Regulation of ABCG2 Expression in the Lactating Mammary Gland

by

Alex Man Lai Wu

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Graduate Department of Pharmacology and Toxicology University of Toronto

© Copyright by Alex Man Lai Wu (2014)

ABSTRACT

Regulation of ABCG2 Expression in the Lactating Mammary Gland

Alex Man Lai Wu, Doctor of Philosophy (2014)

Department of Pharmacology and Toxicology

University of Toronto

The Breast Cancer Resistance Protein (ABCG2) is a multidrug efflux transporter that is upregulated in certain drug-resistant cancer cells and in the mammary gland during lactation.

It is unclear how ABCG2 is regulated in the lactating mammary gland. This thesis sought to understand the role of prolactin (PRL) and its downstream signalling cascades, and epigenetic mechanisms, in the regulation of ABCG2 expression in the mammary gland during lactation.

Using T-47D human breast cancer cells, I showed that PRL upregulated ABCG2, in part, by inducing the recruitment of Signal Transducer and Activator of Transcription-5 (STAT5) to an interferon-γ activation sequence (GAS) motif in the human ABCG2 gene. Pharmacological inhibition of phosphoinositide 3-kinase and mitogen-activated protein kinase signalling further demonstrated the involvement of these pathways in the induction of ABCG2 by prolactin. To further investigate whether STAT5 regulates ABCG2 in vivo, I performed a series of experiments using mammary glands from non-lactating (virgin and forced involution) and lactating mice. These experiments revealed that the E1b alternative mouse Abcg2 mRNA isoform was predominantly expressed and induced in the mouse mammary gland during lactation. Using published STAT5 ChIP-seq datasets and ChIP-qPCR, I showed that STAT5 was bound to multiple regions along the mouse Abcg2 gene during lactation. In particular, one of these STAT5 binding regions, which contains a GAS motif, showed functional activity in luciferase reporter assays. I further investigated epigenetic mechanisms that may regulate

Abcg2 in the mouse mammary gland. Despite the presence of a CpG island at the E1b

ii

promoter, which when hypermethylated in vitro dramatically reduced promoter activity, the

E1b promoter was already hypomethylated in the virgin mammary gland and continued to be hypomethylated during lactation. Analysis of published ChIP-seq data revealed that the E1b promoter region in virgin mouse mammary epithelial cells was already enriched with the open chromatin histone mark H3K4me2 but that there was further accumulation of H3K4me2 during lactation. Collectively, these results suggest that ABCG2 is already poised for expression in the virgin mammary gland and that lactation-associated activation of STAT5, perhaps by prolactin, induces its expression.

iii

ACKNOWLEDGEMENTS I must thank a long list of people who have made it possible for me to endure both the challenges and triumphs of graduate studies:

I must first thank my mentor and supervisor Dr. Shinya Ito for providing me with the opportunitity to work with him. Dr. Ito was always open to new ideas, new experiments, some of which may have failed whereas others succeeded, but it was this openness that has afforded me the opportunity to develop my scientific curiosity – to ‘think outside the box’. I must also thank members of my PhD supervisory committee (Drs. Patricia A Harper, David S. Riddick, and Jason Matthews) for having the foresight at the first year and most criticial time of my PhD training that my thesis was a viable PhD project. I must thank each of them individually: Dr. Harper for going beyond the role of the committee member, often acting as a teacher, mentor, and above all, a listener of all the difficulties of graduate school and life-in-general; Dr. Riddick for being the one to introduce me to Professors at international conferences and for acting as ‘thesis reader’; and Dr. Matthews for generously allowing me to work in his lab for part of my PhD work.

I would also like to thank past and present members of the Ito Lab: most notably Hendrick Tan, Mingdong Yang, Pooja Dalvi, Liana Dedina, Reo Tanoshima, Tohru Kobayashi, Andrew Chuang, Marie-Caroline Delebecque, Hisaki Fujii, and John Leon-Cheon. Some of them have contributed scientifically to this thesis but all of them have made it fun to work in the Ito Lab. Other people who I must sincerely acknowledge for contributing to this thesis are: Dr. Andrei Turinsky (Centre for Computational Medicine, Hospital for Sick Children) for running the ChIP-Seq data pipeline/analysis; Kelvin Wang and Dr. Sean Egan (Developmental and Stem Cell Biology Program, Hospital for Sick Children) for helping with the virgin mammary epithelial cell isolation; Youliang Lou and Drs. Rosanna Weksberg, Darci Butcher, Sanaa Choufani (Genetics and Genome Biology Program, Hospital for Sick Children) for pyrosequencing services and discussing about epigenetics and data interpretation; Dr. Charles V Clevenger (Department of Pathology, Northwestern University) for acting as a resource to learn about PRL signalling and treatment protocols.

iv

Most importantly, I must thank my parents Henry and Lilian, siblings Cecilia and Brian, for being great listeners and providing often well-needed encouragement thoughout my studies. I must thank all my friends who have stayed by me, even when I’m too tired or not in the right mood to socialize.

Lastly, I must thank the taxpayers and others who still believe in the importance of training new PhD scientists. It is a constant reminder that graduate school is a privilege.

v

Table of Contents ABSTRACT ii ACKNOWLEDGEMENTS iv List of Figures x List of Tables xi List of Appendices xiii List of Abbreviations xiv 1. INTRODUCTION 1 1.1 Statement of Research Problem 1 1.2 ABCG2 expression and function 2 1.2.1 Discovery (BCRP/MXR/ABCP/ABCG2) 2 1.2.2 Expression, localization, and function 3 1.2.2.1 Placenta 4 1.2.2.2 Gastrointestinal tract 5 1.2.2.3 Liver 6 1.2.2.4 Kidney 7 1.2.2.5 Brain 8 1.2.2.6 Testis 9 1.2.2.7 Mammary Gland 10 1.2.2.8 Stem Cells 11 1.2.2.9 Cancer 13 1.2.2.10 ABCG2 Substrates and Inhibitors 14 1.3 Promoter Organization of the Human and Mouse ABCG2 gene 18 1.3.1 Human ABCG2 gene promoter 18 1.3.2 Mouse Abcg2 gene promoter 19 1.4 Regulatory control of ABCG2 expression 20 1.4.1 Hormones and inflammatory cytokines 20 1.4.1.1 Estrogen 20 1.4.1.2 Progesterone 21 1.4.1.3 Testosterone 21 1.4.1.4 Glucocorticoids 22 1.4.1.5 Proinflammatory Cytokines 23

vi

1.4.1.6 Other Growth factors 24 1.4.2 Stress and xenobiotics 27 1.4.2.1 Hypoxia 27 1.4.2.2 Aryl hydrocarbon receptor 27 1.4.2.3 Peroxisome-proliferator activated receptor 28 1.4.2.4 Pregnane X Receptor 28 1.4.2.5 Constitutive Androstane Receptor 29 1.4.2.6 Nuclear factor-erythroid 2 p45-related factor 2 29 1.4.2.7 Retinoic acid receptor/retinoid X receptor 29 1.4.2.8 Cyclooxygenase 2 30 1.4.2.9 Extracellular milieu: folate status 30 1.4.3 miRNA and Epigenetics 31 1.4.3.1 Micro-RNA (miRNA) 31 1.4.3.2 Promoter Methylation 32 1.4.3.3 Histone Modification 33 1.4.4 Cell signalling 34 1.4.4.1 PI3K/AKT signalling 34 1.4.4.2 MAPK signalling 36 1.4.5 Species 37 1.5 Mammary Gland Biology 39 1.5.1 General Anatomy 39 1.5.2 Development 39 1.5.2.1 Embryonic 39 1.5.2.2 Prepubertal 40 1.5.2.3 Puberty 40 1.5.2.4 Pregnancy 40 1.5.2.5 Lactation 41 1.5.2.6 Involution 42 1.6 Prolactin and Prolactin Receptors 44 1.6.1 The ligand 44 1.6.2 Prolactin receptor 44 1.6.3 Prolactin receptor signalling 45

vii

1.6.3.1 JAK2/STAT5 pathway 46 1.6.3.2 Mitogen Activated Kinase Pathway 47 1.6.3.3 PI3K/AKT pathway 48 1.6.3.4 Other pathways 48 1.7 Epigenetics 50 1.7.1 CpG Islands 50 1.7.2 Histone Modifications 51 1.8 Research Rationale 52 2. METHODS 54 2.1 Reagents 54 2.2 Animals 54 2.3 Cell Culture and serum starvation 55 2.4 Treatment with hormones or TCDD 56 2.5 Small molecule inhibitors of STAT5, MAPK and PI3K signalling 56 2.6 Isolation of mouse mammary epithelial cells (MEC) 56 2.7 RNA isolation from cells and tissues 58 2.8 cDNA synthesis - reverse transcription 58 2.9 Real-time RT-PCR to assess gene expression 59 2.10 Crude membrane fraction and whole cell lysate from human-derived cells 59 2.11 Preparation of mouse mammary gland tissue lysate 60 2.12 Immunodetection of protein by gel electrophoresis/western blot 60 2.13 Immunohistochemistry 61 2.14 Short-interfering RNA 62 2.15 Plasmid constructs 63 2.15.1 Activity of the human ABCG2 gene promoter and GAS element 63 2.15.2 Activity of the mouse E1b promoter and GAS elements 64 2.16 In vitro methylation of plasmid DNA 65 2.17 Transient transfection and luciferase assay 65 2.17.1 Human ABCG2 gene promoter and GAS element activity 65 2.17.2 Mouse Abcg2 gene promoter and GAS element(s) activity 66 2.18 Genomic DNA isolation, bisulfite treatment, and pyrosequencing 67 2.19 Chromatin Immunoprecipitation (ChIP) 67

viii

2.19.1 T-47D cells 67 2.19.2 Mouse mammary gland tissue 68 2.20 ChIP-seq data analysis 69 2.21 Statistical Analysis 70 3. RESULTS 76 3.1 AIM1: Effect of PRL and PRLR signalling cascades on ABCG2 expression 76 3.1.1 Effect of prolactin on ABCG2 expression 76 3.1.2 JAK2/STAT5-dependency in the induction of ABCG2 by prolactin 77 3.1.3 STAT5 recruitment to the human ABCG2 gene 78 3.1.4 Functional GAS motif in the ABCG2 gene 78 3.1.5 Effect of MAPK and PI3K pathway inhibitors on ABCG2 expression 79 3.1.6 Effect of MAPK and PI3K inhibitors on STAT5 recruitment 80 3.1.7 Growth hormone and ABCG2 expression 80 3.2 AIM2: Role of STAT5 in ABCG2 expression in vivo 95 3.2.1 Abcg2 isoform expression in the mouse lactating mammary gland 95 3.2.2 Effect of prolactin on Abcg2 expression 96 3.2.3 Forced weaning model of involution and STAT5 activity 97 3.2.4 Effect of forced weaning on Abcg2 expression in the mammary gland 98 3.2.5 Stat5 recruitment to the mouse Abcg2 gene in the mammary gland 98 3.2.6 Functional assessment of STAT5 binding regions for enhancer activity 99 3.3 AIM3: Epigenetic profile of the mouse Abcg2 gene 110 3.3.1 Characterization of a CpG island at the E1b promoter region 110 3.3.2 E1b promoter methylation status: in vivo and in vtro 111 3.3.3 Histone modifications at the mouse Abcg2 gene 111 4. DISCUSSION 116 4.1 PRL/PRLR signalling and ABCG2 expression in vitro 116 4.1.1 Prolactin induced ABCG2 in T-47D breast cancer cells 116 4.1.2 JAK2/STAT5 dependence in the prolactin-ABCG2 response 117 4.1.3 STAT5 recruitment to a proximal GAS element 119 4.1.4 Functional significance of the proximal GAS element 119 4.1.5 Growth hormone and regulation of ABCG2 expression 120 4.1.6 Regulation of ABCG2 by MAPK and PI3K signalling 122

ix

4.2 STAT5 and Abcg2 expression in vivo 123 4.2.1 Expression profiling of Abcg2 mRNA isoforms during lactation 123 4.2.2 HC11 and EpH4 cells as models of lactation 124 4.2.3 In vivo models to study STAT5 activation during lactation 125 4.2.4 STAT5 recruitment to the mouse Abcg2 gene during lactation 126 4.2.5 Functional GAS motifs in the intron region 127 4.2.6 Abcg2 expression after forced weaning 127 4.2.7 Sexually dimorphic expression of the E1b/c isoforms 129 4.2.8 E1b isoform is dynamically expressed 130 4.2.9 Activation of STAT5 by mechanisms other than the PRLR 130 4.3 AIM3. Epigenetic profile of the mouse Abcg2 gene 131 4.3.1 Modulation of E1b promoter activity via a CpG island 131 4.3.2 CpG sites as a marker for CpG island/promoter methylation 133 4.3.3 Histone modifications at the E1b promoter 133 4.4 Summary of findings and conclusions 134 4.5 Significance and broader implications 137 4.6 Recommendations for future study 139 5. APPENDICES 142 6. REFERENCES 161 7. LIST OF PUBLICATIONS, ABSTRACTS, AND COPYRIGHT STATEMENTS 201

List of Tables Table I. Examples of substrates transported by ABCG2 16 Table II. ABCG2 inhibitors 17 Table III. Effects of various stimuli and physiological states on ABCG2 expression 38 Table IV. Examples of histone modifications and their associated functions 51 Table V. Primer sequences for gene expression analyses by real-time RT-PCR 71 Table VI. Conditions for western analysis (human-derived samples) 72 Table VII. Conditions for western analysis (mouse-derived samples) 72 Table VIII. Sequence for primers and oligonucleotides used to construct reporter plasmids containing different regions of the human ABCG2 gene 73

x

Table IX. Sequence for primers and double stranded gene fragments used to construct reporter plasmids containing different regions of the mouse Abcg2 gene 74 Table X. Primers for ChIP analysis and Pyrosequencing 75 Table XI. Putative Stat5 binding sequences identified from published ChIP-seq data 107

List of Figures Figure 1. Schematic representation of major alternative untranslated first exons in the human and mouse Abcg2 gene 19 Figure 2. Regulatory mechanisms that modulate human ABCG2 expression 26 Figure 3. Schematic overview of postnatal mammary gland development 43 Figure 4. Major signalling pathways activated by the prolactin receptor 49 Figure 5. Prolactin induces ABCG2/BCRP mRNA and protein expression in T-47D breast cancer cells 81 Figure 6. Actinomycin D abolished prolactin-induced ABCG2 and CISH mRNA expression in T-47D cells 82 Figure 7. Effect of prolactin on ABCG2 expression in human breast cancer cell lines 83 Figure 8. Effect of prolactin on ABCG2 expression in primary human mammary epithelial cells (HMEC) 84 Figure 9. Prolactin receptor expression in human breast cancer cell lines and primary mammary epithelial cells 85 Figure 10. Knockdown of JAK2 attenuates prolactin-induced ABCG2/BCRP mRNA expression 86 Figure 11. Knockdown or pharmacological inhibition of STAT5A/B reduces the effect of prolactin on ABCG2/BCRP expression 87 Figure 12. Time-dependent recruitment of STAT5 to the 5’flanking region of the human ABCG2 gene after prolactin treatment 88 Figure 13. Knockdown or pharmacological inhibition of STAT5 attenuates prolactin-induced recruitment of STAT5 to the proximal GAS element of the ABCG2 gene 89 Figure 14. Prolactin induces ABCG2 promoter-driven luciferase reporter activity in T-47D cells 90

xi

Figure 15. Prolactin induces reporter activity in T-47D cells transfected with luciferase constructs containing tandem repeats of the proximal ABCG2 GAS element 91 Figure 16. Pharmacological inhibition of MAPK and PI3K signalling attenuates the induction of ABCG2 by prolactin 92 Figure 17. Prolactin-induced STAT5 recruitment to the ABCG2 proximal GAS element is not attenuated by inhibitors of MAPK and PI3K signalling 93 Figure 18. Growth hormone induces ABCG2 expression and STAT5 recruitment to the proximal GAS element of the ABCG2 gene in T-47D cells 94 Figure 19. The E1B Abcg2 mRNA isoform is the predominant isoform expressed in the mouse mammary gland during lactation 100 Figure 20. Sexually dimorphic expression of Abcg2 isoforms in the mouse liver 101 Figure 21. Prolactin does not induce Abcg2 mRNA in mammary epithelial cell lines 102 Figure 22. Loss of activated STAT5 in the mammary gland after forced weaning 103 Figure 23. Forced weaning induces rapid reduction in Abcg2 mRNA in the mouse mammary gland 104 Figure 24. ABCG2 protein remains highly expressed in the mammary gland of mice after forced weaning 105 Figure 25. Transcription factor (TF) binding sites in the region of the Abcg2 gene, based on ChIP-seq data from Yamaji et al. (2013) 106 Figure 26. STAT5 recruitment to the Abcg2 gene during lactation 108 Figure 27. Functional activity of STAT5-binding regions in the mouse Abcg2 gene 109 Figure 28. The -71/+199 region contains the minimal E1b promoter 112 Figure 29. Methylation of the mouse Abcg2 E1b promoter region reduces promoter activity 113 Figure 30. The E1b promoter in mouse mammary gland-derived samples are hypomethylated in vivo but partially methylated in vitro 114 Figure 31. Histone mark H3K4me2 in the region of the Abcg2 gene, based on ChIP-seq data from Rijnkels et al. (2013) 115 Figure 32. Framework of the molecular mechanisms governing ABCG2 expression in the lactating mammary gland 136

xii

List of Appendices Figure A1. Effect of AG490 on prolactin-induced expression of ABCG2 and CISH mRNA in T-47D cells 142 Figure A2. AG490 induced CYP1A1 mRNA expression in T-47D and HepG2 cells 143 Figure A3. STAT5 inhibitor reduced prolactin-induced STAT5 phosphorylation 144 Figure A4. A single mutation to the proximal ABCG2 GAS element significantly reduced prolactin-induced reporter activity 145 Figure A5. Prolactin does not induce Abcg2 mRNA in prolactin-responsive HC11 mouse mammary epithelial cells 146 Figure A6. Prolactin treatment failed to induce mammary gland gene expression in virgin FVB mice 147 Figure A7. Effect of differentiation on Abcg2 mRNA isoform and Csn2 mRNA expression in HC11 cells 148 Figure A8. Relative mRNA expression of drug and nutrient transporters in the mammary gland and liver of virgin and lactating C57BL/6 mice 149 Figure A9. Representative pyrogram of methylation levels at CpG sites within the mouse Abcg2 CpG island 150 Figure A10. Histone mark H3K4me3 in the region of the Abcg2 gene, based on ChIP-seq data from Yamaji et al. (2013) 151 Figure A11. Transcription factor (TF) binding sites in the region of the Abcb1a/Abcb1b genes, based on ChIP-seq data from Yamaji et al. (2013) 152 Figure A12. Transcription factor (TF) binding sites in the region of the Abcc1 gene, based on ChIP-seq data from Yamaji et al. (2013) 153 Figure A13. Transcription factor (TF) binding sites in the region of the Abcc2 gene, based on ChIP-seq data from Yamaji et al. (2013) 154 Figure A14. Transcription factor (TF) binding sites in the region of the Slc52a2 gene, based on ChIP-seq data from Yamaji et al. (2013) 155 Figure A15. Transcription factor (TF) binding sites in the region of the Slc52a3 gene, based on ChIP-seq data from Yamaji et al. (2013) 156 Figure A16. Histone mark H3K4me2 in the region of the Abcb1a/b genes, based on ChIP-seq data from Rijnkels et al. (2013) 157

xiii

Figure A17. Histone mark H3K4me2 in the region of the Abcc1 and Abcc2 genes, based on ChIP-seq data from Rijnkels et al. (2013) 158 Figure A18. Histone mark H3K4me2 in the region of the Slc52a2 and Slc52a3 genes, based on ChIP-seq data from Rijnkels et al. (2013) 159 Table AI. Regions of the mouse Abcg2 gene (chr. 6) with significant STAT5A recruitment 160

List of Abbreviations ABCB1, 1st member of the B subfamily of the ABC transporter superfamily (i.e. MDR1/P-gp) ABCC1, 1st member of the C subfamily of the ABC transporter superfamily (i.e. MRP1) ABCC2, 2nd member of the C subfamily of the ABC transporter superfamily (i.e. MRP2) ABCG2, 2nd member of the G subfamily of the ABC transporter superfamily (i.e. BCRP) AG490, JAK2 inhibitor 2-Cyano-3-(3,4-dihydroxyphenyl)-N-(benzyl)-2-propenamide, 2- Cyano-3-(3,4-dihydroxyphenyl)-N-(phenylmethyl)-2-propenamide AKT, protein kinase B BCRP, breast cancer resistance protein ChIP, chromatin immunoprecipitation CISH, cytokine-inducible SH2-containing protein CK-18, Krt18, cytokeratin-18 CP-671, 305, (+)-2-[4-({[2-(benzo[1,3] dioxol-5-yloxy)-pyridine-3-carbonyl]-amino)-methyl)- 3-fluoro-phenoxyl-propionic acid CpG, dinucleotide sequence of CG CSN2, β-casein – a breast milk protein DMEM, Dulbecco's modified Eagle's medium DMSO, dimethyl sulfoxide E3040, 6-hydroxy-5,7-dimethyl-2-methylamino-4-(3-pyridylmethyl) benzothiazole ERK1/2, extraceullar signal-regulated kinases 1 and 2 FBS, fetal bovine serum GAS, intereferon-gamma activated sequence motif GDC-0941, 2-(1H-indazol-4-yl)-6-(4-methanesulfonyl-piperazin-1- ylmethyl)-4-morpholin-4- yl-thieno[3,2-d]pyrimidine GF120918, elacridar (inhibitor of ABCB1 and ABCG2)

xiv

GH, growth hormone GSK3, Glycogen synthase kinase 3 IL, interleukin IQ, 2-Amino-3-methylimidazo[4,5-f]quinoline JAK, Janus Kinase (isoforms 1, 2, etc.) LY294002, 2-(4-Morpholinyl)-8-phenyl-1(4H)-benzopyran-4-one MAPK, mitogen-activated protein kinase MEC, mammary epithelial cells p50, nuclear factor of kappa light polypeptide gene enhancer in B-cells 1, component of NF-κB p53, tumor suppressor p65, RELA, reticuloendotheliosis viral oncogene homolog A, component of NF-κB PBS, phosphate buffered saline PD98059, 2-(2-Amino-3-methoxyphenyl)-4H-1-benzopyran-4-one PhIP, 2-Amino-1-methyl-6-phenylimidazo(4,5-b)pyridine PI3K, phosphoinositide 3-kinase poly (I:C), Polyinosinic:polycytidylic acid PRL, prolactin PRLR, prolactin receptor PTEN, Phosphatase and tensin homolog RPMI, Roswell Park Memorial Institute SFK, Src family of kinases SP, side population STAT, Signal Transducer and Activator of Transcription (-isoforms 1, 3, 5, etc.) TCDD, 2,3,7,8-Tetrachlorodibenzo-p-dioxin TiPARP, TCDD-inducible poly(ADP-ribose) polymerase Trp-P-1, 3-Amino-1,4-dimethyl-5H-pyrido[4,3-b]indole U0126, 1,4-diamino-2,3-dicyano-1,4-bis[2-aminophenylthio] butadiene WAP, whey acid protein – a breast milk protein XBP-1, X-box binding protein 1 YFP, yellow fluorescent protein

xv

1 1. INTRODUCTION 1.1 Statement of Research Problem The Breast Cancer Resistance Protein (BCRP or referred to in this this thesis by its gene name ABCG2) is a multidrug (substrate) efflux transporter that is traditionally associated with chemoresistance in certain cancer cells. However, ABCG2 is also highly expressed at the apical membrane of epithelial cells in normal tissues, particularly at important biological barriers such as the intestine, brain, and placenta, where it functions as a protective efflux transporter. For example, Abcg2/Bcrp1-null mice exhibit increased systemic exposure to dietary carcinogens (e.g. aflatoxin B1) (van Herwaarden et al., 2006), phototoxin (e.g. pheophorbide a) (Jonker et al., 2002), drugs (e.g. topotecan) and phytoestrogens (e.g. daidzein) (Enokizono et al., 2007a) compared to wild-type mice. These mice also demonstrate higher fetal-to-maternal plasma concentration ratio for genistein and topotecan (Enokizono et al., 2007). Contradictory to the role of ABCG2 as protective efflux transporter, a rather puzzling observation is that ABCG2 is dramatically upregulated in the mammary epithelium during lactation where it concentrates drugs and toxins into breast milk (Jonker et al., 2005), which is potentially harmful to the nursing offspring. While this apparent paradoxical role has since been clarified, in part, by more recent findings that ABCG2 is the major transporter of riboflavin into breast milk (van Herwaarden et al., 2007), it is not clear what induces the expression of ABCG2 during lactation. This thesis aims to fill this knowledge gap with the ultimate goal that by having a better understanding of how ABCG2 is regulated during lactation, we may be able to counter unwanted ABCG2-mediated drug efflux in certain pathophysiological conditions.

2 1.2 ABCG2 Expression and Function

1.2.1 Discovery (BCRP/MXR/ABCP/ABCG2) The discovery of ABCG2 was preceded by two very notable and important discoveries in drug transporter biology. In 1976, Juliano and Ling identified a plasma membrane glycoprotein with an apparent molecular weight of 170 kDa that was overexpressed in drug resistant Chinese Hamster Ovary (CHO) cells (Juliano and Ling, 1976). This protein, designated as P-glycoprotein (P-gp), was also overexpressed in other drug resistant mammalian cells (mouse, human) (Kartner et al., 1983). The cDNA clone that encoded P-gp was eventually identified from drug-resistant CHO cells (Riordan et al., 1985). This was followed subsequently by cloning of the human multidrug resistance (MDR1) gene that corresponds to P-gp (Roninson et al., 1986). The discovery of P-gp, while crucial, did not provide a complete explanation for drug resistance. Cole et al. noted that multidrug resistant H69AR cells, a subline of NCI-H69 small cell lung carconima cells that were selected for resistance to doxorubicin, did not express P-gp and nor were these cells effectively sensitized by co-treating with P-gp inhibitors (Cole et al., 1989; Mirski et al., 1987). By screening cDNA libraries, Cole et al. (1992) identified a cDNA clone that was overexpressed in H69AR cells. Based on the deduced amino acid sequence, it was predicted that this protein was a member of the ATP- binding cassette (ABC) transporter family. This protein, designated as multidrug resistance- associated protein (MRP), became the second transporter identified to confer drug resistance. Both of these discoveries, along with advancements in molecular and computational biology, were fundamental to the discovery of ABCG2 by three independent laboratories. Doyle et al. (1998) reasoned that since P-gp and MRP were not upregulated in drug resistant MCF-7/AdrVp cells, which were derived by selecting MCF-7 cells for resistance to doxorubicin in the presence of the P-glycoprotein inhibitor verapamil, that alternative drug resistant mechanisms must exist. Indeed, using RNA fingerprinting, these authors observed a PCR product/cDNA clone that was much more abundant in drug resistant MCF-7/AdrVp cells compared to the partial revertant (resistant) and parental MCF-7 cells. Furthermore, MCF-7 cells transfected with this cDNA showed reduced intracellular accumulation of daunorubucin and increased resistance to mitoxantrone, daunorubucin and doxorubicin. The protein encoded by this cDNA sequence was termed Breast Cancer Resistance Protein (BCRP). Using a different experimental approach, Miyake et al. (1999) identified two cDNA clones which were overexpressed in mitoxantrone-resistant S1-M1-80 human colon carcinoma

3 cells compared to the parent S1 cell line. These clones, called MXR1 and MXR2 for mitoxantrone resistance, showed a high degree of similarity in their sequence. Northern analyses using MXR1 cDNA further revealed that it was overexpressed in MCF7-cells that were selected for resistance to doxorubicin. The third laboratory that discovered ABCG2 took a more computational approach. In 1996, Allikmets et al. characterized 21 new ABC genes by searching the expressed sequence tag database for sequences that showed similarities to P-gp and cystic fibrosis transmembrane conduction regulator (CFTR) (Allikmets et al., 1996). This analysis eventually led them to identify a cDNA clone that was expressed at very high level in the placenta relative to other tissues (Allikmets et al., 1998). For this reason, the authors called this gene ABCP (P for placenta). This gene was mapped to human chromosome 4q22. The murine Abcp was also identified and was mapped to mouse chromosome 6. The deduced amino acid sequences and predicted protein structures described by these three laboratories showed striking similarities: presence of several transmembrane regions, an ATP binding domain, and an overall structure consistent with a ‘half-transporter’ (i.e., containing half the number of domains as the prototypical drug transporter P-glycoprotein) (Allikmets et al., 1998; Doyle et al., 1998; Miyake et al., 1999). Indeed, all three laboratories had characterized the same gene. Based on the amino acid sequence, BCRP/MXR/ABCP is formally classified as the second member of the G subfamily of the ATP-binding cassette (ABC) transporter superfamily, and will be referred to in this thesis as ABCG2. It is beyond the scope of this thesis to discuss the work that empirically characterized the structure of ABCG2, but many of the predictions made by the authors that discovered ABCG2 hold true. ABCG2 has 6 transmembrane domains and one nucleotide (ATP) binding domain, and functions as a dimer or a higher order oligomer [extensively reviewed by McDevitt et al. 2009].

1.2.2 Expression, localization, and function Although ABCG2 is expressed at variable levels in almost all normal tissues, this section will focus on tissues that function as biological barriers or as routes of elimination. The expression of ABCG2 in cancer and cancer stem cells will also be briefly reviewed.

4 1.2.2.1 Placenta Expression. Allikmets et al. (1998) noted at the time of discovery that ABCG2 mRNA was most highly expressed in the human placenta relative to other normal tissue/organs examined. Various studies have confirmed this, and have extended this observation to the protein level. Specifically, ABCG2 in the human placenta is mainly localized to the apical membrane of syncytiotrophoblasts (Maliepaard et al., 2001), although cytoplasmic localization has also been observed (Fetsch et al., 2006); ABCG2 may be present in some fetal blood vessels (Yeboah et al., 2006). It is interesting to note that unlike in human, the rodent placenta only moderately expresses Abcg2 mRNA - the highest expression is observed in kidney (Jonker et al., 2000; Tanaka et al., 2005). There is some evidence that placental expression of ABCG2 may depend on gestational age. Yeboah et al. observed that ABCG2 protein expression was highest in the human placenta at term (38-41weeks) compared to earlier gestational age (<35weeks), but others did not observe this (Mathias et al., 2005). In contrast, mouse placental Abcg2 mRNA (Cygalova et al., 2008) and protein (Wang et al., 2006a) expression was highest at mid-late gestation (day 15). Function. The importance of placental ABCG2 to fetal protection was first proposed by Allikmets et al., but Jonker and colleagues were the first to demonstrate this. Using pregnant Abcb1a/b (P-gp) knockout mice, Jonker et al. (2000) showed that fetus from mothers treated with the ABCB1/ABCG2 inhibitor GF120918 had significantly higher exposure to topotecan compared to vehicle treated mothers. Fetal protection by placental ABCG2 was further demonstrated by the observation that there is increased fetal penetration of glyburide (Zhou et al., 2008), nitrofurantoin (Zhang et al., 2007a), and bile acids (Blazquez et al., 2012) in Abcg2 knockout mice compared to wild-type mice. A number of factors, however, may compromise the ability of ABCG2 to protect the fetus. Systemic inflammation elicited by bacterial endotoxin lipopolysaccharide has been shown to reduce Abcg2 expression in pregnant rats, which ultimately increased glyburide penetration into fetal tissue (Petrovic et al., 2008). Non-synonymous allelic polymorphisms such as C421A that substitutes glutamine 141 to lysine, may also affect ABCG2-mediated fetal protection. Placental tissue that is homozygous for A421 expresses significantly less ABCG2 protein (but not mRNA) compared to homozygous C421 (Kobayashi et al., 2005). Given that the A421 variant also has reduced ability to transport glyburide in HEK293 expression systems (Pollex et al., 2010), one may

5 infer that individuals homozygous for this polymorphism may have a compromised blood- placental barrier.

