p21-activated protein kinase gamma-PAK in pituitary secretory granules phosphorylates prolactin. FEBS Lett., 515(1-3), 84-8. Tworoger SS, Eliassen AH, Rosner B, Sluss P & Hankinson SE. (2004). Plasma prolactin concentrations and risk of postmenopausal breast cancer. Cancer Res, 64(18), 6814-9. van Genderen C, Okamura RM, Farinas I, Quo RG, Parslow TG, Bruhn L & Grosschedl R. (1994). Development of several organs that require inductive epithelial- mesenchymal interactions is impaired in LEF-1-deficient mice. Genes Dev., 8(22), 2691- 703. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, Sutton GG, et al. (2001). The sequence of the human genome. Science., 291(5507), 1304-51. Verger A & Duterque-Coquillaud M. (2002). When Ets transcription factors meet their partners. Bioessays., 24(4), 362-70. Villar-Palasi C & Larner J. (1970). Glycogen metabolism and glycolytic enzymes. Annu Rev Biochem., 39 639-72. Visvader JE & Lindeman GJ. (2003). Transcriptional regulators in mammary gland development and cancer. Int J Biochem Cell Biol., 35(7), 1034-51. Vomachka AJ, Pratt SL, Lockefeer JA & Horseman ND. (2000). Prolactin gene- disruption arrests mammary gland development and retards T-antigen-induced tumor growth. Oncogene., 19(8), 1077-84. Vonderhaar BK. (1989). Estrogens are not required for prolactin induced growth of MCF-7 human breast cancer cells. Cancer Letters., 47(1-2), 105-10. Vonderhaar BK. (1999). Prolactin involvement in breast cancer. Endocrine-Related Cancer., 6(3), 389-404. Wagner KU, Krempler A, Triplett AA, Qi Y, George NM, Zhu J & Rui H. (2004). Impaired alveologenesis and maintenance of secretory mammary epithelial cells in Jak2 conditional knockout mice. Mol Cell Biol., 24(12), 5510-20.
202 Wakao H, Gouilleux F & Groner B. (1995). Mammary gland factor (MGF) is a novel member of the cytokine regulated transcription factor gene family and confers the prolactin response. Embo J., 14(4), 854-5. Wakefield LM, Piek E & Bottinger EP. (2001). TGF-beta signaling in mammary gland development and tumorigenesis. J Mammary Gland Biol Neoplasia., 6(1), 67-82. Walden PD, Ruan W, Feldman M & Kleinberg DL. (1998). Evidence that the mammary fat pad mediates the action of growth hormone in mammary gland development. Endocrinology., 139(2), 659-62. Walker AM, Montgomery DW, Saraiya S, Ho TW, Garewal HS, Wilson J & Lorand L. (1995). Prolactin-immunoglobulin G complexes from human serum act as costimulatory ligands causing proliferation of malignant B lymphocytes. Proc Natl Acad Sci U S A., 92(8), 3278-82. Wang YF & Walker AM. (1993). Dephosphorylation of standard prolactin produces a more biologically active molecule: evidence for antagonism between nonphosphorylated and phosphorylated prolactin in the stimulation of Nb2 cell proliferation. Endocrinology., 133(5), 2156-60. Welch CL, Xia YR, Shechter I, Farese R, Mehrabian M, Mehdizadeh S, Warden CH & Lusis AJ. (1996). Genetic regulation of cholesterol homeostasis: chromosomal organization of candidate genes. J Lipid Res., 37(7), 1406-21. Welsch CW, Louks G, Fox D & Brooks C. (1975). Enhancement by prolactin of carcinogen induced mammary cancerigenesis in the male rat. British Journal of Cancer, 32(4), 427-31. Welsch CW & Nagasawa H. (1977). Prolactin and murine mammary tumorigenesis: a review. Cancer Research., 37(4), 951-63. Wennbo H, Gebre-Medhin M, Gritli-Linde A, Ohlsson C, Isaksson OG & Tornell J. (1997). Activation of the prolactin receptor but not the growth hormone receptor is important for induction of mammary tumors in transgenic mice. J Clin Invest, 100(11), 2744-51. Wennbo H & Tornell J. (2000). The role of prolactin and growth hormone in breast cancer. Oncogene, 19(8), 1072-6.
203 Wicks JR & Brooks CL. (1995). Biological activity of phosphorylated and dephosphorylated bovine prolactin. Molecular & Cellular Endocrinology., 112(2), 223-9. Wiesen JF, Young P, Werb Z & Cunha GR. (1999). Signaling through the stromal epidermal growth factor receptor is necessary for mammary ductal development. Development., 126(2), 335-44. Williams JM & Daniel CW. (1983). Mammary ductal elongation: differentiation of myoepithelium and basal lamina during branching morphogenesis. Dev Biol., 97(2), 274- 90. Williams TM, Lee H, Cheung MW, Cohen AW, Razani B, Iyengar P, Scherer PE, Pestell RG & Lisanti MP. (2004). Combined loss of INK4a and caveolin-1 synergistically enhances cell proliferation and oncogene-induced tumorigenesis: role of INK4a/CAV-1 in mammary epithelial cell hyperplasia. J Biol Chem., 279(23), 24745-56. Epub 2004 Mar 25. Williams TM & Lisanti MP. (2005). Caveolin-1 in oncogenic transformation, cancer, and metastasis. Am J Physiol Cell Physiol., 288(3), C494-506. Williams TM, Sotgia F, Lee H, Hassan G, Di Vizio D, Bonuccelli G, et al. (2006). Stromal and epithelial caveolin-1 both confer a protective effect against mammary hyperplasia and tumorigenesis: caveolin-1 antagonizes cyclin d1 function in mammary epithelial cells. Am J Pathol., 169(5), 1784-801. Wu W, Chen YH, Ueda E, Tan D, Bartolini P & Walker AM. (2006). Different forms of prolactin have opposing effects on the expression of cell cycle regulatory proteins in differentiated mammary epithelial cells. Oncol Res., 16(2), 75-84. Wu W, Coss D, Lorenson MY, Kuo CB, Xu X & Walker AM. (2003). Different biological effects of unmodified prolactin and a molecular mimic of phosphorylated prolactin involve different signaling pathways. Biochemistry., 42(24), 7561-70. Wynick D, Hammond PJ, Akinsanya KO & Bloom SR. (1993). Galanin regulates basal and oestrogen-stimulated lactotroph function. Nature, 364(6437), 529-32. Wynick D, Small CJ, Bacon A, Holmes FE, Norman M, Ormandy CJ, et al. (1998). Galanin regulates prolactin release and lactotroph proliferation. Proc Natl Acad Sci U S A., 95(21), 12671-6. Wysolmerski JJ, Philbrick WM, Dunbar ME, Lanske B, Kronenberg H & Broadus AE. (1998). Rescue of the parathyroid hormone-related protein knockout mouse
204 demonstrates that parathyroid hormone-related protein is essential for mammary gland development. Development., 125(7), 1285-94. Yager JD & Davidson NE. (2006). Estrogen carcinogenesis in breast cancer. NEngl J Med., 354(3), 270-82. Yang H & Churchill G. (2006). Estimating p-values in small microarray experiments. Bioinformatics, 30 30. Yang Y, Spitzer E, Meyer D, Sachs M, Niemann C, Hartmann G, Weidner KM, Birchmeier C & Birchmeier W. (1995). Sequential requirement of hepatocyte growth factor and neuregulin in the morphogenesis and differentiation of the mammary gland. J Cell Biol., 131(1), 215-26. Yaniw D & Hu J. (2005). Epithelium-specific ets transcription factor 2 upregulates cytokeratin 18 expression in pulmonary epithelial cells through an interaction with cytokeratin 18 intron 1. Cell Res., 15(6), 423-9. Yu TX & Rillema JA. (1998). The MEK inhibitor PD 098059 inhibits prolactin- induced Nb2 cell mitogenesis but not milk product synthesis in cultured mouse mammary tissues. Biochimica et Biophysica Acta., 1448(1), 126-34. Zafonte BT, Hulit J, Amanatullah DF, Albanese C, Wang C, Rosen E, et al. (2000). Cell-cycle dysregulation in breast cancer: breast cancer therapies targeting the cell cycle. Front Biosci., 5 D938-61. Zhou J, Chehab R, Tkalcevic J, Naylor MJ, Harris J, Wilson TJ, et al. (2005). Elf5 is essential for early embryogenesis and mammary gland development during pregnancy and lactation. Embo J, 24(3), 635-44. Zhou J, Ng AY, Tymms MJ, Jermiin LS, Seth AK, Thomas RS & Kola I. (1998). A novel transcription factor, ELF5, belongs to the ELF subfamily of ETS genes and maps to human chromosome 11p13-15, a region subject to LOH and rearrangement in human carcinoma cell lines. Oncogene, 17(21), 2719-32. Zhou Z, Yon Toh S, Chen Z, Guo K, Ng CP, Ponniah S, Lin SC, Hong W & Li P. (2003). Cidea-deficient mice have lean phenotype and are resistant to obesity. Nat Genet., 35(1), 49-56.
205 Zutter MM, Sun H & Santoro SA. (1998). Altered integrin expression and the malignant phenotype: the contribution of multiple integrated integrin receptors. J Mammary Gland Biol Neoplasia., 3(2), 191-200.
206 Appendices
207 0013-7227/03/$15.00/0 Endocrinology 144(7):3196–3205 Printed in U.S.A. Copyright © 2003 by The Endocrine Society doi: 10.1210/en.2003-0068
Prostate Development and Carcinogenesis in Prolactin Receptor Knockout Mice
FIONA G. ROBERTSON, JESSICA HARRIS, MATTHEW J. NAYLOR, SAMANTHA R. OAKES, JON KINDBLOM, KARIN DILLNER, HÅKAN WENNBO, JAN TO¨ RNELL, PAUL A. KELLY, JEFF GREEN, AND CHRISTOPHER J. ORMANDY Cancer Research Program (F.G.R., J.H., M.J.N., S.R.O., C.J.O.), Garvan Institute of Medical Research, St. Vincent’s Hospital, Darlinghurst, 2010 Sydney, Australia; Department of Physiology (J.K., K.D., H.W., J.T.), Research Center for Endocrinology and Metabolism, Go¨teborg University, S-405 30 Go¨teborg, Sweden; Institut National de la Sante´etdela Recherche Me´dicale Unite´584 (P.A.K.), Faculte´deMe´decine Necker-Enfants Malades, 75730 Paris, France; and Laboratory of Cell Regulation and Carcinogenesis (J.G.), National Cancer Institute, Bethesda, Maryland 20892-1402
Hyperprolactinemia results in prostatic hypertrophy and hy- not the dorsal lobe, and no tumors were seen in 20 prolactin perplasia, but it is not known whether prolactin plays an receptor knockout animals, compared with 1 of 11 detected in essential role in these processes in the prostate. To address wild-type and 4 of 21 found in heterozygous animals. Oligo- this question, we investigated prostate development, gene ex- nucleotide microarrays were used to identify essential tran- pression, and simian virus 40 (SV40)T-induced prostate car- scriptional roles of prolactin and revealed a small set of genes cinogenesis in prolactin receptor knockout mice. These ani- with decreased expression involved in sperm/oocyte interac- mals showed a small increase in dorsolateral and ventral tion and copulatory plug formation. Infertility or reduced prostate weight but no change in the weight of the anterior fertility was apparent in these animals. These findings estab- prostate. The dorsal but not ventral or lateral lobes showed a lish essential though subtle roles for prolactin in the regula- 12% loss of epithelial cells; all other morphological parameters tion of prostate morphology, gene expression, SV40T-induced were normal. The area of SV40T-induced prostate intraepi- neoplasia, and reproductive function. (Endocrinology 144: thelial neoplasia was reduced by 28% in the ventral lobe but 3196–3205, 2003)
HE CRITICAL ROLE of androgens in the development and rodent prostate (9) by the epithelial cells with a weak T and maintenance of the prostate is demonstrated by the signal in the fibromuscular stroma (10). In vitro prolactin is dramatic prostate regression that follows castration (1). Early mitogenic for cultured prostate epithelial cells (11), and in work showed that other endocrine factors are also involved organ cultures normal morphology was best maintained by in prostate development, particularly those of the pituitary. the addition of androgen and prolactin (12). Expression in the Greater prostate regression occurs with additional hypoph- prostate of both prolactin and the prolactin receptor is in- ysectomy (2), androgen replacement after castration restores creased by androgen treatment in vivo, and prolactin receptor prostate weight less effectively with hypophysectomy (3, 4), level is also increased by prolactin (9, 13). Prolactin has been and pituitary grafts reduce the rate and extent of prostate viewed as an autocrine/paracrine growth factor (9, 14) or a regression induced by castration (5) via androgen-indepen- survival factor (15) for prostate epithelial cells. dent mechanisms (6). Prolactin may be the pituitary factor Prolactin may also play a role in disease of the prostate, responsible for these effects because studies to date have again both indirectly and directly. Human prostate hyper- shown that treatment of intact and castrated animals with plasia and early cancer continue to express the prolactin prolactin is growth stimulatory for the normal prostate, both receptor (9, 16). Prolactin alone and in synergy with testos- in synergy and independently of androgen. Prolactin is ca- terone causes cell proliferation in benign hyperplastic human pable of producing stromal hyperplasia and intraepithelial prostate (17) and human prostate cancer cell lines (18). Clin- dysplastic features with long exposure (7). ical investigation has shown that serum prolactin levels in The mechanism of prolactin action on the prostate is com- plex. Prolactin can influence the prostate indirectly via reg- men with prostate disease are generally not different from ulation of the level of testicular luteinizing hormone recep- age-matched controls when measured at presentation (19) or tors and steroidogenic enzymes to increase testicular up to 13 yr before diagnosis (20). During estrogen, anties- testosterone production. In humans (but not rodents) the trogen, antiandrogen, or GnRH analog therapy for prostate adrenal glands produce androgen precursors DHEA and cancer, prolactin levels can increase and are predictive of dehydroepiandrosterone sulfate in large quantities in re- poor prognosis (21). This has led to 11 reported small un- sponse to prolactin (8). Prolactin can also operate directly; controlled clinical trials of adjuvant bromocriptine during prolactin and the prolactin receptor are expressed in human antiandrogen therapy, some of which report increased re- sponse rates when bromocriptine is used (22), despite the enrollment of late-stage patients in whom tumor prolactin Abbreviations: H&E, Hematoxylin-eosin; PIN, prostatic intraepithe- ϩ ϩ Ϫ Ϫ lial neoplasia; PRLR / , wild-type mice; PRLR / , prolactin receptor receptor levels can be diminished or lost (23). In addition, a knockout mice; SV40, simian virus 40. recent study reported the use of a molecular mimic of phos-
3196 Robertson et al. • Prolactin and Prostate Endocrinology, July 2003, 144(7):3196–3205 3197
phorylated prolactin in vivo to successfully inhibit tumor fymetrix U74A/U74Av2), and analysis (Affymetrix Inc., Santa Clara, initiation and the growth of DU145-derived tumors in nude CA; MicroArray Suite 4 and 5) according to the manufacturer’s instruc- mice (24). tions. The experiment was replicated three (ventral) times and two (dorsal) times using a different set of animals for each replication. Iden- Together this body of work suggests a role for prolactin in tified were genes that showed consistent change in expression level the growth of the normal and cancerous prostate; however, across the replicates, the significance of which was tested using paired a number of fundamental questions remain unanswered. t test. Alterations were confirmed by quantitative PCR using the Light- Chiefly, does prolactin at physiological levels have an es- Cycler (Roche, Basel, Switzerland) and TaqMan (Applied Biosystems, Foster City, CA) instruments according to the manufacturer’s protocols. sential role in prostate development and function, or is pro- DChip software (29) was a gift of Dr. Wong, and the JMP statistical lactin active only during the abnormal hyperprolactinemic software used for principal component analysis was from SAS Institute states produced by either pathophysiological conditions or (Cary, NC). The BioNavigator platform was used for bulk blastx experimental or therapeutic manipulation? We have used searches of SwissProtϩTrEMBL and blastn searches of GenBank, Uni- histomorphological and transcript profiling techniques to Gene, and HomoloGene were also used. Medline was used to identify ϩ ϩ publications relating to gene function. compare wild-type (PRLR / ) prostates with the prostates Ϫ/Ϫ from prolactin receptor knockout mice (PRLR ). To search Statistical analysis for essential roles for prolactin in prostate carcinogenesis, we Ϫ Ϫ produced PRLR / mice that carry the simian virus 40 Comparisons were made using an unpaired, two-tailed t test; Kaplan- Meier survival analysis; and simple regression using Statview 4.5. Array (SV40)T oncogene under the C3 promoter. data were sorted and graphically analyzed using Excel (Microsoft Corp., Mountain View, CA) and the JMP statistical package. Materials and Methods Mice Quantitative PCR All animals were rederived by embryo transfer and SPF barrier RNA (1 g) from the prostates was reverse transcribed using AMV housed. All animal use was approved and supervised by the Garvan reverse transcriptase (Promega Corp., Madison, WI). PCR primers were Institute/St. Vincent’s Hospital Animal Experimentation Ethics Com- designed using MacVector to span an intron. The PCR were performed Ϫ Ϫ mittee. The PRLR / mouse was generated by replacement of exon 5 of in a LightCycler (Roche) using 1 l cDNA diluted 1:2, 5 pmol primers, the PRLR gene, which encodes cysteine residues essential for ligand and the FastStart DNA master SYBR Green I enzyme mix (Roche) in a binding and receptor activation, with the NEO cassette driven by the Tk 10- l reaction volume. Relative quantification of the product was per- Ϫ Ϫ promoter in the opposite direction to PRLR transcription (25). PRLR / formed by comparing the crossing points of different samples above ϩ ϩ and PRLR / mice were derived from E14 ES cells (129/OlaHsd) bred background. to 129/Sv Pas mice and then subsequently intercrossed. The core colony was maintained by heterozygous matings to produce all genotypes. Results Castration surgery was via the abdominal route. The C3-SV40T animals Gross prostate morphology (26, 27) were on an inbred FVB/N background, and the core colony was maintained by homozygous matings. The required mixed genotypes Ϫ/Ϫ Ϫ/Ϫ In the PRLR animals, the seminal vesicles, coagulating were produced by mating 10 PRLR males with 10 homozygous gland (anterior prostate), ventral prostate, dorsal prostate, C3-SV40T females and then mating the resultant females with the Ϫ Ϫ PRLR / males to produce animals that were heterozygous or wild type and lateral prostate were all present and of normal size and ϩ Ϫ Ϫ Ϫ ϩ ϩ for C3-SV40T and PRLR / or PRLR / . Control PRLR / animals appearance, at both 2 wk of age (not shown) and in mature ϩ ϩ were produced by the use of 10 PRLR / in an identical but separate animals of 16–36 wk of age (Fig. 1, top panel). The lobes were scheme to ensure similar genetic diversity between groups. Animals dissected and weighed. The seminal vesicle and ventral pros- were aged to 50 wk before prostate collection. Significant numbers could Ϫ/Ϫ Ͻ not be aged past this point because of the onset of tumors in the sub- tate were 25% heavier in PRLR animals (P 0.0001), maxillary glands and sebaceous glands of the paws. corrected for body weight or not. The dorsolateral lobes, dissected as one for this measurement, were 10% heavier Microdissection and analysis of ductal morphogenesis with borderline (P ϭ 0.07) statistical significance. In contrast, the coagulating gland (anterior prostate) showed no differ- The genital tract was removed en bloc to a Petri dish and cleared of ϭ periprostatic fat and connective tissue, then anterior prostate and sem- ence in weight (P 0.57). The weights of these organs were inal vesicles. A section of the urethra containing the connections of the also examined as a function of age, and an identical pattern prostate ducts was then removed and photographed. Individual lobes was observed for the seminal vesicle and the ventral and were removed by cutting the ducts at their urethral connection, and dorsolateral prostates. Young animals (10–11 wk) showed no microdissected in depression slides as previously described (28). weight differences, but mature animals (16–30 wk) showed a highly significant (P Ͻ 0.001) increase in weight in the Histology and morphometric analysis Ϫ Ϫ PRLR / animals. As the animals aged, the strength of the Hematoxylin-eosin (H&E)-stained sections were taken from through- P values for this weight difference weakened and became ϫ out the prostate and photographed digitally at 10 magnification. Ge- nonsignificant around 1 yr of age. The coagulating gland notype information was not recorded so that the subsequent analysis was blind to genotype. Tissue areas were manually defined and areas showed no difference in weight at any age (data not shown). measured using calibrated software. Epithelial cell nuclei in defined Thus, prolactin plays an essential role in maintaining correct areas were manually counted using images captured at ϫ20 magnifi- prostate weight during sexually mature life, but its effect is cation. Areas of prostatic intraepithelial neoplasia (PIN) were identified lost in old animals. by their hyperchromatic, enlarged, and elongated nuclei (27). We did not attempt to split the PIN lesions into low and high grades. Ductal morphogenesis Transcript profiling and data analysis Microdissection was used to examine branching morpho- RNA from five to eight prostate lobes of each genotype was pooled genesis of individual prostate lobes and quantified by count- in equimolar ratios before probe preparation, chip hybridization (Af- ing urethral ducts and ductal tips and branch points (Fig. 1, 3198 Endocrinology, July 2003, 144(7):3196–3205 Robertson et al. • Prolactin and Prostate
lower panels). This analysis clearly defined the distinctive ductal branching patterns of each lobe. The ventral and lat- eral prostate lobes were attached to the urethra by two or three main ducts that showed extensive oak tree branching morphology, and the dorsal prostate consisted of many ducts attached to the urethra that showed less extensive palm tree branching morphology (28). There was no difference be- ϩ ϩ Ϫ Ϫ tween PRLR / and PRLR / animals in the number of ducts, number of branch points per duct, or number of ductal tips present in each lobe. The more simply branched coag- ulating gland also showed no differences between genotypes (data not shown).
Histology and quantitative morphometric analysis H&E-stained sections (Fig. 2) were compared by measure- ment of tissue areas using manual tracing of epithelium, peri- ductal stroma (including the duct basement membrane), inter- ductal stroma, and lumen using calibrated image analysis software. Results are presented as pie charts (Fig. 2). Cell nuclei in these areas were also counted manually. The sensitivity of this technique is demonstrated by its ability to unequivocally distinguish the distinctive morphology of the various prostate lobes, in which the differences in epithelial, lumen, and peri- ductal stromal areas showed P values less than 0.0001. ϩ ϩ Ϫ Ϫ In PRLR / and PRLR / animals, there were no differ- ences in tissue areas of the ventral prostate. A small but significant difference in the dorsal prostate was detected. Epithelial and interductal stromal area decreased and lume- nal area increased. When the total ductal cross-sectional area was calculated by adding the areas of lumen, epithelium, and periductal stroma, no difference was seen between geno- types, indicating that the changes were not due to an increase in duct size. Epithelial content was examined by a separate analysis of the relative lumen and epithelial areas of 200 ϩ ϩ ductal cross-sections of each genotype. PRLR / epithelium occupied 67.1% of the ductal cross-sectional area and the Ϫ Ϫ PRLR / epithelium occupied 58.7% (P Ͻ 0.0001), a loss of 12.5% of the epithelial content. Epithelial nuclei were counted and the area they occupied measured in a further 50 random ductal cross-sections; no difference in epithelial cell density was seen, demonstrating that no change in cell size had occurred. Thus, epithelial cells were lost.
Role of androgens Castration was used to examine the role androgens play in Ϫ Ϫ the altered tissue ratios observed in the dorsal lobe of PRLR / animals (Fig. 2). The loss of PRLR has no effect on testosterone levels (30). Castration resulted in the loss of epithelium and a gain in lumenal area that was accompanied by an overall in- FIG. 1. Prostate morphology. The ventral (V), lateral (L), and dorsal (D) lobes of the prostate were dissected free of periprostatic fat and crease in the area of the stroma. This was due to an increase in connective tissue but left attached to a small section of the urethra periductal stroma despite the loss of interductal stroma so that (U). Seminal vesicles (SV) and coagulating gland (CG) were removed the most visually striking change was the ratio of epithelium to and are not shown because of their much larger size. Lobes are from periductal stroma. In the ventral lobe, these changes were iden- Ϫ/Ϫ ϩ ϩ Ϫ Ϫ mature animals (16–20 wk). The gross morphology of PRLR pros- tical between PRLR / and PRLR / animals. In the dorsal tates are indistinguishable from controls. Lobe weights (milligrams Ϯ SE) for the ventral (VP) and combined dorsal-lateral (DLP) prostate lobe, however, significant differences were detected between are indicated, and P values (t test) for comparison between PRLRϩ/ϩ and PRLRϪ/Ϫ are shown. Individual lobes from 16- to 36-wk animals were microdissected further in collagenase under a dissecting micro- branching patterns. Quantification of ducts (D) attached to the ure- scope to desegregate the interductal stoma, allowing the ducts to be thra, ductal tips (T), and duct branch points (B) are given with number disentangled and arranged in two dimensions, revealing ductal of lobes (n) analyzed and P value from t test vs. control. Robertson et al. • Prolactin and Prostate Endocrinology, July 2003, 144(7):3196–3205 3199
that placed the SV40T antigen under the control of the C3 promoter. Animals were intercrossed using a scheme de- signed to ensure equivalent background genetic diversity of the test groups (see Materials and Methods). Animals were aged to 50 wk and then killed and their prostates analyzed by H&E-stained serial sections for PIN and tumors. The area of PIN was measured using manual tracing of H&E images in ImageQuant software (Leica Corp.) using the descriptors of PIN used by Shibata et al. (27) (Fig. 3, top two panels). As described by Shibata et al., the amount of PIN was higher in ϩ ϩ ϩ ϩ the PRLR / ventral lobe than the PRLR / dorsal lobe (Fig. 3, panel 3). Loss of PRLR resulted in a 28% decrease (P ϭ 0.026) in the area of PIN in the ventral lobe but was without detectable effect in the dorsal lobe. The level of SV40T antigen expression in the ventral lobe was measure by quantitative RT-PCR (Fig. 3, bottom panel). There was no difference (P ϭ 0.21) between PRLR genotypes in SV40T expression level, consistent with identical androgen levels in these animals and indicating that the C3 promoter fragment used is insen- sitive to prolactin signaling. Only five tumors were detected in the cohort and all occurred in the ventral lobe. One tumor ϩ ϩ was found in the 11 PRLR / animals examined and four in ϩ Ϫ Ϫ Ϫ 21 PRLR / animals. None of the 20 PRLR / animals showed a detectable prostate tumor.
FIG. 2. Prostate lobe histology. Representative histology for each lobe and genotype from intact 16- to 36-wk-old animals is shown. Tissue ratios were measured by tissue area measurements using image anal- ysis of H&E-stained histology. Ratios are presented as pie charts from intact mature animals or mature animals 21 d following castration (histology not shown, pie charts denoted C). Measurements (n ϭ 4–5 animals per group) were compared with wild-type control or castrated wild-type control using t test. *, P Ͻ 0.05; **, P Ͻ 0.005; ***, P Ͻ 0.0005; ****, P Ͻ 0.0001.
ϩ ϩ Ϫ Ϫ PRLR / and PRLR / animals. Importantly, the difference in epithelial area seen in intact animals was lost in castrated an- imals. In addition, the amount of lumenal area showed a greater decrease, and the area of stroma (especially periductal stroma) Ϫ/Ϫ showed a greater increase in PRLR animals. These mea- FIG. 3. Formation of PIN lesions in the prostate by expression of surements demonstrate that castration-induced regression of SV40T. The upper two panels show examples of PIN lesions (p) in the Ϫ Ϫ ϩ ϩ the dorsal lobe was greater in PRLR / animals than PRLR / dorsal and ventral lobe induced by expression of SV40T. The area of these lesions was measured in 50-wk-old animals and expressed as a controls. Prolactin and androgen cooperate to regulate dorsal percentage of total tissue area. No difference between PRLR geno- lobe composition. types was detected in the dorsal lobe; however, a significant difference occurred in the ventral lobe. Levels of SV40T expression were mea- Carcinogenesis sured in the ventral lobe by quantitative RT-PCR and are shown for individual animals of the indicated genotypes, expressed as the cross- We crossed the mouse line carrying the null mutation of ing point for each corrected for a -actin control reaction. A t test was the PRLR with the mouse line carrying a transgenic insert used to test for a significant difference, which was not found. 3200 Endocrinology, July 2003, 144(7):3196–3205 Robertson et al. • Prolactin and Prostate
Transcript profiling time (20-d gestation ϩ 4 d estrus), compared with less than ϩ/ϩ ϩ/Ϫ To determine whether loss of the PRLR resulted in altered 20% of either PRLR or PRLR matings. Analysis using a Cox proportional hazards model indicated the chance of prostate gene expression, we used the Affymetrix GeneChip Ϫ Ϫ pregnancy for PRLR / matings to be 40.9% (P ϭ 0.0005) of system to measure the expression of approximately 12,000 ϩ/ϩ genes. The ventral lobe showed no alteration in relative tissue PRLR animals. The latency to second pregnancy (Fig. 6B) areas and the dorsal lobe a 12% loss of epithelium (Fig. 2). was indistinguishable among PRLR genotypes, as were the Thus, altered tissue ratios will have no effect on ventral gene latencies for third, fourth, and fifth pregnancies (data not expression patterns and a small effect on dorsal gene ex- shown). pression patterns. The experiment was replicated three times First-pregnancy latency was compared by regression anal- ysis to second-pregnancy latency to determine whether a for the ventral prostate. Analysis showed that duplication Ϫ/Ϫ resulted in a very large decrease in false positives and a subpopulation of low fertility PRLR males persisted (Fig. 6C). No correlation was found (R2 ϭ 0.003). To determine further replication did not dramatically improve the false Ϫ/Ϫ positive rate. Consequentially, the dorsal lobe was analyzed whether reduced fertility in PRLR males was a function in duplicate. RNA was pooled from five to eight different of delayed onset of full fertility, we correlated the age of the animals per replicate. Results were analyzed by GeneChip male when first housed with a female to first pregnancy 2 ϭ 3.1 and 5.1 software, and the sources of variation in gene latency (Fig. 6D), and again no relationship was found (R Ϫ/Ϫ Ϯ expression levels were explored using principal components 0.00005). Fertile PRLR males produced 3.3 1.8 litters with 5.8 Ϯ 0.3 pups, compared with 4.0 Ϯ 1.9 litters (P ϭ analysis (Fig. 4). The greatest cause of variance in the data set ϩ Ϫ 0.190) and 4.8 Ϯ 0.5 pups (P ϭ 0.222) for PRLR / males. was the relative abundance of different genes, for example, Ϫ/Ϫ the level of one of the most abundant genes such as ubiquitin, Thus, the ability of PRLR males to produce a first preg- compared with the level of one of the least abundant genes nancy is just 40% of wild-type animals but is 100% for sub- such as transcription factor NF-ATca. Examination of the sequent pregnancies, with no persistent subpopulation of next two (dorsal) or three (ventral) principal components poorly fertile males and no effect of age on fertility. showed that the experimental replication provided the next most influential source of variance in gene expression. Ex- Discussion amination of the higher principal components identified the third most influential factor as PRLR genotype (Fig. 4), re- In the experiments detailed in this report, we have used vealing a small cluster of genes differentially expressed be- genetically modified mice to search for essential roles of tween genotypes. prolactin in the prostate, using morphological, gene tran- Functional annotation of these differentially expressed script, and functional end points. Detailed morphological genes (Fig. 4, lower panels) revealed that the common genes investigations showed that a loss of prolactin receptor caused comprising these clusters are principally involved in sperm/ a 20% increase in prostate weight, a loss of some epithelial area from the dorsal lobe, but no alteration in branching oocyte interaction or copulatory plug formation (see Discus- Ϫ/Ϫ sion for details). These results were confirmed by quantitative morphology. Castration of PRLR animals showed that PCR using TaqMan and LightCycler methodologies by two castration-induced regression is more extreme in dorsal lobes independent laboratories (Sydney and Go¨teborg) using dif- without prolactin receptor, indicating independent mecha- ferent sets of mouse tissues. Figure 5 demonstrates the nisms of complementary action of androgen and prolactin. method used on the LightCycler instrument and compares These data show that prolactin plays a subtle but essential the results with those obtained by 30 cycles of RT-PCR (Fig. role in the control of normal dorsal prostate morphology. In 5A) and from the Affymetrix chips for the gene AEG-1 (Fig. contrast, the effects of hyperprolactinemia are pronounced. 5C). All methods show diminished AEG-1 expression. The A transgenic model of hyperprolactinemia showed a 10- to results for the mRNA levels of other genes obtained in this 20-fold increase in prostate weight, 5-fold increase in cell way are given in Fig. 5D. content, altered relative amounts of epithelium to stroma, and increased cellularity of the interductal stroma. As these animals aged, intraepithelial dysplastic features became ap- Fertility parent (7). A recent transgenic model using prostate-specific In the light of the transcript profiling results, we examined prolactin expression showed a similar prostate phenotype Ϫ Ϫ male PRLR / fertility in the breeding colony held at the (31) in the absence of systemic hormonal effects, conclusively ϩ ϩ ϩ Ϫ Garvan Institute. PRLR / (n ϭ 31), PRLR / (n ϭ 28), and demonstrating the direct effects of hyperprolactinemia on Ϫ Ϫ PRLR / (n ϭ 52) males were first joined monogamously the prostate. These results indicate that many previous ex- and continuously at ages between 40 and 120 d. The rates of perimental investigations, using the general paradigm of ϩ ϩ infertility, defined as no litters in 120 d, were PRLR / 1of prolactin endocrine ablation with prolactin readdition, have ϩ Ϫ Ϫ Ϫ 31, PRLR / 0 of 28, and PRLR / 5 of 52, confirming an measured the effects of hyperprolactinemia but attributed Ϫ Ϫ increased rate of total infertility in the PRLR / genotype, as them to an essential role of prolactin. Our results show the previously reported (25). essential morphological roles of prolactin contrast to those Ϫ Ϫ It is not known whether PRLR / males capable of pro- produced by hyperprolactinemia, limiting the essential ac- ducing litters are fully fertile. Latency to pregnancy was tions to maintenance of the dorsal epithelium and fine reg- examined using Kaplan-Meier survival analysis (Fig. 6A). ulation of secretion. Latency to first litter was significantly delayed (log rank P ϭ These observed effects vary a little from those observed in Ϫ Ϫ 0.0007) in PRLR / matings, with 50% outside the minimal other genetically modified models of prolactin action. Loss of Robertson et al. • Prolactin and Prostate Endocrinology, July 2003, 144(7):3196–3205 3201
ϩ ϩ Ϫ Ϫ FIG. 4. Transcript profiles of ventral prostate. RNA from five to eight PRLR / or PRLR / ventral or dorsal prostates was pooled in equal ratios and labeled before hybridization to Affymetrix U74A or U74A1 chips. The experiment was replicated three times for the ventral lobe and twice for the dorsal lobe. Analysis was undertaken using Affymetrix GeneChip 3.1/5.1 and principal component analysis. The original axis positions are indicated, light gray behind the plane of the page, black in front. Genes called present in at least one of three PRLRϩ/ϩ replicates but without a change in expression level are shown as black squares. Genes called decreasing in two of three replicates (ventral A) or two of two replicates (dorsal, B) are red diamonds. The only two genes called increasing are green diamonds. These genes are labeled in the lower panels (ventral C, dorsal D) in which their average fold change across all replicates is greater than 2. x-axis, P.C. 6; y-axis, P.C. 5; z-axis P.C. 4. All plots are mean centered.
Ϫ Ϫ Ϫ Ϫ Stat5a results in prostate hypersecretion but with disorga- mon between PRLR / and Stat5a / animals and suggests Ϫ Ϫ nization of the ventral epithelium not seen in PRLR / an- Stat5a as the mediator of prolactin’s control of secretion, but imals (32). Stat5a forms one branch of the prolactin receptor effects on ventral prostate organization may be caused by signaling pathway but also transduces signals from other epidermal growth factor or other activators of Stat5. Loss of receptors such as the epidermal growth factor receptor. Thus, the prolactin gene resulted in a reduction in prostate weight only a subset of phenotypic abnormalities should be com- (33), not a gain in weight as would be expected from our 3202 Endocrinology, July 2003, 144(7):3196–3205 Robertson et al. • Prolactin and Prostate
results. The rodent prostate begins development at d 17 of gestation (1) and has complete architecture (but not size or ductal branching) at 2 wk of age. During this period prolactin knockout animals receive maternal prolactin via the mater- nal circulation, placental lactogen (a prolactin receptor li- gand) from the embryonic placenta, and prolactin from milk until weaning at 3 wk. Thus, the prolactin knockout is not prolactin deficient until after basic prostate development is complete, and this may be responsible for the contrasting observations made in prolactin and prolactin receptor knock- outs (34). The introduction of the SV40T antigen demonstrated that loss of the PRLR produced a small but significant reduction in the level of PIN in the ventral but not dorsal lobe. In our model the ventral lobe showed higher levels of PIN than the dorsal lobe, as does the original C3-SV40T model. Ventral ϩ Ϫ ϩ ϩ lobe tumor incidence was 16% (PRLR / and PRLR / com- bined), compared with the original model of around 40%, and the dorsal lobes did not produce any tumors, indicating that the alteration in genetic background has reduced the tumor incidence from levels reported in the original model. Ϫ Ϫ No prostate tumors were detected in PRLR / animals, but only five tumors were detected in other genotypes. Although these numbers are not sufficient to draw conclusions regard- ing the role of prolactin in tumor formation, the observation is consistent with the reduction in PIN seen in the ventral Ϫ Ϫ lobes of PRLR / animals. To determine whether prolactin was exerting essential effects on gene expression that were not expressed as altered growth or morphology, we used high-density oligonucleo- tide arrays to transcript profile the ventral and dorsal pros- ϩ ϩ Ϫ Ϫ tates from PRLR / and PRLR / males. The ventral lobe showed no changes in ductal branching or relative tissue areas, allowing altered gene expression to be attributed solely to transcriptional regulation. The dorsal lobe showed a 12% loss of epithelium, which must be taken into account when analyzing the resulting expression profiles. The use of replicates from pooled samples allowed a small number of consistently altered genes to be identified. Analysis of the ventral data showed that the third replication offered only small additional discriminatory power, compared with the major effect of duplication, and so the dorsal lobe was pro- filed in duplicate only. Literature searches showed the most differentially regu-
FIG. 5. Confirmation of chip data by quantitative PCR. Gene expres- sion changes detected by the oligonucleotide chips were examined using quantitative PCR. The method used is demonstrated for AEG-1. A, RT-PCR for AEG-1 is analyzed after 30 cycles by ethidium bromide staining of an agarose gel. B, Quantitative PCR using the LightCycler instrument. Upper panel shows increase in the concentration of the AEG-1 amplicon with cycle number for samples derived from PRLRϩ/ϩ (purple lines) or PRLRϪ/Ϫ (red lines) ventral or dorsal pros- tates. Amplification of standard curve samples is shown using dashed blue lines, and the standard curve is plotted in the lower panel. A -actin standard curve is used to control the reverse transcription step and is not shown. C, Quantitative PCR results expressed as AEG-1 mRNA copies per microliter in the original samples, compared with their AEG-1 signal intensity measured by the oligonucleotide chip. Fold change is indicated for the bars linked by lines. D, Results using quantitative PCR for a subset of functionally related genes with reduced expression in PRLRϪ/Ϫ prostate. Robertson et al. • Prolactin and Prostate Endocrinology, July 2003, 144(7):3196–3205 3203
lated genes to be involved in fertility. Two subgroups were apparent. The first involves genes mediating sperm-egg in- teraction; acidic epididymal glycoprotein, HE5, and zonad- hesin. Acidic epididymal glycoprotein is also known as AEG1 or protein DE and is the product of the CRISP-1 gene. This androgen-regulated glycoprotein is secreted by the ep- ithelium of the epididymis and associates with the principal piece of the sperm tail, the sperm middle piece, and the postacrosomal region of the sperm head during sperm mat- uration. It is thought to be involved in sperm-oocyte plasma membrane fusion (35–37). HE5 is a glycoprotein that is also inserted onto maturing spermatozoa and shows variable sur- face presentation during sperm capacitation (38). Zonadhe- sin functions as a sperm-zona pellucida-binding protein, conferring species specificity. It is synthesized by primary spermatocytes and located to the anterior acrosome (39). The second subgroup contains the antiinflammatory and proco- agulant seminal vesicle proteins, which function in copula- tory plug formation and protection of sperm from female immune response; SVP2/SV-IV, semenoclotin and seminal vesicle F protein/SVP5 (40–42). Expression of the prolactin receptor was also reduced, as expected from the targeting strategy employed (25). Small proline-rich protein 2A was detected by two independent probe sets, and small proline-rich proteins 2B and 2E were also found in the cluster revealed by the higher principal components but were not detected by GeneChip 3.1 or 5.1. These proteins are expressed in stratified squamous epithelia as components of the cornified cell envelope providing an epidermal barrier function (43). Their function in prostate, if any, is unknown. Quantitative PCR using oligos shown by blast searches to be specific for 2A demonstrated that 2A expression did not change. The probe sets present on the chip for 2A, 2B, and 2E are directed at regions of these genes that show high sequence homology among family members. Thus, the GeneChip result may be due to the integration of signals from multiple small proline-rich protein genes. The reduction in expression level of genes required for plug formation and sperm-oocyte interaction may affect fer- tility, which we checked carefully in records obtained from ϩ ϩ Ϫ Ϫ multiple years of PRLR / and PRLR / breeding. Overall Ϫ Ϫ infertility of PRLR / males was seen in 5 of 52 animals, compared with 1 of 59 of the other genotypes, supporting results of our initial investigation using a small number of Ϫ Ϫ animals (25). The potential subfertility of PRLR / males has not previously been investigated. In fertile animals latency to Ϫ Ϫ first pregnancy was longer in PRLR / animals, and overall they had only 40% probability of producing a first preg- ϩ ϩ nancy, compared with PRLR / animals. Intriguingly, these animals show a return to full fertility following the first pregnancy. Mating behavior appears normal in these males, discounting a learning defect that is overcome by experience.
PRLRϩ/Ϫ males with PRLRϩ/Ϫ females (n ϭ 28 pairs, circles), and PRLRϩ/ϩ males with PRLRϩ/ϩ females (n ϭ 31 pairs, triangles). A and B, Kaplan-Meier survival analysis of latency to first pregnancy (A) and second pregnancy (B). Log rank P value for PRLRϪ/Ϫ males with PRLRϩ/Ϫ females vs. PRLRϩ/Ϫ males with PRLRϩ/Ϫ females. C, Cor- FIG. 6. Male fertility in PRLRKO breeding colony. Three types of relation between latency at first pregnancy and second pregnancy (R2 monogamous, continuous pairings were analyzed for their breeding correlation coefficient). D, Correlation of male age at pairing with life; PRLRϪ/Ϫ males with PRLRϩ/Ϫ females (n ϭ 52 pairs, squares), latency to first pregnancy (R2 correlation coefficient). 3204 Endocrinology, July 2003, 144(7):3196–3205 Robertson et al. • Prolactin and Prostate
Another possibility arises from the breeding method. These References results were obtained from a production colony that exploits 1. Cunha G, Donjacour A, Cooke P, Mee S, Bigsby RM, Higgins SJ, Sugimura the efficiency of postpartum mating. In contrast to the first Y 1987 The endocrinology and development of the prostate. Endocr Rev pregnancy, the second and subsequent pregnancies were 8:338–362 2. Huggins C, Russell P 1946 Endocrinology 39:1–7 achieved in the majority of cases by postpartum mating, 3. Vanderlaan W 1960 Observations on the hormonal control of the prostate. Lab which may have alleviated the physiological deficit present Invest 9:185–190 Ϫ Ϫ in PRLR / animals. Regardless of mechanism these data 4. Grayhack J 1963 Pituitary factors influencing the growth of the prostate. Cancer Inst Monographs 12:198–199 establish a clear role for prolactin in male fertility and pro- 5. Kolbusz WE, Lee C, Grayhack JT 1982 Delay of castration induced regression vide specific candidate genes whose combined loss may un- in rat prostate by pituitary homografts. J Urol 127:581–584 derlie the reduction of fertility. 6. Smith C, Assimos D, Lee C, Grayhack JT 1985 Metabolic action of prolactin in regressing prostate: independent of androgen action. Prostate 6:49–59 In summary we have shown that the prolactin receptor is 7. Wennbo H, Kindblom J, Isaksson OG, Tornell J 1997 Transgenic mice over- essential for full male fertility. Reduced expression of genes expressing the prolactin gene develop dramatic enlargement of the prostate gland. Endocrinology 138:4410–4415 involved in sperm-egg recognition, reduction in maternal 8. Adams JB 1985 Control of secretion and the function of C19-delta 5-steroids immune response to sperm, and copulatory plug formation of the human adrenal gland. Mol Cell Endocrinol 41:1–17 may underlie this effect. Dorsal prostate epithelium is lost 9. Nevalainen MT, Valve EM, Ahonen T, Yagi A, Paranko J, Harkonen PL 1997 Androgen-dependent expression of prolactin in rat prostate epithelium in vivo and fewer ventral PIN lesions are formed in response to and in organ culture. FASEB J 11:1297–1307 SV40T when PRLR is lost. The subtle nature of the essential 10. Ouhtit A, Morel G, Kelly PA 1993 Visualization of gene expression of short developmental roles for prolactin contrast dramatically with and long forms of prolactin receptor in rat reproductive tissues. Biol Reprod 49:528–536 the effects produced by transgenic overexpression of pro- 11. McKeehan WL, Adams PS, Rosser MP 1984 Direct mitogenic effects of insulin, lactin in the prostate, where the prostate is enlarged 10- to epidermal growth factor, glucocorticoid, cholera toxin, unknown pituitary 20-fold because of hypersecretion and stromal hyperplasia factors and possibly prolactin, but not androgen, on normal rat prostate ep- ithelial cells in serum-free, primary cell culture. Cancer Res 44:1998–2010 and exhibits intraepithelial dysplastic features (7). These ef- 12. Nevalainen MT, Valve EM, Makela SI, Blauer M, Tuohimaa PJ, Harkonen fects have been shown to be independent of the elevated PL 1991 Estrogen and prolactin regulation of rat dorsal and lateral prostate in systemic androgen levels present in the transgenic model organ culture. Endocrinology 129:612–622 13. Nevalainen MT, Valve EM, Ingleton PM, Harkonen PL 1996 Expression and (44). Furthermore, generation of a prostate-specific prolactin hormone regulation of prolactin receptors in rat dorsal and lateral prostate. transgenic mouse has demonstrated a similar prostate phe- Endocrinology 137:3078–3088 notype, indicative of a local direct effect of prolactin in the 14. Nevalainen MT, Valve EM, Ingleton PM, Nurmi M, Martikainen PM, Harkonen PL 1997 Prolactin and prolactin receptors are expressed and func- absence of altered systemic hormone levels (31). Thus, pro- tioning in human prostate. J Clin Invest 99:618–627 lactin is developmentally very active in the hyperprolactine- 15. Ahonen TJ, Harkonen PL, Laine J, Rui H, Martikainen PM, Nevalainen MT mic state but has a subtle essential role. Experimental ap- 1999 Prolactin is a survival factor for androgen-deprived rat dorsal and lateral prostate epithelium in organ culture. Endocrinology 140:5412–5421 proaches that involved prolactin treatment, with or without 16. Kadar T, Redding TW, Ben DM, Schally AV 1988 Receptors for prolactin, prior pituitary ablation, may have mimicked hyperpro- somatostatin, and luteinizing hormone-releasing hormone in experimental lactinemia rather than normal prolactin levels and thus may prostate cancer after treatment with analogs of luteinizing hormone-releasing hormone and somatostatin. Proc Natl Acad Sci USA 85:890–894 have wrongly attributed the trophic effects of prolactin as an 17. Syms A, Harper M, Griffiths K 1985 The effect of prolactin on human BPH essential function of the hormone. The implications of these epithelial cell proliferation. Prostate 6:145–153 18. Janssen T, Darro F, Petein M, Raviv G, Pasteels JL, Kiss R, Schulman CC 1996 findings for the treatment of prostate cancer are significant. In vitro characterization of prolactin-induced effects on proliferation in the Antiandrogen therapy results in hyperprolactinemia in a neoplastic LNCaP, DU145, and PC3 models of the human prostate. Cancer significant number of patients (21). It is possible that the 77:144–149 19. Harper ME, Wilson DW, Jensen HM, Pierrepoint CG, Griffiths K 1987 trophic stimulus provided by hyperprolactinemia during an- Steroid hormone concentrations in relation to patient prognosis and prostate tiandrogen therapy allows some cells in early neoplastic tumour grade. J Steroid Biochem 27:521–524 lesions to survive and continue to proliferate following 20. Eaton NE, Reeves GK, Appleby PN, Key TJ 1999 Endogenous sex hormones and prostate cancer: a quantitative review of prospective studies. Br J Cancer androgen withdrawal. These cells, already androgen insen- 80:930–934 sitive, may then continue to accumulate the genetic muta- 21. Matzkin H, Kaver I, Lewyshon O, Aylon D, Braf Z 1988 The role of prolactin tions that will enable the development of disseminated dis- levels under GnRH analogue treatment in advanced prostatic carcinoma. Can- cer 61:2187–2191 ease. This may be the mechanism by which preoperative 22. Horti J, Figg WD, Weinberger B, Kohler D, Sartor O 1998 A phase II study serum prolactin levels have no prognostic value but post- of bromocriptine in patients with androgen-independent prostate cancer. On- operative serum prolactin levels become prognostic. col Rep 5:893–896 23. Leav I, Merk FB, Lee KF, Loda M, Mandoki M, McNeal JE, Ho SM 1999 Prolactin receptor expression in the developing human prostate and in hy- perplastic, dysplastic, and neoplastic lesions. Am J Pathol 154:863–870 Acknowledgments 24. Xu X, Kreye E, Kuo CB, Walker AM 2001 A molecular mimic of phosphor- ylated prolactin markedly reduced tumor incidence and size when DU145 We thank Karl Peters for assistance with the quantitative PCR. human prostate cancer cells were grown in nude mice. Cancer Res 61:6098– 6104 Received January 16, 2003. Accepted April 1, 2003. 25. Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, Buteau H, Edery M, Address all correspondence and requests for reprints to: Dr. Chris- Brousse N, Babinet C, Binart N, Kelly PA 1997 Null mutation of the prolactin topher Ormandy, Garvan Institute of Medical Research, 384 Victoria receptor gene produces multiple reproductive defects in the mouse. Genes Dev Street Darlinghurst, New South Wales 2010, Australia. E-mail: 11:167–178 [email protected]. 26. Maroulakou I, Anver M, Garrett L, Green J 1994 Prostate and mammary adenocarcinomian transgenic mice carrying a rat C3(1) simian virus 40 large This work was supported by grants from the National Health and tumor fusion gene. Proc Natl Acad Sci USA 91:11236–11240 Medical Research Council of Australia (to C.J.O.), Cancer Council of 27. Shibata MA, Ward JM, Devor DE, Liu ML, Green JE 1996 Progression of New South Wales, Institut National de la Sante´ et de la Recherche prostatic intraepithelial neoplasia to invasive carcinoma in C3(1)/SV40 large Me´dicale (to P.A.K.), the Swedish Cancer Foundation (to J.T. and J.K.) T antigen transgenic mice: histopathological and molecular biological alter- and the Gothenburg Medical Society (to J.K.). ations. Cancer Res 56:4894–4903 Robertson et al. • Prolactin and Prostate Endocrinology, July 2003, 144(7):3196–3205 3205
28. Sugimura Y, Cunha GR, Donjacour AA 1986 Morphogenesis of ductal net- 36. Klemme LM, Roberts KP, Hoffman LB, Ensrud KM, Siiteri JE, Hamilton DW works in the mouse prostate. Biol Reprod 34:961–971 1999 Cloning and characterization of the rat Crisp-1 gene. Gene 240:279–288 29. Wong W, Cheng L 2001 Model-based analysis of oligonucleotide arrays: ex- 37. Roberts KP, Hoffman LB, Ensrud KM, Hamilton DW 2001 Expression of pression index computation and outlier detection. Proc Natl Acad Sci USA crisp-1 mRNA splice variants in the rat epididymis, and comparative analysis 98:31–36 of the rat and mouse crisp-1 gene regulatory regions. J Androl 22:157–163 30. Clement-Lacroix P, Ormandy CJ, Lepescheux L, Ammann P, Damotte D, 38. Yeung CH, Perez-Sanchez F, Schroter S, Kirchhoff C, Cooper TG 2001 Goffin V, Bouchard B, Amling M, Gaillard-Kelly M, Binart N, Baron R, Kelly Changes of the major sperm maturation-associated epididymal protein HE5 PA 1999 Osteoblasts are a new target for prolactin: analysis of bone formation (CD52) on human ejaculated spermatozoa during incubation. Mol Hum Re- in prolactin receptor knockout mice. Endocrinology 140:96–105 prod 7:617–624 31. Kindblom J, Dillner K, Robertson F, Ormandy C, Tornell J, Wennbo H, 39. Lea IA, Sivashanmugam P, O’Rand MG 2001 Zonadhesin: characterization, Prostate hyperplasia in a transgenic mouse with prostate-specific expression localization, and zona pellucida binding. Biol Reprod 65:1691–1700 40. Lundwall A, Peter A, Lovgren J, Lilja H, Malm J 1997 Chemical character- of prolactin. Endocrinology, in press ization of the predominant proteins secreted by mouse seminal vesicles. Eur 32. Nevalainen M, Ahonen T, Yamashita H, Chandrashekar V, Bartke A, Grim- J Biochem 249:39–44 ley PM, Robinson GW, Hennighausen L, Rui H 2000 Epithelial defect in 41. Romano-Carratelli C, Bentivoglio C, Nuzzo I, Benedetto N, Buommino E, prostates of Stat5a-null mice. Lab Invest 80:993–1006 Cozzolino A, Carteni M, Morelli F, Costanza MR, Metafora B, Metafora V, 33. Steger R, Chandrashekar V, Zhao W, Bartke A, Horseman N 1998 Neuroen- Metafora S 2002 Effect of protein SV-IV on experimental Salmonella enterica docrine and reproductive functions in male mice with targeted disruption of serovar Typhimurium infection in mice. Clin Diagn Lab Immunol 9:115–125 the prolactin gene. Endocrinology 139:3691–3695 42. Williams L, McDonald C, Higgins S 1985 Sequence organisation of rat seminal 34. Bartke A 1999 Role of growth hormone and prolactin in the control of repro- vesicle F gene: location of transcriptional start point and sequence comparison duction: what are we learning from transgenic and knockout animals? Steroids with six other androgen-regulated genes. Nucleic Acids Res 13:659–672 64:598–604 43. Cabral A, Sayin A, de Winter S, Fischer DF, Pavel S, Backendorf C 2001 SPRR4, 35. Hayashi M, Fujimoto S, Takano H, Ushiki T, Abe K, Ishikura H, Yoshida a novel cornified envelope precursor: UV-dependent epidermal expression and MC, Kirchhoff C, Ishibashi T, Kasahara M 1996 Characterization of a human selective incorporation into fragile envelopes. J Cell Sci 114:3837–3843 glycoprotein with a potential role in sperm-egg fusion: cDNA cloning, im- 44. Kindblom J, Dillner K, Ling C, Tornell J, Wennbo H 2002 Progressive pros- munohistochemical localization, and chromosomal assignment of the gene tate hyperplasia in adult prolactin transgenic mice is not dependent on ele- (AEGL1). Genomics 32:367–374 vated serum androgen levels. Prostate 53:24–33 0888-8809/03/$15.00/0 Molecular Endocrinology 17(12):2436–2447 Printed in U.S.A. Copyright © 2003 by The Endocrine Society doi: 10.1210/me.2003-0199
The Cyclin-Dependent Kinase Inhibitor p27 (Kip1) Regulates Both DNA Synthesis and Apoptosis in Mammary Epithelium But Is Not Required for Its Functional Development during Pregnancy
ELIZABETH A. DAVISON, CHRISTINE S. L. LEE, MATTHEW J. NAYLOR, SAMANTHA R. OAKES, ROBERT L. SUTHERLAND, LOTHAR HENNIGHAUSEN, CHRISTOPHER J. ORMANDY, AND ELIZABETH A. MUSGROVE Cancer Research Program (E.A.D., C.S.L.L., M.J.N., S.R.O., R.L.S., C.J.O., E.A.M.), Garvan Institute of Medical Research, St. Vincent’s Hospital, Sydney, New South Wales 2010, Australia; and Laboratory of Genetics and Physiology (L.H.), National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20892-0822
Decreased expression of the cyclin-dependent ki- thesis was observed during pregnancy in p27؊/؊ nase (CDK) inhibitor p27(Kip1) is common in breast mammary gland transplants, but this was paral- cancer and is associated with poor prognosis. p27 leled by increased apoptosis. During pregnancy is also an important mediator of steroidal regula- and at parturition, development and differentiation -tion of cell cycle progression. We have therefore of p27؉/؉ and p27؊/؊ mammary tissue were indis investigated the role of p27 in mammary epithelial tinguishable. These results demonstrate a role for cell proliferation. Examination of the two major p27 in both the proliferation and survival of mam- functions of p27, assembly of cyclin D1-Cdk4 com- mary epithelial cells. However, the absence of mor- plexes and inhibition of Cdk2 activity, revealed that phological and cellular defects in p27؊/؊ mammary cyclin D1-Cdk4 complex formation was not im- tissue during pregnancy raises the possibility that paired in p27؊/؊ mammary epithelial cells in pri- loss of p27 in breast cancer may not confer an mary culture. However, cyclin E-Cdk2 activity was overall growth advantage unless apoptosis is also increased approximately 3-fold, indicating that the impaired. (Molecular Endocrinology 17: 2436–2447, CDK inhibitory function of p27 is important in mam- 2003) mary epithelial cells. Increased epithelial DNA syn-
ECREASED EXPRESSION OF the cyclin-depen- breast cancer cells, where its induction by progestins Ddent kinase inhibitor (CKI), p27 (Kip1), resulting and antiestrogens is critical to inhibition of proliferation from reduced stability of the protein rather than gene (7–10). In addition, altered posttranslational modifica- deletion or mutation, is common in many epithelial tion of p27 has been implicated in resistance to TGF- cancers, including breast cancer (1). Loss of p27 ex- mediated growth inhibition of normal human mam- pression in breast cancer is associated with poor mary epithelial cells in culture (11). However, relatively prognosis in most studies, although some studies little is known about the role of this cell cycle-regula- have not found a significant association between p27 tory molecule in the mammary gland in vivo. expression and outcome (2). Furthermore, p27 expres- Cell cycle progression is driven by interactions be- sion is decreased in hyperplasias, suggesting that de- tween cyclins, cyclin-dependent kinases (CDKs), and regulation of p27 expression occurs early in the de- CKIs. The D-type cyclins associated with Cdk4 or velopment of breast cancer (3). These data suggest a Cdk6 are critical during G1 phase, as is cyclin E-Cdk2. role for p27 as a tumor suppressor in breast cancer, Once activated by cyclin binding, these CDKs phos- consistent with its function as a negative regulator of phorylate members of the family of pocket proteins cell cycle progression and the demonstration that including pRb to allow progress into S phase (12). An mice with a homozygous or heterozygous disruption additional level of CDK regulation is provided by CKIs of the p27 gene are more susceptible to tumorigenesis including p27 and the related p21 (Waf1, Cip1) and in response to radiation and diverse chemical carcin- p57 (Kip2). Overexpression of the Cip/Kip CKIs leads ogens (4–6). In vitro studies have demonstrated that to cell cycle arrest in G1 phase, because of their potent p27 is an important regulator of cellular proliferation in inhibition of Cdk2 (12). These inhibitors have a dual role: they stabilize complexes between D-type cyclins Abbreviations: BrdU, Bromodeoxyuridine; CDK, cyclin-de- and their CDK partners, and so can also function as pendent kinase; CKI, cyclin-dependent kinase inhibitor; dpc, days post coitum; MEC, mammary epithelial cell; PCNA, prolif- positive regulators of cell cycle progression (13, 14). A erating cell nuclear antigen; TUNEL, terminal deoxynucleotidyl significant fraction of p27 is associated with cyclin D1, transferase-mediated dUTP-biotin nick end-labeling. leading to the hypothesis that a major function of
2436 Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Davison et al. • Role of p27 in Mammary Development Mol Endocrinol, December 2003, 17(12):2436–2447 2437
cyclin D1 is to sequester p27 and thus prevent p27 RESULTS from inactivating cyclin E-Cdk2. The ability of deletion of p27 to restore normal development in tissues, in- Decreased Expression of p27 during Pregnancy cluding the mammary gland, which display defects in mice lacking cyclin D1, illustrates the importance of To establish whether p27 expression was altered dur- this role in vivo (15, 16). Mice lacking p27 are larger ing physiological regulation of proliferation in normal than their wild-type siblings because of an overall in- mammary tissue, we first determined the pattern of crease in cell number, most obvious in tissues that p27 expression during pregnancy. Immunohistochem- normally express p27 at the highest levels (17–19), istry using sections of paraffin-embedded mammary demonstrating that p27 is a critical negative regulator glands from mature female mice at estrus and at var- of cell proliferation in vivo. ious stages though pregnancy revealed intense nu- Mammary gland development is initiated in the em- clear p27 staining in the epithelial cells, as shown in Fig. 1A. At estrus approximately 90% of the ductal bryo and occurs in defined stages associated with epithelium expressed p27, but by 2.5–4.5 dpc (days sexual development and reproduction, under the in- postcoitus) this was reduced to approximately 60% fluence of ovarian, placental, and pituitary hormones (Fig. 1B; 2.5 dpc, P ϭ 0.0008; 3.5 dpc, P ϭ 0.002; 4.5 (20). In the prepubertal mouse, the mammary epithe- dpc, P ϭ 0.0008 compared with estrus). The propor- lium forms a rudimentary ductal structure extending tion of p27-positive cells rose at 5.5 dpc (no significant into the mammary fat pad from the nipple. Initiation of difference at 5.5 dpc compared with estrus, P ϭ 0.1) ovarian hormone secretion at the onset of puberty but subsequently decreased again and remained be- accelerates ductal growth and branching (21, 22). The low that present at estrus until at least 15.5 dpc (Fig. 1; final stages of development do not occur until 7.5 dpc, P ϭ 0.05; 11.5 dpc, P ϭ 0.001; 15.5 dpc P ϭ pregnancy, when the epithelial cells within the ducts 0.04 compared with estrus). In the same mammary undergo extensive proliferation and differentiation, re- glands, bromodeoxyuridine (BrdU) staining was low at sulting in expansion of the lobuloalveolar compart- estrus, but increased significantly by 2.5 dpc (2.5 dpc, ment and differentiation of epithelial cells necessary P ϭ 0.002; 3.5 dpc, P ϭ 0.04; 4.5 dpc, P ϭ 0.01 for milk secretion (21, 22). Recent studies using compared with estrus), returned toward control at 5.5 knockout mice have documented critical roles for a dpc (P ϭ 0.06), but then remained elevated thereafter number of hormones (estrogen, progesterone, and (Fig. 1A; 7.5 dpc, P ϭ 0.009; 11.5 dpc, P ϭ 0.0001; prolactin) as well as transcription factors (signal 15.5 dpc, P ϭ 0.002 compared with estrus). When the transducer and activator of transcription 5A and proportions of BrdU-positive and p27-positive cells CCAAT enhancer binding protein-) in mammary were quantitated, it was clear that p27 expression development during puberty and pregnancy (21, 22). during pregnancy mirrored the amount of DNA syn- However, despite the necessity for precise temporal thesis (Fig. 1B). The inverse relationship between p27 and spatial control of proliferation during mammary and BrdU positivity was statistically significant: P ϭ 2 development, relatively few studies have examined 0.005; r ϭ 0.32. The close inverse relationship the role of cell cycle-regulatory molecules in this between p27 expression and DNA synthesis is con- process. A notable exception is the demonstration sistent with the hypothesis that p27 might play an that cyclin D1 is necessary for normal alveolar de- important role in control of epithelial cell proliferation velopment and lactation (23, 24). In addition, al- as the mammary gland develops during pregnancy. though pRb is not required for mammary develop- Formation of Cyclin D1-Cdk4 Complexes in ment (25), expression of phosphorylation-resistant Mammary Epithelial Cells in Culture Does Not pRb alleles impairs mammary development and Require p27 leads to precocious differentiation (26). In this study we have investigated the role of p27 in proliferation To examine the biochemical consequences of loss of of mammary epithelial cells and development of the p27 in the absence of interference from stromal and mammary gland, using mice lacking p27. The func- other contaminating cell types that would be present tion of p27 as a CKI suggested that, in the absence in lysates of whole mammary glands, we used primary of p27, mammary epithelial cell proliferation might mammary epithelial cell (MEC) cultures from mice be increased, consistent with a role for decreased lacking functional p27 due to deletion of the cyclin- p27 expression in loss of growth control during tu- CDK interaction domain (18). The phenotype of these morigenesis. Our data confirm this hypothesis but mice is apparently identical to that of mice lacking the also demonstrate increased apoptosis. As a result entire coding region of p27 (17, 19), and so they are the overall morphology and histology of mammary referred to here as p27Ϫ/Ϫ. The inguinal mammary tissue are the same in the presence or absence of glands from p27Ϫ/Ϫ and p27ϩ/ϩ mice were dissected, p27. This is in contrast with data published during the epithelial cells were purified by several rounds of the course of this investigation (27), which assigned collagenase digestion, and the resulting MECs were an essential role to p27 in normal mammary harvested after several days culture. Initially, we ex- development. amined whether the absence of functional p27 led to
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 2438 Mol Endocrinol, December 2003, 17(12):2436–2447 Davison et al. • Role of p27 in Mammary Development
alterations in the level of other cell cycle-regulatory proteins. Western blot analysis of MECs derived from p27ϩ/ϩ and p27Ϫ/Ϫ mammary tissue demonstrated no difference in the expression of cyclin D1, cyclin D3, cyclin E, Cdk4, Cdk2, or p21 in samples normalized for expression of cytokeratin 18, an epithelial cell marker (Fig. 2A and Table 1). To investigate whether the ab- sence of functional p27 affected complex formation between cyclin D1 and Cdk4, cyclin D1 immunopre- cipitates from MEC lysates were Western blotted us- ing an anti-Cdk4 antibody. This revealed that neither cyclin D1 nor Cdk4 levels were significantly different in immunoprecipitates of MEC lysates from p27ϩ/ϩ and p27Ϫ/Ϫ glands (Fig. 2B and Table 1), clearly demon- strating that there was no impairment of complex for- mation in MECs lacking p27. In the absence of func- tional p27, however, there was a significant increase in the association between cyclin D1 and p21 (Fig. 2B and Table 1). No corresponding increase in overall p21 abundance was observed (Fig. 2A and Table 1), sug-
Fig. 2. Cyclin, CDK, and CKI Expression and Complex For- mation in Cultured MECs A, MECs were derived from the inguinal mammary glands Fig. 1. Relationship between Expression of p27 and DNA of 16-wk-old mice and harvested after 4 d primary culture. Synthesis in Murine Mammary Tissue during Pregnancy Representative Western blots of MEC lysates from two inde- A, Immunohistochemical analysis of p27 protein expres- pendent cultures (each containing MECs derived from the sion and BrdU incorporation (as a measure of DNA synthesis) pooled mammary glands of two or three animals) examining using paraffin sections of mammary tissue from 16-wk-old cyclin D1, cyclin D3, cyclin E, Cdk4, Cdk2, p27, and p21 C57BL/6;129 ϫ 1/SvJ female mice at estrus (E) and at 2.5 expression are shown. The same membranes were also blot- and 11.5 dpc. Bar,50m. B, Quantitation of p27 protein ted for the epithelial cell marker cytokeratin 18 as a loading expression (diamonds) and BrdU incorporation (squares) control. B, MEC lysates prepared as in panel A were immu- within the luminal epithelium at estrus (E) and during preg- noprecipitated using cyclin D1 antibodies and then Western nancy. Data are presented as mean Ϯ SEM or range of two to blotted using antibodies to cyclin D1, Cdk4, and p21 as four animals for each timepoint, except BrdU incorporation at indicated. Immunoprecipitates of two independent cultures 3.5 dpc (n ϭ 7) and 7.5 dpc (n ϭ 11). In some cases the SEM (each containing MECs from two or three animals) for each or range is smaller than the size of the symbol used. genotype are shown.