1.2.2.2 Gastrointestinal tract Expression. ABCG2 is expressed in the small intestine, colon, and at much more limited levels, in the stomach. In human, ABCG2 is mainly localized to the apical membrane of epithelial cells in the small intestine and colon, although there is some cytoplasmic localization (Fetsch et al., 2006; Maliepaard et al., 2001). Quantitatively, ABCG2 is one of the highest expressed multidrug transporters in the intestine (Tucker et al., 2012). The expression of ABCG2 in the human intestine shows large interindividual variability, which when evaluated at the mRNA level is not associated with any particular polymorphism in the coding sequence of the gene (Urquhart et al., 2008; Zamber et al., 2003). Despite the apparent lack of difference at the mRNA level, patients carrying one allele of the 421C>A polymorphism show increased bioavailability of the ABCG2 substrate sulfasalazine, which may be related to reduced surface expression of this genotype (Urquhart et al., 2008). Single nucleotide polymorphisms at non-coding regions of the ABCG2 gene are associated with different levels of ABCG2 mRNA. Poonkuzhali et al. showed that carriers of -15994C>T have reduced intestinal expression of ABCG2 mRNA, while -15622C>T and 1143G>A was associated with increased expression (Poonkuzhali et al., 2008). Interestingly, there are segmental differences in ABCG2 expression along the intestinal tract. In humans, ABCG2 is expressed at similar levels along different segments of the small intestine (i.e. duondenum, jejunum, and ileum), but the colon expresses less ABCG2 compared to the small intestine (Gutmann et al., 2005; Hilgendorf et al., 2007). In rats and mice, where more detailed analyses have been performed, Abcg2 shows a pattern of increasing expression moving towards the terminal small intestine that peaks at the ileum (Enokizono et al., 2007a; Hosomi et al., 2012; MacLean et al., 2008; Tanaka et al., 2005). Function. At the intestinal tract, ABCG2 serves as an important biological barrier that limits the absorption or re-uptake of drugs and toxins in our diet. This was first demonstrated by the observations that oral bioavailability of topotecan was significantly increased in mice (Allen et al., 1999; Jonker et al., 2000) and humans (Kruijtzer et al., 2002) treated with an ABCG2 inhibitor. More recent studies using Abcg2 knockout mice in both in vivo and ex vivo perfused intestine models have extended these findings to other xenobiotics such as

6 nitrofurantoin (Merino et al., 2005a) and ciprofloxacin (Merino et al., 2006), the phytoestrogen genistein (Zhu et al., 2010), toxins/carcinogens PhIP (van Herwaarden et al., 2003), aflatoxin B1, IQ, and Trp-P-1 (van Herwaarden et al., 2006), and sulfate and glucoronide conjugated 4- methyl-umbelliferone and E3040 (Adachi et al., 2005).

1.2.2.3 Liver Expression. In human liver, ABCG2 is mainly expressed in the bile canalicular membrane of hepatocytes (Fetsch et al., 2006; Maliepaard et al., 2001). However, basolateral localization has also been observed but its role is not clear (Vander Borght et al., 2006). Unlike in the intestine where ABCG2 is one of the predominantly expressed multidrug efflux transporter, ABCG2 is expressed at lower levels compared to ABCC2 and ABCB1 in the liver (Tucker et al., 2012). Interestingly, in the fetal mouse liver two days prior to birth, Abcg2 accounts for one quarter of total mRNA expression of 62 critical efflux and uptake drug transporters (Cui et al., 2012). Right after delivery, this level is reduced to 4% and then plateaus to approximately 2-3% in adulthood. Function. Despite this relatively low expression, ABCG2 plays a very important role in the biliary elimination of certain xenobiotics. This is supported by the observations that Abcg2 knockout mice show significantly reduced biliary excretion of drugs and toxins such as PhIP (Jonker et al., 2002), nitrofurantoin (Merino et al., 2005a), fluoroquinolones (Ando et al., 2007), rosuvastatin (Kitamura et al., 2008), pitavastatin (Hirano et al., 2005), and troglitazone (Enokizono et al., 2007b) compared to wildtype mice. It is interesting that under normal physiological conditions, sex alone can dramatically influence drug disposition mediated by ABCG2. In mouse liver, Abcg2 mRNA and protein are more highly expressed in male compared to female (Merino et al., 2005b; Tanaka et al., 2005), and this difference persists well into late adulthood (27 months) (Fu et al., 2012). This is associated with increased plasma concentration of cimetidine, topotecan, PhIP, and nitrofurantoin when administered intravenously to female mice compared to male mice (Merino et al., 2005b). Female mice also showed reduced biliary excretion of nitrofurantoin (Merino et al., 2005b), and glucoronide- and sulfate-conjugated acetaminophen (Lee et al., 2009). These differences were not observed in Abcg2 knockout mice. It is not clear what is responsible for sexually dimorphic expression of ABCG2. However, since castration reduced Abcg2 expression in the liver of male mice, which was partially rescued by administration of 5α-dihydroxytestosterone, it was suggested

7 that testosterone may be involved in male-predominant expression of Abcg2 in the liver (Tanaka et al., 2005). It is important to note that male Stat5b knockout mice show reduced liver expression of Abcg2 compared to their wildtype counterparts (Clodfelter et al., 2006). This may suggest a possible role for growth hormone, a major activator of Stat5b, in the expression of Abcg2 in the liver. There are conflicting reports on whether ABCG2 expression in the human liver is influenced by sex (Merino et al., 2005b; Prasad et al., 2013). Therefore, more evidence will be required to determine whether male-predominant expression of ABCG2 is species-specific. In addition to sex, systemic inflammation induced by poly (I:C) in rats (Petrovic et al., 2008) or a model of malarial infection in pregnant mice (Cressman et al., 2014) downregulated Abcg2 mRNA expression in the liver. Surprisingly, ABCG2 expression was upregulated in human livers injured by primary biliary cirrhosis or by acetaminophen (Barnes et al., 2007).

1.2.2.4 Kidney Expression. In general, ABCG2 is localized to the apical membrane of tubular structures in the renal cortex, particularly at the proximal tubules (Fetsch et al., 2006; Huls et al., 2006, 2008). However, ABCG2 in human kidney is not always detectable by immunohistochemistry (Maliepaard et al., 2001), which may depend on the antibody used. Function. In the kidney, ABCG2 has been shown to participate in the excretion and eventual elimination of certain drugs/toxins. For example, Abcg2 knockout mice show reduced renal elimination of sulfate-conjugated edaravone (Mizuno et al., 2007) and E3040 (Mizuno et al., 2004), ciprofloxacin and grepafloxacin (Ando et al., 2007). Furthermore, loss of Abcg2 expression resulted in increased accumulation of carcinogens IQ and aflatoxin B1 in the kidney of Abcg2 knockout mice (van Herwaarden et al., 2006). Perhaps one of the most intriguing findings in the past several years was the observation that several genetic polymorphisms at the human ABCG2 gene are associated with serum uric acid levels and gout (Dehghan et al., 2008; Yang et al., 2010). Indeed, several studies have shown that ABCG2 is a urate transporter and that nonsynonymous mutations that alter ABCG2 activity are also associated with increased serum uric acid levels (Nakayama et al., 2011; Woodward et al., 2009). Despite the observation that ABCG2 is a urate transporter, a recent study suggests that it likely plays a minor role in the kidney and that its activity in the intestine as a mechanism for extra-renal elimination of uric acid may be more physiologically relevant. Abcg2 knockout

8 mice, which have elevated serum uric acid levels, actually show increased renal urate excretion but significantly reduced urate excretion via the intestine (Ichida et al., 2012). It is important to note that ABCG2 is only moderately expressed in the human kidney relative to the liver (Fetsch et al., 2006; Maliepaard et al., 2001), but in rodents, expression of ABCG2 in the kidney exceeds that of the liver (Jonker et al., 2000; Tanaka et al., 2005; Zhou et al., 2002). For these reasons, one must be cautious when relating results obtained from rodent models to humans as there may be an overestimation of the relative importance of the kidney in the elimination of ABCG2 substrates. Further complicating the interpretation of the above data is that the expression of various ABC transporters is altered in the kidney of Abcg2 knockout mice (Huls et al., 2008).

1.2.2.5 Brain Expression. ABCG2 is localized to the luminal membrane of endothelial cells in brain microvessels, a phenomen that is conserved at least in several species including human (Aronica et al., 2005; Cooray et al., 2002; Fetsch et al., 2006; Maliepaard et al., 2001; Sisodiya et al., 2003; Zhang et al., 2003b), mouse (Lee et al., 2005; Tachikawa et al., 2005), and rat (Hori et al., 2004). In isolated human brain microvessels, ABCG2 mRNA represented up to 80% of the total mRNA of ten ABC drug transporters (Dauchy et al., 2008). In mice, Abcg2 protein has also been shown to be present in the epithelial cells of the choroid plexus (Tachikawa et al., 2005; Zhuang et al., 2006) and arachnoid barrier (Yasuda et al., 2013). However, ABCG2 was not detected in the human choroid plexus (Daood et al., 2008) and no information is available regarding its expression in the human arachnoid barrier. As such, there is a dearth of information regarding the role of ABCG2 in the blood-to-cerebral spinal fluid (CSF) barrier. Function. ABCG2 is well recognized as an important component of the blood brain barrier (BBB) that limits the distribution of drugs and toxins into the brain. Using Abcg2 knockout mice, it was demonstrated that loss of Abcg2 expression at the BBB increased brain distribution of phytoestrogens (daidzein, genistein, and coumestrol) (Enokizono et al., 2007a), MeIQx, PhIP, dantrolene, prazosin, triamterene (Enokizono et al., 2008), imatinib (Breedveld et al., 2005), flavopiridol (Zhou et al., 2009), sorafenib (Lagas et al., 2010) and erlotinib (Kodaira et al., 2010). While the Abcg2 knockout mice were tremendously useful, the presence of Abcb1a/b in the BBB made it difficult to demonstrate the functional importance of

9 ABCG2 in the BBB for substrates that were co-transported by Abcb1a/b. Indeed, the brain penetration of ABCG2 substrates such as topotecan (de Vries et al., 2007), lapatanib (Polli et al., 2009), dasatinib (Lagas et al., 2009), mitoxantrone (Kodaira et al., 2010), sunitinib (Tang et al., 2012) and vemurafenib (Mittapalli et al., 2012), which show either a lack of, or modest effect, in Abcg2 or Abcb1a/b knockout mice, were dramatically increased in triple (Abcg2/Abcb1a/b) knockout mice. Although ABCG2 has long been considered an important drug transporter at the BBB, until recently it was not known what role it plays physiologically. Xiong et al. found that Alzheimer Disease (AD) patients with cerebral amyloid angiopathy and transgenic mouse models of AD had increased expression of ABCG2 protein in the brain (Xiong et al., 2009). Furthermore, Abcg2 knockout mice showed greater accumulation of amyloid β (aβ) peptide in the brain after an intravenous administration of the peptide compared to wildtype mice. In addition, amyloid β peptide increased the accumulation of Hoescht 33342 (ABCG2 substrate) in HEK293 cells overexpressing ABCG2. These results along with the more recent observation that overexpression of ABCG2 increased the efflux of aβ peptide out of HEK293 cells (Do et al., 2012) suggest that ABCG2 may prevent circulating amyloid β from accumulating in the brain.

1.2.2.6 Testis Expression. There are conflicting reports regarding the specific cellular localization of ABCG2 in the testes. Maliepaard et al. (2001) did not detect any ABCG2 protein in the normal human testis by immunohistochemistry but Fetsch et al. (2006) observed ABCG2 localization to Sertoli and Leydig cells. Further contrasting these reports is the observation that ABCG2 was expressed at the luminal membrane of endothelial cells and the apical side of myoid cells but not in Sertoli cells (Bart et al., 2004). It is not clear whether these discrepancies are due to the use of different antibodies to immunodetect ABCG2, but this appears unlikely since the antibodies described have been extensively validated. In mouse testis, it was demonstrated that ABCG2 was localized to the luminal membrane of endothelial cells (Enokizono et al., 2007a). In the rat testis, ABCG2 is localized to endothelial cells of microvessels and also to the myoid cells (Qian et al., 2013) , which is consistent with at least one report in humans. In mouse and cow, ABCG2 was also localized to the acrosomal region of sperm cells (Caballero et al., 2012; Scharenberg et al., 2009).

10 Function. While it is clear that more studies will be required to clarify the localization of ABCG2 in the testis, there is a large body of evidence that ABCG2 functions to protect the testis from harmful drugs and toxins in the blood. This was first demonstrated by the observation that following an intravenous dose of PhIP, Abcg2 knockout mice showed a 2.6- fold higher accumulation of PhIP in the testis compared to wildtype mice (van Herwaarden et al., 2003). Subsequent studies further showed that Abcg2 knockout mice had increased testicular distribution for ABCG2 substrates such as phytoestrogens (Enokizono et al., 2007a), MeIQx, PhIP metabolites, dantrolene, prazosin (Enokizono et al., 2008), erlotinib, flavopiridol, and mitoxantrone (Kodaira et al., 2010) compared to wildtype mice.

1.2.2.7 Mammary Gland Expression. Unlike other organs in which ABCG2 is usually constitutively expressed, the expression of ABCG2 in the mammary gland is largely dependent on the stage of mammary gland development. Jonker et al. (2005) showed that the expression of ABCG2 protein is induced in the mouse mammary gland during late pregnancy, peaks at lactation, and quickly returns to levels in the virgin gland as quickly as one week after the cessation of lactation (i.e. involution). This upregulated expression of ABCG2 during lactation was also observed in human and cow. It is not clear whether this induction is purely at the translational level as transcript expression was not assessed. Others have reported a 30-fold increase in transcript expression of ABCG2 in the bovine mammary gland from late gestation to early-mid lactation (60 day post parturition) (Bionaz and Loor, 2008). Function. In the mammary gland, ABCG2 is localized to the luminal membrane of milk producing alveolar epithelial cells where it participates in the secretion of drugs and toxins into breast milk. This is supported by the observations that Abcg2 knockout mice showed significantly reduced distribution of ABCG2 substrates such as PhIP, topotecan, cimetidine, acyclovir (Jonker et al., 2005), nitrofurantoin (Merino et al., 2005a), ciprofloxacin (Merino et al., 2006), IQ, Trp-P-1, and aflatoxin-B1 (van Herwaarden et al., 2006) into milk compared to wildtype mice. In the case of nitrofurantoin, this mechanism for ‘drug elimination’ may even surpass that of the liver. Noteworthy, other drug transporters (ABCB1, ABCC1 and ABCC2) are either downregulated or not expressed in the lactating mouse mammary gland, which further highlights the relative importance of ABCG2 (Jonker et al., 2005). There is not yet an entirely clear understanding of the biological role of ABCG2 in the

11 mammary gland during lactation since the secretion of harmful drugs and toxins into breast milk could potentially harm the nursing offspring/infant. However, it is likely that these chemicals simply hijack ABCG2, which normally serves a physiological role. Indeed, Abcg2 knockout mice show a 63-fold reduction in riboflavin levels in breast milk compared to wildtype mice, which suggests that ABCG2 is a major transporter of riboflavin into breast milk (van Herwaarden et al., 2007). Interestingly, there are no reported developmental abnormalities or deficiencies in pups nurtured by Abcg2 knockout mice. This is believed to be due to high levels of flavin adenine dinucleotide that are still present in the breast milk. While the biological reason for lactation-associated upregulation of ABCG2 is partially answered, it is still not clear ‘how’ ABCG2 is upregulated during lactation.

1.2.2.8 Stem Cells Expression. Prior to the discovery of ABCG2, it was thought that ABCB1 (or P- glycoprotein) was responsible for the Hoescht 33342 dye exclusion phenotype that characterizes side population (SP) cells which are enriched with stem/progenitor cells (Bunting et al., 2000). However, Zhou et al., (2001) noted that since bone marrow from Abcb1a/b knockout mice and wildtype mice contained similar numbers of SP cells, that an alternative transporter was responsible for dye exclusion. Further examination revealed that SP cells express high levels of ABCG2 and that overexpression of ABCG2 in mouse bone marrow cells resulted in expansion of SP cells (Zhou et al., 2001). Given that ABCG2 is able to actively remove Hoescht dye from the cell, it not surprising that SP cells isolated from various tissues such as pancreas (Lechner et al., 2002), lung (Summer et al., 2003), heart (Martin et al., 2004) and muscle (Tanaka et al., 2009) express ABCG2. The fundamental questions, however, are: (1) is ABCG2 a good stem cell marker and (2) what is the functional significance of ABCG2 in stem/progenitor cells? The fact that many differentiated epithelial cells express ABCG2 implies that SP cells likely represent a very heterogenous cell population, which for some tissues may mean that only a small fraction of SP cells are true stem/progenitor cells. It is interesting that despite the near complete loss of SP cells in Abcg2 knockout mice, the remaining SP cells lacked stem-like properties. For example, SP cells isolated from bone marrow of Abcg2 knockout mice did not express other stem cell markers (Zhou et al., 2002). Additionally, cardiac SP cells from Abcg2 knockout mice show impaired cell cycle progression and proliferation potential (Pfister et al., 2008; Sereti et al., 2013). These

12 observations may suggest that only SP cells that express ABCG2 are enriched with stem cells and that true stem/progenitor cells express ABCG2. The latter was elegantly demonstrated recently using a transgenic mouse model that contained a stretch of DNA encoding tamoxifen inducible cre recombinase inserted downstream of the mouse Abcg2 gene, which results in a bicistronic mRNA that translates both Abcg2 and cre recombinase protein (Fatima et al., 2012). By crossing these mice with transgenic mice that would express YFP or LacZ when cre recombinase is expressed, Fatima et al., was able to trace which cells were derived from cells that originally expressed Abcg2. These authors found that at 12 months after tamoxifen treatment, multiple hematopoietic cells from different lineages were marked with YFP. Furthermore, even at 20 months after tamoxifen treatment, some villi in the intestine and seminiferous tubules in the testes were labelled with LacZ. Considering the quick turnover of differentiated cells that express Abcg2 in these tissues, cells which are still labelled >12 months after tamoxifen treatment likely represent cells that were derived from progenitors that expressed Abcg2 at the time of tamoxifen treatment. Collectively, a body of evidence suggests that stem cells likely express ABCG2, but given that so many differentiated cells also express ABCG2, it is not an appropriate stem cell marker. Function. The ultimate question to address is whether ABCG2 serves a biological role in stem cells. The answer to this question may be tissue and state specific. When bone marrow cells were sorted using stem cell markers (c-Kit+, Sca-1+, Lineage-) instead of Hoescht dye exclusion, Zhou et al., found no difference in the number of stem cells between Abcg2 knockout and wildtype mice, which implies that ABCG2 is simply a marker that serves no role in maintaining stem cells (Zhou et al., 2002). Further supporting this conclusion was the lack of abnormalities with respect to blood count in Abcg2 knockout mice. In contrast, others have noted that ABCG2 may play a role in maintaining stem cells. Bhattacharya et al. demonstrated that forced overexpression of Abcg2 in mouse retinal neurospheres isolated from embryonic day 18 increased the number of proliferating cells that expressed stem cell markers and reduced the expression of differentiated neural/glial cell markers (Bhattacharya et al., 2007). Neurosphere SP cells overexpressing Abcg2 also displayed increased expression of the mRNAs that code for proteins involved in cell proliferation. In cardiac SP cells, lentiviral based shRNA mediated silencing of Abcg2 resulted in reduced expression of cell cycle proteins and increased expression of negative regulators of cell cycle, with a shift towards favouring asymmetric cell division (Sereti et al., 2013). It is not clear how ABCG2 may help

13 maintain ‘stemness’, but ABCG2 may be important for the removal of harmful endo- biotics/toxins that modulate stem cell activity. For example, pharmacological inhibition or shRNA-mediated downregulation of Abcg2 in mouse embryonic stem cells resulted in a modest increase in intracellular protoporphyrin (PPIX) accumulation, the effect of which was even more dramatic when the PPIX precursor aminolevulinic acid was given (Susanto et al., 2008). This increase in PPIX is associated with increased ROS production and double strand breaks that activate p53, which transcriptionally repress nanog expression and ultimately, loss of stem-like phenotype. Therefore, ABCG2 expression in stem cells may be crucial for protecting stem cells from harmful xeno- and endo-biotics. This may be particularly important in cancer chemotherapy in which cancer stem cells that have high expression of ABCG2 are able to resist treatment and eventually contribute to relapse.

1.2.2.9 Cancer Solid tumours. Despite the large body of experimental evidence that ABCG2 can dramatically alter cellular resistance to certain chemotherapeutic agents in vitro, there is by comparison much less knowledge of the importance of ABCG2 in chemotherapy resistance in the clinical setting. ABCG2 is expressed at varying degrees in a variety of solid cancers (Diestra et al., 2002). In the majority of solid tumours, such as small (Kim et al., 2009b) and non-small cell lung carcinoma (Ota et al., 2009; Yoh et al., 2004), esophageal squamous cell carcinoma (Hang et al., 2012; Tsunoda et al., 2006), laryngeal squamous cell carcinoma (Shen et al., 2011), pancreatic ductal adenocarcinoma (Lee et al., 2012), and colorectal cancer (Wang et al., 2013), ABCG2 expression is correlated with a shorter period of disease-free survival and/or overall survival. Others did not find any relationship between ABCG2 expression and clinical outcomes (i.e. survival) in non-small cell lung carcinoma (Herpel et al., 2011; Li et al., 2009, 2010). These studies mainly focused on the utility of ABCG2 as a prognostic marker for survival, therefore they fail to address the crucial question of whether ABCG2 expression is correlated with response rate to chemotherapy. Of the two known studies on solid tumours that have addressed this in detail, the results are contradictory. Faneyte et al. did not observe any difference in ABCG2 mRNA expression between responders and non-responders of chemotherapy in patients with breast cancer (Faneyte et al., 2002). In contrast, Burger et al. found that non-responders had a significantly higher median ABCG2 mRNA expression compared to responders, but that response rate to chemotherapy was not

14 significantly different between patients with high versus low ABCG2 expression (Burger et al., 2003). Importantly, after stratifying patients based on treatment, Burger et al. (2003) observed a trend toward a significantly lower response rate to combination treatments consisting of doxorubicin/epirubicin in breast cancer patients with high ABCG2 expression. This demonstrates that it is important to consider the specific treatment regimen when evaluating the significance of ABCG2 expression in response to chemotherapy. Liquid tumours. Unlike solid tumours, the role of ABCG2 expression in response rate to chemotherapy is well studied in hematological malignancies. Acute myeloid leukemia (AML) patients with high expression of ABCG2 in leukemic cells showed a lower rate of remission after induction therapy (Benderra et al., 2004; Steinbach et al., 2002) and significantly higher relapse rate (Damiani et al., 2010). Interestingly, ABCG2 mRNA expression was significantly higher in AML cells isolated after relapse versus cells isolated at diagnosis (van den Heuvel-Eibrink et al., 2002), which suggests that ABCG2 may play a role in therapeutic failure. Perhaps an important consideration for therapeutic failure is the overexpression of ABCG2 in the stem cell fraction compared to stem cell depleted fraction in AML (Raaijmakers et al., 2005). Ho et al. (2008) demonstrated that ABCG2 protein was expressed at significantly higher levels in the stem cell enriched fraction from AML patients that did not respond to chemotherapy compared to patients that achieved complete remission. More studies will be required to address the significance of ABCG2 to therapeutic failure in solid tumours, and how ABCG2 expression in cancer stem cells may affect response to chemotherapy.

1.2.2.10 ABCG2 Substrates and Inhibitors As already demonstrated above, one of the defining features of ABCG2 is its broad substrate specificity. Examples of ABCG2 substrates and inhibitors that have been characterized by varying degrees of stringency are shown in Table I and II, respectively. Substrates for ABCG2 have been traditionally characterized using cell systems overexpressing ABCG2, all of which rely on the researcher’s ability to quantify the amount of substrate by measuring fluorescence intensity either inherent to the compound or by tagging the compound with a fluorophore. Alternatively, compounds may be labelled with radioactive isotopes that allow it to be quantified using scintillation counters. Earlier studies relied on the use of cell lines transiently overexpressing ABCG2 to determine whether inhibition of ABCG2 results in reduced cellular

15 efflux (i.e. increase accumulation) of the suspected ABCG2 substrate. A common variation of this technique is the preparation of everted membrane vesicles from these cells which allow for the measurement of uptake/accumulation into membrane vesicles. The creation of polarized MDCKII cell lines transduced with mouse (Jonker et al., 2000) or human (Pavek et al., 2005) ABCG2 has since provided a cell system that could be used routinely by different laboratories to study ABCG2 mediated transport, making the results more consistent and comparable. Furthermore, these polarized cell lines can be further exploited to understand the vectorial transport of substrates across the apical and basolateral membrane.

It is important to note that certain mutations to ABCG2 can dramatically alter its substrate specificity. For example, drug selected human cell lines with acquired mutations to amino acid 482 that substitutes arginine for either a glycine or threonine results in dramatically enhanced ability to transport anthracyclines and rhodamine 123 (Honjo et al., 2001; Robey et al., 2003), but reduced ability to transport methotrexate and folic acid (Chen et al., 2003). In general, similar results were observed in drug-resistant mouse cell lines (Allen et al., 2002a). However, the transport activity of other ABCG2 substrates such as mitoxantrone was comparable between wildtype and mutant, which suggest that the choice of substrate is crucial when evaluating ABCG2 activity.

For more than a decade since the discovery of ABCG2, there was an inherent assumption that human and mouse ABCG2 had comparable substrate specificity. This was not shown empirically until recently. Using two independently drug-selected cell lines that overexpress wildtype ABCG2, Bakhsheshian et al. showed that human and mouse ABCG2 had very similar ability to transport various structurally unrelated ABCG2 substrates (Bakhsheshian et al., 2013). These results suggested a high degree of overlap in substrate specificity between mouse and human ABCG2.

16 Table I. Examples of substrates transported by ABCG2. Intracell. Polarized Substrates Vesicles Other References Accum. Cells line Broad Spectrum

Chemotherapeutics Daunorubicin x Litman et al., 2000 Belotecan x Li et al., 2008 Topotecan x x Li et al., 2008; Litman et al., 2000 Gimatecan x x Marchetti et al., 2007 SN-38 x Nakatomi et al., 2001 Methotrexate & di,tri- Chen et al., 2003; Volk and Schneider, 2003; x glutamate conjugates Volk et al., 2002 Mitoxantrone x Kodaira et al., 2010; Litman et al., 2000 Flavopiridol x x Kodaira et al., 2010; Robey et al., 2001

Tyrosine Kinase Inhibitors

Dasatinib x x Chen et al., 2009 Lapatinib x Polli et al., 2008 Imatinib mesylate x Burger et al., 2004 CI1033 x Erlichman et al., 2001 Sorafenib x x Agarwal et al., 2011; Lagas et al., 2010 Gefitinib x x Agarwal et al., 2010 Erlotinib x Kodaira et al., 2010 Sunitinib x x Tang et al., 2012 Cediranib x x Wang et al., 2012

Serine/Threonine Kinase inhibitors

Dabrafenib x x Mittapalli et al., 2013 Vemurafenib x x Mittapalli et al., 2012 GDC-0941 x Salphati et al., 2010

Antibiotics & Antivirals

Ciprofloxacin, Ofloxacin, x x Merino et al., 2006 Norfloxacin Nitrofurantoin x x Merino et al., 2005a Zidovudine, Lamivudine x x Wang et al., 2003 Acyclovir x x Gunness et al., 2011; Jonker et al., 2005 Puromycin, Cladibrine, Bis x Takenaka et al., 2007 (POM) PMEA

Other xenobiotics

Rosuvastatin x x Kitamura et al., 2008 BODIPY-Prazosin x Litman et al., 2000 Glyburide x Zhou et al., 2008 Cimetidine x x Pavek et al., 2005 CP-671,305 x Kalgutkar et al., 2007 Moxidectin x x Perez et al., 2009 Edaravone sulfate x x Mizuno et al., 2007

Carcinogens, Toxins, & Plant

Derivatives PhIP x x x van Herwaarden et al., 2003 Aflatoxin B1, IQ, Trp-P1 x x van Herwaarden et al., 2006 Pheophorbide a x x Jonker et al., 2002; Robey et al., 2004 Genistein (phytoestrogen) x Imai et al., 2004 Phenethyl isothiocyanate x Ji and Morris, 2005

17 Table I cont’d. Intracell. Polarized Substrates Vesicles Other References Accum. Cells

Endobiotics & Nutrients

Amyloid beta 1-40 x x Do et al., 2012; Tai et al., 2009 Estrone sulfate (E1S) x x Imai et al., 2003; Suzuki et al., 2003 17β-estradiol sulfate x Imai et al., 2003 17β-estradiol glucoronide x Imai et al., 2003; Suzuki et al., 2003 Vitamin B2 (Riboflavin) x x van Herwaarden et al., 2007 Vitamin K3 & Plumbagin x Shukla et al., 2007 Dehydroepiandrosterone x Suzuki et al., 2003 sulfate (DHEAS) Protoporphyrin IX x Zhou et al., 2005 Folic Acid x Chen et al., 2003

Fluoroscein/Dye/Probe

Rhodamine 123 x Litman et al., 2000 D-luciferin x Zhang et al., 2007b Hoescht 33342 x Scharenberg et al., 2002; Zhou et al., 2001 Sulfate conjugated E3040 x Suzuki et al., 2003 and 4-MU

Intracell. Accum., intracellular accumulation Vesicles, everted membrane vesicles prepared from cells overexpressing ABCG2 Polarized Cells, mainly polarized MDCKII or LLC-PK1cells overexpressing ABCG2 that are used for vectorial transport Other: transwell studies or in vivo experiments using ABCG2 -/- mice.

Table II. ABCG2 inhibitors. Inhibitor Pharmacological Properties References GF120918 (Elacridar) Inhibitor of ABCG2 and ABCB1, useful in vitro and Allen et al., 1999; de Bruin et al., 1999; in vivo (at 1 – 10 µM) Jonker et al., 2000

Fumetrimorgin C (FTC) Used in vitro at 1-10 µM Rabindran et al., 2000 Neurotoxic therefore not used for in vivo experiments

Ko143 Analogue of FTC, Allen et al., 2002b Used in vitro at 25 nM to 10 µM Safe to use in mouse in vivo (10 mg/kg, limited toxicity even at 50 mg/kg p.o.)

Note. While many substrates and xenobiotics have been shown to inhibit ABCG2, the inhibitors listed above are most commonly used in the literature.

18 1.3 Promoter Organization of the Human and Mouse ABCG2 Gene

1.3.1 Human ABCG2 gene promoter First described by Bailey-Dell et al., the human ABCG2 gene is thought to be transcribed from a TATA-less and CpG-rich promoter (Bailey-Dell et al., 2001). However, given the existence of multiple human ABCG2 mRNA isoforms that differ at the 5’- untranslated region (5’-UTR), it is recognized that transcription likely arises from various promoters (Fig. 1A). To date, at least four human ABCG2 mRNA isoforms have been described: E1U, E1A, E1B, and E1C (Campbell et al., 2011; Nakanishi et al., 2006). The E1B and E1C isoforms are adjacent to one another and are often indistinguishable. As such, these isoforms are sometimes referred to as E1B/C. The E1B/C isoform is expressed in most normal human tissues and cells whereas the E1A and E1U isoforms show a more limited expression profile (Campbell et al., 2011; Nakanishi et al., 2006). It has been suggested that the utilization of different promoters may be responsible, at least in part, for tissue- and cell- specific expression of ABCG2. For example, while E1A, E1B, and E1C are expressed at similar levels in MCF-7 cells, E1A contributes significantly less to total ABCG2 transcript expression in drug resistant MCF-7/AdrVp cells (Nakanishi et al., 2006). Additionally, others have shown that the E1U isoform is expressed only in paediatric acute megakaryoblastic leukemia (AMKL) cells but not in other AML subtypes (Campbell et al., 2011). The translation start site for ABCG2 is located in exon 2. For this reason, the different mRNA isoforms code for the same ABCG2 protein.

Please note that to avoid confusion, the theoretical transcription start site initially described by Bailey-Dell et al. (2001) will be used as a point of reference to describe the position of regulatory elements.

19 1.3.2 Mouse Abcg2 gene promoter Similar to human, the mouse Abcg2 gene is transcribed from at least four alternative promoters resulting in four mRNA isoforms (E1a, E1b, E1c, and E1u) that differ at the untranslated first exon (Natarajan et al., 2011; Xie et al., 2013; Zong et al., 2006). However, there is a greater distance in the genomic region separating E1b and E1c in mice, producing distinct variants of these isoforms. The E1a isoform is the predominant isoform expressed in haematopoietic stem cells (Zong et al., 2006), whereas E1b is the predominant isoform expressed in the mouse small intestine (Natarajan et al., 2011). Interestingly, E1b in the small intestine shows circadian expression. Evidence suggests that this is mediated by the binding of activating transcription factor-4 (ATF-4) to a cAMP-responsive element (CRE) in the E1b promoter (Hamdan et al., 2012). E1u is predominantly expressed in the testis and is under the control of steroidogenic factor-1 (SF-1) that binds the SF-1 response element in the E1u promoter (Xie et al., 2013)

Figure 1. Schematic representation of major alternative untranslated first exons in the human and mouse ABCG2/Abcg2 gene. Note that the figure is not to scale. Numbers above each exon indicate the position of the splice site that is joined to exon 2 which contains the translation start site (ATG). (A) Human ABCG2 gene. The transcription start site (+1) is based on Bailey- Dell et al., 2001. For simplicity, several E1U sub-exons that are coded by the intervening region between E1U and E1A are not shown (Campbell et al., 2011). (B) Mouse Abcg2 gene. The transcription start site (+1) is based on Natarajan et al., 2011. *In the mouse testis, the E1u exon consists of 2 exons (one variable and one constant) that are approximately 400 bp apart (Xie et al., 2013). The splice site for the constant E1u exon is not clear since the sequence information has not been released.

20 1.4 Regulatory Control of ABCG2 Expression

1.4.1 Hormones and inflammatory cytokines

1.4.1.1 Estrogen. After the initial characterization of the human ABCG2 promoter (Bailey- Dell et al., 2001), Ee et al. identified an imperfect estrogen response element (ERE) at -187 to - 173 in the 5’-flanking region of the ABCG2 gene (Ee et al., 2004a). These authors showed that

ABCG2 mRNA was induced by 17β-estradiol (E2) in estrogen-receptor (ER) positive T47D:A18 cells and in PA-1 ovarian cells stably expressing human ERα. This effect was blocked by co-treatment with the estrogen receptor antagonist, ICI 181, 780. Site-directed mutagenesis of the ERE in an ABCG2 promoter-driven luciferase reporter assay further demonstrated that the ERE was necessary for promoter activity in response to E2 (Ee et al.,

2004a). ABCG2 mRNA and/or protein expression was also induced by E2 in human MCF-7 breast cancer cells (Zhang et al., 2006) and human BeWo choriocarcinoma cells (Ee et al., 2004b). In contrast, there are multiple reports that estrogen receptor agonists negatively regulate ABCG2 expression. Wang et al. showed that β-estradiol decreased ABCG2 mRNA and protein expression in BeWo cells (Wang et al., 2006b, 2008a). Imai et al. demonstrated that β-estradiol downregulates ABCG2 protein expression but does not affect endogenous and exogenous ABCG2 mRNA levels in MCF-7 or MCF-7/ABCG2 and T-47D/ABCG2 cells transfected with a plasmid constitutively expressing ABCG2 (Imai et al., 2005). This suggested a mechanism for estrogen induced post-transcriptional (non-genomic) inhibition of ABCG2 expression and further examination revealed that estrogen suppressed the production of mature ABCG2 protein (Imai et al., 2005). These non-genomic actions of estrogen and its receptor have also been proposed as a mechanism for estrogen-induced downregulation of ABCG2 transport activity and expression in mouse and rat brain capillaries (Hartz et al., 2010a, 2010b). The characterized mechanism, which involves PI3K/AKT signalling (Hartz et al., 2010b), will be described in more detail in section 1.4.4.1 PI3K/AKT signalling. It is not clear why estrogen has such different effects on ABCG2 expression in different cells/tissues. Imai et al. suggested that this discrepancy may be due to differences in the cell clone used (T-47D:A18 versus the more heterogenous parental T-47D) while others have attributed it to variation in expression of ERα versus ERβ (Hartz et al., 2010b). In addition, as will be discussed shortly,

21 the presence of other hormones/cytokines in the medium or differences in the expression of nuclear receptors between cell lines may affect how estrogen regulates ABCG2 expression.

1.4.1.2 Progesterone. Wang et al. demonstrated that progesterone treatment induced ABCG2 mRNA and protein expression in BeWo choriocarcinoma cells (Wang et al., 2006b). However, co-treatment with the progesterone receptor antagonist RU-486 failed to inhibit this response. This led the authors to suggest that the observed upregulation of ABCG2 expression by progesterone may be mediated by a non-classical progesterone receptor (PR) signalling event. Most interestingly, cells concomitantly treated with progesterone and β-estradiol upregulated ABCG2 mRNA and protein to levels significantly higher than levels seen in cells treated with progesterone alone (Wang et al., 2006b). Examination of the nuclear receptor mRNA in response to treatment revealed that estrogen induced PRB mRNA but downregulated ERβ mRNA. Thus, Wang et al. proposed that estrogen indirectly upregulated the expression of

ABCG2 by increasing PRB expression. However, this mechanism does not address the issue of estrogen-mediated post-transcriptional inhibition of ABCG2 expression proposed earlier by Sugimoto and colleagues (Imai et al., 2005). Nevertheless, it may be possible that transcriptional upregulation by progesterone overwhelms the suppressive effects of estrogen on ABCG2 protein expression. More recently, using deletion and site-directed mutagenesis analysis of an ABCG2-promoter driven luciferase reporter, Wang et al. (2008b) identified a progesterone response element (PRE) between -187 and -173 of ABCG2 promoter region. Perhaps with important biological implications, this PRE overlaps the previously identified

ERE (Ee et al., 2004a). Furthermore, the authors demonstrated that overexpression of PRB in

BeWo cells enhanced progesterone-induced ABCG2 promoter activity while PRA co- expression repressed this effect. In contrast, Yasuda et al. did not observe any increase in ABCG2 promoter-driven luciferase reporter activity in BeWo cells after progesterone treatment (Yasuda et al., 2009). Noteworthy, in the former study, Wang et al. (2008b) used

BeWo cells that overexpressed PRB whereas in the latter study, normal BeWo cells were used. This suggests that progesterone-induced upregulation of ABCG2 expression may be highly dependent on the level of PRB expression.

1.4.1.3 Testosterone. Despite interests in understanding the role of sex hormones in the regulation of ABCG2, few studies have investigated the role of testosterone in regulating

22 ABCG2 expression. Tanaka et al. showed that male rats and mice express significantly higher levels of Abcg2 mRNA in kidney and liver, respectively, compared to their female counterparts (Tanaka et al., 2005). Furthermore, ovariectomized rats and castrated mice showed upregulated and downregulated Abcg2 mRNA expression, respectively. Interestingly, only hormone replacement therapy with dihydroxytestosterone but not E2 was able to restore Abcg2 mRNA levels in gonadectomized mice. These observations led Tanaka et al. (2005) to conclude that male-predominant expression of Abcg2 in rat kidney and mouse liver was due to the suppressive effects of β-estradiol and inductive effects of testosterone. In contrast, testosterone alone did not affect ABCG2 expression in human placental BeWo cells (Wang et al., 2008a). However, when cells were co-treated with β-estradiol, which alone decreased ABCG2 mRNA and protein expression, ABCG2 mRNA and protein levels were induced by 2- fold (Wang et al., 2008a). This effect was inhibited by ICI-182,780 or flutamide, an androgen receptor antagonist. As it was observed that E2 induced androgen receptor mRNA, the authors suggested that this may be a possible mechanism by which co-treatment of E2 and testosterone cooperatively induce ABCG2 expression. However, the exact molecular mechanism awaits further investigation. Nevertheless, these studies together implicate testosterone as a potentially important regulator of ABCG2 expression.

1.4.1.4 Glucocorticoids. A number of studies have examined the effects of glucocorticoids on ABCG2 expression. While these efforts have generated a rather large body of data, the overall effect of glucocorticoids and the underlying mechanism by which they regulate ABCG2 expression remains a mystery. The glucocorticoid receptor (GR) agonist dexamethasone induced Abcg2 mRNA and protein expression in primary rat brain endothelial cells (Narang et al., 2008). The specific mechanism for this induction is unclear since co-treatment with the GR antagonist RU-486 only partially attenuated the induction of ABCG2 by dexamethasone, and dexamethasone also induced the expression of PXR (discussed below), which is a known regulator of ABCG2. These observations led the authors to suggest that the regulation of ABCG2 by dexamethasone in rat brain endothelial cells may involve both the GR and PXR pathways. Petropoulos et al. demonstrated that the effect of dexamethasone on Abcg2 mRNA expression in mouse fetal brain (Petropoulos et al., 2011a) and placental tissue (Petropoulos et al., 2011b) depended on the dose, time of treatment during gestation, and sex of the fetus. However, more importantly, protein expression was unchanged. Unlike the results in rodents,

23 ABCG2 expression was downregulated in human MCF-7 (Honorat et al., 2008) and HepG2 (Rosales et al., 2013) cells treated with glucocorticoids. Given the inconsistencies in the literature, which may result from species or tissue differences, it is clear that more studies will be required to fully elucidate the role of glucocorticoids in ABCG2 regulation.

1.4.1.5 Proinflammatory Cytokines. Inflammation has a profound effect on the expression of drug transporters and drug metabolizing enzymes. While in general, proinflammatory cytokines downregulate the expression of ABCG2, accumulating evidence shows that the effect of any given proinflammatory cytokine ultimately depends on the experimental system (i.e. cells). Interleukin 1β (IL-1β) and tumor necrosis factor alpha (TNF-α) both repress ABCG2 expression in primary human trophoblasts (Evseenko et al., 2007), hCMEC/D3 brain endothelial cells (Poller et al., 2010), HeLa ovarian cancer cells (Mosaffa et al., 2012), and MDA-MB-435 breast cancer cells, but induce ABCG2 expression in other human breast cancer cells (MCF-7, BT-474, and CAL51, 84A1) (Malekshah et al., 2012). In primary human hepatocytes, IL-1β treatment decreased ABCG2 mRNA expression (Le Vee et al., 2008) but TNF-α induced ABCG2 protein expression without observable changes to mRNA levels (Le Vee et al., 2009). Interleukin-6 (IL-6) did not affect ABCG2 expression in primary trophoblasts (Evseenko et al., 2007) but decreased ABCG2 expression (transcript and protein) in primary hepatocytes (Le Vee et al., 2009), hCMEC/D3 (Poller et al., 2010) and HeLa cells (Mosaffa et al., 2012). Interestingly, oncostatin M, a member of the IL-6 family, induced ABCG2 expression in HepaRG and human primary hepatocytes (Le Vee et al., 2011). The highly variable response observed with pro-inflammatory cytokines may relate to the modulatory effect of these cytokines on other pathways known to regulate ABCG2 expression. Pradhan et al. showed that IL-1β and TNF-α, which alone do not affect ABCG2 expression, potentiated the inductive effects of estrogen on ABCG2 expression in MCF-7 cells (Pradhan et al., 2010). This is attributed to the recruitment of p65, a member of the NF-κB family, to an imperfect p65 response element at -28/-19 that helps stabilize the binding of ER to an adjacent ERE at the ABCG2 promoter. Interestingly, TNF-α alone was not able to induce p65 recruitment to ABCG2 but required recruitment of ER to the promoter by co-stimulation with E2. These authors suggested that p65 and ER cooperatively stabilize one another’s DNA- interaction. Consistent with this study, ectopic expression of p65 in MCF-7 cells failed to induce ABCG2 promoter-driven reporter activity (Wang et al., 2010b). In contrast, p50,

24 another member of the NF-B family, alone induced ABCG2 promoter activity and was recruited to the ABCG2 promoter. Downregulation of NF-B has been suggested as a potential mechanism for the reduction in ABCG2 expression observed after overexpressing wildtype p53 in MCF-7 cells (Wang et al., 2010b). An additional example that cytokines may modulate the effect of other signalling pathways is the observation that IL-6 synergised with endoplasmic reticulum stress inducers to upregulate ABCG2 expression in human plasma cells (Nakamichi et al., 2009). Mechanistically, IL-6 and endoplasmic reticulum stress inducers activated HIF-1 and XBP-1, respectively, which were then recruited to an overlapping DNA binding sequence, part of which is the hypoxia response element identified by Krishnamurthy et al. (2004), at the ABCG2 promoter. Despite the considerable overlap between the HIF-1 and XBP-1 binding sequences, there was no steric hindrance but instead, cooperative binding.

1.4.1.6 Other Growth factors. While growth factors in general have received considerable attention for their role in cancer, information with regards to how they modulate ABCG2 expression is by comparison sparse. Insulin-like growth factor-II (IGF-II) and epidermal growth factor (EGF) have been shown to induce ABCG2 mRNA and protein expression in primary human trophoblasts (Evseenko et al., 2007). EGF also increased the mRNA and protein expression of ABCG2 in BeWo, MCF-7 and primary human cytotrophoblasts (Meyer zu Schwabedissen et al., 2006). In addition to EGF, which binds and activates ERBB1/EGFR, overexpression of the dimerization partner ERBB2 in MCF-7 cells also induced ABCG2 expression (Gilani et al., 2012; Zhang et al., 2011). The proposed mechanisms that mediate the response to EGF, its receptor, and ERBB2/HER2 will be discussed in more detail below (section 1.4.4.2 MAPK signalling). In contrast to IGF-II and EGF, ABCG2 expression was downregulated by transforming growth factor-β (TGF-β) in gastric cancer cells (OCUM- 2MLN, OCUM-8, OCUM-9, HSC-43) and cell lines from other tissues including A549 (lung), HuH7 (liver), Hela (ovarian), and MDA-MB-231 (breast) (Ehata et al., 2011). Using the OCUM-2MLN cell line, it was demonstrated that siRNA targeting SMAD4 attenuated this response to TGF-β. Most importantly, TGF-β induced SMAD2/3 recruitment to the promoter region and intron 1 of the human ABCG2 gene. This led Ehata et al. (2011) to suggest that recruitment of SMAD2/3 to the ABCG2 gene is responsible for the transcriptional repression of ABCG2 by TGF-β.

25 In addition to the “growth factors” described above, the morphogen sonic hedgehog (SHH) and its signalling pathway is involved in the regulation of ABCG2 in certain human cancer cells. Sims-Mourtada et al., demonstrated that knockdown of Gli1, a transcription factor activated by hedgehog signalling, resulted in decreased ABCG2 protein expression in Seg-1, PC3, and DM14 carcinoma cells with high basal level of hedgehog signalling (Sims- Mourtada et al., 2007). The same effect was observed using cyclopamine to pharmacologically inhibit Smoothened (SMO), a G-protein coupled receptor, that initiates a signalling cascade that ultimately results in nuclear localization of the Gli family of transcription factors. Treatment with SHH induced ABCG2 protein expression in LnCap and Seg-1 cells. Clinically, SHH expression is positively associated with ABCG2 expression in diffuse large B-cell lymphoma (Kim et al., 2009a). More recently, Chen et al. (2014) also showed siRNA-mediated knockdown of Gli1 resulted in reduced expression of ABCG2 protein in A2780 ovarian carcinoma cells. Using electromobility shift assays, these authors further identified a putative consensus Gli1 binding sequence in the human ABCG2 promoter. Future work will be necessary to determine if Gli1 can bind the ABCG2 gene in the native chromatin context.

26

Figure 2. Regulatory mechanisms that modulate human ABCG2 expression. Please see text for details. DR5, CAR binding response element -7933/-7917; DRE, Dioxin Response Element -2347/-2343 and -193/-189; XBP1/HRE, overlapping X-box binding protein 1 and hypoxia response element at -117/-113 and -116/-112, respectively; E/PRE, overlapping estrogen and progesterone response element -187/-173; ARE, antioxidant response element - 376/-366; NFκBRE, NFκB response element -28/-19; and DR1, three motifs that bind PPAR - 3939/-3818.

27 1.4.2 Stress and xenobiotics

1.4.2.1 Hypoxia. Krishnamurthy et al. (2004) demonstrated that hypoxic conditions induced the expression of ABCG2 mRNA and increased the efflux of Hoechst dye (a ABCG2 substrate) in various human cancer cell lines (placental JAR cells, osteosarcoma Saos-2 cells, and myeloid leukemia OCI-AML3 cells). Further investigation revealed a hypoxia response element (HRE) at -115 bp upstream of the transcription start site which was bound by the hypoxia-inducible factor-1 (HIF-1) complex under hypoxic conditions.

1.4.2.2 Aryl hydrocarbon receptor (AHR). While it is well established that aromatic hydrocarbons induce phase I and II drug metabolizing enzymes, Ebert et al. demonstrated in the human colon cancer cell line Caco-2 that AHR agonists also induce ABCG2 expression and thereby potentially affect phase III elimination of xenobiotics (Ebert et al., 2005, 2007). Our laboratory extended these findings to show that the prototypical AHR agonist, TCDD, induced ABCG2 expression in human secondary cell lines of the colon (C2bbe1, LS180, LS174T), liver (HepG2), and breast (MCF7) and identified a dioxin response element (DRE) at -193/-189 base pairs upstream of the human ABCG2 gene (Tan et al., 2010). In contrast to our study, using reporter assays driven by the ABCG2 promoter, Tompkins and colleagues found that the -193/-189 DRE contributed negligibly, if any, to reporter activity (Tompkins et al., 2010). Instead, a more distal DRE at -2347/-2343 from the transcription start site was critical for aromatic hydrocarbon-induced reporter activity in LS174T cells. It is unclear what may account for this difference. However, since different cell lines were used to interrogate this pathway, the regulatory mechanism may be cell line specific. For example, the specific response observed may depend on the utilization of alternative promoters as was demonstrated recently that only the E1B isoform, but not E1A, was induced by TCDD in HepG2 cells (de Boussac et al., 2012). Additionally, it is interesting to note that while aryl hydrocarbon receptor agonists are inducers of ABCG2 in various human-derived cell lines, and more recently in primary bovine mammary epithelial cells (Halwachs et al., 2013), Abcg2 mRNA was not upregulated in the liver and small intestine of mice treated with TCDD (Tan et al., 2010) or 3-methylcholanthrene (Han and Sugiyama, 2006). This lack of response may not be generalizable to all rodents since ABCG2 protein was upregulated in the spinal cord capillaries of rats treated with TCDD (Campos et al., 2012).

28 1.4.2.3 Peroxisome-proliferator activated receptor (PPAR). Peroxisome-proliferator activated receptor gamma (PPARγ) agonists (troglitazone, GW347845X and rosiglitazone) have been shown to induce ABCG2 mRNA and protein expression in human dendritic cells. This is attributed to three PPAR response elements located within a 150-base pair region approximately 3900-base pairs upstream of the transcription start site (Szatmari et al., 2006). Similar findings were observed when human cerebral microvascular endothelial cells (hCMEC/D3) were treated with the PPAR alpha (PPARα) selective agonist clofibrate (Hoque et al., 2012). Mice treated with clofibrate for 10 days also showed upregulated expression of Abcg2 mRNA and protein in the liver (Moffit et al., 2006). As further evidence that Abcg2 is regulated by the PPAR pathway in mice, PPARα knockout mice express significantly less ABCG2 protein in the liver compared to wildtype mice (Moffit et al., 2006).

1.4.2.4 Pregnane X Receptor (PXR). A number of studies have demonstrated that PXR is a regulator of ABCG2 expression. Among the earliest evidence was the observation that the aromatic amine 2-acetylaminofluorene (2-AAF) induced Abcg2 mRNA expression in wildtype mice but not in PXR-null mice (Anapolsky et al., 2006). PXR activators such as rifampicin have also been shown to induce ABCG2 expression in primary human hepatocytes (Jigorel et al., 2006) and porcine brain capillary endothelial cells (Lemmen et al., 2013a). The specific PXR ligand pregnenolone 16α-carbonitrile (PCN) induced ABCG2 protein in primary rat brain microvascular endothelial cells (Narang et al., 2008) but did not induce Abcg2 expression in the placenta (Gahir and Piquette-Miller, 2011) or liver (Teng and Piquette-Miller, 2005) of mice treated with PCN for four days. It is unclear why different PXR activators (2-AAF vs PCN) can have such different effects, but it does suggest that alternative pathways activated by these chemicals can contribute to the overall effect. Overexpression of PXR in HepG2 cells reportedly increases basal ABCG2 mRNA expression (Naspinski et al., 2008). In general, while PXR appears to be a positive regulator of ABCG2 expression, basal Abcg2 mRNA expression in the mouse placenta is inversely related to the copy number of PXR (Gahir and Piquette-Miller, 2011). Collectively, these studies suggest that modulation of ABCG2 expression by PXR is highly dependent on the chemical agent used and shows cell line and tissue specificity.

29 1.4.2.5 Constitutive Androstane Receptor (CAR). As part of a larger study to investigate the role of xenobiotic receptors in the expression of hepatic transporters, Jigorel et al. (2006) showed that phenobarbital, a CAR activator, induced ABCG2 expression in human primary hepatocytes. This provided some of the earliest evidence that CAR may be a regulator of ABCG2 (Jigorel et al., 2006). However the molecular mechanism was not elucidated until recently when it was demonstrated that activated CAR was recruited as part of a heterodimer with retinoid X receptor (RXR) to a direct repeat 5 (DR5) motif located -7988 to -7972 in the 5’-flanking region of the human ABCG2 gene (Benoki et al., 2012). The effect of CAR activation on ABCG2 expression appears to be a relatively well conserved phenomenon. Activation of CAR by phenobarbital in isolated rat brain capillaries, by TCPOBOP in mouse brain capillaries, and by CITCO in porcine brain capillary endothelial cells all induced ABCG2 expression (Lemmen et al., 2013b; Wang et al., 2010c).

1.4.2.6 Nuclear factor-erythroid 2 p45-related factor 2 (Nrf2). In response to oxidative/electrophilic stress, many cytoprotective genes (e.g. phase I and II mechanisms) are upregulated by Nrf2. It is becoming apparent that, in addition to traditional mechanisms of cytoprotection via phase I and II metabolizing enzymes, phase III genes (i.e. membrane transporters) are also induced. In particular, ABCG2 mRNA was induced in HepG2 hepatocarcinoma cells subjected to oxidative/electrophilic stress using tert-butylhydroquinone (Adachi et al., 2007) or photoactivated porphyrins (Hagiya et al., 2008), the effect of which was attenuated by knockdown of Nrf2. More recently, a detailed investigation was conducted by Singh et al. to further characterize the relationship between Nrf2 and ABCG2 (Singh et al., 2010). These authors demonstrated that knockdown of Nrf2 in A549 and H460 lung cancer cells with high basal activation of Nrf2 due to KEAP1 mutations downregulated ABCG2 expression. Conversely, activation of Nrf2 by knockdown of KEAP1 in normal airway epithelial NuLi cells induced the expression of ABCG2. Mechanistically, Nrf2 regulates ABCG2 by binding to an Antioxidant Responsive Element (ARE) located at -376/-366 of the human ABCG2 gene.

1.4.2.7 Retinoic acid receptor (RAR)/retinoid X receptor (RXR). ABCG2 expression was induced by all-trans retinoic acid in Caco-2, Hepa-RG, and primary human hepatocytes (Hessel and Lampen, 2010; Le Vee et al., 2013). Most importantly, AM580 and CD2608,

30 which are selective activators of RXRα and retinoid X receptor (RXR), respectively, also induced ABCG2 expression. In combination, these agonists further upregulated ABCG2 compared to the individual treatment alone (Hessel and Lampen, 2010). Because many nuclear receptors known to regulate ABCG2 expression function as heterodimers with retinoid X receptor (RXR) (Lefebvre et al., 2010), RXR may be one of the most important regulators of ABCG2.

1.4.2.8 Cyclooxygenase 2 (COX-2) activation. Due to the growing body of evidence that COX-2 may be an important target for cancer treatment (Menter et al., 2010), there has been interest in understanding the role of COX-2 activation in ABCG2 expression. COX-2 expression is positively correlated with ABCG2 expression in non-Hodgkin’s lymphomas (Szczuraszek et al., 2009). Induction of COX-2 activity by tetradecanoyl phorbol acetate (TPA) increased ABCG2-mediated mitoxantrone efflux activity in several breast cancer cells lines (Kalalinia et al., 2011). This induction was only very modestly, almost negligibly, inhibited by co-treatment with the selective COX-2 inhibitor celecoxib. In a latter study, it was demonstrated that TPA induced ABCG2 expression, but these changes did not agree with the enhanced ABCG2 activity observed in the former study (Kalalinia et al., 2012). Contrary to the positive role for COX-2 activation on ABCG2 expression, the non-selective COX inhibitor naproxen was shown to induce ABCG2 mRNA and protein in human Caco-2 cells after subchronic treatment (24-72h) but not after chronic treatment (four passages) (Zrieki et al., 2008). It is not clear if these discrepancies reflect a lack of a suitable model for the COX-2 inflammatory cascade in vitro. Certainly, the data thus far are sparse and inconsistent and therefore the role of COX-2 activation in ABCG2 expression remains inconclusive.

1.4.2.9 Extracellular milieu: folate status. Folates and anti-folates such as methotrexate are well-characterized substrates of ABCG2. Folate supplementation is used clinically to counteract methotrexate toxicity. Therefore, there has been interest to explore the effect of folates on ABCG2 expression. MCF-7 breast cancer cells and the mitoxantrone-resistant subline MCF7/MR cells cultured in low folate conditions showed reduced ABCG2 expression and function, with a corresponding increase in sensitivity to mitoxantrone and methotrexate (Ifergan et al., 2004). These cells also showed elevated capacity to accumulate folic acid and upregulated activity of folypoly-γ-glutamate synthetase (FPGS), an enzyme involved in the

31 conjugation of glutamates to folates, which increases cellular retention. This led the authors to suggest that loss of ABCG2 function coupled with increased FPGS activity may be a mechanism for protection against low folate conditions. In contrast to the above study, Caco-2 cells that were grown in low folate conditions actually showed induced ABCG2 expression (Lemos et al., 2008). As expected, these cells displayed increased resistance to mitoxantrone but for unclear reasons showed increased sensitivity to methotrexate. The effect of folate deprivation appears to be cell line specific. Folate deprivation did not induce ABCG2 expression in various cancer cell lines from various tissues: KB (nasopharyngeal), OVCAR-3 and IGROV1 (ovarian), ZR75 (breast), SCC-11B and SCC-22B (head and neck) (Lemos et al., 2009) but was induced in another colon cancer cell line WiDr after folate deprivation, suggesting that it may be a tissue specific phenomenon. Interestingly, ABCG2 expression was mainly localized to the plasma membrane in MCF-7 cells but in Caco-2 and WiDr cells, ABCG2 was expressed in the intracellular compartment. Perhaps, the localization of ABCG2 may be a predictor of the effects of folate deprivation on ABCG2 expression.

1.4.3 miRNA and Epigenetics

1.4.3.1 micro-RNA (miRNA). Several miRNAs have been shown to regulate the expression of ABCG2. To et al. showed that ABCG2 mRNA was more stable in drug resistant cancer cell lines (S1MI80, MCF-7 FLV1000, SF295 MX2000, H460 MX20, and A549 Beca250) compared to the parent cell line (S1, MCF7, SF295, H460, and A549, respectively) (To et al., 2008, 2009). This was attributed to a truncation in the 3’UTR of ABCG2 mRNA, which normally serve as a binding site for miR-519c (To et al., 2008a, 2009). This is supported by the observation that while miR-519c mimic and inhibitor decreased and increased ABCG2 expression, respectively, in A549 (To et al., 2008a) lung carcinoma cells and S1 colon cancer cells (To et al., 2009), ABCG2 expression was unaffected by these mimic/inhibitor agents in drug resistant S1MI80 cells. Therefore the regulation of ABCG2 expression by miR-519c was “uncoupled” in this drug resistant cell line and may be a reason why drug resistant cell lines overexpress miR-519c even though this miRNA negatively regulates ABCG2 in drug sensitive cells (To et al., 2009). In contrast, miR-520h, which may play a role in haematopoietic stem cell differentiation by downregulating ABCG2 (Liao et al., 2008), is expressed at lower levels in drug resistant cells compared to the respective parent cell lines (To et al., 2009). Using

32 siRNA to downregulate ABCG2 mRNA, the authors showed that miR-520h levels increased when ABCG2 mRNA levels decreased. This led To et al. to suggest that the low level of miR- 520h in drug resistant cells may be due to the sequestration of miR-520h by high levels of ABCG2 mRNA (To et al., 2009). It should be noted that miR-520h should not simply be regarded as a marker but is functionally significant since transfection of miR-520h into PANC- 1 pancreatic cancer cells significantly reduced ABCG2 expression and consequently the Hoechst 33342 excluding side-population cells (Wang et al., 2010a). In addition to miR-519c and miR-520h, miR-328, miR-181a and miR-487a are downregulated in mitoxantrone resistant MCF-7/MX cells (Jiao et al., 2013; Ma et al., 2013; Pan et al., 2009). Ectopic expression of these miRNA, which targets the 3’UTR of ABCG2 mRNA, reduced ABCG2 expression and sensitized MCF-7/MX cells to mitoxantrone in vitro. Intravenous administration of miR-487a or miR-181a significantly enhanced the inhibitory effects of mitoxantrone on tumour growth in mice implanted with MCF-7/MX xenografts (Jiao et al., 2013; Ma et al., 2013). Interestingly, the expression of miR-328 is also inversely correlated with ABCG2 expression in human placenta (Saito et al., 2013).