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Davison et al. • Role of p27 in Mammary Development Mol Endocrinol, December 2003, 17(12):2436–2447 2439
Table 1. Cyclin, CDK, and CKI Levels in MEC Lysates and Immunoprecipitates p27ϩ/ϩ p27Ϫ/Ϫ Western Blota P Mean SEM Mean SEM Cyclin D1 75 3.5 76 5.3 0.94 Cyclin D3 49 5.0 44 2.3 0.39 Cyclin E 62 4.7 61 2.2 0.76 Cdk4 122 8.5 132 13 0.59 Cdk2 77 4.7 88 6.3 0.25 p21 29 4 30 0.91 0.7
Cyclin D1 Immunoprecipitatesb P Mean Range Mean Range Cyclin D1 129 8 126 5 0.73 Cdk4 136 8 134 3.6 0.83 p21 26 1.3 57 1.6 0.0042 ϩ ϩ a Data are the mean and SEM for each protein relative to cytokeratin 18, in arbitrary units. p27 / ,nϭ 3 independent cultures representing a total of seven animals; p27Ϫ/Ϫ,nϭ 4 independent cultures representing a total of eight animals. b Data are expressed as the mean and range, in arbitrary units. p27ϩ/ϩ,nϭ 2 independent cultures representing a total of five animals; p27Ϫ/Ϫ,nϭ 2 independent cultures representing a total of four animals.
gesting altered distribution of p21 among its binding shown), consistent with the low levels of p27 in p27ϩ/ϩ partners in the absence of p27 and indicating a role for MECs at these time points. p21 in the stabilization of cyclin D1-Cdk4 complexes in this context. Loss of p27 Has No Effect on the Development, Architecture, and Differentiation of Cyclin E-Cdk2 Activity Is Increased in the Mammary Epithelium Absence of p27 Data presented in Figs. 2 and 3 indicated that the The data in Fig. 2 indicate that the assembly factor principal role of p27 in mammary epithelial cell prolif- function of p27 is not essential in MECs. To examine eration in vitro was likely to be inhibition of CDKs the other major function of p27, CDK inhibition, cyclin including cyclin E-Cdk2, rather than acting as an as- E-Cdk2 activity was measured in MECs in culture. This sembly factor for cyclin D1-Cdk4. This implied that kinase is active during G1 phase and is an important loss of p27 in the mammary gland in vivo might lead to target for p27. However, it is tightly regulated in con- increased cell cycle progression and consequent cert with changes in proliferation rate, and we there- mammary hyperplasia. Since female p27Ϫ/Ϫ mice are fore initially examined regulation of cyclin E-Cdk2 ac- infertile, it was necessary to perform mammary trans- tivity during culture of p27ϩ/ϩ MECs. Freshly purified plants to investigate the effect of lack of p27 on mam- MECs displayed low proliferation rates (measured by mary development during pregnancy. Mammary frag- BrdU incorporation) consistent with the low prolifera- ments from p27Ϫ/Ϫ females were transplanted into tion rates of mammary epithelium in vivo in virgin an- host mammary fat pads that had been cleared of en- imals. However, the proportion of BrdU-positive cells dogenous epithelium, while the contralateral cleared increased to 10–15% by 3 d (data not shown). The mammary fat pads received mammary fragments from abundance of p27 was initially high (Fig. 3A), consis- p27ϩ/ϩ females. The host animals were prepubertal tent with the widespread expression of p27 detected Rag 1Ϫ/Ϫ animals, which lack T and B cells (28) and by immunohistochemistry in virgin mice (Fig. 1), but will thus accept allografts, but are endocrinologically decreased with increasing time in culture (Fig. 3A). normal. After a 10-wk interval, the host animals were Over the same timeframe, cyclin E-Cdk2 activity in- mated and the transplanted mammary glands ana- creased (Fig. 3A). These data suggested that p27 in- lyzed at intervals during pregnancy. hibition of cyclin E-Cdk2 was likely to be of most Under the influence of pubertal hormones from the significance in the initial stage of culture, before BrdU host, mammary epithelium from the transplanted frag- incorporation had reached its peak, and we therefore ments formed a mammary ductal structure that was compared cyclin E-Cdk2 activity in p27Ϫ/Ϫ and morphologically similar to the endogenous mammary p27ϩ/ϩ MEC cultures harvested after 1–2 d of culture. glands, but lacked connection to the nipple. Mature Cyclin E-Cdk2 activity was increased 3-fold in p27Ϫ/Ϫ virgin animals displayed a branched ductal tree, which MECs compared with p27ϩ/ϩ MECs (P ϭ 0.013, Fig. often extended to the edges of the host fat pad (Fig. 3B). In longer-term cultures there was no apparent 4A). The transplanted epithelium underwent further difference in the cyclin E-Cdk2 activity of MECs de- development during pregnancy to form numerous al- rived from p27Ϫ/Ϫ and p27ϩ/ϩ glands (data not veolar buds by 13.5 dpc (Fig. 4A), but no differences in
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 2440 Mol Endocrinol, December 2003, 17(12):2436–2447 Davison et al. • Role of p27 in Mammary Development
The lack of any alteration in morphology in p27Ϫ/Ϫ transplants was unexpected given the increased cy- clin E-Cdk2 activity documented in Fig. 3. To establish whether the lack of aberrant development was due to the specific gene knockout mouse strain, one of the authors (L.H.) investigated mammary development in p27Ϫ/Ϫ mice, which lack the entire coding region (17) rather than lacking the cyclin-CDK-interacting region alone. These mice are on a 129 genetic background and were purchased from the Jackson Laboratory (Bar Harbor, ME). Mammary transplantation experiments were again performed. Tissue fragments from p27Ϫ/Ϫ females and p27ϩ/ϩ littermates were transplanted into cleared contralateral fat pads of nude mice, which were mated after 8 wk. At parturition, p27Ϫ/Ϫ and p27ϩ/ϩ mammary tissue were histologically indistin- guishable (Fig. 5), confirming the results of the mam- mary transplants in Fig. 4.
Increased DNA Synthesis in the Absence of p27
To determine whether the absence of p27 led to alter- ations in cell proliferation despite the lack of effect on overall morphology, markers of proliferation were measured in p27Ϫ/Ϫ and p27ϩ/ϩ mammary trans- plants. BrdU immunohistochemistry was performed to determine the number of epithelial cells undergoing active DNA synthesis. Serial sections were stained for expression of the proliferating cell nuclear antigen Fig. 3. Increased Cyclin E-Cdk2 Activity in p27Ϫ/Ϫ MECs (PCNA), which is expressed in early G1 and S phase A, MECs were derived from the inguinal mammary glands and thus has been used as a marker of actively pro- of 16-wk-old p27ϩ/ϩ mice. Lysates were prepared after 2 or liferating cells. BrdU-positive and PCNA-positive epi- 3 d primary culture and Western blotted using antibodies to p27. The kinase activity of cyclin E immunoprecipitates was thelial cells were present in the epithelium of trans- Ϫ/Ϫ measured using histone H1 substrate. B, Lysates were pre- planted mammary glands from both p27 and ϩ/ϩ Ϫ/Ϫ pared from MECs derived from p27ϩ/ϩ or p27Ϫ/Ϫ mice after p27 mice, but were more frequent in p27 trans- 1–2 d primary culture. The kinase activity of cyclin E immu- plants (Fig. 6A). Comparison of data from 12 glands noprecipitates was measured using histone H1 substrate and with 0–40% PCNA-positive epithelium demonstrated quantitated. Data are the mean Ϯ SEM of five independent that the number of BrdU-positive cells was propor- ϩ ϩ cultures of p27 / MECs (representing 10 animals) and three tional to the number of PCNA-positive cells (r2 ϭ 0.6; Ϫ/Ϫ independent cultures of p27 MECs (five animals). P ϭ 0.006), and thus only the results from PCNA- stained sections are shown in Fig. 6B. Quantitation of the proportion of epithelial cells displaying nuclear gross morphology between p27ϩ/ϩ and p27Ϫ/Ϫ staining confirmed consistent elevation of both BrdU glands were apparent. Quantitation of side branching and PCNA positivity in p27Ϫ/Ϫ epithelium (Fig. 6B and and alveolar bud development in whole mounts of data not shown). The approximately 2.5-fold increase transplanted mammary glands showed, as expected in PCNA-positive cells was statistically significant at in mature animals, little increase in the number of both 5.5 dpc (P ϭ 0.05) and 13.5 dpc (P ϭ 0.03). Thus, ductal branches but a progressive increase in the although loss of p27 did not affect the overall mor- number of alveolar buds as the gland developed phology of the gland, it led to an increase in prolifer- through pregnancy (Fig. 4B). There was, however, no ation, consistent with the increased cyclin E-Cdk2 significant difference between p27ϩ/ϩ and p27Ϫ/Ϫ activity observed in p27Ϫ/Ϫ MECs. One possible ex- transplants for either of these parameters at any time planation for the apparently normal development of point examined. Microscopic examination of hema- p27Ϫ/Ϫ mammary transplants despite altered CDK ac- toxylin and eosin-stained sections of these mammary tivity in p27Ϫ/Ϫ MECs in culture is that, because MEC glands did not reveal any differences in tissue archi- proliferation during pregnancy is strongly steroid- tecture between p27Ϫ/Ϫ and p27ϩ/ϩ glands either in dependent but MEC proliferation in vitro is not, p27 virgin glands or throughout pregnancy, i.e. there were may not perform the same function in both experimen- no apparent increases in the number of cells consti- tal models. However, the increase in proliferation of tuting the epithelium of individual ducts (data not p27Ϫ/Ϫ epithelium during pregnancy suggests that shown). p27 is playing a similar role in both contexts, princi-
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Davison et al. • Role of p27 in Mammary Development Mol Endocrinol, December 2003, 17(12):2436–2447 2441
Fig. 5. Histology of p27ϩ/ϩ and p27Ϫ/Ϫ Mammary Tissue at Parturition Transplanted p27ϩ/ϩ or p27Ϫ/Ϫ mammary tissue was col- lected at parturition, paraffin embedded, and stained using hematoxylin and eosin. Bar in upper panels, 300 m; bar in lower panels,50m.
pally as an inhibitor of cyclin E-Cdk2, i.e. that its func- tion in mammary epithelium is not dependent on the factor(s) stimulating proliferation, consistent with the established role of p27 as a target for multiple mito- genic stimuli.
Increased Apoptosis in the Absence of p27
The increased proliferation but normal morphology of the p27Ϫ/Ϫ mammary transplants raised the possibility that increased apoptosis was also occurring in the absence of p27. We therefore quantitated apoptosis in the transplanted mammary glands. Both in situ end labeling of DNA strand breaks [terminal deoxynucleo- tidyl transferase-mediated dUTP-biotin nick end- labeling (TUNEL)] and immunohistochemical staining for active caspase 3, an effector caspase involved in the DNA fragmentation that characterizes apoptosis, were increased in p27Ϫ/Ϫ transplants (Fig. 7A). The number of caspase 3-positive cells (Fig. 7B) increased
Fig. 4. Morphology of p27ϩ/ϩ and p27Ϫ/Ϫ Mammary Trans- plants A, Whole mounts of mammary tissue derived from trans- planted p27ϩ/ϩ or p27Ϫ/Ϫ epithelium at estrus and at 5.5 and 13.5 dpc were carmine stained. Bar, 2.5 mm. B, The number of ductal branches and lobuloalveolar buds per 1 mm2 area of transplanted p27ϩ/ϩ (open boxes)orp27Ϫ/Ϫ (closed boxes) mammary tissue were quantitated at estrus and at 5.5 and 13.5 dpc. Data were obtained from three to four animals for each time point.
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 2442 Mol Endocrinol, December 2003, 17(12):2436–2447 Davison et al. • Role of p27 in Mammary Development
Fig. 7. Increased Apoptosis in p27Ϫ/Ϫ Mammary Trans- plants A, Representative active caspase 3 staining in p27ϩ/ϩ and Ϫ Ϫ Fig. 6. Increased Proliferation in p27Ϫ/Ϫ Mammary Trans- p27 / epithelium at 5.5 dpc and TUNEL staining at 13.5 plants dpc. Bar,50m. B, Quantitation of TUNEL positivity (open A, Immunohistochemical detection of BrdU incorporation boxes) and active caspase 3 staining (hatched boxes)in ϩ ϩ Ϫ Ϫ and PCNA expression in p27ϩ/ϩ and p27Ϫ/Ϫ mammary trans- p27 / (open boxes) and p27 / (shaded boxes) mammary plants at 13.5 dpc. Bar,50m. B, Quantitation of PCNA transplants. Data were obtained from four to seven trans- expression in p27ϩ/ϩ (open boxes) and p27Ϫ/Ϫ (shaded planted glands per genotype at each time point except at 5.5 ϩ ϩ boxes) mammary transplants. Data were obtained from three dpc (n ϭ 2 for each genotype) and TUNEL-stained p27 / to five transplants per genotype at each time point, except virgin transplants (n ϭ 2). Transplants examined at 13.5 and p27ϩ/ϩ at 5.5 dpc (n ϭ 2). 15.5 dpc gave similar results and consequently data have been pooled. approximately 3-fold, an increase that was statistically significant at all time points (virgin, P ϭ 0.025; 5.5 dpc, P ϭ 0.01; 13.5–15.5 dpc, P ϭ 0.05). Similarly, the DISCUSSION increase in TUNEL-positive cells (Fig. 7B) was statis- tically significant in both virgin animals (P ϭ 0.05) and We have examined the consequences of loss of the during pregnancy (13.5–15.5 dpc, P ϭ 0.02). The pro- CKI, p27, in mammary epithelial cells as a means of portion of nuclei displaying the morphological features gaining insight into its role in normal mammary devel- of apoptosis (i.e. condensed nuclei) was also in- opment and the potential consequences of its de- creased from 0.2% in p27ϩ/ϩ transplants to 0.8% in creased expression in breast cancer. Biochemical p27Ϫ/Ϫ transplants (P ϭ 0.003 overall). All three mea- analysis of p27Ϫ/Ϫ MECs in culture revealed that loss sures indicate increased apoptosis in p27Ϫ/Ϫ mam- of p27 resulted in increased cyclin E-Cdk2 activity, but mary tissue and suggest that, in the absence of p27, had no effect on cyclin D1-Cdk4 assembly. DNA syn- increased apoptosis balances increased rates of epi- thesis during pregnancy was increased approximately thelial cell proliferation to result in a mammary ductal 2.5-fold in transplanted p27Ϫ/Ϫ mouse mammary ep- structure that is apparently indistinguishable from wild ithelium compared with transplanted wild-type epithe- type in terms of numbers of epithelial cells, overall lium. However, the increased mammary epithelial DNA tissue architecture, and macroscopic morphology. synthesis during pregnancy was accompanied by an
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Davison et al. • Role of p27 in Mammary Development Mol Endocrinol, December 2003, 17(12):2436–2447 2443
increase in apoptosis, indicating that p27 functions in of normal morphology during pregnancy. It is possible regulation of both proliferation and survival in this tis- that differences in the details of the experimental pro- sue. These observations are consistent with the in- tocols used in each laboratory contribute to differ- creased Cdk2 activity and hyperplasia observed in ences in the morphology of transplanted mammary other organs in p27Ϫ/Ϫ mice (17–19) but contrast with glands, although this would not also account for dif- results from another laboratory reported while this ferences in expression of cell cycle-regulatory proteins study was in progress (27). Although they reported in MECs from p27Ϫ/Ϫ females. One possible factor increased proliferation resulting from loss of a single investigated was the age of the animals used, since p27 allele, Muraoka and colleagues (27) observed im- Muraoka et al. (27) used 6-wk-old animals as trans- paired development of transplanted p27Ϫ/Ϫ mammary plant donors and prepared whole mammary lysates epithelium during pregnancy, accompanied by de- from 3-wk-old animals, whereas our initial studies creased rates of proliferation and delayed differentia- used older animals, typically approximately 16 wk old. tion. In whole mammary gland extracts from p27Ϫ/Ϫ However, when we repeated our studies using 6- to mice, they found significantly decreased levels of cy- 8-wk-old animals, the development of transplanted clin D1 expression (relative to an epithelial cell marker, p27Ϫ/Ϫ epithelium during pregnancy was not im- cytokeratin 14) and failure to form cyclin D1-Cdk4 paired, and there were no differences in cyclin D1 complexes with consequent marked decreases in levels in MEC cultures from 5- to 6-wk-old animals. Cdk4 activity (27). In contrast, Cdk2 activity was com- Further experimentation, e.g. direct comparison of parable in p27Ϫ/Ϫ and p27ϩ/ϩ mammary gland ex- p27Ϫ/Ϫ epithelium from different laboratories within a tracts (27). Thus in their hands the absence of p27 single experiment, will be required to resolve this principally results in lack of cyclin D1 function, such issue. that the phenotype of p27Ϫ/Ϫ mammary glands mir- The striking inverse relationship between p27 ex- rors that of cyclin D1Ϫ/Ϫ glands (23, 24, 29), while in pression and DNA synthesis during pregnancy is con- our hands the major consequences were increases in sistent with the established role for p27 as a CKI with cyclin E-Cdk2 activity, proliferation, and apoptosis. a fundamental role in maintenance of quiescence and Ϫ/Ϫ Ϫ/Ϫ Crosses between p27 and cyclin D1 mice inhibition of transition from G1 into S phase (12). The have been generated in two laboratories (15, 16). Anal- change in p27 expression during pregnancy may sim- ysis of these mice reveals that ablation of p27 restores ply be a consequence of altered proliferation. How- normal development to cyclin D1-dependent tissues, ever, because MECs derived from the mammary including the retina and mammary gland. These data glands of p27Ϫ/Ϫ mice have increased levels of cyclin are in contrast with what would be predicted if both E-Cdk2 activity, and p27Ϫ/Ϫ mammary epithelium dis- cyclin D1Ϫ/Ϫ and p27Ϫ/Ϫ mammary epithelia exhibit plays increased DNA synthesis during pregnancy, de- impaired mammary development. In that case an ad- creased p27 expression during pregnancy may be a ditional, mammary-specific, pathway would need to cause of changes in the DNA synthetic rate. Given the be activated to allow mammary epithelial cell prolifer- steroid hormone dependence of proliferation during ation and the development of the mammary gland in pregnancy, this points to p27 as a potential mediator the absence of both cyclin D1 and p27 (22). However, of steroid hormone regulation of proliferation in mam- if loss of p27 activates Cdk2 in mammary epithelium, mary epithelium in vivo, consistent with data from as demonstrated here, and can do so in the absence breast cancer cells in vitro (7–10). In another steroid- of cyclin D1, proliferation and consequent mammary responsive tissue, uterine epithelium, estrogen- development in mice lacking both cyclin D1 and p27 mediated cell cycle progression is accompanied by could be driven by cyclin E-Cdk2, analogous to the decreased p27 expression, and both cyclin E-Cdk2 ability of cyclin E to replace cyclin D1 in the mammary and cyclin A-Cdk2 activities are increased in the ab- gland (30) and consistent with our findings of in- sence of p27 (31). Although the absence of p27 did not creased proliferation in p27Ϫ/Ϫ mammary epithelium. impair the ability of progesterone to inhibit estrogen The basis of the differences between our studies stimulation of these cells, these data again suggest and those of Muraoka et al. (27) remains to be deter- p27 as a steroid-regulated CDK inhibitor. mined. The differences do not appear to arise because The cyclin D1-Cdk4 complexes that formed in the of differences in animal strains, because we have used absence of p27 contained increased levels of the re- the same strain of p27Ϫ/Ϫ mice as Muraoka et al. and lated CKI, p21, indicating that p21 can substitute for present additional data from an independent strain of p27 as a cyclin D1-Cdk4 assembly and stabilization p27Ϫ/Ϫ mice that also displayed normal mammary factor in MECs. This was not the case for p27 inhibition morphology when transplanted onto immunocompro- of MEC cyclin E-Cdk2 activity, perhaps because the mised recipients. Our data are also in accord with the relatively low levels of p21 expressed in the mammary results of experiments performed in a third laboratory gland are insufficient to perform both functions. The as controls in an investigation of the phenotype of pRb-family member p130 can bind to cyclin E-Cdk2, mice lacking both cyclin D1 and p27 (15). Thus, data compensating for absent CKIs and allowing normal from independent experiments in three different labo- CDK regulation in fibroblasts lacking both p27 and p21 ratories and using different p27Ϫ/Ϫ mouse strains in- (32). However, there was no evidence for p130 binding dicate that p27Ϫ/Ϫ mammary epithelium forms a gland cyclin E-Cdk2 in MEC lacking p27, perhaps also be-
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 2444 Mol Endocrinol, December 2003, 17(12):2436–2447 Davison et al. • Role of p27 in Mammary Development
cause p130 is expressed at low levels in these cells cepted standards of humane animal care. Mice were geno- (our unpublished data). typed by PCR analysis of genomic DNA using the following primers: ACGTGAGAGTGTCTAACGG, AGTGCTTCTC- Increased apoptosis in the absence of p27 is not CAAGTCCC and GCGAGGATCTCGTCGTGAC (40). These observed in all tissues, although increasing evidence yielded products of 120 bp (wild-type p27) and 400 bp (neo- suggests that p27 may provide protection from apo- mycin). Rag1Ϫ/Ϫ mice on a C57/BL6 genetic background (28) ptosis under conditions of cellular or physiological were purchased from the Animal Resource Centre, Perth, stress (33–35), and decreased rates of apoptosis have WA, Australia. All animals were housed with food and water ad libitum with a 12-h light, 12-h dark cycle at 22 C. been described in some cancer cell lines after en- forced overexpression of p27 (36, 35). One possible mechanism for increased apoptosis in the absence of Cell Culture p27 is that it is triggered by deregulated Cdk2 activity. Primary cultures of MECs were prepared from p27ϩ/ϩ and However, because increased apoptosis has been ob- p27Ϫ/Ϫ animals by limited collagenase digestion essentially Ϫ/Ϫ served in some (33, 37), but not all (17–19), p27 cell as described previously (41). Briefly, both inguinal mammary types exhibiting increased Cdk2 activity and prolifer- glands were surgically removed from mature (16 wk old) ation, this does not seem to be a general mechanism. female mice. Each gland was finely chopped and washed in A further possibility is that the increased rates of apo- HEPES-buffered RPMI 1640 medium supplemented with Ϫ/Ϫ 2.5% fetal calf serum. The tissue was then subjected to ptosis observed in p27 mammary epithelium may approximately four 1-h rounds of digestion at 37 C with 1–2 be a consequence of increased proliferation. Signifi- mg/ml collagenase type L (Sigma-Aldrich Corp., Castle Hill, cant apoptosis occurs in the lumen of the terminal end New South Wales, Australia) in HEPES-buffered RPMI 1640. buds of the developing mammary gland, and this is The resulting mammary organoids were passed through ster- thought to be important for ductal morphogenesis ile 400- m polyester mesh (Small Parts, Inc., Miami Lakes, FL). Organoids derived from two to three animals (i.e. four to (38). A recent study has shown that enhanced prolif- six mammary glands) were pooled and then plated into fi- eration after expression of cyclin D1 or the HPV16 E7 bronectin-coated 25-cm2 tissue culture flasks in primary cul- oncoprotein is not sufficient to impair lumen formation ture medium (1:1 DMEM/Hams F-12 catalog no. 11330 , Life in mammary acini in vitro, but rather is balanced by Technologies, Inc., Melbourne, Victoria, Australia) containing 5 g/ml insulin, 10 ng/ml epidermal growth factor (Promega increased apoptosis to maintain a hollow glandular Corp., Annandale, New South Wales, Australia), 5 g/ml hy- architecture (39). The authors suggested that the drocortisone (Sigma), 5 ng/ml cholera toxin (Sigma), 0.01 apoptosis might be triggered by the inability of some U/ml penicillin (Life Technologies, Inc.), 10 ng/ml streptomy- cells to maintain basement membrane contact, with cin (Life Technologies, Inc.), 20 g/ml gentamicin), and 10% associated loss of polarity (39). Similar mechanisms fetal calf serum. could possibly lead to apoptotic cell death of the ex- cess ductal cells produced by increased proliferation Western Blot Analysis, Immunoprecipitation, and in p27Ϫ/Ϫ mammary epithelium. Kinase Assays In summary, our studies of MEC proliferation and MEC cultures were harvested by trypsinization and pelleted mammary gland development in the absence of by centrifugation. The cell pellets were lysed by incubation on Ϫ/Ϫ p27 provide evidence for a role for p27 in control- ice (5 min) in lysis buffer [50 mM HEPES (pH 7.5), 150 mM ling both the proliferation and survival of normal mam- NaCl, 10% (vol/vol) glycerol, 0.1% Tween-20, 1 mM EDTA, mary epithelial cells and raise questions over the im- 2.5 mM EGTA, 10 mM -glycerophosphate, 10 g aprotinin/ plications of decreased p27 expression in breast ml, 10 g leupeptin/ml, 0.1 mM phenylmethylsulfonyl fluoride, 0.1 mM sodium orthovanadate, 1 mM NaF]. Cellular debris cancer. If reduced p27 expression increases both pro- was cleared by centrifugation, and the lysates stored at liferation and apoptosis in breast cancers it may not Ϫ80 C. lead to increased tumor burden unless the apoptotic For immunoprecipitation, 100 g of MEC lysate were pre- machinery is disrupted. It would thus be of interest to cleared by incubation for1hat4Cwith protein G-Sepharose beads (Zymed Laboratories, Inc., South San Francisco, CA) investigate the relationships between p27 levels, apo- and then immunoprecipitated by incubation for2hat4Cwith ptotic rates, and outcome in clinical breast cancer. cross-linked anticyclin D1 antibody (72–13G, Santa Cruz Bio- technology, Inc., Santa Cruz, CA). Antibody cross-linking was performed as described previously (42). The immunoprecipi- tates were washed with lysis buffer and resuspended in SDS- MATERIALS AND METHODS PAGE sample buffer [63 mM Tris-HCl (pH 6.8), 10% (vol/vol) glycerol, 2% SDS, 5% -mercaptoethanol]. Samples of immunoprecipitated or total protein in SDS- Mice PAGE sample buffer were denatured for 3 min at 95 C and then separated by SDS-PAGE and transferred to Immuno- A colony of p27Ϫ/Ϫ and p27ϩ/ϩ mice was established at the Blot polyvinylidene difluoride membrane (Bio-Rad Laborato- Garvan Institute from founder animals kindly provided by Dr. ries, Inc., Hercules, CA). Transfer was confirmed by staining Andrew Koff (Memorial Sloan-Kettering Cancer Center, New with 0.5% Ponceau S in 10% acetic acid. The membranes York, NY). The p27 coding sequence in these mice is dis- were incubated for2hatroom temperature with the following rupted so that it produces an amino-truncated version of the primary antibodies: cyclin D1 (Ab-3; Neomarkers, Fremont, p27 protein (⌬51) that lacks the cyclin-CDK interaction do- CA); cyclin D3 (C-16), cyclin E (M-20), Cdk2 (M2), Cdk4 main and is nonfunctional (18). Wild-type and knockout ani- (C-22), and p21 (M-19) from Santa Cruz Biotechnology, Inc.; mals were maintained on a mixed C57Bl/6–129/SvJ back- p27 (K25020; Transduction Laboratories, Inc., Lexington, ground. Littermates were used when possible. All animal KY); and cytokeratin 18 (RDI-PRO 61028; Research Diagnos- experimentation was conducted in accordance with ac- tics, Inc., Cleveland, OH).
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Davison et al. • Role of p27 in Mammary Development Mol Endocrinol, December 2003, 17(12):2436–2447 2445
The histone H1 kinase activity of cyclin E immunoprecipi- were injected ip with BrdU dissolved in PBS (100 g BrdU/g tates (100 g MEC lysate) was measured as previously de- body weight), 2 h before euthanasia by cervical dislocation. scribed (43) using 10 g histone H1 as substrate. Little or no Immunohistochemical detection of p27, BrdU, PCNA, and background phosphorylation was detected in samples im- activated caspase 3 was performed using a DAKO au- munoprecipitated by using beads without antibody. tostainer (DAKO A/S, Glostrup, Denmark). Paraffin sections (4 m) were dewaxed and rehydrated. Antigen retrieval was performed using citrate EDTA buffer (DAKO) boiled under Mammary Gland Transplants pressure for p27, low pH target retrieval solution (DAKO) at 100 C for 30 min for BrdU and PCNA detection, and 10 mM Mammary epithelial transplants were performed as follows. trisodium citrate boiled under pressure for 2 min for active 3 Pieces (1 mm ) of mammary gland excised from mature (Ն16 caspase 3. Endogenous peroxidase activity was inhibited wk old) donors were transplanted into a cleared mammary fat with 3% H O , and blocked in 10% normal horse serum. The Ϫ/Ϫ 2 2 pad in a 3-wk-old Rag 1 recipient, i.e. after removal of the ARK peroxidase kit (DAKO) was used for biotinylation of endogenous mammary rudiment. Mammary gland portions mouse monoclonal antibodies. Slides were incubated with Ϫ/Ϫ from one p27 donor were transplanted into the cleared the following biotin-conjugated mouse monoclonal antibod- inguinal (no. 4) fat pads of several recipients. The contralat- ies: anti-p27 at 1:100 (Kip 1/p27; Transduction Laboratories, eral inguinal fat pads were also cleared and transplanted with Inc.), anti-BrdU at 1:50 (Bu20a; DAKO), anti-PCNA at 1:1250 ϩ/ϩ p27 mammary gland sections. In all, 96.4% of transplants (PC10; DAKO). The active caspase 3 rabbit polyclonal anti- were successful. At least 8 wk after transplant, mammary body (R&D Systems, Inc., Minneapolis, MN) was used at glands were analyzed in virgin animals, at 5.5 and 13.5–16.5 1:6000. The sections were counterstained with Witlock’s he- dpc. The morning of plug detection was designated 0.5 dpc. matoxylin and Scott’s blue. Detection of apoptosis in paraffin In one set of transplant experiments, mammary transplants sections by TUNEL analysis was performed using the Dead- Ϫ/Ϫ from 6- to 8-wk-old p27 donor animals were compared End Colorimetric TUNEL System (Promega Corp., Annan- ϩ/ϩ with transplants from age-matched p27 animals or 17- dale, New South Wales, Australia), according to the manu- Ϫ/Ϫ wk-old p27 animals by implantation into contralateral facturer’s instructions. At least 1000 epithelial cells per gland glands. Analysis of these transplants revealed that the age of were counted for each animal at ϫ400 magnification using a the donor animal did not affect the morphology, proliferation, Leica DMRB microscope and photographed using a Leica or apoptotic rate of the transplanted glands; therefore, data DC200 Camera and Leica DC viewer (Leica Microsystems). from donor animals of different ages have been pooled. One set of transplant experiments used p27Ϫ/Ϫ animals obtained from The Jackson Laboratory. These mice lack the Statistical Analysis entire coding region of p27 (17) and were in a 129 back- Ϫ Ϫ ϩ ϩ ground. Mammary tissue from p27 / and p27 / littermates ANOVAs were performed using StatView statistical software was transplanted into cleared fat pads of 3-wk-old nude (Abacus Concepts, Inc., Berkeley, CA), with P values deter- mice, which were then mated after 8 wk. Mammary tissue mined using Fisher’s projected least significant difference was isolated at parturition for histological analysis. test. Morphological examination of the transplanted glands was used to determine that the observed ductal outgrowth was the result of transplanted epithelium and not the result of Acknowledgments endogenous epithelium. Ductal outgrowth resulting from transplanted epithelium originates from the transplant site at We thank Dr. Andrew Koff for providing a breeding pair of the center of the gland, whereas a ductal structure originating p27ϩ/Ϫ mice, Dr. Danielle Lynch for advice on the preparation from the edge of the gland indicates outgrowth that is derived of MEC cultures, and Melanie Trivett for advice on caspase 3 from endogenous epithelium (44). staining. We are grateful to Drs. Darren Saunders and Keiko Miyoshi for assistance with some experiments, and thank Dr. Whole-Mount Analysis Julie Ferguson, Tony Chaplin, Eric Schmied, and other staff of the Garvan Institute animal care facility for advice and expert animal care. For morphological evaluation of transplanted mammary glands, the endogenous third mammary gland and the fourth glands (containing the mammary transplants) were dissected from the skin, spread onto a SuperFrost Plus glass slide Received May 30, 2003. Accepted August 15, 2003. (Menzel-Glaser, Braunscheig, Germany), and fixed in 10% Address all correspondence and requests for reprints to: neutral buffered formalin. For examination of p27 expression Elizabeth A. Musgrove, Cancer Research Program, Garvan during pregnancy, the fourth mammary glands were dis- Institute of Medical Research, 384 Victoria Street, Darling- sected and fixed in the same way. Mammary gland whole hurst, New South Wales 2010, Australia. E-mail: e.musgrove@ mounts were defatted in acetone before carmine alum (0.2% garvan.org.au. carmine, 0.5% aluminium sulfate) staining overnight. The whole mount was dehydrated using a graded ethanol series This research was supported by the US Army Medical followed by xylene treatment for 60 min and storage in methyl Research and Materiel Command (under Grants DAMD17- salicylate (45). Ductal development was examined by low- 00-1-0252 and BC022015), the National Health and Medical power microscopy using a Leica M212 microscope and pho- Research Council of Australia, and The Cancer Council, New tographed with a Leica DC200 Camera and Leica DC viewer South Wales. M.J.N. was the recipient of a University of (Leica Microsystems, Wetzlar, Germany). New South Wales Faculty of Medicine Dean’s Research Scholarship.
Histological and Immunohistochemical Analysis
After morphological evaluation, mammary gland whole REFERENCES mounts were removed from the glass slide, embedded in paraffin, and sectioned for histological and immunohisto- chemical analysis. Histology was examined in paraffin sec- 1. Slingerland J, Pagano M 2000 Regulation of the cdk tions that had been dewaxed, rehydrated, and stained with inhibitor p27 and its deregulation in cancer. J Cell Physiol hematoxylin and eosin. To measure DNA synthesis, animals 183:10–17
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 2446 Mol Endocrinol, December 2003, 17(12):2436–2447 Davison et al. • Role of p27 in Mammary Development
2. Chiarle R, Pagano M, Inghirami G 2001 The cyclin de- lacking the cyclin-dependent kinase inhibitor function of pendent kinase inhibitor p27 and its prognostic role in p27(Kip1). Cell 85:721–732 breast cancer. Breast Cancer Res 3:91–94 19. Nakayama K, Ishida N, Shirane M, Inomata A, Inoue T, 3. De Paolo F, Vecci AM, Granoto AM, Liverani M, Monti F, Shishido N, Horii I, Loh DY 1996 Mice lacking Innoceta AM, Gianni L, Saragoni L, Ricci M, Falcini F, p27(Kip1) display increased body size, multiple organ Amadori D, Volpi A 2002 p27/Kip1 expression in normal hyperplasia, retinal dysplasia, and pituitary tumors. epithelium, benign and neoplastic breast lesions. Cell 85:707–720 J Pathol 196:26–31 20. Topper YJ, Freeman CS 1980 Multiple hormone interac- 4. Fero ML, Randel E, Gurley KE, Roberts JM, Kemp CJ tions in the developmental biology of the mammary 1998 The murine gene p27Kip1 is haplo-insufficient for gland. Physiol Rev 60:1049–1106 tumour suppression. Nature 396:177–180 21. Hennighausen L, Robinson GW 1998 Think globally, act 5. Philipp J, Vo K, Gurley KE, Seidel K, Kemp CJ 2001 locally: the making of a mouse mammary gland. Genes Tumor suppression by p27/Kip1 and p21/Cip1 during Dev 12:449–455 chemically induced skin carcinogenesis. Oncogene 18: 22. Hennighausen L, Robinson GW 2001 Signaling pathways 4689–4698 in mammary gland development. Dev Cell 1:467–475 6. Philipp-Staheli J, Kim K-H, Payne SR, Gurley KE, Liggitt 23. Fantl V, Stamp G, Andrews A, Rosewell I, Dickson C D, Longton G, Kemp CJ 2002 Pathway-specific tumor 1995 Mice lacking cyclin D1 are small and show defects suppression: reduction of p27 accelerates gastrointesti- in eye and mammary gland development. Genes Dev nal tumorigenesis in Apc mutant mice, but not in Smad3 9:2364–2372 mutant mice. Cancer Cell 1:355–368 24. Sicinski P, Donaher JL, Parker SB, Li T, Fazeli A, Gardner 7. Groshong SD, Owen GI, Grimison B, Schauer IE, Todd H, Haslam SZ, Bronson RT, Elledge SJ, Weinberg RA MC, Langan TA, Sclafani RA, Lange CA, Horwitz KB 1997 1995 Cyclin D1 provides a link between development Biphasic regulation of breast cancer cell growth by and oncogenesis in the retina and breast. Cell 82: progesterone: role of the cyclin-dependent kinase inhib- Kip1 621–630 itors, p21 and p27 . Mol Endocrinol 11:1593–1607 25. Robinson GW, Wagner K-U, Hennighausen L 2001 Func- 8. Cariou S, Donovan JC, Flanagan WM, Milic A, Bhatta- tional mammary gland development and oncogene- charya N, Slingerland JM 2000 Down-regulation of induced tumor formation are not affected by the absence p21WAF1/CIP1 or p27Kip1 abrogates antiestrogen- of the retinoblastoma gene. Oncogene 20:7115–7119 mediated cell cycle arrest in human breast cancer cells. 26. Jiang Z, Zacksenhaus E 2002 Activation of retinoblas- Proc Natl Acad Sci USA 97:9042–9046 toma protein in mammary gland leads to ductal growth 9. Carroll JS, Prall OW, Musgrove EA, Sutherland RL 2000 suppression, precocious differentiation, and adenocarci- A pure estrogen antagonist inhibits cyclin E-Cdk2 activity noma. J Cell Biol 156:185–198 in MCF-7 breast cancer cells and induces accumulation 27. Muraoka RS, Lenferink AE, Simpson J, Brantley DM, of p130–E2F4 complexes characteristic of quiescence. Roebuck LR, Yakes FM, Arteaga CL 2001 Cyclin-depen- J Biol Chem 275:38221–38229 dent kinase inhibitor p27(Kip1) is required for mouse 10. Swarbrick A, Lee CS, Sutherland RL, Musgrove EA 2000 mammary gland morphogenesis and function. J Cell Biol Cooperation of p27(Kip1) and p18(INK4c) in progestin- 153:917–932 mediated cell cycle arrest in T-47D breast cancer cells. 28. Mombaerts P, Iacomini J, Johnson RS, Herrup K, Mol Cell Biol 20:2581–2591 Tonegawa S, Papaioannou VE 1992 RAG-1-deficient 11. Ciarallo S, Subramaniam V, Hung W, Lee JH, Kotchetkov R, Sandhu C, Milic A, Slingerland JM 2002 Altered mice have no mature B and T lymphocytes. Cell 68: p27(Kip1) phosphorylation, localization, and function in 869–877 human epithelial cells resistant to transforming growth 29. Fantl V, Edwards PA, Steel JH, Vonderhaar BK, Dickson  C 1999 Impaired mammary gland development in Cyl- factor -mediated G(1) arrest. Mol Cell Biol 22: Ϫ Ϫ 2993–3002 1( / ) mice during pregnancy and lactation is epithelial 12. Sherr CJ, Roberts JM 1999 CDK inhibitors: positive and cell autonomous. Dev Biol 212:1–11 negative regulators of G1-phase progression. Genes Dev 30. Geng Y, Whoriskey W, Park MY, Bronson RT, Medema 15:1501–1512 RH, Li T, Weinberg RA, Sicinski P 1999 Rescue of cyclin 13. LaBaer J, Garrett MD, Stevenson LF, Slingerland JM, D1 deficiency by knockin cyclin E. Cell 97:767–777 Sandhu C, Chou HS, Fattaey A, Harlow E 1997 New 31. Tong W, Pollard JW 1999 Progesterone inhibits estro- functional activities for the p21 family of CDK inhibitors. gen-induced cyclin D1 and cdk4 nuclear translocation, Genes Dev 11:847–862 cyclin E- and cyclin A-cdk2 kinase activation, and cell 14. Cheng M, Olivier P, Diehl JA, Fero M, Roussel MF, Rob- proliferation in uterine epithelial cells in mice. Mol Cell erts JM, Sherr CJ 1999 The p21(Cip1) and p27(Kip1) Biol 19:2251–2264 CDK ’inhibitors’ are essential activators of cyclin D- 32. Coats S, Whyte P, Fero ML, Lacy S, Chung G, Randel E, dependent kinases in murine fibroblasts. EMBO J 18: Firpo E, Roberts JM 1999 A new pathway for mitogen- 1571–1583 dependent cdk2 regulation uncovered in p27(Kip1)- 15. Geng Y, Yu Q, Sicinska E, Das M, Bronson RT, Sicinski deficient cells. Curr Biol 9:163–173 P 2001 Deletion of the p27Kip1 gene restores normal 33. Ophascharoensuk V, Fero ML, Hughes J, Roberts JM, development in cyclin D1-deficient mice. Proc Natl Acad Shankland SJ 1998 The cyclin-dependent kinase inhibi- Sci USA 98:194–199 tor p27Kip1 safeguards against inflammatory injury. Nat 16. Tong W, Pollard JW 2001 Genetic evidence for the in- Med 4:575–580 teractions of cyclin D1 and p27(Kip1) in mice. Mol Cell 34. Hiromura K, Pippin JW, Fero ML, Roberts JM, Shankland Biol 21:1319–1328 SJ 1999 Modulation of apoptosis by the cyclin-depen- 17. Fero ML, Rivkin M, Tasch M, Porter P, Carow CE, Firpo dent kinase inhibitor p27Kip1. J Clin Invest 103:597–604 E, Polyak K, Tsai LH, Broudy V, Perlmutter RM, Kaush- 35. Masuda A, Osada H, Yatabe Y, Kozaki K, Tatematsu Y, ansky K, Roberts JM 1996 A syndrome of multiorgan Takahashi T, Hida T 2001 Protective function of hyperplasia with features of gigantism, tumorigenesis, p27(KIP1) against apoptosis in small cell lung cancer and female sterility in p27(Kip1)-deficient mice. Cell 85: cells in unfavorable microenvironments. Am J Pathol 733–744 158:87–96 18. Kiyokawa H, Kineman RD, Manova-Todorova KO, 36. Kim SG, Kim SN, Jong HS, Kim NK, Hong SH, Kim SJ, Soares VC, Hoffman ES, Ono M, Khanam D, Hayday AC, Bang YJ 2001 Caspase-mediated Cdk2 activation is a Frohman LA, Koff A 1996 Enhanced growth of mice critical step to execute transforming growth factor-1-
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Davison et al. • Role of p27 in Mammary Development Mol Endocrinol, December 2003, 17(12):2436–2447 2447
induced apoptosis in human gastric cancer cells. Onco- 41. Edwards PA, Ward JL, Bradbury JM 1988 Alteration of gene 20:1254–1265 morphogenesis by the v-myc oncogene in transplants of 37. Dyer MA, Capko CL 2001 p27Kip1 and p57Kip2 regulate mammary gland. Oncogene 2:407–412 proliferation in distinct retinal progenitor cell populations. 42. Musgrove EA, Swarbrick A, Lee CS, Cornish AL, Suth- J Neurosci 21:4259–4271 erland RL 1998 Mechanisms of cyclin-dependent kinase 38. Humphreys RC, Krajewska M, Krnacik S, Jaeger R, inactivation by progestins. Mol Cell Biol 18:1812–1825 Weiher H, Krajewski S, Reed JC, Rosen JC 1996 Apo- 43. Musgrove EA, Lee CS, Cornish AL, Swarbrick A, Suth- ptosis in the terminal endbud of the murine mammary erland RL 1997 Antiprogestin inhibition of cell cycle pro- gland: a mechanism of ductal morphogenesis. Develop- gression in T-47D breast cancer cells is accompanied by ment 122:4013–4022 induction of the cyclin-dependent kinase inhibitor p21. 39. Debnath J, Mills KR, Collins NL, Reginato MJ, Muth- Mol Endocrinol 11:54–66 uswamy SK, Brugge JS 2002 The role of apoptosis in 44. Lewis MT, Ross S, Strickland PA, Sugnet CW, Jimenez creating and maintaining luminal space within normal E, Scott MP, Daniel CW 1999 Defects in mouse mam- and oncogene-expressing mammary acini. Cell 111: mary gland development caused by conditional haploin- 29–40 sufficiency of Patched-1. Development 126:5181–5193 40. Zhang P, Wong C, DePinho RA, Harper JW, Elledge SJ 45. Bradbury JM, Edwards PA, Niemeyer CC, Dale TC 1995 1998 Cooperation between the Cdk inhibitors p27(KIP1) Wnt-4 expression induces a pregnancy-like growth pat- and p57(KIP2) in the control of tissue growth and devel- tern in reconstituted mammary glands in virgin mice. Dev opment. Genes Dev 12:3162–3167 Biol 170:5535–5563
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 414
^ REVIEW ARTICLE ^
Prolactin and the prolactin receptor: new targets of an old hormone Jessica Harris, Prudence M Stanford, Samantha R Oakes and Christopher J Ormandy
Prolactin (PRL) is one of a family of related hormones pancreas also exists (reviewed in (1, 2)). A proximal including growth hormone (GH) and placental lactogen (PL) promoter region directs PRL expression in the that are hypothesized to have arisen from a common pituitary (3), while a distal promoter controls extra- ancestral gene about 500 million years ago. Over 300 pituitary sites of expression (4). Although the different functions of PRL have been reported, highlighting majority of pituitary PRL exists as an unmodi®ed the importance of this pituitary hormone. PRL is also 23 kDa protein, a number of variants of PRL have synthesized by a number of extra-pituitary tissues including been found which are a result of alternative splicing, the mammary gland and the uterus. Most of PRL's actions proteolytic cleavage, phosphorylation, glycosylation are mediated by the unmodi®ed 23 kDa peptide, however, and other post-translational modi®cations (for review PRL may be modi®ed post-translation, thereby altering its see (5)). biological effects. PRL exerts these effects by binding to its receptor, a member of the class I cytokine receptor super- family. This activates a number of signaling pathways Splice variants resulting in the transcription of genes necessary for the tissue speci®c changes induced by PRL. Mouse knockout There is evidence for alternative splicing of the Prl models of the major forms of the PRL receptor have transcript in the brain (6) and proteins smaller than con®rmed the importance of PRL's role in reproduction. 23 kDa that cross-react with a PRL antibody are Further knockout models have provided insight into the present in the pituitary (7). One cross-reactive variant importance of PRL signaling intermediates and the advent is 21 kDa in size, has a tyrosine peptide map that of transcript pro®ling has allowed the elucidation of a resembles that of Prl (8) and is present in non- number of PRL target genes. lactating post-partum women (9). However, these proteins have not been linked to a transcript and the Keywords: prolactin; prolactin receptor majority of PRL variants result from post-transla- Ann Med 2004; 36: 414±425 tional modi®cation.
Post-translational modi®ed variants Prolactin Proteolytic cleavage of Prl occurs at sites of Prl Prolactin (PRL) is a polypeptide hormone synthesized production (5) via Cathespsin D, resulting in two and secreted primarily by the lactotroph cells of the fragments of 16 kDa and 6 kDa (10), or via kallikerin anterior pituitary. PRL is also synthesized at extra- giving rise to a 22 kDa fragment (11). The 16 kDa pituitary sites including the mammary epithelium, fragment of PRL has anti-angiogenic properties (12) placenta, uterus, brain and the immune system. and inhibits prostate tumor growth in mice, possibly Evidence for PRL production in the lacrimal gland, by suppressing blood vessel formation (13). adrenal gland, corpus luteum, prostate, testes, and Phosphorylation of PRL occurs within the secre- tory granules of the lactotroph and is mediated by p21-activated protein kinase 2 (PAK2) (14). The From the Garvan Institute of Medical Research, Darlinghurst, function of PRL phosphorylation is widely debated, NSW, Australia. and it has been shown that phosphorylated PRL Correspondence: Jessica Harris, Garvan Institute of Medical (pPRL) can have both agonistic and antagonistic Research, 384 Victoria St, Darlinghurst, 2010 NSW, Australia. effects. Prl activity is commonly measured in Nb2 Fax: 61 2 9295 8321. E-mail: [email protected] lymphoma cells that proliferate following treatment
# 2004 Taylor & Francis. ISSN 0785-3890 DOI 10.1080/07853890410033892 Annals of Medicine 36 PROLACTIN AND THE PROLACTIN RECEPTOR 415
Abbreviations and acronyms Key messages . All abbreviations are based on Unigene (http://www. PRL is a pituitary hormone that has direct ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene) effects on the development of the mammary names. Where possible, italics denote the gene and epithelium during pregnancy. non-italics the protein product. Uppercase denotes the human form while lower case denotes the rodent form. . There is growing evidence for the role of PRL 20a-HSD 20 alpha- hydroxysteroid dehydrogenase in carcinogenesis with a number of animal 3b-HSD 3-beta-hydroxysteroid dehydrogenase BRCA1 Breast cancer susceptibility gene 1 models implicating PRL as a promoter of breast C/EBP CCAAT enhancer-binding protein and prostate tumors. CL Corpus luteum . Elucidation of the transcription targets of PRL CypB Cyclophilin B ERa Estrogen receptor alpha is proceeding due to the advent of new genomic F3-SPrlr Short form 3 of the rat Prlr techniques. Gab2 Growth factor receptor bound protein 2-associated protein 2 GH Growth hormone Grb2 Growth factor receptor bound protein 2 Igf1 Insulin-like growth factor 1 Igf2 Insulin-like growth factor 2 with lactogenic hormones (15, 16). Phosphorylated Irf1 Interferon regulatory factor 1 Prl fails to stimulate proliferation of Nb2 lymphoma Jak2 Janus kinase 2 cells to the extent of non-phosphorylated Prl (17). Lcn2 Lipocalin 2 Map Mitogen activated protein kinase This inhibition of Prl stimulated proliferation by pPrl Mt Metallothionein occurs in a dose-dependent, antagonistic manner (18). Nb2 lymphoma Treatment of Nb2 lymphoma cells with a molecular Lymphoma derived from the Nb rat mimic of phosphorylated Prl (S179D) also antago- Nrl Neu-related lipocalin PAK2 p21-activated protein kinase 2 nizes Prl activity (19). Furthermore, S179D inhibits PI3 Phosphoinositide-3-kinase growth in the rat mammary gland (20) and antag- Pias3 protein inhibitor of activated Stat 3 onizes PRL induced proliferation in breast cancer cell PIN Prostatic intraepithelial neoplasia PL Placental lactogen lines (21). S179D may preferentially activate the Map Ppar Peroxisome proliferator-activated receptor kinase pathway in HC11 mouse mammary epithelial gamma cells, resulting in increased b-casein expression (22). pPRL Phosphorylated prolactin (human) pPrl Phosphorylated prolactin (rodent) An increase in b-casein expression also occurs in the PRL prolactin (human) mammary gland of S179D treated rats (20). However, Prl prolactin (rodent) S179D has also been reported to act as an agonist of / Prl prolactin wild-type Nb2 proliferation and activation of the Jak/Stat Prl / prolactin knockout Prlr prolactin receptor pathway, indicating that the action of this Prl mutant Prlr / prolactin receptor wild-type is not fully understood (23). Prlr / prolactin receptor heterozygote PRL is also modi®ed by glycosylation, deamination Prlr / prolactin receptor knockout PyMT Polyoma middle-T antigen of asparagine or glutamate residues, sulfonation of Raf proto-oncogene serine/threonine-protein tyrosine residues, polymerization and formation of kinase complexes with other molecules (reviewed in (5)), Rankl Receptor activator of nuclear factor kappa B ligand however, the physiological functions of these modi- Ras Oncogene isolated from rat sarcoma ®cations are not well characterized. S1,2,3 Short form 1, 2, and 3 of the Prlr S179D Molecular mimic of phosphorylated PRL Shc Src homology 2 domain containing trans- forming protein Prolactin receptor Socs Suppressor of cytokine signaling SRC Tyrosine protein kinase pp60-c-src The PRL receptor (PRLR) is a member of the class I Stat Signal transducer and activator of tran- scription cytokine receptor superfamily that includes the Stat5 Signal transducer and activator of tran- growth hormone receptor, leptin receptor, erythro- scription 5 poietin receptors and receptors for several interleu- Stat5a Signal transducer and activator of tran- scription 5, subunit a kins (24). Alternative splicing of the PRLR gene leads Stat5b Signal transducer and activator of tran- to multiple isoforms which differ in the length and scription 5, subunit b composition of their cytoplasmic tail and are referred SV40T Simian virus 40 large T antigen to as the short (291aa), and long (591aa) PRLR iso- TEB Terminal end bud Tnsf11 Tumor necrosis factor (ligand) superfamily, forms. Both isoforms exist in the rat and human, member 11 while three short and one long form are present in the mouse (reviewed in (2, 25)). In human and cancer tissues an additional receptor isoform with a deleted
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 416 HARRIS . STANFORD . OAKES . ORMANDY extracellular domain, in addition to two alternatively but not proliferation. The discrepancy in severity spliced short isoforms exists (reviewed in (26)). A between Stat5 / and Prlr / mammary gland soluble PRLR isoform is present in human serum and phenotypes may be explained by the ability of other milk and serves as a PRL binding protein (27). hormones and growth factors to stimulate Stat5a activation. Growth hormone and epidermal growth factor are able to stimulate Stat5a activation in PRLR activation and signaling Prlr / mammary epithelium (31), a scenario that PRL, as well as placental lactogen and primate may also apply to activation of Jak2. growth hormone, binds the PRLR. The ®rst step in Tyrosine phosphorylation sites in the Prlr serve as receptor activation is the binding of a single ligand to docking sites for the adaptor proteins Src homology 2 the receptor resulting in receptor dimerization domain containing transforming protein (Shc) (32), (reviewed in (2, 25) which activates a number of Growth factor receptor bound protein 2 (Grb2) (33) signaling cascades through which PRL exerts its and growth factor receptor bound protein 2-asso- effects. ciated protein 2 (Gab2) (34), resulting in binding and Dimerization of the Prlr induces phosphorylation activation of Ras and Raf and subsequently, the of Janus kinase 2 (Jak2), which in turn phos- mitogen activated protein (Map) kinase pathway. phorylates the receptor. Although animals lacking Activation of the Map kinase pathway may mediate Jak2 are embryonic lethal, transplantation of the the effect of Prl on proliferation of mammary embryonic mammary anlage to mammary fat pads of epithelial cells (35). Prlr also facilitates docking of adult mice enabled examination of the role of Jak2 SRC family kinases that in turn activate protein during mammary gland development. Mammary kinase B and the phosphatidylinositol 3-kinase (PI3 glands containing Jak2 / epithelium in virgin kinase) pathway (reviewed in (26)). Activation of this animals develop normally, however at parturition pathway mediates Prl's anti-apoptotic (36) and cell the mammary glands fail to develop lobuloalveoli. proliferation effects (37). Furthermore, terminal end buds (TEBs) persist at the Attenuation of Prlr signaling is mediated by ends of ducts indicating an additional defect in the members of the suppressor of cytokine signaling earlier stage of differentiation to alveolar buds. This (Socs) gene family, in the case of Socs1 and Socs3 by phenotype is accompanied by a reduction in prolif- inhibition of the Jak/Stat pathway (38). Socs1 / eration, maintenance of ductal cell markers and lack mice rescued from embryonic lethality by deletion of of secretory cell markers in Jak2 / mammary the interferon gamma gene exhibit accelerated lobu- epithelium (28, 29). loalveolar development characterized by an increase Signal transducer and activator of transcription in lobuloalveoli, dramatic enlargement of the lumen (Stat) docks at the tyrosine residues of the Prlr and an increase in milk protein expression and Stat5 phosphorylated by Jak2 and is itself phosphorylated phosphorylation during pregnancy and lactation. by Jak2. The activated Stat molecules dissociate from Deletion of a single Socs1 allele is able to restore the receptor, form heterodimers and translocate to the lobuloalveolar development in Prlr / mammary nucleus where they activate gene transcription. Abla- epithelium, demonstrating that Socs1 is a key regu- tion of Stat5a in mice results in impaired mammary lator of Prl signaling in the mammary gland (39). gland development due to reduced formation of lobuloalveoli (30). This knockout model demon- strated a non-redundant role for Stat5a considering the expression of its closely related and co-expressed isoform Stat5b. Stat5 null mammary epithelium (in PRL nuclear action which both Stat5a and Statb isoforms are knocked out) develops normally in virgin animals but displays Prl and a number of other polypeptide hormones are a lack of lobuloalveolar development at parturition, translocated to the nucleus (reviewed in (40)) where demonstrating a role for Stat5b in mammary gland they exert a number of effects. The Prlr is constitu- development. This phenotype appeared more extreme tively expressed in the nucleus of Nb2 lymphoma cells than the Prlr / mammary phenotype due to the where Prl is translocated following internalization presence of a closed epithelial lumen and aberrant (41), assisted by tyrosine kinase and protein kinase C cell-cell contacts (31). Stat5 null epithelium retains activation (42). Once internalized, Prl interacts with markers of ductal epithelium and fails to express the peptidyl prolyl isomerase cyclophilin B (CypB), markers of secretory epithelium (29, 31). The Prlr / facilitating its nuclear retrotransport (43). Further- mammary epithelium has a greater defect in prolif- more, this Prl/CypB complex induces transcription by eration in response to estrogen and progesterone than interacting with Stat5 and enhancing its function by the Stat5 / mammary epithelium (31), indicating mediating release of the Stat-repressor protein inhi- that Stat5 mediates the effect of Prl on differentiation, bitor of activated Stat3 (Pias3) (44).
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 PROLACTIN AND THE PROLACTIN RECEPTOR 417
Tissues targeted by prolactin puberty, however secondary side branching does not occur as the animals age and the ducts become more Reproductive system distended (53, 54). Closer analysis revealed a further defect in virgin animals. In wild-type animals ductal Ablation of the Prlr in mice results in a number of elongation results from the formation of terminal end / reproductive defects. Female Prlr animals are buds (TEBs), specialized structures with high rates of unable to maintain pseudopregnancy, have a dis- cell division at their invasive leading edge and a rupted estrus cycle and are sterile despite regular central zone of apoptotic cells that produce the / mating, due to the inability of the Prlr uterus to canalized duct. This structure differentiates into a provide an environment conducive to blastocyst terminal alveolar bud once ductal elongation has development and implantation (45). Similar repro- ceased. In mature Prlr / animals (46) a unique TEB- / ductive defects are seen in Prl female mice (46). like structure persists. This structure has the overall / Serum progesterone levels in Prlr mice are sub- shape of a TEB but lacks the cap cell layer and stantially reduced due to the inability of Prl to multiple layers of highly mitotic and apoptotic cells, maintain trophic support for the corpora lutea. demonstrating its quiescent state and suggesting a Restoration of progesterone levels enabled egg devel- defect in the ®nal stage of differentiation (54). opment and implantation (47), however, restoration Progesterone acts to promote side-branching in / of sex steroid levels in Prlr animals was not able to virgin animals (55) and reduced progesterone levels in maintain pregnancy in all animals (47) indicating a Prlr / and Prl / animals results in reduced ductal role for prolactin in maternal decidual transforma- side branching in the mammary gland of virgin tion. animals. Progesterone treatment restores the defect / Maternal behavior in Prl mice is normal (46), in ductal side branching in Prlr / virgin animals however pup-induced maternal behavior in both (47, 56) and Prl / animals (57). / / virgin and pregnant Prlr or Prlr mice is Transplantation of Prlr / mammary epithelium substantially reduced, independent of other behavior- to the mammary fat pad of an immunocompromized al activities and olfactory function, implicating the but otherwise normal animal allows examination of Prlr as an important regulator of maternal behavior Prlr / mammary gland development in a normal (48). The molecular mimic of phosphorylated Prl, endocrine environment, allowing development during S179D, is able to delay the onset of normal maternal pregnancy to be assessed. The host fat pad is cleared behavior in rats (49). of endogenous epithelium and the transplanted / A proportion (20%) of Prlr males are infertile, epithelium is allowed to penetrate and ®ll the fat however this is not due to a defect in mating behavior pad (58). Transplanted Prlr / epithelium develops / (45). Two subsequent studies using separate Prlr side branches in virgin animals, as the epithelium mouse lines that originated from this initial study experiences normal levels of sex steroid hormones, population, but which have since diverged, have re- however during pregnancy, a complete failure of examined this phenotype. One study found no effect lobuloalveolar development is observed and only on fertility (50), while the other showed that 10% of alveolar-buds are formed (54). This illustrates a direct / Prlr males are infertile and those that are fertile role for epithelial Prlrs in mediating the formation of only have a 40% chance of producing a successful ®rst the milk producing structures during pregnancy. In pregnancy (51). These contrasting reports may be due agreement, the mammary glands of Prlr / animals to the effect of Prl on male reproduction being that could maintain pregnancy following estrogen modi®ed by divergence in the genetic background of and progesterone treatment, also experience failure of / the Prlr model, which was derived on a mixed lobuloalveolar development (47). / 129SvPas/129OlaHsd background. Prl males are A cell autonomous function for a number of not infertile, although Prl does play a role in hormone receptors has been demonstrated in the reproductive neuroendocrine function by control of mammary stroma (59±63). Prlrs are present in the leutinizing hormone release (52). Overall, it is clear mammary stroma (64, 65), but play no direct role in that Prl exerts, at most, a modest effect on male mammary gland development in virgin or pregnant fertility. However, the existence of this effect points animals as demonstrated by the normal development to its utility at some point in the evolutionary history that occurs in mammary glands consisting of Prlr / of the mouse. In contrast, Prl is essential for many stroma and Prlr / epithelium (56). This technique aspects of female fertility and lactation. however, does not allow an analysis of lactation as the ductal network is not connected to a nipple, and so engorgement-induced involution of the gland Mammary gland begins following parturition. As a consequence we In mammary glands from Prlr / mice, ductal do not know if stromal Prlrs are required for elongation occurs normally as the animals enter lactogenesis or the maintenance of lactation.