1.4.3.2 Promoter Methylation. The promoter region of the human ABCG2 gene contains a putative CpG island (Bailey-Dell et al., 2001). The effect of promoter methylation on ABCG2 expression was first demonstrated in leukemia cells. Turner et al. showed that 5-aza- deoxycytidine (5-Aza-dc) induced ABCG2 expression in HL-60 leukemia cells by demethylating the ABCG2 promoter (Turner et al., 2006). Clinically, ABCG2 promoter methylation is inversely correlated with ABCG2 expression in patient myeloma samples (Turner et al., 2006). It is not entirely clear how methylation directly affects ABCG2 gene expression. One proposed mechanism is that methylation may impair recruitment of c-myc, which is often deregulated in cancer, to the ABCG2 promoter (Porro et al., 2011). This is based on the observation that the myc binding site near the ABCG2 transcription start site was methylated in CD34+ progenitor cells from five out of twenty-one patients with chronic myeloid leukemia and that ABCG2 expression was lowest in these samples. Notably, this mechanism is only speculative since there is no evidence to support a causal relationship between methylation status and c-myc recruitment on ABCG2 expression. Methylation at the ABCG2 promoter may explain some of the inherent differences in ABCG2 expression among cell lines. UOK121 and UOK143 renal carcinoma cells express

33 half the level of ABCG2 compared to UOK181 cells (To et al., 2006). This is attributed to increased promoter CpG methylation, increased acetylation and decreased methylation at lysine 9 on histone 3, and increased recruitment of transcription repressing methyl-CpG binding proteins to the ABCG2 promoter. Demethylation of the ABCG2 promoter induced ABCG2 expression in UOK121 and UO143 cells. In addition to renal carcinoma cell lines, small and non-small lung cancer cell lines also demonstrate an inverse relationship between promoter methylation and ABCG2 expression (Nakano et al., 2008). Most interestingly, the ABCG2 promoter is hypomethylated in a number of drug-resistant cell lines that overexpress ABCG2 (Bram et al., 2009; Nakano et al., 2008). In contrast, the corresponding parent cell lines that express very low levels of ABCG2 have a highly methylated ABCG2 promoter. While promoter hypomethylation and the associated upregulation of ABCG2 is a concern for drug resistance, promoter methylation has recently been described as a mechanism by which melatonin augments the effect of chemotherapeutics (Martín et al., 2013). Treatment with melatonin induced ABCG2 promoter methylation and reduced ABCG2 expression in brain tumour stem cell lines and A172 human malignant glioblastoma cells. These effects were inhibited by 5-Aza-cytidine. Collectively, there is abundant evidence to support the important regulatory role of CpG methylation in ABCG2 expression.

1.4.3.3 Histone Modification. Many cell lines derived by multi-step selection for drug resistance show increased accumulation of permissive histone marks (acetylated lysine 9/14 and methylated lysine 4 on histone 3) at the ABCG2 promoter (To et al., 2008b). Clones of MCF-7 cells derived by a single step selection procedure for resistance to doxorubicin and etoposide also display increased histone 3 acetylation at the ABCG2 promoter (Calcagno et al., 2008). These drug resistant cells all overexpress ABCG2 compared to their parent cell lines and as such it is not entirely unexpected that there is an increase in permissive histone marks at the promoter. However, more controversial is whether these changes at the promoter represent an adapation to the more hostile environment or simply selection of pre-existing cells that already display a more open chromatin organization at the ABCG2 gene and therefore inherently express higher levels of ABCG2. In addition to drug selection, the effect of histone deacetylase inhibitors (HDACi) on ABCG2 expression has also been studied. To et al. showed that the HDACi romidepsin (depsipeptide) induced ABCG2 mRNA and protein expression in S1 colon cancer cells and

34 H460 lung cancer cells, downregulated ABCG2 mRNA in MCF-7 and SF295 cells, and had no effect on SW260 cells (To et al., 2008b). A much more consistent effect (i.e. induction of ABCG2) was observed when myeloid leukemia cells (CMK, KG-1a, K-562m and HL-60) were treated with various HDACi (SAHA, TSA, VPA, and PB) (Hauswald et al., 2009). Interestingly, SAHA and TSA downregulated the expression of ABCG2 mRNA in the squamous cell carcinoma of the head and neck cell lines HSC-2 and KUMA-1 (Chikamatsu et al., 2013). The effect elicited by specific HDACi appears to be cell line specific. This may be attributed to the observations that the effect of some HDACi on ABCG2 expression may not be solely related to their traditional mechanism of action. For example, in S1 cells, romidepsin increased the acetylation of heat shock protein (HSP)-70 which through an unclear mechanism reduced the association of Hsp90 with AHR (To et al., 2011). Consequently, AHR was able to enter the nucleus and bind the ABCG2 gene to induce gene expression. Additionally, various HDACi has been shown to upregulate the expression and surface location of the polymorphic Q141K ABCG2 variant by facilitating ABCG2 gene expression and inhibiting retrograde transport of the protein into aggresomes (Basseville et al., 2012).

1.4.4 Cell signalling To date, the role of cell signalling in the regulation of ABCG2 expression is under-studied. While some data are available, the evidence is sparse, and the downstream molecular targets (e.g. transcription factors) that directly modulate ABCG2 expression remain to be identified. However, precisely because there is a dearth of information on how ABCG2 is regulated by cell signalling, there is a need to conduct more research in this regard.

1.4.4.1 PI3K/AKT signalling. The importance of the AKT pathway in ABCG2 expression was first demonstrated by Mogi and colleagues (Mogi et al., 2003). Mogi et al. (2003) observed that bone marrow cells isolated from Akt1-/- mice but not Akt2-/- mice had significantly reduced ABCG2-high expressing “side population” (SP) cells. Pharmacological inhibition of basal Akt activation using the PI3K inhibitor LY294002 resulted in a significant reduction in the SP fraction and conversely, lentivirus-mediated overexpression of constitutively active Akt increased the SP fraction. Using fluorescence microscopy, the authors attributed this change in SP fraction to translocation of ABCG2 and not to an overall change in ABCG2 expression. Some of these findings were again demonstrated by Dr. Yuichi

35 Sugiyama’s group using LLC-PK1 porcine kidney epithelial cells stably expressing human ABCG2 (Takada et al., 2005). More recently, it was reported that neurospheres deficient in PTEN (a negative regulator of PI3K) contained a higher fraction of SP cells and treatment with LY294002 effectively reduced this SP fraction, with a shift in ABCG2 localization from the membrane to cytoplasm, with no change in ABCG2 protein and mRNA expression (Bleau et al., 2009). There is a growing body of evidence to suggest that in addition to regulation of ABCG2 subcellular localization, the PI3K/AKT pathway also plays a very important role in modulating ABCG2 expression. Hartz et al. (2010b) showed that isolated rat brain capillaries treated with LY294002 or the AKT inhibitor triciribine had significantly reduced expression of functional ABCG2 dimers (with modest effects on monomeric ABCG2). These results were similarly observed after E2 treatment. It was further revealed that through non-genomic mechanisms, estrogen inhibited the activation of Akt and induced the activation of PTEN and GSK3. The effects of estrogen on ABCG2 expression were blocked using PTEN inhibitor bpV (Hopic), GSK3 inhibitor XIII, and proteasome inhibitor Lactacystin. These results led the authors to suggest that the PTEN/PI3K/AKT/GSK3 pathway mediates estrogen-induced proteasomal degradation of functional ABCG2 dimer (Hartz et al., 2010b). It is not apparent how and why degradation is specific to dimeric ABCG2 and spares the monomeric form of the protein. The PI3K/AKT pathway has recently been implicated in the transcriptional regulation of ABCG2. Wang et al. observed that ectopic overexpression of OCT4 in human hepatocellular carcinoma PLC cells increased phosphorylated/activated AKT and increased ABCG2 transcript expression. This effect was blocked by the PI3K inhibitor LY294002. It should be noted that from this study, since only a PI3K inhibitor was used, it is not clear which downstream effectors/transcription factors are involved. One candidate may be NF-κB (p65). As mentioned above, it has been demonstrated that ABCG2 expression is upregulated in MCF- 7 cells transfected with HER2/ERBB2. Zhang et al. showed that these cells have elevated AKT and MAPK activity, and increased nuclear localization of p65 (Zhang et al., 2011). Interestingly, ABCG2 protein expression was downregulated after treatment with the NF-κB inhibitor BAY11-7082 and PI3K inhibitor LY294002 but not with MEK inhibitor PD98059. Similarly, BAY11-7082 and LY294002 reduced luciferase reporter activity driven by the human ABCG2 promoter. For these reasons, it was suggested that elevated PI3K/AKT/ NF-κB signalling contributed to the upregulation of ABCG2 in HER2 overexpressing MCF-7 cells

36 (Zhang et al., 2011). However, more work is required to support this claim since it was not demonstrated whether NF-κB is recruited to the native ABCG2 gene and if so, whether it could be modulated by inhibiting AKT activity. The AKT signalling pathway is also implicated in the upregulation of ABCG2 expression in gefitinib resistant A431/GR cells. Huang et al. (2011) showed that knockdown or pharmacological inhibition of AKT reduced ABCG2 expression (mRNA and protein) in these cells. These authors further demonstrated that AKT phosphorylated EGFR which promoted the localization of EGFR to the nucleus where it was bound to an AT-rich minimal consensus sequence in the proximal ABCG2 promoter.

1.4.4.2 MAPK signalling. Induction of ABCG2 expression (mRNA and protein) by EGF involved activation of the MAPK signalling cascade. Meyer zu Schwabedissen al. (2006) showed that co-treatment with the EGFR/ERBB1 tyrosine kinase inhibitor AG1478 and MEK inhibitor PD98059 attenuated the induction of ABCG2 expression by EGF in BeWo cells. These results were similarly observed in MCF-7 cells and human cytotrophoblasts. In mouse kidney side population cells, overexpression of MEK was sufficient to increase MAPK activity and induce mouse Abcg2 protein and mRNA expression (Liu et al., 2013). It is important to note that MAPK activation alone should not be regarded as a predictor of ABCG2 upregulation. For example, while PD98059 was able to block the inductive effects of EGF on ABCG2 expression (Meyer zu Schwabedissen et al., 2006), it had no effect on ABCG2 expression in ERBB2/HER2-overexpressing MCF-7 cells with elevated MAPK signalling (Zhang et al., 2011). It is likely that multiple signalling networks must integrate to regulate ABCG2 gene expression.

37 1.4.5 Species. As indicated above, the genomic distance separating the promoters of the alternative mRNA isoforms differs quite substantially between mouse and human. For this reason, it is still not entirely clear whether the mouse is a good in vivo model for studying the regulation of ABCG2 in human. Indeed, certain stimuli do show clear species-specific responses. For example, aryl hydrocarbon receptor agonists are robust inducers of human ABCG2 but have no effect on mouse ABCG2 expression in vitro and in vivo (Tan et al., 2010). However, a survey of the limited data available for the response observed for various stimuli or physiological states under different experimental systems (Table III) shows that in the majority of cases, there is good agreement between species. Particularly, similar responses were observed when species were compared based on ex vivo and in vivo experimental systems. Therefore, in the absence of further experimental evidence, the mouse in vivo may still serve as a good model for studying regulation of ABCG2 in humans.

38 Table III. Effects of various stimuli and physiological states on ABCG2 expression. Response observed under different experimental systems* In vitro Ex vivo (and 1° cells) In vivo Stimuli/State Homo. Mus Rat Other Homo. Mus Rat Other Homo. Mus Rat Other

Estrogen - = + - -

Progesterone = +

Testosterone = +

Glucocorticoids - + - = +

IL-1β - + -

TNFa - + -

IL-6 - - =

Oncostatin M + +

IGF-II +

EGF + +

TGF-B -

SHH +

AHR agonists + = + = +

PPAR agonists + +

PXR agonists + + + + = +

CAR agonists + + + +

Nrf2 agonists +

RAR/RXR + + activators

COX-2 = +

Hypoxia +

Low folate - = +

miRNA -

Promoter - methylation Histone + modifications

Male + = + +

Lactation + + +

Inflammation + - - (liver) * Response is categorized as induced (+), reduced (-), or unchanged (=) in experimental systems from human (homo.), mouse (mus), rat, or in some cases bovine or porcine (other). Note that often a variable response is observed and may depend on whether the assessment is at the mRNA or protein level. Blank boxes indicate data not available.

39 1.5 Mammary Gland Biology

1.5.1 General Anatomy The mammary gland basically serves only one biological purpose and that is to produce milk (i.e. lactate) to nurture the offspring. It is most structurally and functionally developed during lactation and mainly consists of two compartments: epithelial (ductal) and connective (stromal) tissues (Hennighausen and Robinson, 2005). The epithelial compartment includes the ducts that are important for the transport or storage of milk and the alveoli which contain differentiated luminal epithelial cells that produce and secrete milk. The luminal epithelial cells that line the ducts and alveoli sit above a basal layer of myoepithelial cells that contract in response to oxytocin to eject milk in a process known as milk let-down (Richert et al., 2000). These ductal and alveolar structures are surrounded by connective tissues mainly made of adipocytes, blood vessels, and immune cells. How the mammary gland develops into this ultimate state has been an active area of research, particularly since this may uncover mechanisms responsible for breast cancer.

1.5.2 Development Mammary gland development can be divided into several key stages: embryonic, prepubertal, pubertal, pregnancy, lactation, and involution (schematically shown in Fig. 3 and reviewed by Watson and Khaled, 2008). This section will give a macroscopic overview of mammary gland development with a few examples of major player(s) involved at each stage of development.

1.5.2.1 Embryonic. In mice, mammary gland development begins around embryonic day 10 (E10) with the formation of the milk line from the ectoderm that will give rise to 5 pairs of mammary placodes (E12) (Cowin and Wysolmerski, 2010). These mammary placodes then grow and penetrate into the underlying mesenchyme, forming mammary buds made of epithelial cells (E15). These buds further proliferate to form simple epithelial ductal structures that penetrate into part of a fat pad, forming a rudimentary gland consisting of a central (primary) duct with several branches (E18). At this point, development of the mammary gland is almost arrested until puberty (Watson and Khaled, 2008).

40 1.5.2.2 Prepubertal. As indicated above, there are very little changes to the mammary gland during this stage of development except for allometric/isomorphic growth in which the size of the gland and its ductal structures increases in proportion to the body (McNally and Martin, 2011; Watson and Khaled, 2008).

1.5.2.3 Puberty. Up until this point, mammary gland development has mainly occurred independent of hormones (Brisken and O’Malley, 2010). With the increase in serum ovarian hormones during puberty, club/spoon shaped structures called terminal end buds (TEB) appear at the ends of the ducts (Hennighausen and Robinson, 2005; Watson and Khaled, 2008). These TEBs contain an outer layer of cap cells that proliferate to form the basal myoepithelial cells and a dense inner clump of body cells that proliferate to form the luminal epithelial cells. These TEBs continue to elongate and branch into the mammary fat pad until it reaches the limits of the fat pad where it regresses to form terminal ducts (Richert et al., 2000). In mice, this process starts around 3-5 weeks and continues until 10-12 weeks by which a mature virgin mammary gland is formed (Richert et al., 2000). Pubertal mammary gland development is highly dependent on the interplay of various hormones and growth factors but most would agree that estrogen and its receptor likely play a very major role (McNally and Martin, 2011). This is based on the observation that donor mammary epithelium from estrogen receptor alpha (ER α) knockout mice that were transplanted into the fat pad of pre-pubertal wild-type mice, failed to form ductal structures after puberty or even during pregnancy (Mallepell et al., 2006). While estrogen appears to be crucial for normal ductal outgrowth during puberty, in the mature non-lactating mammary gland, progesterone plays a role in the formation of side branches and alveolar buds that grow and regress with each estrous cycle (Fata et al., 2001; Hennighausen and Robinson, 2005; Lydon et al., 1995).

1.5.2.4 Pregnancy. With the onset of pregnancy, the mammary gland undergoes dramatic developmental changes. The early phase of pregnancy is characterized by the proliferation of ductal structures to form side branches and alveolar buds that are reminiscent of that observed during diestrus of the estrous cycle (Richert et al., 2000). This phase of development appears to be highly dependent on progesterone since donor mammary epithelium from progesterone receptor knockout mice fail to form side branches and alveolar structures even after being exposed to the hormone environment of the pregnant recipient mice (Brisken et al., 1998). In

41 the later phase of pregnancy, the alveolar buds develop into well-formed clusters of alveoli in a process often described as “lobuloalveolar” growth. The mammary gland also becomes more vascularized with comparatively fewer adipocytes. Near the end (day 18) of the approximately 21 day pregnancy period in mice, the alveoli are fully differentiated and capable of producing milk (Richert et al., 2000). However, the alveoli are still ‘leaky’ without well-formed tight junctions and as such milk constituents can still enter the blood stream and vice versa. The proliferation and differentiation of the alveoli is dependent on the actions of the major lactogenic hormone prolactin. Heterozygous mice with only one copy of the prolactin receptor gene have impaired lobuloalveolar growth (Brisken et al., 1999). Furthermore, mammary epithelium implants from prolactin receptor knockout mice were able to develop side branches but displayed a lack of lobuloalveolar growth and functional differentiation after pregnancy (Brisken et al., 1999).

1.5.2.5 Lactation. Although the mammary gland at parturition is already fully developed to produce milk (i.e. has undergone complete ‘secretory differentiation’), the main trigger for the production of copious amount of milk (i.e. ‘secretory activation’ characterized by a change in milk composition that resembles established lactation) is progesterone withdrawal (Neville and Morton, 2001; Neville et al., 2002; Pang and Hartmann, 2007). Progesterone withdrawal is crucial for the closure of the tight junctions around the alveoli to establish a physical barrier between the milk compartment and the blood. This is supported by the observation that ovariectomy or treatment with the progesterone receptor antagonist RU-486 triggered tight junction closure in the mammary alveoli of mice at late-gestation (Nguyen et al., 2001). During pregnancy, progesterone also represses prolactin receptor and milk protein β-casein expression, therefore progesterone withdrawal enhances the expression of these genes (Nishikawa et al., 1994). After secretory activation (by progesterone withdrawal in the presence of prolactin), the maintenance of lactation is mainly due to prolactin. Support for this comes from well-documented observations (reviewed by Neville and Walsh, 1995) that lactation is suppressed by drugs that inhibit the secretion of prolactin but it is enhanced/promoted by drugs that induce prolactin secretion. Furthermore, direct evidence from mice has shown that the suppression of lactation by bromocriptine, a drug that inhibits prolactin secretion, can be rescued by prolactin administration (Knight et al., 1986). During established lactation, as more fat becomes utilized for milk synthesis, adipocytes becomes less

42 visible and the mammary gland becomes mainly made up of alveolar epithelial cells (Richert et al., 2000).

1.5.2.6 Involution. Loss of suckling stimulus due to weaning of the offspring triggers the cessation of lactation. The importance of the suckling stimulus is best demonstrated by the observation that lactation in mice can be artificially prolonged by continually fostering nursing pups to the lactating mother (Hadsell et al., 2005). However, while suckling is an inducer of prolactin secretion, the main trigger for involution appears to be a loss of milk removal from the mammary gland (i.e. milk stasis). This was elegantly demonstrated using a mouse model in which one side of the fourth inguinal mammary gland was sealed and other side was left open (Li et al., 1997). Despite the presence of suckling stimuli which occurred through other glands that remained opened and even within the same hormone environment, the sealed gland involuted. This demonstrated that local factors, which are activated by milk stasis, contribute to involution (Li et al., 1997). Multiple mechanisms appear to be involved in involution, which follows a sequence of events starting with apoptosis of the epithelial cells, subsequent tissue remodelling by proteases and active removal of tissue debris by immune cells (Watson, 2006; Watson and Kreuzaler, 2011). Ultimately, the entire ductal/alveolar network is remodelled such that by the end of involution, the gland resembles that of pre-pregnancy (Richert et al., 2000).

43

Figure 3. Schematic overview of postnatal mammary gland development. The rudimentary mammary gland developed in the embryo is arrested from further development but grows only proportional to the body (allometric growth) until puberty. At puberty, with increased serum estrogen, terminal end buds (TEB) appear and grow outward and penetrate further into the mammary fat pad. As these TEBs elongate and bifurcate, they create a tree of ductal structures that ends at the limit of the mammary fat pad. With the onset of pregnancy, the surge in progesterone causes the formation of side branches and alveolar buds. With prolactin stimulation, these buds further differentiate into alveolar structures capable of milk secretion (secretory differentiation). At term, progesterone withdrawal results in secretory activation and the mammary gland is now capable of producing copious amounts of milk. As lactation continues, the ratio of epithelial content to fat cells increases. After the cessation of lactation, the mammary gland remodels to a form similar to that before pregnancy but may contain more ductal structures. Based on information described in section 1.5 (Mammary Gland Biology) and in particular, in Brisken and O’Malley (2010), and in Hennighausen and Robinson (2005).

44 1.6 Prolactin and Prolactin Receptors

1.6.1 The ligand. Prolactin (PRL) is a 23-kDa peptide that shows sequence and structural similarities to placental lactogen and growth hormone (Goffin et al., 1996a). It is principally produced and secreted by lactotrophs (or mammotrophs) in the anterior lobe of pituitary gland (reviewed by Freeman et al., 2000), but may also be produced at extra-pituitary sites such as the mammary gland (Clevenger et al., 1995a; Shaw-Bruha et al., 1997; Touraine et al., 1998) . The controlled secretion of prolactin from the pituitary is very complex and can be modulated by both external (light, stress, odour, etc.) and internal stimuli (e.g. hormones) (Freeman et al., 2000). The best-characterized regulatory mechanism of PRL secretion is the hypothalamic- pituitary axis. It is generally thought that PRL is constitutively produced by lactrotrophs but that its secretion is constantly suppressed by PRL inhibiting factors (most notably the neurotransmitter dopamine) that are released from neuroendocrine neurons in the hypothalamus (Freeman et al., 2000; Ben-Jonathan et al., 2008). Stimuli such as those during nursing, which alter the secretion of these PRL-inhibiting factors, result in increased PRL secretion. Alternatively, PRL secretion and synthesis may be enhanced by PRL releasing factors (such as oxytocin, estrogen, thyrotropin releasing factors) that act directly on the lactotrophs. A normal mean serum PRL concentration in adults is around 5-10 ng/mL and does not deviate significantly from this range throughout the menstrual cycle (Guyda and Friesen, 1973). The serum concentration of PRL steadily increases during pregnancy from approximately 30 ng/mL in the first trimester to approximately 200 ng/mL at term (Hwang et al., 1971). After delivery, basal PRL concentrations decrease to approximately 30 ng/mL (1- 1.5 mo. postpartum) to 10 ng/mL (2-6 mo. post-partum) during lactation (Noel et al., 1974). However, depending on the stage of lactation, serum PRL concentration increases to a mean range of 80-250 ng/mL at 30 min following suckling/nursing and returns to baseline values after 2-3 h (Noel et al., 1974).

1.6.2 Prolactin receptor (PRLR). While radioligand binding assays and antibodies targeting the PRLR have revealed the existence of a high affinity receptor, it was not until 1988 that the prolactin receptor gene was cloned and characterized from rat liver cDNA (Boutin et al., 1988). This was followed by the cloning of the long form of the human PRLR from HepG2 and T- 47D cells (Boutin et al., 1989). Since these initial discoveries, there are now at least six

45 different human PRLR isoforms, all of which are the result of alternative RNA splicing. Due to alternative splicing and a frameshift in the open reading frame that resulted in a stop codon 13 residues following the splice site, the intermediate isoform lacks a very large part of the intracellular domain (Kline et al., 1999). Unlike the long form of the PRLR, the intermediate isoform failed to induce proliferation when transfected into Ba/F3 cells, which suggested that the loss of the intracellular domain resulted in functional deficits. Indeed, despite similar ability to activate JAK2 (although the ability to activate STAT5 was not studied), the intermediate isoform lacked the ability to activate Fyn kinase (Kline et al., 1999). The short isoforms of the human PRLR partially (S1a) or completely (S1b) lack exon 10, and are alternatively spliced to exon 11 (Hu et al., 2001). Similar to the intermediate isoform, these short isoforms lack part of the intracellular domain and therefore lack the ability to activate JAK2/STAT5 signalling. However, perhaps most interesting is the ability for these short isoforms to act as dominant negative forms of the PRLR. Unlike the intermediate and short isoforms, the circulating prolactin binding protein (PRLBP), identified in serum and with an apparent molecular weight of 32 kDa, contains only the extracellular domain of the long form of the PRLR (Kline and Clevenger, 2001). Therefore, this isoform does not play a role in intracellular PRL signalling, but instead antagonizes PRL actions in vitro by preventing PRL from activating cell surface PRLR. In contrast to the isoforms described above, which differ from the long form of the PRLR only at the intracellular domain and therefore have similar affinity for PRL, the delta S1 (∆S1) isoform lacks the S1 (ligand binding) region within the extracellular domain of the PRLR (Kline et al., 2002). As a result, the ∆S1 isoform has a 7- fold reduction in affinity to bind PRL. However, since the intracellular domains are intact, this isoform can activate similar signalling cascades activated by the long form of the PRLR (described below).

1.6.3 Prolactin receptor signalling While the lactogenic properties of PRL and its receptor have been known for many years, the signalling cascade elicited by PRL was only characterized within the last two decades. Much of what is known about PRLR signalling today is based on discoveries made during the period from the late 1990s to early 2000s using rat and mouse cell lines. In particular, the PRL- dependent pre-T lymphoma (Nb2) rat cell line and the mouse interleukin-3 dependent pro-B cell (Ba/F3) were instrumental to characterizing the PRLR/JAK2, MAPK, and PI3K pathways

46 whereas STAT5 activation was characterized using HC11 mouse mammary epithelial cells. Please note that only well characterized (or loosely defined here as ‘classical’) signalling cascades in prolactin signalling will be reviewed (Fig. 4). Other minor signalling pathways will be discussed, where applicable, in the discussion section of this thesis.

1.6.3.1 JAK2/STAT5 pathway. The JAK2/STAT5 pathway was one of the earliest pathways shown to be activated by PRL and still remains the best characterized pathway in PRL signalling. The tyrosine kinase JAK2 is constitutively associated with the PRLR even prior to prolactin stimulation (Dusanter-Fourt et al., 1994; Lebrun et al., 1994; Rui et al., 1994). Binding of PRL to its receptor induces the autophosphorylation of tyrosine residues on Janus Kinase 2 (JAK2) and phosphorylation of tyrosine residues on the PRLR. Early studies on the kinetics of this phosphorylation event had quickly established that phosphorylation of JAK2 preceded that of the prolactin receptor and therefore is a proximal protein in the PRL signalling cascade (Lebrun et al., 1994; Rui et al., 1994). Indeed, JAK2 can certainly be regarded as the major kinase involved in many aspects of PRLR signalling. The association of JAK2 with the PRLR appears to mediated by a BOX-1 proline rich motif in the intracellular domain of the prolactin receptor (Lebrun et al., 1995a). However, the presence of this region alone, which binds JAK2, is not sufficient for transcriptional upregulation by prolactin. Instead, more C- terminal intracellular domains are required. In particular, the C-terminal tyrosine residue of the PRLR is crucial for efficient PRL-mediated gene transcription (Lebrun et al., 1995b). These tyrosine residues serve as docking sites for the association of signal transducer and activator of transcription-5 (STAT5) to the PRLR where it is phosphorylated by JAK2 (Pezet et al., 1997). The tyrosine phosphorylation of STAT5 by JAK2 induces its dimerization and DNA binding activity (Gouilleux et al., 1994). STAT5, originally named mammary gland factor (MGF) for its ability to bind the β-casein (milk protein) gene promoter (Schmitt-Ney et al., 1991), was first cloned from the cDNA of lactating sheep mammary gland (Wakao et al., 1994). This was followed by the cloning of the mouse Stat5a and the alternative isoform Stat5b (Liu et al., 1995). STAT5A and STAT5B are encoded by separate genes but have a greater than 94% similarity in their protein sequences that mainly differ by a few amino acid residues in the C- terminus (Ambrosio et al., 2002; Liu et al., 1995). Both isoforms are known to bind gamma interferon activated sequences (GAS motifs with consensus TTCNNNGAA) to induce gene transcription (Liu et al., 1995, 1996). Furthermore, STAT5A and STAT5B are expressed in

47 the mammary gland and show increased activation during late pregnancy and early lactation (Liu et al., 1996). STAT5A is the predominant isoform expressed in the mammary gland (Liu et al., 1995) and is crucial for mammary gland development. Stat5a knockout mice have retarded lobuloalveolar outgrowth after parturition and therefore are not able to rear their pups (Liu et al., 1997). In addition, Stat5a knockout mice have dramatically reduced expression of the Wap gene but similar expression of Csn2, which was attributed to the limited degree of compensation by Stat5b. Interestingly, Stat5b knockout mice display a different set of deficits compared to Stat5a knockout mice and anecdotally appear to have less severe impairment of mammary gland development (Udy et al., 1997). Therefore, in the mammary gland, STAT5A is considered the more important isoform. It is now recognized that JAK2/STAT5 is the major pathway involved in the induction of milk proteins (Burdon et al., 1994; Gouilleux et al., 1994; Li and Rosen, 1995) in the mammary gland during lactation and in general, mediates the expression of the majority of genes induced by prolactin.

1.6.3.2 Mitogen Activated Kinase Pathway. In addition to STAT5, phosphorylated tyrosine residues on the PRLR can also serve as docking sites for adaptor proteins. One such protein is the Src-homology 2 containing protein (SHC) (Erwin et al., 1995). Recruitment of SHC to the PRLR complex is followed by the phosphorylation of tyrosine residues on SHC by JAK2 (or other associated kinases). This allows for Growth factor receptor-bound protein 2 (GRB2) to bind via its SH2 domain to SHC (Das and Vonderhaar, 1996a). GRB2 forms a complex with son of sevenless (SOS), which as a guanine nucleotide exchange factor, induces the activity of rat sarcoma protein (Ras) by promoting the exchange of GDP to GTP (Das and Vonderhaar, 1996a; Erwin et al., 1995). Activation of Ras results in downstream serine/threonine phosphorylation and activation of Rapidly Accelerated Fibrosarcoma (Raf or MAP3K) serine/threonine kinase, which is the first member of the MAPK pathway (Clevenger et al., 1994). Activated Raf then phosphorylates and activates MEK1/2 (MAP2K) tyrosine/threonine kinase, which then activates extracellular signal regulated kinase1/2 (ERK1/2, p42/p44, or MAPK) (Das and Vonderhaar, 1996b). ERK1/2 then phosphorylates serine/threonine residues on downstream effectors such as transcription factors, and other protein kinases or phosphatases, to modulate their activity (Cargnello and Roux, 2011; Yoon and Seger, 2006). Although not as well studied, in addition to ERK1/2, prolactin has also been shown to stimulate the phosphorylation and activation of c-Jun N-terminal Kinase (JNK) (Cheng et al.,

48 2000; Olazabal et al., 2000; Schwertfeger et al., 2000). Activated JNK phosphorylates c-Jun of the AP-1 complex, which ultimately induces AP-1 mediated transcriptional activity.

1.6.3.3 PI3K/AKT pathway. The PI3K pathway is the third major pathway activated in PRLR signalling. Prolactin treatment has been shown to induce recruitment and tyrosine phosphorylation of insulin receptor substrate 1 (IRS-1) and phosphatidylinositide 3-kinase (PI3K) to the PRLR, resulting in increased activity of PI3K (Berlanga et al., 1997). It is not entirely clear which kinase is ultimately responsible for the activation of this pathway. The expression of functional JAK2 in COS and CHO cells is crucial for the tyrosine phosphorylation and activation of PI3K by prolactin (Yamauchi et al., 1998). However, the dependence of PI3K activation on JAK2 may be cell line specific. This is because in some cells, certain members (e.g. Fyn, c-Src) of the Src family of kinases (SFK), which are bound to the PRLR even prior to PRL stimulation, are activated by PRL treatment or by antibody- mediated dimerization of the PRLR (Clevenger and Medaglia, 1994; Fresno Vara et al., 2000, 2001). This effect appears to be independent of JAK2 (Fresno Vara et al., 2000). In PRL treated rat T-lymphoma Nb2 cells, PI3K was not co-immunoprecipitated with JAK2 (al-Sakkaf et al., 1996) but instead was co-immunoprecipitated with Fyn (src kinase) (al-Sakkaf et al., 1997). This suggested that the activation of PI3K by prolactin could also be mediated by SFKs. Indeed, inhibition of SFKs has been shown to attenuate prolactin-induced activation of the PI3K pathway (Fresno Vara et al., 2001). Regardless of the means of activation, activated PI3K catalyzes the phosphorylation of membrane associated phosphatidylinositol-4,5- bisphosphate to phosphatidylinositol-3,4,5-trisphosphate, which recruits protein kinase B (also known as AKT) to the plasma membrane where it is phosphorylated and activated by 3- phosphoinositide-dependent kinases (PDK) (Al-Sakkaf et al., 2000). Activated AKT then phosphorylates downstream effectors to modulate their activity.

1.6.3.4 Other pathways. In addition to the three major pathways described above (JAK2/STAT5, MAPK, and PI3K), prolactin has also been shown to activate Protein Kinase C in NOG-8 mammary epithelial cells (Banerjee and Vonderhaar, 1992) and Focal Adhesion Kinase in T-47D human breast cancer cells (Canbay et al., 1997). Furthermore, experiments using a combination of Nb2, T-47D, and CHO expression cell systems have demonstrated that prolactin induces the activation of Tec tyrosine kinase (Kline et al., 2001) and Nek3

49 serine/threonine kinase (Miller et al., 2005). These kinases modulate the nucleotide exchange factor Vav, which together are associated as part of a complex with the PRL-bound prolactin receptor (Clevenger et al., 1995b; Kline et al., 2001; Miller et al., 2005). Activated Vav then induces Rac GTPase activity. It is important to note that these ‘other pathways’ may also be extensions of the major pathways described above but, in the context of PRL signalling/biology, have not received much attention. As such, their overall importance remains unclear.

Figure 4. Major signalling pathways activated by the prolactin receptor. The binding of prolactin (PRL) induces dimerization of the prolactin receptor (PRLR), which is followed by the autophosphorylation and activation of receptor-associated kinases JAK2 and SFK. Activated JAK2 then phoshorylates tyrosine residues on the PRLR to create docking sites for proteins to bind. STAT5 is recruited to one of these sites where it too becomes phosphorylated by JAK2. Phosphorylated STAT5 then dissociates from the PRLR, dimerizes, and translocates to the nucleus to modulate gene transcription. These sites may also bind adaptor proteins such as SHC and IRS that ultimately leads to the activation of the MAPK (GRB2/SOS/Ras/Raf/MEK/ERK) and the PI3K/AKT pathway, respectively. By an unclear mechanism, the PI3K/AKT pathway can also be activated by receptor associated SFK. Please note that these pathways are likely not mutually exclusive and that there may be significant cross-talk. Schematic is based on information from references described in section 1.6.3 (Prolactin receptor signalling).

50 1.7 Epigenetics

It is well recognized that besides the sequence encoded in the genome, “epigenetic” mechanisms also play a role in regulating gene expression. To date, various epigenetic mechanisms have been characterized. A detailed review of these different mechanisms and their function is beyond the scope of this thesis. The field is too large to ascribe what we know about epigenetics to particular paper(s). As such, this section will refer to a number of review papers and only provide a brief generalized overview of epigenetic mechanisms with specific focus on CpG islands and histone modifications.

1.7.1 CpG Islands CpG islands are stretches of DNA (<1 kb) that are rich in CpG dinucleotides (Jones, 2012). They are often found at promoters but may also be found outside of promoters in the intragenic/intergenic region (called orphan CpG islands) (Deaton and Bird, 2011). While the cytosine of CpG dinucleotides at promoter CpG islands are usually unmethylated, orphan CpG islands have a higher frequency of being methylated (5-methylcytosine), but their function is not as well known. Promoter CpG islands can become methylated during differentiation and this is associated with the tissue-specific expression of certain genes. Active mechanisms mediated by DNA methyltransferases (DNMT) catalyze the de novo and maintenance of CpG methylation. In contrast, the removal of these methyl CpG marks appears to be mediated by both passive and active mechanisms. Passive loss of CpG methylation may result from successive rounds of replication without maintenance of the CpG methylation pattern by DNMTs whereas the active removal of methylated CpG may involve demethylases that are not well characterized (Jones, 2012). Methylation of CpG islands results in transcriptional repression possibly through the direct inhibition of the transcription factor-to-DNA interaction (Deaton and Bird, 2011). However more complex mechanisms are also involved. This includes the recruitment of proteins such as methyl-CpG-binding domain (MBD) proteins, which recruit histone modifying enzymes that alter the accessibility of the chromatin, or by the recruitment of polycomb receptor complexes that inhibit the activity of RNA polyermase (Deaton and Bird, 2011).

51 1.7.2 Histone Modifications A nucleosome, the basic structure of chromatin, contains an octamer of one pair of each histone (H3, H4, H2A, H2B) wrapped around by DNA (Kouzarides, 2007). Histones may be modified at specific residues by methylation, phosphorylation, ubiquitination, sumoylation, ADP ribosylation, deamination, and proline isomerization (Kouzarides, 2007). Different modifications are associated with different functions. Certain histone modifications change the charge of the histone, thereby altering the histone-to-DNA interaction such that the DNA becomes more accessible, while other modifications serve as docking sites for proteins that participate in transcription, RNA splicing, DNA replication and repair (Kouzarides, 2007; Li et al., 2007; Zhou et al., 2014). It should be noted that it is very difficult to establish causality between a histone modification and a particular function since histone modifications are only a snapshot (i.e. a mark) of the state of the chromatin at a particular point in time and cannot be indicative of what events precede it or what may occur afterwards. Please refer to Table IV for examples of histone modifications and their associated function(s).

Table IV. Examples of histone modifications and their associated functions Histone Associated Functionb modificationsa H3K4me1 Poised or active distal enhancers (but may also mark promoters) H3K4me2 Poised or active enhancers and promoters H3K4me3 Mainly active promoters but may also mark active enhancers H3K9me1 Marks 5’UTR and promoter of genes H3K9me3 Inactive/transcriptional repression H3K9Ac Active promoters H3K27me3 Inactive/transcriptionally repressed promoter H3K27Ac Active enhancers H3K36me3 Elongation, active transcription, mark 3’ to the transcription start site a me, methylation; Ac, acetylation, at lysine residues (K) b Note that a histone mark can be associated with both “poised” (i.e. permissive) and active chromatin. Poised is when the chromatin is open to the recruitment of transcription factors or components of the transcriptional machinery, whereas active is when these factors are already bound to the DNA. Table includes information from multiple sources: Black and Whetstine, 2011; Calo and Wysocka, 2013; ENCODE Project Consortium et al., 2012; Harmston and Lenhard, 2013; Spicuglia and Vanhille, 2012

52 1.8 Research Rationale

In this thesis, I sought to understand how ABCG2 is upregulated in the mammary gland during lactation. Despite that we now have a good understanding of how ABCG2 is regulated (reviewed in Section 1.4 ‘Regulatory Control of ABCG2 expression’), none of these mechanisms offer an explanation as to how ABCG2 is upregulated during lactation. Hormones such as progesterone and estrogen, for which the molecular mechanisms involved in regulating ABCG2 expression are partly known, are likely candidates, but their serum concentration profile and physiological importance are mainly restricted to the estrus cycle and early-late pregnancy. There is evidence that the major lactogenic hormone prolactin induces ABCG2 in BeWo carcinoma cells, however the molecular mechanism is not known (Wang et al., 2008a). This would suggest that PRL may be responsible for the lactation-associated upregulation of ABCG2 in the mammary epithelium. For this reason, the first aim of this thesis (AIM1) was to determine whether prolactin induces ABCG2 in normal and cancer cells derived from the human breast and if so, what signalling mechanisms are involved. The effect of prolactin on ABCG2 expression was extensively investigated using human T-47D breast cancer cells since these cells are known to respond to PRL treatment and have an intact signalling network downstream of the PRLR. The role of JAK2/STAT5, MAPK, and PI3K pathways on the PRL- ABCG2 response was further studied using small molecule inhibitors or siRNAs targeting specific components of these three major signalling pathways. Lastly, the recruitment of STAT5 to a putative GAS motif (defined in silico) proximal to the ABCG2 promoter was examined by chromatin immunoprecipitation and the functional activity of this STAT5 binding site was interrogated using luciferase reporter assays.

Based on the results of AIM1, my second aim (AIM2) was to examine whether STAT5 also plays a role in the regulation of ABCG2 in the lactating mammary gland in vivo. Due to obvious ethical reasons, the mouse model was used. However, since there is no prior knowledge whether the Abcg2 transcript(s) is induced in the mouse mammary gland during lactation, I first investigated which of the mouse Abcg2 mRNA isoforms were expressed in the non-lactating (virgin) and lactating mammary gland. This information would help me gain insight into the genomic region of potential regulatory interest. Next, a forced-weaning model where pups were removed from lactating mice to trigger the cessation of lactation (involution)

53 was used to rapidly turn-off STAT5 activity in the mammary gland while preserving tissue morphology. The transcript and protein expression of mouse Abcg2 in the mammary gland of lactating mice was compared with mice at 24 h and 48 h after pup removal. Differential STAT5 recruitment to the mouse Abcg2 gene was also compared between lactating and non- lactating mouse mammary glands. Similar to the approach described in AIM1, the functional activity of these STAT5 binding sites was further interrogated using luciferase reporter assays.

There is accumulating evidence that in addition to prolactin/PRLR/JAK2/STAT5, milk protein gene expression is also under epigenetic control (Rijnkels et al., 2010). There is an inverse relationship between the level of DNA methylation at specific CpG sites and the level of expression of milk protein genes (Johnson et al., 1983; Rijnkels et al., 2013; Thompson and Nakhasi, 1985; Vanselow et al., 2006), a phenomenon that is conserved across several species (rat, mouse, and cow) and has been proposed as a mechanism to explain the tissue- and developmental stage-dependent expression of milk proteins. In addition to DNA methylation, there is also evidence of epigenetic control at the level of histone modifications. Specifically, compared to the liver, which does not express high levels of milk proteins during lactation, the casein gene cluster and Wap gene promoters in the lactating mouse mammary gland are enriched with the open chromatin histone markers dimethylated histone 3 on lysine 4 (H3K4me2) and acetylated histone 3 (H3Ac) (Rijnkels et al., 2013). Given that the pattern of expression of mouse Abcg2 gene parallels that of milk protein genes in the mammary gland, it would suggest that Abcg2 may also be under epigenetic control. Therefore, the third aim (AIM3) of this thesis was to determine the epigenetic profile (permissive or repressive) of the mouse Abcg2 gene in the virgin and lactating mammary gland. To this end, the effect of promoter methylation on transcription activity was assessed in vitro using luciferase reporter assays. The methylation level of the CpG island that encompasses the Abcg2 E1b mRNA isoform promoter was also assessed by bisulfite pyrosequencing. Finally, using previously published and publicly available ChIP-seq datasets, the Abcg2 gene locus was interrogated for enrichment of the open chromatin histone-mark H3K4me2 in both lactating mammary glands and mammary epithelial cells isolated from the virgin mammary gland.

54 2. METHODS

2.1 Reagents Human cell lines T-47D, MCF-10A, and MCF-7 cells were purchased from the American Type Culture Collection (Manassas, VA). Primary human mammary epithelial cells (HMEC) were purchased from Clonetics, Lonza (Walkersville, MD). Mouse HC11 cells were a gift from Dr. Jason Matthews (Department of Pharmacology & Toxicology, University of Toronto). Mouse EpH4 cells were generously provided by Senthil Muthuswamy (Princess Margaret Cancer Centre, University Health Network, Toronto, Ontario, Canada). Fetal Bovine Serum (FBS), recombinant human insulin, RPMI-1640, DMEM, and phenol-red free DMEM/F12 were from Wiscent Inc. (Montreal, Quebec, Canada). DMEM/F12, phenol-red free high glucose DMEM, sodium pyruvate, recombinant human epidermal growth factor (EGF), horse serum, and GlutaMAX were from Gibco, Invitrogen (Life Technologies, Inc., Burlington, Canada). TCDD, Actinomycin D, and cholera toxin were purchased from Sigma-Aldrich Canada Ltd (Oakville, Ontario). Growth factor reduced Matrigel (Cat# 354230) was from BD Biosciences. STAT5 inhibitor (N′-((4-Oxo-4H-chromen-3-yl)methylene)nicotinohydrazide ), PD98059, U0126, wortmannin, and LY294002 were from EMD Bioscience (San Diego, CA) and were dissolved to working concentrations in DMSO. FugeneHD and pGL4.23[luc2/minP] were purchased from Promega (Madision, WI). Short-interfering RNAs were purchased from Dharmacon, ThermoScientific (Lafayette, CO). Recombinant human and mouse PRL, and recombinant human growth hormone (GH) were obtained from Dr. A. F. Parlow at the NIDDK’s National Hormone and Peptide Program, Harbor-UCLA Medical Center (Torrance, CA), stored as lyophilized powder, and dissolved in PBS to working concentrations that were stored at -20°C (for prolactin) or -80°C (for growth hormone). All other chemicals were purchased from Sigma.

2.2 Animals FVB/N and C57BL/6 mice originally from Jackson Laboratory and Charles River Laboratories, respectively, were bred and housed at the Toronto Centre for Phenogenomics (TCP). All protocols performed were approved by the TCP Animal Care Committee and are in accordance with the Canadian Council on Animal Care guidelines. Female virgin and lactating (7d-10d) mice were between 12-14 weeks old. All lactating mice were first time mothers and litter sizes were controlled to 6 pups immediately after parturition. For gene expression analyses

55 comparing virgin and lactating mice, pups were removed 2-3 h prior to tissue collection. FVB/N mice were used for all forced weaning experiments. For forced weaning experiments to trigger involution of the mammary gland, pups were either kept with the mother (control) or removed from the mothers at 24 h and 48 h prior to tissue collection. Tissues were collected between 9:00AM to 12:00PM and either snap frozen in liquid nitrogen or immersed in RNAlater (QIAGEN Inc.). Tissues were subsequently stored at -80°C.

2.3 Cell Culture and serum starvation T-47D cells were maintained in RPMI-1640 supplemented with 10% FBS. For serum- starvation, T-47D cells were washed twice with phenol-red free RPMI-1640, and then incubated with serum-free and phenol red-free RPMI-1640 supplemented with 0.05% w/v fatty-acid free BSA and 0.01 mg/mL holo-transferrin. MCF-7 cells were grown in high glucose DMEM containing sodium pyruvate supplemented with 10% FBS. For serum starvation, MCF-7 cells were washed twice and incubated with phenol red-free DMEM supplemented with 1 mM sodium pyruvate, 4 mM GlutaMAX, 0.05% w/v fatty-acid free BSA, and 0.01 mg/mL holo-transferrin. MCF-10A cells were maintained in DMEM/F12 supplemented with 20 ng/mL EGF, 0.5 µg/mL hydrocortisone, 100 ng/mL cholera toxin, 10 µg/mL insulin, and 5% horse serum. For serum-starvation, MCF-10A cells were washed twice and incubated with phenol-red free DMEM/F12. HMEC were grown as per instructions provided by Clonetics, Lonza (Walkersville, MD). In brief, HMEC were grown in Mammary Epithelial Basal Medium (MEBM, Clonetics) supplemented with 0.4% bovine pituitary extract, 10 ng/mL EGF, 5 µg/mL insulin, 0.5 µg/mL hydrocortisone, 15 ng/mL amphotericin B and 30 µg/mL Gentamicin and used in the second passage after thawing (approximately 6th cell division). HMEC were serum-starved in phenol red-free MEBM or phenol-red-free MEBM supplemented with 5 µg/mL insulin and 0.5 µg/mL hydrocortisone. HC11 mammary epithelial cells were grown in RPMI-1640 supplemented with 10% FBS, 10 ng/mL EGF, 5 µg/mL insulin, 4.7 g/L HEPES, 2.2 g/liter sodium bicarbonate (growth medium). EpH4 cells were grown in DMEM/F12 supplemented with 2% FBS and 5 µg/mL insulin. All cells were maintained in a humidified incubator at 37°C under a 5% CO2 atmosphere.

56 2.4 Treatment with hormones or TCDD Prior to treatment, T-47D, MCF-7, MCF-10A, and where applicable, HMEC, were serum- starved overnight for 18-20 h. After overnight serum-starvation, cells were treated with 0-1000 ng/mL recombinant human prolactin in starvation medium for up to 24 h. HC11 cells were seeded at 1.5 x 105 cells/well on 12 well plates and grown for 2-3 days to 100% confluence. Cells were then subsequently incubated in maintenance medium as per Hovey et al., (2003) (RPMI-1640 supplemented with 8% FBS, 5 µg/mL insulin, 1 uM dexamethasone, 4.7 g/L HEPES, 2.2 g/liter sodium bicarbonate) containing 0.1% v/v PBS or induced with PRL (0.1-1 µg/mL) for 24 h and 72 h. Alternatively, after reaching 100% confluence, cells were grown for an additional 3 days in growth medium before treating with PRL for 24 h in maintenance medium. EpH4 mammary epithelial cells were grown and treated as previously described (Xu et al., 2009) with modifications. EpH4 cells were grown in DMEM/F12 supplemented with 2% FBS and 5 µg/mL insulin. For prolactin treatment, cells were first serum-starved for 24 h and then plated at 6 x 104 cells/well on poly-HEMA coated 24-well plates (Corning) in 0.5 mL of RPMI-1640 medium. After 24 h, 0.5 mL of treatment medium was added to each well such that EpH4 cells were treated in RPMI-1640 supplemented with 2% v/v Matrigel, 10 µg/mL insulin and 2 µg/mL hydrocortisone in the presence or absence of recombinant mouse prolactin (1 µg/mL). Forty-eight hours after treatment, cells from three wells were sequentially transferred into a 1.5-mL microfuge tube, pelleted by short centrifugation using a table top centrifuge, and pooled samples from 3 wells were used for subsequent analyses.

2.5 Small molecule inhibitors of STAT5, MAPK and PI3K signalling T-47D cells were serum-starved overnight and pre-treated with STAT5 inhibitor (200 µM), MEK inhibitors PD98059 (20 µM) and U0126 (10 µM), or PI3K inhibitors LY294002 (10 µM) and Wortmannin (25 nM), and appropriate concentration of DMSO vehicle control for 1 h. After 1 h, cells were treated with PBS (0.1% v/v) or 100 ng/mL PRL for 15 min to 6 h.

2.6 Isolation of mouse mammary epithelial cells (MEC) Mouse mammary epithelial cells were isolated using immunomagnetic negative selection with help from Kelvin Wang in Dr. Sean Egan’s Laboratory (Program in Developmental and Stem Cell Biology, Hospital for Sick Children, Toronto, Canada). The third and fourth pair of mammary glands were excised from virgin FVB mice and immediately placed into 5 mL

57 Complete EpiCult-B medium (Cat#05610, STEMCELL Technologies) supplemented with 5% FBS, and 300 U/mL collagenase and 100 U/mL hyaluronidase (Cat# 07912, STEMCELL Technologies). Mammary glands were dissociated at 37°C for 5 h with agitation. The insoluble lymph nodes were removed 2 h into the dissociation protocol. The resulting cell suspension was centrifuged at 350 g at room temperature for 5 min. The cell pellet was resuspended in 5 mL ice cold 1:4 mixture of HF solution (Hanks Balanced Salt Solution

Modified supplemented with 2% FBS, Cat#37150) and 4 M NH4Cl (Cat#07800, STEMCELL Technologies), and vortexed to lyse red blood cells. Cells were pelleted at 350 g for 5 min at room temperature and resuspended in 3 mL of pre-warmed Trypsin-EDTA (Cat#07901, STEMCELL Technologies) by triturating 30 times with a p1000 micropipettor. An additional 10 mL of ice-cold HF solution was added and the cell suspension was centrifuged at 350g for 5 min at room temperature. After removal of the supernatant, the cell pellet was resuspended in 2 mL of Dispase (5 mg/mL, Cat#07913, STEMCELL Technologies) supplemented with 0.5 mg/mL DNase I (BioBasic Inc. Markham, ON) by pipetting up and down 30 times with a p1000 micropipettor. The cell suspension was diluted with 10 mL of ice-cold HF solution and filtered through a 40 µm nylon strainer to remove cell clumps. The single cell suspension was centrifuged at 350 g for 5 min at room temperature. The cell pellet was resuspended in 170 µL HF solution supplemented with 0.5 mg/mL DNase I and incubated with 30 µL of antibody cocktail. The antibody cocktail was prepared by combining anti-mouse antibodies purchased from eBioscience (San Diego, CA) for mouse CD31 (Cat#14-0311-85), CD45 (Cat#14-0451- 85), Ter-119 (Cat#14-5921-85), and CD140a (Cat#14-1401-82) together to achieve a final concentration of 0.15 mg/mL each for anti-mouse CD31, CD45, and Ter-119 and 0.06 mg/mL anti-mouse CD140a. Cells were mixed gently and incubated at 4°C for 30 min, after which the cells were diluted in 1 mL of HF solution containing 0.5 mg/mL DNase I and centrifuged at 300 g for 10 min at room temperature. The cell pellet was resuspended in 80 µL of HF solution supplemented with DNase I and incubated with 20 µL of anti-rat IgG MicroBeads (Miltenyi Biotec Inc., San Diego, CA) for 15 min at 4°C. The cells were washed once more with 1 mL HF solution/DNase I and centrifuged at 300 g for 10 min at room temperature. The resulting cell pellet was resuspended in 500 µL of HF solution/DNase I and separated using autoMACS Pro Separator (Miltenyi Biotec Inc.).

58 2.7 RNA isolation from cells and tissues RNA was isolated from cells (cell lines and isolated mouse MEC) using the RNeasy Kit (QIAGEN Inc. Toronto, Canada) with on-column DNase treatment according to manufacturer instructions. Total RNA was isolated from mouse mammary gland and liver tissues using Qiazol lysis reagent (QIAGEN Inc.) as per manufacturer protocol. The resulting RNA pellet was re-dissolved in 87.5 µL of RNase-free H2O and treated with DNase I by adding 10 µL RDD and 2.5 µL DNase I stock solution (at 2.7 Kunitz units/µL; RNase-free DNase Set, QIAGEN Inc., Toronto, Canada). After incubating at room temperature for 15 min, the RNA was further purified using RNeasy mini columns. RNA was eluted from RNeasy columns using DEPC-treated H2O or RNase-free H2O. The RNA concentration was determined spectrophotometrically using Nanodrop 2000. RNA integrity was assessed by agarose gel electrophoresis and ethidium bromide staining for the presence of sharp distinct bands corresponding to 28s RNA and 18s RNA. RNA samples were either immediately reverse- transcribed or aliquoted and stored at -80°C.

2.8 cDNA synthesis - reverse transcription For samples from human cell lines, RNA (2 µg) was reverse transcribed to cDNA in a 20 µL buffered reaction (50 mM Tris-HCl pH 8.3, 75 mM KCl, 3 mM MgCl2, 10 mM DTT) containing MMLV-reverse transcriptase (200 U), dNTP (0.5 mM of each), random primers (120 ng) and RNaseOUT (20 U) (Invitrogen, Life Technologies Inc., Burlington, Canada). The reaction was incubated at 25°C for 10 min, then 42°C for 50 min, and then heat inactivated by incubating at 70°C for 15 min. For samples from mouse cell lines or tissues, RNA (2 µg) was first incubated at 65°C for 5 min in a solution containing Oligod(T)16 (from Applied

Biosystems) and dNTP made up to 12 µL with RNase-free H2O. Immediately following this incubation, the samples were placed on ice and reverse transcriptase buffer, DTT, and RNaseOUT was added. The sample was then heated to 42°C for 2 min after which 200 U of SuperScript II (Invitrogen, Life Technologies Inc.) was added. The final reaction (2 µg RNA, 2.5 µM Oligo(dT), 0.5 mM each dNTP, 40 U RNaseOUT, 50 mM Tris-HCl pH 8.3, 75 mM

KCl, 3 mM MgCl2, 10 mM DTT) was incubated at 42°C for 50 min and then heat inactivated at 70°C for 15 min. The cDNA samples were then stored at -80°C until analyzed by real-time PCR.

59 2.9 Real-time RT-PCR to assess gene expression For quantitation of gene expression, cDNA derived from human (100 ng) or mouse (40 ng) samples were amplified in a 20 µL reaction by real-time PCR using inventoried TaqMan probe and primer sets (Table V) and 1X TaqMan Universal PCR Master Mix, no AmpErase® UNG, from Applied Biosystems (Life Technologies, Burlington, Canada). Mouse Abcg2 alternative mRNA isoforms (E1a, E1b, and E1c) were amplified in a 20 µL reaction containing 1X Power SYBR Green PCR Master Mix and 400 nM each of a unique forward and common reverse primer, previously published by Zong et al., (2006) (sequence given in Table V). The PCR products from these alternative isoforms were inserted into pCR2.1-TOPO (Invitrogen) by TA- cloning, transformed into One Shot Chemically Competent TOP10 E. coli (Invitrogen), and plated into LB plates containing 100 µg/mL ampicillin and X-gal for blue/white screening. Colonies were screened by sequencing and plasmids were prepared by QIAprep Spin Miniprep Kit (QIAGEN Inc.). Plasmids containing one copy of the PCR product (pCR2.1-E1a, pCR2.1- E1b, and pCR2.1-E1c) were linearized using HindIII and serial diluted (10-fold) to generate standard curves for absolute quantitation of Abcg2 mRNA isoforms using the standard curve method. Plasmid copy number was determined using OligoCalc (http://basic.northwestern.edu/biotools/OligoCalc.html) (Kibbe, 2007). Relative mRNA expression was determined using the 2-∆∆Ct method (Livak and Schmittgen, 2001). For experiments addressing the role of prolactin in human-derived cells, ∆∆Ct values were calculated by normalizing to the cycle threshold (Ct) value for GAPDH. For samples derived from mice, ∆∆Ct values were calculated by normalizing to the Ct of Krt-18, Gapdh, or Actb. The specific gene used for normalization is indicated in the figure legends. All real-time RT- PCR was performed using the ABI 7500 system (Applied Biosystems) set at 50°C for 2 min, 95°C for 10 min, and then 40 cycles of 95°C for 15 s and 60°C for 1 min.

2.10 Preparation of crude membrane fraction and whole cell lysate from human-derived cells Crude membrane fractions were prepared as described previously (Wang et al., 2003) with modifications. Each T225 flask of T-47D cells was washed twice with ice cold PBS and cells detached with a rubber scraper in ice-cold PBS containing 1 mM PMSF. The cell suspension was pelleted by centrifugation at 150 g for 5 min at 4°C, the cell pellet was resuspended in hypotonic lysis buffer (10 mM KCl, 1.5 mM MgCl2, 10 mM Tris-HCl, 1 mM EDTA, 1 mM

60 PMSF, 1X protease inhibitor cocktail, pH 7.4) and incubated on ice for 10 min. Cells were homogenized with a Dounce Homogenizer, and removal of plasma membrane was confirmed by visual inspection under a microscope. The homogenate was centrifuged at 1500 g for 10 min at 4°C, and the resulting supernatant was collected and centrifuged at 100,000 g for 60 min at 4°C. The crude membrane pellet was resuspended in 200 µL resuspension buffer (10 mM Tris-HCl, 0.25 M Sucrose, 150 mM NaCl, 1 mM PMSF, 1X protease inhibitor cocktail, pH 7.4), passed through a 26-gauge needle 10 times, and aliquoted. Aliquots were stored at -80°C. For preparation of whole cell lysate from T-47D, MCF-7, MCF-10A, and HMEC, cells were washed twice with ice-cold PBS (containing 1 mM Na3VO4 and 2.5 mM NaF if lysate is to be used for detection of phosphorylated proteins) and lysed with RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1% w/v NP-40, 0.5% w/v sodium deoxycholate, 0.1% w/v SDS) containing 2 mM Na3VO4, 5 mM NaF, 2 mM EDTA, 1 mM PMSF, and 1X protease inhibitor cocktail. Cells were scraped with a rubber policeman, transferred to a microcentrifuge tube and rotated end-over-end at 4°C for 20 min. Cellular debris was pelleted by centrifugation at 10000 rpm for 10 min at 4°C. The supernatant was collected, aliquoted, and stored at -80°C. Protein concentration was measured using the Bradford Assay.

2.11 Preparation of mouse mammary gland tissue lysate Approximately 50-60 mg of snap frozen mammary gland tissue was cut into small pieces over dry ice with a razor blade and then homogenized in 300 µL RIPA buffer supplemented with 2 mM Na3VO4, 5 mM NaF, 2 mM EDTA, 1 mM PMSF, and 1X protease inhibitor cocktail for approximately 10 s using a Polytron PT 2100 (Kinematica AG) set at 15000 rpm. The tissue lysate was incubated end-over-end at 4°C for 20 min and insoluble debris was pelleted by centrifugation at 10000 rpm for 10 min at 4°C. The supernatant (excluding the top fatty layer) was transferred to a new microfuge tube and centrifuged again. The resulting supernatant was aliquoted and stored at -80°C. Protein concentration was determined using the Bradford Assay.

2.12 Immunodetection of protein by gel electrophoresis/western blot Crude membrane preparations (30 µg), whole cell lysates (10 µg), and mouse mammary gland tissue lysates (10 µg or 20 µg) were resolved in 4-12% Bis-Tris gradient gels using MES running buffer and transferred to Hybond-C nitrocellulose membranes (GE Healthcare) using the Novex NuPAGE SDS-PAGE system (Invitrogen, Life Technologies Inc., Burlington,

61 Canada). Membranes were blocked overnight at 4°C prior to incubation with primary antibody. In general, blots were incubated with primary antibody either at room temperature for 1 h or overnight at 4°C. Blots were then washed three times for 10 min each with PBST (PBS, 0.05% v/v Tween 20) or TBST (TBS, 0.1% v/v Tween 20) prior to incubation with horseradish peroxidase conjugated secondary antibody for 1 h at room temperature. After incubation with secondary antibody, the blots were washed three times for 10 min each with PBST or TBST and bands were visualized by enhanced chemiluminescence (ECL, GE Healthcare Life Sciences) followed by film exposure. Blots were stripped with Restore Western Blot stripping buffer (ThermoScientific) at room temperature for 20 min and re- blocked overnight before incubating with a different primary antibody. For densitometric analysis of band intensity, films were captured using the Fluorchem 8000 Gel Documental System coupled to a Fisher Scientific CCD camera. Captured images were then analyzed using Image J (NIH). Please refer to Table VI (human) and Table VII (mouse) for complete information on antibodies used and specific incubation conditions.