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 418 HARRIS . STANFORD . OAKES . ORMANDY
Two processes must occur within the alveolar bud pregnancy (45), restored mammary gland develop- to form functional lobuloalveoli during pregnancy: 1) ment at the level of proliferation and differentiation, proliferation to increase cell number so as to provide indicating that this form of the receptor acts in a greater epithelial surface area for milk production and similar way to the long form (69). The function of secretion (lobuloalveolar development), and 2) ®nal S2Prlr has not been investigated to date. stage differentiation to allow the cells to produce and The degeneracy of the class I cytokine receptor secrete milk (lactogenesis). Activation of the Prlr is superfamily is demonstrated by expression of a vital for proliferation in the mammary gland. This is chimeric receptor consisting of the extracellular demonstrated by the inability of Prlr / epithelium to domain of the Prlr and the trans-membrane and proliferate to form lobuloalveoli in transplants (54) cytoplasmic domains of the erythropoietin receptor in and in the Prlr / mammary gland following treat- Prlr / mammary epithelial cells in vivo. The rescue ment with estrogen and progesterone (66). In con- of failed mammary gland development in trans- trast, mice heterozygous for the Prlr show a block in planted Prlr / mammary epithelium by this chimeric mammary development at the later lactogenesis stage, receptor indicates that the receptors may mediate the indicating a defect in differentiation. Prlr / mice same signaling events that lead to tissue-speci®c gene mated at 8 weeks of age have less than a 50% chance expression and developmental outcome (70). of successfully nursing their young, although these S179D, the molecular mimic of phosphorylated Prl, mothers have a greater chance of successful lactation acts as a repressor of mammary epithelial ductal following their second pregnancy (45). This failure of growth but activates b-casein expression when lactation is associated with a decrease in maternal administered to rats. Lactation is prevented in mice behavior (48), however, the phenotype is intrinsic to by S179D administration (Naylor MJ, Ormandy CJ et the Prlr / mammary epithelium as it is recapitulated al. in preparation). Hence, the ratio of phosphory- by transplantation of the mammary epithelium to a lated to unmodi®ed Prl during mammary gland wild-type mammary fat pad cleared of endogenous development may be crucial for the balance between epithelium (56). The mammary glands of mothers growth and differentiation within the mammary that fail to lactate show the formation of lobules, but gland (20). they are unable to form an open lumen or secrete milk (45, 54). The Prlr / phenotype suggests that the level Mammary gland carcinogenesis of signaling ¯ux initiated by Prl can modulate mammary gland development, with the later devel- Prl has a profound positive effect on mammary gland opmental stage of lactogenesis requiring high levels of carcinogenesis in rodent models (71). Results from Prlr signaling that cannot be met from a single more recent studies of a number of transgenic and functional Prlr allele. knockout mouse models are consistent with a role for Prl produced by the mammary gland may also play Prl in mammary carcinogenesis. Polyoma middle-T a role in proliferation of mammary epithelial cells antigen (PyMT) induced tumors in Prl / mice are during lactogenesis. Prl / mammary epithelium detected an average of 9 days later than those in wild- transplanted to a wild-type cleared mammary fat type littermates (57) and transgenic animals expres- pad displayed a decrease in proliferation despite sing Prl from the metallothionein promoter (Mt-Prl) normal appearance of whole mount and thin-section develop mammary tumors (72). These studies how- histology (67). This suggests that mammary-produced ever, did not take into account the altered systemic Prl has a role in sustaining milk production, rather hormonal environment of the knockout and trans- than development of the gland during pregnancy. genic animals. In the overexpression models, prolac- In the mouse, alternative splicing of the Prlr gene tin may be acting via the pituitary/ovarian axis to results in four forms of the Prl receptor each having a induce estrogen and progesterone secretion, thus cytoplasmic domain of different length. The function affecting carcinogenesis indirectly. Transplantation of the long form of the receptor is described above, of the mammary epithelium into the mammary fat however the functions of the three short forms of the pads of host animals with normal hormonal environ- receptor (S1-3Prlr) are not well characterized. Trans- ments will determine if Prl is acting directly on the genic expression of the short form of the rat Prlr mammary epithelium to promote mammary carcino- (F3-SPrlr) in the mammary gland results in reduced genesis. lobuloalveolar development, Stat5 phosphorylation Prl is produced by Mt-Prl transgenic mammary and milk protein gene expression but no change in epithelial tumor cells in culture and is able to promote proliferation (68). This suggests that this form of the lobuloalveolar development in an explant culture receptor may act as a dominant negative to control system, indicating that Prl may be acting in an the rate of differentiation occurring in the mammary autocrine/paracrine manner to promote carcino- gland. Transgenic expression of the S1Prlr in Prlr / genesis (72). This observation was supported by the mice, normally unable to lactate following their ®rst development of a transgenic mouse expressing Prl
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 PROLACTIN AND THE PROLACTIN RECEPTOR 419 speci®c to the mammary gland under control of the ity, family history and exposure to exogenous hormone-independent neu-related lipocalin 2 (Lcn2/ hormones (82). A number of studies have correlated Nrl) promoter. Mammary neoplasms develop in the the relationship between these risk factors and levels majority of Nrl-Prl transgenic animals independent of of PRL in the serum. PRL levels in women decrease circulating Prl levels, and Prl stimulates proliferation following ®rst pregnancy and each subsequent preg- in Nrl-Prl transgenic preneoplastic mammary glands nancy (83) suggesting a role for PRL in promoting (73). However, an undetected increase in serum breast cancer in nulliparous women. Women taking prolactin may occur in these animals, below the the oral contraceptive pill (OCP) have a small sensitivity of the available assays for serum prolactin increased risk of breast cancer diagnosis (84) and levels and/or in a non-pulsatile pattern of secretion. increased secretion of prolactin from the pituitary Conversion from pulsatile to continuous growth (85). Age at ®rst pregnancy is the only factor hormone secretion is suf®cient to induce acromegaly differentiating malignant from benign disease (86) without serum growth hormone levels exceeding the and the relative risk of breast cancer is higher in normal range (74), indicating that normal range but parous women who have never breastfed (87). As PRL continuous Prl secretion could have dramatic effect on is important for establishing terminal differentiation mammary carcinogenesis. Transgenic mouse models of the mammary gland during pregnancy and lacta- have generally failed to reproduce the presence of ER tion, it is possible that PRL action plays a protective positive and negative breast tumors seen in the human role in the normal breast and has a role in prolifera- population, as tumors from transgenic mouse models tion at high serum levels and in breast cancer. are generally ER negative. For example, ERa is Alternatively, pregnancy alters subsequent hormone expressed in the normal mammary epithelium of levels, reducing the risk of breast cancer. The asso- C3(1)/SV40 large T-antigen transgenic mice, but it is ciation between these risk factors and prolactin is dramatically decreased in tumors (75). The Nrl-Prl circumstantial and a strong case can be made for transgenic mouse is able to produce tumors that are many other hormones, especially estrogen, in the either positive or negative for ERa, better modeling underlying mechanisms behind these risk factors. the human disease (73). Although the role for Prl in mammary carcinoma in Prostate rodents is well established, its role in humans has been dismissed by some on the basis of the lack of PRL and the PRLR are expressed in human and rat ef®cacy in late stage patients of drugs that lower prostate epithelial cells where their level is increased serum prolactin. There is now strong evidence for a by androgen treatment (88). Prl has been viewed as an causal role for PRL in human breast cancer. Recent autocrine/paracrine growth factor (89) or a survival reviews have highlighted the body of work providing factor (90) for prostate epithelial cells in vitro. evidence for the role of PRL in breast cancer Expression of Prl in transgenic mice from the (reviewed in (26, 76±78)). Speci®cally, serum PRL metallothionein promoter results in systemic hyper- levels positively correlate with breast cancer risk in prolactinemia accompanied by enlargement of the post-menopausal women as identi®ed by the large prostate gland due to an increase in collagenous scale prospective Nurses Health Study. Women stroma, secretion volume and hyperplasia of the within the top quartile serum PRL level have a two glandular epithelium (91). Prostate-speci®c Prl fold increased risk of developing breast cancer, expression from the probasin promoter in mice independent of circulating estrogen levels (79). A results in an identical phenotype but does not result separate study showed that high serum PRL levels in the elevated serum Prl, androgen and insulin-like (>20 ng/ml) correlates with tumor size, stage and growth factor 1 (Igf1) levels seen in the mice utilizing lymph node status and those women with high serum the metallothionein promoter. The increase in pros- PRL levels have a greater risk of developing meta- tate ductal branching and subsequent enlargement of static disease (80). Three quarters of tumors screened the prostate gland in these animals is thus Prl speci®c in this study express PRL, which correlates with (92). expression of the PRLR (80), suggesting that PRL may Prostate weight is slightly increased in young play an autocrine/paracrine role in breast cancer in Prlr / animals and subtle changes in epithelial and addition to its endocrine effects. Furthermore, the stromal content of the dorsal lobe occur, probably PRLR protein is expressed in over two thirds of breast regulated by cooperation between Prl and androgens cancers (81). (51). The phenotypes of the knockout model com- Currently the mechanism by which prolactin pared to the transgenic model indicate that Prl may affects breast cancer is unknown. There are a number have a subtle role in the prostate at physiological of risk factors for developing breast cancer, including levels, but that hyperprolactinemia has profound and age, geographical location, age at menarche, age at direct effect on hyperplasia of the prostate. menopause, age at ®rst full-term pregnancy, nullipar- In rats treated with S179D the prostate epithelium
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 420 HARRIS . STANFORD . OAKES . ORMANDY gains a more differentiated phenotype, while changes in the way Prl regulates weight gain (101). Prl also in the epithelium of animals treated with unmodi®ed plays a subtle role in establishing sexual dimorphism Prl are consistent with an effect on proliferation (93). in lacrimal glands. Similar effects are seen in the harderian gland, accompanied by cessation of por- phyrin accretion in Prlr / animals (102). Prlr / Prostate carcinogenesis mice have reduced islet numbers and b-cell mass in Atypical glandular epithelium is present in the the pancreas and experience a reduced insulin secre- prostates of transgenic animals expressing Prl from tion in response to glucose, accompanied by a modest the metallothionein or probasin promoters. These increase in insulin resistance (103), establishing a regions have morphological similarities to low grade physiological role for Prl in b-cell development. prostatic intraepithelial neoplasia (PIN), a lesion seen The subtlety of most of the phenotypes seen in infrequently in wild-type animals (92) and thought to Prlr / and Prl / mice illustrates the way Prl works be the precursor lesion of prostate cancer. In Prlr / in concert with other hormones and growth factors to animals transgenic for the simian virus 40 large T exert its effects. It also highlights the major role that antigen (SV40T) oncogene, the area of PIN is Prl plays in mammary gland development and the decreased and no tumors present in the ventral lobe. potential for developing speci®c breast cancer ther- This demonstrates a role for Prl in the establishment apeutics against molecules downstream of Prl action of preneoplastic regions and indicates that Prl may in the mammary gland. also play a role in progression of tumors in the prostate (51). PRL treatment of human prostate cancer cell lines Transcriptional targets of prolactin increases their proliferation (94, 95). The establish- ment rate of human prostate cancer xenografts in rats Mammary gland treated with S179D is substantially reduced, as is the growth rate of the tumors. This indicates that the Stat5 is known to activate transcription from the phosphorylated form of the receptor may act as an b-casein promoter (104, 105), the b-lactoglobulin antagonist of the growth promoting effects of un- promoter (106) and the whey acidic promoter (107) modi®ed PRL on the prostate (96). in a Prl dependent manner. Despite the well estab- lished role of Prl in mammary gland differentiation and milk protein production, only a few studies have Others tissues been able to elucidate other genes transcribed by Prl As the Prlr is a member of the cytokine receptor that exert its effects on proliferation and differentia- superfamily it is thought that Prl may have a role in tion. Genes expressed following treatment of mouse lymphohemopoietic development. There is no detect- mammary epithelial cells with Prl include the p100 co- able change in the cellular composition of the activator (108) and the CBP/p300 transactivator lymphoid organs of Prlr / mice. Furthermore, Cited4 (109). PRL also increases transcription of Prlr / mice are able to mount a normal immune BRCA1 in MCF-7 and T47-D cells (110) and activates response against pathogens (97) and have normal B the cyclin D1 promoter via the JAK2/STAT5 signal- and T cell development and myelopoiesis (46). ing pathway in T47D cells (111). However, elevated Prl is able to maintain survival Comprehensive studies have been performed on the of T-lymphocytes under conditions of increased Prl mediated regulation of two genes important for glucocorticoids, suggesting a novel role for Prl in mammary gland development, Igf2 and Rankl. Igf2 the immune system during times of stress (98). expression is decreased in Prlr / mammary epithe- Loss of the Prlr in bone slows bone formation with lium during pregnancy (112, 113) but remains at a no change in bone resorption, leading to an overall level similar to wild-type mammary glands in cyclin loss of bone mass. Although this phenotype may be D1 / mammary epithelium, which have a defect in indirect via systemic endocrine effects, it is interesting lobuloalveolar development similar to that seen in the to note that osteoblasts express the Prlr while Prlr / epithelium (112). This indicates that Igf2 acts osteoclasts do not (99). downstream of Prl but upstream of CyclinD1. Igf2 In female mice lacking the Prlr an advanced ®rst expression is also induced by Prl in primary mam- molt abolishes the sexual dimorphism seen in pelage mary epithelial cells in culture (112), in cultured replacement, and both genders have longer and pregnant mammary gland explants, and in HC11 coarser hair (100). Weight gain in Prlr de®cient mice mammary epithelial cells (113). Igf2 is able to induce decreases after the age of 16 weeks compared to their lobuloalveolar development in cultured mammary wild-type littermates. This is due to reduced abdom- gland explants (113) and ablation of a single Igf2 allele inal fat stores and reduced plasma leptin levels in in mouse mammary epithelium decreases lobulo- female Prlr / mice, indicating a sexual dimorphism alveolar development (112), con®rming that Igf2 has
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 PROLACTIN AND THE PROLACTIN RECEPTOR 421 a Prl-like role in the mammary epithelium during feron-regulatory factor 1 (Irf1) is transcribed in T pregnancy. Ectopic expression of Igf2 in Prlr / cells treated with Prl (121). In the reproductive system mammary epithelium partially restores lobuloalveo- PRL induces transcription of the estrogen receptor lar development (112) con®rming that Igf2 acts (122) and 3b-hydroxysteroid dehydrogenase (3b- downstream of Prl to induce lobuloalveolar develop- HSD), a molecule involved in progesterone synthesis ment. in the corpus luteum (CL) during pregnancy (123). Prl Tnsfs11 or Rankl expression decreases in Prlr / can also negatively regulate transcription as it mammary epithelium during early pregnancy (56) and represses expression of 20 alpha-hydroxysteroid Rankl / animals display a lack of alveolar develop- dehydrogenase (20a-HSD), a molecule known to play ment in the mammary gland during pregnancy (114). a role in the termination of pregnancy in the rat (124). Furthermore, Rankl mRNA is undetectable in mam- Finally, Prl treatment stimulates the transcription of mary glands from Prl / animals and Prl is able to the transcription factors CCAAT enhancer-binding induce expression of Rankl in primary mammary protein (Cebp) and peroxisome proliferator-activated epithelial cells in a Stat5a/Jak2 dependent manner receptor (Ppar ) in NIH-3T3 cells (125). (115). A number of transcript pro®ling studies have Transcript pro®ling of Prlr / and Prlr / mam- recently been undertaken using models of PRL action mary epithelium transplanted to a wild-type fat pad that should reveal a large number of candidates during early pregnancy demonstrated a decrease in potentially regulated by PRL at the level of transcrip- expression in the Prlr / epithelium of a number tion (reviewed in (119)). molecules important for mammary epithelial cell function. These molecules included cytokeratins, cell adhesion molecules, transcription factors, compo- Conclusions nents of cell junctions and a number of growth factors (56) previously shown to be important for The construction of a number of knockout and mammary gland development (114, 116±118). A simi- transgenic animal models has allowed the elucidation lar study also found mRNA encoding for components of PRL target tissues and molecules. PRL has a of the cytoskeleton and the extracellular matrix to be profound and direct effect on the mammary gland decreasing in Prlr / mammary epithelial transplants during pregnancy and in almost all other aspects of (112), con®rming a role for prolactin in the regulation female reproduction. PRL also plays a role in breast of transcription of these components of a mammary and prostate carcinogenesis, and therapeutic strate- epithelial cell. gies targeting this signaling system may be of bene®t in the future. Modi®ed PRL molecules are rapidly being identi®ed and should enable a greater under- Other tissues standing of how PRL mediates its effects. The A number of genes are transcribed following PRL discovery of a non-toxic small molecule antagonist action on other target tissues (119). In cultured of the PRL receptor signaling pathway would allow hepatocytes Prl regulates transcription of the Na / the role of this signaling system in cancer to be taurocholate cotransporting polypeptide (120). Inter- thoroughly tested.
References
1. Ben-Jonathan N, Mershon JL, Allen DL, Steinmetz RW. tively spliced prolactin messenger RNA. Mol Endocrinol Extrapituitary prolactin: distribution, regulation, functions, 1992;6:35±42. and clinical aspects. Endocr Rev1996;17:639±69. 7. Sinha YN, Ott K, Vanderlaan WP. Detection of multiple 2. Freeman ME, Kanyicska B, Lerant A, Nagy G. Prolactin: PRL- and GH-like proteins in human pituitary by Western structure, function, and regulation of secretion. Physiol Rev blot analysis. Am J Med Sci 1987;294:15±25. 2000;80:1523±631. 8. Sinha YN, Jacobsen BP. Structural and immunologic evi- 3. Berwaer M, Monget P, Peers B, Mathy-Hartert M, Bellefroid dence for a small molecular weight 21K variant of prolactin. E, Davis JR, et al. Multihormonal regulation of the human Endocrinology 1988;123:1364±70. prolactin gene expression from 5000 bp of its upstream 9. Liu JH, Lee DW, Markoff E. Differential release of prolactin sequence. Mol Cell Endocrinol 1991;80:53±64. variants in postpartum and early follicular phase women. J 4. Berwaer M, Martial JA, Davis JR. Characterization of an Clin Endocrinol Metab 1990;71:605±10. up-stream promoter directing extrapituitary expression of 10. Baldocchi RA, Tan L, King DS, Nicoll CS. Mass spectro- the human prolactin gene. Mol Endocrinol 1994;8:635±42. metric analysis of the fragments produced by cleavage and 5. Sinha YN. Structural variants of prolactin: occurrence and reduction of rat prolactin: evidence that the cleaving enzyme physiological signi®cance. Endocr Rev1995;16:354±69. is cathepsin D. Endocrinology 1993;133:935±8. 6. Emanuele NV, Jurgens JK, Halloran MM, Tentler JJ, 11. Anthony PK, Powers CA. Characterization of cysteamine Lawrence AM, Kelley MR. The rat prolactin gene is induction of the 22K prolactin variant in the rat pituitary. expressed in brain tissue: detection of normal and alterna- Neuroendocrinology 1993;57:167±76.
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 422 HARRIS . STANFORD . OAKES . ORMANDY
12. Clapp C, Martial JA, Guzman RC, Rentier-Delure F, Weiner markers: a tool to de®ne developmental lesions. J Histochem RI. The 16-kilodalton N-terminal fragment of human Cytochem 2003;51:555±65. prolactin is a potent inhibitor of angiogenesis. Endocrinology 30. Liu X, Robinson GW, Wagner KU, Garrett L, Wynshaw- 1993;133:1292±9. Boris A, Hennighausen L. Stat5a is mandatory for adult 13. Kim J, Luo W, Chen DT, Earley K, Tunstead J, Yu-Lee LY, mammary gland development and lactogenesis. Genes Dev et al. Antitumor activity of the 16-kDa prolactin fragment in 1997;11:179±86. prostate cancer. Cancer Res 2003;63:386±93. 31. Miyoshi K, Shillingford JM, Smith GH, Grimm SL, Wagner 14. Tuazon PT, Lorenson MY, Walker AM, Traugh JA. p21- KU, Oka T, et al. Signal transducer and activator of activated protein kinase gamma-PAK in pituitary secretory transcription (Stat) 5 controls the proliferation and differ- granules phosphorylates prolactin. FEBS Lett 2002;515:84±8. entiation of mammary alveolar epithelium. J Cell Biol 2001; 15. Gout PW, Beer CT, Noble RL. Prolactin-stimulated growth 155:531±42. of cell cultures established from malignant Nb rat lympho- 32. Erwin RA, Kirken RA, Malabarba MG, Farrar WL, Rui H. mas. Cancer Res 1980;40:2433±6. Prolactin activates Ras via signaling proteins SHC, growth 16. Tanaka T, Shiu RP, Gout PW, Beer CT, Noble RL, Friesen factor receptor bound 2, and son of sevenless. Endocrinology HG. A new sensitive and speci®c bioassay for lactogenic 1995;136:3512±8. hormones: measurement of prolactin and growth hormone in 33. Das R, Vonderhaar BK. Involvement of SHC, GRB2, SOS human serum. J Clin Endocrinol Metab 1980;51:1058±63. and RAS in prolactin signal transduction in mammary 17. Wicks JR, Brooks CL. Biological activity of phosphorylated epithelial cells. Oncogene 1996;13:1139±45. and dephosphorylated bovine prolactin. Mol Cell Endocrinol 34. Ali S. Recruitment of the protein-tyrosine phosphatase SHP- 1995;112:223±9. 2 to the C-terminal tyrosine of the prolactin receptor and to 18. Wang YF, Walker AM. Dephosphorylation of standard the adaptor protein Gab2. J Biol Chem 2000;275:39073±80. prolactin produces a more biologically active molecule: 35. Das R, Vonderhaar BK. Activation of raf-1, MEK, and MAP evidence for antagonism between nonphosphorylated and kinase in prolactin responsive mammary cells. Breast Cancer phosphorylated prolactin in the stimulation of Nb2 cell Res Treat 1996;40:141±9. proliferation. Endocrinology 1993;133:2156±60. 36. Tessier C, Prigent-Tessier A, Ferguson-Gottschall S, Gu Y, 19. Chen TJ, Kuo CB, Tsai KF, Liu JW, Chen DY, Walker AM. Gibori G. PRL antiapoptotic effect in the rat decidua Development of recombinant human prolactin receptor involves the PI3K/protein kinase B-mediated inhibition of antagonists by molecular mimicry of the phosphorylated caspase-3 activity. Endocrinology 2001;142:4086±94. hormone. Endocrinology 1998;139:609±16. 37. Fresno Vara JA, Caceres MA, Silva A, Martin-Perez J. Src 20. Kuo CB, Wu W, Xu X, Yang L, Chen C, Coss D, et al. family kinases are required for prolactin induction of cell Pseudophosphorylated prolactin (S179D PRL) inhibits proliferation. Mol Biol Cell 2001;12:2171±83. growth and promotes beta-casein gene expression in the rat 38. Pezet A, Favre H, Kelly PA, Edery M. Inhibition and mammary gland. Cell Tissue Res 2002;309:429±37. restoration of prolactin signal transduction by suppressors 21. Schroeder MD, Brockman JL, Walker AM, Schuler LA. of cytokine signaling. J Biol Chem 1999;274:24497±502. Inhibition of prolactin (PRL)-induced proliferative signals in 39. Lindeman GJ, Wittlin S, Lada H, Naylor MJ, Santamaria M, breast cancer cells by a molecular mimic of phosphorylated Zhang JG, et al. SOCS1 de®ciency results in accelerated PRL, S179D-PRL. Endocrinology 2003;144:5300±7. mammary gland development and rescues lactation in 22. Wu W, Coss D, Lorenson MY, Kuo CB, Xu X, Walker AM. prolactin receptor-de®cient mice. Genes Dev2001;15:1631±6. Different biological effects of unmodi®ed prolactin and a 40. Clevenger CV, Rycyzyn MA. Translocation and action of molecular mimic of phosphorylated prolactin involve differ- polypeptide hormones within the nucleus. Relevance to ent signaling pathways. Biochemistry 2003;42:7561±70. lactogenic transduction. AdvExp Med Biol 2000;480:77±84. 23. Bernichtein S, Kinet S, Jeay S, Llovera M, Madern D, Martial 41. Rao YP, Olson MD, Buckley DJ, Buckley AR. Nuclear co- JA, et al. S179D-human PRL, a pseudophosphorylated localization of prolactin and the prolactin receptor in rat human PRL analog, is an agonist and not an antagonist. Nb2 node lymphoma cells. Endocrinology 1993;133:3062±5. Endocrinology 2001;142:3950±63. 42. Rao YP, Buckley DJ, Olson MD, Buckley AR. Nuclear 24. Bazan JF. A novel family of growth factor receptors: a translocation of prolactin: collaboration of tyrosine kinase common binding domain in the growth hormone, prolactin, and protein kinase C activation in rat Nb2 node lymphoma erythropoietin and IL-6 receptors, and the p75 IL-2 receptor cells. J Cell Physiol 1995;163:266±76. beta-chain. Biochem Biophys Res Commun 1989;164:788±95. 43. Rycyzyn MA, Clevenger CV. Role of cyclophilins in 25. Bole-Feysot C, Gof®n V, Edery M, Binart N, Kelly PA. somatolactogenic action. Ann N Y Acad Sci 2000;917:514±21. Prolactin (PRL) and its receptor: actions, signal transduction 44. Rycyzyn MA, Clevenger CV. The intranuclear prolactin/ pathways and phenotypes observed in PRL receptor knock- cyclophilin B complex as a transcriptional inducer. Proc Natl out mice. Endocr Rev1998;19:225±68. Acad Sci USA 2002;99:6790±5. 26. Clevenger CV, Furth PA, Hankinson SE, Schuler LA. The 45. Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, role of prolactin in mammary carcinoma. Endocr Rev2003; Buteau H, et al. Null mutation of the prolactin receptor gene 24:1±27. produces multiple reproductive defects in the mouse. Genes 27. Kline JB, Clevenger CV. Identi®cation and characterization Dev1997;11:167±78. of the prolactin-binding protein in human serum and milk. J 46. Horseman N, Zhao W, Montecino-Rodriquez E, Tanaka M, Biol Chem 2001;276:24760±6. Nakashima K, Engle S, et al. Defective mammopoiesis, but 28. Shillingford JM, Miyoshi K, Robinson GW, Grimm SL, normal hematipoiesis, in mice with targeted disruption of the Rosen JM, Neubauer H, et al. Jak2 is an essential tyrosine prolactin gene. EMBO J 1997;16:101±110. kinase involved in pregnancy-mediated development of 47. Binart N, Helloco C, Ormandy CJ, Barra J, Clement-Lacroix mammary secretory epithelium. Mol Endocrinol 2002; 16: P, Baran N, et al. Rescue of preimplantatory egg develop- 563±70. ment and embryo implantation in prolactin receptor-de®- 29. Shillingford JM, Miyoshi K, Robinson GW, Bierie B, Cao Y, cient mice after progesterone administration. Endocrinology Karin M, et al. Proteotyping of mammary tissue from trans- 2000;141:2691±7. genic and gene knockout mice with immunohistochemical 48. Lucas BK, Ormandy CJ, Binart N, Bridges RS, Kelly PA.
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 PROLACTIN AND THE PROLACTIN RECEPTOR 423
Null mutation of the prolactin receptor gene produces a 66. Grimm SL, Seagroves TN, Kabotyanski EB, Hovey RC, defect in maternal behavior. Endocrinology 1998;139:4102±7. Vonderhaar BK, Lydon JP, et al. Disruption of steroid and 49. Bridges R, Rigero B, Byrnes E, Yang L, Walker A. Central prolactin receptor patterning in the mammary gland corre- infusions of the recombinant human prolactin receptor lates with a block in lobuloalveolar development. Mol antagonist, S179D-PRL, delay the onset of maternal behavior Endocrinol 2002;16:2675±91. in steroid-primed, nulliparous female rats. Endocrinology 67. Naylor MJ, Lockefeer JA, Horseman ND, Ormandy CJ. 2001;142:730±9. Prolactin regulates mammary epithelial cell proliferation via 50. Binart N, Melaine N, Pineau C, Kercret H, Touzalin AM, autocrine/paracrine mechanism. Endocrine 2003;20:111±4. Imbert-Bollore P, et al. Male reproductive function is not 68. Saunier E, Dif F, Kelly PA, Edery M. Targeted expression of affected in prolactin receptor-de®cient mice. Endocrinology the dominant-negative prolactin receptor in the mammary 2003;144:3779±82. gland of transgenic mice results in impaired lactation. 51. Robertson FG, Harris J, Naylor MJ, Oakes SR, Kindblom J, Endocrinology 2003;144:2669±75. Dillner K, et al. Prostate development and carcinogenesis in 69. Binart N, Imbert-Bollore P, Baran N, Viglietta C, Kelly PA. prolactin receptor knockout mice. Endocrinology 2003;144: A short form of the prolactin (PRL) receptor is able to rescue 3196±205. mammopoiesis in heterozygous PRL receptor mice. Mol 52. Steger RW, Chandrashekar V, Zhao W, Bartke A, Horseman Endocrinol 2003;17:1066±74. ND. Neuroendocrine and reproductive functions in male 70. Brisken C, Socolovsky M, Lodish HF, Weinberg R. The mice with targeted disruption of the prolactin gene. Endo- signaling domain of the erythropoietin receptor rescues crinology 1998;139:3691±5. prolactin receptor-mutant mammary epithelium. Proc Natl 53. Ormandy CJ, Binart N, Kelly PA. Mammary gland devel- Acad Sci USA 2002;99:14241±5. opment in prolactin receptor knockout mice. J Mammary 71. Welsch CW, Nagasawa H. Prolactin and murine mammary Gland Biol Neoplasia 1997;2:355±64. tumorigenesis: a review. Cancer Res 1977;37:951±63. 54. Brisken C, Kaur S, Chavarria TE, Binart N, Sutherland RL, 72. Wennbo H, Gebre-Medhin M, Gritli-Linde A, Ohlsson C, Weinberg RA, et al. Prolactin controls mammary gland Isaksson OG, Tornell J. Activation of the prolactin receptor development via direct and indirect mechanisms. Dev Biol but not the growth hormone receptor is important for 1999;210:96±106. induction of mammary tumors in transgenic mice. J Clin 55. Brisken C, Park S, Vass T, Lydon JP, O'Malley BW, Invest 1997;100:2744±51. Weinberg RA. A paracrine role for the epithelial progester- 73. Rose-Hellekant TA, Arendt LM, Schroeder MD, Gilchrist K, one receptor in mammary gland development. Proc Natl Sandgren EP, Schuler LA. Prolactin induces ERalpha-positive Acad Sci USA 1998;95:5076±81. and ERalpha-negative mammary cancer in transgenic mice. 56. Ormandy CJ, Naylor M, Harris J, Robertson F, Horseman Oncogene 2003;22:4664±74. ND, Lindeman GJ, et al. Investigation of the transcriptional 74. Ho KY, Weissberger AJ. Characterization of 24-hour growth changes underlying functional defects in the mammary hormone secretion in acromegaly: implications for diagnosis glands of prolactin receptor knockout mice. Recent Prog and therapy. Clin Endocrinol (Oxf) 1994;41:75±83. Horm Res 2003;58:297±323. 75. Yoshidome K, Shibata MA, Couldrey C, Korach KS, Green 57. Vomachka AJ, Pratt SL, Lockefeer JA, Horseman ND. JE. Estrogen promotes mammary tumor development in Prolactin gene-disruption arrests mammary gland develop- C3(1)/SV40 large T-antigen transgenic mice: paradoxical ment and retards T-antigen-induced tumor growth. Onco- loss of estrogen receptoralpha expression during tumor gene 2000;19:1077±84. progression. Cancer Res 2000;60:6901±10. 58. Deome KB, Faulkin LJ, Jr, Bern HA, Blair PB. Development 76. Vonderhaar BK. Prolactin: the forgotten hormone of human of mammary tumors from hyperplastic alveolar nodules breast cancer. Pharmacol Ther 1998;79:169±78. transplanted into gland-free mammary fat pads of female 77. Ben-Jonathan N, Liby K, McFarland M, Zinger M. Prolactin C3H mice. Cancer Res 1959;19:515±20. as an autocrine/paracrine growth factor in human cancer. 59. Mueller S, Clark J, Myers P, Korach K. Mammary gland Trends Endocrinol Metab 2002;13:245±50. development in adult mice requires epithelial and stromal 78. Vonderhaar BK. Prolactin involvement in breast cancer. estrogen receptors. Endocrinology 2002;143:2357±2365. Endocr Relat Cancer 1999;6:389±404. 60. Silberstein GB. Tumour-stromal interactions. Role of the 79. Hankinson SE, Willett WC, Michaud DS, Manson JE, stroma in mammary development. Breast Cancer Res 2001;3: Colditz GA, Longcope C, et al. Plasma prolactin levels and 218±23. subsequent risk of breast cancer in postmenopausal women. 61. Wiesen JF, Young P, Werb Z, Cunha GR. Signaling through J Natl Cancer Inst 1999;91:629±34. the stromal epidermal growth factor receptor is necessary for 80. Bhatavdekar JM, Patel DD, Shah NG, Vora HH, Suthar TP, mammary ductal development. Development 1999;126:335± Ghosh N, et al. Prolactin as a local growth promoter in 44. patients with breast cancer: GCRI experience. Eur J Surg 62. Humphreys RC, Lydon J, O'Malley BW, Rosen JM. Oncol 2000;26:540±7. Mammary gland development is mediated by both stromal 81. Gill S, Peston D, Vonderhaar BK, Shousha S. Expression of and epithelial progesterone receptors. Mol Endocrinol 1997; prolactin receptors in normal, benign, and malignant breast 11:801±11. tissue: an immunohistological study. J Clin Pathol 2001;54: 63. Wiseman BS, Werb Z. Stromal effects on mammary gland 956±60. development and breast cancer. Science 2002;296:1046±9. 82. McPherson K, Steel CM, Dixon JM. ABC of breast diseases. 64. Hovey RC, Trott JF, Ginsburg E, Goldhar A, Sasaki MM, Breast cancer-epidemiology, risk factors, and genetics. BMJ Fountain SJ, et al. Transcriptional and spatiotemporal 2000;321:624±8. regulation of prolactin receptor mRNA and cooperativity 83. Wang DY, de Stavola BL, Bulbrook RD, Allen DS, Kwa HG, with progesterone receptor function during ductal branch Verstraeten AA, et al. The permanent effect of reproductive growth in the mammary gland. DevDyn 2001;222:192±205. events on blood prolactin levels and its relation to breast 65. Camarillo IG, Thordarson G, Moffat JG, Van Horn KM, cancer risk: a population study of postmenopausal women. Binart N, Kelly PA, et al. Prolactin receptor expression in the Eur J Cancer Clin Oncol 1988;24:1225±31. epithelia and stroma of the rat mammary gland. J Endocrinol 84. Collaborative Group on Hormonal Factors in Breast Cancer. 2001;171:85±95. Breast cancer and hormonal contraceptives: collaborative
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 424 HARRIS . STANFORD . OAKES . ORMANDY
reanalysis of individual data on 53 297 women with breast weight and fat deposition in prolactin receptor-de®cient cancer and 100 239 women without breast cancer from 54 mice. Endocrinology 2001;142:532±7. epidemiological studies. Lancet 1996;347:1713±27. 102. McClellan KA, Robertson FG, Kindblom J, Wennbo H, 85. Mishell DR, Jr, Kletzky OA, Brenner PF, Roy S, Nicoloff J. Tornell J, Bouchard B, et al. Investigation of the role of The effect of contraceptive steroids on hypothalamic- prolactin in the development and function of the lacrimal and pituitary function. Am J Obstet Gynecol 1977;128:60±74. harderian glands using genetically modi®ed mice. Invest 86. Lagiou A, Lagiou P, Vassilarou DS, Stoikidou M, Tricho- Ophthalmol Vis Sci 2001;42:23±30. poulos D. Comparison of age at ®rst full-term pregnancy 103. Freemark M, Avril I, Fleenor D, Driscoll P, Petro A, Opara between women with breast cancer and women with benign E, et al. Targeted deletion of the PRL receptor: effects on islet breast diseases. Int J Cancer 2003;107:817±21. development, insulin production, and glucose tolerance. 87. Collaborative Group on Hormonal Factors in Breast Cancer. Endocrinology 2002;143:1378±85. Breast cancer and breastfeeding: collaborative reanalysis of 104. Wakao H, Gouilleux F, Groner B. Mammary gland factor individual data from 47 epidemiological studies in 30 (MGF) is a novel member of the cytokine regulated countries, including 50302 women with breast cancer and transcription factor gene family and confers the prolactin 96973 women without the disease. Lancet 2002;360:187±95. response. EMBO J 1994;13:2182±91. 88. Nevalainen MT, Valve EM, Ahonen T, Yagi A, Paranko J, 105. Gouilleux F, Wakao H, Mundt M, Groner B. Prolactin Harkonen PL. Androgen-dependent expression of prolactin induces phosphorylation of Tyr694 of Stat5 (MGF), a in rat prostate epithelium in vivo and in organ culture. prerequisite for DNA binding and induction of transcription. FASEB J 1997;11:1297±307. EMBO J 1994;13:4361±9. 89. Nevalainen MT, Valve EM, Ingleton PM, Nurmi M, 106. Burdon TG, Maitland KA, Clark AJ, Wallace R, Watson CJ. Martikainen PM, Harkonen PL. Prolactin and prolactin Regulation of the sheep beta-lactoglobulin gene by lactogenic receptors are expressed and functioning in human prostate. J hormones is mediated by a transcription factor that binds an Clin Invest 1997;99:618±27. interferon-gamma activation site-related element. Mol En- 90. Ahonen TJ, Harkonen PL, Laine J, Rui H, Martikainen PM, docrinol 1994;8:1528±36. Nevalainen MT. Prolactin is a survival factor for androgen- 107. Li S, Rosen JM. Nuclear factor I and mammary gland factor deprived rat dorsal and lateral prostate epithelium in organ (STAT5) play a critical role in regulating rat whey acidic culture. Endocrinology 1999;140:5412±21. protein gene expression in transgenic mice. Mol Cell Biol 1995;15:2063±70. 91. Wennbo H, Kindblom J, Isaksson OG, Tornell J. Transgenic 108. Broadhurst MK, Wheeler TT. The p100 coactivator is mice overexpressing the prolactin gene develop dramatic present in the nuclei of mammary epithelial cells and its enlargement of the prostate gland. Endocrinology 1997;138: abundance is increased in response to prolactin in culture and 4410±5. in mammary tissue during lactation. J Endocrinol 2001;171: 92. Kindblom J, Dillner K, Sahlin L, Robertson F, Ormandy C, 329±37. Tornell J, et al. Prostate hyperplasia in a transgenic mouse 109. Yahata T, Takedatsu H, Dunwoodie SL, Braganca J, with prostate-speci®c expression of prolactin. Endocrinology Swingler T, Withington SL, et al. Cloning of mouse Cited4, 2003;144:2269±78. a member of the CITED family p300/CBP-binding transcrip- 93. Xu X, Wu W, Williams V, Khong A, Chen YH, Deng C, et tional coactivators: induced expression in mammary epithe- al. Opposite effects of unmodi®ed prolactin and a molecular lial cells. Genomics 2002;80:601±13. mimic of phosphorylated prolactin on morphology and the 110. Favy DA, Rio P, Maurizis JC, Hizel C, Bignon YJ, Bernard- expression of prostate speci®c genes in the normal rat Gallon DJ. Prolactin-dependent up-regulation of BRCA1 prostate. Prostate 2003;54:25±33. expression in human breast cancer cell lines. Biochem 94. Syms AJ, Harper ME, Grif®ths K. The effect of prolactin on Biophys Res Commun 1999;258:284±91. human BPH epithelial cell proliferation. Prostate 1985;6:145± 111. Brockman JL, Schroeder MD, Schuler LA. PRL activates the 53. cyclin D1 promoter via the Jak2/Stat pathway. Mol 95. Janssen T, Darro F, Petein M, Raviv G, Pasteels JL, Kiss R, Endocrinol 2002;16:774±84. et al. In vitro characterization of prolactin-induced effects on 112. Brisken C, Ayyannan A, Nguyen C, Heineman A, Reinhardt proliferation in the neoplastic LNCaP, DU145, and PC3 F, Tan J, et al. IGF-2 is a mediator of prolactin-induced models of the human prostate. Cancer 1996;77:144±9. morphogenesis in the breast. DevCell 2002;3:877±87. 96. Xu X, Kreye E, Kuo CB, Walker AM. A molecular mimic of 113. Hovey RC, Harris J, Hadsell DL, Lee AV, Ormandy CJ, phosphorylated prolactin markedly reduced tumor incidence Vonderhaar BK. Local insulin-like growth factor-II mediates and size when DU145 human prostate cancer cells were prolactin-induced mammary gland development. Mol En- grown in nude mice. Cancer Res 2001;61:6098±104. docrinol 2003;17:460±71. 97. Bouchard B, Ormandy CJ, Di Santo JP, Kelly PA. Immune 114. Fata JE, Kong YY, Li J, Sasaki T, Irie-Sasaki J, Moorehead system development and function in prolactin receptor- RA, et al. The osteoclast differentiation factor osteoprote- de®cient mice. J Immunol 1999;163:576±82. gerin-ligand is essential for mammary gland development. 98. Krishnan N, Thellin O, Buckley DJ, Horseman ND, Buckley Cell 2000;103:41±50. AR. Prolactin suppresses glucocorticoid-induced thymocyte 115. Srivastava S, Matsuda M, Hou Z, Bailey JP, Kitazawa R, apoptosis in vivo. Endocrinology 2003;144:2102±10. Herbst MP, et al. Receptor activator of NF-kB ligand 99. Clement-Lacroix P, Ormandy C, Lepescheux L, Ammann P, (RANKL) induction via Jak2 and Stat5a in mammary Damotte D, Gof®n V, et al. Osteoblasts are a new target for epithelial cells. J Biol Chem 2003;2:2. prolactin: analysis of bone formation in prolactin receptor 116. Brisken C, Heineman A, Chavarria T, Elenbaas B, Tan J, knockout mice. Endocrinology 1999;140:96±105. Dey SK, et al. Essential function of Wnt-4 in mammary gland 100. Craven AJ, Ormandy CJ, Robertson FG, Wilkins RJ, Kelly development downstream of progesterone signaling. Genes PA, Nixon AJ, et al. Prolactin signaling in¯uences the timing Dev2000;14:650±4. mechanism of the hair follicle: analysis of hair growth cycles 117. Kenney NJ, Smith GH, Rosenberg K, Cutler ML, Dickson in prolactin receptor knockout mice. Endocrinology 2001; RB. Induction of ductal morphogenesis and lobular hyper- 142:2533±9. plasia by amphiregulin in the mouse mammary gland. Cell 101. Freemark M, Fleenor D, Driscoll P, Binart N, Kelly P. Body Growth Differ 1996;7:1769±81.
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 PROLACTIN AND THE PROLACTIN RECEPTOR 425
118. Luetteke NC, Qiu TH, Fenton SE, Troyer KL, Riedel RF, 122. Frasor J, Gibori G. Prolactin regulation of estrogen receptor Chang A, et al. Targeted inactivation of the EGF and expression. Trends Endocrinol Metab 2003;14:118±23. amphiregulin genes reveals distinct roles for EGF receptor 123. Feltus FA, Groner B, Melner MH. Stat5-mediated regulation ligands in mouse mammary gland development. Develop- of the human type II 3beta-hydroxysteroid dehydrogenase/ ment 1999;126:2739±50. delta5-delta4 isomerase gene: activation by prolactin. Mol 119. Gass S, Harris J, Ormandy CJ, Brisken C. Using gene Endocrinol 1999;13:1084±93. expression arrays to elucidate transcriptional pro®les under- 124. Albarracin CT, Parmer TG, Duan WR, Nelson SE, Gibori G. lying prolactin function. J Mammary Gland Biol Neoplasia Identi®cation of a major prolactin-regulated protein as 20 2003;8:269±285. alpha-hydroxysteroid dehydrogenase: coordinate regulation 120. Ganguly TC, O'Brien ML, Karpen SJ, Hyde JF, Suchy FJ, of its activity, protein content, and messenger ribonucleic Vore M. Regulation of the rat liver sodium-dependent bile acid expression. Endocrinology 1994;134:2453±60. acid cotransporter gene by prolactin. Mediation of transcrip- 125. Nanbu-Wakao R, Fujitani Y, Masuho Y, Muramatu M, tional activation by Stat5. J Clin Invest 1997;99:2906±14. Wakao H. Prolactin enhances CCAAT enhancer-binding 121. Yu-Lee LY, Hrachovy JA, Stevens AM, Schwarz LA. protein-beta (C/EBP beta) and peroxisome proliferator- Interferon-regulatory factor 1 is an immediate-early gene activated receptor gamma (PPAR gamma) messenger RNA under transcriptional regulation by prolactin in Nb2 T cells. expression and stimulates adipogenic conversion of NIH- Mol Cell Biol 1990;10:3087±94. 3T3 cells. Mol Endocrinol 2000;14:307±16.
# 2004 Taylor & Francis. ISSN 0785-3890 Annals of Medicine 36 0888-8809/05/$15.00/0 Molecular Endocrinology 19(7):1868–1883 Printed in U.S.A. Copyright © 2005 by The Endocrine Society doi: 10.1210/me.2004-0254
Transcriptional Changes Underlying the Secretory Activation Phase of Mammary Gland Development
Matthew J. Naylor,* Samantha R. Oakes,* Margaret Gardiner-Garden, Jessica Harris, Katrina Blazek, Timothy W. C. Ho, Foo C. Li, David Wynick, Ameae M. Walker, and Christopher J. Ormandy Development Group (M.J.N., S.R.O., M.G.-G., J.H., K.B., C.J.O.), Cancer Research Program, Garvan Institute of Medical Research, St. Vincent’s Hospital, Sydney, New South Wales 2010, Australia; Division of Biomedical Sciences (T.W.C.H., A.M.W.), University of California, Riverside, California 92521; and University Research Centre Neuroendocrinology (F.C.L., D.W.), Bristol University, Bristol BS2 8HW, United Kingdom
The secretory activation stage of mammary gland representing milk proteins, metabolism,lipid, cho- development occurs after parturition and converts lesterol and fatty acidbiosynthetic enzymes, im- inactive lobuloalveoli to active milk secretion. This mune response, and key transcription factors. A process istriggered by progestinwithdrawal and set of 35 genes, commonly regulated in all three depends upon augmented prolactin (Prl) signaling. models, was identified and their role in lactogene- Little is known about the Prl-induced transcrip- sis was validated by examining their expression in tional changes that occur in the mammary gland to response to Prl stimulation or signal transducer drive this process. To examine changes in the and activator of transcription 5 knockdown in the mammary transcriptome responsible for secretory HC11 mouse mammary cell culture model. The activation, we have used transcript profiling of transcript profiles provided by these experiments three mouse models that exhibitfailure of secre- identify 35 key genes (many for the first time) in- tory activation: knockout of galanin (a regulator of volved in the secretory activation phase of mam- pituitary Prl production and a mammary cell auton- mary gland development, show that S179D acts as omous modulator of Prl action); treatment with an antagonist of Prl action, and provide insight into S179D Prl (a phosphoprolactin mimic); and knock- the partial penetrance of failed lactation in Prl re- out of a single Prl receptor allele. Asignificant ceptor heterozygous females. (Molecular Endocri- reduction in expression was observed in genes nology 19: 1868–1883, 2005) belonging to 46 gene ontologies including those
NLIKE MOST DEVELOPMENTAL processes, branching gives way to the formation of lobuolalveloar Umammary gland development occurs after birth structures. During this phase the basic architecture of and indiscrete phases dependent on the endocrine the gland is established, with lobuloalveoli replacing state of the organism. These phases have been de- the previously predominant adipose tissue. The ap- fined as 1) ductal morphogenesis, occurring during pearance of cytoplasmiclipid droplets signals the on- puberty and with each estrous cycle; 2) alveolar mor- set of lactogenesis, which isdivided into the secretory phogenesis, occurring during pregnancy and consist- initiation and activation phases (2). ing of an initial proliferation phase, followed by the The initiation phase, beginning around midpreg- secretory initiation phase and secretory activation nancy, results in the acquisition of limited secretory phase, 3) lactation postpartum, and then 4) involution capacity. Milk protein gene expression commences in during weaning of the offspring (1, 2). a programmed pattern such that Wdnm1 is expressed Proliferative alveolar morphogenesis occurs in re- first, followed by the caseins, whey acidic protein sponse to increased levels of estrogen, progesterone, (Wap) and lactalbumin (3). Secretory capacity is and prolactin (Prl). An early increase in ductal side gained only by a subset of epithelial cells. Prl, proges- First Published Online February 10, 2005 terone, and estrogen are permissive for this stage of *M.J.N and S.R.O. contributed equally to this work. development, and further development is held in Abbreviations: Egf, Epidermal growth factor; Gal, galanin; check by high progesterone levels controlled by the GFP, green fluorescent protein; PI3, phosphatidylinositol 3; placenta (4). This occurs directly inhumans via pla- P-Prl, phosphorylated Prl; Prl, prolactin; Prlr, Prl receptor; QPCR, quantitative RT-PCR; RMA, robust multiarray aver- cental progesterone secretion or indirectly in mice via age; siRNA, short interfering RNA; Stat, signal transducer and placental lactogen support of the progesterone-pro- activator of transcription; U-Prl, unmodified Prl; Wap, whey ducing ovarian corpus luteum. acidic protein. The secretory activation phase of lactation istrig- Molecular Endocrinology is published monthly by The gered by falling progesterone levels, which in mice Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine also trigger parturition, with the result that milk is im- community. mediately available to the pups. In humans, secretory
1868 Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1869
activation commences after parturition and so lacta- mammary gland transcriptome that result in secretory tion is delayed until1dor2dpostpartum. Prl levels activation of the mammary gland. also rise in all species at thistime and are essential for lactation (2). The mammary epithelium also synthe- sizes Prl, which is essential for the increase inepithelial cell proliferation that accompanies secretory activa- RESULTS tion (5). Although the hormonal control of the concomitant A Molecular Mimic of Phosphorylated Prl (P-Prl) morphological and functional events of secretory ac- Inhibits Secretory Activation tivation have been well described, the underlying al- terations occurring in gene expression in the mam- Serine 179 is the major phosphorylation site of hu- mary gland that drive these events are not well man Prl (27, 28), and substitution of this position understood. Experimental models used to examine with aspartate (S179D) mimics the effects of P-Prl secretory activation have been limited to the exami- (21). We treated mice using 28-d miniosmotic nation of the response of candidate genes to hormonal pumps loaded with S179D, unmodified Prl (U-Prl), or manipulation of whole animals. The combination of vehicle (saline) inserted interscapularly on the day of mouse gene knockout and microarray technology (6) observation of a vaginal plug after copulation. Of 15 now offers a new experimental approach to this saline-treated control animals who became preg- question. nant, all had normal pregnancies and deliveries, and We have made two knockout models that experi- all successfully suckled a total of 120 pups, as in- ence failure of secretory activation. These are animals dicated by the presence of milk in the stomach of ϩ/ϩ withasingle functional Prl receptor (Prlr) allele (7) and the pups (Fig. 1A). Treatment of Gal animals with animals inwhich the neuropeptide galanin (Gal) is lost S179D caused lactational failure in four of five fe- (8). Loss of a Prlr allele reduces mammary Prlr expres- males and, despite continued suckling, the stom- sion during pregnancy (9) and, intriguingly, also results achs of 30 pups from these four S179D-treated mice in poor maternal behavior, reducing or abolishing milk did not contain milk. Significantly, treatment of five Ϫ Ϫ delivery to the pups (10). Gal is an autocrine/paracrine Gal / females with S179D, at rates that produced ϩ ϩ growth factor for the Prl-secreting pituitary lactotroph the lactational failure in Gal / animals, was unable cells (8, 11) and, consequently, GalϪ/Ϫ mice display a to rescue lactation, and all 40 pups showed no milk failure of estrogen-induced lactotroph proliferation in their stomachs, demonstrating no agonist activity and reduced Prl serum levels. As a result, secretory of S179D inthis assay. S179D treatment prevented activation fails in GalϪ/Ϫ mothers and the pups die the large rise in the level of -casein gene expres- ϩ ϩ (12). The levels of other pituitary hormones, such as sion seen in Gal / animals (Fig. 1B). Levels in GH, LH, FSH, and TSH are normal in GalϪ/Ϫ mice. Gal S179D-treated animals were comparable to those Ϫ Ϫ ϩ Ϫ and its receptors are also expressed by the mammary measured in Gal / animals and Prlr / animals, epithelium and we have recently demonstrated that indicating that all models passed through secretory Gal exerts a direct developmental effect via the mam- initiation but were developmentally blocked during mary epithelium to modulate Prl action during preg- the secretory activation phase. A reduction in ex- nancy (13). pression in response to S179D was also seen in Several posttranslational modifications of Prl are Wdnm1 and Wap milk protein mRNAs (Fig. 1C) and known to occur (14) with phosphorylation of Prl being ␣- and -casein by Western blot (Fig. 1D). The com- quantitatively the most important (15–17). This phos- parison of milk protein content by Western analysis phorylation can be mimicked by mutation of the nor- iscomplicated by the retention of colostrumin the Ϫ Ϫ ϩ Ϫ mally phosphorylated serine to an aspartate (S179D in S179D, Gal / , and Prlr / animals and its expul- most species). Treatment of female rats with S179D sion in the wild-type control animals capable of resulted infailed lactation and slower onset of mater- lactation in response to suckling. Colostrum has 5 nal behavior (18, 19). The in vitro mechanism of S179D times the milk protein content of milk and is seen as action is currently controversial, with both weak ago- the strong pink staining of luminal content by hema- nist and potent antagonist activities described (20–23). toxylin and eosinhistology in the S179D-treated In vivo treatment with S179D mimicsanumber of glands. Morphological examination of the fourth phenotypes previously described in PrlrϪ/Ϫ mice (10, mammary gland at the first day postpartum also 19, 24–26). For the purposes of the current study, showed that S179D treatment produced a block in S179D provides another model of failed secretory ac- development at the same developmental stage as tivation. Unlike the knockouts, its effects are exerted that caused by loss of a Prlr allele (29) or loss of Gal from the onset of pregnancy, providing a control for (13). Compared with controls (Fig. 1E), lobules had any undiscovered prepregnancy developmental ef- formed, but appeared smaller and less dense than fects in the germline models. those seen inwild-type animals (Fig. 1F). Histolog- We used these experimental models combined with ical examination showed that, compared with vehi- oligonucleotide arrays (Affymetrix, Santa Clara, CA; cle-treated controls (Fig. 1G), the S179D-treated an- U74A2 GeneChip) to examine the alterations in the imals showed reduced alveolar density and failed
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 1870 Mol Endocrinol, July 2005, 19(7):1868–1883 Naylor et al.•Transcriptome Change during Secretory Activation
S179D Prl Inhibits Signal Transducer and Activator of Transcription 5 (Stat5) Activation in the Mammary Gland
Binding of Prl to the Prlr results in receptor dimer- ization and subsequent activation of the Jak/Stat pathway (30), the major signaling pathway used by the long Prlr. Both Jak2 and Stat5 have been dem- onstrated as essential for mammary gland develop- ment and milk protein gene expression (31, 32). Activation of Prlr can also result in the induction of the Mapk and phosphatidylinositol 3 (PI3) kinase signaling pathways (33–36). To investigate the mechanism of S179D-induced failure of lobuloal- veolar development and lactation, the activation of these signaling pathways in the mammary gland was examined by Western blot. The proportions of Stat5a in total protein were equal in the mammary glands of saline- and S179D- treated Galϩ/ϩ mice; however, the amount of Stat5 that was phosphorylated was markedly reduced in the mammary glands of mice treated with S179D (Fig. 2A). The levels of phosphorylated and total ERK were variable among animal samples, and there was no apparent difference (Fig. 2B). Likewise, the amount of total and phosphorylated Akt, a down- stream target of the PI3 kinase-signaling pathway, did not significantly change between the two groups of mice (Fig. 2B). Asmall diminution in the level of cyclin D1, previously reported to be up-regulated by U-Prl and down-regulated by S179D (20), was de- tected insome experiments.
؉ ؊ Fig. 1. S179D Inhibits Lobuloalveolar Development and Prlr / Show Diminished Stat5 Activation Lactation inMice A, Percentage of mothers with surviving pups after treat- We undertook the samecomparison of signaling ϩ/ϩ ment with vehicle (saline), U-Prl ,or S179D of Gal or pathways in Prlrϩ/Ϫ mice. About two thirds of these GalϪ/Ϫ mice throughout pregnancy by 21-d miniosmotic animals experience failure of lactogenesis (NL) dur- pump. B, -Casein mRNA levels in the fourth inguinal mammary glands from Galϩ/ϩ or GalϪ/Ϫ mice measured by ing late pregnancy (29), due to a reduction in Prlr QPCR and expressed as a ratioofGalϩ/ϩ levels at the first numbers (37) and thus the level of Prlr signaling. day postpartum. Comparison is made with levels seen in These animals provide a second model of develop- Prlrϩ/Ϫ mice incapable of lactation. C, Examination of the mental arrest during the secretory activation phase. effects of S179D treatment on Wdnm1 and Wap expres- The Prlrϩ/Ϫ phenotype is partially penetrant and sion by QPCR at the first day postpartum. Fold difference thus dependent upon the presence or absence of in expression levels expressed as S179D Prl-treated ϩ ϩ ϩ ϩ allelic variants or haplotypes that segregate in the Gal / mice vs. vehicle-treated Gal / mice. D, Measure- ment of milk protein level (␣-casein, -casein, and Wap) by mixed genetic background. Some individuals are Western blot and densitometry in S179D-treated Galϩ/ϩ able to lactate (L) at a level sufficient for pup sur- ϩ Ϫ mice. E and F, Carmine-stained whole-mount analysisof vival. Comparison of Prlr / animals that experi- mammary gland development in vehicle or S179D-treated enced failure of secretory activation with Prlrϩ/ϩ ϩ/ϩ Gal mice at the first day postpartum. G and H, Hema- animals showed a reduction in Stat5 phosphoryla- toxylin and eosin-stained 5- m sections from the same tion, with no detectable alteration in the other path- glands. ways examined (Fig. 2C). Variable levels of Stat5 phosphorylation were seen in Prlrϩ/Ϫ animals capa- lactation, seen as the pink-stained accumulation of ble of lactation sufficient for pup survival, but these colostrum proteinwithin the ducts and alveoli of Fig. were all higher than in the animals that could not 1H. Together these data show that secretory acti- lactate. We have previously shown mammary Stat5 vation had failed in S179D-treated animals. phosphorylation in response to Gal in culture (13).
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1871
broad overview of the functional groups contained within the set of genes exhibiting altered expression in association with failed lactation, we used OntoExpress (38) to identify gene ontologies with statistically signif- icant overrepresentation in the set of genes showing altered expression associated with failed lactation. These ontologies are shown in supplemental Table 1 (published as supplemental data on The Endocrine Society’s Journals Online web site at http://mend. endojournals.org) where x denotes the significant overrepresentation of that ontology in one or more of the animal models used, and ontologies are organized into functional groups. Genes involved in cholesterol, sterol, fatty acid, and lipidbiosynthesis, metabolism, and glycolysis are overrepresented. Interestingly, genes involved in immune responses are also over- represented to a sufficient degree to be detected at this broad-overview level, probably indicative of suc- cessful preventative control of organismswith poten- tial to cause mastitis, as we did not detect any cases of the infection. Other notable ontologies include the IGF-binding proteins, actin cytoskeleton, and arginine metabolic enzymes.
Venn Analysis of the Alterations in Gene Expression Patterns
We undertook a Venn analysis of gene expression among the models of failed lactation. We searched for Fig. 2. S179D Prl inhibits STAT5 Activation in the Mammary genes that showed altered expression between Gland Prlrϩ/ϩ and Prlrϩ/Ϫ (lactating or nonlactating) in both Western analysis of STAT5 (A), and MAPK and PI3 kinase experimental replicates. We denoted this set as (B) signaling pathways in the mammary glands of saline- Ϫ ϩ/ϩ “ Prlr,” the genes showing changed expression due vehicle and S179D-treated Gal mice at the first day post- ϩ ϩ to loss of the Prlr. Similarly, we compared Gal / to partum. C, The same analysis conducted using mammary Ϫ/Ϫ glands from Prlrϩ/ϩ and Prlrϩ/Ϫ animals that exhibited no Gal and identified the genes changing expression Ϫ lactation (No lact.) or lactation sufficient for pup survival (Par- due to loss of Gal, denoted “ Gal.”Wecompared ϩ/ϩ ϩ/ϩ tial lact.). Gal treated with saline to Gal treated with S179D to identify genes that changed in response to S179D treatment denoted “ϩS179D.” We then com- Investigation of the Altered Patterns of Gene bined these sets in the Venn analysis shown inFig. 3A. Expression Underlying Failed Lactogenesis There were 7,278 probe sets that were not expressed Produced by Loss of Gal, by Loss of a Prlr Allele, in the mammary gland of the 12,488 probe sets on the or by Treatment with S179D chip. There were 939 probe sets that showed changed expression in at least one of the models from a total of We measured the alterations in gene expression dur- 5210 probe sets with detectable expression. The Venn ing the secretory activation phase of lactogenesis us- analysis shows the number of genes with increased (I) ing oligonucleotide expression arrays to compare the or decreased (D) expression in each subset. The global patterns of altered gene expression in our three squares below the Venn diagram show these data for models of failed lactogenesis: loss of a Prlr allele, loss the intersections of two sets, and the pattern of gene of Gal, and by treatment with S179D. To perform these expression change for the 35 genes of the center experiments, RNA was pooled from four to six repli- intersection set can be found inFig. 5. The false dis- cate animals from each of the seven different geno- covery rate for each set is indicated in brackets. types or treatment groups, and expression profiles We estimated the false discovery rate due to multi- were obtained using MGU74Av2 Affymetrix Gene- ple testing by first comparing identical experimental Chips. The entire experiment, including all animal replicates (e.g. Prlrϩ/ϩ with Prlrϩ/ϩ)tofind genes with treatments and RNA pooling, was repeated at a later changed expression due to random events or experi- time to provide complete experimental duplication. mental error, and then randomizing these sets to pro- The results are available as supplemental data pub- duce a three-set Venn diagram, recalculating the Venn lished on The Endocrine Society’s Journals Online analysis inthis way 240 times (twice the number of web site at http://mend.endojournals.org. To gaina possible combinations). The median false discovery
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 1872 Mol Endocrinol, July 2005, 19(7):1868–1883 Naylor et al.•Transcriptome Change during Secretory Activation
ϩS179D sets had such high false discovery rates that they could be considered as biologically irrelevant. These irrelevant sets correspond biologically to unique and non-Prl-mediated actions. The unique actions of Gal (set of 110 I and 18 D) and the unique actions of S179D (set of 42 I and 86 D) each contained 102 members erroneously placed there by the multiple testing error, an 80% error rate. Similarly, the intersec- tion set of 29 contained a 30% error rate. This analysis indicated first that S179D has virtually no detectable off-target effects and second that Gal acts to influence mammary gland development predominantly via mod- ulation of Prl action, at least at thispoint in develop- ment. Conversely, the false discovery rate was very low for all sets and intersection sets within the ϪPrlr set, showing these subsets to be of high biological relevance. Prlr was clearly shown to be the dominant force in secretory activation, as loss of a Prlr allele resulted in 654 of the 939 changes detected, with a strong bias toward positive action. The expression of most of these genes (602) was not influenced by S179D, greatly limiting the role of S179D and P-Prl as global modulators of Prl action, but the fact that S179D in- duced failure of lactation demonstrated that it regu- lated key genes. Knockout of Gal regulated 109 genes incommon with loss of a Prlr allele. This set is indic- ative of Gal control of pituitary Prl secretion (12) and Gal modulation of Prl action at the mammary epithelial cell (13). Examination of the ϩS179D sets shows that S179D shares actions incommon with ϪGal and ϪPrlr (central set of 35 genes) and with Prl indepen- dent of Gal (set of 52 genes). Where ϩS179D ex- Fig. 3. Transcript Profiling of Three Models of Failed Lacta- Ϫ tion erted effects incommon with Prlr (87 genes com- Global patterns of gene expression change were analyzed prised of sets containing 52 and 35 genes), the in 1 d postpartummammary glands frommice experiencing observed pattern overwhelmingly demonstrated lactational failure due to treatment with S179D (ϩS179D), S179D to be an antagonist of Prl action, as just three loss of a Prlr allele (ϪPrlr), or loss of the Gal gene (ϪGal). A, of these 87 genes increased with S179D treatment Venn diagram depicting the patterns of unique and overlap- and loss of Prl signaling, whereas 84 responded to ping changes in gene expression among the models of failed a loss of Prl signaling flux and treatment with S179D lactation. The number of genes showing increasing expres- sion (I) and decreasing expression (D) isgiven for the unique inasimilar way. sets within the Venn, and for the two-way intersections in the squares below the Venn. Estimate of the false discovery rate Validation of Array Results by Quantitative RT- due to multiple testing is indicated in brackets. The identities, PCR (QPCR) behavior, and postulated function of the 35 genes of the central set can be found inFig. 5. B, Comparison of the We used QPCR on the LightCycler platform to con- change in gene expression measured by Affymetrixchip (du- firm the array results for the nine genes selected plicates) and QPCR (triplicates) for nine genes chosen at random from the central set of 35 genes. Both techniques from the central set of 35. In all samples the showed high reproducibility. The I or D call was verified in all Affymetrix increasing or decreasing calls were con- cases, and the magnitude of the fold change was generally firmed by QPCR, and overall it was apparent that the smaller when measured by Affymetrix. Affymetrix estimate of the magnitude of change was conservative. Both the Affymetrix and QPCR tech- niques gave highly reproducible and consistent re- rate of the 240 measurements for each subset is indi- sults (Fig. 3B) as we have observed previously in cated in brackets. The most important outcomeofthis other studies (13, 25, 39), allowing the use of the analysis was the demonstration that the sets outside Affymetrix profiles as an accurate measure of gene the ϪPrlr set and its intersections with the ϪGal and expression level.