2.13 Immunohistochemistry Mouse mammary glands were fixed overnight in 10% neutral buffered formalin, dehydrated, and paraffin-embedded. Paraffin-embedded tissue sections mounted onto glass slides were rehydrated by incubating in xylene three times for 10 min each and then in 100% ethanol three times for 10 min each, followed sequentially by 5 min incubations in 95% ethanol, 70% ethanol, and 50% ethanol. Tissue sections were then washed twice with ddH2O for 5 min each on a rotating shaker. Heat-induced antigen retrieval was performed by immersing tissue sections into sodium citrate buffer (10 mM sodium citrate, 0.05% Tween 20 pH 6.0) in a pressure cooker and microwaving for 21 min. Slides were then washed in ddH2O twice for 5 min each and then tissues were outlined with a hydrophobic barrier pen. Tissue sections were washed once more in PBS for 5 min and then blocked in 10% goat serum in incubation buffer (1% w/v BSA, 0.3% v/v Triton X-100 and 0.05% v/v Tween 20) for 1 h. Slides were then sequentially incubated with different antibodies under the conditions indicated below. Unless specified, all incubations with blocking buffer and antibodies were performed at room temperature in a humidified chamber. Slides were washed five times for 5 min each with PBST before incubating with the next antibody. Negative slides prepared from tissue sections of lactating mouse mammary glands which were only incubated with secondary antibodies

62 were processed in parallel. To immunodetect ABCG2/Cytokeratin 18 combination, tissues sections were first incubated overnight at 4°C with 1:100 anti-ABCG2 (BXP-53, Cat# Ab24115, Abcam) in 10% goat serum blocking buffer. Slides were then washed with PBST and then incubated with 1:200 Cy3 conjugated anti-rat (Cat# 112-165-003, Jackson ImmunoResearch) for 1 h. Slides were washed with PBST, blocked for 40 min with 10% goat serum blocking buffer supplemented with 20 µg/mL AffiniPure F(ab’)2 goat anti-mouse (Jackson ImmunoResearch) and then incubated overnight with 1:50 anti-cytokeratin 18 (Cat#Ab668, Abcam). Slides were washed with PBST and then incubated with 1:100 Alexa 488 conjugated goat anti-mouse (Cat# A21121, Invitrogen) for 1 h. The nuclei were stained with Draq5 as indicated below. To immunodetect phospho-STAT5/Cytokeratin 18 combinations, after blocking with 10% goat serum in incubation buffer, slides were first incubated with the antibody combination described above for the detection of cytokeratin-18. Tissue sections were then blocked with 10% donkey serum in incubation buffer for 1.5 h followed by an overnight incubation with 1:100 anti-phospho-STAT5 antibody (Cat# 716900, Invitrogen) at 4°C. Slides were washed with PBST and incubated with 1:200 Cy3-conjugated donkey anti-rabbit (Cat#711-165-152, Jackson ImmunoResearch) in 5% donkey serum in incubation buffer for 1 h. After the last antibody incubation, slides were washed with PBST five times for 10 min each and then incubated with 5 µM Draq5 (Cat# 4084S, Cell Signaling) for 2 h. Slides were mounted with Fluorescence Mounting Medium (Dako, Agilent Technologies) and stored at 4°C in the dark. Spinning disk confocal images were acquired with a Zeiss AxioVert 200M microsope and analyzed using Volocity 6.3 (Perkin Elmer).

2.14 Short-interfering RNA T-47D cells were seeded at 5 x105 cells per well on 6 well plates, grown for 24 h, and transfected with siRNA using Dharmafect 1 as per manufacturer protocol (Dharmacon, ThermoScientific, Lafayette, CO). For simultaneous knockdown of both isoforms STAT5A and STAT5B, cells were transfected with 25 nM custom siRNA duplex sense 5’- CUACAGUCCUGGUGUGAGAUU-3’ and antisense 5’- UCUCACACCAGGACUGUAGUU-3’ used previously by others (Gutzman et al., 2007). A non-targeting siRNA was used as a control (siGENOME non-targeting RNA #3, Dharmacon). For knockdown of JAK2, cells were transfected with 20 nM each of two predesigned siRNA (J-003146-12 and J-003146-13) against JAK2, and for control, 20 nM each of non-targeting

63 siRNA 1 and siRNA 2 (ON-TARGETplus siRNA, Dharmacon). Twenty four hours after transfection, cells were incubated with fresh growth medium, grown for an additional 24 h, then serum-starved overnight. Cells were either lysed to obtain whole cell lysate for assessing knockdown of protein expression by western blot, or treated with PBS (0.1% v/v) or PRL for 6 h to assess mRNA expression.

2.15 Plasmid constructs 2.15.1 Plasmids used to investigate the activity of the human ABCG2 gene promoter and GAS element. Please refer to Table VIII for primer and oligonucleotide sequence. The - 1285/+362 ABCG2 construct reported previously (Bailey-Dell et al., 2001), which contains the ABCG2 proximal GAS element, surrounding 5’-flanking region and promoter, inserted into the pGL3-basic construct (Promega) is referred to here as pGL3-ABCG2. Fragments of DNA containing the ABCG2 distal GAS element were prepared by either annealing chemically synthesized oligonucleotides containing MluI sticky ends corresponding to -4476/-4442ABCG2 or by PCR amplification of the -4565/-4414ABCG2 region using human BAC DNA and primers that contain MluI restriction sites. Annealed oligos and MluI digested PCR product was subsequently cloned into pGL3-ABCG2 construct. Single mutations were introduced into the proximal GAS element of the pGL3-ABCG2 construct by site directed mutagenesis using ABCG2/GASmut1 and ABCG2/GASmut2 primer sets with the Pfu Turbo DNA polymerase as per manufacturer instructions (Stratagene). Successful mutagenesis was confirmed by sequencing and the mutated insert was released by digestion with MluI/BglII and re-cloned back into the pGL3-basic vector. The pGL4-CISH construct which contains the -1304 to +1 region of the human CISH gene cloned into the pGL4.10 vector was obtained from Dr. Charles V Clevenger (Department of Pathology, Northwestern University) (Fang et al., 2008). Tandem GAS constructs were cloned using chemically synthesized oligonucleotides that contain variable numbers of copies of the putative proximal GAS element (-448 to -422 region) of the human ABCG2 gene. For GAS1v2.1, annealed oligos were cloned into KpnI/BglII digested pGL4.23[Luc2/minP] construct (Promega). GAS3v2.5 was constructed by inserting HindIII digested annealed oligos into pGL4.23. GAS6v2.4 was constructed by ligating undigested annealed oligos used to construct GAS3v2.5, which contain KpnI/BglII sticky ends, to GAS3v2.5. The same approach was used to construct GAS6v2mut.8 using annealed oligos that contain double mutations to the GAS element. The pCMV-wtStat5a construct contains the

64 rat Stat5a cloned into the pCMV-Tag 3B expression vector and was obtained from Dr. Charles V Clevenger (Department of Pathology, Northwestern University) (Fang et al., 2009). Mouse Stat5a and Stat5b expression plasmids (pcDNA3-Stat5a and pcDNA3-Stat5b) were generous gifts from Dr. Dwayne L. Barber (Ontario Cancer Institute, Princess Margaret Hospital, Toronto, Ontario, Canada). 2.15.2 Plasmids used to investigate the activity of the mouse E1b promoter and GAS elements. Please refer to Table IX for primer and oligonucleotide sequence. Firefly luciferase reporter plasmids driven by different lengths of the E1b promoter and 5’-flanking region were constructed by first amplifying the region of interest from C57BL/6 genomic DNA with various forward primers and a common reverse primer using Pfu Turbo (Stratagene). Forward primers contained a KpnI restriction site and the common reverse primer contained a XhoI restriction site. PCR products were treated with KpnI/XhoI, purified using QIAquick Spin Columns (Qiagen) and cloned into pGL4.10[Luc2] (Promega). Firefly luciferase reporter plasmids driven by a minimum promoter under the control of various STAT5 binding regions in the mouse Abcg2 gene were prepared by amplifying the region of interest from mouse C57BL/6 DNA or BAC (RP363-L24) DNA using sequence-specific primers and cloning the PCR product via KpnI or XhoI restriction sites into pGL4.23[Luc2] (Promega). The mAbcg2- GAS4mut1 and mAbcg2-GASmut2 plasmids, which contain a single mutation to the GAS4 element, were constructed by digesting chemically synthesized double-stranded DNA (gBlocks® Gene Fragments, Integrated DNA Technologies) with XhoI and inserting the purified digested fragment into pGL4.23. E1b promoter-driven lucia luciferase reporter constructs were generated by amplifying the -377/+199 and -71/+199 E1b promoter region with forward primers and reverse primers that contain HindIII and SpeI restriction sites, respectively. The PCR product was treated with HindIII and SpeI and inserted into pCpGfree- basic-Lucia (InvivoGen) which is devoid of CpG sites throughout the entire backbone. All firefly reporter plasmids (pGL3 and pGL4 series) were introduced into chemically competent Subcloning EfficiencyTM DH5α E. coli. Bacteria transformed with pGL3 and pGL4 were grown up in LB agar or broth containing 100 µg/mL ampicillin. The pCpGfree-basic- Lucia plasmids were transformed into and propagated in ChemiComp GT115 E. coli grown in TB broth or LB agar containing 25 µg/mL Zeocin. All plasmids were sequenced at The Centre for Applied Genomics (SickKids, Toronto, Ontario) and prepared using HiSpeed Plasmid Midi and Maxi Kit (QIAGEN Inc.) for transfection.

65 2.16 In vitro methylation of plasmid DNA Plasmid DNA (4 µg) was incubated with 20 units of CpG methyltransferase, M.SssI, and 640 µM S-adenylmethionine (SAM) in a 50 µL reaction containing 1X NEBuffer 2 (50 mM NaCl,

10 mM Tris-HCl, 10 mM MgCl2, 1 mM dithiothreitol). Mock methylation reactions in which CpG methyltransferase was not added were performed in parallel. The reactions were first incubated for 4 h at 37°C followed by heat inactivation of the methyltransferase by incubating for 20 min at 65°C. An additional round of methylation was performed by adding 10 units of methyltransferase and 16 nmol of SAM to the methylation reaction and incubating for 5 h at 37°C. The treated DNA were purified using QIAquick Spin Columns and eluted with 25 µL of

Mili-Q H2O. Methylation status was assessed by agarose gel electrophoresis, testing for the absence of linearized plasmid after incubation with the CpG-methylation sensitive AatII that would otherwise cut a non-methylated CpG site within the insert.

2.17 Transient transfection and luciferase assay For all transfections, plasmids were first diluted in OPTIMEM and then incubated with 3:1 Fugene HD: DNA for 15 min at room temperature. Fugene HD-DNA complexes were then diluted with an appropriate amount of OPTIMEM and added drop-wise to each well containing complete (i.e. growth) medium. 2.17.1 Assessment of human ABCG2 gene promoter and GAS element activity. For experiments using firefly luciferase reporter constructs driven by the ABCG2 and CISH promoter, T-47D cells were seeded at 3 x 105cells/well on 12-well plates and grown for 24 h. Cells were then transfected in growth medium with 500 ng/well pGL3-basic, variations of the pGL3-ABCG2 constructs, or 75 ng/well pGL4.10 or pGL4-CISH. Twenty four hours after transfection, cells were serum-starved overnight and treated with 100-500 ng/mL recombinant human PRL for 24 h. For experiments using the tandem GAS constructs, T-47D cells were transfected as above with slight modifications. T-47D cells were seeded at 1.25 x 105cells per well on 24-well plates and transfected with 300 ng/well pGL4.23 or variations of this constructs containing different copies of the proximal ABCG2 GAS element. To overexpress Stat5, cells were co-transfected with 300 ng/well pCMV-wtSTAT5a, pcDNA3-Stat5a, pcDNA3-Stat5b, or their respective empty vector for control. At the end of treatment, cells were lysed with 100 µL (24-well plates) or 250 µL (12-well plates) phosphate lysis buffer, and 10 µL of lysate was used to measure luciferase activity using the Dual Luciferase Reporter

66 Assay Kit from Promega on a Sirius Luminometer (Berthold Detection System, Pforzheim, Germany) . Firefly luciferase reporter activity was normalized to Renilla luciferase activity from a co-transfected pRL-TK plasmid (5 ng/well and 10 ng/well for 24- and 12-well plates, respectively) and presented as fold change over vehicle and empty vector control. All experiments were conducted in triplicate wells at least three times. 2.17.2 Assessment of mouse Abcg2 gene promoter and GAS element(s) activity. For promoter deletion analysis, HC11 cells were seeded at 9 x 104 cells/well on 24-well plates and grown for 24 h in growth medium prior to transfection. HC11 cells were then transfected with 100 ng/well pGL4.10 and variations of this construct driven by different lengths of the E1b promoter region. Transfection efficiency was monitored by co-transfection with 1 ng/well pRL-TK that constitutively expresses Renilla luciferase. Twenty-four hour after transfection, cells were lysed with 100 µL phosphate lysis buffer (PLB) and 2.5 µL of lysate was used to measure luciferase activity using the Dual Luciferase Reporter Assay Kit (Promega) on a Sirius Luminometer (Berthold Detection Systems; Pforzheim, Germany). For promoter methylation experiments, cells were transfected with 50 ng/well mock or CpG methylated pCpG-free-basic- lucia reporter driven by -71/+199 or -377/+199 of the Abcg2 E1b promoter region. To control for transfection efficiency, cells were co-transfected with 2 ng/well pGL3-control (a gift from Dr. David S Riddick, Department of Pharmacology & Toxicology, University of Toronto) that constitutively express firefly luciferase. After 24 h, medium containing secreted lucia luciferase was collected and cells were lysed with 100 µL of PLB. Lucia luciferase activity was quantitated using 5 µL of medium and 50 µL of QUANTI-Luc (InvivoGen). Firefly luciferase activity was quantitated using 10 µL of lysate and 50 uL of luciferase assay reagent II (LARII; Promega). For the functional assessment of enhancer activity of STAT5 binding regions/GAS elements, T-47D cells were seeded and transfected as per conditions described above for the tandem GAS constructs. Briefly, T-47D cells were seeded at 1.25 x 105cells per well on 24-well plates and transfected with 300 ng/well pGL4.23 or variations of this constructs containing different STAT5 binding regions (or GAS4 mutation) of the mouse Abcg2 gene. Cells were co-transfected with 300 ng/well pcDNA3-Stat5a to overexpress mouse STAT5A. After transfection, cells were serum-starved overnight and treated with 200 ng/mL rhPRL for 24 h. Luciferase activity was measured as described above using the Dual Luciferase Reporter Assay Kit (Promega).

67 2.18 Genomic DNA isolation, bisulfite treatment, and pyrosequencing Genomic DNA (gDNA) was isolated from snap frozen mammary glands or cells using the DNA QiaAMP Mini Kit (Qiagen). Samples were treated with RNase A during the lysis step according to manufacturer instructions. Genomic DNA or in some cases, bisulfite treated DNA, was submitted to the laboratory of Dr. Rosanna Weksberg, Pyrosequencing Service (Program in Genetics and Genome Biology, The Hospital for Sick Children, Toronto, Ontario) for pyrosequencing. The protocol used was as follows. Approximately 1 µg of gDNA was treated with bisulfite using the EpiTect Bisulfite Kit (Qiagen) and eluted from the DNA purification column using 20 µL of EB buffer. Sequence specific primers to the mouse Abcg2 CpG island (assay amplified the -62/+18 region, see Table IX for primer sequences) and M13 biotin-labeled universal primers were used to amplify 1 µL of bisulfite-treated DNA in a 25 µL reaction. The PCR product (20 µL) was subsequently pyrosequenced using PyroMark Q24 (Qiagen).

2.19 Chromatin Immunoprecipitation (ChIP) 2.19.1 ChIP for STAT5 in T-47D cells. T-47D cells were grown to 80-90% confluence in 10 cm-petri dishes, serum-starved in 10 mL of starvation medium for 18-20 h, and then treated with PBS (0.1% v/v), indicated concentrations of PRL, or 200 ng/mL GH for up to 6 h. For study using T-47D cells transfected with siSTAT5A/B or siRNA control, cells were transfected as per protocol for 6-well plates but scaled up to account for the larger surface area. For STAT5 inhibitor study, cells were pre-treated with 200 µM STAT5 inhibitor or vehicle control (0.2% DMSO v/v) for 1 h prior to prolactin treatment. Cells were fixed with 1% formaldehyde by adding 37% formaldehyde dropwise into the medium and incubating at room temperature for 10 min. The cross-linking reaction was quenched with 125 mM glycine and mixed by rotation at room temperature for 5 min. Fixed cells were scraped in PBS containing 1 mM PMSF and 1X protease inhibitor cocktail, and pelleted at 10000 rpm, for 3 min at 4°C. The cell pellet was frozen in an ethanol-dry ice bath and stored at -80°C until use. To immunoprecipitate STAT5, cell pellets were resuspended in 400 µL TSEI (20 mM Tris, 150 mM NaCl, 2 mM EDTA, 1% TritonX-100, 0.1% SDS, pH 8.0) and sonicated at 30% amplitude, 10 pulses (at 0.9 s on and 0.1 s off) for 10 s, twice, and cycled 5 times to a combined sonication of 100 s using a digital Branson 450 sonifier. Insoluble debris was pelleted at 13200 rpm, at 4°C for 10 min, and the supernatant was pre-cleared with 10 µL 50% slurry protein-A Agarose

68 beads (Sigma). A small volume (4.25 µL) was saved to quantify total input. The pre-cleared lysate (85 µL) was then incubated with 2 µg anti-STAT5 antibody (N-20, sc-836x, Santa Cruz) or non-specific rabbit IgG (Sigma) at 4°C overnight with rotation. Immunocomplexes were precipitated with 20 µL pre-blocked 50% Protein-A agarose beads/salmon sperm DNA for 1.5 h at 4°C, and subsequently washed for 5 min three times with TSEI, once each with TSEII (20 mM Tris, 500 mM NaCl, 2 mM EDTA, 1% TritonX-100, 0.1% SDS, pH 8.0) and LiCl buffer (20 mM Tris, 250 mM LiCl, 1 mM EDTA, 1% NP-40, 1% sodium deoxycholate, pH 8.0), and twice with TE buffer as previously described (Ahmed et al., 2009). Immunocomplexes were eluted with 1% SDS in TE buffer at room temperature for 30 min and cross-links were reversed overnight at 65°C. DNA was purified using EZ-10 Spin Column (BioBasic Inc.) and eluted with 50 µL ddH2O. For quantification of STAT5 binding, real-time PCR was performed using 1 µL eluted DNA in a 10 µL reaction containing SsoFAST EvaGreen Supermix (BioRAD) and 100 nM each of forward and reverse primer. Primer sequences are listed in Table X. Recruitment was calculated as % of total input. 2.19.2 ChIP for STAT5 in mouse mammary gland tissue. A modified ChIP protocol from that described above was used for mouse mammary gland tissues. Approximately 70 mg of snapped frozen mouse mammary gland was cut into small pieces (<3 mm2) over dry ice and placed into 2 mL microfuge tubes. To fix the tissue, 600 µL of 1% formaldehyde in PBST (0.1% v/v Tween-20) was added to the tissue and the tissue was homogenized for 30 s using a Polytron 2100 set at 11000 rpm. Immediately following homogenization, an additional 600 µL of 1% formaldehyde was added to the homogenate. Tissues were fixed by incubating end- over-end at room temperature for 10 min. The fixation was quenched by adding 75 µL of 2 M glycine (final 125 mM glycine) to the homogenate and incubating end-over-end at room temperature for 5 min. The fixed tissue homogenate was centrifuged at 1000 g for 5 min at 4°C. The tissue pellet was washed twice with ice-cold PBST containing protease inhibitor cocktail and then resuspended in 1 mL of ice cold cell lysis buffer (5 mM PIPES, 85 mM KCl, 0.5% NP-40, and 1X protease inhibitor cocktail). After 20 min on ice, the nuclei were pelleted at 1000 g for 5 min at 4°C. The resulting pellet was resuspended in 400 µL TSE I supplemented with protease inhibitor cocktail and incubated on ice for at least 20 min before sonication. Samples were sonicated 10 times at 20% amplitude, 0.5 s on, 1.0 s off, for 20 s to a total sonication time of 200 s using a Digital Branson Sonifier 450. After sonication, the samples were centrifuged at 13000 rpm for 10 min at 4°C and the supernatant (soluble

69 chromatin) was pre-cleared with 40 µL 50% protein A agarose beads for 1 h rotating end over end at 4°C. Pre-cleared soluble chromatin (150 µL) was then incubated with 2 µg IgG or Stat5 antibody (N-20x) end-over-end at 4°C overnight. Immunocomplexes were incubated end- over-end with 40 µL of 50% protein A agarose beads for 2 h at 4°C. The beads were pelleted and then washed 3x with TSEI, once with TSEII, once with LiCl buffer, and twice with TE buffer. Immuncomplexes were eluted from the beads by incubating with 120 µL 1%SDS in TE buffer on a rotating wheel at room temperature for 4 h. Eluted complexes (with beads) and input DNA (4%) was incubated at 65°C overnight to reverse cross-links and then treated with 55 µg Proteinase K (ThermoScientific) for 1 h at 52°C. The DNA was purified using EZ-10 spin columns (BioBasic Inc) and eluted twice with 50 µL of EB buffer (10 mM Tris-HCl, pH 8.5, QIAGEN). Recruitment was determined by real-time PCR using 2 µL eluted ChIP DNA or input DNA diluted 1:4 in a 20 µL reaction containing Power SYBR green mix (Applied Biosystems) and 400 nM each of forward and reverse primer. Primer sequences are listed in Table X. Recruitment was calculated as % of total input.

2.20 ChIP-seq data analysis This work was performed in collaboration with Dr. Andrei Turinsky at the Centre for Computational Medicine, The Hospital for Sick Children. Pre-processed ChIP-seq datasets from Rijnkels et al. (2013), already aligned to the reference genome by the authors, were downloaded from the Gene Ontology Omnibus site (http://www.ncbi.nlm.nih.gov/geo; GEO Series GSE25105). Multiple replicates for MEC and Liver samples were concatenated before the processing, with the matching control data scaled up appropriately. Data reads mapped to chromosome 6 were extracted and processed using MACS peak-finding algorithm version 1.4.2 (Zhang et al., 2008), comparing the histone-mark data to the matching background control data. The data for lactating mammary cells were also compared directly to the matching data from virgin cells. The algorithm parameters were adjusted for the tag size 36 bp and for the effective-genome size reduced by the ratio 68.6% as recommended by the MACS usage guidelines for the mm9 genome. Due to the input data being comprised of multiple replicates, all duplicate tags were retained. The paired-peaks model building was switched off, which is appropriate for potentially long histone-mark regions. Other algorithm parameters were left with default values. Raw ChIP-sequencing datasets corresponding to the Yamaji et al. (2013) study (GEO data series GSE40930) were retrieved in Fastq format via the DDBJ resource

70 (http://www.ddbj.nig.ac.jp; 21062823). The data were first aligned to the reference mouse genome mm9 using Bowtie algorithm (Langmead et al., 2009). Sequencing reads that were mapped to chromosome 6 were then processed using MACS, comparing the data to the matching background control sets. The data corresponding to the wild-type mammary cells were also compared to the matching Stat5a knockouts. The tag size and the effective-genome size were set as described above, and the paired-peaks model building was switched off for the H3K4me3 datasets. Other parameters were left with default values. The results generated by the MACS algorithm were visually mapped onto the genomic neighborhood of the Abcg2 gene using the UCSC Genome Browser (Karolchik et al., 2014). The visualization included the locations of statistically significant peaks of DNA occupancy, as well as the WIGGLE data representing the sequencing-fragment pileup in the original ChIP treatment data (calculated by MACS at every 10 bps, with a further smoothing window of 5 pixels applied to the image in the browser). Whenever multiple replicates were present in the input data, the WIGGLE data values were scaled down appropriately. The data analysis used publically available resources Galaxy (Blankenberg et al., 2010) and Cistrome (Liu et al., 2011), as well as custom-made software scripts for data pre- and post-processing. Nucleotide sequences extracted from ChIP- seq datasets were interrogated for STAT5 binding sites (matrix similarity score ≥ 0.8) using MatInspector (www.genomatrix.de).

2.21 Statistical Analysis Data from all animal experiments are presented as mean ± standard deviation (SD) whereas in vitro results are expressed as mean ± standard error of the mean (SEM). Test of significance was performed using the student’s t-test, or analysis of variance (ANOVA) followed by post- hoc pairwise multiple comparisons using the Tukey’s or Dunnett’s test. A p-value <0.05 was considered significant. All statistical analyses were performed using SPSS Statistics versions 19 to 21 (International Business Machines Corp.).

71 Table V. Primer sequences for gene expression analyses by real-time RT-PCR Gene/mRNA isoform Sequence or Accession number Reference or Assay IDa Human ABCG2 NM_001257386.1 and NM_004827.2 Hs00184979_m1 CISH NM_013324.5 and NM_145071.2 Hs00367082_g1 CYP1A1 NM_000499.3 Hs00153120_m1 TIPARP NM_001184717.1, NM_001184718.1, and NM_015508.4 Hs00296054_m1 GAPDH NM_002046.4 Hs99999905_m1 JAK2 NM_004972.3 Hs01078124_m1 STAT5A NM_003152.3 Hs00559643_m1 STAT5B NM_012448.3 Hs00273500_m1 18s RNA X03205.1 Hs99999901_s1

Mouse Abcb1a NM_011076.2 Mm00440761_m1 Abcb1b NM_011075.2 Mm00440736_m1 Abcc1 NM_008576.3 Mm00456156_m1 Abcc2 NM_013806.2 Mm00496899_m1 Abcg2 NM_011920.3 Mm00496364_m1 Slc52a2 NM_029643.3 Mm00518631_g1 Slc52a3 NM_001164819.1 and NM_027172.3 Mm00510189_m1 Cish NM_009895.3 Mm00515488_m1 Csn2 NM_00128602(0, 1, 2, 3, or 4).1 and NM_009972.2 Mm00839664_m1 Wap NM_011709.5 Mm00839913_m1 Krt18 NM_010664.2 Mm01601702_g1 Gapdh NM_008084.2 Mm99999915_g1 Actb (β-actin) NM_007393.3 Mm00607939_s1 E1a forward 5’-CCTTTTGACACCTCATTACACATAGC-3’ Zong et al., 2006 E1b forward 5’-CCAGGGACCGCGAGAAAG-3’ Zong et al., 2006 E1c forward 5’-GGAAATTTTAACTGCACATTGAGAGA-3’ Zong et al., 2006 E2 common reverse 5’-GTCCTAACGGCTCTGGAGTTCA-3’ Zong et al., 2006 a Assay ID is from Applied Biosystems.