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1873
Comparison of Milk Protein Expression and Lipid Biosynthetic Enzymes among the Models of Failed Lactation
We compared the changes in the expression of a panel of milk proteins and lipidbiosynthetic enzymes to determine whether the arrest in secretory activation had occurred at the samepoint in development among our models of failed lactation. We used the gene pan- els defined by Rudolph et al. (6) for the milk proteins (Fig. 4A) and lipidbiosynthetic enzymes (Fig. 4B) inthis analysis. It is clear from thiscomparison that all mod- els have proceeded through secretory initiation as all models express all of the milk proteins including ␣-lactalbumin, the last of the milk proteins to be ex- pressed (3). All express the lipidbiosynthetic machin- ery required for the production of lipid droplets, con- sistent with the histology of these models. Strikingly, the reduction in expression of these genes compared with wild-type animals is very similar among the mod- els, with the exception of casein-␦,which showed the most heterogeneous response among the models. Thus these models arrest development at very similar points during secretory activation.
Key Lactational Regulators Identified from the Intersection Set
The 35 genes (Venn Fig. 3A) that were commonly altered in the three models of lactational failure rep- resent a small set of key genes involved in lacto- genesis. This is a very stringent analysis and relax- ation of the conditions by requiring, for example, a gene to be represented in just one of two experi- mental replicates, expands the central set to many hundreds of genes. Clustering using the small set of 35 unequivocally separated those animals able to Fig. 4. Developmental Arrest Occurs at the Same Stage in lactate from those unable to lactate. Analysisofthe All Models of Failed Lactogenesis function of these genes by extensive literature The expression of a panel of milk protein genes (A) and searches showed that most were involved in the lipidbiosynthesis genes (B) developed by Ruldolph et al. (6) metabolic processes underlying milk production, was examined in the models of failed lactation. Results are such as the synthesisoftriacylglycerols and cho- expressed as a percentage of theirwild-type controls inan lesterol from glucose, transport of fatty acids from analysisusing epithelial-specific keratin18tonormalize gene the circulation, and lactose synthesis. Although this expression across groups. Gray bars indicate calls of de- creasing expression, and the white bar a call of no change in finding is not surprising, many of the genes con- expression by MAS5. All decreases in expression were sta- tained inthis set have not previously been impli- tistically significant at P Ͻ 0.01. cated in lactogenesis and they are now placed within the lactation pathway for the first time(Fig. 5). We have presented these data in a functional Genes Associated with Partial Rescue of scheme based upon extensive literature searching, Lactation in Prlr؉/؊ Mice together with a heat map to indicate expression pattern, to allow a succinct summary of the large This data set also allowed us to examine the alterations amount of information on which thisfigure is based. in gene expression that occur between Prlrϩ/Ϫ females Unigene names are used and Unigene should be that could not lactate and those able to lactate suffi- used as the access point to the gene structure, full ciently for pup survival. Here we are examining the ability gene name, and aliases, and thus the Medline pub- of segregating genetic elements within mixed 129 ge- lication data used to construct thisfigure. The Dis- netic backgrounds to rescue lactation insome individu- cussion contains details of the more immediately als. Using an analysis of the data by the robust multiarray interesting genes inthis set. average (RMA)/penalized T Statistic method, of the top
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 1874 Mol Endocrinol, July 2005, 19(7):1868–1883 Naylor et al.•Transcriptome Change during Secretory Activation
Fig. 5. The Common Lactation Signature: The Identities, Behavior, and Postulated Function of the 35 Genes with Altered Expression in all Models of Lactation Failure The 35 genes of the central Venn set are identified by their gene name, which should be used to retrieve theirUnigene entry for more detail (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?dbϭunigene). Relative gene expression level is shown in all of the experiments analyzed using a heat map (darker shading indicates higher expression), and the experiments are clustered using these genes. Groups of animals experiencing lactational failure are indicated by green or blue shading, and lactating groups are shaded pink or yellow,using the clustering inFig. 7. The postulated function of these genes has been placed where possible in abiochemical scheme of a lactating mammary epithelial cell. Unplaced genes are Tparl, Cuta, Gbp3, and the expressed sequence tags (ESTs). PDH, Pyruvate dehydrogenase; Cyp51, cytochromeP450 family 51; NEFA, nonesterified fatty acid.
100 genes ranked on significance of their P value for AnalysisbyHierarchical Clustering changed expression between Prlrϩ/Ϫ animals that lac- tated compared with those that did not, an extraordinary We used hierarchical clustering to examine the simi- 25% were found to play a role in the initiation of DNA larity in gene expression changes among our various replication (Fig. 6). These genes are presented according samples. Initial clustering using the 939 genes with to their function in the assembly of the DNA replication changed expression grouped the samples according machinery, adapted from the review of Bell and Dutta to the genetic background of the sample, which varies
(41). In addition to these genes, a number of G2/M phase between Gal and Prlr knockouts, obscuring the effect genes were also found. As demonstrated by the heat of genotype or treatment. The Prlr and Gal knockout map, the expression levels of these genes were not only models were made using chimera breeding partners higher than nonlactating animals but were higher than derived from different 129 mouse strains, and gene wild-type animals as well. Thus, the most prominent expression changes due to this background difference feature of lactational rescue in these animals is abnor- were dominant over changes induced by genotype. mally high expression of genes involved in cell prolifer- We have previously demonstrated the dramatic effects ation and mitosis. of genetic background on mammary ductal patterning
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1875
Fig. 6. Increased Mammary Cell Proliferation Is Associated with Partial Penetrence of Prlrϩ/Ϫ Lactational Failure Comparison of mammary gene expression profiles between Prlrϩ/Ϫ that were capable (L) or incapable (NL) of lactation, and ϩ/ϩ Prlr animals, revealed a predominant functional group of genes involved in the initiation of DNA synthesis and G2/M progression. The relative expression levels of these genes are shown in a heat map (darker indicates higher expression). Groups of animals experiencing lactational failure are indicated by green or blue shading, lactating groups are shaded pink or yellow,using the clustering inFig. 7. Their functions are shown inadiagram depicting the assembly of the prereplicative and replicative complex adapted from Bell and Dutta (41), which should be consulted for further detail. The black lines represent the DNA and the colored shapes represent the various proteincomponents of the complex. Proteins active at G2/M phase are shown in the box below, where binding partners are indicated by short black lines. Proteins named in red show increased expression in Prlrϩ/Ϫ animals capable of lactation over both Prlrϩ/Ϫ NL and Prlrϩ/ϩ. dNTP, Deoxynucleotide triphosphate; rNTP, ribonucleoside triphosphates.
(42). To overcomethis inherent problem,weidentified closely related to a second cluster containing the a set of genes that showed a change in expression Prlrϩ/Ϫ duplicates that were capable of lactation. Dis- between the wild-type (ϩ/ϩ)animals of the different tant to these clusters were the Prlrϩ/Ϫ that did not strains and removed them from the analysis. This re- lactate and a fourth cluster containing animals treated duced our set of genes with genotype-specific with S179D and the two GalϪ/Ϫ. The GalϪ/Ϫ mice changes to 316, and clustering using this set now treated with U-Prl were split, indicating that the U-Prl allowed the effect of genotype to emerge. Four clus- treatment was less effective in rescuing lactation in ters were found (Fig. 7). The wild-type animals of both one of the treatment groups. It is clear from this anal- strains now formed a cluster together. This cluster was ysis that the changed pattern of gene expression pro-
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 1876 Mol Endocrinol, July 2005, 19(7):1868–1883 Naylor et al.•Transcriptome Change during Secretory Activation
Fig. 3. When we clustered using the genes that changed in the Prlr model (ϪPrlr), we again produced the same four clusters, but the effects of S179D and loss of Gal were less distinguishable than Prlr geno- type (Fig. 7B). Using the ϩS179D gene set to cluster separated the GalϪ/Ϫ and S179D-treated samples from the rest (Fig. 7C), and clustering using the ϪGal set (Fig. 7D) split the wild type and Prlrϩ/Ϫ that could lactate from the S179D and Prlrϩ/Ϫ that experienced lactational failure. This analysis suggests a close rela- tionship between the transcriptional changes caused by the loss of Gal and those caused by treatment with the S179D mimic of P-Prl. We investigated the P-Prl pituitary secretion of phosphoprolactin in GalϪ/Ϫ vs. Galϩ/ϩ mice.
The Ratio of Phosphorylated to Unphosphorylated Prl Is Altered in the Pituitaries of Gal؊/؊ Mice
We sought to determine whether expression of Gal regulates the ratio of P-Prl to U-Prl that is released by the pituitary. To eliminate the effects of hormonal sta- tus on the degree of Prl phosphorylation, male animals were used. In Galϩ/ϩ male mice, 80.0 Ϯ 4.1% of Prl was present in the unmodified form, whereas 20.0 Ϯ 1.9% was in the phosphorylated form (Fig. 8). GalϪ/Ϫ mice, however, had 68.9 Ϯ 3.2% of Prl as the unmod- ified form and 31.1 Ϯ 2.1% as the phosphorylated form (Fig. 1, P Ͻ 0.0001 Student’s (unpaired) t test). Thus, the relative ratio of U-Prl to P-Prl was 4:1 in
Fig. 7. ϪGal and ϩS179D Models Show Very Similar Pat- terns of Transcriptional Change Hierarchical clustering was used to draw a dendogram that clustered the separate experiments according to the similar- ities in their pattern of gene expression. A, Clustering using a set of 316 genes that changed in at least one of the models. Clusters of highly related experiments are colored. Clusters experiencing lactational failure are green or blue; lactating clusters are pink or yellow. Panel B, Clustering using those Fig. 8. Levels of Modified Forms of Prl Secreted from the Ϫ genes that changed in the Prlr model. C, Clustering using Pituitaries of GalϪ/Ϫ and Galϩ/ϩ Mice ϩ Ϫ the S179D gene set. D, Clustering using the Gal set. L, Pituitaries were isolated from Galϩ/ϩ and GalϪ/Ϫ mice and Lactation; NL, no lactation. incubated in media before analysisofthemedia for phos- phorylated (P-Prl, P) and unmodified (U-Prl, U) forms of Prl by silver-stained two-dimensional gel electrophoresis. Gels duced by S179D treatment far more closely resembles were quantified by densitometry, and the proportion of U-Prl the pattern of change produced by loss of Gal than and P-Prl are expressed as a percentage Ϯ SEM. P-Prl was loss of Prlr. 20 Ϯ 1.9% in Galϩ/ϩ and 31 Ϯ 2.1% in GalϪ/Ϫ, P Ͻ 0.0001 We then examined relatedness between these sam- by Student’s t test, changing the ratio of P-Prl:U-Prl from 4:1 ples from the point of view of each of the Venn sets in in Galϩ/ϩ to 2:1 in GalϪ/Ϫ.
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1877
Galϩ/ϩ mice, compared with 2:1 in GalϪ/Ϫ mice (Fig. 8).
Behavior of the Members of the Set of 35 Key Lactational Regulators Using the HC11 Mouse Mammary Cell Model
To examine the ability of Prl to directly regulate the expression of the 35 genes presented inFig. 5, we used the HC11 mouse mammary cell line. These cells initially proliferate to confluent density under the influ- ence of epidermal growth factor during the first3dof culture and then differentiate in response to Prl added atd4byforming domes and synthesizing -casein. Milk proteins, as an assay endpoint, are measured at d8. We measured the expression level of the 35 genes using Affymetrixchipsatd2,d3,d4,andd8. Of the 35 genes, we detected expression of 28, and their fold change in expression in response to Prl and the P value for this change are shown inFig. 9A. Genes are named where they showed a significant P value for changed expression, or the case of Cidea and Angptl4 where they showed a large fold change but nonsignif- icant P value due to a low level of expression at the edge of detectability. The genes Slc39a8, Slc34a2, Car2, Aldo3, Kcnk1, Ctsc, Elovl5, and Ugalt, which showed decreased expression in the models of failed secretory activation, all showed increased expression in response to Prl in HC11 cells with significant P values. Psmb9 and the transcription factors Cebpd and Sox4, which showed increased expression in all models of failed secretory activation, showed de- creased expression in response to Prl in HC11 cells. Four genes (Fdps, Acly, Gbp3, and Ctgf) showed the opposite pattern of regulation. Of the remaining 14 genes that were not directly regulated by Prl between d 4 and d 8, Erbb, Folr1, Sqle, Siat, G0s2, and Scd2 showed expression levels that increased from d2tod 4 and then reached a plateau, or dropped to zero (G0s2), indicating their regulation during the growth phase of these cells and confirming their role in the proliferative phase of mammary development. To examine the role of Stat5A we used short inter- fering RNA (siRNA) directed against Stat5A to reduce Stat5A expression. Cells were grown using the Fig. 9. Examination of the Expression of the 35 Genes Com- scheme outlined above but were transfected during mon to All Models of Failed Secretory Activation A, The expression of the 35 genes identified inFig. 5 were the growth phase atd2with either the Stat5 siRNA or measured in HC11 cells at 2 d, 3 d, 4 d, and8dusing the asiRNA directed against the green fluorescent protein experimental scheme outlined in Materials and Methods sec- (GFP), used as a control. Gene expression was mea- tion. RNA from three independent experiments was pooled in sured at d 4, d 6, and d 8. Stat5A protein levels at d 4 equal ratios, and gene expression was measured using Af- fymetrixchips. Results are expressed as fold change relative to the expression level at d 4 (when Prl was added), and the P value for the gene expression change from d4tod8is was used as a siRNA control, Stat5B was used as a speci- indicated. B, siRNA was used to knock down expression of ficity control, and the Stat5A-regulated milk proteins Wdnm1 Stat5A in HC11 cells. Representative Western blot for Stat5A and Csnb were used as positive controls. Results are ex- is shown in the top panel,with P value for the knockdown pressed as fold change relative to GFP control and P values calculated from densitometric quantification of three inde- indicate statistically significant differences at d 4 for three pendent experiments indicated in the bar graph. QPCR was independent experiments. Csnb was not expressed atd4but used to measure the effect of the loss of Stat5 on the ex- issignificantly reduced atd6tod8(P ϭ 0.05). NS, Nonsig- pression of the key lipogenic enzymes, Aldo3 and Scd2. GFP nificant (P ϭ 0.3).
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 1878 Mol Endocrinol, July 2005, 19(7):1868–1883 Naylor et al.•Transcriptome Change during Secretory Activation
were significantly (P ϭ 0.014) reduced to 25% of levels entiation. This conclusion is also supported by the seen in cells transfected with siRNA directed against observation that S179D treatment reduces Stat5 GFP (Fig. 9B), but recovered overd6tod8. We have phosphorylation (Fig. 3). observed a similar rapid recovery in HC11 cells with a These experiments are consistent with a study pub- number of other siRNAs, indicating this may be a lished during the course of these investigations in general feature of these cells. The level of the highly which Walker and colleagues (18) reported that S179D homologous Stat5B proteinremained unchanged, inhibited lactation in rats. In this report Walker and demonstrating the specificity of this approach. Reduc- colleagues used Northern blotting of S179D-treated tion of Stat5 caused a fall in the levels of the milk mouse mammary HC-11 cells to document a greater proteins Wdnm-1 (P ϭ 0.018, d 4; and P ϭ 0.06, induction of -casein relative to the ribosomal sub- combined d 6 and d 8) and Csnb (no expression,d4; units in response to S179D vs. U-Prl. Although it is and P ϭ 0.05 for combined d 6 and d 8 data) demon- initially tempting to directly contrast the induction of strating that we had reduced Stat5 levels sufficiently to -casein by S179D in vitro with the repression ob- have an effect on the expression of these Stat5-regu- served in vivo,suchacomparison involves the direct lated genes. We chose to examine whether reduction effect of S179D and U-Prl on the HC-11 -casein of Stat5 could reduce the levels of key lipogenic en- promoter in vitro compared with the developmental zymes Aldo3, the key step in the conversion of glucose effects of S179D and U-Prl on a complex tissue in vivo, to pyruvate, and Scd-2, the rate-limiting step in the making conclusions difficult to draw. The basisofthis production of acyl-CoenzymeAs. Knockdown of Stat5 inconsistency remains to be investigated but could be significantly reduced mRNA expression of both these a consequence of altered short to long receptor ratios enzymes at d 4 (Aldo3, P ϭ 0.06; and Scd-2, P ϭ (18) or the high levels of insulin and/or hydrocortisone 0.000016), and their mRNA levels recovered as Stat5 used in the in vitro system. levels recovered, with no significant expression differ- The in vitro antagonism of Prl action by S179D has ence at d 6–d 8. These experiments show that key been disputed by some (22, 23) and supported by points inlipogenesis fall under the control of Stat5. others (20), indicating that in vitro the action of S179D Overall, these experiments with the HC11 cells show remains to be fully explored. This is in stark contrast to the existence of direct mechanisticlinks between Prl, the in vivo situation inwhich S179D reproduces a Stat5, and the 35 genes identified as common to our number of phenotypes seen in the Prlr knockout three models of failed secretory activation. A large model (10, 19, 24–26), demonstrating a clear Prl an- proportion of the genes identified inFig. 5askey tagonist activity. To examine thispoint, we used global lactation genes fall under either the direct control of Prl gene expression as an endpoint to further analyze or under the control of other regulatory processes whether S179D acted as an agonist or as an antago- during HC11 cell growth. nist of Prl-induced alterations in gene expression in our model of S179D-induced failure of lactogenesis. We compared the pattern of altered mammary gene expression caused by treatment with S179D to the DISCUSSION alterations produced by the loss of a single Prlr allele, and to the alterations caused by the loss of Gal. All We have compared three models of failed secretory ac- three models exhibit a phenotypically indistinguish- tivation in mice to defineasmall set of genes that display able failure of lactogenesis. There were 87 genes (35 ϩ altered gene expression in all models, revealing a key set 52) that showed altered expression in response to loss of genes involved in the initiation of lactation. This data of a Prlr allele and treatment with S179D. Of these, 75 set has also allowed us to gain insight into the mecha- showed decreased expression and nine showed in- nisms underlying the partially penetrant Prlrϩ/Ϫ pheno- creased expression, in response to loss of a Prlr allele type and suggests a previously unknown mechanism of and treatment with S179D, a pattern demonstrating regulating Prl phosphorylation. S179D as an antagonist of Prl action. Two genes We have established a new mouse model of failure showed increased expression with S179D, but de- of secretory activation, using S179D Prl to inhibitthis creased expression in response to the loss of a Prlr phase of mammary gland development. In experi- allele, and one gene showed the inverse pattern; both ments reported inFigs. 1, 2, and 4, S179D treatment are patterns that indicate S179D was acting as a Prl during pregnancy reduced the expression of a broad agonist. Thus S179D is overwhelmingly acting as an range of genes involved in lactation, such as lipid and antagonist of Prl action on lactogenesis but, at the cholesterol synthesis enzymes, solute carriers, and level of the expression of specific genes, it has detect- lipid transport enzymes. Milk proteins were also dra- able agonist activity for three genes. This conclusion matically reduced. Specifically, -casein mRNA must also be viewed in the light of total Prl action. showed a large decrease in expression in S179D Prl- Overall, the loss of a Prlr allele affected the expression treated mice compared with saline-treated mice (Figs. of 654 genes, of which just 87 (13%) were altered by 1 and 4). Analysisofmilk protein levels also demon- S179D. Given the failure of lactogenesis in S179D- strated a reduction (Fig. 1). Together these findings treated animals this 13% was clearly a functionally indicate that lactation failed due to a failure of differ- important subset of Prl-regulated genes, but thissmall
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1879
subset indicates either that S179D is not a global to prolong the activity of Stat family members by con- antagonist of Prl action or that there are developmen- jugation (44). The key transcription factors Srebf1, tal effects of the absence of one Prlr allele that result controlling lipid metabolism genes (45), Cebp␦, in- in the disparity. Given the induced failure of lactation, volved inlipogenic responses and mammary develop- it cannot be argued that thissmall subset simply re- ment (46), and Sox4, a progesterone-responsive tran- sults from a suboptimal dose of S179D resulting in scription factor (47), were also found inthis set. The submaximal U-Prl antagonism, or a partial agonist ac- latter two transcription factors were expressed at a tivity of S179D. higher level in nonlactating animals, and our results Another aspect of thiscomparison was the discov- suggest that their loss of expression is required for ery of 128 genes that responded to S179D treatment, secretory activation. Other genes showing this pattern but not to loss of a Prlr allele or loss of Gal. This set included Erbb3, two probe sets interrogating Angptl4 containsahigh false discovery rate (102), demonstrat- expression, which isaninhibitor of lipoproteinlipase ing that most of these genes are false positives. Thus (48), proangiogenic Ctgf (49), and antiangiogenic S179D has a very restricted unique activity (off-target Thbs1 (50). A large number of genes have been impli- activity). The same caveat applies to the 128 (110 ϩ cated in the process of secretory activation by a de- 18) genes specific to the GalϪ/Ϫ set that could repre- tailed time-course study using wild-type FVB mice (6), sent a direct action of Gal that is independent of Prl andacomparison of data sets will allow Prl-regulated action, if not for the high false discovery rate. This genes to be distinguished from those regulated by the indicates this effect issmall and that almost all of Gal developmental process in general. We used the HC11 action during secretory activation isvia modulation of mammary epithelial cell model to show that a large Prl action by 1) control of serum Prl levels (12), 2) Gal proportion of these 35 genes were directly regulated modulation of Prl action at the mammary epithelial cell by Prl, and we knocked down Stat5A expression to (13), and 3) possibly also via regulation of Prl phos- demonstrate that two key genes in de novo synthesis phorylation (Fig. 8). It is during the transition from the of lipids from glucose, Aldo3 and Scd2, were respon- proliferative to secretory initiation phase at midpreg- sive to the levels of Stat5A expression, providing a nancy that Gal serum levels and mammary Gal recep- further mechanisticlink between our observations and tors are at theirhighest, conditions most suitable for the endocrine control of secretory activation. direct Gal action independent of Prl (13). Our data sets also allowed us to examine the To further examine the similarity between the effects changes in gene expression that resulted in lactation of S179D, the loss of Gal, and the loss of a Prlr allele in Prlrϩ/Ϫ mice. We detected a very strong proliferation we examined our profiles using hierarchical clustering signal in lactating Prlrϩ/Ϫ mammary glands, with al- to group them based on the similarity of changes in most all of the genes involved in the initiation of DNA their transcript profiles. Whereas the Venn analysis replication (41), and a number from the G2/M phase of was based on a change in gene expression among our the cell cycle, showing elevation in expression not only models of failed lactation irrespective of magnitude, above nonlactating Prlrϩ/Ϫ glands but also above hierarchical clustering groups experimental replicates Prlrϩ/ϩ glands. Notable among this set was the ele- together based upon the level of gene expression, vated expression of proliferating cell nuclear antigen allowing an estimate of overall similarity between ex- and Ki67, widely used markers of proliferation. pression profiles from their relative position in the Searching for a potential cause for this we found that computed dendogram. We also used principal-com- epidermal growth factor (Egf) was 4-fold higher and ponents analysis to cluster the experimental repli- ras family members Rab-18 and K-ras were 2-fold cates, with very similar results to those found with higher in lactating glands. As the Egf signaling path- hierarchical clustering. GalϪ/Ϫ and S179D-treated way results in Stat5 phosphorylation, this provides a glands consistently fell within the same cluster. This potential Prl-independent growth factor signal that approach showed that the pattern of gene expression could account for the increased levels of Stat5 phos- found in glands experiencing lactational failure due to phorylation seen in lactating Prlrϩ/Ϫ animals. Other S179D treatment was very similar to the pattern seen genes found to be elevated in lactating Prlrϩ/Ϫ glands in nonlactating glands from GalϪ/Ϫ mice and that both included many of the key lactational genes shown in were distant from the pattern of gene expression seen Fig. 5. Another gene elevated inthis group was in nonlactating Prlrϩ/Ϫ mice. This suggested that al- synuclein-␥, otherwise known as persyn or breast teration in the ratio of U-Prl to P-Prl may form part of cancer-specific gene 1, the increased expression of Gal’s modulation of lactogenesis. which is associated with aggressive breast cancer, A functional analysis of the 35 genes common to all increased metastasis, and activation of estrogen- three models of failed lactation was undertaken by driven transcription (51). Interestingly, apoptosis extensive literature searches, allowing almost all of genes were not prominent inthis group. these genes to be placed in the model of a lactating In summary, treatment of mice with a molecular mammary epithelial cell presented inFig. 6. This anal- mimic of phosphorylated Prl resulted infailed lactation ysis implicated a number of genes in lactation for the and impaired lobuloalveolar development that was as- first time. Examples include Cidea, a key metabolic sociated with reduced Stat5 activation in the mam- gene (43), and the ubiquitin-like Isg15, which is known mary gland. Transcriptome analysis of the secretory
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 1880 Mol Endocrinol, July 2005, 19(7):1868–1883 Naylor et al.•Transcriptome Change during Secretory Activation
activation phase of mammary gland development us- miniosmoticpump (Alzet OsmoticPumps Durect Corp., Cu- ing three models of failed lactation identified potential pertino, CA) containing either U-Prl or S179D, the molecular mimic of P-Prl; both hormones were prepared as described key regulatory genes for this process. Transcript pro- elsewhere (21). Either 0.6or1.2 g was delivered per 24 h. filing showed that S179D has actions that are predom- On the first day postpartum, maternal behavior of mothers inantly, but not exclusively, antagonistic to U-Prl-reg- was observed, pups were examined for the presence of milk, ulated gene expression in the mammary gland. and glands were taken for histological analysis. Increased cell proliferation was observed in Prlrϩ/Ϫ females able to lactate, providing mechanistic insight Histological Analysis into the partial penetrance of this phenotype. Prl treat- ment of HC11 cells demonstrated that many of the 35 Mammary whole mounts were made by spreading the gland on a glass slide before fixing in a 10% formalin solution. key genes were under direct Prl regulation and that Glands were defatted in acetone before carmine alum (0.2% others were associated with mammary epithelial cell carmine, 0.5% aluminum sulfate) staining overnight. The proliferation. Stat5 mediated Aldo3 and Scd2, key en- whole mount was dehydrated using a graded ethanol series zymes in de novo lipidbiosynthesis from glucose. followed by xylene treatment for 60 min and storage and Together these results provideasmall list of key genes photography in methyl salicylate (54). involved in secretory activation for whichanumber of applications can be envisaged. For example, these mRNA Isolation genes provide an excellent starting point for the iden- The fourth inguinal mammary gland was frozen inliquidni- tification of alleles that may provide enhanced lacta- trogen before storage at Ϫ80 C before use. Total RNA was tional performance inamarker-assisted selection pro- extracted using TRIZOL Reagent (Life Technologies, Gaith- cess incommercially valuable agricultural species. ersburg, MD) according to the manufacturer’s instructions. Alternatively, their study may help our understanding of lactation failure and other disorders of the breast. QPCR
QPCR was performed using LightCycler technology (Roche MATERIALS AND METHODS Clinical Laboratories, Indianapolis, IN). Primers were de- signed on the basisofmismatch to other genes. PCR reac- tions were performed in 10-l volumes with 1 l cDNA, 5 Animals pmol of each primer, and FastStart DNA Master SYBR Green I enzyme mix (Roche) as per manufacturer’s instructions. All experiments involving mice were performed under the Relative quantitation of the product was performed by com- supervision and in accordance with the regulations of the paring the crossing points of different samples normalized to Garvan/St Vincent’s Hospital Animal Experimental Ethics an internal control (-actin). Each cycle in the linear phase of Committee. GalϪ/Ϫ mice (12) were inbred on the 129OlaHsd the reaction corresponds to a 2-fold difference in transcript genetic background. Prlrϩ/Ϫ mice (7) were of mixed levels between samples. Each reaction was performed in 129SvPas/129OlaHsd genetic background. All animals were duplicate using pooled RNA from the three to six mammary specific pathogen free and housed with food and water ad glands per experiment. libitum with a 12-h light, 12-h dark cycle at 22 C and 80% mUgalt2 F a GGTGGTTGGAATAGAAGAGCACAC relative humidity. mUgalt2 R a CAAGACCGAGACCCAGGAAAAC mFolr1 F a TGGAGTTGGCGATTAGAGTCTGAC Two-Dimensional PAGE mFolr1 R a GAGGCAGGTGTCTTGGATAAAGTG mSiat1 F a TGTAAAATGGGGGTGACAATCC mSiat1 R a CTCTTGCTGACCTCTTGAAGGAAC Following decapitation the anterior pituitary was removed, mCyp51 F a AAAGGTAATGGGGTCGTGTAGTTG cut into 1-mm pieces, rinsed in PBS to remove material from mCyp51 R a GCACAGAATACGGGCAATGATAC damaged cells, and incubated in DMEM containing 0.1% mCuta F a TGTCCCAACGAAAAAGTCGC BSAfor2hat37Cinanatmosphere of water-saturated 5% mCuta R a AAAGGCATCAGGAGCAGGAGAG CO2. At the end of the 2-h incubation period, the medium was mCopz1 F a CAGCACAAGTGGGTTTGGAGTG removed and frozen before preparation for two-dimensional mCopz1 R a TGAGGAGAAGGAACACGGCAAG gel analysis. Two pituitaries were used per 2 mlofincubation mCsnd F a TATTACCCATCTACCCCCAGCC medium to allow sufficient Prl accumulation in the samples Ϫ/Ϫ mCsnd R a GAAACCCACAAGCAGACCTAACAC from the Gal mice. The proteins in the incubation medium mCsnb F a TTCACCTCCTCTCTTGTCCTCCAC Ϫ were precipitated in 4 vol of 20 C acetone overnight, col- mCsnb R a GGGGCATCTGTTTGTGCTTG lected by centrifugation, and then dissolved in urea lysis mWDMN1 TGACAATGACTACTGCCTGGGC buffer containing 9 M urea, 5% 2-mercaptoethanol, 4% am- mWDMN1 TTCCAAAACTGCGTGGGGGC pholines (pH 4–6.5) (Sigma Chemical Co.,St. Louis, MO). mWAP F a TGCCTCATCAGCCTCGTTCTTG Electrophoresis was performed according to the method of mWAP R a CTGGAGCATTCTATCTTCATTGGG Ho et al. (17). After electrophoresis the gel was silver stained mCIDEA F a GACTTCCTCGGCTGTCTCAATG (52), and the spots were identified by reference to standards mCIDEA R a GAAACTGATTCGTATCCACGCAG as described previously (53) and by reference to a corun mErbb3 F a TCTACCAAGTGGAACAGGAGAGGC sample that was subject to Western blot analysis (53). Spot mErbb3 R a CACCAACAAACGGAGTCTGGAAG intensity was analyzed using a Kodak image analysis system mKeratin18 F a CAAGATCATCGAAGACCTGAGGGC (Eastman Kodak Co., Rochester, NY). mKeratin18 R a TGTTCATAGTGGGCACGGATGTCC mAldo3 F TGCCAGTATGTTACAGAGAAGGTCC Phosphorylated and Unmodified Prl Treatment of Mice mAldo3 R CCGCTTGATAAACTCCTCAGTAGC mScd2 F GCTGGGGCGAGACTTTTGTAAAC On the morning of the observation of a vaginal plug, 6- to mScd2 R TGGCTTCTGGAACAGGAACTGC 8-wk-old mice were implanted witha0.25 l/h, 28-d Alzet mStat5a F CACAGGTGGAAGATTGGGGTTC
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1881
mStat5a R CCACTCCCCATCCAAAAACC the Oligofectamine protocol (Invitrogen). Serum free medium mStat5b R CGAATGGAGAAAAGGGATGGTG (800 l) and 200 l of the siRNA/Lipofectamine complexes mStat5b F GTTCCTCTGCCAGGTAGTCCATAG were added to each well of a six-well plate 24 h after plating the cells, and 4 h later 500 l of RPMI/30% fetal calf serum was added to each well. Western Analysis Acknowledgments Following RNA extraction frommammary glands using TRIZOL Reagent, protein was extracted according to the manufacturer’s instructions. Protein was separated using Received July 1, 2004. Accepted February 2, 2005. SDS-PAGE (Bio-Rad Laboratories, Hercules, CA), transferred Address all correspondence and requests for reprints to: to polyvinylidine difluoride (Millipore Corp., Bedford, MA), and Christopher J. Ormandy, Development Group, Cancer Re- blocked overnight with 2% fetal bovine serum,50mM sodium search Program, Garvan Institute of Medical Research, St. phosphate, 50 mM NaCl, and 0.1% Tween 20. Membranes Vincent’s Hospital, Sydney, New South Wales 2010, Austra- were incubated with one of the following primary antibodies: lia. E-mail: [email protected]. ␣-milk protein (Accurate Chemical&Scientific Corp., West- This work was supported by the National Health and Med- bury, NY), ␣-Stat5A (Upstate Biotechnology, Inc., Lake ical Research Council Australia, the New South Wales Cancer Placid, NY), ␣-phospho-Stat5, ␣-phospho-Erk1/2, ␣-Erk2, Council, the Clinical Research Center for Innovative Dairy ␣-phospho-Akt (S473), ␣-phospho-Akt (T308), ␣-Akt (Cell Products and the U.S. Army Breast Cancer Research Pro- Signaling Technology, Beverly, MA) or ␣--actin(Sigma). gram (BCRP) (to C.J.O.); National Institutes of Health Grant Protein (20 g) was loaded per lane except for ␣-milk protein DK 61005 and Grant 10PB-0127 from the California BCRP (to where 400 ng of protein was loaded. Specificbinding was A.M.W.). detected using horseradish peroxidase-conjugated second- Present address for M.J.N.: Developmental Biology Pro- ary antibodies (Amersham Biosciences, Arlington Heights, IL) gram,Victor Chang Cardiac Research Institute, 384 Victoria with Chemiluminescence Reagent (PerkinElmer, Norwalk, Street, Darlinghurst, New South Wales, 2010, Australia. CT) and Biomax Light Film (Eastman Kodak).
Transcript Profiling REFERENCES
Total RNA was extracted using TRIZOL Reagent (Invitrogen Life Science, Carlsbad, CA), purified using RNeasy Mini Kit 1. Hovey RC, Trott JF, Vonderhaar BK 2002 Establishing a (QIAGEN, Chatsworth, CA); cDNA synthesis was performed framework fort he functional mammary gland: from en- using Superscript II (Invitrogen Life Technologies), and syn- docrinology to morphology. JMammary Gland Biol Neo- thesisofbiotin-labeled cRNA was performed using BioArray plasia7:17–38 HighYield RNA Transcript labeling kit (Enzo Life Sciences, 2. Neville MC, McFadden TB, Forsyth I 2002 Hormonal Farmingdale, NY) and hybridized to Affymetrix MGU74Av2 regulation of mammary differentiation and milk secretion. GeneChips overnight according to manufacturer’s instruc- JMammary Gland Biol Neoplasia17:49–66 tions. Arrays were performed in duplicate using four to six 3. Robinson GW, McKnight RA, Smith GH, Hennighausen L glands per treatment group from two separate replicate ex- 1995 Mammary epithelial cells undergo secretory differ- periments. Analysis was performed using the Affymetrix Ge- entiation in cycling virgins but require pregnancy for the neChip version 5 software (MAS 5). Data were also analyzed establishment of terminal differentiation. Development as follows: Signal intensities of each gene were obtained 121:2079–2090 using the RMA function in the Affy package in R (http:// 4. Djiane J, Durand P 1977 Prolactin-progesterone antag- www.bioconductor.org). The RMA function, which involves onismin self regulation of prolactin receptors in the quantile normalization of oligonucleotide signals followed by mammary gland. Nature 266:641–643 estimation of the average perfect match signal intensity for 5. Naylor MJ, Lockefeer JA, Horseman ND, Ormandy CJ each probe set, has been shown to reduce variability and bias 2003 Prolactin regulates mammary epithelial cell prolif- when compared with the MAS 5.0 software (40, 55). Differ- eration via autocrine/paracrine mechanism. Endocrine ential expression was then assessed by ranked penalized t 20:111–114 statistics using lm.series and ebayes functions in the limma 6. Rudolph MC, McManaman JL, Hunter L, Phang T, Neville package in R (http://www.bioconductor.org). MC 2004 Initiation of lactation in the murine mammary gland: temporal analysisofacomplex biological switch with expression profiling and trajectory clustering. J HC11 Cell Culture Mammary Gland Biol Neoplasia8:287–308 7. Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, HC11 cells (a kind gift from Dr. Nancy Hynes) were main- Buteau H, Edery M, Brousse N, Babinet C, Binart N, Kelly tained in RPMI 1640 medium (Invitrogen) with 10% heat- PA 1997 Null mutation of the prolactin receptor gene inactivated fetal calf serum,5g/ml insulin, and 10 ng/ml Egf produces multiple reproductive defects in the mouse. (Sigma). Differentiation assays were performed by plating Genes Dev 11:167–178 1.5 ϫ 105 cells inasix-well plate in the above medium for 3d, 8. Wynick D, Hammond PJ, Akinsanya KO, Bloom SR 1993 during which time the cells grow to confluent density, fol- Galanin regulates basal and oestrogen-stimulated lac- lowed by incubation in mediawithout EGF for 24 h and then totroph function. Nature 364:529–532 Ϫ treatment with 10 6 M dexamethasone (Sigma) and 5 g/mlovine 9. Gallego MI, Binart N, Robinson GW, Okagaki R, Coschi- Prl (Sigma) in mediawithout EGF for4dfrom d4tod8. gano KT, Perry J, Kopchick JJ, Oka T, Kelly PA, Hen- nighausen L 2001 Prolactin, growth hormone, and epi- dermal growth factor activate Stat5 indifferent siRNA Transfection of HC11 Cells compartments of mammary tissue and exert different and overlapping developmental effects. Dev Biol 229: siRNA molecules were synthesized using the Silencer siRNA 163–175 Construction Kit(Ambion, Inc., Austin, TX). Two microliters of 10. Lucas BK, Ormandy CJ, Binart N, Bridges RS, Kelly PA 500 nM siRNA (final concentration of 1 nM) was mixed with 1998 Null mutation of the prolactin receptor gene pro- 200 l serum free RPMI, 2.5 lLipofectamine 2000 (Invitro- duces a defect in maternal behavior. Endocrinology 139: gen), and 200 l OPTI-MEM media (Invitrogen) according to 4102–4107
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 1882 Mol Endocrinol, July 2005, 19(7):1868–1883 Naylor et al.•Transcriptome Change during Secretory Activation
11. Kaplan LM, Gabriel SM, Koenig JI, Sunday ME, Spindel 28. Tuazon PT, Lorenson MY, Walker AM, Traugh JA 2002 ER, Martin JB, Chin WW 1988 Galanin is an estrogen- p21-activated proteinkinase ␥-PAK inpituitary secretory inducible, secretory product of the rat anterior pituitary. granules phosphorylates prolactin. FEBS Lett 515:84–88 Proc Natl Acad Sci USA 85:7408–7412 29. Brisken C, Kaur S, Chavarria TE, Binart N, Sutherland RL, 12. Wynick D, Small CJ, Bacon A, Holmes FE, Norman M, Weinberg RA, Kelly PA, Ormandy CJ 1999 Prolactin con- Ormandy CJ, Kilic E, Kerr NC, Ghatei M, Talamantes F, trols mammary gland development viadirect and indirect Bloom SR, Pachnis V 1998 Galanin regulates prolactin mechanisms. Dev Biol 210:96–106 release and lactotroph proliferation. Proc Natl Acad Sci 30. Gouilleux F, Moritz D, Humar M, Moriggl R, Berchtold S, USA 95:12671–12676 Groner B 1995 Prolactin and interleukin-2 receptors inT 13. Naylor MJ, Ginsburg E, Iismaa TP, Vonderhaar BK, Wyn- lymphocytes signal through a MGF-STAT5-like tran- ick D, Ormandy CJ 2003 The neuropeptide galanin aug- scription factor. Endocrinology 136:5700–5708 ments lobuloalveolar development. JBiol Chem 278: 31. Miyoshi K, Shillingford JM, Smith GH, Grimm SL, Wagner 29145–29152 KU, Oka T, Rosen JM, Robinson GW, Hennighausen L 14. Sinha YN 1995 Structural variants of prolactin: occur- 2001 Signal transducer and activator of transcription rence and physiological significance. Endocr Rev 16: (Stat) 5 controls the proliferation and differentiation of 354–369 mammary alveolar epithelium. J Cell Biol 155:531–542 32. Shillingford JM, Miyoshi K, Robinson GW, Grimm SL, 15. Oetting WS, Tuazon PT, Traugh JA, Walker AM 1986 Phos- Rosen JM, Neubauer H, Pfeffer K, Hennighausen L 2002 phorylation of prolactin. JBiol Chem 261:1649–1652 Jak2 is an essential tyrosine kinase involved in pregnancy- 16. Ho TWC, Kawaminani M, Walker AM 1993 Secretion of mediated development of mammary secretory epithelium. phosphorylated and nonphosphorylated monomer pro- Mol Endocrinol 16:563–570 lactin isoforms during rat pregnancy and pseudopreg- 33. Das R, Vonderhaar BK 1996 Involvement of SHC, GRB2, nancy. Endocr J 1:435–439 SOS and RAS in prolactinsignal transduction in mam- 17. Ho TW, Leong FS, Olaso CH, Walker AM 1993 Secretion mary epithelial cells. Oncogene 13:1139–1145 of specific nonphosphorylated and phosphorylated rat 34. Das R, Vonderhaar BK 1996 Activation of raf-1, MEK, prolactin isoformsatdifferent stages of the estrous cy- and MAP kinase in prolactin responsive mammary cells. cle. Neuroendocrinology 58:160–165 Breast Cancer Res Treat 40:141–149 18. Kuo CB, Wu W, Xu X, Yang L, Chen C, Coss D, Birdsall 35. Berlanga JJ, Fresno Vara JA, Martin-Perez J, Garcia-Ruiz B, Nasseri D, Walker AM 2002 Pseudophosphorylated JP 1995 Prolactin receptor is associated with c-src ki- prolactin (S179D PRL) inhibits growth and promotes nase in rat liver. Mol Endocrinol 9:1461–1467 -casein gene expression in the rat mammary gland. Cell 36. al-Sakkaf KA, Dobson PR, Brown BL 1997 Prolactin in- Tissue Res 309:429–437 duced tyrosine phosphorylation of p59fyn may mediate 19. Bridges R, Rigero B, Byrnes E, Yang L, Walker A 2001 phosphatidylinositol 3-kinase activation in Nb2 cells. J Central infusions of the recombinant human prolactin Mol Endocrinol 19:347–350 receptor antagonist, S179D-PRL, delay the onset of ma- 37. Binart N, Imbert-Bollore P, Baran N, Viglietta C, Kelly PA ternal behavior in steroid-primed, nulliparous female rats. 2003 A short form of the prolactin (PRL) receptor is able Endocrinology 142:730–739 to rescue mammopoiesis in heterozygous PRL receptor 20. Schroeder MD, Brockman JL, Walker AM, Schuler LA mice. Mol Endocrinol 17:1066–1074 2003 Inhibition of prolactin (PRL)-induced proliferative 38. Draghici S, Khatri P, Martins RP, Ostermeier GC, Krawetz signals in breast cancer cells by a molecular mimicof SA 2003 Global functional profiling of gene expression. phosphorylated PRL, S179D-PRL. Endocrinology 144: Genomics 81:98–104 5300–5307 39. Ormandy CJ, Naylor M, Harris J, Robertson F, Horseman 21. Chen TJ, Kuo CB, Tsai KF, Liu JW, Chen DY, Walker AM ND, Lindeman GJ, Visvader J, Kelly PA 2003 Investiga- 1998 Development of recombinant human prolactin re- tion of the transcriptional changes underlying functional ceptor antagonists by molecular mimicry of the phos- defects in the mammary glands of prolactin receptor phorylated hormone. Endocrinology 139:609–616 knockout mice. Recent Prog Horm Res 58:297–323 22. BernichteinS,Kinet S, Jeay S, Llovera M, Madern D, Martial 40. Bolstad BM, Irizarry RA, Astrand M, Speed TP 2003 A JA, Kelly PA, Goffin V 2001 S179D-Human PRL, a pseu- comparison of normalization methods for high density dophosphorylated human PRL analog, is an agonist and oligonucleotide array data based on variance and bias. not an antagonist. Endocrinology 142:3950–3963 Bioinformatics 19:185–193 23. Bernichtein S, Kayser C, Dillner K, Moulin S, Kopchick 41. Bell SP, Dutta A 2002 DNA replication in eukaryotic cells. JJ, Martial JA, Norstedt G, Isaksson O, Kelly PA, Goffin Annu Rev Biochem 71:333–374 42. Naylor MJ, Ormandy CJ 2002 Mouse strain-specific pat- V 2003 Development of pure prolactin receptor antago- terns of mammary epithelial ductal side branching are nists. JBiol Chem 278:35988–35999 elicited by stromal factors. Dev Dynam 225:100–105 24. Xu X, Kreye E, Kuo CB, Walker AM 2001 A molecular 43. Zhou Z, Yon Toh S, Chen Z, Guo K, Ng CP, Ponniah S, mimic of phosphorylated prolactin markedly reduced tu- Lin SC, Hong W, Li P 2003 Cidea-deficient mice have mor incidence and size when DU145 human prostate lean phenotype and are resistant to obesity. Nat Genet cancer cells were grown in nude mice. Cancer Res 61: 35:49–56 6098–6104 44. Kim KI, Zhang DE 2003 ISG15, not just another ubiquitin- 25. Robertson FG, Harris J, Naylor MJ, Oakes SR, Kindblom like protein. Biochem Biophys Res Commun 307:431–434 J, Dillner K, Wennbo H, Tornell J, Kelly PA, Green J, 45. Horton JD, Goldstein JL, Brown MS 2002 SREBPs: ac- Ormandy CJ 2003 Prostate development and carcino- tivators of the complete program of cholesterol and fatty genesis in prolactin receptor knockout mice. Endocrinol- acid synthesis in the liver. JClin Invest 109:1125–1131 ogy 144:3196–3205 46. Mason RJ, Pan T, Edeen KE, Nielsen LD, Zhang F, Long- 26. Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, phre M, Eckart MR, Neben S 2003 Keratinocyte growth Buteau H, Edery M, Brousse N, Babinet C, Binart N, Kelly factor and the transcription factors C/EBP␣, C/EBP␦, PA 1997 Null mutation of the prolactin receptor gene and SREBP-1c regulate fatty acid synthesis in alveolar produces multiple reproductive defects in the mouse. type II cells. JClin Invest 112:244–255 Genes Dev 11:167–178 47. McGowan EM, Clarke CL 1999 Effect of overexpression 27. Wang YF, Liu JW, Mamidi M, Walker AM 1996 Identifi- of progesterone receptor A on endogenous progestin- cation of the major site of rat prolactin phosphorylation sensitive endpoints in breast cancer cells. Mol Endocri- as serine 177. JBiol Chem 271:2462–2469 nol 13:1657–1671
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 Naylor et al.•Transcriptome Change during Secretory Activation Mol Endocrinol, July 2005, 19(7):1868–1883 1883
48. Yoshida K, Shimizugawa T, Ono M, Furukawa H 2002 52. Oakely BR, Kirsh DR, MorrisNR1980Asimplified ultra- Angiopoietin-like protein4is a potent hyperlipidemia- sensitive silver stain for detecting proteins in polyacryl- inducing factor in mice and inhibitor of lipoproteinlipase. amide gels. Anal Biochem 105:361–363 JLipid Res 43:1770–1772 53. Oetting WS, Walker AM 1985 Intracellular processing of 49. Brigstock DR 2003 The CCN family: a new stimulus prolactin. Endocrinology 117:1565–1569 package. J Endocrinol 178:169–175 54. Bradbury JM, Edwards PA, Niemeyer CC, Dale TC 1995 50. Watnick RS, Cheng YN, Rangarajan A, Ince TA, Wein- Wnt-4 expression induces a pregnancy-like growth pat- berg RA 2003 Ras modulates Myc activity to repress tern in reconstituted mammary glands invirgin mice. Dev thrombospondin-1 expression and increase tumor an- Biol 170:553–563 giogenesis. Cancer Cell 3:219–231 55. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, 51. Wu K, Weng Z, Tao Q, LinG,WuX,Qian H, Zhang Y, Antonellis KJ, Scherf U, Speed TP 2003 Explora- Ding X, Jiang Y, Shi YE 2003 Stage-specific expression tion, normalization, and summaries of high density of breast cancer-specific gene ␥-synuclein. Cancer Epi- oligonucleotide array probe level data. Biostatistics demiol Biomarkers Prev 12:920–925 4:249–264
Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.
Downloaded from mend.endojournals.org at Univ New South Wales Biomedical Library on December 6, 2006 0888-8809/06/$15.00/0 Molecular Endocrinology 20(5):1177–1187 Printed in U.S.A. Copyright © 2006 by The Endocrine Society doi: 10.1210/me.2005-0473
Socs2 and Elf5 Mediate Prolactin-Induced Mammary Gland Development
Jessica Harris, Prudence M. Stanford, Kate Sutherland, Samantha R. Oakes, Matthew J. Naylor, Fiona G. Robertson, Katrina D. Blazek, Michael Kazlauskas, Heidi N. Hilton, SergioWittlin, Warren S. Alexander, Geoffrey J. Lindeman, Jane E. Visvader, and Christopher J. Ormandy Garvan Institute of Medical Research (J.H., P.M.S., S.R.O., M.J.N., F.G.R., K.D.B., M.K., H.N.H., C.J.O.), St. Vincent’s Hospital Darlinghurst, New South Wales 2010, Australia; and Walter and Eliza Hall Institute of Medical Research (K.S., S.W., W.S.A., G.J.L., J.E.V.), Parkville, Victoria 3050, Australia
The proliferative phase of mammary alveolar mor- of cytokine signaling 2), and the ets transcription phogenesis is initiated during early pregnancy by factor, E74-like factor 5 (Elf5). Homozygous null rising levels of serum prolactin and progesterone, mutation of Socs2 rescued the failure of lactation establishing a program of gene expression that is and reduction of mammary signal transducer and ultimately responsible for the development of the activator of transcription 5 phosphorylation that lobuloalveoli and the onset of lactation. To explore characterizes Prlr heterozygous mice, demonstrat- this largely unknown genetic program, we con- ing that mammary Socs2 is a key regulator of the structed transcript profiles derived from trans- prolactin-signaling pathway. Reexpression of Elf5 planted mammary glands formed by recombina- in Prlr nullizygous mammary epithelium restored tion of prolactin receptor (Prlr) knockout or wild- lobuloalveolar development and milk production, type mammary epithelium with wild-type demonstrating that Elf5 is a transcription factor mammary stroma. Comparison with profiles de- capable of substituting for prolactinsignaling. rived from prolactin-treated Scp2 mammary epi- Thus, Socs2 and Elf5 are key members of the set of thelial cells produced a small set of commonly pro- prolactin-regulated genes that mediate prolactin- lactin-regulated genes that included the negative driven mammary development. (Molecular Endo- regulator of cytokine signaling, Socs2 (suppressor crinology 20: 1177–1187, 2006)
AMMARY GLAND development differs from the and secretory activation after parturition. At weaning Mdevelopment of most other organs as it pro- the gland commences involution, with the loss of most ceeds in adults in response to endocrine changes of the epithelial component gained during the preced- associated with the timing of reproductive events. The ing lactation (1–3). hormonal changes of puberty induce the ductal mor- Mouse knockout models have revealed both the phogenesis phase of development, when the mam- requirement for these hormones for mammary devel- mary rudiment developed in utero forms terminal end opment and theircomplex interactions, which involve buds that elongate and bifurcate to fill the mammary receptors located in the ovary and pituitary, in addition fat pad with a branched ductal network. Ductal density to the mammary epithelium and stroma. Ductal mor- then increases due to secondary and tertiary side phogenesis is initiated by rising estrogen viaacom- branching in response to each estrous or menstrual plex mechanism (4–6). Progesterone is required for cycle. The hormonal changes of pregnancy cause the ductal side branching and alveolar bud formation after gland to enter the alveolar morphogenesis phase, puberty and the formation of lobuloalveolar structures characterized by an initial proliferation phase, during during pregnancy (7). Loss of prolactin (8), or the pro- which the alveolar architecture is established, followed lactin receptor (Prlr) (9, 10), also prevents side branch- by the onset of lactation comprised by phases of ing after puberty, but indirectly viafailure of proges- secretory initiation during the later part of pregnancy terone secretion from the corpora lutea (11, 12). During pregnancy, complete loss of Prlr stalls development at First Published Online February 9, 2006 an early point in the alevolar proliferation phase, after Abbreviations: Elf5, E74-like factor 5; Expi, extracellular the formation of alveolar buds (9). In contrast, loss of proteinase inhibitor; H&E, hematoxylin and eosin; IRES, in- ternal ribosome entry site; MEC, mammary epithelial cell; justasingle Prlr allele has no effect during the early Plet-1, placenta-expressed transcript 1; Prlr, prolactin recep- alveolar proliferation phase, and development pro- tor; Socs2, suppressor of cytokine signaling 2; Stat, signal ceeds past the formation of alveolar buds to produce transducer and activator of transcription. lobuloalveolar structures with normal architecture. Molecular Endocrinology is published monthly by The These structures fail to fully differentiate late in the Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine proliferative phase or early in the secretory initiation community. phase, however, and lactation fails during secretory
1177 Downloaded from mend.endojournals.org on August 27, 2006 1178 Mol Endocrinol, May 2006, 20(5):1177–1187 Harris et al.•Elf5, Socs2, and Mammary Development
activation (9). The effects of prolactin on lobuloalveolar of Prlrϩ/ϩ to PrlrϪ/Ϫ material provides a very large development are exerted via the mammary epithelium contrast in prolactin action, much greater than can be and not the stroma (9, 13). Ahierarchy of action is achieved using prolactin treatment of wild-type epi- apparent from these and other studies. Estrogen is thelium. This strategy also allowed epithelial patterns essential from the earliest stage of ductal develop- of gene expression to be distinguished from genes ment, and progesterone is required for subsequent expressed in the entire gland. The use of transplanta- side branching. Prolactin and progesterone are nec- tion removes the confounding effects of the loss of the essary for the proliferative phase of alveolar develop- Prlr from the endocrine system, as the only cells car- ment. A further increase in prolactin-generated signal- rying the null mutation of the Prlr occur in the mam- ing is required for functional differentiation during mary epithelium of our experimental animals. It also secretory initiation and activation. prevents the lymph node diluting the mammary RNA Little is known about the program of altered gene pool. Asmall selected list of genes from this experi- expression that drives these developmental events. To ment has been previously published (13), and the screen for key members of this program,wecombined complete list of epithelial genes that decreased in transcript profiling with two contrasting models of pro- PrlrϪ/Ϫ epitheliumis supplied in supplemental Fig. 1 lactin action: an in vivo model inwhich prolactin action published as supplemental data on The Endocrine is ablated specifically in the mammary epithelial cells Society’s Journals Online web site at http://mend. (MECs) and the Scp2 cell model of augmented pro- endojournals.org. To further increase the discrimina- lactin action. These models provide multiple contrasts, tory power of our analysis we have added a model of such as negative prolactin action vs. positive prolactin positive prolactin action. SCp2 cells grown on matrigel action, whole tissue vs. cultured cells, and proliferative were treated for 48 h with insulin and hydrocortisone, phase vs. secretory activation phase, and offer a and either with or without prolactin. Supplemental Fig. highly selective set of overlapping criteria that we have 2 published as supplemental data on The Endocrine exploited to reduce the normally large set of genes Society’s Journals Online web site at http://mend. produced by transcript-profiling experiments to a endojournals.org. lists the genes that increased in re- small and focused set of prolactin-regulated genes, sponse to prolactin. To distillasmall set of genes for which were validated by quantitative PCR. Two of further investigation we combined the Scp2 cell data these genes, Socs2 and Elf5 (E74-like factor 5), are with that of the mammary epithelial transplants. We shown by geneticcomplementation to rescue the de- screened for genes that showed decreased expres- ϩ/Ϫ Ϫ/Ϫ velopmental defects seen in Prlr and Prlr mam- sion at any day in the PrlrϪ/Ϫ epithelial transplants, mary glands, respectively, demonstrating their crucial that increased expression in one of the two SCp2 roles in mediating the developmental signal delivered transcript profiles, and that exhibited epithelial expres- by prolactin. sion at any day in the PrlrϪ/Ϫ epithelial transplants. This three-way selection criteria resulted inasmall set of genes. Quantitative RT-PCR was used to verify the RESULTS expression patterns of thissmall set, allowing the ex- clusion of a number and resulting inasmall validated Identification of Elf5 and Socs2 set that comprised the milk proteins caseins ␣, , and , Expi (extracellular proteinase inhibitor), the ets tran- Both the epithelium and stromaofthemurine mam- scription factor Elf5, and Plet-1 (placenta-expressed mary gland express the Prlr (14, 15), but Prlr is only transcript 1) (Table 1). Plet-1 is expressed poorly in required in the epithelium for lobuloalveolar develop- humans due to degradation of its splice acceptor and ment (13). We transcript profiled wild-type mammary donor sequences (16) and so was not investigated fat pads cleared of endogenous epithelium,orwild- further. Elf5 was chosen from this set for further type mammary fat pad cleared of endogenous epithe- analysis. lium and transplanted with Prlrϩ/ϩ or PrlrϪ/Ϫ epithe- In the Scp2 cells only five genes showed a suffi- lium,allat2,4,and6dofpregnancy. The comparison ciently robust increase of expression in response to
Table 1. Genes Selected by Searching for Probe Sets that Showed Epithelial-Specific Expression, Decreased in Glands Formed from PrlrϪ/Ϫ Compared with Prlrϩ/ϩ Epithelium, and Increased in Response to Prolactin Treatment of Scp2 Cells Probe Set ID Sequences Derived from Gene Title Gene Symbol Function 103051_at X93037 Extracellular proteinkinase inhibitor Expi Milk protein 96030_at M36780 Casein-alpha Csn␣ Milk protein 99065_at M10114 Casein-kappa Csn Milk protein 99130_at X04490 Casein-beta Csn Milk protein 97413_at AI121305 Placenta-expressed transcript 1 Plet-1 Unknown 103283_at AF049702 E74-like factor 5 Elf5 Transcription Expressed sequence tags not shown.
Downloaded from mend.endojournals.org on August 27, 2006 Harris et al.•Elf5, Socs2, and Mammary Development Mol Endocrinol, May 2006, 20(5):1177–1187 1179
prolactin to allow detection in both experimental rep- licates. We believe this was due to variation in culture conditions between replicates, which serendipitously identified only the most robust changes in gene ex- pression in response to prolactin. This set consisted of four milk proteins (caseins , ␣, and and Expi) and Socs2. Examination of the transplant data showed that Socs2 decreased expression in response to the loss of the Prlr (average –1.8-fold), but was only called decreasingatd4byMAS4. Socs2 did not show epithelial expression as a signal was detected in the stroma. Given these data indicating prolactin regula- tion of Socs2 expression, and our previous demon- stration of the involvement of the family member Socs1 in prolactin-directed mammary development (17), we chose to also analyze the role of Socs2 as a mediator of the mammary response to prolactin.
Expression of Elf5 and Socs2 in the Transcript- Profiling Experiments
We used quantitative PCR to examine Elf5 and Socs2 expression in the transcript-profiling experiments. Elf5 expression was lower in glands formed from PrlrϪ/Ϫ epithelium compared with Prlrϩ/ϩ epithelium at all times, and was absent from the mammary stroma. Elf5 Fig. 1. Candidate Genes Selected by Transcript Profiling levels increased in Scp2 cells when they were treated Expression pattern of Elf5 (panel A) and Socs2 (panel B) in the transplants and Scp2 experiments by quantitative RT- with prolactin(Fig. 1A). Socs2 expression was also PCR. Solid bars show average difference, the MAS4 mea- decreased in response to a loss of the Prlr from the surement of gene expression level. Hatched bars show gene mammary and was increased by prolactin treatment of expression measured by quantitative PCR using an absolute Scp2 cells. Socs2 expression was clearly seen in both quantification method that reports transcripts per microgram the epithelium and stroma(Fig. 1B). We sought to of total RNA. Stippled bars show gene expression measured demonstrate that these genes are essential members by quantitative PCR using a relative method that reports fold of the prolactin-directed program of gene expression change relative to control. CF, Cleared fat pad; Preg, preg- that results in mammary development during preg- nancy; QPCR, quantitative PCR. nancy by a geneticcomplementation approach.
Socs2 Expression during Mammary Development genetic background, failure of lactation occurs in 100% of Prlrϩ/Ϫ females (18). Whole-mount and his- We examined the pattern of Socs2 expression in the tological analyses of mammary glands from wild-type mammary gland using in situ hybridization. Socs2 was animals (Socs2ϩ/ϩ Prlrϩ/ϩ) revealed normal lobuloal- expressed throughout mammary ontogeny, with in- veolar development (Fig. 3, A and B) whereas tense staining of epithelial cells, particularly during Socs2ϩ/ϩ Prlrϩ/Ϫ females (four of four) showed mark- pregnancy and lactation (Fig. 2). In contrast, the sense edly reduced lobuloalveolar development and failed control showed little background staining (Fig. 2A). lactation (Fig. 3, C and D). In contrast, deletion of both Socs2 expression was not restricted to the epithelium Socs2 alleles resulted incomplete rescue of lactation as a moderate signal intensity was detected in the in Socs2Ϫ/Ϫ Prlrϩ/Ϫ females (seven of seven), with all surrounding stroma, adipocytes, and vasculature. pups surviving. There was no evidence of rescue by Socs2 expression increased in pregnant samples deletion of a single Socs2 allele, because pups of compared with virgin. Socs2ϩ/Ϫ Prlrϩ/Ϫ mothers contained little milk in their stomachs. Western blot analysistomeasure milk pro- Loss of Socs2 Restores Lactogenesis in Prlr؉/؊ tein expression in mammary gland lysates revealed Females that Socs2 deficiency led to restoration of whey acidic protein, ␣- and -casein production, comparable to To determine whether loss of the negative regulator that seen inwild-type mammary glands (Fig. 4A). Anal- Socs2 could rescue the failure of lactation observed in ysis of Stat5 activation showed that heterozygous loss Prlrϩ/Ϫ mice, we generated females that were null for of the Prlr greatly reduced Stat5 phosphorylation and Socs2 and heterozygous for the Prlr gene (Socs2Ϫ/Ϫ, that this was restored inanimals that also carried a Prlrϩ/Ϫ)byinterbreeding the respective targeted mice, loss of the Socs2 gene (Fig. 4B) These results dem- both of which are on a C57Bl6 background. On this onstrate that loss of Socs2 rescued the lactational
Downloaded from mend.endojournals.org on August 27, 2006 1180 Mol Endocrinol, May 2006, 20(5):1177–1187 Harris et al.•Elf5, Socs2, and Mammary Development
Fig. 3. Null Mutation of the Socs2 Gene Rescues Mammary Gland Development in Prlrϩ/Ϫ Mice C57Bl6 mice carrying null mutations of Socs2 or Prlr were crossed, and mammary gland development was compared among the resulting genotypes at 1 d post partum by whole- mount (left panels A, C, and E) or H&E histology (right panels B, D, and F).
Elf5 Expression inMammary Gland and Breast
Examination of Elf5 expression by quantitative RT- PCR showed that a massive increase in Elf5 levels occurs in the mammary gland during pregnancy. In the virgin mammary gland, Elf5 was expressed at levels within the same order of magnitude as other epithelial tissues, approximately 104 copies per 25 ng total RNA (Fig. 5A, bars). Pregnancy induced a very large in- crease in Elf5 expression to a peak of about 600,000 transcripts per 25 ng total RNA atd2oflactation. Levels remained high throughout lactation. These RNA levels were mirrored by a large increase in Elf5 protein, Fig. 2. Mammary Socs2 Expression Increases during Preg- from low levels that were inconsistently detectable by nancy and Lactation Western blot invirgin and early pregnant glands to In situ hybridization was used to examine the pattern of high levels during lactation. Involution of the mammary Socs2 expression during the indicated stages of mammary gland induced by pup removal at d 15 of lactation had gland development. Sense control hybridizations (panel A) an immediate effect on Elf5 protein levels, which be- showed no signal. Socs2 was seen at increased levels during came undetectable within a day and preceded the fall pregnancy and lactation (compare panel B with panels C and D) in the epithelium with a weaker signal in the stroma. Socs2 in mRNA expression (Fig. 5A, blot). Immunohisto- expression remained high during early involution (panel E). chemistry showed that Elf5 was located predomi- ISH, In situ hybridization. nantly in the nuclei of the keratin 18-expressing lumi- nal mouse MECs, with approximately half of these cells clearly positive for the protein(Fig. 5B). Elf5 was failure produced by loss of a single Prlr allele viaa not expressed in cells located at the basal membrane mechanisminvolving Stat5 activation, establishing that expressed high molecular weight keratin. Asimilar Socs2 as an attenuator of prolactinsignaling in the pattern of expression was seen in normal human mammary gland in vivo. breast (data not shown).
Downloaded from mend.endojournals.org on August 27, 2006 Harris et al.•Elf5, Socs2, and Mammary Development Mol Endocrinol, May 2006, 20(5):1177–1187 1181
Fig. 4. Null Mutation of the Socs2 Gene Rescues Milk Pro- tein Expression and Mammary Stat5 activation in Prlrϩ/Ϫ Mice Prlrϩ/Ϫ mice showed a failure of mammary development associated with reduced milk protein synthesis(top panels) and Stat5 phosphorylation (lower panels); however, mice that were Prlrϩ/Ϫ and Socs2Ϫ/Ϫ showed near-normal levels of milk protein synthesis and Stat5 phosphorylation. p-Stat-5, phosphorylated Stat5; WAP, whey acidic protein.