72 Table VI. Conditions for western analysis of protein expression in human-derived samples Protein (µg loaded) Primary Antibody Secondary Wash Blocking Buffer Antibodya Buffer (in wash buffer)

ABCG2 1:200 BXP-21 1:1500 HRP-goat PBST 5% skim-milk (30 µg crude membrane) (Millipore), anti-mouse overnight at 4°C β-actin 1:10000 (clone AC-15, 1:2000 HRP-goat PBST 5% skim-milk (condition for crude membrane) Sigma), 1h at RT anti-mouse β-actin 1:20000 (clone AC-15, 1:5000 HRP-goat TBST 5% skim-milk (condition for whole cell lysate) Sigma), 1h at RT anti-mouse PRLR 1:500 clone 1A2B1 1:5000 HRP-goat PBST 5% skim-milk (1-10 µg whole cell lysate) (35-9200; Invitrogen) anti-mouse overnight at 4°C JAK2 1:1000, D2E12, 1:1500 HRP-goat TBST 5% BSA (10 µg whole cell lysate) (Cell Signaling), anti-rabbit overnight at 4°C STAT5A 1:2000 sc-1081 1:3000 HRP-goat TBST 5% skim-milk (10 µg whole cell lysate) (Santa Cruz), overnight anti-rabbit at 4°C STAT5B 1:1000 sc-1656 1:2000 HRP-goat TBST 5% skim-milk (10 µg whole cell lysate) (Santa Cruz) overnight anti-mouse at 4°C Phospho-STAT5 1:4000 71-6900 1:5000 HRP-goat TBST 5% BSA (10 µg whole cell lysate) (Invitrogen), overnight anti-rabbit at 4°C ERK1/2 1:2000 L34F12 (Cell 1:3000 HRP-goat TBST 5% skim-milk (10 µg whole cell lysate) Signaling), anti-mouse overnight at 4°C Phospho-ERK1/2 1:1000 197G2 1:5000 HRP-goat TBST 5% BSA (10 µg whole cell lysate) (Cell Signaling), anti-rabbit overnight at 4°C AKT 1:10000 C67E7 (Cell 1:5000 HRP-goat TBST 5% BSA (10 µg whole cell lysate) Signaling), overnight at anti-rabbit 4°C Phospho-AKT 1:4000 C31E5E (Cell 1:5000 HRP-goat TBST 5% BSA (10 µg whole cell lysate) Signaling), overnight at anti-rabbit 4°C a Secondary antibody incubations were all performed at room temperature for 1 h

Table VII. Conditions for western analysis of protein expression in mouse-derived samples Protein (µg loaded) Primary Antibody Secondary Wash Blocking Buffer Antibodya Buffer (in wash buffer)

Abcg2 1:5000 BXP-53 1:10000 HRP-goat TBST 5% skim-milk (10 µg tissue homogenate) (Abcam), anti-rat overnight at 4°C β-actin 1:10000 anti- β-actin 1:5000 HRP-goat TBST 5% skim-milk (10 or 20 µg tissue homogenate) (Sigma), 1h at RT anti-mouse Stat5a 1:10000 sc-1081 1:20000 HRP-goat TBST 5% skim-milk (20 µg tissue homogenate) (Santa Cruz), 1h at RT anti-rabbit STAT5B 1:4000 sc-1656 1:4000 HRP-goat TBST 5% skim-milk (20 µg tissue homogenate) (Santa Cruz), 1h at RT anti-mouse Phospho-STAT5 1:10000 71-6900 1:20000 HRP-goat TBST 5% BSA (20 µg tissue homogenate) (Invitrogen), 1h at RT anti-rabbit a Secondary antibody incubations were all performed at room temperature for 1 h

73 Table VIII. Sequence for primers and oligonucleotides used to construct reporter plasmids containing different regions of the human ABCG2 gene. Synthesized oligonucleotides annealed for cloning Sequence (5’ to 3’, GAS sequence underlined) -4476/-4442ABCG2 forward CGCGTCCAAATGAATATTTTCTCTGAACCCTAAGATAGCCA reverse CGCGTGGCTATCTTAGGGTTCAGAGAAAATATTCATTTGGA GAS1v2 forward CAGCATCCACTTTCTCAGAATCCCATTCA reverse GATCTGAATGGGATTCTGAGAAAGTGGATGCTGGTAC GAS3v2 forward CAGACAAGCTTAGCATCCACTTTCTCAGAATCCCATTCAGACAGCATCCAC TTTCTCAGAATCCCATTCAGACAGCATCCACTTTCTCAGAATCCCATTCAA GCTTAGACA reverse GATCTGTCTAAGCTTGAATGGGATTCTGAGAAAGTGGATGCTGTCTGAATG GGATTCTGAGAAAGTGGATGCTGTCTGAATGGGATTCTGAGAAAGTGGAT GCTAAGCTTGTCTGGTAC GAS3v2mut (to make GAS6v2.8) forward CAGACAAGCTTAGCATCCACTGGCTCAGAATCCCATTCAGACAGCATCCAC TGGCTCAGAATCCCATTCAGACAGCATCCACTGGCTCAGAATCCCATTCAA GCTTAGACA reverse GATCTGTCTAAGCTTGAATGGGATTCTGAGCCAGTGGATGCTGTCTGAATG GGATTCTGAGCCAGTGGATGCTGTCTGAATGGGATTCTGAGCCAGTGGATG CTAAGCTTGTCTGGTAC

Primers Sequence (5’ to 3’) -4565/-4414ABCG2 forward TGATACTAACGCGTTAAGGGACCTGACACTACCAATAAC reverse TGATACTAACGCGTGGAAGGTGAGAGAAAGGAATATAGC ABCG2/GASmut1 forward CATCCACTTTCTCAGCATCCCATTCACCAG (site of mutation underlined) reverse CTGGTGAATGGGATGCTGAGAAAGTGGATG (site of mutation underlined) ABCG2/GASmut2 forward GCAAGCATCCACTTACTCAGAATCCCATTC (site of mutation underlined) reverse GAATGGGATTCTGAGTAAGTGGATGCTTGC (site of mutation underlined)

74 Table IX. Sequence for primers and double stranded gene fragments used to construct reporter plasmids containing different regions of the mouse Abcg2 gene. Primer/gene fragment Sequence (5’ to 3’)a Firefly luciferase reporter Promoter Analysis -1651E1b forward TTATTTGGTACCCCCTTTGGACAAGCACAGAG -1277E1b forward TTATTTGGTACCGGCTCACAAAGGGTCAAGTG -978E1b forward TTATTTGGTACCGATTCTTGCCCTGTTTGAGC -377E1b forward TTATTTGGTACCCTCTCATCAAAAGGCCATCC -164E1b forward TTATTTGGTACCCTCTCTGCTCTCCGCTCTCC -100E1b forward TTATTTGGTACCACATGTGTCCAGCTGCTCCT -71E1b forward TTATTTGGTACCTTCTCCTGCCGGCGAGCTG -40E1b forward TTATTTGGTACCAGCTGACGTCACGGCGGTC +23E1b forward TTATTTGGTACCGTGCAGGTCTGAGTGTGTGC +199E1b reverse TTATTTCTCGAGTCACTTTCACAACCCACACC

Enhancer (GAS) analysis mAbcg2-GAS1 forward TTATTTGGTACCGGAAATAAGTTTGATAGCAAGTCCCAAG reverse TTATTTGGTACCGTACCTTCAGTGACCCACCAGCATTG mAbcg2-GAS2 forward TTATTTGGTACCCAGGAAACCCAAAAACCTTCTCAC reverse TTATTTGGTACCGTTGATGAAGTCAAACATTGTACCCTG mAbcg2-Peak2 forward TTATTTGGTACCGAATGCATGGCTGATATGTCCAG reverse TTATTTGGTACCCCCTAAGCAATGAACAGACACC mAbcg2-GAS3 forward TTATTTCTCGAGCTTCCCTCTGATGGCATTCC reverse TTATTTCTCGAGCTATGAGCCAGGCCAGTTTC mAbcg2-GAS4 forward TTATTTCTCGAGGCCACCTTAGGGAGTGGAT reverse TTATTTCTCGAGGAGCACATTTTCCCGGAGGA mAbcg2-GAS5 forward TTATTTCTCGAGTTCTTGGGATGTTCTCACAAG reverse TTATTTCTCGAGTATAACCTCTTAATTCTACGTGT mAbcg2-Peak2 forward TTATTTGGTACCGAATGCATGGCTGATATGTCCAG reverse TTATTTGGTACCCCCTAAGCAATGAACAGACACC gBlocks® Gene Fragments mAbcg2-GAS4mut1 TCTGTGCTCGAGGCCACCTTAGGGAGTGGATCCATTCTCCTACAGA (mutation underlined) TCTTTTTTTTTTCTTGCTGCTTCATTTCTATAAATATACATTACATCT TGAAGTCCAGGAATAGATACAGATCTCCCTTCCTCCGGGAAAATGT GCTCCTCGAGTTATTT mAbcg2-GAS4mut2 TCTGTGCTCGAGGCCACCTTAGGGAGTGGATCCATTCTCCTACAGA (mutation underlined) TCTTTTTTTTTTCTTGCTGCTTCATTTCTATAAATATACATTACATCT TGAATTCCAGGGATAGATACAGATCTCCCTTCCTCCGGGAAAATGT GCTCCTCGAGTTATTT

Lucia luciferase reporter CpG-377E1b forward TTATTTACTAGTCTCTCATCAAAAGGCCATCC CpG-71E1b forward TTATTTACTAGTTTCTCCTGCCGGCGAGCTG CpG+199E1b reverse TTATTTAAGCTTTCACTTTCACAACCCACACC aKpnI (GGTACC), XhoI (CTCGAG), HindIII (ACTAGT) and SpeI (AAGCTT) restriction sites are indicated in bold.

75 Table X. Primers for ChIP analysis and Pyrosequencing Primer Sequence (5’ to 3’) STAT5 ChIP – T-47D CISH GAS elements forward CCCCTCTGGGTAGCTTCAG reverse CCCTGAGCAGTGAAAGGAAA ABCG2 proximal GAS forward AAGTTTCTCCCCTTTCCTTCC reverse ACAGGTTGCCCAGTCACAAG ABCG2 distal GAS forward GCTTCCTAAGGGACCTGACAC reverse GGAAGGTGAGAGAAAGGAATATAGC

STAT5 ChIP – mouse mammary gland tissue Peak1 (GAS1&2) forward CTGCCAGTTCTTAGTGTGGG reverse GTGCGAGCTGAAGAAAAGGG GAS_AB forward ACATCAAAGCTTCTGGGTGGT reverse TGCCTAAGACCCTGAGAGGA Peak2 forward TTGGCCAAAGAGCAGACACT reverse GTGCAATACGATCACCTGGC GAS_C forward TTTATGGGCCAGCGATTAAC reverse CACTGAATTCTGAGGGAAAAAC GASneg forward CAGTTCAGTCCTTGGGGTCC reverse ATGGGCCACTGATTGACAGG Peak3 (GAS3) forward CTTCCCTCTGATGGCATTCC reverse CTATGAGCCAGGCCAGTTTC Peak4 (GAS4) forward CTTCCTCCGGGAAAATGTG reverse TTGAAAACTCCAGCCACTTG Peak5 (GAS5) forward AGCTAACCCAAGGGAGAAGC Reverse ATACCAGCTGGGCATGTGTT Csn2 GAS promoter forward TGGGCAAGTTCCTTAACCAG reverse TCCCTTCAATTCCAAGAAGTC Wap GAS2 forward CATCTCTTCCTGCCCATGAC reverse TCGGGCATACATTGAAAAGG

Pyrosequencing CGCCAGGGTTTTCCCAGTCACGACGGAAGTTTTTAGTTTATATGTGTT Abcg2-mouse-F1 TAGTTG Abcg2-mouse-R1 AAAACTCCCACACACTCAAACCTA Abcg2-mouse-S1 (sequencing) ACACTCAAACCTACAC

76 3. RESULTS

3.1 AIM1: to determine whether prolactin induces ABCG2 in normal and cancer cells derived from the human breast and if so, what signalling mechanisms are involved – Please note that certain passages and figures within this section 3.1 contain materials reprinted from Wu et al. (2013) with permission from the American Society for Pharmacology and Experimental Therapeutics.

3.1.1 Prolactin induces ABCG2 in T-47D cells To examine if prolactin regulates ABCG2 expression, I first treated serum-starved PRLR-expressing T-47D human breast cancer epithelial cells with varying concentrations of prolactin for 24 h. Prolactin dose-dependently increased ABCG2 mRNA and the mRNA of cytokine-inducible SH2 containing protein (CISH) (Fig. 5A), a known prolactin-responsive gene. ABCG2 mRNA was maximally induced at 6 h whereas CISH mRNA was induced as early as 2 h after prolactin treatment (Fig. 5B). In addition, immunoblotting for ABCG2 protein in crude membrane preparations from T-47D cells treated with prolactin for 24 h showed that ABCG2 protein is also induced in a dose-dependent manner (Fig. 5C). The induction of ABCG2 mRNA by prolactin was blocked by co-treatment with the transcription inhibitor actinomycin D (Fig. 6). In contrast to the response observed in T-47D cells, prolactin did not induce the expression of ABCG2 and CISH mRNA in other human breast epithelial cell lines (MCF-7 and MCF-10A, Fig. 7) and primary human mammary epithelial cells (Fig. 8) despite that under the current treatment conditions (i.e. serum starvation), these cells responded to the aryl hydrocarbon receptor agonist TCDD, a known inducer of ABCG2, CYP1A1, and TIPARP (Kress and Greenlee, 1997; Ma et al., 2001; Tan et al., 2010). Upon further examination using western analysis, it was revealed that these cells expressed significantly less PRLR (long form) compared to T-47D cells (Fig. 9). This may explain the lack of responsiveness to prolactin treatment, which I define here by the induction of CISH mRNA. Based on these results, since T-47D cells responded to prolactin treatment, I used this cell line to interrogate the signalling cascade(s) responsible for the induction of ABCG2.

77 3.1.2 JAK2- and STAT5-dependency in the induction of ABCG2 by prolactin As one of the most proximal proteins and the major kinase associated with the prolactin signalling pathway, JAK2 plays an important role in phosphorylating and hence modulating downstream signalling cascades. However, as was described briefly in the introduction, not all pathways activated by the prolactin receptor are dependent on JAK2. Therefore, I first used AG490, a well-characterized small molecule inhibitor of JAK2 (Meydan et al., 1996) to examine the role of JAK2 in the induction of ABCG2 by prolactin. Unexpectedly, at lower concentrations (25 µM and 50 µM), AG490 alone induced ABCG2 expression but had no apparent effect on prolactin-induced expression of ABCG2 or CISH mRNA (Fig. A1). At a higher concentration (100 µM), AG490 nearly completely abolished the prolactin response. The lack of effect on the prolactin response observed with the lower concentrations of AG490 was quite surprising given that Nielson et al. (2007) showed that 50 µM AG490 effectively inhibited JAK2 activation in T-47D cells treated with 20 nM PRL [equivalent to ~ 460 ng/mL] (Neilson et al., 2007). However, Nielson et al. did not investigate the functional consequence of this inhibition. Therefore it is possible that the inhibition was not complete and that residual JAK2 activity could sufficiently activate downstream signalling proteins like STAT5 to modulate gene expression. Most interestingly, AG490 also induced the aryl hydrocarbon receptor target gene CYP1A1 at the concentrations used (Fig. A2). This suggests that the induction of ABCG2 by AG490 may be due to activation of the AHR. Given this unreported effect of AG490, which confounded my results, I used a combination of two siRNAs that target JAK2 to knockdown JAK2 expression in T-47D cells. These JAK2-targeting siRNAs reduced JAK2 mRNA expression by approximately 80% compared to T-47D cells transfected with a combination of two non-targeting siRNAs (data not shown). This reduction in mRNA was reflected at the protein level (Fig. 10A). The knockdown of JAK2 expression significantly attenuated the induction of ABCG2 and CISH mRNA by prolactin treatment (Fig. 10B). These results suggest that the induction of ABCG2 by prolactin is dependent on JAK2 expression. Noted above, STAT5 is one of the major effector(s) of activated JAK2. Therefore, to determine if STAT5 mediates prolactin-induced expression of ABCG2, T-47D cells were transfected with a siRNA (siSTAT5A/B) that targets both isoform STAT5A and STAT5B. This siRNA reduced STAT5A and STAT5B mRNA expression by approximately 75% and 85%, respectively (data not shown), but more importantly it effectively knocked down STAT5A/B at the protein level (Fig. 11A). The induction of ABCG2 and CISH mRNA by PRL was

78 significantly reduced in siSTAT5A/B-transfected T-47D (Fig. 11B). Consistent with these results, co-treatment with a novel STAT5 inhibitor (Müller et al., 2008) that partially reduced prolactin-stimulated STAT5 phosphorylation/activation (Fig. A3) significantly blunted the induction of ABCG2 and CISH mRNA by prolactin (Fig. 11C). These experiments demonstrated that the JAK2/STAT5 pathway was important in the induction of ABCG2 by prolactin.

3.1.3 STAT5 recruitment to a putative proximal GAS element in the human ABCG2 gene As a transcription factor, STAT5 modulates gene expression by binding DNA motifs called GAS elements with the consensus sequence 5’-TTCNNNGAA-3’. The in silico examination of an arbitrarily defined 10-kb 5’-flanking region of the human ABCG2 gene revealed a putative distal GAS element at -4459 and a proximal GAS element at -434 (Fig. 12A). Chromatin immunoprecipitation (ChIP) for STAT5 was performed to explore whether these regions were bound by STAT5 after prolactin stimulation. ChIP analyses showed that STAT5 was recruited to the proximal GAS element in a time-dependent manner with peak recruitment achieved at 1 h after PRL treatment, but the distal GAS motif did not show significant STAT5 binding (Fig. 12B). This time-course of STAT5 recruitment was consistent with that observed for the CISH gene suggesting a similar mechanism of regulation. Further, STAT5 recruitment to the proximal ABCG2 GAS element and CISH GAS elements was attenuated by siRNA mediated knockdown of STAT5 expression (Fig. 13A) or by pharmacological inhibition of STAT5 activation (Fig. 13B).

3.1.4 Evidence for a functional proximal GAS element in the ABCG2 gene To test if the putative GAS elements were functional, I transfected T-47D with the luciferase reporter construct pGL3-ABCG2 (Bailey-Dell et al., 2001). This construct, which contained the proximal GAS element and surrounding region, was induced approximately 2- fold by prolactin (Fig. 14A). This was quite modest compared to the 14- to 16-fold induction observed using the pGL4-CISH reporter construct, which I used to confirm that the T-47D cells were responsive to prolactin treatment under my transfection protocol. However, it should be noted that pGL4-CISH is driven by the human CISH promoter, which contains four GAS elements, and therefore is known to be exceptionally sensitive to prolactin treatment (Fang et al., 2008). The addition of the distal GAS element and the neighbouring region to

79 pGL3-ABCG2 did not further enhance reporter activity (Fig. 14B), which was consistent with the ChIP data showing that the distal GAS site does not bind STAT5 (Fig. 12B). In contrast, by introducing a single mutation into the proximal GAS element, reporter activity was significantly reduced (Fig. 14C). Similar results were observed when a different single mutation was introduced into the proximal GAS element (Fig. A4). To investigate whether this proximal GAS element was functional in isolation, various numbers of copies of the GAS element were cloned into the minimum promoter-driven pGL4.23 luciferase reporter construct (Fig. 15A). The reporter activity of these constructs was induced by prolactin in a dose- and copy number-dependent manner (Fig. 15B). In particular, GAS6v2.4 containing six tandem repeats of the proximal GAS element was very sensitive to prolactin treatment. Prolactin- induced reporter activity was synergistically increased by co-transfection with rat STAT5A (Fig. 15C) or mouse STAT5A and STAT5B expression vector (Fig. 15D). As evidence that this reporter activity was dependent on an intact GAS element, double mutations to each of the GAS elements, as shown using the GAS6v2mut.8 construct, completely abolished reporter activity (Fig. 15C).

3.1.5 Prolactin-induced ABCG2 expression is attenuated by MAPK and PI3K pathway inhibitors. In addition to JAK2/STAT5, the MAPK and PI3K pathways are also activated by prolactin signalling. The effect of these pathways on the induction of ABCG2 by prolactin was examined using MAPK pathway inhibitors U0126 and PD98059, and PI3K pathway inhibitors LY294002 and wortmannin. These inhibitors were used at concentrations that consistently blocked activation of the MAPK or PI3K pathway as assessed by reduction of prolactin stimulated phosphorylation of ERK1/2 (Fig. 16A and 16B) or AKT (Fig. 16C and 16D) as shown by western blot. Pharmacological inhibition of MAPK and PI3K signalling attenuated the inductive effect of prolactin on ABCG2 but not CISH mRNA after 6 h of prolactin treatment (Fig. 16E-H), demonstrating a role for these signalling pathways in the prolactin- ABCG2 response.

80 3.1.6 STAT5 recruitment to the proximal GAS element is not affected by MAPK and PI3K inhibitors One potential mechanism by which small molecule-mediated inhibition of MAPK and PI3K pathways can attenuate the induction of ABCG2 by prolactin is by repressing STAT5 recruitment. To test this, I used ChIP to assess the effect of U0126, PD98059, LY294002 and wortmannin on prolactin stimulated STAT5 recruitment. Co-treatment with these inhibitors did not affect STAT5 recruitment to the ABCG2 proximal GAS element and the CISH GAS elements (Fig. 17A and 17B). This suggests that MAPK and PI3K pathways modulate prolactin-induced ABCG2 expression by a mechanism distinct from STAT5 recruitment.

3.1.7 Growth hormone induces ABCG2 mRNA and STAT5 recruitment to the ABCG2 proximal GAS element. Growth hormone (GH) is a member of the peptide hormone family to which prolactin belongs and can activate similar signalling networks including JAK2/STAT5. For this reason, to examine the generalization that JAK2/STAT5 activators are potential inducers of ABCG2, T-47D cells were treated with different concentrations of recombinant human GH for 6 h. The mRNA expression of ABCG2 and CISH was upregulated by GH in a dose-dependent manner (Fig. 18A). To determine if GH, like prolactin, stimulates STAT5 recruitment to the ABCG2 proximal GAS element, a ChIP for STAT5 was performed at 1 h after GH treatment (200 ng/mL). GH treatment significantly increased STAT5 occupancy at the ABCG2 proximal GAS element (Fig. 18B), demonstrating a potential role for STAT5 in GH-induced ABCG2 expression.

81

Figure 5. Prolactin induces ABCG2 mRNA and protein expression in T-47D breast cancer cells. (A) T-47D cells were treated with indicated concentrations of recombinant human prolactin (PRL) for 24 h. Relative mRNA expression compared to vehicle control (PBS 0.1% v/v) was quantified by real-time RT-PCR. (B) T-47D cells were treated with 500 ng/mL PRL for the times indicated. Relative mRNA expression compared to time-matched vehicle control (0.1% v/v PBS) was quantified by real-time RT-PCR. (C) T-47D cells were treated with indicated concentrations of PRL for 24 h. Crude membrane preparations were resolved by SDS-PAGE and immunoblotted for ABCG2 and β-actin. Human placental tissue lysate ran in parallel was used to confirm the specificity of the ABCG2 antibody (data not shown). Results shown are means ± SEM (n = 3).

82

Figure 6. Actinomycin D abolished prolactin-induced ABCG2 and CISH mRNA expression in T-47D cells. T-47D cells were pre-treated with the transcription inhibitor Actinomyin D (ActD, 5 μg/mL) or DMSO (0.1%v/v) for 30 min and subsequently treated with PBS (0.1%v/v) or 500 ng/mL recombinant human prolactin (PRL) for 6 h. Relative mRNA expression was quantified by real-time RT-PCR, normalized to 18s ribosomal RNA, and expressed as % of untreated cells (PBS-DMSO). Data shown are mean ± SEM, n =3. *p<0.05, ***p < 0.001, n.s. not significant, Student’s t-test

83

Figure 7. Effect of prolactin on ABCG2 expression in human breast cancer cell lines. T-47D, MCF-7, and MCF-10A cells were treated with 100 ng/mL or 1000 ng/mL recombinant human prolactin (100PRL and 1000PRL, respectively), or 10 nM TCDD for 6 h. Transcript expression was quantified by real-time RT-PCR. Results shown are fold change over control (PBS+DMSO, indicated by a dashed line), mean ± SEM, n = 3-4. Significance was tested by ANOVA followed by Dunnett’s post-hoc pairwise comparison. *p<0.05, **p<0.01, compared to control.

84

Figure 8. Effect of prolactin on ABCG2 expression in primary human mammary epithelial cells (HMEC). HMEC were grown to 70% confluence and subsequently incubated overnight in (A) growth medium, (B) phenol red- free basal medium, or (C) phenol red-free basal medium supplemented with 0.5 μg/mL hydrocortisone and 5 μg/mL insulin. After overnight incubation, HMEC were treated with recombinant human prolactin (PRL: 100 ng/mL and 1000 ng/mL), 10 nM TCDD, or with vehicle control (0.1% DMSO, 0.1% PBS) for 6 h and 24 h. Relative mRNA expression was quantified by real-time RT-PCR and presented as fold change relative to time-matched vehicle control, n =1.

85

Figure 9. Prolactin receptor expression in human breast cancer cell lines and primary mammary epithelial cells. Indicated amount of whole cell lysate prepared from (A) serum- starved T-47D, MCF-7, and MCF-10A, and (B) human mammary epithelial cells (HMEC) cultured in growth medium (growth), phenol red-free basal medium (basal) or phenol red-free basal medium supplemented with hydrocortisone and insulin (Basal + HI) overnight, were resolved by SDS-PAGE and immunoblotted for prolactin receptor (PRLR) and β-actin. L, long isoform of the PRLR. ∆S1, delta S1 isoform of the PRLR, which lacks an extracellular domain found in the long form of the PRLR and has reduced affinity for prolactin.

86

Figure 10. Knockdown of JAK2 attenuates prolactin-induced ABCG2 mRNA expression. T47D cells transfected with non-targeting siRNA (siCtrl) or siRNA targeting JAK2 (siJAK2) were either (A) lysed to assess JAK2 protein expression or (B) treated with PBS (0.1%v/v, vehicle) or PRL (recombinant human prolactin, ng/mL) for 6 h to assess mRNA expression by real-time RT-PCR. Results are presented as fold change from siCtrl+PBS, mean ± SEM (n = 3). A mean fold change of ABCG2 or CISH under each treatment was compared between siCtrl and siJAK2 using Student’s t-test (two-tailed). * p < 0.05, ** p < 0.01, *** p < 0.001, as compared to corresponding siCtrl treatment group.

87

Figure 11. Knockdown or pharmacological inhibition of STAT5A/B reduces the effect of prolactin on ABCG2 expression. T-47D cells transfected with non-targeting siRNA (siCtrl) or siRNA targeting STAT5A and STAT5B (siSTAT5A/B) were either (A) lysed to assess STAT5 protein expression by SDS-PAGE/western blot or (B) treated with PBS (0.1%v/v) or 100 ng/mL recombinant human prolactin (PRL) for 6 h to assess transcript expression using real-time RT-PCR. B. Transcript expression is presented as fold change to siCtrl+PBS group, mean ± SEM (n = 3). A mean fold change of ABCG2 or CISH mRNA levels after PRL treatment was compared between siCtrl and siSTAT5A/B using Student’s t-test (two-tailed). *** p < 0.001. (C) T-47D cells were pretreated with DMSO (0.2% v/v) or 200 µM STAT5 inhibitor for 1 h, and then treated with PBS (0.1% v/v) or 100 ng/mL PRL for 1 h or 6 h to assess transcript expression. Relative mRNA expression was quantified by real-time RT-PCR, and results shown are fold change relative to the DMSO+PBS group at each time point, mean ± SEM (n = 3). A mean fold change of ABCG2 or CISH expression after PRL treatment was compared to that in the absence of the inhibitor at each time point by Student’s t-test (two- tailed). * p < 0.05, **p<0.01, Student’s t-test.

88

Figure 12. Time-dependent recruitment of STAT5 to the 5’flanking region of the human ABCG2 gene after prolactin treatment. (A) Schematic representation of the 5’flanking region of the human ABCG2 and CISH gene and regions amplified (flanked by arrows) using real- time PCR after chromatin immunoprecipitation (ChIP). (B) T-47D cells were starved overnight and treated with 100 ng/mL recombinant human prolactin (PRL) for the times indicated. STAT5 recruitment to the regions of interest was analyzed by ChIP. Results shown represent mean ± SEM of three independent experiments.

89

Figure 13. Knockdown or pharmacological inhibition of STAT5 attenuates prolactin-induced recruitment of STAT5 to the proximal GAS element of the ABCG2 gene. (A) T-47D cells transfected with 25 nM non-targeting siRNA (siCtrl) or siRNA targeting STAT5A and STAT5B (siSTAT5A/B) were treated with PBS (0.1% v/v) or 100 ng/mL recombinant human prolactin (PRL) for 1 h. STAT5 recruitment was assessed by ChIP, and expressed as % input, mean ± SEM (n=3). Means of STAT5 recruitment values were compared between PRL- treated siCtrl and siSTAT5A/B by Student’s t-test (two-tailed). ABCG2: p = 0.003. CISH: p = 0.001. No significant difference was observed with IgG. ** p < 0.01, Student’s t-test. (B) T- 47D cells were pre-treated with DMSO (0.2% v/v) or 200 µM STAT5 inhibitor for 1 h prior to treatment with PBS (0.1% v/v) or 100 ng/mL recombinant human prolactin (PRL) for 1 h. STAT5 recruitment was assessed by ChIP and presented as % input, mean ± SEM (n=3). Means of STAT5 recruitment values were compared between PRL-treated DMSO and STAT5 inhibitor by Student’s t-test (two-tailed). ** p < 0.01; n.s., not significant (CISH: p = 0.055). No significant difference was observed with IgG.

90

Figure 14. Prolactin induces ABCG2 promoter-driven luciferase reporter activity in T-47D cells. T-47D cells transfected with firefly luciferase reporter constructs and a Renilla luciferase control construct were treated with PBS (0.1% v/v) or indicated concentration of recombinant human prolactin (PRL). Data presented are fold change over PBS and empty vector control (pGL3-basic for ABCG2 constructs, or pGL4 for pGL4-CISH), normalized to Renilla luciferase activity, and expressed as means ± SEM (n =3). (A) T-47D cells were transfected with luciferase reporter constructs pGL3-ABCG2 or pGL4-CISH. ** p < 0.01, *** p < 0.001, One-way ANOVA followed by Dunnett’s post-hoc pairwise comparison to vehicle control. (B) T-47D cells were transfected with pGL3-ABCG2, or constructs containing the distal GAS element in a fragment (-4565/-4414 or -4476/-4442) of the human ABCG2 gene ligated to pGL3-ABCG2. Cells were treated with PBS (0.1% v/v) or 100 ng/mL PRL. One- way ANOVA and Dunnett’s post-hoc pairwise comparison to pGL3-ABCG2 was not significant. n.s. not significant. (C) T-47D cells were transfected with pGL3-ABCG2 or ABCG2/GASmut1 that contains a single mutation (bold italics) to the conserved/consensus putative proximal GAS element and treated with PBS (0.1%v/v) or 100 ng/mL PRL. * p < 0.05, Student’s t-test.

91

Figure 15. Prolactin induces reporter activity in T-47D cells transfected with luciferase constructs containing tandem repeats of the proximal ABCG2 GAS element. (A) Schematic representation of the luciferase reporter constructs that contain variable numbers of copies of ABCG2 proximal GAS element (-448/-422 fragment) inserted into the pGL4.23 construct (GAS1v2.1, GAS3v2.5, GAS6v2.4). GAS6v2mut.8 contains six tandem repeats of the mutated ABCG2 proximal GAS element -448/-422 fragment (represented with an X in the schematic). Numbers denote position relative to nucleotide 1 in the backbone pGL4.23 construct. (B) T-47D cells transiently transfected with the indicated reporter constructs were treated with the indicated concentration of recombinant human prolactin (PRL, ng/mL) for 24 h. Renilla-normalized luciferase activity was expressed as fold-change over PBS treatment. Significance was tested by One-way ANOVA followed by Dunnet’s post-hoc pairwise comparison to PBS treatment for each reporter. *p<0.05, **p<0.01. (C) T-47D cells were transfected with GAS6v2.4 or GAS6v2mut.8 along with pCMV-Tag3B (empty vector) or a plasmid constitutively expressing wild-type rat STAT5A (pCMV-wtStat5a) and treated with PBS (0.1% v/v) or 100 ng/mL PRL for 24 h. Firefly luciferase reporter activity was normalized to Renilla luciferase activity from co-transfected pRL-TK plasmid, and presented as fold change over vehicle and empty expression vector control (pCMV-Tag3B+PBS). Mean fold changes in reporter activity after PRL treatments were compared between pCMV-Tag3B and pCMV-wtStat5a using the Student’s t-test (two-tailed). ** p < 0.01. (D) T-47D cells were transfected with GAS6v2.4 along with pcDNA3 (empty vector) or a plasmid constitutively expressing wild-type mouse Stat5a or Stat5b and treated with PBS (0.1% v/v) or 100 ng/mL PRL for 24 h. Reporter activity is presented as fold change over vehicle and empty expression vector control (pcDNA3+PBS). ***p<0.001, One-way ANOVA, Tukey’s post-hoc pairwise comparison. B to D. Results shown are mean ± SEM (n = 3).

92

Figure 16. Pharmacological inhibition of MAPK and PI3K signalling attenuates the induction of ABCG2 by prolactin. T-47D cells were pre-treated with pathway specific inhibitors U0126 (10 µM, MEK1/2 inhibitor, panel A and E), PD98059 (20 µM, MEK1 inhibitor, panel B and F), LY294002 (10 µM, PI3K inhibitor, panel C and G), wortmannin (WT, 25 nM, PI3K inhibitor, panel D and H) or DMSO (0.1% v/v) for 1h, and subsequently treated with PBS (0.1% v/v) or 100 ng/mL recombinant human prolactin (PRL) for 15 min to assess ERK1/2 phosphorylation, 30 min to assess AKT phosphorylation, or 6 h to assess mRNA expression. A to D. Whole cell lysates (10 µg) were resolved by SDS-PAGE and immunoblotted for phosphorylated ERK1/2 (p-ERK1/2) and total ERK1/2 or phosphorylated AKT (p-AKT) and total AKT. E to G. ABCG2 or CISH transcript expression was quantified by real-time RT- PCR, and GAPDH-normalized values were shown as fold change relative to DMSO+PBS group, mean ± SEM (n = 3-5). PRL-induced fold change was compared between DMSO and an inhibitor of the signalling pathway using Student’s t-test (two tailed). * p<0.05, **p<0.01, n.s., not significant.