Reexpression of Elf5 Restores Normal Development to Prlr؊/؊ Mammary Epithelium
To investigate whether Elf5 could compensate for loss of the Prlr in the mammary epithelium, we reexpressed Ϫ/Ϫ Elf5 in Prlr MECs by infection with the PolyPOZ Fig. 5. Elf5 Expression during Mammary Gland Develop- retrovirus (19) encoding Elf5. We then transplanted the ment resulting heterogeneous population of infected and A, The level of Elf5 was measured by quantitative RT-PCR uninfected MECs into the mammary fat pad (previ- using an absolute method, and by Western blot at various ously cleared of endogenous epithelium)ofimmuno- stages during mammary gland development as indicated. B, compromised Rag1Ϫ/Ϫ animals. As a control we in- Immunohistochemistry was used to examine the sites of Elf5 fected and transplanted PrlrϪ/Ϫ MECs with empty expression in the mammary gland at d 12 of pregnancy. An Elf5-blocking peptide was used as a control. Luminal cells polyPOZ. The animals were mated 12 wk after the and basal cells were distinguished by the expression pattern transplant, and the mammary glands collected at 1 d of keratin 18 (K18) and high-molecular weight keratin post partum. The experiment was repeated on four (HMWK) respectively. IHC, Immunochemistry. separate occasions. In total we attempted to recon- stitute nine control PrlrϪ/Ϫ mammary glands using empty polyPOZ-infected MECs. Of these, five glands veoli along the ducts, to lobules that appeared normal were successfully reconstituted, and all showed the but smaller insize than those seen in endogenous typical PrlrϪ/Ϫ defect of stalled lobuloalveolar devel- glands. Figure 6D shows an example that combines opment after the formation of alveolar buds (Fig. 6A). complete rescue (area immediately to the left of the Hematoxylin and eosin (H&E) histology showed the arrow)with incomplete rescue, seen as ducts covered formation of alveolar buds but no lobuloalveoli (Fig. with single alveoli (arrow). In this example both regions 6B), and the lumen of these structures stained very were connected viaacommon ductal network. H&E weakly with an antibody directed against mouse milk histology showed that the large lobules contained oil (Fig. 6C). We successfully reconstituted 28 mammary droplets and colostrum (Fig. 6E), and these lobules glands from a total of 38 attempts using PrlrϪ/Ϫ MECs stained intensely with the antibody raised against that we had infected with Elf5-polyPOZ. Of these, 10 mouse milk (Fig. 6F). The ducts covered with single showed restoration of alveolar morphogenesiswith alveoli (Fig. 6, A–C), were more developed than any- examples seen in all four experiments. The pattern of thing present in PrlrϪ/Ϫ glands as they showed multi- restoration was chimeric, comprised of completely ple single alveoli (Fig. 6E, arrow), but the lumen of rescued lobules and areas that showed varying de- these structures did not stain strongly with the antimilk grees of incomplete rescue. The areas of incomplete antibody (Fig. 6F, arrow). Figure 6, G–I, shows a rescued varied from the induction of small single al- PrlrϪ/Ϫ gland displaying complete rescue in most ar-
Downloaded from mend.endojournals.org on August 27, 2006 1182 Mol Endocrinol, May 2006, 20(5):1177–1187 Harris et al.•Elf5, Socs2, and Mammary Development
Fig. 6. Expression of Elf5 in PrlrϪ/Ϫ Mammary Gland Rescues Mammary Gland Development PrlrϪ/Ϫ MECs were infected with the polyPOZ retrovirus and then transplanted to the cleared mammary fat pad of host Rag1Ϫ/Ϫ mice and made pregnant 12 wk after transplant. A–C, Use of polyPOZ without the Elf5 construct (empty vector) had no effect on the failure of development seen in PrlrϪ/Ϫ transplants during pregnancy in all cases. D–F, Use of a polyPOZ Elf5 construct resulted in a rescue of mammary gland development that showed a chimeric pattern of complete rescue mixed with partial rescue (arrows). G–I, Some glands showed near-complete rescue that mimicked development seen in the endogenous glands. J–L, Endogenous glands. IHC, Immunohistochemistry. eas, but scanning of the tissue showed occasional cases often had smaller alveoli with smaller lumens small areas with structures identical to those indicated than wild-type transplants or endogenous glands. Milk by arrows inFig. 6, D–F. In examples showing the least staining showed that these lobules synthesized milk at degree of rescue, the whole mounts appear as a duc- wild-type levels but that the single alveoli located tal tree covered insingle alveoli,similar to that indi- along the ducts showed less milk protein expression. cated by arrow in 6D, with a just small portion con- This is presumably due to development of a mammary taining lobules that stain intensely with the antimilk tree from a heterogeneous population of infected and antibody (data not shown). Of the 10 rescued glands, uninfected MECs, which have contributed to a differ- two showed a chimeric rescue similar to Fig. 6, D–F, ent extent to the developing epithelium. Where the four showed almost complete rescue similar to Fig. 6, ratio of cells expressing high levels of Elf5 ishigh, G–I, and four showed a chimeric rescue with just a lobuloalveolar development proceeds further. Our small region of lobules. In all cases, development had vector contained a LacZ marker under the control of proceeded past the production of the sparse alveolar an internal ribosome entry site (IRES). We were able to buds that characterize the PrlrϪ/Ϫ defect. H&E staining visualize a weak Lacz signal in MECs after PolyPOZ showed that the lobules that had formed in these infection, which showed the cultures to be a mixture of
Downloaded from mend.endojournals.org on August 27, 2006 Harris et al.•Elf5, Socs2, and Mammary Development Mol Endocrinol, May 2006, 20(5):1177–1187 1183
Elf5-expressing and nonexpressing cells, but we were unable to successfully visualize LacZ in the resulting transplants. The use of an IRES results in much lower expression of the marker compared with the test gene, and so we believe that lack of LacZ sensitivity pre- vented visualization. RT-PCR investigation of PrlrϪ/Ϫ glands that showed rescued development by infection with PolyPOZ-Elf5-LacZ retrovirus demonstrated the presence of LacZ in these glands. We examined the level of Elf5 expression in these glands. In PrlrϪ/Ϫ glands infected with the empty vec- tor, endogenous Elf5 was seen as a weak nuclear signal of the luminal epithelial cells with a columnar shape (Fig. 7, A and B). About half of the luminal epithelial cells stained positive for Elf5. In PrlrϪ/Ϫ glands showing partial rescue of alveolargenesis, higher levels of Elf5 were detected in the nuclei,which now showed an oval shape, and weak cytoplasmic Elf5 staining was seen (Fig. 7, C and D). Still higher levels of nuclear and cytoplasmic Elf5 were detected in areas of complete rescue that stained strongly for milk (Fig. 7, E and F). Both the level of expression and the nuclear/cytoplasmic localization were similar to the levels seen in endogenous glands, where Elf5 was seen as a strong cytoplasmicsignal and an intense red-brown staining of large rounded nuclei that com- pletely obscured the hematoxylin nuclear stain(Fig. 7, G and H). Use of a peptide block or no Elf5 antibody (Fig. 7, I and J) showed the nuclear and cytoplasmic staining to be specific but showed the adipocyte mar- gin staining to be nonspecific. Thus, complete rescue of lobuloalveolargenesisbyinfection with an Elf5-pro- ducing retrovirus was associated with restoration of Elf5 expression and subcellular distribution in a way that mimicked the level and pattern seen in endoge- nous glands. Incomplete rescue was associated with a conversion of columnar to oval nuclei and a level of Elf5 expression intermediate between PrlrϪ/Ϫ and en- dogenous levels.
DISCUSSION Fig. 7. Expression of Elf5 in PrlrϪ/Ϫ Glands Rescued by Re- expression of Elf5 Immunohistochemistry was used to examine Elf5 levels. A From our transcript-profiling experiments we chose to andB(top row), Elf5 expression in PrlrϪ/Ϫ epitheliuminfected focus on Socs2 and Elf5. We report here that both with empty PolyPOZ. Weak nuclear staining is seen. C and D, Socs2 and Elf5 can recapitulate prolactin function in A region of partial rescue showing increased Elf5 expression vivo by geneticcomplementation using Prlr-deficient over PrlrϪ/Ϫin the nuclei. Note change in nuclear shape. E and mammary epithelium. F, A region of complete rescue showing a further increase in Signaling initiated by cytokine receptors via the Jak- nuclear Elf5 level and expression in the cytoplasm, rounding Stat pathway is attenuated via three mechanisms, the of the epithelial nuclei, multiple alveoli with lumen formation, protein inhibitor of activated STAT proteins, which pre- and the synthesisofmilk protein(purple color in the lumen). vent Stat dimerization or DNA interaction, the SH2- G and H, Elf5 levels in endogenous glands show identical histological features and levels and patterns of Elf5 staining containing protein tyrosine phosphatases, which de- as produced by PolyPOZ in panels E and F. I and J, Controls phosphorylate activating tyrosine phosphorylations, demonstrating nuclear Elf5 staining to be specific but Elf5 and the Socs proteins, which are transcribed in re- staining of the adipocyte margins and alveolar lumen to be sponse to cytokine signaling and which interact with nonspecific. IHC, Immunohistochemistry. the receptors or the receptor-associated Jak kinase to prevent Stat activation and to promote degradation via the proteasome (20). Knockout of Socs2 in mice re-
Downloaded from mend.endojournals.org on August 27, 2006 1184 Mol Endocrinol, May 2006, 20(5):1177–1187 Harris et al.•Elf5, Socs2, and Mammary Development
sulted in a loss of growth control. The long bones and that the retroviral-transplantation model does not al- nose to tail length is increased, and organs appear low secretory activation to be evaluated, because the normal, but are larger (although in proportion) due to mammary tree is not connected to the nipple and so increased cell numbers with increased collagen dep- the gland undergoes engorgement-induced involution osition, especially in the organ ducts and vessels (21). post partum. Therefore we do not know whether Elf5 Socs2 has not previously been implicated in the at- alone can rescue lactation in PrlrϪ/Ϫ mammary gland tenuation of prolactin action in the mammary gland in to a level sufficient for pup survival. Prolactin has a vivo, and it is now clear that both Socs1 (17) and pro-proliferative effect during alveolargenesis and an Socs2 perform this function via modulation of Stat5 additional differentiation action during secretory acti- phosphorylation. Although the studies of Flint and col- vation. Thus, although Elf5 can substitute for the pro- leagues (22) show that GH treatment could enhance liferative action of prolactin, our data allow no conclu- alveolar development in Prlrϩ/Ϫ mice, it further sup- sions regarding the differentiative and lactogenic pressed secretory activity and prevented lactation. actions of this hormone. Because Elf5 alone can res- Because we observed normal lactation, we can dis- cue alveolar development, it is very likely that Elf5 count a GH-based mechanism of Socs2 rescue of mediates mammary development in response to many Prlrϩ/Ϫ development. of the hormones and growth factors known to be Elf5 is an ETS transcription factor, a large and di- essential for this process. Given this, it is also likely verse family homologous to the v-ets oncogene en- that the regulation of Elf5 will involve the interaction of coded by the E26 avian erythroblastosisvirus, with a the signaling pathways activated by many of the hor- conserved DNA-binding domain of the winged helix- mones controlling mammary development and will not turn-helix superfamily. They are involved in cell prolif- be solely regulated by the Prl/Prlr/Jak2-Stat5 path- eration, differentiation (23), and carcinogenesis (24). way. Our findings demonstrate that expression of Elf5 Most ETS factors are expressed in MECs (25), where is sufficient to do the work of building the lobuloalveoli, PEA3 influences branching morphogenesis (26) and is and itwill intriguing to discover whether Elf5 expres- implicated in the initiation of ERB2-positive breast sion can rescue mammary development in knockout cancer inhumans (27, 28) and mice (29). ETS2 in the models of other pathway members, such as Stat5, and fat pad is necessary for development of tumors initi- in other nonpathway members, such as progesterone ated by ERRB2 in the mammary epithelium (30) and is receptor. Placing Rank ligand relative to Elf5 will also necessary for anchorage-independent growth of shed further light on the composition of this pathway. breast cancer cell lines (31). The Elf subfamily consists Only two genes have been reported to recapitulate of five members (Elf1 to -5). Elf5 acts as an activator of prolactin action by geneticcomplementation. Retrovi- transcription in the mouse (32) and human (33). Elf5 ral reexpression of Igf2 has been shown to allow partial directly activates a GGAA site in the whey acidic pro- development of Prlr knockout mammary epithelium tein promoter (34). It has been suggested that Elf5 has during pregnancy (37), and heterozygous loss of a negative regulatory domain that inhibits DNA binding Socs1 has been demonstrated to fully rescue lacta- (33). Elf5 is located on human chromosome 11p13–15 tional failure in Prlr heterozygous glands (17). We can (32), a region of the genome known to experience loss now add reexpression of Elf5 in Prlr knockout mam- of heterozygosity insome breast cancers. Elf5 mRNA mary epithelium and homozygous loss of Socs2 from expression is also lost inanumber of breast cancers Prlr heterozygous mammary epithelium,tothissmall compared with adjacent normal tissue (35). Elf5 isa list of genes that have been demonstrated by comple- key regulator of lobuloalveolar development, as dem- mentation assays to recapitulate prolactin action. onstrated by the formation of lobules capable of milk production after retroviral reexpression of Elf5 in Prlr knockout mammary epithelium. Not only were mor- phologically normal alveoli produced, histological ex- MATERIALS AND METHODS amination showed correct cellular architecture, the formation of lipid droplets within the cells of these Mice, Tissue Recombination, and Epithelial Transplantation alveoli, and milk production, demonstrating that the secretory initiation phase had been entered. Rescued All animal experimentation was conducted under the supervi- portions of Prlr knockout glands showed greatly in- sion and within the guidelines of the Garvan Institute/St Vin- creased expression of Elf5. The conclusion that Elf5 is cent’s Hospital Animal Experimentation Ethics Committee. a key regulator of mammary development is sup- Glands for transcript profiling were prepared by clearing both the fourth mammary fat pads from a Rag1Ϫ/Ϫ mouse of endog- ported by our recent finding that heterozygous null Ϫ Ϫ enous mammary epithelium and then transplanting Prlr / or mutation of the Elf5 gene caused lactational failure Prlrϩ/ϩ epithelium (C57BL6ϫ129SVPas) to either fourth mam- (36). Analysis of the effect of complete loss of Elf5 has mary fat pad of the sameanimal (38). The animals were then been prevented by early embryonic lethality of ho- aged for 12 wk. Animals were mated and checked for vaginal mozygous Elf knockout mice (36). plugs in the morning. At 2, 4, and 6 d after observation of a plug the mammary glands were collected, frozen inliquidnitrogen, It is surprising that a single transcription factor can and stored at Ϫ70 C. The Socs2Ϫ/Ϫ and Prlrϩ/Ϫ mice used for produce such a comprehensive rescue of develop- the interbreeding study were inbred C57Bl6 and generation 6 Ϫ Ϫ ment in Prlr / mammary epithelium. A caveat here is backcrossed to C57Bl6, respectively.
Downloaded from mend.endojournals.org on August 27, 2006 Harris et al.•Elf5, Socs2, and Mammary Development Mol Endocrinol, May 2006, 20(5):1177–1187 1185
SCp2 Cell Differentiation Assay tions onto Superfrost Plus slides (Menzel-Glaser, Singapore). Antigen retrieval used DAKO (DAKO, Carpinteria, CA) Target SCp2 MECs (39) were passaged in DMEM-F12 medium con- Retrieval Solution, pH 9.9, inaboiling water bath for 20 min taining DMEM-HAM, 10% fetal calf serum, and insulin5 followed by H2O2 treatment. Counterstaining was performed g/ml(Sigma Chemical Co.,St. Louis, MO). Briefly, 7 ϫ 105 with hematoxylin and 1% acid alcohol incubation with pri- cells were plated in 35-mm dishes precoated with extracel- mary antibody. Elf5 immunohistochemistry was performed lular matrix (ECM) (Matrigel; Collaborative Research) in using the Elf-5 (N-20) affinity-purified goat polyclonal anti- DMEM-F12 containing 2% fetal calf serum and insulin(5 body (sc-9645) from Santa Cruz Biotechnology, Inc. (Santa g/ml) (Sigma). The next day, cells were washed twice, and Cruz, CA) at a dilution of 1:300, and a secondary biotinylated then placed indifferentiation media containing DMEM-F12 horse antigoat (Vector Laboratories, Inc., Burlingame, CA), plus insulin(5g/ml), hydrocortisone (1 g/ml), and prolactin LSABϩ (DAKO) and was detected using Liquid DABϩ Sub- (3 g/ml)akind gift of Dr. A. Parlow, National Hormone and strate Chromogen (DAKO). The antimilk (1:12,000) primary Pituitary Program. Cells were induced to differentiate for 48 h antibody (Accurate Chemical&Scientific Corp., Westbury, (10 dishes). Untreated cells (10 dishes) were incubated with NY) was incubated for 30 min, and bound antibody was insulin(5g/ml) and hydrocortisone (1 g/ml) but not pro- detected using the Envision System (DAKO) and 3,3Ј-diami- lactin. Cells were harvested directly for RNA extraction using nobenzidine Plus (DAKO) as substrate. Rabbit IgG without Trizol (Life Technologies, Gaithersburg, MD). antibody provided the negative technical control. The anticy- tokeratin18mouse monoclonal (Research Diagnostics, Flanders, NJ), and the high-molecular weight keratin mouse Target Preparation, GeneChip Hybridization, and monoclonal antibody (DAKO) were used at 1:100 and were Scanning biotinylated with the DAKO ARK kit antibody and detected using LSABϩ/DAB as above. Total RNA was prepared from frozen mammary glands by homogenization with a Polytron inTrizol for 30 sec, chloroform In Situ Hybridization extraction, and isopropanol precipitation. RNA was further purified using QIAGEN RNeasy Mini Kit (QIAGEN, Chatsworth, CA). The cRNA targets were generated as recommended Full-length mouse SOCS2 cDNA was cloned into Bluescript by Affymetrix (Santa Clara, CA) and hybridized to MG-U74A SKII (Stratagene, La Jolla, CA). Antisense and sense ribo- GeneChips (Affymetrix). The GeneChips were scanned using probes were generated using T3 or T7 RNA polymerase the GeneArray Scanner, and the hybridization intensities and (Promega) with digoxigenin-UTP (Roche). Standard in situ fold change between experiments was obtained using Microar- hybridizations were performed as described elsewhere (22). ray Suite 4.0 (Affymetrix) and the MGU74A mask. Western Blotting GeneChip Analysis and Database Interrogation Mouse mammary gland protein lysates were prepared by Data were analyzed using Microarray Suite 4.0 (Mas4, Af- grinding the tissue intoafine powder inliquidnitrogen and fymetrix) Mas5, and Spotfire visualization software. Known subsequent solubilization in 1% TEB [150 mM NaCl; 5 mM genes were searched in PubMed for relevance to mammary EDTA, 50 mM Tris (pH 7.5); 0.1% Nonidet P-40] supple- gland development. Expressed sequence tags were associ- mented with complete protease inhibitor (Roche, Mannheim, ated with known genes or other expressed sequence tags by Germany), 10 mM sodium fluoride, and 1 mM sodium or- querying Unigene (http://www.ncbi.nlm.nih.gov/UniGene/) thovanodate. Protein lysates (50 g) were separated by SDS- and verified by ClustalW alignment in Macvector (Oxford PAGE. After transfer, filters were blocked and incubated with Molecular, Inc., Palo Alto, CA). Further sequence and struc- rabbit polyclonal antiserum raised against mouse milk-spe- tural information, ontologies, and human orthologs were ob- cific proteins (Accurate Chemical&Scientific Corp.)oranti- tained from Resourcerer (http://pga.tigr.org/tigr-scripts/mag- ␣-tubulin monoclonal antibody (Sigma). Antibodies specific ic/r1.pl) and Netaffx (https://www.affymetrix.com), and exons for a-phospho-STAT5 (Upstate Biotechnology, Inc., Lake were identified in the mouse genome by searching the En- Placid, NY) and a-STAT5a (Santa Cruz Biotechnology) were sembl database (http://www.ensembl.org/Mus_musculus/). also used. Antibody binding was visualized with peroxidase- conjugated antirabbitorantimouse (Amersham Pharmacia Biotech, Arlington Heights, IL) using the enhanced chemilu- Quantitative PCR minescence system (Amersham).
Total RNA was prepared using the Trizol method. RNA was Retroviral Infection of MECs reverse transcribed using AMV reverse transcriptase (Pro- mega Corp., Madison, WI). PCR primers were designed using Macvector so that the product spanned an intron. The PCRs Mouse Elf5 cDNA was isolated frommammary gland cDNA were performed inaLightCycler using the FastStart DNA by PCR and cloned into the retroviral vector polyPOZ [a gift master SYBR Green I enzyme mix (Roche Clinical Laborato- from Dr. T. Dale, (19)]. Elf5-IRES-LacZ-polyPOZ and LacZ- ries, Indianapolis, IN) in a 10-l reaction volume. Absolute polyPOZ ecotropic retroviruses were packaged in Phoenix- quantification was performed by comparing transcript levels Eco cells (a gift of Philip Achacoso and Garry Nolan, Stanford insamples to a standard curve constructed by performing University Medical Center, Stanford, CA) by transient trans- serial dilutions of PCR product purified using QIAquick Gel fection using FuGENE-6 Reagent (Roche). Viral supernatant Extraction Kit (QIAGEN) and analyzed using the Second De- was harvested by filtration through a 0.45-m filter. Primary rivative Maximummethod (Roche). All data were normalized mouse MECs were harvested frommammary glands of 11- to Ϫ/Ϫ to expression of the housekeeping gene -actin. 13-wk-old virgin Prlr mice. Briefly, the no. 4 mammary glands were dissected out under sterile conditions, finely chopped, and subjected to three to four rounds of collage- Immunohistochemistry nase (10 mg/ml) digestion in2.5% fetal bovine serum/ HEPES-buffered RPMI 1640. The purified epithelial cells were Mammary glands were dissected frommice, rinsed in cold plated in DMEM: Ham’s F12 (GIBCO) supplemented with PBS, and fixed overnight in 10% neutral buffered formalinor 10% fetal bovine serum,5g/ml insulin, 10 ng/mlepidermal for2hin 4% paraformaldehyde at 4 C. Tissues were pro- growth factor, 5 g/ml hydrocortisone, and 10 ng/ml cholera cessed and embedded in paraffin wax and cut in5-m sec- toxin (all additives from Sigma). Primary MECs were sub-
Downloaded from mend.endojournals.org on August 27, 2006 1186 Mol Endocrinol, May 2006, 20(5):1177–1187 Harris et al.•Elf5, Socs2, and Mammary Development
jected to four rounds of retroviral infection by addition of viral in breast development and tumorigenesis—as revealed supernatant plus 8 g/ml polybrene. 0.5–1 ϫ 106 MECs were by progesterone receptor “knockout” and “knockin” injected into the no. 4 mammary fat pad of a 3-wk-old mouse models. Steroids 68:779–787 RAG1Ϫ/Ϫ recipient female mouse prepared as described 8. Horseman N, Zhao W, Montecino-Rodriguez E, Tanaka above. M, Nakashima K, Engle SJ, Smith F, Markoff E, Dorsh- kind K 1997 Defective mammopoiesis, but normal hema- topo Whole-Mount and H&E Histology iesis, in mice with targeted disruption of the prolactin gene. EMBO J 16:6926–6935 9. Brisken C, Kaur S, Chavarria TE, Binart N, Sutherland RL, The mammary glands were fixed in 10% formalin solution, Weinberg RA, Kelly PA, Ormandy CJ 1999 Prolactin con- defatted in acetone stained in carmine alum (0.2% carmine, trols mammary gland development viadirect and indirect 0.5% aluminum sulfate) or hematoxylin, dehydrated in etha- mechanisms. Dev Biol 210:96–106 nol followed by SlideBrite, and then cleared in methyl salic- 10. Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, ylate. Hematoxylin and eosin staining used 4- m sections of Buteau H, Edery M, Brousse N, Babinet C, Binart N, Kelly mammary glands incubated in two changes of SlideBrite PA 1997 Null mutation of the prolactin receptor gene (SASKO, Stuttgart, Germany), hydrated, stained with hema- produces multiple reproductive defects in the mouse. toxylin, washed in water, dehydrated, stained with eosin, and Genes Dev 11:167–178 dehydrated in ethanol and SlideBrite before mounting with 11. Vomachka A, Pratt S, Lockefeer J, Horsemann N 2000 Britemount. Prolactin gene-disruption arrests mammary gland devel- opment and retards T-antigen-induced tumor growth. Acknowledgments Oncogene 19:1077–1084 12. Binart N, Helloco C, Ormandy CJ, Barra J, Clement- Data inFigs. 1 and 5–9 and Table 1 were supplied by the Lacroix P, Baran N, Kelly PA 2000 Rescue of preimplan- Garvan Institute with support from The National Health and tatory egg development and embryo implantation in pro- Medical Research Council of Australia, the Cancer Council lactin receptor-deficient mice after progesterone New South Wales, The Cooperative Research Center for administration. Endocrinology 141:2691–2697 Innovative Dairy Products, and the United States Department 13. Ormandy CJ, Naylor M, Harris J, Robertson F, Horseman of Defense Breast Cancer Research Program [DAMD17-01- ND, Lindeman GJ, Visvader J, Kelly PA 2003 Investiga- 1-0240 (to J.H.) and DAMD17-03-1-0686 (to S.R.O.)]. Data in tion of the transcriptional changes underlying functional Figs. 2–4 were supplied by the Walter and Eliza Hall Institute defects in the mammary glands of prolactin receptor with support from the Victorian Breast Cancer Research Con- knockout mice. Recent Prog Horm Res 58:297–323 sortium, the Cancer CouncilVictoria, and by The National 14. Camarillo IG, Thordarson G, Moffat JG, Van Horn KM, Health and Medical Research Council of Australia. Binart N, Kelly PA, Talamantes F 2001 Prolactin receptor expression in the epithelia and stroma of the rat mam- mary gland. J Endocrinol 171:85–95 15. Hovey RC, Trott JF, Ginsburg E, Goldhar A, Sasaki MM, Received November 24, 2005. Accepted February 1, 2006. Fountain SJ, Sundararajan K, Vonderhaar BK 2001 Tran- Address all correspondence and requests for reprints to: scriptional and spatiotemporal regulation of prolactin re- Christopher J. Ormandy, Garvan Institute of Medical Re- ceptor mRNA and cooperativity with progesterone re- search, St. Vincent’s Hospital, Darlinghurst, New South ceptor function during ductal branch growth in the Wales 2010, Australia. E-mail: [email protected]. mammary gland. Dev Dyn 222:192–205 W.A. received support from Amrad Corp. from 1997–2004 16. Zhao SH, Simmons DG, Cross JC, Scheetz TE, Casavant and isaninventor on patents in the socs area. There are no TL, Soares MB, Tuggle CK 2004 PLET1 (C11orf34), a other potential conflicts of interest involving the authors. highly expressed and processed novel gene inpig and mouse placenta, is transcribed but poorly spliced in hu- man. Genomics 84:114–125 17. Lindeman GJ, Wittlin S, Lada H, Naylor MJ, Santamaria REFERENCES M, Zhang JG, Starr R, Hilton DJ, Alexander WS, Or- mandy CJ, Visvader J 2001 SOCS1 deficiency results in accelerated mammary gland development and rescues 1. Hovey RC, Trott JF, Vonderhaar BK 2002 Establishing a lactation in prolactin receptor-deficient mice. Genes Dev framework for the functional mammary gland: from en- 15:1631–1636 docrinology to morphology. JMammary Gland Biol Neo- 18. Gallego MI, Binart N, Robinson GW, Okagaki R, Coschi- plasia7:17–38 gano KT, Perry J, Kopchick JJ, Oka T, Kelly PA, Hen- 2. Hennighausen L, Robinson GW 2001 Signaling pathways nighausen L 2001 Prolactin, growth hormone, and epi- in mammary gland development. Dev Cell 1:467–475 dermal growth factor activate Stat5 indifferent 3. Neville MC, McFadden TB, Forsyth I 2002 Hormonal compartments of mammary tissue and exert different regulation of mammary differentiation and milk secretion. and overlapping developmental effects. Dev Biol 229: JMammary Gland Biol Neoplasia7:49–66 163–175 4. Bocchinfuso WP, Korach KS 1997 Mammary gland de- 19. Naylor S, Smalley MJ, Robertson D, Gusterson BA, Ed- velopment and tumorigenesis in estrogen receptor wards PA, Dale TC 2000 Retroviral expression of Wnt-1 knockout mice. JMammary Gland Biol Neoplasia and Wnt-7b produces different effects in mouse mam- 2:323–334 mary epithelium. J Cell Sci 113:2129–2138 5. Bocchinfuso WP, Lindzey JK, Hewitt SC, Clark JA, Myers 20. Larsen L, Ropke C 2002 Suppressors of cytokine PH, Cooper R, Korach KS 2000 Induction of mammary signalling: SOCS. APMIS 110:833–844 gland development in estrogen receptor-␣ knockout 21. Metcalf D, Greenhalgh CJ, Viney E, Willson TA, Starr R, mice. Endocrinology 141:2982–2994 Nicola NA, Hilton DJ, Alexander WS 1069 Gigantismin 6. Mueller S, Clark J, Myers P, Korach K 2002 Mammary mice lacking suppressor of cytokine signalling-2. Nature gland development in adult mice requires epithelial and 405:1069–1073 stromal estrogen receptors. Endocrinology 143: 22. Allan GJ, Tonner E, Barber MC, Travers MT, Shand JH, 2357–2365 Vernon RG, Kelly PA, Binart N, Flint DJ 2002 Growth 7. IsmailP,Amato P, Soyal SM, DeMayo FJ, Conneely OM, hormone, acting in part through the insulin-like growth O’Malley BW, Lydon JP 2003 Progesterone involvement factor axis, rescues developmental, but not metabolic,
Downloaded from mend.endojournals.org on August 27, 2006 Harris et al.•Elf5, Socs2, and Mammary Development Mol Endocrinol, May 2006, 20(5):1177–1187 1187
activity in the mammary gland of mice expressing a stimulated invasion by BT20 breast carcinoma cells. single allele of the prolactin receptor. Endocrinology 143: Cancer Res 58:1027–1033 4310–4319 32. Zhou J, Ng AY, Tymms MJ, Jermiin LS, Seth AK, Thomas 23. Sharrocks AD 2001 The ETS-domain transcription factor RS, Kola I 1998 A novel transcription factor, ELF5, be- family. Nat Rev Mol Cell Biol 2:827–837 longs to the ELF subfamily of ETS genes and maps to 24. Gilliland DG 2001 The diverse role of the ETS family of human chromosome 11p13–15, a region subject to LOH transcription factors in cancer. Clin Cancer Res and rearrangement inhuman carcinoma cell lines. On- 7:451–453 cogene 17:2719–2732 25. Galang CK, Muller WJ, Foos G, Oshima RG, Hauser CA 33. Oettgen P, Kas K, Dube A, Gu X, Grall F, Thamrongsak U, 2003 Changes in the expression of many Ets family tran- Akbarali Y, Finger E, Boltax J, Endress G, Munger K, scription factors and of potential target genes in normal Kunsch C, Libermann TA 1999 Characterization of mammary tissue and tumors. JBiol Chem 297: ESE-2, a novel ESE-1-related Ets transcription factor that 11281–11292 is restricted to glandular epithelium and differentiated 26. Chotteau-Lelievre A, Montesano R, Soriano J, SoulieP, keratinocytes. JBiol Chem 274:29439–29452 Desbiens X, de Launoit Y 2003 PEA3 transcription fac- 34. Thomas RS, Ng AN, Zhou J, Tymms MJ, Doppler W, Kola tors are expressed intissues undergoing branching mor- I 2000 The Elf group of Ets-related transcription factors. phogenesis and promote formation of duct-like struc- ELF3 and ELF5. Adv Exp Med Biol 480:123–128 tures by mammary epithelial cells invitro. Dev Biol 259: 35. Ma XJ, Salunga R, Tuggle JT, Gaudet J, Enright E, Mc- 241–257 Quary P, Payette T, Pistone M, Stecker K, Zhang BM, 27. Benz CC, O’Hagan RC, Richter B, Scott GK, Chang CH, Zhou YX, Varnholt H, Smith B, Gadd M, Chatfield E, Xiong X, Chew K, Ljung BM, Edgerton S, Thor A, Hassell Kessler J, Baer TM, Erlander MG, Sgroi DC 2003 Gene JA 1997 HER2/Neu and the Ets transcription activator expression profiles of human breast cancer progression. PEA3 are coordinately upregulated inhuman breast can- Proc Natl Acad Sci USA 100:5974–5979 cer. Oncogene 15:1513–1525 36. Zhou J, Chehab R, Tkalcevic J, Naylor MJ, HarrisJ, 28. Bieche I, Tozlu S, Girault I, Onody P, Driouch K, Vidaud Wilson TJ, Tsao S, Tellis I, Zavarsek S, Xu D, Lapinskas M, Lidereau R 2003 Expression of PEA3/E1AF/ETV4, an EJ, Visvader J, Lindeman GJ, Thomas R, Ormandy CJ, Ets-related transcription factor, in breast tumors: positive Hertzog PJ, Kola I, Pritchard MA 2005 Elf5 is essential for links to MMP2, NRG1 and CGB expressions. Carcino- early embryogenesis and mammary gland development genesis25:405–411 during pregnancy and lactation. EMBO J 24: 635–644 29. Shepherd TG, Kockeritz L, Szrajber MR, Muller WJ, Has- 37. Brisken C, Ayyannan A, Nguyen C, Heineman A, Rein- sell JA 2001 The pea3 subfamily ets genes are required hardt F, Tan J, Dey SK, Dotto GP, Weinberg RA 2002 for HER2/Neu-mediated mammary oncogenesis. Curr IGF-2 isamediator of prolactin-induced morphogenesis Biol 11:1739–1748 in the breast. Dev Cell 3:877–887 30. Man AK, Young LJ, Tynan JA, Lesperance J, Egeblad M, 38. Naylor MJ, Ormandy CJ 2002 Mouse strain-specific pat- Werb Z, Hauser CA, Muller WJ, Cardiff RD, OshimaRG terns of mammary epithelial ductal side branching are 2003 Ets2-dependent stromal regulation of mouse mam- elicited by stromal factors. Dev Dyn 225:100–105 mary tumors. Mol Cell Biol 23:8614–8625 39. Desprez PY, Roskelley C, Campisi J, Bissell MJ 1993 31. Sapi E, Flick MB, Rodov S, Kacinski BM 1998 Ets-2 Isolation of functional cell lines from a mouse mammary transdominant mutant abolishes anchorage-indepen- epithelial cell strain: the importance of basement mem- dent growth and macrophage colony-stimulating factor- brane and cell-cell interaction. Mol Cell Differ 1:99–110
Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.
Downloaded from mend.endojournals.org on August 27, 2006 Available online http://breast-cancer-research.com/content/8/2/207
Review Key stages in mammary gland development The alveolar switch: coordinating the proliferative cues and cell fate decisions that drive the formation of lobuloalveoli from ductal epithelium Samantha R Oakes, Heidi N Hilton and Christopher J Ormandy
Garvan Institute of Medical Research, St Vincent’s Hospital, Darlinghurst, NSW 2010, Australia
Corresponding author: Christopher J Ormandy, [email protected]
Published: 25 April 2006 Breast Cancer Research 2006, 8:207 (doi:10.1186/bcr1411) This article is online at http://breast-cancer-research.com/content/8/2/207 © 2006 BioMed Central Ltd
Abstract the epithelial cells occurs within the ductal branches and Massive tissue remodelling occurs within the mammary gland developing alveoli. This increases both epithelial cell number during pregnancy, resulting in the formation of lobuloalveoli that are and epithelial surface area, actions essential for sufficient milk capable of milk secretion. Endocrine signals generated production during lactation. Cell differentiation becomes predominantly by prolactin and progesterone operate the alveolar dominant from mid-pregnancy as the gland moves into the switch to initiate these developmental events. Here we review the secretory initiation phase [3]. The developing alveoli cleave current understanding of the components of the alveolar switch and the alveolar cells become polarised and form a sphere- and conclude with an examination of the role of the ets transcription factor Elf5. We propose that Elf5 is a key regulator of like single layer of epithelial cells that envelopes a circular the alveolar switch. lumen, connected to the ductal network via a single small duct. Each individual alveolus is surrounded by a basket-like architecture of contractile myo-epithelial cells. The myo- Introduction: the alveolar switch epithelium of the alveoli is discontinuous so that the luminal Massive tissue remodelling within the mammary gland during cells directly contact the underlying basement membrane, pregnancy results in the formation of the secretory which forms part of the extracellular matrix. Some cells of the lobuloalveolar units in preparation for lactation. The initial ductal network also contact the basement membrane. proliferative phase of alveolar morphogenesis is instigated by Contact is required for complete lobuloalveolar differentiation an increase in the level of serum prolactin (Prl) and [4,5], seen morphologically by the appearance of lipid progesterone (Pg) [1]. These hormones activate the alveolar droplets [6] and by the initiation of gene expression in a switch, a genetic program that coordinates changes in defined order [7]. Nearing parturition, alveolar tight junctions mammary epithelial cell proliferation, migration, differentiation close and milk and colostrum proteins move into the alveolar and deletion within the many tissue types of the mammary lumen, in preparation for active milk secretion post-partum, gland. Here we review our current understanding of the which marks the onset of the secretory activation phase [8] genetic program controlling alveolar morphogenesis, using (Figure 1). the mouse as a model of the human breast [2]. We then examine the role played by the ets transcription factor Elf5 in The epithelial expansion is paralleled by equally dramatic coordinating this program in epithelial cells, and propose that changes in other tissue compartments. Adipocytes lose their Elf5 is a central component of the alveolar switch. lipid content and remain as long projections scattered throughout the alveolar epithelium [9]. A huge expansion of Tissue remodelling during pregnancy the vasculature also occurs within the stroma, to provide the The most striking aspect of mammary development during large quantities of energy, sugars, amino acids and solutes pregnancy is massive tissue remodelling. During the alveolar required for milk production [10]. Developmental events are morphogenesis phase [3], rapid and global proliferation of also elicited elsewhere in the animal; for example, the gut and
Ccnd1 = cyclin D1; Gal = galanin; MEC = mammary epithelial cell; NF = nuclear factor; Pg = progesterone; Pgr = progesterone receptor; Prl = prolactin; Prlr = prolactin receptor; RankL/Opgl = receptor activator of NF-κB ligand/osteoprotegrin ligand; Socs = suppressor of cytokine sig- nalling; Tgf = transforming growth factor.
Page 1 of 10 (page number not for citation purposes) Breast Cancer Research Vol 8 No 2 Oakes et al.
Figure 1
Alveolar morphogenesis. Mammary wholemounts (Carmine alum stain top row) and mammary cellular architecture (low power, middle row; high power, bottom row) in virgin, 12 days posts coitus (dpc), 18 dpc and 1 day post partum (1 dpp) murine mammary glands. Ductal epithelial cells (arrow) and myoepithelial cells (arrowhead) arise from a common mammary epithelial stem cell. Massive epithelial cell proliferation occurs at the onset of pregnancy, which is co-ordinated predominantly by prolactin and progesterone. At mid-pregnancy (12 dpc), developing alveoli continue to proliferate and polarise to form a sphere-like single layer of epithelial cells enveloping a circular lumen (indicated by X). This is followed by further cell proliferation and differentiation categorised by the expression of milk genes and the formation of cytoplasmic lipid droplets (indicated by asterisks). At 18 dpc, alveoli have large amounts of lipid and milk protein expression is increased. At parturition, tight junctions between alveolar cells close and milk proteins and lipid are secreted into the alveolar lumen (X). An expansion of the vasculature (open arrows) and reduction in the adipocyte (A) area is also apparent in the stroma. liver enlarge dramatically to cope with the energy needs of hypothesised, based on a paradigm developed in the gestation and lactation. The brain is programmed for correct hematopoietic system, that a primary mammary epithelial stem maternal behaviour by Prl [11]. Thus, the alveolar switch is cell gives rise to a hierarchy of epithelial progenitor cell part of a larger mechanism controlling all aspects of lineages to ultimately produce the different cells found in the adaptation to pregnancy and lactation. mammary epithelium [16,17]. The flux of cells through these lineages is likely to be controlled by, and in turn control, the Another striking aspect of tissue remodelling during patterns of gene expression that comprise the alveolar switch. pregnancy is its cyclical nature. Following weaning nearly all Integrating our knowledge of gene expression patterns with of the development induced by the alveolar switch is removed the emerging knowledge regarding stem cell lineages and by programmed cell death during the involution phase, only to their interactions offers us an unprecedented opportunity to redevelop with the next pregnancy. This observation first led understand this phase of mammary development. researchers to hypothesise that mammary tissue must contain persistent self-renewing mammary stem cells Prolactin and progesterone initiation of (reviewed in [12]). The ability of small epithelial transplants to alveolar morphogenesis recapitulate a complete and fully functional epithelial The formation of the milk secreting structures during mammary gland reinforced this view [13]. The presence of a pregnancy is dependent on a synergy between Prl and Pg single mammary stem cell was indicated by limiting dilution signalling [6]. These hormones trigger an initial wave of cell experiments and the existence of committed progenitor cells proliferation during days two to six of pregnancy [18]. The was demonstrated by transplants that showed limited progesterone receptor (Pgr) knockout mouse demonstrated developmental capacity [14]. This cell was recently isolated that Pg is required for alveolar morphogenesis, and epithelial and elegantly demonstrated to be capable of producing a recombination experiments demonstrated that Pgr in the renewable and complete mammary epithelium [15]. Thus it is mammary epithelium, not the stroma, was essential for
Page 2 of 10 (page number not for citation purposes) Available online http://breast-cancer-research.com/content/8/2/207
epithelial cell proliferation [19]. Not all mammary epithelial Pituitary Prl stimulation of ovarian Pg assists in maintaining cells express Pgr and, therefore, are unable to respond to Pg the required levels of Pg during early pregnancy [28]. In directly. Mammary gland chimeras made from Pgr+/+ and addition, up-regulation of Pgr expression by Prl, and Prl Pgr–/– mammary epithelial cells (MECs) demonstrated that receptor (Prlr) expression by Pg, suggests that these Pgr–/– epithelial cells proliferate in response to Pg and, hormones may interact in a synergistic manner to control therefore, must respond to a paracrine factor from Pgr+/+ alveolar development. Prolactin receptor knockout mice cells [1]. Indeed, in the epithelium, proliferating cells (Prlr–/–) have demonstrated the importance of this receptor segregate with Pgr positive cells [20]. This is also true for during mammary development [29]. Like Pgr, experiments estrogen receptor positive cells [21]. Further, steroid with Prlr–/– mice have shown that the presence of Prlr in the receptor positive cells are in close proximity to proliferating epithelial cells, not the stroma, is essential for normal cells, indicating that proliferation is mediated, at least in part, lobuloalveolar differentiation [30]. Prlr–/– mammary by a paracrine mechanism. This heterogeneous receptor transplants fail to develop lobuloalveoli and produce milk patterning observed in the luminal epithelium is required for proteins during pregnancy, illustrating that Prlr is essential in complete lobuloalveolar development [22]. the mammary epithelium during alveolar morphogenesis. The downstream targets of prolactin signalling will be discussed Wingless-related MMTV integration site 4 (Wnt4) and in more detail later in this review. receptor activator of nuclear factor (NF)-κB ligand (RankL) are targets of the Pgr signalling pathway and may be the The neuronal peptide galanin (Gal) regulates Prl secretion paracrine factors responsible for cellular proliferation in from the pituitary lactotrophs [31]. In addition, the mammary steroid receptor negative cells. Over-expression of the proto- epithelium is responsive to Gal, as it augments alveolar oncogene Wnt1 can rescue pregnancy-induced ductal side morphogenesis in mammary explants in the presence of Prl branching in Pgr knockout mice, indicating that a Wnt factor [32]. Gal–/– mice show increased levels of the inhibitory may be an important paracrine mediator of Pg-induced ductal phosphorylated form of Prl [33], and are unable to nurse side branching during early pregnancy [23]. Mammary pups due to failed secretory activation [34]. Therefore, Gal transplants of Wnt4–/– epithelium have demonstrated that has dual actions: firstly, an indirect role by modulating Wnt4 acts in a paracrine fashion to stimulate epithelial ductal pituitary Prl and phosphorylated Prl release; and secondly, a side branching during early pregnancy. In these experiments, direct cell autonomous role in the formation of lobuloalveoli normal lobuloalveolar proliferation was observed during the during pregnancy. later half of pregnancy, indicating that other factors mediating proliferation in late pregnancy may be involved [23]. Other hormones can influence alveolar morphogenesis. Growth hormone may act in combination with Prl to mediate The RankL target, NF-κB, is required for cyclin D1 (Ccnd1) alveolar proliferation. Growth hormone treatment restores activation via the kinase IκB (IKKα) in neighbouring alveolar morphogenesis but inhibits lactation in Prlr+/– proliferating cells. Germ line deletion of both RankL and its mammary glands [35]. Placental lactogen is released from receptor (Rank) in mice resulted in failed alveolar the placenta during pregnancy and can fully compensate for morphogenesis due to reduced proliferation and increased Prl, allowing alveolar morphogenesis in Prl–/– mice [36]. apoptosis of alveolar epithelial cells [24]. These effects were mediated by protein kinase B (PKB/Akt), demonstrating that Molecular modulators of Prl induced alveolar this pathway is essential for the formation of lobuloalveolar morphogenesis structures [24]. The RankL/NF-κB/Ccnd1 pathway is now Members of the Prl-signalling pathway are essential for known to be crucial for the formation of alveolar structures normal alveolar morphogenesis [37]. Prlr dimerization occurs during pregnancy [25], and NF-κB is essential for Pg driven after Prl binding and leads to the phosphorylation of the proliferation within alveoli [20]. RankL also co-localises with associated Janus kinase (Jak2) [38,39], which in turn Pgrs in response to pregnancy levels of estrogen and Pg, phosphorylates specific residues on the Prlr [40]. Stat5 is indicating this is an important part of the response. In primary then recruited to the receptor and is phosphorylated by Jak2 MEC cultures, Pg acts in synergy with estrogen to increase [41]. Phosphorylated Stat5 is then translocated to the Ccnd1 transcription, resulting in increased proliferation [26]. nucleus where it can activate transcription of multiple genes Together, these data indicate that Pg may drive the [42] involved a variety of processes during alveolar proliferation of neighbouring cells via RankL/NF-κB, resulting morphogenesis, including establishment of epithelial polarity in Ccnd1 transcription (Figure 2a,b). Pgr consists of two and cell-cell interactions, stromal epithelial interactions and isoforms, PgrA and PgrB, which are expressed from a single milk protein expression during lactation (Figure 2c). Both gene. The PgrB isoform is essential and sufficient for alveolar isoforms of Stat5, Stat5a and Stat5b, when knocked out in morphogenesis during pregnancy. Alveoli in PgrB knockout mice, result in lobuloalveolar defects [43-45]. The phenotype mice fail to develop due to impaired proliferation of the ductal is more severe in combined Stat5a/Stat5b knockout animals. and alveolar compartment, which is possibly mediated via One class of genes activated by the prolactin-signalling activation of RankL [27]. pathway are the suppressor of cytokine signalling (Socs)
Page 3 of 10 (page number not for citation purposes) Breast Cancer Research Vol 8 No 2 Oakes et al.
Figure 2
Molecular control of alveolar morphogenesis. Signalling from the progesterone receptor (Pgr) and prolactin receptor (Prlr) is essential for alveolar morphogenesis in pregnancy. Increases in serum progesterone (Pg) and prolactin (Prl) result in luminal cell proliferation during early pregnancy, which continues throughout gestation. (a,b) Heterogenous receptor patterning is essential for complete alveolar morphogenesis. (a) Transforming growth factor (Tgf)-β1 signalling via phosphorylation of Smad results in the transcription of target genes, which act to control proliferation in steroid receptor positive cells. Wnt4 and RankL are transcribed in response to Pgr signalling, probably in cooperation with Prl signalling, and appear to stimulate proliferation of neighbouring cells via paracrine mechanisms. (b) RankL binds to its receptor Rank in a neighbouring cell and activates the RankL/nuclear factor (NF)-κB pathway, resulting in cyclin-D1 (Ccnd1) transcription and proliferation. Wnt4 binds and activates its target β-catenin, which has specific roles for both luminal and myo-epithelium for cell fate decisions involving both proliferation and differentiation. (a,c) Prl binds to Prlr and activates the Jak2/Stat5 cascade, resulting in the transcription of genes, including various transcription factors (TF) involved in epithelial morphogenesis and branching (Wnt4), establishment of epithelial polarity and cell-cell interactions (claudins and connexins), stromal epithelial interactions (collagen and laminin), proteins that regulate their own pathway (Socs1/2) and lactation (serotonin and milk proteins). Prl signalling also results in the transcription of cyclin D1 via an insulin growth factor 2 dependent mechanism. The ets transcription factor Elf5, transcribed in response to Prl, can completely compensate for the loss of Prlr signalling. Laminin in the extracellular matrix binds to β1-integrin when contact between the basement membrane and the luminal epithelium is established, and is essential for the maintenance of alveolar cell polarity and differentiation. ErbB4 and its ligands complement Prlr signalling as activation of ErbB4 results in Stat5 phosphorylation and translocation to the nucleus. GJ, gap junction; L, lipid droplet; TJ, tight junction.
members, which act to shut down the Prl-signalling pathway. also identified and is involved in the exchange of small ions Socs1 knockout mice show precocious development during and metabolites [49]. Recently, Connexin-26 was shown to pregnancy, and Socs1+/– mice can restore the lobuloalveolar be important in full lobuloalveolar development and in the defects present in Prlr+/– mice due to Prlr haplo-insufficiency prevention of alveolar cellular apoptosis [50]. [46]. Similarly, loss of Socs2 can also rescue lactation in Prlr+/– females [47]. Wnt4 was also down-regulated in Prlr–/– transplants, indicating that it is potentially a target of Prlr signalling [46]. Transcript profiling of Prlr knockout mammary glands The downstream target of Wnt, β-catenin, has specific actions identified a panel of genes that require Prlr-mediated in both the luminal and myoepithelial compartments of the signalling for increased expression during early pregnancy epithelium, and as a component of cell-cell junctions appears [46,48] (Figure 2c). Two members of the collagen family and to have a role in signalling to luminal epithelial cells [51,52]. laminin were identified. These molecules are cell adhesion Indeed, activation of β-catenin within the basal epithelial cells components of the extracellular matrix and play an important results in premature differentiation of the luminal epithelium part in the epithelial-stromal signalling required for full during pregnancy and persistent proliferation resulting in lobuloalveolar differentiation and gene expression [4,7]. tumors. These tumors consisted predominantly of Alveolar morphogenesis induced by Prl involves the undifferentiated basal cells, which were amplified in response establishment of polarity and cell-cell communication. The to β-catenin activation, thereby implicating this molecule in cell maintenance of cellular polarity is regulated by the closure of fate decisions in the mammary gland [52]. tight junctions, and the expression of tight junction proteins Claudin-3 and Claudin-7 was reduced in Prlr–/– mammary The gene encoding RankL was also identified as potentially transplants [46]. The gap junction protein Connexin 26 was regulated by Prl [46,53]. Ccnd1 null mutants exhibit
Page 4 of 10 (page number not for citation purposes) Available online http://breast-cancer-research.com/content/8/2/207
significantly delayed alveolar cell proliferation and impaired The NF1 family of transcription factors also play a role in lactation, which was shown to be epithelial cell autonomous functional differentiation as they regulate the transcription of [54]. Interestingly, Prl can induce Ccnd1 expression via milk protein genes such as those encoding whey acidic induction of insulin growth factor 2, independent of RankL protein, α-lactalbumin and β-lactoglobulin [61]. The NF1-C2 induction [55]. The similarities between Prl- and Pg-mediated isoform member of this family induces the expression of the effects on both RankL and Wnt signalling is further evidence milk genes encoding carboxyl ester lipase and whey acidic of the co-operation of these pathways for alveolar cell protein. Prl regulates the protein expression of NF1-C2 in proliferation during early pregnancy (Figure 2a). NmuMG cells, and its expression is reduced in the nucleus of Prlr–/– luminal cells at mid-pregnancy, indicating that NF1-C2 Gene expression profiling of Prl–/– mice has also identified may be regulated by Prl signalling during pregnancy and unique targets of mammary development. Expression of involved in expression of milk genes in preparation for tryptophan hydroxylase, the rate-limiting enzyme in serotonin lactation [62]. biosynthesis, is increased by Prl during pregnancy and lactation. Accumulation of serotonin due to milk engorgement The helix-loop-helix transcription factors Id1 and Id2 have experienced during weaning or experimentally via teat sealing varying expression in the mammary gland. Id1 expression is inhibits milk gene expression and can induce involution, increased during early pregnancy, remains low during providing a mechanism that is put in place by Prl to stop lactation and rises again at involution. Unlike Id1, Id2 remains lactation at weaning [56]. high during lactation, indicating that these isoforms have specific functional roles during alveolar morphogenesis [63]. Transcription factors involved in alveolar Id1 is specifically expressed by the expanding epithelium morphogenesis during the alveolar proliferative phase and is inversely Prl and Pg and other factors induce the transcription of correlated with the expression of β-casein; it appears, genes via activation of target transcription factors. These therefore, to be an important factor during early alveolar include Stat5 and the steroid hormone receptors as proliferation. Id1 also regulates Clusterin, which is involved in discussed previously, which bind to DNA and result in the the regulation of cell-cell interactions. Additionally, transcription of genes involved in many aspects of alveolar lobuloalveolar development is severely impaired in Id2 morphogenesis. Further, some of these target genes are knockout mice. Reduced proliferation and increased transcription factors also, which act to induce the expression apoptosis has been observed in mammary epithelium lacking of genes or groups of genes involved in lobuloalveolar Id2, resulting in the failure to form alveolar structures and development. An example is the transcription factor Srebf1, consequently failure of lactation [64]. Id2 also promotes which was identified from transcript profiling experiments on differentiation in MEC cultures, indicating Id2 is essential for three mouse models of failed secretory activation [33]. the differentiation of the mammary epithelium [63]. Srebf1 controls the expression of a number of key lipid metabolism genes [57] that showed reduced expression The transcription factor NF-κB discussed earlier in this review concomitantly with decreased Srebf1 expression [33]. Some is essential for Pg induced alveolar cell proliferation resulting transcription factors that appear to be involved in alveolar in Ccnd1 transcription [20,25]. NF-κB can also induce the morphogenesis include the homeobox genes, helix-loop-helix transcription of many genes involved in the regulation of genes, Stats, Tcf/Lef family, NF-κB, the Ceb/p family, the apoptosis. NF-κB levels are induced during pregnancy, nuclear factor family and the Ets transcription factors. The decline during lactation and are re-induced during lactation regulation of cellular proliferation during mammary develop- implying a role in mammary gland remodelling. It is also ment by the homeobox genes, helix-loop-helix genes, stats hypothesised that NF-κB is an essential ‘checkpoint’ of and ets transcription factors has been reviewed previously [58]. apoptosis, whose actions are dependent on association with specific transcriptional regulators. Thus NF-κB is an Pg and Prl are hypothesised to influence the expression of β- important transcription factor controlling both proliferation catenin via induction of the Wnt pathway, as discussed and apoptosis in the epithelium during pregnancy [65]. earlier. β-Catenin regulates the activity of the Tcf/Lef family of transcription factors, which appear to mediate β-catenin The C/ebp family of proteins appear to be important signalling and, therefore, may play a role during alveolar regulators of alveolar morphogenesis (for a review, see [66]). morphogenesis [59]. Inhibition of β-catenin results in alveolar C/ebpβ and C/ebpδ isoforms are increased during apoptosis and greatly reduced milk production capacity. Mice pregnancy and decline during lactation, indicating that they lacking Lef-1 demonstrate a failure to form the alveolar bud at play a critical role in alveolar morphogenesis and early milk embryonic day 13. The expression of Lef-1 was co-expressed gene expression. Transplantation experiments have revealed with β-catenin, and shows a similar expression pattern in that C/ebpβ is required in epithelial cells for normal response to parathyroid hormone-related protein [60]. Thus, lobuloalveolar development during pregnancy, and C/ebpβ Lef-1 may act to mediate the actions of β-catenin, although its knockout mice display phenotypes similar to Pgr, Prlr, effects during alveolar morphogenesis are still unclear. Stat5a/b, Ccnd1, Id2 and RankL knockouts [66]. Interestingly
Page 5 of 10 (page number not for citation purposes) Breast Cancer Research Vol 8 No 2 Oakes et al.
Pgr expression was dramatically increased in the mammary reduced milk gene expression. Alveolar proliferation was glands of C/ebpβ null mice and, in addition, the expression of attenuated and Stat5 phosphorylation was abolished. The Pgr was unusually uniform within the epithelium [67]. These ErbB4 ligand neuregulin/heregulin-1 (Nrg) promotes lobulo- effects were associated with a 10-fold decrease in the rate of alveolar development and the expression of milk genes when proliferation. There was, however, no change in the used in mammary gland explants [73], indicating a role for expression of C/ebpβ in the mammary glands of Pgr this ligand in lobuloalveolar development. In addition, mice knockout mice, indicating that C/ebpβ is upstream of Pgr and that lack the alpha form of Nrg show a similar phenotype to possibly controls the spatial distribution of epithelial cells, ErbB4 knockout, with reduced alveolar proliferation and which influence proliferation in alveolar progenitors [67]. differentiation, demonstrated by reduced β-casein expression C/ebpβ null epithelium significantly increased Tgf-β and in reduced alveoli expansion [74]. Smad2 signalling, and this pathway is known to inhibit cellular proliferation [68]. Cell cycle progression in C/ebpβ null Other ErbB ligands also appear to have overlapping functions MECs was blocked at the G1/S transition, preventing these for mammary gland development. Amphiregulin null animals cells from proliferating in response to early pregnancy levels have reduced alveolar development, although the phenotype of Pg and estrogen [69]. Therefore, C/ebpβ is essential for was much more severe in a triple mutant including knockouts controlling cell fate decisions within the mammary gland, of Tgfα and epidermal growth factor (all ligands of the ErbB including attenuating Pgr expression resulting in mammary family), indicating overlapping and compensatory roles for epithelial cell differentiation during pregnancy. these ligands during alveolar morphogenesis [75]. Triple mutants developed poorly organised and differentiated alveoli, The expression of the Ets transcription factor subfamily Pea3 had reduced milk protein expression and often pups born to is elevated at the onset of pregnancy but declines during mid- these mice did not survive. Amphiregulin loss was also pregnancy to low levels at lactation and involution, suggesting associated with reduced Stat5 phosphorylation. Our transcript a role in early pregnancy induced ductal outgrowth. Three profiling experiments demonstrated that amphiregulin was members of the Pea3 subfamily are expressed by both the down-regulated in Prlr–/– epithelium [46], indicating that myoepithelium and the luminal cells, although their expression amphiregulin may be modulated by Prlr signalling. These data varies during pregnancy, suggesting multiple signalling roles together indicate important roles for the ErbB receptors and during alveolar morphogenesis. The expression of all members ligands during alveolar morphogenesis. The overlapping of the family remains in the myoepithelium during pregnancy, phenotypes observed in Prlr, Pgr and ErbB knockout mice although the expression of the ER81 member declines in the suggest there may be some cross-talk between these luminal epithelium seven days after impregnation. Increased receptors, which is yet to be fully understood. numbers of dividing cells were observed in the terminal end buds of Pea3 knockout mice, and mammary gland transplants The cell surface receptor β1 integrin, which is present on of Pea3 knockout epithelium displayed reduced mammary luminal epithelial cells, is an essential mediator of extracellular branching during pregnancy, suggesting a role for Pea3 in matrix signalling via its ligands collagen and laminin [76]. progenitor cell differentiation [70]. Mammary epithelium in mice lacking β1 integrin in the luminal cells, displayed reduced proliferation and alveolar Other factors involved in alveolar disorganisation [77]. The focal adhesion kinase, which is morphogenesis important in protein complexes that connect the extracellular The receptor tyrosine kinase ErbB (epidermal growth factor) matrix to the actin cytoskeleton, was also reduced in these family and their ligands are important mediators of all aspects mice. Conditional deletion of β1 integrin during early of mammary development. There are four receptors: epidermal pregnancy and late pregnancy demonstrates that this growth factor receptor/ErbB/Her1, ErbB2/Her2/neu, ErbB3/ molecule was important for both the formation of Her3 and ErbB4/Her4, which are activated by a variety of lobuloalveolar structures and for functional differentiation ligands inducing activation via dimerisation and cross [78]. In these mammary glands, luminal epithelium becomes phosphorylation. ErbB ligands share a 50 amino acid domain, dissociated from the basement membrane, and cellular which is homologous to epidermal growth factor. Mice polarity is compromised as luminal epithelial cells protrude expressing a truncated dominant negative allele of ErbB2 did into the alveolar luminal space. In addition, Prl-stimulated milk not exhibit a phenotype until late pregnancy, when alveoli protein expression via phosphorylation of Stat5 was largely failed to expand and distend, indicating that ErbB2 is critical absent in primary mammary epithelial cells lacking β1-integrin, for secretory activation, and will be discussed later in this indicating that it is essential for Prl-induced activation of review series [71]. Conditional deletion of ErbB4 within the Stat5 [79]. mammary gland at pregnancy demonstrated a critical role for this receptor during alveolar morphogenesis [72]. Alveolar The cytokine Tgf-β1 is an important regulator of mammary cell expansion was reduced from 13.5 days post coitus in proliferation during pregnancy [68]. Tgf-β1 is restricted to the mammary epithelium lacking ErbB4, resulting in incomplete luminal epithelial cells and can control cell proliferation via alveolar development and failure to nurse pups due to phosphorylation of Smad following Tgf-β receptor activation
Page 6 of 10 (page number not for citation purposes) Available online http://breast-cancer-research.com/content/8/2/207
[80]. Tgf-β1 heterozygote mice display accelerated lobulo- This article is part of a review series on alveolar development due to increased proliferation, indicating Key stages in mammary gland development, that the expression of Tgf-β1 restricts alveolar cell edited by Charles Streuli. proliferation. Epithelial cell proliferation was increased more than 15-fold in Tgf-β1 null ovariectomised animals treated Other articles in the series can be found online at with estrogen and Pg compared to wild-type mice [81]. In http://breast-cancer-research.com/articles/ animals treated with estrogen and Pg, Tgf-β1 expression was review-series.asp?series=bcr_keystages restricted to the steroid receptor positive epithelial cells, indicating that Tgf-β1 may play an important role in restricting epithelial cell proliferation in these cells [82]. completely compensate for the loss of the Prlr signalling cascade. Prlr–/– MECs expressing lower levels of Elf5 The ets transcription factor Elf5 showed development that passed alveolar formation but Our transcript profiling experiments identified a number of failed during secretory initiation, mimicking the situation seen transcription factors that showed reduced expression in in Elf5+/– and Prlr+/– mice. Elf5 is a key mediator of structural response to a loss of Prlr, but profiling of a cell based model and functional development of lobuloalveoli [47]. Elf5 would of positive Prl action identified the ets transcription factor Elf5 thus appear to be a master-regulator of the alveolar switch [47]. Ets transcription factors are identified by a highly required for alveolar morphogenesis. conserved DNA binding domain (the ets domain), which binds to sites containing a central GGA motif [83]. Ets Conclusion transcription factors regulate gene expression during the It is apparent that a large number of genes can influence differentiation of multiple tissues including vascular, lymphoid, alveolar morphogenesis during pregnancy, some of which are muscle and bone (reviewed in [84]). Elf5 (e74-like factor 5 or shown in Figure 2. A better understanding of the components ESE-2) is an epithelial specific member of the Elf subfamily of of the alveolar switch, and thus the regulation of mammary Ets transcription factors, and is closely related to the cell proliferation and differentiation, has direct application to epithelial specific Elf3 (ESE-1) and Ehf (ESE-3) [85,86]. The the regulation of lactation in agricultural species and the predicted protein products of mouse Elf5 and human ESE-2 prevention and control of breast cancer. The key question is are 95% identical and are expressed as two isoforms how the expression of these numerous proteins is organised produced by alternative start sites. Such high conservation of and regulated by the alveolar switch. One potential model is a sequence implies similar conservation of function [86]. hierarchy of transcription factors that are each responsible for regulating an aspect of development. A precedent for this Elf5 is expressed specifically in the luminal cells of mammary model is provided by the action of the transcription factor tissue [47], and its expression is increased dramatically Srebf1, which regulates the expression of lipogenic enzymes during pregnancy, to levels that far exceed those seen in during secretory initiation [33]. In this model, Elf5 would be other tissues. Elf5 can also bind to an ets-like domain in the placed close to the origin of the hierarchy, as a master proximal promoter of whey acidic protein and induce its regulator of the transcriptional cascade controlling alveolar expression independently of lactogenic hormones, indicating morphogenesis. that Elf5 may be an important mediator of alveolar differentiation during mid-pregnancy [87]. Elf5–/– mice die in Competing interests utero due to a placentation defect [88]. Elf5+/– mice did not The authors declare that they have no competing interests. lactate due to failed alveolar development and, in some mice where alveoli had formed, differentiation into functional Acknowledgments This work is supported by The National Health and Medical Research secretory units was severely impaired [89]. Mammary Council of Australia, the Cancer Council New South Wales, United epithelial cell proliferation was reduced throughout alveolar States DoD BCRP (DAMD17-03-1-0686) (SRO), and the University of morphogenesis and secretory activation, and mammary New South Wales University Postgraduate Award (HNH). epithelial transplants demonstrated that this effect was cell References autonomous. The levels of Elf5 are reduced in Prlr+/– glands 1. Brisken C: Hormonal control of alveolar development and its implications for breast carcinogenesis. J Mammary Gland Biol and there is no similar reduction in the expression of Prlr in Neoplasia 2002, 7:39-48. Elf5+/–, indicating that Elf5 is downstream of the Prlr [89]. 2. Cardiff RD, Wellings SR: The comparative pathology of human MECs from Prlr–/– mammary glands fail to form lobuloalveoli and mouse mammary glands. J Mammary Gland Biol Neoplasia 1999, 4:105-122. during pregnancy when transplanted into the cleared fat pad 3. Richert MM, Schwertfeger KL, Ryder JW, Anderson SM: An atlas of hosts with a normal endocrine milieu. Retroviral re- of mouse mammary gland development. J Mammary Gland expression of Elf5 in Prlr–/– MECs followed by transplantation Biol Neoplasia 2000, 5:227-241. 4. Streuli CH, Edwards GM, Delcommenne M, Whitelaw CB, Burdon to a cleared fat pad resulted in a rescue of alveolar TG, Schindler C, Watson CJ: Stat5 as a target for regulation by morphogenesis [47]. MECs expressing high levels of Elf5 extracellular matrix. J Biol Chem 1995, 270:21639-21644. 5. Fata JE, Werb Z, Bissell MJ: Regulation of mammary gland proliferated and differentiated into distended, milk filled alveoli branching morphogenesis by the extracellular matrix and its [47]. Thus, re-expression of Elf5 in Prlr–/– MECs could remodeling enzymes. Breast Cancer Res 2004, 6:1-11.
Page 7 of 10 (page number not for citation purposes) Breast Cancer Research Vol 8 No 2 Oakes et al.
6. Neville MC, McFadden TB, Forsyth I: Hormonal regulation of 28. Binart N, Helloco C, Ormandy CJ, Barra J, Clement-Lacroix P, mammary differentiation and milk secretion. J Mammary Gland Baran N, Kelly PA: Rescue of preimplantatory egg develop- Biol Neoplasia 2002, 7:49-66. ment and embryo implantation in prolactin receptor-deficient 7. Rudolph MC, McManaman JL, Hunter L, Phang T, Neville MC: mice after progesterone administration. Endocrinology 2000, Functional development of the mammary gland: use of 141:2691-2697. expression profiling and trajectory clustering to reveal 29. Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, Buteau H, changes in gene expression during pregnancy, lactation, and Edery M, Brousse N, Babinet C, Binart N, et al.: Null mutation of involution. J Mammary Gland Biol Neoplasia 2003, 8:287-307. the prolactin receptor gene produces multiple reproductive 8. Nguyen DA, Parlow AF, Neville MC: Hormonal regulation of defects in the mouse. Genes Dev 1997, 11:167-178. tight junction closure in the mouse mammary epithelium 30. Brisken C, Kaur S, Chavarria TE, Binart N, Sutherland RL, Wein- during the transition from pregnancy to lactation. J Endocrinol berg RA, Kelly PA, Ormandy CJ: Prolactin controls mammary 2001, 170:347-356. gland development via direct and indirect mechanisms. Devel- 9. Neville MC, Medina D, Monks J, Hovey RC: The mammary fat opmental Biol 1999, 210:96-106. pad. J Mammary Gland Biol Neoplasia 1998, 3:109-116. 31. Wynick D, Hammond PJ, Akinsanya KO, Bloom SR: Galanin reg- 10. Djonov V, Andres AC, Ziemiecki A: Vascular remodelling during ulates basal and oestrogen-stimulated lactotroph function. the normal and malignant life cycle of the mammary gland. Nature 1993, 364:529-532. Microsc Res Tech 2001, 52:182-189. 32. Naylor MJ, Ginsburg E, Iismaa TP, Vonderhaar BK, Wynick D, 11. Lucas BK, Ormandy CJ, Binart N, Bridges RS, Kelly PA: Null Ormandy CJ: The neuropeptide galanin augments lobuloalve- mutation of the prolactin receptor gene produces a defect in olar development. J Biol Chem 2003, 278:29145-29152. maternal behavior. Endocrinology 1998, 139:4102-4107. 33. Naylor MJ, Oakes SR, Gardiner-Garden M, Harris J, Blazek K, Ho 12. Smalley M, Ashworth A: Stem cells and breast cancer: A field in TW, Li FC, Wynick D, Walker AM, Ormandy CJ: Transcriptional transit. Nat Rev Cancer 2003, 3:832-844. changes underlying the secretory activation phase of 13. DeOme KB, Faulkin LJJ, Bern HA, Blair PB: Development of mammary gland development. Mol Endocrinol 2005, 19:1868- mammary tumors from hyperplastic alveolar nodules trans- 1883. planted into gland-free mammary fat pads of female C3H 34. Wynick D, Small CJ, Bacon A, Holmes FE, Norman M, Ormandy mice. Cancer Research 1959, 19:515-520. CJ, Kilic E, Kerr NC, Ghatei M, Talamantes F, et al.: Galanin regu- 14. Smith GH, Boulanger CA: Mammary epithelial stem cells: lates prolactin release and lactotroph proliferation. Proc Natl transplantation and self-renewal analysis. Cell Prolif 2003, 36: Acad Sci USA 1998, 95:12671-12676. 3-15. 35. Allan GJ, Tonner E, Barber MC, Travers MT, Shand JH, Vernon 15. Shackleton M, F. V, Simpson KJ, Stingl J, Smythe GK, Asselin- RG, Kelly PA, Binart N, Flint DJ: Growth hormone, acting in part Labat M-L, Wu L, Lindeman GJ, Visvader JE: Generation of a through the insulin-like growth factor axis, rescues develop- functional mammary gland from a single stem cell. Nature mental, but not metabolic, activity in the mammary gland of 2006, 439:84-88. mice expressing a single allele of the prolactin receptor. 16. Smalley MJ, Clarke RB: The mammary gland “side population”: Endocrinology 2002, 143:4310-4319. a putative stem/progenitor cell marker? J Mammary Gland 36. Horseman ND: Prolactin and mammary gland development. J Biol Neoplasia 2005, 10:37-47. Mammary Gland Biol Neoplasia 1999, 4:79-88. 17. Stingl J, Raouf A, Emerman JT, Eaves CJ: Epithelial progenitors 37. Hennighausen L, Robinson GW: Information networks in the in the normal human mammary gland. J Mammary Gland Biol mammary gland. Nature Rev Mol Cell Biol 2005, 6:715-725. Neoplasia 2005, 10:49-59. 38. Han Y, Watling D, Rogers NC, Stark GR: JAK2 and STAT5, but 18. Traurig H: Cell Proliferation in the mammary gland during late not JAK1 and STAT1, are required for prolactin-induced beta- pregnancy and lactation. Anatomical Record 1967, 157:489- lactoglobulin transcription. Mol Endocrinol 1997, 11:1180- 503. 1188. 19. Brisken C, Park S, Vass T, Lydon JP, O’Malley BW, Weinberg RA: 39. Pezet A, Buteau H, Kelly PA, Edery M: The last proline of Box 1 A paracrine role for the epithelial progesterone receptor in is essential for association with JAK2 and functional activation mammary gland development. Proc Natl Acad Sci USA 1998, of the prolactin receptor. Mol Cell Endocrinol 1997, 129:199- 95:5076-5081. 208. 20. Conneely OM, Jericevic BM, Lydon JP: Progesterone receptors 40. Lebrun JJ, Ali S, Goffin V, Ullrich A, Kelly PA: A single phospho- in mammary gland development and tumorigenesis. J tyrosine residue of the prolactin receptor is responsible for Mammary Gland Biol Neoplasia 2003, 8:205-214. activation of gene transcription. Proc Natl Acad Sci USA 1995, 21. Clarke RB, Howell A, Potten CS, Anderson E: Dissociation 92:4031-4035. between steroid receptor expression and cell proliferation in 41. DaSilva L, Rui H, Erwin RA, Howard OM, Kirken RA, Malabarba the human breast. Cancer Res 1997, 57:4987-4991. MG, Hackett RH, Larner AC, Farrar WL: Prolactin recruits 22. Grimm SL, Seagroves TN, Kabotyanski EB, Hovey RC, Vonder- STAT1, STAT3 and STAT5 independent of conserved receptor haar BK, Lydon JP, Miyoshi K, Hennighausen L, Ormandy CJ, Lee tyrosines TYR402, TYR479, TYR515 and TYR580. Mol Cell AV, et al.: Disruption of steroid and prolactin receptor pattern- Endocrinol 1996, 117:131-140. ing in the mammary gland correlates with a block in lobulo- 42. Wakao H, Gouilleux F, Groner B: Mammary gland factor (MGF) alveolar development. Molecular Endocrinology 2002, 16: is a novel member of the cytokine regulated transcription 2675-2691. factor gene family and confers the prolactin response. EMBO 23. Brisken C, Heineman A, Chavarria T, Elenbaas B, Tan J, Dey SK, J 1995, 14:854-855. McMahon JA, McMahon AP, Weinberg RA: Essential function of 43. Liu X, Robinson GW, Wagner KU, Garrett L, Wynshaw-Boris A, Wnt-4 in mammary gland development downstream of prog- Hennighausen L: Stat5a is mandatory for adult mammary esterone signaling. Genes Dev 2000, 14:650-654. gland development and lactogenesis. Genes Dev 1997, 11: 24. Fata JE, Kong YY, Li J, Sasaki T, Irie-Sasaki J, Moorehead RA, 179-186. Elliott R, Scully S, Voura EB, Lacey DL, et al.: The osteoclast dif- 44. Liu X, Robinson GW, Gouilleux F, Groner B, Hennighausen L: ferentiation factor osteoprotegerin-ligand is essential for Cloning and expression of Stat5 and an additional homologue mammary gland development. Cell 2000, 103:41-50. (Stat5b) involved in prolactin signal transduction in mouse 25. Cao Y, Karin M: NF-kappaB in mammary gland development mammary tissue. Proc Natl Acad Sci USA 1995, 92:8831- and breast cancer. J Mammary Gland Biol Neoplasia 2003, 8: 8835. 215-223. 45. Teglund S, McKay C, Schuetz E, van Deursen JM, Stravopodis D, 26. Said TK, Conneely OM, Medina D, O’Malley BW, Lydon JP: Prog- Wang D, Brown M, Bodner S, Grosveld G, Ihle JN: Stat5a and esterone, in addition to estrogen, induces cyclin D1 expres- Stat5b proteins have essential and nonessential, or redun- sion in the murine mammary epithelial cell, in vivo. dant, roles in cytokine responses. Cell 1998, 93:841-850. Endocrinology 1997, 138:3933-3939. 46. Ormandy CJ, Naylor M, Harris J, Robertson F, Horseman ND, Lin- 27. Mulac-Jericevic B, Lydon JP, DeMayo FJ, Conneely OM: Defec- deman GJ, Visvader J, Kelly PA: Investigation of the transcrip- tive mammary gland morphogenesis in mice lacking the prog- tional changes underlying functional defects in the mammary esterone receptor B isoform. Proc Natl Acad Sci USA 2003, glands of prolactin receptor knockout mice. Recent Prog 100:9744-9749. Hormone Res 2003, 58:297-323.
Page 8 of 10 (page number not for citation purposes) Available online http://breast-cancer-research.com/content/8/2/207
47. Harris J, Stanford PM, Sutherland K, Oakes SR, Naylor MJ, 68. Barcellos-Hoff MH, Ewan KB: Transforming growth factor-beta Robertson FG, Blazek KD, Kazlauskas M, Hilton HN, Wittlin S, et and breast cancer: Mammary gland development. Breast al.: Socs2 and Elf5 mediate prolactin-induced mammary gland Cancer Res. 2000, 2:92-99. development. Mol Endocrinol 2006, 9. [epub ahead of print] 69. Grimm SL, Contreras A, Barcellos-Hoff MH, Rosen JM: Cell cycle 48. Harris J, Stanford PM, Oakes SR, Ormandy CJ: Prolactin and the defects contribute to a block in hormone-induced mammary prolactin receptor: new targets of an old hormone. Ann Med gland proliferation in CCAAT/enhancer-binding protein 2004, 36:414-425. (C/EBPbeta)-null mice. J Biol Chem 2005, 280:36301-36309. 49. Monaghan P, Moss D: Connexin expression and gap junctions 70. Kurpios NA, Sabolic NA, Shepherd TG, Fidalgo GM, Hassell JA: in the mammary gland. Cell Biol Int 1996, 20:121-125. Function of PEA3 Ets transcription factors in mammary gland 50. Bry C, Maass K, Miyoshi K, Willecke K, Ott T, Robinson GW, development and oncogenesis. J Mammary Gland Biol Neopla- Hennighausen L: Loss of connexin 26 in mammary epithelium sia 2003, 8:177-190. during early but not during late pregnancy results in unsched- 71. Jones FE, Stern DF: Expression of dominant-negative ErbB2 in uled apoptosis and impaired development. Dev Biol 2004, the mammary gland of transgenic mice reveals a role in 267:418-429. lobuloalveolar development and lactation. Oncogene 1999, 51. Imbert A, Eelkema R, Jordan S, Feiner H, Cowin P: Delta N89 18:3481-3490. beta-catenin induces precocious development, differentiation, 72. Long W, Wagner KU, Lloyd KC, Binart N, Shillingford JM, Hen- and neoplasia in mammary gland. J Cell Biol 2001, 153:555- nighausen L, Jones FE: Impaired differentiation and lactational 568. failure of Erbb4-deficient mammary glands identify ERBB4 as 52. Teuliere J, Faraldo MM, Deugnier MA, Shtutman M, Ben-Ze’ev A, an obligate mediator of STAT5. Development 2003, 130:5257- Thiery JP, Glukhova MA: Targeted activation of beta-catenin 5268. signaling in basal mammary epithelial cells affects mammary 73. Yang Y, Spitzer E, Meyer D, Sachs M, Niemann C, Hartmann G, development and leads to hyperplasia. Development 2005, Weidner KM, Birchmeier C, Birchmeier W: Sequential require- 132:267-277. ment of hepatocyte growth factor and neuregulin in the mor- 53. Srivastava S, Matsuda M, Hou Z, Bailey JP, Kitazawa R, Herbst phogenesis and differentiation of the mammary gland. J Cell MP, Horseman ND: Receptor activator of NF-kappaB ligand Biol 1995, 131:215-226. induction via Jak2 and Stat5a in mammary epithelial cells. J 74. Li L, Cleary S, Mandarano MA, Long W, Birchmeier C, Jones FE: Biol Chem 2003, 278:46171-46178. The breast proto-oncogene, HRGalpha regulates epithelial 54. Fantl V, Edwards PA, Steel JH, Vonderhaar BK, Dickson C: proliferation and lobuloalveolar development in the mouse Impaired mammary gland development in Cyl-1(-/-) mice mammary gland. Oncogene 2002, 21:4900-4907. during pregnancy and lactation is epithelial cell autonomous. 75. Luetteke NC, Qiu TH, Fenton SE, Troyer KL, Riedel RF, Chang A, Dev Biol 1999, 212:1-11. Lee DC: Targeted inactivation of the EGF and amphiregulin 55. Brisken C, Ayyannan A, Nguyen C, Heineman A, Reinhardt F, Jian genes reveals distinct roles for EGF receptor ligands in T, Dey SK, Dotto GP, Weinberg RA, Jan T: IGF-2 is a mediator mouse mammary gland development. Development 1999, of prolactin-induced morphogenesis in the breast. Dev Cell 126:2739-2750. 2002, 3:877-887. 76. Zutter MM, Sun H, Santoro SA: Altered integrin expression and 56. Matsuda M, Imaoka T, Vomachka AJ, Gudelsky GA, Hou Z, Mistry the malignant phenotype: the contribution of multiple inte- M, Bailey JP, Nieport KM, Walther DJ, Bader M, et al.: Serotonin grated integrin receptors. J Mammary Gland Biol Neoplasia regulates mammary gland development via an autocrine- 1998, 3:191-200. paracrine loop. Dev Cell 2004, 6:193-203. 77. Li N, Zhang Y, Naylor MJ, Schatzmann F, Maurer F, Wintermantel 57. Horton JD, Goldstein JL, Brown MS: SREBPs: activators of the T, Schuetz G, Mueller U, Streuli CH, Hynes NE: Beta1 integrins complete program of cholesterol and fatty acid synthesis in regulate mammary gland proliferation and maintain the the liver. J Clin Invest 2002, 109:1125-1131. integrity of mammary alveoli. EMBO J 2005, 24:1942-1953. 58. Coletta RD, Jedlicka P, Gutierrez-Hartmann A, Ford HL: Tran- 78. Naylor MJ, Li N, Cheung J, Lowe ET, Lambert E, Marlow R, Wang scriptional control of the cell cycle in mammary gland devel- P, Schatzmann F, Wintermantel T, Schuetz G, et al.: Ablation of opment and tumorigenesis. J Mammary Gland Biol Neoplasia β1 integrin in mammary epithelium reveals a key role for inte- 2004, 9:39-53. grin in glandular morphogenesis and differentiation. J Cell 59. Hatsell S, Rowlands T, Hiremath M, Cowin P: Beta-catenin and Biol 2005, 171:717-728. Tcfs in mammary development and cancer. J Mammary Gland 79. Naylor M, Streuli C Integrin regulation of mammary gland Biol Neoplasia. 2003, 8:145-158. development. In Integrins and Development. Edited by Danin E. 60. Foley J, Dann P, Hong J, Cosgrove J, Dreyer B, Rimm D, Dunbar Georgetown, TX: Landes Bioscience; 2006:176-185. M, Philbrick W, Wysolmerski J: Parathyroid hormone-related 80. Massague J, Chen YG: Controlling TGF-beta signaling. Genes protein maintains mammary epithelial fate and triggers nipple Dev 2000, 14:627-644. skin differentiation during embryonic breast development. 81. Ewan KB, Shyamala G, Ravani SA, Tang Y, Akhurst R, Wakefield Development 2001, 128:513-525. L, Barcellos-Hoff MH: Latent transforming growth factor-beta 61. Murtagh J, Martin F, Gronostajski RM: The Nuclear Factor I (NFI) activation in mammary gland: regulation by ovarian hormones gene family in mammary gland development and function. J affects ductal and alveolar proliferation. Am J Pathol 2002, Mammary Gland Biol Neoplasia 2003, 8:241-254. 160:2081-2093. 62. Johansson EM, Kannius-Janson M, Gritli-Linde A, Bjursell G, 82. Ewan KB, Oketch-Rabah HA, Ravani SA, Shyamala G, Moses HL, Nilsson J: Nuclear factor 1-C2 is regulated by prolactin and Barcellos-Hoff MH: Proliferation of estrogen receptor-alpha- shows a distinct expression pattern in the mouse mammary positive mammary epithelial cells is restrained by transform- epithelial cells during development. Mol Endocrinol 2005, 19: ing growth factor-beta1 in adult mice. Am J Pathol 2005, 167: 992-1003. 409-417. 63. Desprez PY, Sumida T, Coppe JP: Helix-loop-helix proteins in 83. Sharrocks AD, Brown AL, Ling Y, Yates PR: The ETS-domain mammary gland development and breast cancer. J Mammary transcription factor family. Int J Biochem Cell Biol 1997, 29: Gland Biol Neoplasia 2003, 8:225-239. 1371-1387. 64. Mori S, Nishikawa SI, Yokota Y: Lactation defect in mice lacking 84. Oikawa T, Yamada T: Molecular biology of the Ets family of the helix-loop-helix inhibitor Id2. Embo J 2000, 19:5772-5781. transcription factors. Gene 2003, 303:11-34. 65. Clarkson RW, Watson CJ: NF-kappaB and apoptosis in 85. Oettgen P, Kas K, Dube A, Gu X, Grall F, Thamrongsak U, Akbar- mammary epithelial cells. J Mammary Gland Biol Neoplasia ali Y, Finger E, Boltax J, Endress G, et al.: Characterization of 1999, 4:165-175. ESE-2, a novel ESE-1-related Ets transcription factor that is 66. Grimm SL, Rosen JM: The role of C/EBPbeta in mammary restricted to glandular epithelium and differentiated ker- gland development and breast cancer. J Mammary Gland Biol atinocytes. J Biol Chem 1999, 274:29439-29452. Neoplasia 2003, 8:191-204. 86. Zhou J, Ng AY, Tymms MJ, Jermiin LS, Seth AK, Thomas RS, Kola 67. Seagroves TN, Lydon JP, Hovey RC, Vonderhaar BK, Rosen JM: I: A novel transcription factor, ELF5, belongs to the ELF sub- C/EBPbeta (CCAAT/enhancer binding protein) controls cell family of ETS genes and maps to human chromosome fate determination during mammary gland development. Mol 11p13-15, a region subject to LOH and rearrangement in Endocrinol 2000, 14:359-368. human carcinoma cell lines. Oncogene 1998, 17:2719-2732.
Page 9 of 10 (page number not for citation purposes) Breast Cancer Research Vol 8 No 2 Oakes et al.
87. Thomas RS, Ng AN, Zhou J, Tymms MJ, Doppler W, Kola I: The Elf group of Ets-related transcription factors. ELF3 and ELF5. Adv Exp Med Biol 2000, 480:123-128. 88. Donnison M, Beaton A, Davey HW, Broadhurst R, L’Huillier P, Pfeffer PL: Loss of the extraembryonic ectoderm in Elf5 mutants leads to defects in embryonic patterning. Develop- ment 2005, 132:2299-2308. 89. Zhou J, Chehab R, Tkalcevic J, Naylor MJ, Harris J, Wilson TJ, Tsao S, Tellis I, Zavarsek S, Xu D, et al.: Elf5 is essential for early embryogenesis and mammary gland development during pregnancy and lactation. EMBO J 2005, 24:635-644.
Page 10 of 10 (page number not for citation purposes) Oncogene (2006), 1–11 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE Loss of mammary epithelial prolactin receptor delays tumor formation by reducing cell proliferation in low-grade preinvasive lesions
SR Oakes1, FG Robertson1, JG Kench2, M Gardiner-Garden1, MP Wand3, JE Green4 and CJ Ormandy1
1Cancer Research Program, Garvan Institute of Medical Research, Darlinghurst, Sydney, NSW, Australia; 2Department of Tissue Pathology, Institute of Clinical Pathology and Medical Research, Westmead, NSW, Australia; 3Department of Statistics, School of Mathematics, The University of New South Wales, Sydney, NSW, Australia and 4Transgenic Oncogenesis Group, Laboratory of Cell Regulation and Carcinogenesis, Bethesda, Maryland, MD, USA
Top quartile serum prolactin levels confer a twofold Introduction increase in the relative risk of developing breast cancer. Prolactin exerts this effect at an ill defined point in the The Nurses Health Study I (http://www.channing. carcinogenic process, via mechanisms involving direct harvard.edu/nhs/) is a large prospective study begun in action via prolactin receptors within mammary epithelium 1976. A case–control study conducted using this cohort and/or indirect action through regulation of other examined the risk of breast cancer conferred by elevated hormones such as estrogen and progesterone. We have serum prolactin levels. Blood samples were collected addressed these questions by examining mammary carci- between 1989 and 1990, and 306 postmenopausal nogenesis in transplants of mouse mammary epithelium women were subsequently diagnosed with breast cancer expressing the SV40T oncogene, with or without the before 1994. These women were matched to 448 control prolactin receptor, using host animals with a normal subjects. Measurement of serum prolactin demonstrated endocrine system. In prolactin receptor knockout trans- that top quartile serum prolactin (PRL) conferred a plants the area of neoplasia was significantly smaller higher relative risk (2.03-fold 95%CI 1.24–3.31 (7 versus 17%; Po0.001 at 22 weeks and 7 versus 14%; P ¼ 0.01) of developing breast cancer compared to P ¼ 0.009 at 32 weeks). Low-grade neoplastic lesions women with bottom quartile serum prolactin (Hankin- displayed reduced BrdU incorporation rate (11.3 versus son et al., 1999). The effect was independent of plasma 17% P ¼ 0.003) but no change in apoptosis rate. Tumor sex steroid hormones and exclusion of cases diagnosed latency increased (289 days versus 236 days, Po0.001). within 2 years of blood collection resulted in the same Tumor frequency, growth rate, morphology, cell prolif- conclusion. The cohort was updated with 851 cases eration and apoptosis were not altered. Thus, prolactin diagnosed by 2000, matched to 1275 controls (Tworoger acts directly on the mammary epithelial cells to increase et al., 2004). Overall the same positive correlation cell proliferation in preinvasive lesions, resulting in more between breast cancer risk and serum prolactin levels neoplasia and acceleration of the transition to invasive was seen (1.35-fold 95%CI 1.02–1.76 P ¼ 0.01), and this carcinoma. Targeting of mammary prolactin signaling association varied by sex hormone receptor status, with thus provides a strategy to prevent the early progression of ER þ PR þ tumors having an increased relative risk of neoplasia to invasive carcinoma. 1.78 (95% CI, 1.28, 2.50; P-trend o0.001) compared to Oncogene advance online publication, 24 July 2006; ER tumors (0.76 95% CI, 0.43, 1.32; P-trend ¼ 0.28). doi:10.1038/sj.onc.1209838 Rodent cancer models recapitulate the sensitivity of the human breast to PRL (Welsch and Nagasawa, 1977; Keywords: prolactin receptor; mouse; mammary; carcino- Wennbo and Tornell, 2000). Pituitary grafts or trans- genesis; C3(1)/SV40T genic strategies that increase serum Prl levels result in mammary cancer. For example, transgenic mice that overexpress human growth hormone, which binds both Prlr and growth hormone receptors, develop mammary carcinoma while mice over expressing the growth hormone receptor-restricted ligand bGH do not (Wennbo et al., 1997). Overexpression of rat Prl using the lipocalin promoter to drive expression predomi- Correspondence: Professor CJ Ormandy, Garvan Institute of Medical nantly in mammary epithelium produces (ER) positive Research, St Vincent’s Hospital, 384 Victoria St, Darlinghurst, tumors at a higher rate than other mouse mammary Sydney, NSW 2010, Australia. cancer models (Rose-Hellekant et al., 2003). Prl as a E-mail: [email protected] Financial Support: US CDMRP Predoctoral Scholarship DAMD modulator of preinitiated cancer has been examined in 17-03-1-0686, National Health and Medical Research Council. chemical carcinogenesis (Welsch et al., 1975) and Received 22 March 2006; revised 8 June 2006; accepted 22 June 2006 transgenic oncogene models. Prl and Prlr mRNA were Prolactin and mammary carcinogenesis SR Oakes et al 2 detected in nitrosomethylurea (NMU) carcinogen- We have scant knowledge regarding the point in the induced tumors and a rat Prl antiserum inhibited carcinogenic process where prolactin exerts its effect. It NMU-induced tumor cell proliferation by up to 70%, may exert an effect as a continuation of its normal compared to normal rabbit serum and GH antiserum developmental role, or acquire a novel role due to (Mershon et al., 1995). Mice expressing the polyoma dysregulation in cancer. There is evidence for the latter. middle-T antigen oncogene develop tumors in the first Prlr expression is increased in cancer compared to weeks of life, but when crossed with Prl knockout mice adjacent normal tissue (Touraine et al., 1998). Alter- they developed tumors significantly later (Vomachka natively altered ratios of a number of different splicing et al., 2000). These, and other experiments, have variants and isoforms could modulate signaling from demonstrated that Prl alone at high levels is sufficient this receptor in cancer (Clevenger et al., 1995). The to produce mammary cancer, and that its loss can retard C3(1)SV40T model of mammary cancer provides a tumor formation in response to an oncogenic stimulus. reproducible series of defined neoplastic lesions that The mechanisms behind these important observations progress to invasive carcinoma and resemble the human are not clear. It has generally been assumed that disease (Maroulakou et al., 1994; Shibata et al., 1998; prolactin exerts its effects via direct modulation of the Green et al., 2000). In combination with genetic deletion mammary epithelial cell (reviewed Vonderhaar, 1999; of the Prlr, this provides an ideal model to investigate Clevenger et al., 2003) and there is evidence consistent where in the carcinogenic process prolactin acts. By with this. Thus, prolactin receptors (PRLR) are using mammary epithelial transplantation to remove the expressed at high levels predominantly by steroid disruption to the ovarian–pituitary endocrine axis that is hormone receptor positive breast cancer cells and caused by systemic deletion of the Prlr, we can tumors (Bonneterre et al., 1987; Ormandy et al., distinguish the direct and indirect actions of Prl. We 1997c), but at low levels by most tumors (Reynolds have used this approach to examine the questions of et al., 1997). Prolactin causes an increase in proliferation direct or indirect action and point of influence in the (Ginsburg and Vonderhaar, 1995; Das and Vonderhaar, carcinogenic process. 1997; Llovera et al., 2000) and cyclin D1 expression (Schroeder et al., 2003) in breast cancer cell lines selected for prolactin sensitivity (Schroeder et al., 2002). Like- wise the use of PRLR antagonists can reduce prolifera- Results tion (Goffin et al., 1996; Chen et al., 1999; Llovera et al., Tumor formation in Prlr / /C3(1)/SV40T and 2000) in some breast cancer cell lines. PRL is also produced by mammary epithelial cells and has been WT/C3(1)/SV40T mice The development of palpable tumors was monitored in hypothesized to act via an autocrine mechanism / (Clevenger et al., 1995; Ginsburg and Vonderhaar, Prlr and WT mice that carried the C3(1)/SV40T 1995; Reynolds et al., 1997). Transplants of mouse construct. C3(1)/SV40T mice which lacked Prlr had significantly increased latency to palpable tumor forma- mammary epithelium lacking the Prl gene show normal 7 development during pregnancy, but show a threefold tion (200 9 days) compared to control C3(1)/SV40T 7 P ¼ / reduction in cell proliferation at parturition, the time at mice (175 7 days, logrank 0.033; Figure 1a). Prlr / C3(1)/SV40T mice also reached a tumor burden of 10% which PRL production by the epithelium becomes 7 apparent (Naylor et al., 2003). PRL may also exert a body weight significantly later (243 15 days) compared to WT/C3(1)/SV40T mice (217715 days, logrank direct effect via the modulation of the sensitivity of the ¼ epithelial cell to the action of other hormones. For P 0.032; Figure 1b). To determine whether Prlr example, exogenous Prl modulates the expression of signaling affected tumor growth rate, a mixed effects progesterone receptors by human breast cancer cells linear model was applied to the cubed root of tumor (Ormandy et al., 1997c), while endogenous PRL can volume for each experimental animal group. Tumors influence estrogen receptor (ER) alpha levels (Gutzman that were filled with fluid at the ethical end point were et al., 2004). excluded from the analysis. Results are plotted with day Recent work in mice, however, demonstrates that 0 as the day the tumor was detected (Figure 1c). No prolactin also exerts potent indirect effects on the significant difference in the rate of change in tumor volume was detected in Prlr / /C3(1)/SV40T mice mammary gland via modulation of the systemic endo- ¼ crine system. Null mutation of Prl or Prlr in mice compared to control C3(1)/SV40T mice (P 0.45). (Horseman et al., 1997; Ormandy et al., 1997a) results in disruption of ovarian, pituitary and other endocrine Tumor morphology in Prlr / /C3(1)/SV40T and systems (Clement-Lacroix et al., 1999). Thus, failed WT/C3(1)/SV40T mice mammary ductal side branching during ductal morpho- Histological examination of tumor tissues collected at genesis is restored by transplanting Prlr / epithelium the ethical end point was undertaken (Table 1). Some of into the cleared fat pads of hosts with normal endocrine the smaller tumors exhibited areas that had not yet function (Brisken et al., 1999) or by progesterone pellet invaded through the basement membrane and represent administration (Binart et al., 2000). Modulation of the a stage similar to human ductal carcinoma in situ. The endocrine system by prolactin provides a potential invasive tumors demonstrated a high-grade morpho- mechanism underlying the results to date in mouse logy, high mitotic index, coarse chromatin structure, models and the Nurses Health Study. pleomorphic nuclei and foci of necrosis. Tumors often
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 3
Figure 1 Tumor formation in Prlr / /C3(1)/SV40T mice (a) Tumor-free survival curve (Kaplan–Meier). Time is represented in days after birth. Prlr / /C3(1)/SV40T mice (circles) develop palpable tumors significantly later compared to WT/C3(1)/SV40T mice (squares, P ¼ 0.033). (b) Survival curve. Time is represented in days after birth. Prlr / /C3(1)/SV40T mice (circles) reach the ethical end point of 10% tumor burden significantly later than WT/C3(1)/SV40T mice (squares, P ¼ 0.032). (c) Tumor volume trellis plot. The cube root of tumor volume (volume1/3) is plotted with respect to time after detection (days). Time 0 is the day of initial detection. There was no significant difference in the rate of change in tumor volume from WT/C3(1)/SV40T mice (left box) and Prlr / /C3(1)/SV40T mice (right box) (P ¼ 0.45). (d) Bar graph summary of mammary whole-mount analysis. Prlr / /C3(1)/SV40T mice (white bars) showed a trend towards reduced area of lesions as a percentage of total mammary fat pad area compared to WT/C3(1)/SV40T mice (black bars, P ¼ 0.08). The area of lesions when classified into neoplasia and tumor was reduced in Prlr / /C3(1)/SV40T mice compared to WT/C3(1)/SV40T mice although this was not statistically significant (P ¼ 0.19 and P ¼ 0.15, respectively).
Table 1 Tumor morphology and cellular structure in tumors from WT/C3(1)/SV40T and Prlr / /C3(1)/SV40T mice % Area of tumor Cellular structure
Epithelium genotype n Acinar/Glandular Papillary Solid Type A Type B
Reached 10% tumor burden WT/C3(1)/SV40T 30 17731173727567 33 Prlr / /C3(1)/SV40T 29 18731975637674 26 P-value 0.86 0.19 0.25 w2 P ¼ 0.28
Did not reach 10% tumor burden WT/C3(1)/SV40T 12 23741677617877 23 Prlr / /C3(1)/SV40T 20 15731074757667 33 P-value 0.14 0.46 0.17 w2 P ¼ 0.12
All tumors collected WT/C3(1)/SV40T 42 19731273697470 30 Prlr / /C3(1)/SV40T 49 17721574687471 29 P-value 0.49 0.47 0.84 w2 P ¼ 0.88 n; number of tumors. The average percent of tumor area classified as acinar and/or glandular, papillary and solid is given7s.e. The percentage of tumors classified as having type A or type B cellular structure is also shown. displayed more than one low-power architectural detail previously (Cardiff et al., 2000). Cellular mor- pattern including acinar and/or glandular, papillary phology demonstrated two main variants: firstly tumor and solid areas. Acinar and glandular patterns were cells with a high nucleocytoplasmic ratio, hyperchro- grouped together for analysis as they often tended to matic nuclei, coarse chromatin and mild to moderate merge into each other rendering reliable distinction pleomorphism designated (Type A); and secondly cells problematic. These pathologies have been described in with a lower nucleocytoplasmic ratio, a smaller amount
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 4
Figure 2 SV40T expression, body weight and mammary morphology in Prlr / /C3(1)/SV40T mice (a) Bar graph of relative SV40T mRNA expression. There was no significant difference in the relative expression of SV40T mRNA in the inguinal mammary glands from WT/C3(1)/SV40T mice and Prlr / /C3(1)/SV40T (P ¼ 0.94) (b) Body weight trellis plot for WT (top left) and Prlr / (top right) mice and WT/C3(1)/SV40T (bottom left) and Prlr / /C3(1)/SV40T (bottom right) transplants. Age (days) is represented on the horizontal axis and body weight (grams) on the vertical axis. The increase in body weight is significantly more gradual in Prlr / and Prlr / /C3(1)/SV40T compared to control WT and WT/C3(1)/SV40T (P ¼ 0.006 and P ¼ 0.005, respectively). (c) Prlr / /C3(1)/SV40T inguinal mammary gland, Carmine stain. (d) WT/C3(1)/SV40T inguinal mammary gland, Carmine stain.
of eosinophilic cytoplasm, more vesicular chromatin mRNA. Expression of SV40T mRNA was detected in and more marked pleomorphism (Type B). Some all mammary glands from 12 week C3(1)/SV40T tumors demonstrated intermediate forms and were inguinal mammary glands (Figure 2a). There was no classified according to the predominant features. Areas significant difference in the relative expression of SV40T of necrosis were observed in 98% of tumors at between mammary glands from 12-week-old WT/C3(1)/ collection. Statistical analysis of tumor classification SV40T (1870.5 U) and Prlr / /C3(1)/SV40T mice in a blinded fashion revealed no significant difference in (1871.2 U, t-test P ¼ 0.94). tumor architecture or cellular morphology between tumors derived from Prlr / /C3(1)/SV40T mice or control C3(1)/SV40T mice. Body weight and mammary gland morphology in Prlr / /C3(1)/SV40T and WT/C3(1)/SV40T mice / To determine if Prlr mice differ in body weight Mammary neoplasia in Prlr / /C3(1)/SV40T and compared to control mice, both WT and C3(1)/SV40T WT/C3(1)/SV40T mice animals were aged and weighed weekly. A mixed effects Mammary whole mounts were analysed for the devel- linear model demonstrated that both Prlr / and Prlr / / opment of neoplastic lesions. Analysis of mammary C3(1)/SV40T animals gained weight at a reduced rate whole mounts collected at the ethical end point compared to WT and WT/C3(1)/SV40T mice (P ¼ 0.006 (Figure 1d) demonstrated that C3(1)/SV40T mice that and P ¼ 0.005, respectively; Figure 2b). Control Prlr / lacked Prlr displayed a trend toward reduced lesion-area mice were approximately 17% lighter (average measured as a percentage of total mammary fat pad area 27.870.8 g) at 50 weeks of age compared to control (8.971.3%) compared to WT/C3(1)/SV40T mice at the WT mice (average 33.671.3 g). Reduced body weight in ethical end point (13.672.2%, P ¼ 0.08). female Prlr / mice is due to reduced abdominal fat stores via a mechanism that includes altered endocrine SV40T levels in Prlr / /C3(1)/SV40T and environment (Freemark et al., 2001) and possibly Prlr WT/C3(1)/SV40T mice expression by adipocytes (Ling et al., 2000). Mammary Some mammary specific promotors such as MMTV and whole mounts (Figure 2c) collected from these mice at WAP are sensitive to pregnancy and/or hormone stimuli 50 weeks demonstrated a failure of ductal side branch- (Hutchinson and Muller, 2000). This complicates ing in Prlr / animals compared to WT (Figure 2d), as investigation of endocrine-mediated carcinogenesis as reported previously in prolactin and prolactin receptor observed effects may simply be due to changes in knockout (Horseman et al., 1997; Ormandy et al., transgene expression. The C3(1) rat prostatic steroid- 1997a). These results are due to Prl modulation of binding protein (PSBP) promoter used here is not progesterone levels via the pituitary–ovarian axis steroid hormone responsive (Shibata et al., 1998; Green (Binart et al., 2000). These changes in body weight et al., 2000). To confirm that SV40T expression was not and mammary epithelial cell content potentially altered by Prlr genotype, quantitative real time PCR was confound our results regarding altered tumor latency. used to examine the relative expression of SV40T These problems also potentially confound the results
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 5 obtained in many other rodent models of Prl action. SV40T epithelium compared to tumors from WT/C3(1)/ Increased Prl levels produced via pituitary graft, Prl SV40T epithelium (P ¼ 0.33). WT/C3(1)/SV40T trans- injection or transgenic methods, or loss of Prl produced plants produced a total of 29 tumors while Prlr / /C3(1)/ by knockout or pituitary ablation, may have altered the SV40T transplants produced 27, with mean tumors per systemic hormonal environment causing undetected transplant of 1.670.2 and 1.270.3, respectively, reveal- changes in body weight and mammary epithelial cell ing no detectable significant (P ¼ 0.23) difference in number. To investigate this problem, we utilized tumor frequency between genotypes. mammary epithelial transplantation. This procedure rescues the defect in ductal side branching and negates / the body weight issue by placing the test glands in a SV40T expression is unaltered by loss of Prlr normal endocrine environment (Brisken et al., 1999). We examined SV40T levels in 8 (56 days), 22 (154 days) and 32 week (224 days) old C3(1)/SV40T mammary glands formed by transplantation (Figure 4a). No / Tumor formation in Prlr /C3(1)/SV40T and significant difference was observed at 8, 22 and 32 WT/C3(1)/SV40T transplants weeks post surgery between WT/C3(1)/SV40T (1670.8, / Mammary glands made from Prlr epithelium devel- 1871.0 and 1970.2 U) and Prlr / /C3(1)/SV40T epi- oped palpable tumors significantly later than mammary thelial transplants (1770.5, 1870.3 and 1970.3 U, 7 glands with WT/C3(1)/SV40T epithelium (289 22 days P ¼ 0.10, P ¼ 0.64, P ¼ 0.78, respectively). Western 7 o verses 236 24 days, logrank P 0.001; Figure 3a). We blotting using an antibody against SV40T protein was looked directly for an effect of Prlr genotype on tumor used to determine the protein expression of SV40T in 8 growth rate using a mixed effects linear model described week (56 day) old transplants (Figure 4b). Detection of above (Figure 3b). Tumors that were filled with fluid at ethical end point were excluded from the analysis. Overall there was no significant difference in the rate of tumor growth in tumors derived from Prlr / /C3(1)/
Figure 4 SV40T expression in Prlr / /C3(1)/SV40T transplants (a) Relative expression of SV40T mRNA in 55 day (8 week), 154 (22 week) and 224 (32 week) old transplants. There was no significant Figure 3 Tumor formation in Prlr / /C3(1)/SV40T transplants (a) difference in the expression of SV40T at 8, 22 and 32 weeks Tumor-free survival curve. Time is represented in days after between Prlr / /C3(1)/SV40T (white bars) and WT/C3(1)/SV40T transplantation. Palpable tumors were detected in Prlr / /C3(1)/ (black bars) transplants (P ¼ 0.10, P ¼ 0.64, P ¼ 0.78, respectively). SV40T transplants (circles) significantly later compared to WT/ (b) Western blot of SV40T protein expression in 55 day (8 week) C3(1)/SV40T transplants (Po0.001). (b) Tumor volume trellis plot. old transplants. b-actin is shown as a loading control. Four donor The cube root of tumor volume (volume1/3) is plotted with respect animals are indicated by the numbers above the blot. (c) Average to time (days). Time 0 is the day of initial detection. There was no volume of SV40T protein normalised to b-actin. There was no significant difference in the rate of change in tumor volume from significant difference in the expression of SV40T protein between WT/C3(1)/SV40T transplants (left box) and Prlr / /C3(1)/SV40T WT/C3(1)/SV40T (black bar) and Prlr / /C3(1)/SV40T (white bar) transplants (right box) (P ¼ 0.33). 8 week old transplants.
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 6 b-actin protein was used as a loading control. The that the small increase in acinar/glandular tumors may average volume of SV40T protein in WT/C3(1)/SV40T simply reflect the longer latency in Prlr / /C3(1)/SV40T transplants was 8.670.7% which was not significantly transplants rather than an effect of Prlr. Overall, Prlr different to Prlr / /C3(1)/SV40T transplants (10.67 null epithelium does not appear to change the mechan- 2.4%, P ¼ 0.45), indicating that like the mRNA expres- ism determining the morphology of SV40T-induced sion of SV40T, the protein expression of the transgene is tumors. not altered by the presence of the Prlr. Univariate regression analysis demonstrated that SV40T mRNA SV40T-induced neoplasia is delayed in Prlr / mammary expression was not a predictor for age of detection epithelial cells ¼ ¼ (P 0.52), tumor latency (P 0.95) and days with In order to determine whether the presence of Prlr in ¼ tumor (P 0.52). mammary epithelium can modulate the development of SV40T-induced neoplasia, we collected mammary Loss of Prlr within epithelium does not change the glands made from WT/C3(1)/SV40T and Prlr / /C3(1)/ histology and morphology of SV40T-induced tumors SV40T epithelial transplants at 22 (154 days) (Figure 5a We then investigated whether loss of Prlr in the and b, respectively) and 32 weeks (224 days) (Figure 5c epithelium changed the histological appearance of and d, respectively). The Rag1 / C57BL/6J mouse SV40T-induced tumors. Tumor tissues were collected strain used as our transplant host develops very few and hematoxylin and eosin (H&E) histology was mammary ductal side branches (Figure 5e), a feature of undertaken in a similar manner to C3(1)/SV40T mice this mouse strain that is dependent on factors from the described above. There was similar diversity in the stroma and not the epithelial donor (Naylor and microscopic features of lesions observed in tumors taken Ormandy, 2002). Thus, donor tissue from a mixed from mammary gland transplants, the histopathology FVB/N and 129Ola/Pas strain develops a mammary tree was comparable to that observed in C3(1)/SV40T mice that shows a predominantly primary ductal branching described above. Areas of necrosis were found in 100% pattern, formed by bifurcation during ductal elongation of tumors taken from mice that had reached 10% tumor at pregnancy (Y-shaped junctions), with sparse side burden. Only 16/25 and 2/14 palpable tumors collected branches (T-shaped junctions), when transplanted into for histological investigation reached the ethical end C57BL/6J Rag1 / hosts. Abnormal development in point from control C3(1)/SV40T epithelium and Prlr / / WT/C3(1)/SV40T transplants first appears as an in- C3(1)/SV40T epithelium, respectively. The large laten- creased number of short side branches at abnormally cies observed in the formation of palpable tumors from close spacing (Figure 5a), a feature that is not seen in these transplants resulted in the lengthening of the control transplants without the C3(1)/SV40T construct. experiment beyond the normal healthy life span of a In contrast, Prlr / /C3(1)/SV40T epithelial transplants Rag1 / immune-compromised host. Therefore, a large at the same age exhibit the same developmental proportion of tumors were collected before reaching the abnormality, but at a greatly reduced frequency predetermined end point size. There was little variation (Figure 5b). in tumor architecture or cellular morphology between To quantify neoplastic area, we assessed the area tumors derived from mammary glands made from occupied by SV40T-induced lesions in carmine alum Prlr / /C3(1)/SV40T epithelium and control C3(1)/ stained Prlr / /C3(1)/SV40T mammary whole mounts. SV40T epithelium (Table 2). There was no significant C3(1)/SV40T mammary epithelium lacking Prlr had a difference in percentage areas of papillary and solid or significantly smaller area of total lesions at 22 cellular structure. A small increase was detected in the (7.671.7%; Figure 5f) and 32 weeks (11.772.4%; percentage of tumors that displayed acinar/glandular Figure 5g) compared to control C3(1)/SV40T epithelium characteristics in tumors from Prlr / /C3(1)/SV40T (18.471.2 and 24.172.8%, Po0.001 and P ¼ 0.005, epithelium, but we detected no difference in tumor type respectively). We divided the lesions into neoplasia and (acinar/glandular, papillary and solid) as a function tumor. Prlr / /C3(1)/SV40T transplants also had less of high and low Prlr expression level in tumors from neoplastic or tumor area at 22 weeks (7.571.6 and WT/C3(1)/SV40T epithelium (P ¼ 0.22, P ¼ 0.15 and 0.270.1%, respectively) and 32 weeks (7.371.1 and P ¼ 0.54, respectively; data not shown). This suggests 4.471.6%, respectively) than control C3(1)/SV40T
Table 2 Tumor morphology and cellular structure in tumors from WT/C3(1)/SV40T and Prlr / /C3(1)/SV40T transplants Tumor morphology (%) Cellular structure (%)
Epithelium Genotype n Acinar/Glandular Papillary Solid Type A Type B
All tumors collected WT/C3(1)/SV40T 25 2174177661776040 Prlr / /C3(1)/SV40T 14 3576207645775743 P-value 0.042 0.76 0.10 w2 P ¼ 0.67
n; number of tumors. The average percent of tumor area classified as acinar and/or glandular, papillary and solid is given7s.e. The percentage of tumors classified as having type A or type B cellular structure is also shown.
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 7
Figure 5 Neoplastic development in Prlr / /C3(1)/SV40T transplants. Mammary whole mounts of WT/C3(1)/SV40T transplants (a, c) and Prlr / /C3(1)/SV40T transplants (b, d) at 22 weeks (a, b) and 32 weeks (c, d) post-transplantation. Scale bars represent 500 mm. Virgin Rag1 / endogenous inguinal mammary gland whole mount (e), L denotes lymph node. Bar graph summary of mammary whole-mount analysis at 22 weeks (f) and 32 weeks (g). Prlr / /C3(1)/SV40T transplants (white bars) had less area of total lesions, neoplasia and tumor as a percentage of total mammary gland area compared to WT/C3(1)/SV40T transplants (black bars) at 22 weeks (Po0.001, Po0.001 and P ¼ 0.061, respectively) and 32 weeks (P ¼ 0.005, P ¼ 0.009 and P ¼ 0.11, respectively). transplants at 22 weeks (17.471.2 and 1.070.4%; ductal epithelium displaying a normal epithelial Po0.001 and P ¼ 0.06, respectively) and 32 weeks morphology (7.871.8%, Figure 6e; t-test P ¼ 0.019). (14.472.0 and 9.872.6%; P ¼ 0.009 and P ¼ 0.11, We detected significantly less proliferation in low-grade respectively). The ratio of neoplasia to tumor in Prlr / / and high-grade MIN lesions from Prlr / /C3(1)/SV40T C3(1)/SV40T transplants at 22 weeks was greater than transplants (11.371.2 and 13.071.4%) compared to in WT/C3(1)/SV40T transplants, however, this ratio control C3(1)/SV40T transplants (17.071.2 and equalled control levels by 32 weeks. 17.572.0% Figure 6e; t-test P ¼ 0.003 and P ¼ 0.067), demonstrating that loss of Prlr results in reduced SV40T induced proliferation in early stage lesions. There Cellular proliferation in SV40T-induced neoplasia is was no significant difference in the number of prolife- mediated by Prlr within mammary epithelium rating cells in invasive lesions from Prlr / /C3(1)/SV40T H&E histology allows the division of neoplasia into transplants compared to control C3(1)/SV40T trans- low-grade mammary intraepithelial neoplasia (LGMIN) plants. and high-grade mammary intraepithelial neoplasia (HGMIN). LGMIN displayed the presence of stratified atypical ductal epithelial cells with elongated, hyper- Apoptosis via activation of Caspase-3 is unaltered by Prlr chromatic and pleomorphic nuclei. HGMIN were within mammary epithelium present at multiple foci, and showed greater cellular We investigated the effect of a loss of Prlr signaling on crowding, more stratification, loss of polarity and apoptosis using an antibody raised against the cleaved increased pleomorphism and hyperchromatism. Often and active form of Caspase-3, a marker of cellular neoplastic cells completely filled the ductal lumen. apoptosis (Figure 6c and d), We detected no cleaved Invasive lesions are distinguished from HGMIN by Caspase-3 positive cells in typical epithelium from both breaching the basement membrane and stromal inva- Prlr / /C3(1)/SV40T and control C3(1)/SV40T trans- sion. We used BrdU immunocytochemisty to investigate plants. A low level of apoptosis was detected in the effect of Prlr on SV40T-induced cellular prolifera- LGMIN, which increased in HGMIN lesions and was tion within these lesion types (Figure 6a and b). A maintained at the same level in invasive lesions. There significant increase was detected in the proliferation rate was no significant difference in the rates of apoptosis in of cells from WT/C3(1)/SV40T preinvasive lesions LGMIN, HGMIN and invasive lesions from Prlr / / (17.071.2%) compared to ‘typical’ WT/C3(1)/SV40T C3(1)/SV40T (0.270.1, 1.470.3 and 1.270.5%) and
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 8 control C3(1)/SV40T transplants (0.370.1, 1.270.2 and Tworoger et al., 2004), estrogen (Missmer et al., 2004) 1.370.2%, Figure 6F; t-tests P ¼ 0.42, P ¼ 0.59 and and the androgenic precursors of estrogen (Hankinson P ¼ 0.92, respectively). These results indicate that Prlr et al., 1998) are hormones that increase the risk of breast signaling has no effect on cellular survival via apoptotic cancer in women who experience serum levels within the mechanisms involving cleavage of Caspase-3 in SV40T- top quartile of the population range. We have examined induced lesions. how prolactin acts to modulate carcinogenesis, using a model in which mammary cancer is initiated by the SV40T oncogene in the absence or presence of an intact prolactin signaling pathway. Using whole animals or Discussion transplanted glands, we have been able to contrast the direct action of prolactin on the mammary epithelial cell Large-scale prospective studies of breast cancer have with its indirect actions. By a combination of long- demonstrated that prolactin (Hankinson et al., 1999; itudinal survival analysis and cross-sectional histological studies, we have defined the influence of prolactin over latency, numbers and types of tumors produced by SV40T, and we have identified the stage in this carcinogenic process where prolactin acts. Tumor latency increased as a consequence of the loss of Prlr by 26 days in whole animals (12.0%) and by 53 days (22.5%) in transplants. This comparison shows that the direct effect of prolactin via the mammary epithelial prolactin receptor is the predominant mechan- ism by which prolactin modulates mammary carcino- genesis. This is also the first demonstration of a mammary cell autonomous effect of prolactin outside of pregnancy. The indirect effects of prolactin are complex. Loss of the prolactin receptor caused reduced estrogen and progesterone levels but increased para- thyroid hormone (Clement-Lacroix et al., 1999). Prlr / mice also had reduced insulin levels and sensitivity (Freemark et al., 2001) and decreased body weight. Despite these changes in the endocrine environment, comparison of the difference in relative tumor latency between tumor genotypes in the whole animal (26 days) and transplant experiments (53 days) indicates that the combined effects of these indirect actions on carcino- genesis are negligible. The stage of the carcinogenic process that is influenced by prolactin has not previously been defined. Transplants lacking Prlr showed greatly reduced areas of neoplasia and longer latency to the first palpable tumor. An analysis of cell proliferation showed a reduction in cell proliferation in the neoplasias as a
Figure 6 Cellular proliferation and apoptosis in Prlr / /C3(1)/ SV40T epithelium. Immunohistochemistry using antibodies against 5-bromo-20-dexyuridine (BrdU) in WT/C3(1)/SV40T transplants (a) and Prlr / /C3(1)/SV40T transplants (b) and cleaved Caspase-3 in WT/C3(1)/SV40T transplants (c) and Prlr / /C3(1)/SV40T transplants (d). (e) The percentage of BrdU positive cells in areas of typical ductal epithelium, LGMIN, HGMIN and invasion was reduced in Prlr / /C3(1)/SV40T transplants (white bars) compared to WT/C3(1)/SV40T transplants (black bars). Prlr / /C3(1)/SV40T transplants had significantly reduced proliferation via detection of BrdU staining in lesions classified as LGMIN and HGMIN (P ¼ 0.003 and P ¼ 0.067, respectively). (f) There was an undetect- able level (ND) of cleaved Caspase-3 staining in typical ductal epithelium from WT/C3(1)/SV40T and Prlr / /C3(1)/SV40T trans- plants. There was no significant difference in the rate of apoptosis via measurement of cleaved Caspase-3 in LGMIN, HGMIN and invasive lesions from Prlr / /C3(1)/SV40T transplants (white bars) compared to WT/C3(1)/SV40T transplants (black bars; P ¼ 0.42, P ¼ 0.59 and P ¼ 0.92, respectively).
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 9 result of a loss of the Prlr. Apoptosis was unaffected by Materials and methods genotype at every stage examined. Thus, prolactin acts at the very earliest stages of carcinogenesis to increase Mice and breeding cell proliferation in neoplastic lesions, resulting in a All experiments involving mice were performed under the greater area of neoplasia and a more rapid emergence of supervision and in accordance with the regulations of the Garvan/St Vincent’s Animal Experimentation Committee. The invasive tumors. Prlr / mouse was generated as previously described (Ormandy We also found that in this model of carcinogenesis, et al., 1997b). The C3(1)/SV40T animals (Maroulakou et al., loss of Prlr did not influence the growth of invasive 1994) were on an inbred FVB/N background, and the core lesions. Overall there was no difference in the prolifera- colony was maintained by homozygous matings. Mice hetero- tion rate or the growth rate of invasive lesions between zygous for both Prlr / and C3(1)/SV40T were produced by prolactin genotypes in either the whole animals or the mating homozygous Prlr / males and homozygous C3(1)/ transplants. An analysis of WT tumor transplants SV40T females. Female heterozygous progeny were then back showed no relationship between Prlr expression level crossed to homozygous Prlr / males to produce mice heterozygous or wildtype (WT) for C3(1)/SV40T, and and tumor growth rate despite the detection of the / expected difference between genotypes of latency to Prlr . Control WT mice were produced by using WT males in an identical but separate scheme to ensure similar genetic palpable tumor. Close inspection of Figure 3b shows a diversity between groups (Robertson et al., 2003). Rag1 / / dichotomy of tumor growth rates in Prlr transplants mice (Mombaerts et al., 1992) of the C57BL/6J strain were compared to WT. Two distinct types of growth rate are purchased from Animal Resource Centre, Perth, Australia. All seen in Prlr / tumors, a majority with slow growth and animals were housed with food and water ad libitum with a three tumors that showed initial slow growth followed 12 h day/night cycle at 221C and 80% relative humidity. by a dramatic increase in growth rate. This dichotomy is Rag1 / mice were administered Resprim (Alphapharm, reflected in the size of the BrdU error bar for Prlr / Carole Park, Australia), containing Sulfamethoxazole/Tri- tumors (Figure 6e). WT tumors show a broad spectrum methoprim via drinking water (5 mg/1 mg per 5 ml drinking of growth rates. Although it is tempting to speculate water) in alternate weeks. Mice wild-type for C3(1)/SV40T that this difference in growth pattern may reflect a were weighed weekly and aged to 50 weeks. fundamental difference in tumor biology between genotypes, this effect was not seen in whole animals or Experimental groups in relation to Prlr expression level. A few advanced Twenty WT/C3(1)/SV40T and 25 Prlr / /C3(1)/SV40T mice tumors may remain sensitive to prolactin, but in our were observed twice weekly for tumor formation (whole experiments their frequency was not sufficient to animal study). The date of the first palpable tumor was influence the analysis. recorded (age of detection) and tumor size monitored using Thus, prolactin may facilitate tumor formation in two vernier callipers. The volume of each tumor was estimated ways: by increasing the number of neoplastic cells using the major and minor axes of palpable tumors (Attia and prolactin increases the chance of a tumorigenic event; Weiss, 1966). Mice were killed when the tumor burden reached and by driving the proliferation of neoplastic cells 10% of the animal’s body weight (ethical end point) or earlier if the animal became unhealthy. At killing, tumors were prolactin forces cell divisions that may replicate collected for routine H&E histochemistry, and remaining tumorigenic genetic or epigenetic events. Once these mammary glands were collected for whole mount histology as events occur the resulting invasive lesions generally described below. become independent of prolactin. This observation Mammary epithelium transplants were performed as pre- explains why bromocriptine treatment of patients with viously described (DeOme et al., 1959). Approximately 1 cm3 advanced breast cancer was not successful (Peyrat et al., section of the fourth mammary gland was excised from 5- to 1984; Bonneterre et al., 1988; Manni et al., 1989), and 8-week old WT/C3(1)/SV40T and Prlr / /C3(1)SV40T donors why prolactin treatment of breast cancer cell lines does (before the onset of neoplasia) and transplanted into the / not have a generalized and potent effect on prolifera- cleared fourth mammary fat pad of 3-week-old Rag1 mice. Two cohorts were generated. The first consisted of 32 Rag1 / tion. Our finding that prolactin acts primarily on mice with Prlr / /C3(1)/SV40T donor epithelium and 32 mice neoplastic lesions, and not on subsequent invasive with WT/C3(1)/SV40T donor epithelium. This cohort was lesions, challenges the prevailing assumption that investigated for the development of palpable tumors as prolactin acts primarily during late-stage disease to described above. The second cohort comprised 21 mice with drive invasive tumor growth. Prlr / /C3(1)/SV40T and WT/C3(1)/SV40T epithelium in These results have important implications for the alternate inguinal fat pads. Mammary epithelial transplants treatment of human disease. Agents antagonizing were collected at 8, 22 and 32 weeks for whole-mount analysis prolactin action, such as S179D prolactin and G129R of early neoplastic lesions. prolactin (Goffin et al., 2005) may prove to be useful in preventing the progression of hyperplastic and neoplas- tic lesions to invasive cancer. Improvements in imaging mRNA and protein isolation and diagnostic techniques are currently under develop- Mammary glands from 12-week-old C3(1)/SV40T animals (n ¼ 13) and 8 (n ¼ 4), 22 (n ¼ 2) and 32 week (n ¼ 2) old ment to allow the identification of these early lesions. mammary glands formed from C3(1)/SV40T epithelium as well Prolactin receptor antagonists should be considered as as tumors at ethical end point were collected and frozen in agents for their treatment, both as an adjuvant to liquid nitrogen before storage at 801C before use. Total surgery and hormonal therapy, or as a component of RNA was extracted using TRIZOL Reagent (Gibco BRL) preventative therapy. according to the manufacturer’s instructions.
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 10 Whole-mount histology Protein (20 mg per lane) was separated using sodium Mammary whole-mounts were made using the Carmine alum dodecyl sulfate–polyacrylamide gel electrophoresis (SDS- technique as described before (Bradbury et al., 1995). PAGE) (Bio-Rad Laboratories, CA, USA), transferred to Quantitative analysis of neoplasia and tumor was performed polyvinylidine difluoride (Millipore Corp., MA, USA), and using the public domain NIH Image (developed at the US blocked overnight with 1% skim milk powder, 50 nM sodium National Institutes of Health and available on the Internet at phosphate, 50 mM NaCl, and 0.1% Tween 20. Membranes http://rsb.info.nih.gov/nih-image/). Briefly, the area of neopla- were incubated with a-SV40T (Santa Cruz, CA, USA) and a-b- sia and tumor (distinguished using whole-mount gross actin (Sigma). Specific binding was detected using horseradish morphology) was estimated by tracing the perimeter of each peroxidase-conjugated secondary antibodies (Amersham lesion manually, using photomicrographs of Carmine alum Briosciences, IL, USA) with Chemiluminescence Reagent stained whole-mounts imported into NIH Image software. (PerkinElmer, CT, USA) and Fuji Medical X-ray Film These areas were then converted into a percentage of total (Fujifilm, Tokyo). mammary gland area. Mammary whole-mounts were then peeled off the slides and paraffin embedded. Sections (4 mm) Statistics were cut for routine H&E cytochemistry and immunohisto- The effect of Prlr genotype on the rate of weight gain in whole chemistry. animals was assessed by the coefficients corresponding to the interactions between genotype and time (b5, b6 and b7) in the Quantitative PCR following mixed effects model (Laird and Ware, 1982) using Single-stranded cDNA was produced by reverse transcription the nlme package in R (Pinheiro et al., 2005): using 1 mg RNA in a 20 ml reaction (Promega, WI, USA). Quantitative PCR was performed using LightCycler techno- ¼ b þ þðb þ Þ þ b logy (Roche Diagnostics, Basel, Switzerland). PCR reactions weightij 0 Ui 1 Vi timeij 2genoBi were performed in a 10 ml volume with 1 ml cDNA, 5 pmoles of þ b þ b þ b each primer (SV40T forward CCTGGAATAGTCACCATG, 3genoCi 4genoDi 5timeijgenoBi reverse CAATGCCTGTTTCATGCC, Prlr forward GAGA þ b time genoC þ b time genoB þ error AAAACACCTATGAATGT, reverse GAAGAGCAAGATC 6 ij i 7 ij i ij TCAAGAAC and Keratin-18 forward TGTTCATAGTGGG CACGGATGTC, reverse CAAGATCATCGAAGACCTGA where i indexes animal and j indexes measurement. yij is weight GGGC (as an epithelial house keeping control). FastStart measured in grams, genoA (reference), genoB, genoC and DNA Master SYBR Green I enzyme mix (Roche) as per genoD represent indicator variables for the four genotypes. manufacturer’s instructions. Relative quantitation of the 2 Time is measured as weeks since birth and UiBN(0, sU) and product was performed by comparing the crossing points of V BN (0, s2 ). The effect of genotype on the rate of increase in different samples normalised to Keratin-18. Each cycle in the i V tumor volume was assessed using a similar model where yij is linear phase of the reaction corresponds to a two-fold the cuberoot of volume, time is days since the detection of a difference in transcript levels between samples. Each reaction palpable tumor. The difference between Prlr genotypes in was performed in duplicate. terms of mean SV40T mRNA levels, wholemount analyses, BrdU and cleaved Caspase-3 immuno-cytochemistry and Immunocytochemistry tumor morphology was examined by an unpaired t-statistic Mammary gland sections were baked at 801C for 5 min and (Statview, SAS Institute, NC, USA). The difference between placed in xylene for deparaffinisation. Antigen retrieval was Prlr geneotypes in tumor multiplicity was examined by an performed using target retrieval solution low pH (S1699) unpaired t-statistic (Statview). The effect of Prlr genotype on 20 min water bath (anti-5-bromo-20-dexyuridine (BrdU) clone time until detection of a tumor of a defined size, or attainment Bu20a) and high pH (S2367) 30 s pressure cooker (anticleaved of ethical end point, was determined by Kaplan Meier survival Caspase-3 Asp175, Cell Signaling Technologies, Beverley MA, analysis (logrank statistic) (Statview). The effect of genotype w2 USA). Slides were blocked in 3% H2O2 and a protein block on cellular structure in tumors was determined by a statistic (anticleaved Caspase-3 only) before application of 1:200 anti- (Statview). In all analyses, a Po0.05 corresponded to BrdU primary antibody or 1:100 (anticleaved Caspase-3). statistical significance. Secondary antibody was Envision mouse (anti-BrdU) and Envision rabbit (anticleaved Caspase-3) applied for Acknowledgements 30 min. Visualization was via diaminobenzidine (DAB þ ). All immunocytochemistry reagents were purchased from Dako The Garvan Institute group was supported during the course Cytomation (Botany, Australia) unless otherwise stated. of this work by The National Health and Medical Research Council of Australia, the Cancer Council New South Wales, Western blotting The Cooperative Research Center for Innovative Dairy Following RNA extraction using TRIZOL reagent, protein Products and the United States DoD BCRP (DAMD17-03- was extracted according to the manufacturer’s instructions. 1-0686 to SRO).
References
Attia MA, Weiss DW. (1966). Cancer Res 26: 1787–1800. Bonneterre J, Peyrat JP, Beuscart R, Lefebvre J, Demaille A. Binart N, Helloco C, Ormandy CJ, Barra J, Clement-Lacroix (1987). Cancer Res 47: 4724–4728. P, Baran N et al. (2000). Endocrinology 141: 2691–2697. Bradbury JM, Edwards PA, Niemeyer CC, Dale TC. (1995). Bonneterre J, Mauriac L, Weber B, Roche H, Fargeot P, Dev Biol 170: 553–563. Tubiana-Hulin M et al. (1988). Eur J Cancer Clin Oncol 24: Brisken C, Kaur S, Chavarria TE, Binart N, Sutherland RL, 1851–1853. Weinberg RA et al. (1999). Dev Biol 210: 96–106.
Oncogene Prolactin and mammary carcinogenesis SR Oakes et al 11 Cardiff RD, Anver MR, Gusterson BA, Hennighausen L, Mershon J, Sall W, Mitchner N, Ben-Jonathan N. (1995). Jensen RA, Merino MJ et al. (2000). Oncogene 19: 968–988. Endocrinology 136: 3619–3623. Chen WY, Ramamoorthy P, Chen N, Sticca R, Wagner TE. Missmer SA, Eliassen AH, Barbieri RL, Hankinson SE. (1999). Clin Cancer Res 5: 3583–3593. (2004). J Natl Cancer Inst 96: 1856–1865. Clement-Lacroix P, Ormandy C, Lepescheux L, Ammann P, Mombaerts P, Iacomini J, Johnson RS, Herrup K, Tonegawa Damotte D, Goffin V et al. (1999). Endocrinology 140: S, Papaioannou VE. (1992). Cell 68: 869–877. 96–105. Naylor MJ, Lockefeer JA, Horseman ND, Ormandy CJ. Clevenger CV, Chang WP, Ngo W, Pasha TL, Montone KT, (2003). Endocrine 20: 111–114. Tomaszewski JE. (1995). Am J Pathol 146: 695–705. Naylor MJ, Ormandy CJ. (2002). Dev Dyn 225: 100–105. Clevenger CV, Furth PA, Hankinson SE, Schuler LA. (2003). Ormandy CJ, Binart N, Kelly PA. (1997a). J Mammary Gland Endocrine Rev 24: 1–27. Biol Neoplasia 2: 355–364. Das R, Vonderhaar BK. (1997). J Mammary Gland Biol Ormandy CJ, Camus A, Barra J, Damotte D, Lucas B, Buteau Neoplasia 2: 29–39. H et al. (1997b). Genes Dev 11: 167–178. DeOme KB, Faulkin LJJ, Bern HA, Blair PB. (1959). Cancer Ormandy CJ, Hall RE, Manning DL, Robertson JF, Blamey Res 19: 515–520. RW, Kelly PA et al. (1997c). J Clin Endocrinol Metab 82: Freemark M, Fleenor D, Driscoll P, Binart N, Kelly P. (2001). 3692–3699. Endocrinology 142: 532–537. Peyrat JP, Vennin P, Bonneterre J, Hecquet B, Vandewalle B, Ginsburg E, Vonderhaar BK. (1995). Cancer Res 55: 2591–2595. Kelly PA et al. (1984). Eur J Cancer Clin Oncol 20: Goffin V, Bernichtein S, Touraine P, Kelly PA. (2005). 1363–1367. Endocrine Rev 26: 400–422. Pinheiro J, Bates D, DebRoy S, Sarkar D. (2005). R package Goffin V, Kinet S, Ferrag F, Binart N, Martial JA, Kelly PA. version 3: 1–65. (1996). J Biol Chem 271: 16573–16579. Reynolds C, Montone KT, Powell CM, Tomaszewski JE, Green JE, Shibata MA, Yoshidome K, Liu ML, Jorcyk C, Clevenger CV. (1997). Endocrinology 138: 5555–5560. Anver MR et al. (2000). Oncogene 19: 1020–1027. Robertson FG, Harris J, Naylor MJ, Oakes SR, Kindblom J, Gutzman JH, Miller KK, Schuler LA. (2004). J Steroid Dillner K et al. (2003). Endocrinology 144: 3196–3205. Biochem Mol Biol 88: 69–77. Rose-Hellekant TA, Arendt LM, Schroeder MD, Gilchrist K, Hankinson SE, Willett WC, Manson JE, Colditz GA, Sandgren EP, Schuler LA. (2003). Oncogene 22: 4664–4674. Hunter DJ, Spiegelman D et al. (1998). J Natl Cancer Inst Schroeder MD, Brockman JL, Walker AM, Schuler LA. 90: 1292–1299. (2003). Endocrinology 144: 5300–5307. Hankinson SE, Willett WC, Michaud DS, Manson JE, Schroeder MD, Symowicz J, Schuler LA. (2002). Mol Colditz GA, Longcope C et al. (1999). J Natl Cancer Inst Endocrinol 16: 45–57. 91: 629–634. Shibata MA, Jorcyk CL, Liu ML, Yoshidome K, Gold LG, Horseman ND, Zhao W, Montecino-Rodriguez E, Tanaka M, Green JE. (1998). Toxicol Pathol 26: 177–182. Nakashima K, Engle SJ et al. (1997). EMBO J 16: Touraine P, Martini JF, Zafrani B, Durand JC, Labaille F, 6926–6935. Malet C et al. (1998). J Clin Endocrinol Metab 83: 667–674. Hutchinson JN, Muller WJ. (2000). Oncogene 19: 6130–6137. Tworoger SS, Eliassen AH, Rosner B, Sluss P, Hankinson SE. Laird NM, Ware JH. (1982). Biometrics 38: 963–974. (2004). Cancer Res 64: 6814–6819. Ling C, Hellgren G, Gebre-Medhin M, Dillner K, Wennbo H, Vomachka AJ, Pratt SL, Lockefeer JA, Horseman ND. (2000). Carlsson B et al. (2000). Endocrinology 141: 3564–3572. Oncogene 19: 1077–1084. Llovera M, Pichard C, Bernichtein S, Jeay S, Touraine P, Vonderhaar BK. (1999). Endocrine-Related Cancer 6: 389–404. Kelly PA et al. (2000). Oncogene 19: 4695–4705. Welsch CW, Louks G, Fox D, Brooks C. (1975). Br J Cancer Manni A, Boucher AE, Demers LM, Harvey HA, Lipton A, 32: 427–431. Simmonds MA et al. (1989). Breast Cancer Res Treat 14: Welsch CW, Nagasawa H. (1977). Cancer Res 37: 951–963. 289–298. Wennbo H, Gebre-Medhin M, Gritli-Linde A, Ohlsson C, Maroulakou IG, Anver M, Garrett L, Green JE. (1994). Isaksson OG, Tornell J. (1997). J Clin Invest 100: 2744–2751. Proc Natl Acad Sci USA 91: 11236–11240. Wennbo H, Tornell J. (2000). Oncogene 19: 1072–1076.
Oncogene Appendix VIII
Probe Set Representa Gene Title Gene p p p p FC FC ID tive Public Symbol (Da (Da (Da (Da (Da (Da ID y 4) y 6) y4C y6C y 4) y 6) FP) FP) 1425583_at BC010605 Mammary gland RCB-0526 --- 0.000 0.000 0.000 0.000 -5.59 -4.81 Jyg-MC(A) cDNA, RIKEN full-length enriched library, clone:G830022P11 product:unclassifiable, full insert sequence 1460372_at BC019755 Dual oxidase maturation factor Duox1 0.000 0.000 0.025 0.391 -1.67 -1.52 1 1455301_at BG064092 expressed sequence BQ952480 BQ9524 0.000 0.000 0.034 0.288 1.76 1.71 80 1417089_a NM_0098 creatine kinase, mitochondrial Ckmt1 0.000 0.000 0.005 0.037 1.79 1.88 _at 97 1, ubiquitous 1435998_at BG066504 gene model 288, (NCBI) Gm288 0.000 0.000 0.306 0.422 2.36 2.14 1459253_at AW55659 ------0.000 0.000 0.000 0.000 2.58 3.47 7 1426904_s AV114239 DnaJ (Hsp40) homolog, Dnajc10 0.000 0.000 0.950 0.655 3.70 4.42 _at subfamily C, member 10 1452230_at AV114239 DnaJ (Hsp40) homolog, Dnajc10 0.000 0.000 0.670 0.682 4.17 4.05 subfamily C, member 10 1426905_a AV114239 DnaJ (Hsp40) homolog, Dnajc10 0.000 0.000 0.500 0.273 4.21 4.88 _at subfamily C, member 10 1453089_at AA259365 RIKEN cDNA 3110079O15 3110079 0.000 0.000 0.054 0.423 5.31 4.85 gene O15Rik 1425500_x BC003207 Hypothetical protein LOC625 0.000 0.001 0.000 0.000 -4.42 -2.99 _at LOC625794 794 1425584_x BC010605 Mammary gland RCB-0526 --- 0.000 0.002 0.000 0.000 -4.88 -3.56 _at Jyg-MC(A) cDNA, RIKEN full-length enriched library, clone:G830022P11 product:unclassifiable, full insert sequence 1452565_x M11024 hypothetical protein LOC641 0.000 0.004 0.000 0.000 -4.41 -2.40 _at LOC641050 050 1425502_x BC003207 Mammary gland RCB-0526 --- 0.000 0.006 0.000 0.000 -4.26 -2.58 _at Jyg-MC(A) cDNA, RIKEN full-length enriched library, clone:G830022P11 product:unclassifiable, full insert sequence 1438278_a BB532946 cDNA sequence BC003993 BC0039 0.000 0.022 0.944 0.456 1.68 1.35 _at 93 1451851_a BC004601 casein kappa Csnk 0.000 0.078 0.000 0.005 -4.98 -1.93 _at 1418684_at NM_0271 RIKEN cDNA 2310012P17 2310012 0.000 0.110 0.270 0.203 -1.59 -1.18 42 gene P17Rik 1440320_at BB149299 ------0.000 0.468 0.004 0.343 -1.34 1.05 1423519_at BE457744 RIKEN cDNA 2210412D01 2210412 0.000 0.526 0.000 0.007 -1.49 -1.06 gene D01Rik
297 1423414_at BB520073 prostaglandin-endoperoxide Ptgs1 0.000 0.364 0.699 0.585 -1.46 1.08 synthase 1 1426851_a X96585 nephroblastoma overexpressed Nov 0.000 0.001 0.286 0.774 -2.55 -1.78 _at gene 1426852_x X96585 nephroblastoma overexpressed Nov 0.000 0.006 0.251 0.828 -2.49 -1.61 _at gene 1460194_at NM_0107 phytanoyl-CoA hydroxylase Phyh 0.000 0.063 0.866 0.083 1.42 1.15 26 1425469_a BC003855 RIKEN cDNA 9030208C03 9030208 0.000 0.004 0.000 0.000 -3.73 -2.25 _at gene C03Rik 1417257_at BC006872 carboxyl ester lipase Cel 0.000 0.000 0.000 0.000 -3.70 -3.51 1417569_at BG071381 neurocalcin delta Ncald 0.000 0.000 0.203 0.067 1.62 1.58 1426031_a AF289078 nuclear factor of activated T- Nfatc2 0.000 0.444 0.070 0.067 -1.50 -1.07 _at cells, cytoplasmic, calcineurin- dependent 2 1420338_at L34570 arachidonate 15-lipoxygenase Alox15 0.000 0.065 0.042 0.781 2.02 1.26 1419618_at NM_1304 butyrobetaine (gamma), 2- Bbox1 0.000 0.002 0.006 0.476 2.19 1.70 52 oxoglutarate dioxygenase 1 (gamma-butyrobetaine hydroxylase) 1426039_a U39200 arachidonate lipoxygenase, Alox12e 0.000 0.014 0.871 0.776 7.26 2.18 _at epidermal 1456877_at BB147678 dual oxidase 1 /// similar to Duox1 0.000 0.003 0.064 0.104 -1.57 -1.42 dual oxidase 1 /// LOC640 578 1424890_at U88064 basonuclin 1 Bnc1 0.000 0.001 0.509 0.937 2.91 1.95 1419556_at BC012424 E74-like factor 5 Elf5 0.000 0.481 0.000 0.000 -3.26 -1.20 1419555_at BC012424 E74-like factor 5 Elf5 0.000 0.826 0.000 0.000 -3.68 -1.06 1439382_x BB378700 discoidin domain receptor Ddr1 0.000 0.033 0.321 0.906 1.42 1.18 _at family, member 1 1417616_at BC010208 ST6 (alpha-N-acetyl- St6galna 0.000 0.000 0.000 0.019 1.73 2.53 neuraminyl-2,3-beta- c2 galactosyl-1,3)-N- acetylgalactosaminide alpha- 2,6-sialyltransferase 2 1449005_at NM_0306 solute carrier family 16 Slc16a3 0.000 0.297 0.535 0.018 -1.38 -1.08 96 (monocarboxylic acid transporters), member 3 1449203_at NM_1308 solute carrier organic anion Slco1a5 0.000 0.000 0.009 0.130 3.02 3.61 61 transporter family, member 1a5 1422178_a NM_0089 RAB17, member RAS Rab17 0.000 0.000 0.001 0.011 2.64 3.15 _at 98 oncogene family 1423186_at BM22895 T-cell lymphoma invasion and Tiam2 0.000 0.208 0.461 0.784 1.57 1.14 7 metastasis 2 1427168_a AJ131395 procollagen, type XIV, alpha 1 Col14a1 0.000 0.273 0.512 0.063 -1.77 -1.16 _at 1428455_at BB521934 procollagen, type XIV, alpha 1 Col14a1 0.000 0.318 0.480 0.172 -1.68 -1.12 1457666_s AV229143 interferon activated gene 202B Ifi202b 0.000 0.106 0.091 0.015 4.46 1.66 _at 1427608_a X03802 T-cell receptor gamma, Tcrg-C 0.000 0.438 0.259 0.769 1.39 -1.05 _at constant region 1426153_a BC020144 desmoglein 2 Dsg2 0.000 0.000 0.055 0.257 1.49 1.49 _at 1447551_x BB274232 latrophilin 3 Lphn3 0.000 0.036 0.085 0.965 2.12 1.40 298 _at 1450197_at NM_0299 retinal pigment epithelium 65 Rpe65 0.000 0.000 0.626 0.381 1.91 2.08 87 1451932_a BC016215 ADAMTS-like 4 Adamtsl 0.001 0.002 0.021 0.602 -1.49 -1.42 _at 4 1425312_s BC016190 DNA segment, Chr 11, D11Ertd 0.001 0.002 0.267 0.042 1.45 1.42 _at ERATO Doi 636, expressed 636e 1440339_at AW91050 RIKEN cDNA 4833416E15 4833416 0.001 0.002 0.986 0.549 1.71 1.61 4 gene E15Rik 1425163_at BC006605 expressed sequence AI661453 AI66145 0.001 0.003 0.001 0.003 1.54 1.48 3 1424413_at BE650508 opioid growth factor receptor- Ogfrl1 0.001 0.005 0.000 0.000 -1.94 -1.75 like 1 1446342_at BB137728 RIKEN cDNA 2310001H17 2310001 0.001 0.014 0.067 0.265 1.46 1.29 gene H17Rik 1428391_at AK004767 RAB3A interacting protein Rab3il1 0.001 0.018 0.017 0.052 -1.29 -1.19 (rabin3)-like 1 1440475_at BB740266 expressed sequence AW0117 0.001 0.023 0.861 0.025 1.37 1.22 AW011738 38 1453321_at AK003938 fibronectin type III domain Fndc1 0.001 0.027 0.340 0.975 -1.86 -1.43 containing 1 1425471_x BC003855 ------0.001 0.032 0.000 0.000 -3.56 -2.18 _at 1451610_at BC024561 cDNA sequence BC024561 BC0245 0.001 0.033 0.032 0.313 2.33 1.61 61 1424289_at BB817847 cDNA sequence BC010311 BC0103 0.001 0.089 0.984 0.365 1.68 1.27 11 1445079_at BG066842 DNA segment, Chr 15, D15Ertd 0.001 0.113 0.640 0.565 -1.28 -1.11 ERATO Doi 320, expressed 320e 1427817_at M11024 hypothetical protein LOC641 0.001 0.114 0.002 0.398 -1.90 -1.29 LOC641050 050 1437885_at BF682730 RIKEN cDNA D030029J20 D030029 0.001 0.123 0.070 0.073 1.34 1.12 gene J20Rik 1456881_at BF019797 neurobeachin-like 2 Nbeal2 0.001 0.153 0.876 0.003 -1.28 1.10 1432442_at AK015606 RIKEN cDNA 4930481F22 4930481 0.001 0.203 0.271 0.599 -1.30 -1.08 gene F22Rik 1420411_a NM_0287 phosphatidylinositol 4-kinase Pi4k2b 0.001 0.220 0.471 0.224 1.61 -1.17 _at 44 type 2 beta 1435642_at BB081437 RIKEN cDNA 9630019K15 9630019 0.001 0.804 0.271 0.000 -1.33 -1.02 gene K15Rik 1431425_a AK015972 RIKEN cDNA 4930535B03 4930535 0.001 0.888 0.961 0.319 1.34 -1.01 _at gene B03Rik 1435595_at AV016374 RIKEN cDNA 1810011O10 1810011 0.001 0.938 0.033 0.269 1.49 1.01 gene O10Rik Appendix VIII. Differentially expressed genes in Elf5-/- epithelium at day 4 of pregnancy. The top 69 differentially expressed genes in Elf5-/- epithelium at day 4 of pregnancy. Affymetrix ID, Genebank ID, gene title and gene symbol is listed. The unadjusted p-value for the change in gene expression between Elf5-/- and WT at day 4 (p Day 4) and at day 6 (p Day 6) and the difference between cleared fat pad (CFP) and wildtype at day 4 (p Day4CFP) and at day 6 (p Day6CFP) is tabulated. The fold change at day 6 (FC Day 4) and at day 6 (FC Day 6) is also shown. Negative and positive values indicate decreasing and positive genes respectively.
299 Appendix IX
Probe Set Representa Gene Title Gene p p p p FC FC ID tive Public Symbol (Da (Da (Da (Da (Da (Da ID y 4) y 6) y4C y6C y 4) y 6) FP) FP) 1425583_at BC010605 Mammary gland RCB-0526 --- 0.000 0.000 0.000 0.000 -5.59 -4.81 Jyg-MC(A) cDNA, RIKEN full-length enriched library, clone:G830022P11 product:unclassifiable, full insert sequence 1460372_at BC019755 Dual oxidase maturation Duox1 0.000 0.000 0.025 0.391 -1.67 -1.52 factor 1 1455301_at BG064092 expressed sequence BQ9524 0.000 0.000 0.034 0.288 1.76 1.71 BQ952480 80 1417089_a NM_0098 creatine kinase, Ckmt1 0.000 0.000 0.005 0.037 1.79 1.88 _at 97 mitochondrial 1, ubiquitous 1435998_at BG066504 gene model 288, (NCBI) Gm288 0.000 0.000 0.306 0.422 2.36 2.14 1459253_at AW55659 ------0.000 0.000 0.000 0.000 2.58 3.47 7 1452230_at AV114239 DnaJ (Hsp40) homolog, Dnajc10 0.000 0.000 0.670 0.682 4.17 4.05 subfamily C, member 10 1426904_s AV114239 DnaJ (Hsp40) homolog, Dnajc10 0.000 0.000 0.950 0.655 3.70 4.42 _at subfamily C, member 10 1453089_at AA259365 RIKEN cDNA 3110079O15 3110079 0.000 0.000 0.054 0.423 5.31 4.85 gene O15Rik 1426905_a AV114239 DnaJ (Hsp40) homolog, Dnajc10 0.000 0.000 0.500 0.273 4.21 4.88 _at subfamily C, member 10 1449347_a NM_0213 X-linked lymphocyte- Xlr4b /// 0.002 0.000 0.875 0.863 2.62 3.39 _at 65 regulated 4B /// X-linked Xlr4a /// lymphocyte-regulated 4A /// Xlr4e X-linked lymphocyte- regulated 4E 1439231_at BG228852 ------0.002 0.000 0.655 0.575 2.52 3.78 1420357_s NM_0117 X-linked lymphocyte- Xlr3a /// 0.003 0.000 0.570 0.034 2.47 5.27 _at 26 regulated 3A /// X-linked Xlr3b /// lymphocyte-regulated 3B /// MGC766 hypothetical protein 89 LOC574437 1451287_s BC024599 RIKEN cDNA 2810003C17 2810003 0.004 0.000 0.000 0.134 -1.35 -1.47 _at gene C17Rik 1441213_at AV370141 ------0.004 0.000 0.002 0.207 1.80 2.34 1455216_at BI408433 progestin and adipoQ Paqr6 0.005 0.000 0.052 0.177 1.36 1.51 receptor family member VI 1426550_at AI449329 SID1 transmembrane family, Sidt1 0.009 0.000 0.014 0.079 1.61 2.89 member 1 1445510_at C79601 expressed sequence C79601 C79601 0.011 0.000 0.182 0.645 1.42 1.84 1456481_at BB525078 DNA segment, Chr 9, D9Ertd2 0.014 0.000 0.000 0.001 1.43 1.80 ERATO Doi 280, expressed 80e 1424607_a BM22525 similar to hypothetical LOC432 0.017 0.000 0.270 0.075 1.55 2.05 _at 5 protein MGC37588 823 1449281_at NM_0087 neurturin Nrtn 0.024 0.000 0.004 0.290 1.28 1.60 38
300 1439813_at BB232083 Signal peptide peptidase 3 Sppl3 0.026 0.000 0.171 0.115 -1.53 2.25 1459568_at BG076403 Transmembrane and Tmtc4 0.036 0.000 0.288 0.533 1.23 1.65 tetratricopeptide repeat containing 4 1452861_at AK008485 RIKEN cDNA 2010300C02 2010300 0.039 0.000 0.000 0.006 1.60 2.51 gene /// similar to C02Rik Y59A8B.19 /// LOC639 555 1437136_at AV316469 RIKEN cDNA 5830436I19 5830436I 0.047 0.000 0.132 0.231 -1.25 1.83 gene 19Rik 1424623_at AI462524 serine (or cysteine) peptidase Serpinb5 0.049 0.000 0.000 0.003 1.71 3.02 inhibitor, clade B, member 5 1431053_at AI536236 M-phase phosphoprotein 9 Mphosph 0.071 0.000 0.111 0.022 1.13 1.51 9 1460133_at BF607205 ------0.076 0.000 0.021 0.622 1.19 1.55 1459713_s AU040576 transmembrane protein 16A Tmem16 0.076 0.000 0.000 0.000 1.40 2.47 _at a 1442277_at BB546429 choline kinase alpha Chka 0.080 0.000 0.146 0.346 -1.25 1.70 1430965_at BB098581 RIKEN cDNA 9430064K01 9430064 0.084 0.000 0.003 0.993 -1.21 1.60 gene K01Rik 1456878_at BE910952 expressed sequence AI64602 0.089 0.000 0.000 0.000 1.44 2.63 AI646023 3 1438062_at BM93398 RIKEN cDNA 4832420A03 4832420 0.092 0.000 0.971 0.032 -1.19 1.59 4 gene A03Rik 1427261_at BQ176786 WW, C2 and coiled-coil Wwc1 0.117 0.000 0.011 0.132 1.27 2.09 domain containing 1 1418213_at NM_0333 keratin complex 1, acidic, Krt1-23 0.124 0.000 0.014 0.417 1.46 4.24 73 gene 23 1439252_at AV301185 ------0.151 0.000 0.912 0.213 1.20 1.72 1445128_at BQ176646 CDNA clone --- 0.158 0.000 0.002 0.788 1.15 1.62 IMAGE:30463708 1428732_at AK005774 RIKEN cDNA 1700008J07 1700008 0.159 0.000 0.044 0.515 -1.12 -1.44 gene J07Rik 1441516_a BB527707 RIKEN cDNA C130050O18 C130050 0.169 0.000 0.760 0.019 -1.14 -1.50 _at gene O18Rik 1459078_at BG069377 RNA binding protein gene Rbpms 0.169 0.000 0.001 0.989 1.22 2.15 with multiple splicing 1429401_at BM11439 storkhead box 2 Stox2 0.175 0.000 0.841 0.252 1.19 1.99 8 1459495_at BG075346 Transmembrane protein 16A Tmem16 0.175 0.000 0.000 0.000 1.19 2.77 a 1446147_at BB436856 expressed sequence C79248 C79248 0.193 0.000 0.029 0.157 1.28 2.72 1441177_at AW55207 ------0.198 0.000 0.024 0.098 -1.23 2.50 6 1443230_at BB340127 RIKEN cDNA 5330421F07 5330421 0.199 0.000 0.042 0.667 1.31 2.46 gene F07Rik 1443104_at BB541236 ------0.211 0.000 0.178 0.236 -1.14 1.68 1446148_x BB436856 expressed sequence C79248 C79248 0.219 0.000 0.035 0.262 1.25 2.18 _at 1437872_at BB325565 cDNA sequence AB112350 AB1123 0.225 0.000 0.006 0.008 1.13 -1.56 50 1460328_at BG072367 bromodomain containing 3 Brd3 0.247 0.000 0.276 0.113 -1.16 1.78 1427247_at BB238462 DNA segment, Chr 3, D3Bwg0 0.264 0.000 0.125 0.694 1.18 1.99 Brigham & Women's 562e 301 Genetics 0562 expressed 1453840_at AK005009 poly A binding protein, Pabpc1 0.273 0.000 0.550 0.834 -1.15 1.72 cytoplasmic 1 1441112_at BB394889 similar to Zinc finger BED LOC667 0.281 0.000 0.993 0.350 -1.24 2.58 domain containing protein 4 118 /// /// similar to Zinc finger BED LOC669 domain containing protein 4 797 /// /// similar to Zinc finger BED LOC675 domain containing protein 4 760 1453264_at AK007346 MARVEL (membrane- Marveld 0.281 0.000 0.000 0.000 1.29 2.78 associating) domain 3 containing 3 1445302_at BG070512 immunoglobulin superfamily, Igsf6 0.290 0.000 0.326 0.020 -1.10 1.58 member 6 1453214_at AK017350 leucine rich repeat containing Lrrc15 0.294 0.000 0.507 0.108 -1.31 3.01 15 1452855_at AK010485 lymphocyte antigen 6 Ly6k 0.311 0.000 0.383 0.001 -1.15 -2.03 complex, locus K 1416178_a NM_0137 pleckstrin homology domain Plekhb1 0.328 0.000 0.000 0.352 1.20 2.20 _at 46 containing, family B (evectins) member 1 1457161_at BB111383 RIKEN cDNA 9530029O12 9530029 0.335 0.000 0.299 0.705 -1.17 2.10 gene O12Rik 1448007_at BE847301 bromodomain adjacent to Baz2b 0.336 0.000 0.965 0.035 1.15 2.00 zinc finger domain, 2B 1428947_at AK008016 RIKEN cDNA 2010001M09 2010001 0.340 0.000 0.579 0.001 1.07 -1.33 gene M09Rik 1430392_at AK020657 RIKEN cDNA 9530086O07 9530086 0.348 0.000 0.886 0.062 -1.22 3.22 gene O07Rik 1438277_at BE948602 RIKEN cDNA E130308A19 E130308 0.355 0.000 0.595 0.021 -1.10 1.60 gene A19Rik 1433870_at AV288135 ATPase family, AAA domain Atad4 0.362 0.000 0.130 0.144 1.16 2.42 containing 4 1454254_s AK002767 RIKEN cDNA 1600029D21 1600029 0.366 0.000 0.000 0.000 1.25 3.32 _at gene D21Rik 1445918_at BB035414 Transmembrane protein 2 Tmem2 0.367 0.000 0.004 0.036 1.14 2.09 1435572_at AV232798 RIKEN cDNA 2310014L17 2310014 0.372 0.000 0.010 0.140 1.08 1.47 gene L17Rik 1445329_at BB450122 dystrobrevin, beta Dtnb 0.443 0.000 0.022 0.621 1.11 2.01 1431476_at AV279437 RIKEN cDNA 4933407I05 4933407I 0.452 0.000 0.399 0.828 -1.05 -1.31 gene 05Rik 1459464_at BG069189 Transcribed locus --- 0.453 0.000 0.731 0.036 -1.15 2.34 1439357_at BB291414 interleukin 17 receptor E Il17re 0.460 0.000 0.022 0.112 -1.10 1.80 1437923_at BG070544 expressed sequence AI31476 0.464 0.000 0.836 0.229 -1.15 2.22 AI314760 0 1429410_at AI595744 enhancer of yellow 2 Eny2 0.470 0.000 0.002 0.493 1.08 1.70 homolog (Drosophila) 1459469_at C78516 expressed sequence C78516 C78516 0.470 0.000 0.698 0.123 -1.11 2.11 1445938_at BB037457 RIKEN cDNA 5930427L02 5930427 0.475 0.000 0.011 0.280 1.16 2.80 gene L02Rik 1436061_at BG092263 ------0.476 0.000 0.485 0.330 -1.09 -1.73 1443201_at BM11907 Glypican 6 Gpc6 0.481 0.000 0.286 0.187 -1.14 2.19 8 1451458_at BC019745 transmembrane protein 2 Tmem2 0.502 0.000 0.001 0.012 1.14 2.30 1440611_at BB387793 RIKEN cDNA 6230409E13 6230409 0.504 0.000 0.084 0.555 1.07 1.76 302 gene E13Rik 1457637_at BE992966 Prickle like 1 (Drosophila) Prickle1 0.504 0.000 0.234 0.344 1.19 3.10 1443077_at BB526689 RIKEN cDNA 1700081L11 1700081 0.515 0.000 0.040 0.965 1.08 1.83 gene L11Rik 1448793_a BC005679 syndecan 4 Sdc4 0.528 0.000 0.000 0.000 1.20 4.43 _at 1450505_a NM_0254 RIKEN cDNA 1810015C04 1810015 0.556 0.000 0.000 0.722 1.07 1.82 _at 59 gene C04Rik 1435514_at BB700884 leucine zipper transcription Lztfl1 0.561 0.000 0.012 0.579 1.07 -1.98 factor-like 1 1450541_at NM_0112 plasmacytoma variant Pvt1 0.573 0.000 0.006 0.018 1.08 1.89 22 translocation 1 1436535_at BQ176653 TROVE domain family, Trove2 0.575 0.000 0.113 0.166 1.12 2.59 member 2 1437757_at BB402190 ------0.585 0.000 0.122 0.079 -1.07 1.77 1443286_at BG069560 expressed sequence AU0197 0.588 0.000 0.185 0.064 1.06 1.77 AU019754 54 1423072_at AW54992 RIKEN cDNA 6720475J19 6720475 0.642 0.000 0.004 0.811 1.12 5.26 8 gene /// similar to putative J19Rik retrovirus-related gag protein /// LOC670 480 1429106_at AK014853 RIKEN cDNA 4921509J17 4921509 0.651 0.000 0.876 0.033 -1.04 -1.58 gene J17Rik 1426361_at AV328883 zinc finger CCCH type Zc3h11a 0.668 0.000 0.069 0.584 -1.05 1.68 containing 11A 1440986_at BM19782 RNA polymerase II Rpap1 0.671 0.000 0.091 0.588 -1.06 1.80 3 associated protein 1 1453582_at BB129366 choline kinase alpha Chka 0.681 0.000 0.658 0.401 -1.06 2.12 1423825_at BC018381 G protein-coupled receptor Gpr177 0.689 0.000 0.123 0.246 -1.04 1.61 177 1445559_at BB762499 Retinitis pigmentosa 9 Rp9h 0.689 0.000 0.620 0.855 1.05 1.81 homolog (human) 1437291_at AV131166 RIKEN cDNA 2700081O15 2700081 0.708 0.000 0.000 0.152 1.02 -1.34 gene O15Rik 1453448_at AK010086 RIKEN cDNA 2310067E19 2310067 0.801 0.000 0.358 0.896 1.06 2.57 gene E19Rik 1438807_at BB251000 Heterogeneous nuclear Hnrpr 0.803 0.000 0.933 0.459 -1.05 2.71 ribonucleoprotein R 1447510_at AI467086 RIKEN cDNA C530014P21 C530014 0.807 0.000 0.906 0.399 1.04 1.99 gene P21Rik 1458943_at BB147698 ------0.821 0.000 0.523 0.953 1.04 2.13 1429093_at BI110088 DNA-damage inducible Ddi2 0.871 0.000 0.229 0.318 1.03 2.95 protein 2 1441460_at BB435465 FGFR1 oncogene partner 2 Fgfr1op2 0.889 0.000 0.809 0.744 1.02 1.98 1417029_a BB283676 tripartite motif protein 2 Trim2 0.903 0.000 0.127 0.280 1.02 1.88 _at 1444140_at BB468014 ------0.923 0.000 0.349 0.102 -1.02 2.54 1452837_at AK021389 lipin 2 Lpin2 0.925 0.000 0.929 0.293 -1.01 2.46 1443107_at BB050233 Transmembrane protein 16F Tmem16 0.932 0.000 0.633 0.038 -1.01 1.89 f 1457294_at BM19502 Transmembrane and Tmtc3 0.933 0.000 0.837 0.074 1.01 1.61 1 tetratricopeptide repeat containing 3
303 1441403_at AV339054 RIKEN cDNA 6430501K19 6430501 0.959 0.000 0.810 0.057 -1.01 1.80 gene K19Rik 1418091_at NM_0237 transcription factor CP2-like Tcfcp2l1 0.001 0.000 0.000 0.189 1.90 2.79 55 1 1454120_a AK016709 polycomb group ring finger 6 Pcgf6 0.401 0.000 0.169 0.365 -1.11 1.69 _at 1459470_at AW54641 Runt related transcription Runx1 0.709 0.000 0.000 0.865 -1.05 1.74 2 factor 1 1449880_s NM_0075 bone gamma- Bglap- 0.002 0.000 0.221 0.198 3.10 4.03 _at 41 carboxyglutamate protein, rs1 /// related sequence 1 /// bone Bglap1 gamma carboxyglutamate /// protein 1 /// bone gamma- Bglap2 carboxyglutamate protein 2 1419485_at BB759833 forkhead box C1 Foxc1 0.078 0.000 0.005 0.019 1.40 2.61 1440699_at BB189640 microtubule-associated Mtap2 0.553 0.000 0.003 0.090 1.13 2.51 protein 2 1417257_at BC006872 carboxyl ester lipase Cel 0.000 0.000 0.000 0.000 -3.70 -3.51 1417569_at BG071381 neurocalcin delta Ncald 0.000 0.000 0.203 0.067 1.62 1.58 1435272_at BE196957 inositol 1,4,5-trisphosphate 3- Itpkb 0.317 0.000 0.040 0.600 -1.07 -1.38 kinase B 1440037_at BB480970 pre B-cell leukemia Pbx1 0.757 0.000 0.837 0.366 1.05 2.02 transcription factor 1 1458140_at BB547877 slit homolog 2 (Drosophila) Slit2 0.774 0.000 0.002 0.165 1.07 3.43 1426587_a AI325183 signal transducer and Stat3 0.602 0.000 0.397 0.176 -1.06 1.64 _at activator of transcription 3 1416835_s NM_0096 S-adenosylmethionine Amd1 /// 0.150 0.000 0.064 0.711 1.26 2.48 _at 65 decarboxylase 1 /// S- Amd2 adenosylmethionine decarboxylase 2 1424850_at L13103 mitogen activated protein Map3k1 0.012 0.000 0.960 0.257 1.31 1.52 kinase kinase kinase 1 1437347_at BF100813 endothelin receptor type B Ednrb 0.027 0.000 0.150 0.932 -1.28 -1.62 1456778_at BB409477 Neuropilin 2 Nrp2 0.149 0.000 0.252 0.862 -1.13 -1.44 1435349_at BB752129 neuropilin 2 Nrp2 0.914 0.000 0.107 0.681 1.01 -1.51 1457082_at BM50844 forkhead box P3 Foxp3 0.690 0.000 0.672 0.370 -1.03 -1.42 5 1445798_at BB297161 discs, large homolog 1 Dlgh1 0.226 0.000 0.060 0.382 1.08 1.36 (Drosophila) 1450534_x M58156 MHC (A.CA/J(H-2K-f) class LOC566 0.282 0.000 0.381 0.339 -1.10 -1.61 _at I antigen 28 1418153_at NM_0084 laminin, alpha 1 Lama1 0.198 0.000 0.000 0.010 1.44 3.53 80 1441669_at BB131936 centaurin, beta 2 Centb2 0.138 0.000 0.457 0.074 -1.25 1.86 1442295_at BM24422 Actin related protein 2/3 Arpc2 0.537 0.000 0.022 0.055 1.11 2.00 9 complex, subunit 2 1429222_at AI661697 ------0.682 0.000 0.238 0.479 1.05 1.71 1425558_at BC017147 kinesin light chain 3 Klc3 0.205 0.000 0.000 0.000 1.28 2.55 1460681_at BC024320 CEA-related cell adhesion Ceacam1 0.003 0.000 0.007 0.090 1.45 1.66 molecule 1 1460682_s BC024320 CEA-related cell adhesion Ceacam1 0.015 0.000 0.001 0.007 1.49 2.04 _at molecule 1 /// CEA-related /// cell adhesion molecule 2 Ceacam2 1425675_s M77196 CEA-related cell adhesion Ceacam1 0.045 0.000 0.005 0.046 1.56 2.55 _at molecule 1 304 1441119_at BM24407 GTPase activating RANGAP Garnl1 0.617 0.000 0.000 0.050 -1.04 1.48 4 domain-like 1 1444943_at C77776 SH3 domain containing ring Sh3rf1 0.243 0.000 0.618 0.209 -1.20 2.01 finger 1 1442511_at BG069408 importin 7 Ipo7 0.110 0.000 0.397 0.016 -1.25 1.89 1446892_at AU041374 Leucine rich repeat Lrrc16 0.154 0.000 0.002 0.004 1.20 1.71 containing 16 1445681_at BE852160 Cell division cycle associated Cdca7 0.190 0.000 0.000 0.304 -1.12 1.77 7 1419031_at NM_0196 fatty acid desaturase 2 Fads2 0.072 0.000 0.132 0.876 1.29 2.31 99 1449325_at NM_0196 fatty acid desaturase 2 Fads2 0.489 0.000 0.214 0.347 1.13 2.14 99 1450607_s NM_0312 urocortin 2 Ucn2 0.474 0.000 0.248 0.660 -1.06 1.48 _at 50 1459311_at BB255824 phosphodiesterase 4D, cAMP Pde4d 0.146 0.000 0.000 0.000 1.30 2.59 specific 1442771_at BB799477 RAD51-like 1 (S. cerevisiae) Rad51l1 0.978 0.000 0.003 0.549 -1.00 1.45 1456260_at AV251099 retinoblastoma binding Rbbp4 0.287 0.000 0.323 0.047 -1.27 3.21 protein 4 1450661_x NM_0086 nuclear factor I/C Nfic 0.918 0.000 0.044 0.004 1.01 1.69 _at 88 1439477_at BB047737 ubiquitin-conjugating enzyme Ube2b 0.646 0.000 0.730 0.059 -1.06 1.86 E2B, RAD6 homology (S. cerevisiae) 1447071_at BM21890 Transcription factor 7-like 2, Tcf7l2 0.512 0.000 0.691 0.265 -1.20 4.62 8 T-cell specific, HMG-box 1435670_at AV237028 transcription factor AP-2 beta Tcfap2b 0.068 0.000 0.000 0.020 1.57 2.73 1446529_at BB420672 RIKEN cDNA 1200003I07 1200003I 0.074 0.000 0.030 0.443 -1.23 2.09 gene 07Rik 1418301_at NM_0168 interferon regulatory factor 6 Irf6 0.100 0.000 0.000 0.000 1.49 3.11 51 1459393_at BB438847 Transcription factor AP-2 Tcfap2b 0.154 0.000 0.007 0.311 -1.23 2.28 beta 1426556_at BG066777 suppressor of hairy wing Suhw4 0.314 0.000 0.004 0.618 1.08 1.53 homolog 4 (Drosophila) 1423340_at AV334599 transcription factor AP-2 beta Tcfap2b 0.442 0.000 0.000 0.070 1.29 6.08 1416586_at NM_0086 zinc finger protein 239 Zfp239 0.857 0.000 0.007 0.593 1.02 1.51 16 1438508_at BB224228 Jumonji domain containing Jmjd1b 0.996 0.000 0.758 0.068 -1.00 2.38 1B 1437545_at BM19499 REST corepressor 1 Rcor1 0.249 0.000 0.021 0.067 -1.27 2.40 4 1451689_a BC018551 SRY-box containing gene 10 Sox10 0.334 0.000 0.000 0.000 1.15 1.89 _at 1425369_a BC018551 SRY-box containing gene 10 Sox10 0.998 0.000 0.023 0.207 -1.00 1.53 _at 1419157_at AI428101 SRY-box containing gene 4 Sox4 /// 0.127 0.000 0.159 0.799 -1.66 5.84 /// similar to Transcription LOC672 factor SOX-4 274 1425400_a BC025116 Cbp/p300-interacting Cited4 0.008 0.000 0.000 0.020 1.80 3.22 _at transactivator, with Glu/Asp- rich carboxy-terminal domain, 4 1449530_at NM_0320 trichorhinophalangeal Trps1 0.285 0.000 0.001 0.001 -1.26 2.96 305 00 syndrome I (human) 1438214_at BB483226 trichorhinophalangeal Trps1 0.303 0.000 0.007 0.017 -1.27 3.47 syndrome I (human) 1456983_at BB360644 Trichorhinophalangeal Trps1 0.870 0.000 0.000 0.000 -1.03 2.54 syndrome I (human) 1443161_at BM24064 Trichorhinophalangeal Trps1 0.990 0.000 0.000 0.000 1.00 2.30 8 syndrome I (human) 1447167_at BB713397 Expressed sequence AL03331 0.527 0.000 0.007 0.228 -1.08 1.70 AL033314 4 1419603_at NM_0083 interferon activated gene 204 Ifi204 /// 0.875 0.000 0.029 0.602 -1.02 -1.63 29 /// similar to Interferon- LOC672 activatable protein 204 (Ifi- 547 204) (Interferon-inducible protein p204) 1426114_at L36663 heterogeneous nuclear Hnrpab 0.588 0.000 0.004 0.065 -1.06 1.60 ribonucleoprotein A/B 1424127_at BC003755 eyes absent 2 homolog Eya2 0.009 0.000 0.000 0.019 1.60 2.44 (Drosophila) 1437543_at BB488001 far upstream element (FUSE) Fubp1 0.134 0.000 0.693 0.399 -1.29 2.05 binding protein 1 1456229_at BG073383 homeo box B3 Hoxb3 0.003 0.000 0.082 0.405 1.42 1.71 1439885_at AA051236 homeo box C5 Hoxc5 0.014 0.000 0.019 0.296 -1.33 -1.57 1445757_at BB045423 T-box 3 Tbx3 0.003 0.000 0.102 0.415 1.38 1.55 1449458_at BC007475 forkhead box I1 Foxi1 0.011 0.000 0.003 0.027 2.62 5.58 1448977_at BC003778 transcription factor AP-2, Tcfap2c 0.133 0.000 0.002 0.021 1.40 2.58 gamma 1419475_a BC008249 ets homologous factor Ehf 0.087 0.000 0.004 0.287 1.53 2.88 _at 1451375_at BC006789 ets homologous factor Ehf 0.126 0.000 0.000 0.000 1.63 4.29 1419474_a BC008249 ets homologous factor Ehf 0.141 0.000 0.107 0.118 1.43 4.07 _at 1447360_at AW41316 TSC22 domain family, Tsc22d1 0.060 0.000 0.017 0.656 -1.43 2.18 9 member 1 1427143_at BC019446 jumonji, AT rich interactive Jarid1b 0.385 0.000 0.998 0.605 1.14 2.05 domain 1B (Rbp2 like) 1453290_at AA165746 high mobility group box 2- Hmgb2l1 0.445 0.000 0.338 0.044 -1.10 1.73 like 1 1421072_at NM_0188 Iroquois related homeobox 5 Irx5 0.128 0.000 0.001 0.003 1.43 3.06 26 (Drosophila) 1456258_at BG072869 empty spiracles homolog 2 Emx2 0.204 0.000 0.075 0.443 -1.16 -1.70 (Drosophila) 1438184_a BB017776 ankyrin repeat domain 5 Ankrd5 0.840 0.000 0.521 0.784 1.02 -1.66 _at 1436512_at BI964400 ADP-ribosylation factor-like Arl4c /// 0.009 0.000 0.619 0.489 1.40 1.70 4C /// similar to ADP- LOC632 ribosylation factor-like 433 protein 7 1441414_at BG075856 Eukaryotic translation Eef2 0.064 0.000 0.406 0.012 -1.22 1.63 elongation factor 2 1444855_at AW55353 B-cell CLL/lymphoma 9-like Bcl9l 0.638 0.000 0.280 0.618 1.07 1.88 7 1442384_at BG065559 G elongation factor, Gfm2 0.798 0.000 0.078 0.370 -1.02 -1.40 mitochondrial 2 1443235_at BM24000 Eukaryotic translation Eif2ak4 0.640 0.000 0.152 0.138 -1.07 1.98 5 initiation factor 2 alpha 306 kinase 4 1440551_at BM24086 DnaJ (Hsp40) homolog, Dnajc1 0.323 0.000 0.032 0.545 -1.12 1.72 2 subfamily C, member 1 1457988_at BB257743 SEC63-like (S. cerevisiae) Sec63 0.329 0.000 0.302 0.221 1.15 1.94 1442671_at BB210666 Huntingtin interacting protein Hip2 0.215 0.000 0.093 0.033 -1.23 2.08 2 1457374_at AV377264 neural precursor cell Nedd4l 0.155 0.000 0.341 0.043 1.20 1.86 expressed, developmentally down-regulated gene 4-like 1452514_a X65997 kit oncogene Kit 0.001 0.000 0.000 0.000 2.23 2.91 _at 1415900_a BB333334 kit oncogene Kit 0.008 0.000 0.012 0.217 1.69 2.13 _at 1459588_at AW55552 Kit oncogene Kit 0.126 0.000 0.535 0.215 1.21 1.77 6 1453355_at AK007980 WNK lysine deficient protein Wnk2 0.488 0.000 0.002 0.005 1.10 1.81 kinase 2 1430529_at AK019176 casein kinase 1, alpha 1 Csnk1a1 0.448 0.000 0.486 0.671 1.14 2.14 1430579_at AW04555 TRAF2 and NCK interacting Tnik /// 0.842 0.000 0.051 0.680 1.03 1.96 1 kinase /// similar to [Segment LOC665 1 of 2] Traf2 and NCK- 113 /// interacting kinase /// region LOC674 containing TRAF2 and NCK 765 interacting kinase; RIKEN cDNA C630040K21 gene 1440293_at BB821275 RIKEN cDNA C230081A13 C230081 0.722 0.000 0.028 0.004 1.03 -1.38 gene A13Rik 1445032_at AW49115 Death associated protein Dapk1 0.804 0.000 0.150 0.457 -1.04 2.08 8 kinase 1 1458738_at C87678 Guanine nucleotide binding Gna14 0.516 0.000 0.009 0.323 1.13 2.69 protein, alpha 14 1417616_at BC010208 ST6 (alpha-N-acetyl- St6galna 0.000 0.000 0.000 0.019 1.73 2.53 neuraminyl-2,3-beta- c2 galactosyl-1,3)-N- acetylgalactosaminide alpha- 2,6-sialyltransferase 2 1418320_at BC003851 protease, serine, 8 (prostasin) Prss8 0.191 0.000 0.000 0.001 1.31 2.50 1456449_at BB327279 Suppressor of Ty 16 homolog Supt16h 0.447 0.000 0.035 0.641 -1.14 2.11 (S. cerevisiae) 1427477_at BF138523 transmembrane protease, Tmprss1 0.547 0.000 0.000 0.000 -1.12 2.31 serine 13 3 1449317_at NM_0098 CASP8 and FADD-like Cflar 0.734 0.000 0.925 0.529 1.03 -1.53 05 apoptosis regulator 1457454_at BG071065 ubiquitin specific peptidase Usp47 0.683 0.000 0.250 0.052 -1.08 2.33 47 1447294_at BB730861 F-box and WD-40 domain Fbxw11 0.214 0.000 0.105 0.233 1.15 1.73 protein 11 1448747_at AF441120 F-box only protein 32 Fbxo32 0.620 0.000 0.007 0.832 1.15 4.47 1427121_at BF455337 F-box only protein 4 Fbxo4 0.874 0.000 0.404 0.480 1.01 -1.37 1440573_at BB037335 Erbb2 interacting protein Erbb2ip 0.669 0.000 0.082 0.378 1.05 1.70 1422472_at BB045429 peroxisomal biogenesis factor Pex13 0.253 0.000 0.020 0.966 1.19 -1.94 13 1424306_at BB829575 elongation of very long chain Elovl4 0.003 0.000 0.194 0.185 1.89 2.32 fatty acids (FEN1/Elo2, SUR4/Elo3, yeast)-like 4 307 1448485_at NM_0081 gamma-glutamyltransferase 1 Ggt1 0.231 0.000 0.013 0.081 1.51 4.29 16 1424296_at BC019374 glutamate-cysteine ligase, Gclc 0.379 0.000 0.104 0.729 1.18 2.25 catalytic subunit 1419373_at NM_1341 ATPase, H+ transporting, Atp6v1b 0.001 0.000 0.000 0.000 2.24 3.37 57 lysosomal V1 subunit B1 1 1431176_at AK008733 RIKEN cDNA 2310010I16 2310010I 0.011 0.000 0.000 0.028 1.66 2.26 gene 16Rik 1449203_at NM_1308 solute carrier organic anion Slco1a5 0.000 0.000 0.009 0.130 3.02 3.61 61 transporter family, member 1a5 1445837_at BB461843 Dendritic cell protein GA17 Ga17 0.495 0.000 0.115 0.162 1.09 1.75 1424125_at BC023443 potassium channel, subfamily Kcnk13 0.816 0.000 0.922 0.001 1.02 -1.54 K, member 13 1435721_at BE956674 potassium voltage-gated Kcnq4 0.736 0.000 0.000 0.160 -1.03 -1.41 channel, subfamily Q, member 4 1417623_at BG069505 solute carrier family 12, Slc12a2 0.243 0.000 0.037 0.744 1.40 3.50 member 2 1443332_at BB157520 Solute carrier family 12, Slc12a2 0.603 0.000 0.355 0.941 1.06 1.80 member 2 1434096_at BB283443 solute carrier family 4 (anion Slc4a4 0.061 0.000 0.046 0.555 -1.19 -1.56 exchanger), member 4 1449999_a NM_0097 calcium channel, voltage- Cacna2d 0.060 0.000 0.264 0.003 -1.19 1.53 _at 84 dependent, alpha2/delta 1 subunit 1 1459288_at BB472451 Potassium voltage-gated Kcnd2 0.777 0.000 0.000 0.013 1.07 2.76 channel, Shal-related family, member 2 1422178_a NM_0089 RAB17, member RAS Rab17 0.000 0.000 0.001 0.011 2.64 3.15 _at 98 oncogene family 1435082_at AV232599 synaptophysin-like protein Sypl 0.417 0.000 0.571 0.440 1.11 1.87 1451424_at BC027245 gamma-aminobutyric acid Gabrp 0.008 0.000 0.405 0.772 2.42 4.32 (GABA-A) receptor, pi 1447808_s BB244383 solute carrier family 15 Slc15a2 0.147 0.000 0.000 0.001 1.41 2.76 _at (H+/peptide transporter), member 2 1427225_at AF057286 epsin 2 Epn2 0.680 0.000 0.950 0.010 -1.06 1.79 1432017_at AK011828 huntingtin interacting protein Hip1 0.434 0.000 0.680 0.009 1.06 -1.39 1 1442956_at BG075869 Protein phosphatase 1, Ppp1r13 0.763 0.000 0.062 0.040 -1.03 1.66 regulatory (inhibitor) subunit b 13B 1435448_at BM12092 BCL2-like 11 (apoptosis Bcl2l11 0.109 0.000 0.000 0.001 1.19 1.58 5 facilitator) 1435449_at BM12092 BCL2-like 11 (apoptosis Bcl2l11 0.208 0.000 0.605 0.939 1.22 2.37 5 facilitator) 1456005_a BB667581 BCL2-like 11 (apoptosis Bcl2l11 0.873 0.000 0.037 0.125 1.03 2.64 _at facilitator) 1459453_at BB176702 Roundabout homolog 1 Robo1 0.373 0.000 0.344 0.381 -1.10 2.13 (Drosophila) 1418762_at NM_0100 CD55 antigen Cd55 0.001 0.000 0.227 0.856 -1.44 -1.57 16 1460242_at NM_0100 CD55 antigen Cd55 0.008 0.000 0.240 0.762 -1.39 -1.63 16 308 1449308_at NM_0167 complement component 6 C6 0.079 0.000 0.276 0.438 -1.22 -1.62 04 1443299_at BB480432 PDZ and LIM domain 3 Pdlim3 0.021 0.000 0.000 0.000 1.58 3.72 1458432_at BE948993 Non-catalytic region of Nck2 0.364 0.000 0.081 0.268 1.14 2.23 tyrosine kinase adaptor protein 2 1460102_at BB037161 CLIP associating protein 1 Clasp1 0.530 0.000 0.285 0.329 -1.08 2.11 1418831_at AW47599 plakophilin 3 Pkp3 0.155 0.000 0.000 0.001 1.25 2.09 3 1426153_a BC020144 desmoglein 2 Dsg2 0.000 0.000 0.055 0.257 1.49 1.49 _at 1449740_s C79957 desmoglein 2 Dsg2 0.159 0.000 0.000 0.000 1.50 3.57 _at 1426301_at U95030 activated leukocyte cell Alcam 0.302 0.000 0.023 0.025 1.15 1.97 adhesion molecule 1420630_at NM_0289 RIKEN cDNA 8430419L09 8430419 0.055 0.000 0.010 0.036 1.27 1.72 82 gene L09Rik 1437203_at BB460904 Casitas B-lineage lymphoma- Cbll1 0.681 0.000 0.474 0.128 -1.05 1.70 like 1 1445108_at BB283493 Latrophilin 3 Lphn3 0.003 0.000 0.021 0.824 1.46 1.72 1442276_at AV365231 Latrophilin 3 Lphn3 0.813 0.000 0.354 0.130 1.02 1.90 1422141_s NM_0336 component of Sp100-rs Csprs 0.004 0.000 0.886 0.750 1.80 2.30 _at 16 1445952_at BG076125 Membrane associated Magi1 0.122 0.000 0.890 0.014 -1.20 1.77 guanylate kinase, WW and PDZ domain containing 1 1443337_at BB531021 Glutamate receptor Grip1 0.882 0.000 0.070 0.743 1.03 2.37 interacting protein 1 1438451_at BI408524 Rho GTPase-activating Grit 0.877 0.000 0.636 0.488 -1.02 1.89 protein 1425425_a BC004048 Wnt inhibitory factor 1 Wif1 0.009 0.000 0.189 0.275 1.88 2.52 _at 1422537_a NM_0104 inhibitor of DNA binding 2 Id2 0.180 0.000 0.094 0.005 1.18 1.72 _at 96 1429459_at BB499147 sema domain, Sema3d 0.081 0.000 0.167 0.034 -1.22 -1.73 immunoglobulin domain (Ig), short basic domain, secreted, (semaphorin) 3D 1417419_at NM_0076 cyclin D1 Ccnd1 0.551 0.000 0.936 0.709 1.25 5.02 31 1441042_at BE688115 fibroblast growth factor 1 Fgf1 0.345 0.000 0.169 0.151 1.10 1.81 1444598_at BB735884 Ets variant gene 6 (TEL Etv6 0.665 0.000 0.091 0.413 1.06 2.09 oncogene) 1450197_at NM_0299 retinal pigment epithelium 65 Rpe65 0.000 0.000 0.626 0.381 1.91 2.08 87 1448556_at BC005555 prolactin receptor Prlr 0.279 0.000 0.000 0.001 1.33 3.02 1425853_s M22958 prolactin receptor Prlr 0.367 0.000 0.000 0.005 1.28 3.37 _at 1443164_at BG069263 Vacuolar protein sorting 13D Vps13d 0.284 0.000 0.038 0.014 -1.09 1.43 (yeast) 1449028_at AF378088 ras homolog gene family, Rhou 0.016 0.000 0.316 0.185 1.52 2.71 member U 1417311_at NM_0242 cysteine rich protein 2 Crip2 0.270 0.000 0.000 0.104 1.11 1.62 23 1458525_at BB023775 Amyloid beta (A4) precursor App 0.703 0.000 0.334 0.650 1.06 2.00 309 protein 1431328_at AK017392 protein phosphatase 1, Ppp1cb 0.902 0.000 0.715 0.853 -1.01 1.62 catalytic subunit, beta isoform 1425500_x BC003207 Hypothetical protein LOC625 0.000 0.001 0.000 0.000 -4.42 -2.99 _at LOC625794 794 1432418_a AK018487 creatine kinase, Ckmt1 0.002 0.001 0.002 0.009 1.48 1.52 _at mitochondrial 1, ubiquitous 1450724_at NM_0530 down-regulated by Ctnnb1, a Drctnnb1 0.003 0.001 0.000 0.639 -1.23 -1.26 90 a 1441185_at AV319863 Musashi homolog 2 Msi2 0.014 0.001 0.032 0.674 1.34 1.53 (Drosophila) 1436533_at BQ176653 TROVE domain family, Trove2 0.014 0.001 0.067 0.415 2.32 3.43 member 2 1424477_at BC019731 cDNA sequence BC019731 BC01973 0.017 0.001 0.001 0.492 1.34 1.52 1 1443478_at C86695 expressed sequence C86695 C86695 0.023 0.001 0.194 0.705 2.33 4.25 1459324_at BB554124 ------0.026 0.001 0.088 0.002 -1.36 1.70 1424609_a BM22525 similar to hypothetical LOC432 0.027 0.001 0.346 0.089 1.54 2.08 _at 5 protein MGC37588 823 1456442_at BG065737 RAB3A interacting protein Rab3il1 0.028 0.001 0.006 0.071 -1.30 -1.54 (rabin3)-like 1 1457044_at BB007136 RIKEN cDNA 4732474O15 4732474 0.036 0.001 0.003 0.022 1.80 2.89 gene O15Rik 1424824_at BB704967 RIKEN cDNA 9630044O09 9630044 0.040 0.001 0.008 0.124 1.40 1.89 gene O09Rik 1441793_at BB824091 ring finger protein 39 Rnf39 0.043 0.001 0.001 0.030 1.28 1.60 1427030_at BG067674 coiled-coil domain containing Ccdc52 0.051 0.001 0.986 0.526 1.23 1.49 52 1428168_at AK003513 myelin protein zero-like 1 Mpzl1 0.052 0.001 0.001 0.078 1.34 1.72 1426910_at BB398886 ------0.055 0.001 0.000 0.008 1.34 1.80 1429351_at AK018314 kelch-like 24 (Drosophila) Klhl24 0.057 0.001 0.133 0.323 1.16 -1.34 1437085_at AV370040 RIKEN cDNA D630039A03 D630039 0.058 0.001 0.004 0.070 1.51 2.30 gene A03Rik 1440668_at BM23538 ADAMTS-like 3 Adamtsl 0.059 0.001 0.468 0.103 -1.22 -1.50 9 3 1423071_x AW54992 hypothetical gene supported LOC270 0.059 0.001 0.000 0.016 1.50 2.29 _at 8 by BC019681; BC027236 /// 335 /// 1455555_at BF224468 Expressed sequence AV0716 0.081 0.001 0.006 0.007 1.26 1.67 AV071699 99 1422026_at AB046537 peptidase inhibitor 16 Pi16 0.088 0.001 0.008 0.721 -1.19 -1.46 1458537_at BB449824 ------0.088 0.001 0.023 0.922 -1.14 -1.37 1427492_at AF408412 premature ovarian failure 1B Pof1b 0.088 0.001 0.000 0.001 1.81 3.79 1456210_at AV222628 RIKEN cDNA 5430407P10 5430407 0.109 0.001 0.004 0.072 1.23 1.66 gene P10Rik 1433369_at AK018374 RIKEN cDNA 8430401P03 8430401 0.120 0.001 0.303 0.386 -1.24 1.72 gene P03Rik 1435493_at AV297961 desmoplakin Dsp 0.128 0.001 0.000 0.001 1.80 4.45 1441941_x AV171845 serine (or cysteine) peptidase Serpinb5 0.131 0.001 0.000 0.007 1.70 3.79 _at inhibitor, clade B, member 5 1440604_at BB070941 Riken cDNA 8030494B02 8030494 0.137 0.001 0.333 0.057 -1.27 1.89 gene B02Rik 1438856_x BB230853 serine (or cysteine) peptidase Serpinb5 0.138 0.001 0.000 0.000 1.57 3.17 _at inhibitor, clade B, member 5
310 1421064_at AW25837 membrane protein, Mpp5 0.151 0.001 0.972 0.231 -1.22 1.67 3 palmitoylated 5 (MAGUK p55 subfamily member 5) 1417588_at AK019995 UDP-N-acetyl-alpha-D- Galnt3 0.160 0.001 0.000 0.006 1.55 3.24 galactosamine:polypeptide N- acetylgalactosaminyltransfera se 3 1455686_at BB077342 ------0.172 0.001 0.977 0.418 1.17 -1.52 1427610_at BC026631 desmoplakin Dsp 0.172 0.001 0.040 0.380 1.24 1.84 1437766_at AA177898 Transcribed locus --- 0.173 0.001 0.556 0.030 1.14 -1.45 1418449_at NM_1336 ladinin Lad1 0.174 0.001 0.007 0.130 1.59 3.63 64 1435313_at BB770873 Cd200 receptor 4 Cd200r4 0.181 0.001 0.099 0.062 -1.13 -1.41 1427175_at BC004761 expressed sequence AI42893 0.183 0.001 0.064 0.087 1.10 1.33 AI428936 6 1457174_at BB131588 expressed sequence AU0156 0.189 0.001 0.053 0.005 1.18 1.64 AU015680 80 1447344_at AI447264 Transcribed locus --- 0.208 0.001 0.984 0.816 -1.14 1.50 1440984_at BE943712 bromodomain adjacent to Baz2b 0.217 0.001 0.695 0.524 1.20 1.83 zinc finger domain, 2B 1449816_at NM_0205 sulfotransferase family 5A, Sult5a1 0.232 0.001 0.001 0.482 -1.12 -1.43 64 member 1 1437403_at BB308071 RIKEN cDNA E130306M17 E130306 0.233 0.001 0.035 0.269 1.36 2.81 gene M17Rik 1457692_at BB641320 ------0.236 0.001 0.109 0.224 1.22 1.97 1426571_at AW91049 transmembrane protein 16A Tmem16 0.236 0.001 0.000 0.015 1.26 2.07 9 a 1459314_at BB229373 CDK5 regulatory subunit Cdkal1 0.242 0.001 0.001 0.736 -1.10 1.41 associated protein 1-like 1 1423426_at BB018522 RIKEN cDNA 1300012G16 1300012 0.251 0.001 0.013 0.061 1.08 -1.29 gene G16Rik 1451210_at BC010332 phosphatidic acid Ppap2c 0.257 0.001 0.000 0.001 1.19 1.81 phosphatase type 2c 1455340_at BB763709 Expressed sequence AI85244 0.258 0.001 0.023 0.844 1.13 -1.51 AI852444 4 1429433_at BI083627 BAT2 domain containing 1 Bat2d 0.262 0.001 0.252 0.140 -1.28 2.25 1417613_at BF147705 immediate early response 5 Ier5 0.263 0.001 0.286 0.782 -1.15 1.65 1419066_at NM_0302 immediate early response 5- Ier5l 0.268 0.001 0.712 0.719 -1.15 1.70 44 like 1442348_at BB549862 ------0.270 0.001 0.288 0.216 -1.10 -1.38 1420178_at AI593797 Transcribed locus --- 0.279 0.001 0.294 0.030 -1.08 -1.31 1430251_at BB738310 RIKEN cDNA D330022H12 D330022 0.284 0.001 0.053 0.019 1.11 1.45 gene H12Rik 1438528_at AW54597 Pericentriolar material 1 Pcm1 0.290 0.001 0.419 0.981 1.20 2.08 9 1454681_at BG070112 RNA binding motif protein Rbm35a 0.295 0.001 0.000 0.000 1.39 3.28 35A 1441238_at BG074304 RIKEN cDNA 9030416H16 9030416 0.335 0.001 0.947 0.028 -1.14 1.75 gene H16Rik 1440417_at BG918834 DNA segment, Chr 19, D19Ertd 0.350 0.001 0.505 0.301 -1.23 2.43 ERATO Doi 409, expressed 409e 1419700_a NM_0089 prominin 1 Prom1 0.354 0.001 0.000 0.257 1.21 2.12 _at 35 1449639_at AA597087 RIKEN cDNA 0610040J01 0610040 0.359 0.001 0.072 0.494 -1.11 1.57
311 gene J01Rik 1433622_at BB385039 gem (nuclear organelle) Gemin4 0.360 0.001 0.782 0.462 1.08 1.42 associated protein 4 1430697_at BM23551 Alport syndrome, mental Ammecr 0.364 0.001 0.386 0.427 -1.13 1.67 4 retardation, midface 1 hypoplasia and elliptocytosis chromosomal region gene 1 homolog (human) 1442484_at BG066761 DNA segment, Chr 9, D9Ertd3 0.367 0.001 0.254 0.246 -1.16 1.91 ERATO Doi 306, expressed 06e 1445277_at BB152459 16 days neonate thymus --- 0.376 0.001 0.072 0.178 -1.17 1.99 cDNA, RIKEN full-length enriched library, clone:A130006J10 product:unclassifiable, full insert sequence 1438172_x BB091183 exonuclease domain Exod1 0.382 0.001 0.643 0.051 1.09 1.50 _at containing 1 1439550_at AV321031 Zinc finger protein 469 Zfp469 0.384 0.001 0.711 0.459 -1.06 1.28 1440013_at BG072725 Tripartite motif-containing 44 Trim44 0.401 0.001 0.735 0.038 -1.17 2.05 1457588_at C76213 expressed sequence C76213 C76213 0.411 0.001 0.047 0.485 1.10 1.55 1457218_at BB296225 RIKEN cDNA 6430510M02 6430510 0.425 0.001 0.000 0.595 1.16 2.11 gene M02Rik 1426284_at AF473907 RIKEN cDNA 9030623C06 9030623 0.429 0.001 0.044 0.249 -1.06 -1.34 gene C06Rik 1424968_at BC027185 RIKEN cDNA 2210023G05 2210023 0.436 0.001 0.000 0.393 -1.10 -1.58 gene G05Rik 1433598_at BG093966 RIKEN cDNA 9430010O03 9430010 0.440 0.001 0.122 0.345 1.11 1.76 gene O03Rik 1423672_at BC026507 RIKEN cDNA 2510042P03 2510042 0.441 0.001 0.118 0.046 1.06 -1.35 gene P03Rik 1446592_at BB466044 ------0.459 0.001 0.585 0.085 -1.10 1.69 1424966_at BC019416 transmembrane protein 40 Tmem40 0.465 0.001 0.010 0.541 1.24 3.02 1445161_at BE952061 USP6 N-terminal like Usp6nl 0.488 0.001 0.713 0.142 -1.12 1.95 1429537_at BG277020 RIKEN cDNA 5730406M06 5730406 0.497 0.001 0.330 0.024 -1.12 2.04 gene M06Rik 1451447_at BC024476 CUE domain containing 1 Cuedc1 0.517 0.001 0.001 0.121 1.08 1.65 1424490_at BC024802 RIKEN cDNA 2410005H09 2410005 0.539 0.001 0.750 0.116 1.04 1.30 gene H09Rik 1442473_at BG070998 ------0.544 0.001 0.029 0.080 -1.11 1.96 1459283_at BB208378 RIKEN cDNA 6430510B20 6430510 0.553 0.001 0.056 0.007 1.09 1.73 gene B20Rik 1436150_at BB391675 RIKEN cDNA 1700066J24 1700066 0.556 0.001 0.422 0.495 -1.04 -1.32 gene J24Rik 1428127_at BG975608 RIKEN cDNA 4921506J03 4921506 0.562 0.001 0.699 0.068 -1.10 1.93 gene J03Rik 1439011_at BB333400 Transcribed locus --- 0.565 0.001 0.464 0.957 1.05 -1.39 1444588_at AI592278 ------0.598 0.001 0.176 0.162 1.05 -1.43 1426544_a BB118847 tetratricopeptide repeat Ttc14 0.613 0.001 0.007 0.046 1.09 1.86 _at domain 14 1429784_at BB192700 RIKEN cDNA C130032J12 C130032 0.620 0.001 0.391 0.695 1.11 2.36 gene J12Rik 1439117_at AU067755 calmin Clmn 0.635 0.001 0.294 0.053 1.11 2.49 1436347_a BB501229 RIKEN cDNA 5530601H04 5530601 0.642 0.001 0.214 0.038 -1.04 -1.44
312 _at gene H04Rik 1450984_at NM_0115 tight junction protein 2 Tjp2 0.655 0.001 0.006 0.187 1.05 1.53 97 1452914_at AK010587 RIKEN cDNA 2410024N18 2410024 0.663 0.001 0.118 0.917 -1.05 -1.54 gene N18Rik 1417028_a BB283676 tripartite motif protein 2 Trim2 0.694 0.001 0.289 0.672 1.07 2.14 _at 1433015_at AK018204 RIKEN cDNA 6330436F06 6330436 0.701 0.001 0.009 0.296 -1.09 2.41 gene F06Rik 1444237_at BB068784 RIKEN cDNA C330019G07 C330019 0.706 0.001 0.427 0.566 1.05 1.57 gene G07Rik 1458082_at BB220405 Cd200 antigen Cd200 0.706 0.001 0.003 0.426 1.08 2.11 1459860_x BB466780 tripartite motif protein 2 Trim2 0.723 0.001 0.140 0.256 1.06 2.01 _at 1438880_at AV321778 RIKEN cDNA 1700012D14 1700012 0.732 0.001 0.802 0.506 -1.06 1.97 gene D14Rik 1418006_at NM_0242 RIKEN cDNA 5830416A07 5830416 0.736 0.001 0.442 0.030 -1.04 1.63 68 gene A07Rik 1457751_at BB371102 RIKEN cDNA 4832420A03 4832420 0.747 0.001 0.733 0.035 -1.05 1.94 gene A03Rik 1456272_at AV255313 SET domain containing 5 Setd5 0.750 0.001 0.981 0.042 -1.06 1.98 1453662_at AK020987 hypothetical LOC77413 LOC774 0.759 0.001 0.083 0.079 -1.04 1.65 13 1417849_at AW41362 glucocorticoid induced gene Gig1 0.776 0.001 0.000 0.005 1.05 2.09 0 1 1429212_a AK005758 leucine rich repeat containing Lrrc51 0.781 0.001 0.376 0.582 1.03 -1.44 _at 51 1457791_at BB333194 RIKEN cDNA B830008H07 B830008 0.789 0.001 0.616 0.522 -1.03 -1.61 gene H07Rik 1417087_at BG074127 golgi apparatus protein 1 Glg1 0.804 0.001 0.216 0.459 -1.03 1.56 1433968_a BQ174710 ------0.809 0.001 0.001 0.476 -1.02 -1.46 _at 1436362_x AA798895 RIKEN cDNA 2700079J08 2700079 0.831 0.001 0.936 0.091 -1.03 1.70 _at gene J08Rik 1426541_a BF168366 endonuclease domain Endod1 0.847 0.001 0.479 0.019 1.03 1.91 _at containing 1 1424239_at BC006820 RIKEN cDNA 2310066E14 2310066 0.858 0.001 0.681 0.400 -1.01 -1.36 gene E14Rik 1416442_at NM_0104 immediate early response 2 Ier2 0.867 0.001 0.273 0.496 1.02 1.58 99 1430752_at BB397948 RIKEN cDNA C330006D17 C330006 0.873 0.001 0.005 0.015 -1.02 1.85 gene D17Rik 1437878_s BB711506 coiled-coil domain containing Ccdc39 0.947 0.001 0.526 0.219 -1.01 2.28 _at 39 /// tetratricopeptide repeat /// Ttc14 domain 14 1452768_at AK013971 testis expressed gene 261 Tex261 0.950 0.001 0.654 0.706 -1.01 -1.37 1457059_at BB333562 zinc finger protein 438 Zfp438 0.988 0.001 0.072 0.736 1.00 -1.39 1457184_at BB472173 FERM domain containing 6 Frmd6 0.988 0.001 0.410 0.061 1.00 2.28 1453550_a AK011187 male sterility domain Mlstd2 0.524 0.001 0.184 0.601 1.10 1.79 _at containing 2 1449314_at NM_0117 zinc finger protein, multitype Zfpm2 0.274 0.001 0.000 0.848 -1.13 -1.53 66 2 1417817_a BC014727 WW domain containing Wwtr1 0.621 0.001 0.010 0.013 -1.10 2.13 _at transcription regulator 1
313 1418158_at NM_0116 transformation related protein Trp63 0.021 0.001 0.000 0.100 1.76 2.45 41 63 1426431_at AV264681 jagged 2 Jag2 0.404 0.001 0.054 0.616 1.10 1.55 1437833_at BB173900 latent transforming growth Ltbp3 0.141 0.001 0.000 0.000 1.23 1.70 factor beta binding protein 3 1450673_at NM_0077 procollagen, type IX, alpha 2 Col9a2 0.402 0.001 0.004 0.023 1.12 1.73 41 1419486_at BB759833 forkhead box C1 Foxc1 0.003 0.001 0.000 0.006 1.81 2.11 1426851_a X96585 nephroblastoma Nov 0.000 0.001 0.286 0.774 -2.55 -1.78 _at overexpressed gene 1437422_at AV375653 sema domain, seven Sema5a 0.669 0.001 0.497 0.028 -1.13 2.98 thrombospondin repeats (type 1 and type 1-like), transmembrane domain (TM) and short cytoplasmic domain, (semaphorin) 5A 1436223_at BB504737 integrin beta 8 Itgb8 0.004 0.001 0.000 0.001 1.54 1.65 1439018_at AV329790 RIKEN cDNA 6330505N24 6330505 0.598 0.001 0.001 0.002 1.11 2.31 gene N24Rik 1435551_at BG066491 formin homology 2 domain Fhod3 0.019 0.001 0.000 0.007 1.63 2.09 containing 3 1437821_at BB408240 Diaphanous homolog 1 Diap1 0.692 0.001 0.837 0.025 1.08 2.15 (Drosophila) 1433972_at BB558154 calmodulin binding Camta1 0.117 0.001 0.038 0.814 1.19 1.51 transcription activator 1 1418245_a BG277926 RNA binding motif protein 9 Rbm9 0.375 0.001 0.017 0.009 -1.18 2.13 _at 1434651_a AW61146 claudin 3 Cldn3 0.312 0.001 0.000 0.000 1.28 2.57 _at 2 1426332_a AF087821 claudin 3 Cldn3 0.369 0.001 0.000 0.003 1.32 3.44 _at 1442679_at BM24189 Mitogen activated protein Map2k4 0.962 0.001 0.068 0.489 1.01 1.57 8 kinase kinase 4 1421906_at BB760479 peroxisome proliferator Pparbp 0.642 0.001 0.451 0.205 1.07 1.75 activated receptor binding protein 1457712_at BG073323 chromodomain helicase DNA Chd8 0.738 0.001 0.564 0.557 -1.06 2.09 binding protein 8 1458876_at BM19838 Enabled homolog Enah 0.423 0.001 0.280 0.072 -1.10 1.57 9 (Drosophila) 1417693_a NM_0213 growth factor receptor bound Gab1 0.076 0.001 0.002 0.045 1.23 1.55 _at 56 protein 2-associated protein 1 1417694_at NM_0213 growth factor receptor bound Gab1 0.194 0.001 0.023 0.054 1.18 1.63 56 protein 2-associated protein 1 1423319_at AK014111 hematopoietically expressed Hhex 0.377 0.001 0.035 0.079 -1.08 -1.37 homeobox 1427266_at BG065752 polybromo 1 Pb1 0.127 0.001 0.324 0.152 -1.27 1.82 1449163_at NM_0230 single immunoglobulin and Sigirr 0.694 0.001 0.003 0.406 1.03 1.31 59 toll-interleukin 1 receptor (TIR) domain 1425811_a BF124540 cysteine and glycine-rich Csrp1 0.120 0.001 0.002 0.021 1.27 1.85 _at protein 1 1425810_a BF124540 cysteine and glycine-rich Csrp1 0.440 0.001 0.007 0.206 1.17 2.27 _at protein 1 1431239_at AK013444 non-POU-domain-containing, Nono 0.991 0.001 0.994 0.292 1.00 2.84 314 octamer binding protein 1436898_at BI738328 splicing factor Sfpq 0.486 0.001 0.007 0.047 -1.31 4.98 proline/glutamine rich (polypyrimidine tract binding protein associated) 1444746_at BM19503 Polypyrimidine tract binding Ptbp2 0.810 0.001 0.670 0.567 1.06 2.57 3 protein 2 1442886_at BB667153 RIKEN cDNA 1500010G04 1500010 0.818 0.001 0.063 0.090 -1.05 2.56 gene G04Rik 1416318_at AF426024 serine (or cysteine) peptidase Serpinb1 0.128 0.001 0.027 0.003 -1.16 -1.43 inhibitor, clade B, member 1a a 1425538_x BC016891 CEA-related cell adhesion Ceacam1 0.035 0.001 0.000 0.198 1.85 2.86 _at molecule 1 1440223_at BB031290 RNA binding motif protein 6 Rbm6 0.645 0.001 0.061 0.402 -1.04 1.42 1438816_at BM24720 AT hook containing Ahctf1 0.332 0.001 0.001 0.573 -1.11 1.52 1 transcription factor 1 1425565_at BI153474 RE1-silencing transcription Rest 0.761 0.001 0.479 0.026 -1.07 2.51 factor 1458802_at BB164127 human immunodeficiency Hivep3 0.012 0.001 0.001 0.003 1.50 1.78 virus type I enhancer binding protein 3 1460144_at BB352444 CDNA sequence BC052040 BC05204 0.692 0.001 0.299 0.024 -1.03 1.37 0 1443871_at BB200812 Mannosidase, alpha, class Man1a2 0.325 0.001 0.398 0.070 1.08 -1.36 1A, member 2 1426538_a BB828014 transformation related protein Trp53 0.152 0.001 0.992 0.703 1.19 1.65 _at 53 Appendix IX. Differentially expressed genes in Elf5-/- epithelium at day 6 of pregnancy. The top 429 differentially expressed genes in Elf5-/- epithelium at day 6 of pregnancy. Affymetrix ID, Genebank ID, gene title and gene symbol is listed. The unadjusted p-value for the change in gene expression between Elf5-/- and WT at day 4 (p Day 4) and at day 6 (p Day 6) and the difference between cleared fat pad (CFP) and wildtype at day 4 (p Day4CFP) and at day 6 (p Day6CFP) is tabulated. The fold change at day 6 (FC Day 4) and at day 6 (FC Day 6) is also shown. Negative and positive values indicate decreasing and positive genes respectively.
315 Appendix X
Gene FC(Day FC(Day Molecular Fuction symbol Gene title Potential Function 4) 6) Cellular retinoic acid Cyclin binding, lipid Drug metabolism CRABP2 binding protein 2 binding, transporter -2.404 Prostaglandin I2 Synthesis of cholesterol, PTGIS (prostacyclin) synthase steroids and other lipids -1.65 Prostaglandin-endoperoxide synthase 1 (prostaglandin G/H synthase and PTGS1 cyclooxygenase) Prostaglandin biosynthesis -1.464 Carboxylesterase 2 CES2 (intestine, liver) Lipid hydolysis -1.577 Homeobox, hematopoietically Hematopoietic HHEX expressed differentiation -1.366 Solute carrier family 6 (neurotransmitter transporter, serotonin), SLC6A4 member 4 Serotonin transport -1.468 ATP-binding cassette, sub- family C (CFTR/MRP), ABCC1 member 1 Cellular transporter -1.294 Cellular retinoic acid Cyclin binding, lipid Lipid Metabolism CRABP2 binding protein 2 binding, transporter -2.404 Cholesterol and lipid- Carboxyl ester lipase (bile soluble vitamin ester CEL salt-stimulated lipase) hydrolysis, milk protein -3.704 -3.509 Prostaglandin I2 Synthesis of cholesterol, PTGIS (prostacyclin) synthase steroids and other lipids -1.65 Prostaglandin-endoperoxide synthase 1 (prostaglandin G/H synthase and PTGS1 cyclooxygenase) Prostaglandin biosynthesis -1.464 Oxidation of branched SCP2 Sterol carrier protein 2 chain fatty acids -1.661 N-acyl- phosphatidylethanolamine- NAPE- hydrolyzing phospholipase PLD D Phospholipase -1.562 Huntingtin interacting Phospholipid binding / HIP1 protein 1 actin binding -1.385 UDP-Gal:betaGlcNAc beta 1,3-galactosyltransferase, Galactosyltransferase B3GALT1 polypeptide 1 activity -2.012 ATP-binding cassette, sub- family C (CFTR/MRP), ABCC1 member 1 Cellular transporter -1.294 Small Molecule Cellular retinoic acid Cyclin binding, lipid Biochemistry CRABP2 binding protein 2 binding, transporter -2.404
316 DUOX1 Dual oxidase 1 Peroxidase activity -1.572 Cholesterol and lipid- Carboxyl ester lipase (bile soluble vitamin ester CEL salt-stimulated lipase) hydrolysis, milk protein -3.704 -3.509 Prostaglandin I2 Synthesis of cholesterol, PTGIS (prostacyclin) synthase steroids and other lipids -1.65 Prostaglandin-endoperoxide synthase 1 (prostaglandin G/H synthase and PTGS1 cyclooxygenase) Prostaglandin biosynthesis -1.464 Oxidation of branched SCP2 Sterol carrier protein 2 chain fatty acids -1.661 Phosphodiesterase 4A, cAMP-specific (phosphodiesterase E2 dunce homolog, PDE4A Drosophila) cAMP-mediated Signaling -1.42 Carboxylesterase 2 CES2 (intestine, liver) Lipid hydolysis -1.577 G-Protein Coupled EDNRB Endothelin receptor type B Receptor Signaling -1.618 N-acyl- phosphatidylethanolamine- NAPE- hydrolyzing phospholipase PLD D Phospholipase -1.562 Huntingtin interacting Phospholipid binding / HIP1 protein 1 actin binding -1.385 Homeobox, hematopoietically Hematopoietic HHEX expressed differentiation -1.366 UDP-Gal:betaGlcNAc beta 1,3-galactosyltransferase, Galactosyltransferase B3GALT1 polypeptide 1 activity -2.012 Cholesterol and lipid- Carboxyl ester lipase (bile soluble vitamin ester CEL salt-stimulated lipase) hydrolysis, milk protein -3.704 -3.509 Solute carrier family 6 (neurotransmitter transporter, serotonin), SLC6A4 member 4 Serotonin transport -1.468 Ribonucleotide reductase RRM2B M2 B (TP53 inducible) Enzyme -1.508 ATP-binding cassette, sub- family C (CFTR/MRP), ABCC1 member 1 Cellular transporter -1.294 Achaete-scute complex-like Cell cycle ASCL2 2 (Drosophila) Transcription factor -1.245 Ampty spiracles homolog 2 EMX2 (Drosophila) Transcription factor -1.698 Interferon, gamma- IFI16 inducible protein 16 Transcription factor -1.634 Ribonucleotide reductase RRM2B M2 B (TP53 inducible) Enzyme -1.508
317 Appendix X. Decreasing genes identified in the top 4 over-represented functional categories as a result of loss of epithelial Elf5 The decreasing genes identified in the top 4 over-represented functional categories in Elf5-/- epithelium at day 4 and 6 of pregnancy. Affymetrix ID, gene title and gene symbol is listed. The fold change between Elf5-/- and WT transplants at day 4 (p Day 4) and at day 6 (p Day 6) is tabulated. Negative and positive values indicate decreasing and positive genes respectively. Text in red indicates the probe sets that were also significantly different between the cleared fat pad and wildtype samples (epithelial enriched)
318 Appendix XI
Gene FC(Day FC(Da Molecular Fuction symbol Gene title Potential Function 4) y 6) V-kit Hardy-Zuckerman 4 feline sarcoma viral Type 3 transmembrane Cell Cycle KIT oncogene homolog receptor 2.232 2.13 Cell growth, differentiation, mitotic Protein tyrosine cycle, and oncogenic phosphatase, receptor type, transformation cell-cell PTPRF F contacts 1.784 PRKC, apoptosis, WT1, PAWR regulator Regulation of transcription 1.802 BCL2-like 11 (apoptosis Apoptosis activator/PTEN BCL2L11 facilitator) signalling 1.581 Tumor protein p53 Apoptosis, cell cycle TP53INP1 inducible nuclear protein 1 progression 1.423 Serpin peptidase inhibitor, clade B (ovalbumin), Motility, apoptosis, SERPINB5 member 5 invasion, invasiveness 3.022 SRY (sex determining Regulation of transcription, SOX10 region Y)-box 10 cell fate\ 1.527 V-kit Hardy-Zuckerman 4 Cellular feline sarcoma viral Type 3 transmembrane Compromise KIT oncogene homolog receptor 2.232 2.13 SRY (sex determining Regulation of transcription, SOX10 region Y)-box 10 cell fate 1.527 PRKC, apoptosis, WT1, Cell Death PAWR regulator Regulation of transcription 1.802 Component of tight CLDN3 Claudin 3 junction strands 3.437 BCL2-like 11 (apoptosis Apoptosis activator/PTEN BCL2L11 facilitator) signalling 1.581 Syndecan 4 (amphiglycan, SDC4 ryudocan) Intracellular signaling 4.426 Tumor protein p53 Apoptosis, cell cycle TP53INP1 inducible nuclear protein 1 progression 1.423 Serpin peptidase inhibitor, clade B (ovalbumin), Motility, apoptosis, SERPINB5 member 5 invasion, invasiveness 3.022 SRY (sex determining Regulation of transcription, SOX10 region Y)-box 10 cell fate\ 1.527 Trichorhinophalangeal TRPS1 syndrome I Apoptosis 3.473 V-kit Hardy-Zuckerman 4 feline sarcoma viral Type 3 transmembrane KIT oncogene homolog receptor 2.232 2.13 Cellular Proliferation, Development PRLR Prolactin receptor differentiation, apoptosis, 3.368 319 adhesion, formation LAMA1 Laminin, alpha 1 Integrin Signaling 3.526 FOXC1 Forkhead box C1 Transcription factor 2.609 Serpin peptidase inhibitor, clade B (ovalbumin), Motility, apoptosis, SERPINB5 member 5 invasion, invasiveness 3.022 SRY (sex determining Regulation of transcription, SOX10 region Y)-box 10 cell fate 1.527 V-kit Hardy-Zuckerman 4 feline sarcoma viral Type 3 transmembrane KIT oncogene homolog receptor 2.232 2.13 LAMA3 Laminin, alpha 3 Integrin Signaling 2.631 Appendix XI. Increasing genes identified in the top 4 over-represented functional categories as a result of loss of epithelial Elf5. The increasing genes identified in the top 4 over-represented functional categories in Elf5-/- epithelium at day 4 and 6 of pregnancy. Affymetrix ID, gene title and gene symbol is listed. The fold change between Elf5-/- and WT transplants at day 4 (p Day 4) and at day 6 (p Day 6) is tabulated. Negative and positive values indicate decreasing and positive genes respectively. All probe sets listed in this table were also significantly different between the cleared fat pad and wildtype samples (epithelial enriched).
320 Appendix XII
Probe Set ID Gene Title Gene Symbol p-value Fold Change 1448556_at prolactin receptor Prlr 0.00001 -4.34405 1437397_at expressed sequence AI987712 AI987712 0.00002 -3.85888 1437230_at Potassium voltage-gated channel, Kcna1 0.00009 2.20202 shaker-related subfamily, member 1 (Kcna1), mRNA 1417822_at DNA segment, Chr 17, human D17H6S56E-5 0.00020 -2.46929 D6S56E 5 1458347_s_at transmembrane protease, serine 2 Tmprss2 0.00023 -2.17936 1417156_at keratin complex 1, acidic, gene 19 Krt1-19 0.00027 -1.99045 1425853_s_at prolactin receptor Prlr 0.00029 -9.43127 1417416_at potassium voltage-gated channel, Kcna1 0.00043 2.41945 shaker-related subfamily, member 1 1436555_at ------0.00044 -1.86263 1419154_at transmembrane protease, serine 2 Tmprss2 0.00052 -1.75441 1441102_at expressed sequence AI987712 AI987712 0.00053 -2.97489 1421340_at mitogen activated protein kinase Map3k5 0.00063 1.77449 kinase kinase 5 1418496_at forkhead box A1 Foxa1 0.00076 -1.75833 1437502_x_at CD24a antigen Cd24a 0.00093 -1.96892 1437019_at RIKEN cDNA 2200001I15 gene 2200001I15Rik 0.00120 -1.78120 1450206_at deleted in liver cancer 1 Dlc1 0.00125 1.63074 1448182_a_at CD24a antigen Cd24a 0.00131 -1.70128 1440662_at Ral guanine nucleotide dissociation Rgl1 0.00143 1.58337 stimulator,-like 1, mRNA (cDNA clone MGC:18430 IMAGE:4241244) 1448169_at keratin complex 1, acidic, gene 18 Krt1-18 0.00162 -1.57223 1445008_at ------0.00168 1.60314 1436910_at RAS protein activator like 2 Rasal2 0.00177 1.56831 1440305_at 0 day neonate eyeball cDNA, RIKEN --- 0.00190 1.57799 full-length enriched library, clone:E130307M10 product:unclassifiable, full insert sequence 1451905_a_at myxovirus (influenza virus) Mx1 0.00195 1.66491 resistance 1 1460569_x_at claudin 3 Cldn3 0.00213 -1.55385 1417821_at DNA segment, Chr 17, human D17H6S56E-5 0.00215 -3.44588 D6S56E 5 1432538_a_at replication factor C (activator 1) 3 Rfc3 0.00215 -1.51772 1416034_at CD24a antigen Cd24a 0.00220 -1.93706 1452534_a_at high mobility group box 2 Hmgb2 0.00221 -1.52648 1434553_at transmembrane protein 56 Tmem56 0.00238 -1.69910 1457349_at gene model 69, (NCBI) Gm69 0.00258 -1.74478 1424351_at WAP four-disulfide core domain 2 Wfdc2 0.00265 -1.79208 1443516_at Spinocerebellar ataxia 2 homolog Sca2 0.00274 1.90710 (human) (Sca2), mRNA 1429906_at RIKEN cDNA A930035E12 gene A930035E12Ri 0.00279 -1.48464 k 1435742_at RIKEN cDNA 1110034C04 gene 1110034C04Rik 0.00294 1.48527 1447147_at Autophagy-related 7 (yeast) (Atg7), Apg7l 0.00296 1.57912 321 mRNA 1439925_at ------0.00304 1.50040 1451701_x_at claudin 3 Cldn3 0.00305 -1.64741 1419700_a_at prominin 1 Prom1 0.00317 -1.84661 1427095_at CUB domain containing protein 1 Cdcp1 0.00345 -1.50406 1447951_at RIKEN cDNA 5730403B10 gene, 5730403B10Rik 0.00352 1.57338 mRNA (cDNA clone MGC:8188 IMAGE:3590511) 1417373_a_at tubulin, alpha 4 Tuba4 0.00359 -1.73973 1439079_a_at Erbb2 interacting protein Erbb2ip 0.00395 1.56864 1425469_a_at ------0.00429 -2.08281 1458601_at RIKEN cDNA 8030447M02 gene 8030447M02Ri 0.00431 1.57284 k 1437883_s_at ------0.00440 1.44913 1417896_at tight junction protein 3 Tjp3 0.00443 -1.47392 1434893_at ATPase, Na+/K+ transporting, alpha Atp1a2 0.00464 1.69004 2 polypeptide, mRNA (cDNA clone MGC:36347 IMAGE:4955003) 1449369_at transmembrane protease, serine 2 Tmprss2 0.00466 -2.60076 1423569_at glycine amidinotransferase (L- Gatm 0.00470 1.78041 arginine:glycine amidinotransferase) 1453070_at protocadherin 17 Pcdh17 0.00472 1.46844 1460711_at RIKEN cDNA 4930461P20 gene 4930461P20Rik 0.00479 1.54287 1423341_at chondroitin sulfate proteoglycan 4 Cspg4 0.00506 1.46862 1449178_at PDZ and LIM domain 3 Pdlim3 0.00507 -2.99848 1447062_at Vacuolar protein sorting 24 (yeast) Vps24 0.00514 1.41685 (Vps24), mRNA 1435553_at PDZ domain containing 3 Pdzk3 0.00527 1.52220 1434651_a_at claudin 3 Cldn3 0.00560 -1.56586 1416610_a_at chloride channel 3 Clcn3 0.00570 1.44928 1423432_at pleckstrin homology domain Phip 0.00577 1.40927 interacting protein 1417957_a_at tetraspan 1 Tspan1 0.00585 -1.87812 1415936_at breast cancer anti-estrogen resistance Bcar3 0.00588 1.44627 3 1438841_s_at arginase type II Arg2 0.00596 -1.54554 1424649_a_at tetraspanin 8 Tspan8 0.00627 -1.49054 1439497_at Adult male testis cDNA, RIKEN full- --- 0.00644 1.55449 length enriched library, clone:4933415E08 product:unclassifiable, full insert sequence 1446481_at Amyloid beta (A4) precursor protein- Apbb2 0.00655 1.65632 binding, family B, member 2, mRNA (cDNA clone MGC:100042 IMAGE:30550087) 1427465_at ATPase, Na+/K+ transporting, alpha Atp1a2 0.00656 1.43109 2 polypeptide 1459187_at ST3 beta-galactoside alpha-2,3- Siat4c 0.00694 1.56416 sialyltransferase 4, mRNA (cDNA clone MGC:6096 IMAGE:3497996) 1435494_s_at desmoplakin Dsp 0.00694 -1.46011 1455510_at Spop Spop 0.00701 1.40331 1440028_at ------0.00725 1.41063
322 1455244_at dishevelled associated activator of Daam1 0.00729 1.43166 morphogenesis 1 1417374_at tubulin, alpha 4 Tuba4 0.00734 -1.61162 1455381_at RIKEN cDNA 4921513D23 gene 4921513D23Rik 0.00740 1.39330 1455990_at kinesin family member 23 Kif23 0.00764 -1.43339 1457553_at ------0.00791 2.33927 1433870_at cDNA sequence BC022765 BC022765 0.00804 -1.42408 1455228_at ------0.00804 1.37830 1448595_a_at reduced expression 3 Rex3 0.00813 -1.66040 1452856_at HCF-binding transcription factor MGI:2675296 0.00827 1.37599 Zhangfei 1448377_at secretory leukocyte peptidase Slpi 0.00839 -1.48388 inhibitor 1438116_x_at solute carrier family 9 Slc9a3r1 0.00863 -1.54971 (sodium/hydrogen exchanger), isoform 3 regulator 1 1427225_at gene model 1965, (NCBI) Gm1965 0.00866 1.65454 1418486_at vanin 1 Vnn1 0.00878 1.37235 1452160_at TCDD-inducible poly(ADP-ribose) Tiparp 0.00884 1.54176 polymerase 1460666_a_at early B-cell factor 3 Ebf3 0.00896 1.83306 1428440_at solute carrier family 25 Slc25a12 0.00908 -2.14652 (mitochondrial carrier, Aralar), member 12 1428074_at RIKEN cDNA 2310037P21 gene 2310037P21Rik 0.00913 -1.37052 1447174_at DACH protein (Dach) Dach1 0.00913 1.56287 1428857_at RIKEN cDNA 2610304F09 gene 2610304F09Rik 0.00958 1.39542 1425584_x_at ------0.00961 -1.57062 1452309_at cingulin-like 1 Cgnl1 0.00962 1.50801 1440417_at DNA segment, Chr 19, ERATO Doi D19Ertd409e 0.00979 -1.37475 409, expressed 1418831_at plakophilin 3 Pkp3 0.00993 -1.46158 1457373_at Transcribed locus --- 0.01008 1.49423 1448962_at myosin, heavy polypeptide 11, Myh11 0.01014 1.35206 smooth muscle 1437474_at GATA zinc finger domain containing Gatad2b 0.01022 1.48321 2B 1455703_at thymoma viral proto-oncogene 2 Akt2 0.01038 1.34388 1455007_s_at glutamic pyruvate transaminase Gpt2 0.01046 1.49919 (alanine aminotransferase) 2 1425505_at myosin, light polypeptide kinase Mylk 0.01047 1.54119 1436870_s_at expressed sequence AU041783 AU041783 0.01047 1.43270 1450733_at bicaudal D homolog 2 (Drosophila) Bicd2 0.01056 1.53203 1437066_at RIKEN cDNA 7330412A13 gene 7330412A13Rik 0.01062 1.76346 1419503_at stanniocalcin 2 Stc2 0.01064 -1.47874 1456761_at RIKEN cDNA D630030B22 gene D630030B22Ri 0.01071 1.53631 k 1437318_at p21 (CDKN1A)-activated kinase 3 Pak3 0.01073 1.39637 1444320_at DDHD domain containing 2 Ddhd2 0.01081 1.36069 1424198_at discs, large homolog 5 (Drosophila) Dlg5 0.01097 1.37115 1443736_at PREDICTED: Mus musculus RIKEN 4833412E22Rik 0.01125 -1.33669 cDNA 4833412E22 gene (4833412E22Rik), mRNA 1449405_at tensin 1 Tns1 0.01129 1.61381
323 1445827_at Protein kinase C binding protein 1, Prkcbp1 0.01131 1.37283 mRNA (cDNA clone IMAGE:4194817) 1452350_at ------0.01151 1.56233 1437559_at DNA segment, Chr 13, Brigham & D13Bwg1146e 0.01157 1.35423 Women's Genetics 1146 expressed 1418517_at Iroquois related homeobox 3 Irx3 0.01160 -1.33604 (Drosophila) 1451234_at cDNA sequence BC021381 BC021381 0.01176 1.33224 1434219_at stromal interaction molecule 2 Stim2 0.01196 1.43125 1450401_at nuclear receptor coactivator 6 Ncoa6ip 0.01197 1.60897 interacting protein 1419356_at Kruppel-like factor 7 (ubiquitous) Klf7 0.01217 2.10177 1439527_at Adult female vagina cDNA, RIKEN --- 0.01222 -1.40019 full-length enriched library, clone:9930117J11 product:hypothetical protein, full insert sequence 1418488_s_at receptor-interacting serine-threonine Ripk4 0.01246 -1.47557 kinase 4 1438451_at Rho GTPase-activating protein MGI:2450166 0.01254 1.35312 1428867_at RIKEN cDNA 4933417E01 gene 4933417E01Rik 0.01283 1.36517 1451069_at proviral integration site 3 Pim3 0.01285 1.37196 1453582_at choline kinase alpha Chka 0.01296 1.47819 1453145_at RIKEN cDNA 4933439C20 gene 4933439C20Rik 0.01312 1.35997 1428519_at RIKEN cDNA 2610528E23 gene 2610528E23Rik 0.01316 -1.37399 1419407_at hemolytic complement Hc 0.01317 -1.45922 1459707_at Phosphofurin acidic cluster sorting Pacs1 0.01332 1.38943 protein 1 (Pacs1), mRNA 1434703_at exostoses (multiple)-like 3 Extl3 0.01414 1.39588 1452114_s_at insulin-like growth factor binding Igfbp5 0.01414 1.40859 protein 5 1458637_x_at Ubiquitin protein ligase E3C, mRNA Ube3c 0.01429 1.33926 (cDNA clone IMAGE:5353408) 1452789_at stannin Snn 0.01436 1.44292 1431798_a_at synapse defective 1, Rho GTPase, Syde1 0.01436 1.36547 homolog 1 (C. elegans) 1427878_at RIKEN cDNA 0610010O12 gene 0610010O12Rik 0.01452 -1.53335 1452828_at F-box only protein 21 Fbxo21 0.01498 1.33812 1448393_at claudin 7 Cldn7 0.01505 -1.60059 1454934_at protein phosphatase 1F (PP2C Ppm1f 0.01511 1.43630 domain containing) 1431292_a_at protein tyrosine kinase 9-like (A6- Ptk9l 0.01542 -1.41245 related protein) 1455131_at optic atrophy 3 (human) Opa3 0.01544 1.40305 1456532_at ------0.01552 1.31650 1443489_at ------0.01568 1.31012 1426933_at oxidative-stress responsive 1 Oxsr1 0.01569 1.39823 1446234_at Ubiquitously transcribed Utx 0.01612 1.32631 tetratricopeptide repeat gene, X chromosome, mRNA (cDNA clone IMAGE:5356391) 1418341_at RAB4A, member RAS oncogene Rab4a 0.01614 1.33486 family
324 1435628_x_at cDNA sequence BC005512 /// similar BC005512 /// 0.01636 1.32696 to BC005512 protein /// hypothetical LOC277193 /// LOC432650 /// similar to BC005512 LOC432650 /// protein LOC432898 1437065_at RIKEN cDNA 7330412A13 gene 7330412A13Rik 0.01645 1.40331 1438496_a_at RIKEN cDNA 6330505F04 gene 6330505F04Rik 0.01655 1.47906 1443882_at Mus musculus, clone --- 0.01659 1.43134 IMAGE:3983419, mRNA 1460207_s_at E2F transcription factor 5 E2f5 0.01661 1.32553 1422818_at neural precursor cell expressed, Nedd9 0.01671 1.33075 developmentally down-regulated gene 9 1429166_s_at Calmin (Clmn), mRNA Clmn 0.01680 1.32193 1456960_at Adenosine kinase (Adk), mRNA Adk 0.01703 1.30425 1417817_a_at WW domain containing transcription Wwtr1 0.01721 1.60818 regulator 1 1459485_at Neogenin (Neo1), mRNA Neo1 0.01739 1.42414 1430792_at RIKEN cDNA 3230402G14 gene 3230402G14Rik 0.01752 1.44001 1440966_at membrane-associated ring finger Mar-07 0.01756 1.50969 (C3HC4) 7 1459595_at Transcribed locus --- 0.01757 1.38032 1458934_at DNA segment, Chr 5, ERATO Doi D5Ertd505e 0.01780 1.45246 505 , expressed 1441389_at Transcribed locus --- 0.01790 1.31952 1433993_at RIKEN cDNA 4931406P16 gene 4931406P16Rik 0.01819 1.30396 1416579_a_at tumor-associated calcium signal Tacstd1 0.01824 -1.54325 transducer 1 1455050_at RIKEN cDNA E130203B14 gene E130203B14Ri 0.01855 1.40641 k 1419459_a_at RIKEN cDNA 2610529C04 gene 2610529C04Rik 0.01862 1.48482 1439686_at Muscleblind-like 1 (Drosophila) Mbnl1 0.01867 1.56299 (Mbnl1), mRNA 1425471_x_at ------0.01870 -2.29374 1430191_at RIKEN cDNA 9130004J05 gene 9130004J05Rik 0.01877 1.48846 1440037_at pre B-cell leukemia transcription Pbx1 0.01897 1.32735 factor 1 1425400_a_at Cbp/p300-interacting transactivator, Cited4 0.01930 -1.32921 with Glu/Asp-rich carboxy-terminal domain, 4 1447649_x_at DnaJ (Hsp40) homolog, subfamily C, Dnajc1 0.01955 1.29684 member 1 1423995_at kinesin family member 1B Kif1b 0.01970 1.35663 1418847_at arginase type II Arg2 0.02017 -1.63366 1428239_at ankyrin repeat domain 16 Ankrd16 0.02028 1.29877 1455967_at sorbin and SH3 domain containing 1 Sorbs1 0.02031 1.47761 1449379_at kinase insert domain protein receptor Kdr 0.02038 1.33178 1448127_at ribonucleotide reductase M1 Rrm1 0.02066 -1.64042 1437414_at zinc finger protein 217 Zfp217 0.02071 1.28420 1459981_s_at rosbin, round spermatid basic protein Rsbn1 0.02089 1.30309 1 1429487_at protein phosphatase 1, regulatory Ppp1r12a 0.02103 1.39151 (inhibitor) subunit 12A 1433769_at ALS2 C-terminal like Als2cl 0.02127 1.35076 1428232_at cleavage and polyadenylation Cpsf6 0.02132 1.30660
325 specific factor 6 1448360_s_at angel homolog 2 (Drosophila) Angel2 0.02140 1.28517 1419833_s_at centaurin, delta 3 Centd3 0.02146 1.39442 1438028_at RIKEN cDNA 4930535B03 gene 4930535B03Rik 0.02157 1.55160 1429538_a_at RIKEN cDNA 5730406M06 gene 5730406M06Ri 0.02175 1.35248 k 1457177_at 12 days embryo male wolffian duct --- 0.02188 1.32958 includes surrounding region cDNA, RIKEN full-length enriched library, clone:6720415L05 product:unclassifiable, full insert sequence 1444096_at ------0.02198 1.47030 1443576_at SLIT2 (Slit2) Slit2 0.02201 -1.28582 1448788_at Cd200 antigen Cd200 0.02220 -1.33627 1428834_at dual specificity phosphatase 4 Dusp4 0.02226 1.35824 1454255_at RIKEN cDNA 5430434F05 gene 5430434F05Rik 0.02227 1.29590 1424730_a_at solute carrier family 15 (H+/peptide Slc15a2 0.02242 -1.37475 transporter), member 2 1436984_at abl-interactor 2 Abi2 0.02252 1.31410 1427120_at zinc finger protein 26 Zfp26 0.02286 1.40972 1426332_a_at claudin 3 Cldn3 0.02306 -1.59663 1441779_at RIKEN cDNA 9530006C21 gene 9530006C21Rik 0.02307 1.27571 1438349_at cDNA sequence BC043476 BC043476 0.02309 1.31939 1454871_at RNA binding motif protein 15B Rbm15b 0.02346 -1.30746 1452989_at RIKEN cDNA 2900009J20 gene 2900009J20Rik 0.02368 1.41539 1450630_at queuine tRNA-ribosyltransferase 1 Qtrt1 0.02392 -1.40501 1425470_at ------0.02399 -2.32272 1453698_at RIKEN cDNA 6030451C04 gene 6030451C04Rik 0.02404 1.94939 1447094_at Yamaguchi sarcoma viral (v-yes) Yes 0.02416 1.45109 oncogene homolog 1, mRNA (cDNA clone MGC:18632 IMAGE:3989783) 1417500_a_at transglutaminase 2, C polypeptide Tgm2 0.02422 1.27279 1434026_at Atpase, class I, type 8B, member 2 Atp8b2 0.02423 1.29485 1426147_s_at claudin 10 Cldn10 0.02456 -1.45997 1457510_at PHD finger protein 14 (Phf14), Phf14 0.02472 1.36056 mRNA 1451097_at vasodilator-stimulated Vasp 0.02501 1.33602 phosphoprotein 1427564_at diaphanous homolog 2 (Drosophila) Diap2 0.02511 1.45060 1436216_s_at RIKEN cDNA 2610204M08 gene 2610204M08Ri 0.02517 1.36000 k 1425809_at Fatty acid binding protein 4, Fabp4 0.02520 1.79535 adipocyte (Fabp4), mRNA 1437941_at phosphorylase kinase alpha 2 Phka2 0.02562 1.83160 1439216_at PREDICTED: hypothetical protein 2900075B16Rik 0.02563 1.33007 LOC78506 [Mus musculus], mRNA sequence 1426858_at inhibin beta-B Inhbb 0.02565 1.32828 1416410_at platelet-activating factor Pafah1b3 0.02571 -1.39763 acetylhydrolase, isoform 1b, alpha1 subunit 1426707_at tubulin, gamma complex associated Tubgcp3 0.02592 1.36196 protein 3
326 1454717_at Ankyrin repeat domain 27 (VPS9 Ankrd27 0.02603 1.38303 domain) (Ankrd27), transcript variant 1, mRNA 1445105_at P300/CBP-associated factor (Pcaf), Pcaf 0.02607 1.42121 mRNA 1434834_at suppressor of cytokine signaling 7 Socs7 0.02611 1.47251 1436631_at RIKEN cDNA 2010010M04 gene 2010010M04Ri 0.02611 1.32547 k 1443212_at ------0.02626 1.26696 1423597_at ATPase, aminophospholipid Atp8a1 0.02633 2.00738 transporter (APLT), class I, type 8A, member 1 1417738_at RAB25, member RAS oncogene Rab25 0.02642 -1.35448 family 1444456_at RIKEN cDNA 9030425P06 gene 9030425P06Rik 0.02648 1.29496 1416656_at chloride intracellular channel 1 Clic1 0.02651 -1.40136 1437404_at microtubule associated Mast4 0.02653 1.29165 serine/threonine kinase family member 4 1425242_at RIKEN cDNA 1810006K21 gene 1810006K21Rik 0.02659 -1.26873 1433992_at apical protein, Xenopus laevis-like Apxl 0.02679 1.26661 1449484_at stanniocalcin 2 Stc2 0.02686 -1.34807 1440295_at RIKEN cDNA 1110057K04 gene 1110057K04Rik 0.02692 1.26875 1425675_s_at CEA-related cell adhesion molecule 1 Ceacam1 0.02703 -1.39159 1448261_at cadherin 1 Cdh1 0.02709 -1.41512 1428436_at RIKEN cDNA 2700023B17 gene 2700023B17Rik 0.02719 1.28455 1439949_at Glycogen synthase kinase 3 beta, Gsk3b 0.02733 1.27408 mRNA (cDNA clone MGC:6814 IMAGE:2648507) 1419494_a_at tumor protein D52 Tpd52 0.02739 -1.42654 1444746_at Polypyrimidine tract binding protein Ptbp2 0.02753 1.25977 2, mRNA (cDNA clone MGC:11671 IMAGE:3709255) 1416858_a_at FK506 binding protein 3 Fkbp3 0.02764 -1.30850 1436985_at zinc finger protein 644 Zfp644 0.02776 1.30240 1434480_at RIKEN cDNA 4930402E16 gene 4930402E16Rik 0.02782 1.85283 1426544_a_at tetratricopeptide repeat domain 14 Ttc14 0.02790 1.57872 1419555_at E74-like factor 5 Elf5 0.02798 -1.62891 1439929_at PREDICTED: similar to malignant Mfhas1 0.02801 1.26148 fibrous histiocytoma amplified sequence 1 [Mus musculus], mRNA sequence 1446332_at Protocadherin gamma subfamily A, Pcdhga12 0.02810 1.34935 10, mRNA (cDNA clone MGC:40648 IMAGE:5400956) 1443247_at ------0.02831 1.39591 1450931_at dedicator of cytokinesis 9 Dock9 0.02840 1.28999 1437123_at multimerin 2 Mmrn2 0.02854 1.37191 1417110_at mannosidase 1, alpha Man1a 0.02856 1.42689 1455320_at expressed sequence AI480535 AI480535 0.02877 1.29946 1460694_s_at supervillin Svil 0.02891 -1.29472 1448989_a_at myosin IB Myo1b 0.02897 1.29629 1435989_x_at similar to cytokeratin EndoA - mouse LOC434261 0.02910 -1.58279 1448155_at programmed cell death 6 interacting Pdcd6ip 0.02911 1.39330
327 protein 1434741_at PREDICTED: Mus musculus ras --- 0.02914 1.39961 responsive element binding protein 1 (Rreb1), mRNA 1435088_at ------0.02944 1.26755 1418200_at GLI-Kruppel family member HKR3 Hkr3 0.02953 -1.31382 1425502_x_at ------0.02954 -1.27999 1448484_at S-adenosylmethionine decarboxylase Amd1 0.02964 -1.28886 1 1426839_at polymerase (DNA-directed), delta 3, Pold3 0.02966 -1.37855 accessory subunit 1455393_at ------0.02993 1.30555 1433742_at ankyrin repeat domain 15 Ankrd15 0.02999 1.28426 1434997_at cell division cycle 2-like 6 (CDK8- Cdc2l6 0.03020 -1.31440 like) 1421321_a_at neuroepithelial cell transforming Net1 0.03024 1.35223 gene 1 1439369_x_at solute carrier family 9 Slc9a3r2 0.03029 1.26891 (sodium/hydrogen exchanger), isoform 3 regulator 2 1429310_at fibronectin leucine rich Flrt3 0.03032 1.43132 transmembrane protein 3 1453113_at WD repeat, SAM and U-box domain Wdsub1 0.03044 1.25411 containing 1 1444279_at HECT, UBA and WWE domain Huwe1 0.03049 1.33137 containing 1 1416357_a_at melanoma cell adhesion molecule Mcam 0.03068 1.29404 1421344_a_at ajuba Jub 0.03075 1.27627 1443603_at RIKEN cDNA A030012M09 gene, A030012M09Ri 0.03100 1.64729 mRNA (cDNA clone MGC:76781 k IMAGE:30067935) 1438592_at PREDICTED: Mus musculus NIMA Nek1 0.03114 1.27995 (never in mitosis gene a)-related expressed kinase 1 (Nek1), mRNA 1434361_at SH3 and PX domain containing 3 Sh3px3 0.03126 1.29196 1437181_at pellino 2 Peli2 0.03139 1.37514 1457641_at Adult male aorta and vein cDNA, --- 0.03148 1.26750 RIKEN full-length enriched library, clone:A530003I17 product:unclassifiable, full insert sequence 1448460_at activin A receptor, type 1 Acvr1 0.03165 1.36542 1454685_at G protein-coupled receptor 146 Gpr146 0.03173 1.30987 1459488_at RIKEN cDNA 3110050K21 gene, 3110050K21Rik 0.03176 1.45677 mRNA (cDNA clone IMAGE:5345757) 1422123_s_at CEA-related cell adhesion molecule 1 Ceacam1 /// 0.03185 -1.50626 /// CEA-related cell adhesion Ceacam2 molecule 2 1438400_at RIKEN cDNA 4632411B12 gene 4632411B12Rik 0.03191 1.24859 1420150_at splA/ryanodine receptor domain and Spsb1 0.03202 1.37624 SOCS box containing 1 1416832_at solute carrier family 39 (metal ion Slc39a8 0.03209 1.28907 transporter), member 8
328 1446906_at expressed sequence C81615 C81615 0.03217 1.27861 1426587_a_at signal transducer and activator of Stat3 0.03218 1.38944 transcription 3 1458075_at Bpag1-e mRNA for bullous Dst 0.03251 1.35891 pemphigoid antigen 1-e 1437372_at cleavage and polyadenylation Cpsf6 0.03253 1.71791 specific factor 6 1415767_at YTH domain family 1 Ythdf1 0.03255 1.35280 1417084_at eukaryotic translation initiation factor Eif4ebp2 0.03262 1.72595 4E binding protein 2 1459354_at Strain ICR endothelial and smooth Dcbld2 0.03264 1.34176 muscle cell-derived neuropilin-like protein (Esdn) 1442760_x_at RTN4 (Rtn4) mRNA, complete cds, Rtn4 0.03266 1.28464 alternatively spliced 1421752_a_at serine (or cysteine) peptidase Serpinb5 0.03276 -1.42490 inhibitor, clade B, member 5 1427143_at jumonji, AT rich interactive domain Jarid1b 0.03300 1.40645 1B (Rbp2 like) 1451750_at ------0.03313 1.28957 1448886_at GATA binding protein 3 Gata3 0.03322 -1.35295 1418635_at ets variant gene 3 Etv3 0.03329 1.53483 1420132_s_at Pituitary tumor-transforming 1 Pttg1ip 0.03332 1.31219 interacting protein, mRNA (cDNA clone MGC:38220 IMAGE:5323397) 1435945_a_at potassium intermediate/small Kcnn4 0.03336 -1.72102 conductance calcium-activated channel, subfamily N, member 4 1435336_at cadherin EGF LAG seven-pass G- Celsr2 0.03355 1.27866 type receptor 2 1450854_at proliferation-associated 2G4 Pa2g4 0.03356 1.26055 1451191_at cellular retinoic acid binding protein Crabp2 0.03376 -1.48279 II 1447040_at Activated leukocyte cell adhesion Alcam 0.03378 -1.26405 molecule, mRNA (cDNA clone MGC:27910 IMAGE:3501215) 1437543_at far upstream element (FUSE) binding Fubp1 0.03382 1.34282 protein 1 1440488_at Protein tyrosine phosphatase, non- Ptpn4 0.03404 1.30723 receptor type 4 (Ptpn4), mRNA 1454728_s_at ATPase, aminophospholipid Atp8a1 0.03424 1.29690 transporter (APLT), class I, type 8A, member 1 1438783_at Transmembrane, prostate androgen Tmepai 0.03438 1.38288 induced RNA, mRNA (cDNA clone IMAGE:5038092) 1454064_a_at ring finger protein 138 Rnf138 0.03458 1.25148 1440227_at expressed sequence BF642829 BF642829 0.03560 1.57072 1434806_at metaxin 3 Mtx3 0.03572 1.24583 1438980_x_at RIKEN cDNA 4732466D17 gene 4732466D17Rik 0.03579 1.31411 1454850_at TBC1 domain family, member 10c Tbc1d10c 0.03594 -1.28444 1456514_at RNA binding motif protein 16, Rbm16 0.03603 1.33620 mRNA (cDNA clone IMAGE:3591001)
329 1431131_s_at RIKEN cDNA A630007B06 gene A630007B06Ri 0.03612 1.29645 k 1419549_at arginase 1, liver Arg1 0.03616 -1.64404 1443949_at MKIAA4006 protein Ppp2r5e 0.03619 1.27576 1449411_at Down syndrome cell adhesion Dscam 0.03630 -1.31616 molecule 1416415_a_at H2A histone family, member Z H2afz 0.03639 -1.24172 1435697_a_at pleckstrin homology, Sec7 and Pscdbp 0.03654 -1.33985 coiled-coil domains, binding protein 1429764_at RIKEN cDNA 1500005K14 gene 1500005K14Rik 0.03662 1.53674 1456262_at RNA binding motif protein 5 Rbm5 0.03662 2.21789 1437155_a_at WW domain containing transcription Wwtr1 0.03675 1.57582 regulator 1 1448688_at podocalyxin-like Podxl 0.03677 1.26462 1427188_at ariadne ubiquitin-conjugating enzyme Arih1 0.03688 -1.36814 E2 binding protein homolog 1 (Drosophila) 1419149_at serine (or cysteine) peptidase Serpine1 0.03698 1.39139 inhibitor, clade E, member 1 1425216_at G protein-coupled receptor 43 Gpr43 0.03710 1.24094 1453188_at RIKEN cDNA 6230424C14 gene 6230424C14Rik 0.03726 1.27387 1449146_at Notch gene homolog 4 (Drosophila) Notch4 0.03749 1.34076 1449069_at zinc finger protein 148 Zfp148 0.03769 1.25921 1436043_at sodium channel, voltage-gated, type Scn7a 0.03776 1.33811 VII, alpha 1442254_at Transcribed locus --- 0.03805 1.31866 1420847_a_at fibroblast growth factor receptor 2 Fgfr2 0.03816 -1.24872 1434248_at protein kinase C, eta Prkch 0.03848 1.24300 1452179_at PHD finger protein 17 Phf17 0.03851 1.25598 1424970_at purine-rich element binding protein G Purg 0.03852 1.31911 1423091_a_at glycoprotein m6b Gpm6b 0.03870 1.32278 1434481_at RIKEN cDNA 4121402D02 gene 4121402D02Rik 0.03874 1.31851 1438069_a_at RNA binding motif protein 5 Rbm5 0.03891 2.02844 1439476_at desmoglein 2 Dsg2 0.03894 -1.40877 1432757_at RIKEN cDNA 2900011L18 gene 2900011L18Rik 0.03896 1.29083 1440379_at solute carrier family 1 (neutral amino Slc1a5 0.03934 1.81437 acid transporter), member 5 1451206_s_at pleckstrin homology, Sec7 and Pscdbp 0.03936 -1.26985 coiled-coil domains, binding protein 1448818_at wingless-related MMTV integration Wnt5a 0.03949 1.95637 site 5A 1434115_at ------0.03953 1.24823 1429367_at RIKEN cDNA 2510001I10 gene 2510001I10Rik 0.03957 1.25355 1460689_at DNA segment, Chr 15, Wayne State D15Wsu75e 0.03976 1.29545 University 75, expressed 1426951_at cysteine rich transmembrane BMP Crim1 0.03976 1.50435 regulator 1 (chordin like) 1451210_at phosphatidic acid phosphatase type Ppap2c 0.03982 -1.23231 2c 1434292_at RIKEN cDNA E130013N09 gene E130013N09Ri 0.03997 1.39380 k 1447899_x_at tumor-associated calcium signal Tacstd1 0.04009 1.23586 transducer 1 1453160_at thyroid hormone receptor associated Thrap1 0.04010 1.24363
330 protein 1 1448218_s_at tyrosine 3- Ywhaz 0.04023 1.25592 monooxygenase/tryptophan 5- monooxygenase activation protein, zeta polypeptide 1441425_at MAME mRNA for amelogenin Amelx 0.04027 1.27007 1437403_at RIKEN cDNA E130306M17 gene E130306M17Ri 0.04057 1.23119 k 1420747_at per-pentamer repeat gene MGI:1349458 0.04059 1.33124 1425603_at RIKEN cDNA 0610011I04 gene 0610011I04Rik 0.04081 1.23676 1433574_at cell division cycle 37 homolog (S. Cdc37l1 0.04086 1.25949 cerevisiae)-like 1 1431322_at immunoglobulin superfamily, Igsf3 0.04106 -1.31814 member 3 1454681_at RNA binding motif protein 35A Rbm35a 0.04111 -1.37354 1426500_at isoprenylcysteine carboxyl Icmt 0.04135 1.24389 methyltransferase 1441349_at hypothetical LOC552904 LOC552904 0.04136 1.36324 1425028_a_at tropomyosin 2, beta Tpm2 0.04137 -1.79283 1436893_a_at membrane-associated ring finger Mar-07 0.04156 1.28490 (C3HC4) 7 1427178_at transmembrane channel-like gene Tmc4 0.04183 -1.23090 family 4 1433184_at RIKEN cDNA 6720477C19 gene 6720477C19Rik 0.04184 1.26539 1453473_a_at t-complex testis expressed 1 Tctex1 0.04192 -1.31505 1435632_at RIKEN cDNA 1110001M19 gene 1110001M19Ri 0.04195 1.26772 k 1416835_s_at S-adenosylmethionine decarboxylase Amd1 /// Amd2 0.04197 -1.28069 1 /// S-adenosylmethionine /// LOC432454 decarboxylase 2 /// similar to S- adenosylmethionine decarboxylase 1434712_at expressed sequence AI452372 AI452372 0.04213 1.22846 1434678_at muscleblind-like 3 (Drosophila) Mbnl3 0.04258 1.31293 1438634_x_at LIM and SH3 protein 1 Lasp1 0.04281 -1.24657 1444503_at Glioblastoma amplified sequence Gbas 0.04281 1.28534 (Gbas), mRNA 1437716_x_at kinesin family member 22 Kif22 0.04287 -1.33808 1436101_at ring finger protein 24 Rnf24 0.04305 1.31194 1444828_at Protein phosphatase 2, regulatory Ppp2r5c 0.04322 1.26939 subunit B (B56), gamma isoform (Ppp2r5c), mRNA 1449089_at nuclear receptor interacting protein 1 Nrip1 0.04338 1.45975 1443624_at PREDICTED: similar to CDNA --- 0.04343 1.31827 sequence BC013481 [Mus musculus], mRNA sequence 1439970_at Protein kinase, lysine deficient 1, Prkwnk1 0.04349 1.25533 mRNA (cDNA clone IMAGE:5371931) 1457765_at Proteasome (prosome, macropain) Psmd14 0.04352 1.31561 26S subunit, non-ATPase, 14, mRNA (cDNA clone MGC:5840 IMAGE:3599671) 1457245_at disrupted in renal carcinoma 2 Dirc2 0.04353 1.25244 (human)
331 1435495_at adenosine A1 receptor Adora1 0.04437 1.30303 1434552_at WD repeat domain 77 Wdr77 0.04440 1.23347 1420481_at ------0.04447 1.23356 1421821_at low density lipoprotein receptor Ldlr 0.04452 1.29332 1423306_at RIKEN cDNA 2010002N04 gene 2010002N04Rik 0.04477 -1.43275 1423952_a_at keratin complex 2, basic, gene 7 Krt2-7 0.04486 -1.34866 1442151_at RIKEN cDNA D230040A04 gene D230040A04Ri 0.04488 1.38685 k 1416432_at 6-phosphofructo-2-kinase/fructose- Pfkfb3 0.04501 1.39106 2,6-biphosphatase 3 1449249_at protocadherin 7 Pcdh7 0.04502 1.26731 1431786_s_at RIKEN cDNA 1190003J15 gene 1190003J15Rik 0.04510 -1.39424 1418219_at interleukin 15 Il15 0.04518 1.23796 1425538_x_at CEA-related cell adhesion molecule 1 Ceacam1 0.04530 -1.31648 1452163_at E26 avian leukemia oncogene 1, 5' Ets1 0.04535 1.62962 domain 1446592_at RIKEN cDNA 3526402H21 gene 3526402H21Rik 0.04536 1.24694 (3526402H21Rik), mRNA 1429735_at RIKEN cDNA 1110003F05 gene 1110003F05Rik 0.04544 1.54297 1448254_at pleiotrophin Ptn 0.04545 -1.45281 1450494_x_at CEA-related cell adhesion molecule 1 Ceacam1 0.04547 -1.41905 1457534_at Ubiquitin associated protein 2-like Ubap2l 0.04571 1.24101 (Ubap2l), transcript variant 2, mRNA 1421964_at Notch gene homolog 3 (Drosophila) Notch3 0.04583 1.22938 1455434_a_at kinectin 1 Ktn1 0.04599 -1.28372 1458351_s_at kelch-like 2, Mayven (Drosophila) Klhl2 0.04606 1.38152 1439163_at zinc finger and BTB domain Zbtb16 0.04620 1.41929 containing 16 1454834_at nuclear factor I/B Nfib 0.04627 1.23474 1420444_at solute carrier family 22 (organic Slc22a3 0.04642 1.29282 cation transporter), member 3 1434724_at ubiquitin specific peptidase 31 Usp31 0.04667 1.24030 1438637_x_at splicing factor 3b, subunit 2 Sf3b2 0.04687 -1.22940 1419283_s_at tensin 1 Tns1 0.04729 1.66804 1426680_at selenoprotein N, 1 Sepn1 0.04758 1.26035 1437128_a_at RIKEN cDNA A630033E08 gene A630033E08Ri 0.04762 1.51377 k 1427199_at RIKEN cDNA 2510002A14 gene 2510002A14Rik 0.04763 1.35411 1454878_at RIKEN cDNA 2310047C04 gene 2310047C04Rik 0.04765 1.24260 1427262_at inactive X specific transcripts Xist 0.04779 2.13230 1438900_at SAC1 (supressor of actin mutations Sacm1l 0.04786 1.24623 1, homolog)-like (S. cerevisiae), mRNA (cDNA clone IMAGE:3486559) 1442025_a_at ------0.04820 1.25582 1427284_a_at tocopherol (alpha) transfer protein Ttpa 0.04821 -1.24863 1416359_at sorting nexin associated golgi protein Snag1 0.04821 1.32605 1 1457955_at Golgi associated, gamma adaptin ear Gga2 0.04821 1.24286 containing, ARF binding protein 2 (Gga2), mRNA 1418213_at keratin complex 1, acidic, gene 23 Krt1-23 0.04824 -1.34424 1457447_at Purinergic receptor P2Y, G-protein P2y5 0.04834 1.21774 coupled, 5, mRNA (cDNA clone
332 MGC:78367 IMAGE:30244233) 1427680_a_at nuclear factor I/B Nfib 0.04834 1.34760 1456686_at 0 day neonate eyeball cDNA, RIKEN --- 0.04843 1.52325 full-length enriched library, clone:E130119O22 product:unclassifiable, full insert sequence 1454254_s_at RIKEN cDNA 1600029D21 gene 1600029D21Rik 0.04849 -1.70951 1443185_at lipoma HMGIC fusion partner-like 2 Lhfpl2 0.04855 1.33544 1447808_s_at solute carrier family 15 (H+/peptide Slc15a2 0.04882 -1.36077 transporter), member 2 1457139_at ------0.04894 1.23246 1437577_at RIKEN cDNA 1110064P04 gene, 1110064P04Rik 0.04926 1.26972 mRNA (cDNA clone MGC:70098 IMAGE:30135289) 1439560_x_at hypothetical gene supported by LOC432995 /// 0.04949 -1.31445 BC047216 /// hypothetical gene LOC436333 supported by BC047216 1456112_at translocated promoter region Tpr 0.04950 2.20150 1439477_at ubiquitin-conjugating enzyme E2B, Ube2b 0.04956 1.51457 RAD6 homology (S. cerevisiae) 1459045_at golgi phosphoprotein 4 Golph4 0.04957 1.29311 1437001_at glycogen synthase kinase 3 beta Gsk3b 0.04958 1.77424 1418222_at RIKEN cDNA 2610024G14 gene 2610024G14Rik 0.04981 1.28163 Appendix XII. Differentially regulated genes in Prlr-/-/C3(1)/SV40T mammary epithelium. The 425 epithelial specific genes with differential expression between Prlr-/-/C3(1)/SV40T and WT/C3(1)/SV40T mammary transplants (p<0.05). Affymetrix ID, gene title and gene symbol is listed. The unadjusted p-value and fold change between Prlr-/-/C3(1)/SV40T and WT/C3(1)/SV40T mammary transplants is tabulated. Negative and positive values indicate decreasing and positive genes respectively.
333