93

Figure 17. Prolactin-induced STAT5 recruitment to the ABCG2 proximal GAS element is not attenuated by inhibitors of MAPK and PI3K signalling. T-47D cells were pre-treated with MAPK inhibitors (panel A: 20 µM PD98059 or 10 µM U0126), PI3K inhibitors (panel B: 10 µM LY294002 or 25 nM Wortmannin, WT) or DMSO (0.1% v/v) for 1 h and subsequently treated with PBS (0.1% v/v) or 100 ng/mL recombinant human prolactin (PRL) for an additional 1 h. STAT5 recruitment to the regions of interest was analyzed by ChIP. Data presented are recruitment as mean % input ± SEM (n = 3). Two-way ANOVA for ABCG2 or CISH showed no statistically significant difference in STAT5 recruitment between DMSO and each signalling inhibitor. No significant difference was observed for IgG. n.s. not significant compared to prolactin in DMSO (i.e., in the absence of inhibitor).

94

Figure 18. Growth hormone induces ABCG2 expression and STAT5 recruitment to the proximal GAS element of the ABCG2 gene in T-47D cells. (A) T-47D cells were treated with indicated concentrations of recombinant human growth hormone (GH) for 6 h to assess mRNA expression. Relative mRNA expression was quantified by real-time RT-PCR and presented as fold change over untreated cells (0.1% PBS), mean ± SEM, n = 3. (B) T-47D cells were treated with PBS (0.1% v/v) or 200 ng/mL GH for 1 h and STAT5 recruitment to the regions of interest was assessed by ChIP. A non-specific rabbit IgG antibody was included as negative control. Recruitment is presented as % of total input, mean ± SEM (n = 4). ** p<0.01 compared to PBS treatment, Student’s t-test.

95 3.2 AIM2: to examine whether STAT5 also plays a role in the regulation of ABCG2 in the lactating mammary gland in vivo. Please note that certain passages and figures within this section 3.2 contain materials from Wu et al. (2014) published by the American Physiological Society.

3.2.1 The E1b isoform is the predominant isoform expressed in the mouse lactating mammary gland Given the results observed using human T-47D breast cancer cells in vitro, my next goal was to determine whether STAT5 also plays a role in the expression of ABCG2 in the lactating mammary gland in vivo. Due to obvious ethical reasons, I had to use the mouse model but I first had to understand which Abcg2 mRNA isoforms (E1a, E1b, and E1c) are expressed and induced in the lactating mammary gland. This was important so as to understand which genomic region may be under the greatest regulatory control during lactation. To gain insight into the Abcg2 mRNA isoform expression profile, I quantified the expression of E1a, E1b, and E1c isoforms in the mammary gland of virgin and lactating FVB and C57BL/6 mice (Fig. 19A and 19B). The expression of all three Abcg2 mRNA isoforms was very low or undetectable in the virgin mammary gland. In contrast, all three isoforms, particularly E1b, which was the main transcript expressed, were robustly detected in the lactating mammary gland. This observation might be explained by differences in relative abundance of epithelial cells between the lactating (i.e. epithelia rich) and virgin (i.e. epithelia- depleted) mammary gland, which is almost entirely composed of fat cells, rather than Abcg2 (E1b) upregulation in the mammary epithelia during lactation. Therefore, I further examined the expression profile of these Abcg2 mRNA isoforms in lineage-depleted cell populations enriched with mammary epithelia cells (MEC) isolated from mammary glands of virgin FVB mice. Indeed, all three isoforms were robustly detected in MECs (small panel, Fig. 19A). However, the levels of expression were an order of magnitude lower than that observed in the lactating mammary gland. This demonstrates that the increased expression of Abcg2 mRNA in the mammary gland during lactation was not simply due to an increase in epithelial cell content but perhaps instead due to mechanisms such as transcriptional upregulation. It is important to note that MECs isolated from lactating mammary glands had very low viability, possibly due to removal from the hormone-rich environment that helps sustain these cells. As such, no direct comparisons could be made between virgin and lactating MECs.

96 Noteworthy, the liver also showed E1b predominance, but unlike the mammary gland, expression of the Abcg2 isoforms did not change with lactation (Fig. 19). These results suggest that regulatory mechanisms governing increased Abcg2 transcript expression during lactation is specific to the mammary gland. Interestingly, while the expression of the Abcg2 mRNA isoforms in the liver were comparable between virgin and lactation, the E1b and E1c isoforms showed male-predominant expression in adult FVB mice (Fig. 20). Therefore, in both the mammary gland and liver, the E1b and E1c isoforms are dynamically expressed.

3.2.2 Prolactin does not induce Abcg2 mRNA in mouse mammary epithelial cell lines As was demonstrated above, prolactin upregulated human ABCG2 mRNA in T-47D breast cancer cells by inducing STAT5 recruitment to a GAS motif in the ABCG2 gene. Given that Abcg2 mRNA isoforms are induced in the mouse mammary gland during lactation, I sought to identify an in vitro system that could be first used to to investigate whether the PRL/PRLR/JAK2/STAT5 pathway is involved before attempting to examine it in vivo. Among the very limited mouse cell lines that respond to prolactin, I first focused on HC11 mouse mammary epithelial cells. HC11 cells are a clone of COMMA-1D cells that were originally derived from mammary glands of BALB/c mice at mid-gestation, which functionally differentiates to express β-casein (Csn2) after prolactin treatment (Ball et al., 1988; Danielson et al., 1984). Under various treatment protocols adapted from Hovey et al. (2003), Abcg2 mRNA was not upregulated in prolactin-treated HC11 cells, despite induction of the mRNA for prolactin-responsive genes Cish and Csn2 (Fig. A5). However, this initial study only assessed global/total Abcg2 mRNA expression; therefore, modulation of low expressing or minor isoforms may be masked by high expressing isoforms. Therefore, a second set of independent experiments in which HC11 cells were treated with PRL for three days after reaching 100% confluence was performed to further assess the relative expression of individual isoforms. Consistent with previous results (Fig. A5 panel B), PRL induced the mRNA of Cish and Csn2 but did not induce Abcg2 or individual mRNA isoforms (Fig. 21A and 21B). Instead, there was a modest but statistically significant reduction in total Abcg2 mRNA expression. To determine if this response is cell-line specific, I further treated EpH4 mouse mammary epithelial cells with PRL. Like HC11 cells, EpH4 cells are a subclone of spontaneously immortalized cells derived from the mammary gland of BALB/c mice at mid-gestation (Fialka et al., 1996; Reichmann et al., 1989). The ability of these cells to respond robustly to PRL

97 treatment, characterized by prolonged activation of STAT5 and Csn2 mRNA expression, is highly dependent on the presence of laminin-111 rich extracellular matrix (Xu et al., 2009). Therefore, EpH4 cells were treated in suspension in medium supplemented with matrigel. In agreement with the HC11 results, PRL induced Cish and Csn2 mRNA but not Abcg2 mRNA or the only detectable E1b isoform (Fig. 21C).

3.2.3 Forced weaning (pup removal) model of involution and loss of activated Stat5 in the mammary gland Given the lack of other suitable in vitro models, it was not clear whether the results observed in mouse mammary epithelial cells in vitro are indicative of those in vivo. For this reason, I treated virgin FVB mice with prolactin to determine if Abcg2 expression could be upregulated in the mammary gland. However, not only was Abcg2 expression not induced by prolactin, but other well-characterized prolactin-responsive genes (Cish, Wap, Csn2 – not detectable) were also not induced (Figure A6). This suggested that the treatment protocol failed to induce a prolactin response. This was not entirely unexpected; since these mice were non-lactating, the mammary gland likely expressed much lower levels of PRLR compared to the lactating mammary gland. More importantly, careful review of the literature and through personal communication with Dr. Charles V. Clevenger (collaborator at Northwestern University), it was revealed that prolactin treatment alone is insufficient to elicit a response that mimics lactation. In a very comprehensive and controlled study using hypophysectomised- ovariectomized-adrenalectomized C4H/He Crgl non-lactating female mice, Nandi (1958) showed that prolactin must be given with a battery of hormones (estrogen, progesterone, and corticosteroids) under a defined treatment schedule in order to induce mammary gland development that resembles lactation. While this treatment protocol could artificially induce lactation, the long treatment duration and large number of hormones administered would make it impossible to distinguish between primary and secondary effects. Therefore, I used the forced weaning model as an alternative strategy, rather than exogenous prolactin treatment, to manipulate STAT5 activity in the mammary gland. In the forced weaning model, pups are pre- maturely removed from the lactating mother to trigger the cessation of lactation, causing the mammary gland to undergo involution. During the first 48 h, there are changes to the mammary gland such as loss of activated STAT5 and increased STAT3 activation, with some minor anatomical changes (Li et al., 1997). There is still a high level of epithelial cells present

98 that allow the gland to resume lactation if the suckling stimulus is restored. This model has been used by others to compare STAT5 binding between lactating and non-lactating mammary glands (at early stage involution) that have comparable epithelial content (Creamer et al., 2010). Using this model, I removed pups from lactating FVB mice and collected the glands 24 h and 48 h later. Time-matched animals where pups remained with the mother were included as controls. As expected, and consistent with this model, there was a dramatic loss of activated (phosphorylated) STAT5 as early as 24 h after forced weaning (Fig. 22A). There was also a reduction of total STAT5B at 24 h and a reduction of both STAT5 isoforms at 48 h after forced weaning. Using immunohistochemistry, it was further determined that this differential STAT5 activity was specific to the mammary luminal epithelial cells (Fig. 22B).

3.2.4 Effect of forced weaning on mouse Abcg2 expression in the mammary gland Next, I examined whether this loss of activated STAT5 was accompanied by a corresponding reduction in Abcg2 expression. Forced weaning led to a significant reduction in the mRNA expression of total Abcg2 and milk proteins Csn2 and Wap (Fig. 23A). This decline in total Abcg2 transcripts was reflected in the significant reduction of not only the predominant E1b isoform, but also the other isoforms (Fig. 23B). Interestingly, despite some changes to cell morphology, ABCG2 protein remained highly expressed in the mammary gland after forced weaning (Fig. 24A). In particular, luminal epithelial cells that express ABCG2 protein (Fig. 24B) remained largely intact even at 24 h and 48 h after forced weaning. This supports the validity of the forced weaning model as a means for obtaining mammary gland tissue with relatively intact epithelial integrity that is depleted of active STAT5.

3.2.5 Stat5 is recruited to the mouse Abcg2 gene in the lactating mammary gland To gain insight into whether STAT5 is bound to mouse Abcg2 in the lactating mammary gland, with help from a bioinformatician (Dr. Andrei Turinksy), I first analyzed a recently published ChIP-seq dataset from Yamaji et al., (2013). These authors performed an elegant experiment whereby the mammary fat pads from mice carrying different gene copies of mouse Stat5a and Stat5b were transplanted into the mammary fat pad of nude mice, which were then mated to induce pregnancy-related mammary gland development. Mammary glands were then collected at day 1 of lactation. Analysis of this dataset showed that STAT5A was significantly enriched at nine regions along the Abcg2 gene in donor mammary glands from

99 wild type mice (Fig. 25, and genomic location identified in Table A1). Recruitment of STAT5A to these regions was significantly reduced in mammary glands from Stat5a knockout mice. The top five regions (denoted hereon as Peaks 1 to 5) that showed particularly high affinity for STAT5A were selected for futher analysis. With the exception of Peak2 that contains only one half of a consensus GAS motif, all other peaks contain at least one putative GAS motif (Table XI). Using ChIP, I compared STAT5 recruitment to these regions in the lactating mouse mammary glands to non-lactating mammary glands from mice after forced weaning (Fig. 26B) or from virgin mice (Fig. 26C). STAT5 was recruited to all five regions of the Abcg2 gene in the lactating mammary gland but not to other regions that contain a putative consensus GAS motif (defined in silico) or a region that is completely devoid of GAS motif sequences (see schematic Fig. 26A). STAT5 recruitment to the five peak regions and known STAT5-binding GAS motifs in the Csn2 and Wap genes was reduced in the mammary glands of mice after forced weaning and in the virgin mammary gland. Collectively, these results demonstrate that STAT5 was recruited to Abcg2 in the mouse mammary gland during lactation.

3.2.6 Functional assessment of STAT5 binding regions for enhancer activity Many of the regions that bound STAT5 during lactation were located a distance away from the three alternative Abcg2 gene promoters. Therefore, instead of using the native gene promoters, I used a luciferase reporter construct with a minimal promoter to assess the functional activity of these STAT5 binding regions. It should be noted that this minimal promoter reporter system did not function well in HC11 cells. This is supported by my unpublished observation that prolactin failed to induce the reporter activity of GAS6v2.4 (data not shown), which I had previously shown to be highly inducible by prolactin in T-47D cells. For this reason, I used T-47D cells to assess the functional activity of these STAT5 binding regions. Among the various regions examined for enhancer activity, only the reporter construct containing GAS4 was strongly induced by prolactin treatment (Fig. 27A). A more modest but statistically significant induction was also observed for the reporter construct containing the Peak2. I further examined the effect of overexpressing STAT5A on reporter activity. Ectopic expression of mouse STAT5A further induced reporter activity of the mAbcg2-GAS4 construct after prolactin treatment (Fig. 27B). The reporter activity of mAbcg2-Peak2 and mAbcg2-GAS2 was also enhanced by overexpression of mouse STAT5A. This suggests that in addition to GAS4, other GAS motifs/STAT5 binding regions, under

100 certain defined conditions such as those in vivo, could also be functional. The detailed analysis of these GAS motifs is the subject of future work. Here I show that single mutations to the GAS4 motif significantly abolished reporter activity after prolactin treatment (Fig. 27C). Together, these results show that some of these STAT5 binding regions may serve as functional enhancer elements of the Abcg2 gene in the mammary gland during lactation.

Figure 19. The E1b Abcg2 mRNA isoform is the predominant isoform expressed in the mouse mammary gland during lactation. The expression of Abcg2 mRNA isoforms (E1a, E1b, and E1c) in the mammary gland and liver of virgin and lactating FVB (panel A) and C57BL/6 (panel B) mice. Samples enriched with mammary epithelial cells (MEC) were obtained from virgin FVB mice. Data presented are copies per µg of RNA, mean ± SD, n = 4-6 per group. The difference in mRNA isoform expression within each tissue at each stage was tested for significance using one-way ANOVA followed by Tukey’s post-hoc pairwise comparison. # p < 0.05, ### p <0.001, compared to E1a isoform; ** p < 0.01, *** p < 0.001, compared to E1c isoform. n.d., not detected. No Abcg2 mRNA isoforms were detected in the mammary gland of two C57BL/6 mice. For these animals, a limit of detection was assigned.

101

Figure 20. Sexually dimorphic expression of Abcg2 mRNA isoforms in the mouse liver. Relative mRNA expression was quantitated by real-time RT-PCR, normalized to Gapdh gene expression, and presented as fold change over female (set as 1). Results shown are mean fold change ± SD, n = 3 per group. Significance was tested using the Student’s T-test. ** p < 0.01.

102

Figure 21. Prolactin does not induce Abcg2 mRNA in mammary epithelial cell lines. HC11 cells were treated with the indicated concentration of recombinant mouse prolactin (PRL) for three days and mRNA expression of (A) total Abcg2, Cish, Csn2, and (B) Abcg2 mRNA isoforms was assessed by real-time RT-PCR, using Gapdh as housekeeping gene. Results shown are fold change relative to cells not treated with prolactin, mean ± SEM, n =3. Significance was tested by one-way ANOVA, followed by Dunnett’s post-hoc comparison to untreated cells. * p<0.05, ** p <0.01, ***p<0.001. (C) EpH4 cells were treated with PBS (0.1% v/v) or recombinant mouse PRL (1 μg/mL) for 48 h. Relative expression of mRNA was quantified by real-time PCR, using β-actin as housekeeping gene, and presented as fold change relative to PBS control, mean ± SEM, n = 4. The expression of total Abcg2 and the E1b isoform did not change with treatment (data not shown). E1a and E1c isoforms could not be reproducibly detected. Csn2 was induced in each independent experiment but was variable between experiments, p=0.075. * p< 0.05, Student’s T-test.

103

Figure 22. Loss of activated STAT5 in the mouse mammary gland after forced weaning. The expression of activated (phosphorylated) STAT5 (p-STAT5) in the mammary glands of lactating FVB mice undergoing forced involution by removal of pups for 24 h or 48 h, and time-matched control mice were assessed by (A) western blot and (B) fluorescent immunohistochemistry. (A) Tissue homogenate were resolved by SDS-PAGE and immunoblotted for phosphorylated STAT5 (p-STAT5), STAT5A, STAT5B, and β-Actin. Protein expression was analyzed by densitometry and presented as relative expression to a standard sample present in each gel (not shown), and to β-Actin, mean ± SD, n = 3 per group. Each lane is from a different animal. * p < 0.05, ** p < 0.01, *** p < 0.001, Student’s T-test compared to time-matched controls. (B) p-STAT5 is pseudo-coloured red. Draq5 (blue) and Cytokeratin-18 (CK-18, green) were used to mark the nuclei and epithelial cells, respectively. Representative images of three animals per group. Bar shown indicates 50 μm.

104

Figure 23. Forced weaning induces rapid reduction in Abcg2 mRNA in the mouse mammary gland. Expression of (A) total Abcg2, Csn2, Wap, and (B) individual Abcg2 mRNA isoforms in the mammary gland from lactating FVB mice with pups removed for 24 or 48 h (involution) and corresponding controls (where pups remained with the mother). Expression was normalized to the epithelial marker cytokeratin-18, and presented as fold change over 24 h control animals, mean ± SEM, n = 3 per group. Difference within each time point was tested using the Student’s T-test. * p< 0.05, ** p <0.01, *** p < 0.001.

105

Figure 24. ABCG2 protein remains highly expressed in the mammary gland of mice after forced weaning. (A) Western analysis of mammary gland tissue homogenate prepared from virgin or lactating mice and mice after forced weaning. V: virgin. Control: lactating mice where pups remained with the mother. Involution: mice with pups removed for 24 h and 48 h. Except for virgin tissue (pooled from two mice), each lane is from a different animal. (B) Expression and localization of ABCG2 protein (red) in mammary tissue sections from mice force-weaned for the times indicated (involution) and corresponding time-matched lactating control mice. Cytokeratin18 (CK-18, green) and Draq5 staining (blue) was used to mark epithelial cells and nucleus, respectively. Results shown are representative of 3 animals per group. Bar shown indicates 50 μm.

106

Figure 25. Transcription factor (TF) binding sites in the region of the Abcg2 gene, based on ChIP-seq data from Yamaji et al. (2013). The distribution of sequencing reads are shown in the mammary tissues of the wild type mice (WT) and the Stat5a knockout mice (KO). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) for each TF compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). Similarly, a direct comparison of the wild type to the knockout mice is indicated by brown-scale bars, labeled "WT-KO" (from light beige for less significant, to dark brown for p-values < 1e-95). The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

107

Table XI. Putative Stat5 binding sequences identified from published ChIP-seq data Sequenceb Regiona Positionc (+) (-)

Peak1 GAS1 caagTTCCatgaaaacagg tgttTTCAtggaacttgca - 33778 GAS2 tgatTTCTttgaaaatcat gattTTCAaagaaatcaaa - 33503 Peak2 Half-site gaaaTACCaagaagtatca - 30636 Peak3 Half-site ggcaTTCCaggtagtcatc + 1245 GAS3 atccTTCAaagaaagagtg ctctTTCTttgaaggatga + 1265 Peak4 GAS4 tgaaTTCCaggaatagata tctaTTCCtggaattcaag + 4236 Peak5 GAS5 tctgTTCCttgaacatgtt catgTTCAaggaacagagt +33858 Half-site atagTTCCtagcacagctg +34301

a Based on Stat5 ChIP-seq results from Yamaji et al. (2013). Sequences that contain a STAT5 binding site in both strands and match the consensus sequence TTCNNNGAA are labeled as GAS elements and numbered in order from 5’ to 3’. b Predicted Stat5 binding sequences (uppercase) and flanking regions (lowercase) are shown. Positive (+) and negative (-) strand. c For GAS sites, position of the middle “N” nucleotide relative to the E1b Abcg2 transcription start site (+1) identified by Natarajan et al. (2011). For half-sites, the 5’ nucleotide position in the positive strand was used.

108

Figure 26. STAT5 recruitment to the Abcg2 gene during lactation. (A) Schematic representation of regions along the mouse Abcg2 gene that were amplified by real-time PCR to assess STAT5 recruitment. Triangles depict identified STAT5 (peak) regions. STAT5 recruitment to the Abcg2 gene in the mammary gland tissue isolated from (B) mice whose pups were withdrawn for 24 h or 48 h (involution) and time-matched lactating mice (lactation), and (C) virgin and lactating mice. STAT5 recruitment to established GAS elements in the milk protein genes Csn2 and Wap are shown. Results are mean % input ± SD, n = 3 per group.

109

Figure 27. Functional activity of STAT5-binding regions in the mouse Abcg2 gene. Regions bound by STAT5 during lactation were inserted into a firefly luciferase reporter driven by a minimal promoter (minP) and transiently transfected into T-47D cells. Cells were serum starved and treated with PBS (0.1%v/v) or 200 ng/mL PRL for 24 h. (A) Reporter activity is presented as fold induction with PRL treatment. ** p<0.01, *** p< 0.001, one-way ANOVA, Dunnett’s post-hoc pairwise comparison with prolactin treated empty vector. (B) Cells were transiently transfected with the indicated reporter and an expression vector for mouse STAT5A or empty vector (pcDNA3) control, and then treated with PBS or 200 ng/mL PRL. Reporter activity is presented as fold change over PBS and empty expression vector control. **p<0.01, ***p<0.001, student’s T-test. (C) Cells were transfected with wild-type or variants of the mAbcg2-GAS4 construct containing a single mutation to the GAS motif (indicated in bold italics; GAS4mut1 and GAS4mut2). Data are presented as fold induction after PRL treatment. Significance was tested using one-way ANOVA. **p<0.01, Dunnett’s post-hoc pairwise comparison to mAbcg2-GAS4.

110 3.3 AIM3: to determine the epigenetic profile (permissive or repressive) of the mouse Abcg2 gene in the virgin and lactating mammary gland. Please note that certain passages and figures within this section 3.3 contain materials from Wu et al. (2014) published by the American Physiological Society.

3.3.1 Characterization of a CpG island at the promoter region of the E1b isoform There are numerous reports that in addition to conventional transcriptional control of gene expression mediated largely by STAT5, milk protein genes that are dramatically upregulated during lactation may also be under epigenetic control (Johnson et al., 1983; Rijnkels et al., 2013; Thompson and Nakhasi, 1985; Vanselow et al., 2006). Given that Abcg2 is highly upregulated in the mammary gland during lactation, I hypothesized that a similar epigenetic mechanism may regulate Abcg2. Based on the UCSC Genome Browser (mouse mm9 assembly) default track setting for CpG islands, there is only one CpG island in the entire mouse Abcg2 gene. Interestingly, this CpG island (schematically shown in Fig. 28, right panel) is located at -231/+178 relative to the transcription start site of the E1b isoform, which was the predominant isoform expressed and induced in the lactating mammary gland. To determine the functional significance of this CpG island in vitro, I first investigated E1b promoter activity in HC11 mouse mammary epithelial cells, which express the E1b isoform under normal growth conditions (Fig. A7), using luciferase reporter assays. The E1b promoter region was able to drive luciferase reporter activity, and sequential deletion analysis demonstrated that the minimal promoter is within the -71/+199 region (Fig. 28). To study the effect of CpG methylation on promoter activity, the -377/+199 and -71/+199 E1b promoter region was subcloned into a novel lucia luciferase reporter plasmid that is completely devoid of CpG sites. These constructs were then subjected to methylation at CpG sites that are present only in the E1b promoter inserts. Successful methylation was confirmed by agarose gel electrophoresis for the absence of linearized plasmid after incubation with CpG-methylation sensitive AatII that would otherwise cut a non-methylated CpG site within the insert (Fig. 29A). Methylation of the E1b promoter region almost completely abolished reporter activity in HC11 cells (Fig. 29B). This suggests that methylation of CpG sites could potentially impact the transcriptional activity of the E1b promoter. Therefore, a loss of CpG methylation at the E1b promoter in the mammary gland from virgin to lactation may be a potential mechanism for the upregulated expression of Abcg2 during lactation.

111 3.3.2 E1b promoter methylation status: in vivo and in vitro To investigate whether there is differential CpG methylation at the E1b promoter in vivo, genomic DNA isolated from the mammary gland of virgin and lactating FVB and C57BL/6 mice were treated with sodium bisulfite and pyrosequenced. I also examined CpG methylation in the liver; since this tissue expressed Abcg2 mRNA at similar levels between virgin and lactation, the E1b promoter should not be differentially methylated. Unexpectedly, CpG sites within the E1b promoter region were already hypo-methylated in the mammary gland of virgin mice and continued to be hypo-methylated during lactation (Fig. 30A). Similarly, these CpG sites were hypo-methylated in mammary epithelial cells isolated from mammary glands of virgin FVB mice. This demonstrates that the E1b promoter was not differentially methylated between the virgin and lactating mammary gland. As expected, similar findings were observed in liver. Surprisingly, in contrast to mammary gland tissues or epithelial cells obtained in vivo, which were hypo-methylated at the E1b promoter, the E1b promoter region was partially methylated in HC11 and EpH4 cells even prior to prolactin induction (Fig. 30B). This suggests that the state of the E1b promoter is different between mammary gland cells grown in vitro compared to in vivo, and may explain the lack of induction observed with prolactin treatment in mouse mammary epithelial cells (Fig. 21).

3.3.3 Histone modifications at the mouse Abcg2 gene in the mouse mammary gland In addition to CpG methylation, the state of the chromatin marked by histone modifications can also modulate gene expression. For this reason, recently published ChIP-seq data for H3K4 dimethylation (H3K4me2) in virgin mouse mammary epithelial cells (MEC) and mouse lactating mammary gland were analyzed to investigate whether there is differential enrichement of this open chromatin histone mark at the Abcg2 gene (Rijnkels et al., 2013). H3K4me2 was already enriched at the promoter regions of E1a, E1b, and E1c, in mammary epithelial cells isolated from virgin mammary gland and remained enriched during lactation (Fig. 31). However, a much more significant H3K4me2 enrichment, characterized by a broader genomic coverage, was observed at the E1b and E1c promoter region in the mammary gland during lactation. This was not observed at the E1a promoter, which further supports my observation that E1b and E1c isoforms are dynamically expressed during lactation. Collectively, these results suggests that Abcg2 is likely already poised for expression in the

112 virgin mammary gland and that during lactation the E1b/E1c promoter region adopts a more open chromatin configuration possibly due to transcriptional activation by STAT5.

Figure 28. The -71/+199 region contains the minimal E1b promoter. HC11 cells were transfected with luciferase reporter constructs driven by various lengths of the Abcg2 E1B promoter region. Reporter activity was normalized to the activity of renilla luciferase constitutively expressed from a co-transfected control vector (pRL-TK, 1 ng/well), and expressed as fold over the pGL4.10 empty vector, mean ± SEM, n =3. Significance was tested using one-way ANOVA, followed by Tukey’s post-hoc pairwise comparison. **p<0.01, compared to -1651/+199, Tukey’s test.

113

Figure 29. Methylation of the mouse Abcg2 E1b promoter region reduces promoter activity. (A) Methylated pCpGfree-basic Lucia plasmids containing the indicated region of the mouse Abcg2 E1b promoter is resistant to digestion with the CpG methylation-sensitive AatII enzyme. Plasmids were methylated by incubating with M.SssI methylase or mock methylated in the absence of enzyme. Mock or methylated (Meth) plasmids were purified by QIAspin columns and 1 μL of eluted DNA was digested with AatII. After an overnight digest, the entire reaction was analyzed by agarose gel electrophoresis. (B) HC11 cells were transfected with pCpGfree- basic Lucia plasmids driven by the indicated region of the mouse Abcg2 E1b promoter (- 71/+199 or -377/+199). Results shown are lucia luciferase activity over firefly luciferase activity expressed from a co-transfected control vector (pGL3-control) that constitutively expressed firefly luciferase and represents mean ± SEM of three independent experiments. ** p < 0.01, *** p < 0.001, Student’s t-test.

114

Figure 30. The E1b promoter in mouse mammary gland-derived samples are hypomethylated in vivo but partially methylated in vitro. Bisulfite-treated genomic DNA from (A) tissue or mammary epithelial cells (MEC) of FVB and C57BL/6 mice (n=4-6 per group, mean ± SD) or (B) HC11 and EpH4 cells were pyrosequenced to determine the CpG methylation status at the E1b promoter. HC11 cells at 100% confluence (0d) were treated with PBS or 1 μg/mL prolactin (PRL) for 3 days (3d); n=3, mean ± SEM. EpH4 cells grown in suspension were treated with PBS or 1 μg/mL prolactin (PRL) for 48 h; n =4, mean ± SEM. Results shown are means of average % methylation from 5 CpG sites. n.a., not available.

115

Figure 31. Histone mark H3K4me2 in the region of the Abcg2 gene, based on ChIP-seq data from Rijnkels et al. (2013). The distribution of sequencing reads is shown in lactating mammary gland (MG, top), mammary epithelial cells isolated from 12 week virgin mammary glands (MEC, middle), and liver tissue (Lvr, bottom). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) of H3K4me2 enrichment compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). Similarly, a direct comparison of H3K4me2 in the lactating mammary gland to the virgin MEC is indicated by brown-scale bars, labeled "Mg-Mec" (from light beige for less significant, to dark brown for p-values < 1e-95). The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

116 4. DISCUSSION

4.1 Effect of PRL and the PRLR signalling cascades on ABCG2 expression Please note that certain passages and figures within sections 4.1 to 4.5 contain materials from Wu et al. (2013) and Wu et al. (2014) originally published by the American Society for Pharmacology & Experimental Therapeutics and American Phyisological Society, respectively.

4.1.1 Prolactin induced ABCG2 in T-47D breast cancer cells The multidrug resistance transporter ABCG2 is upregulated starting from late- pregnancy and peaks during established lactation in the mouse mammary gland (Jonker et al., 2005). To date, few studies have addressed the potential regulatory mechanisms that may explain this phenomenon. I have chosen to investigate the effects of prolactin on ABCG2 expression because it is the main hormone responsible for not only the induction, but more importantly for the maintenance of lactation. Here, I showed that ABCG2 was upregulated by PRL in T-47D cells in a dose- and time-dependent manner. This induction was achieved at concentrations (100-200 ng/mL) of prolactin that are commonly observed in normal lactating women. Wang et al. (2008a) was the first to show that PRL induced ABCG2 protein expression in human choriocarcinoma BeWo cells. However, my results extend these findings to T-47D cells and further provide mechanistic insight into the regulation of human ABCG2 by prolactin at the transcriptional level. Among various human-derived breast epithelial cells examined that are tumorigenic (T- 47D and MCF-7) or ‘normal’ (MCF-10A and primary non-lactating HMEC), ABCG2 was only induced by prolactin in T-47D cells. This may be due to the low expression of the long form of the PRLR in these other human mammary-derived cells compared to T-47D (Fig. 9). Others have also noted a much higher PRLR expression in T-47D cells. Radioligand binding assays have demonstrated that T-47D cells expressed two (Canbay et al., 1997) to three fold (Shiu, 1979) more PRLR (binding sites) per cell compared to MCF-7. However, these binding studies did not account for different prolactin receptor isoforms that have similar affinity for prolactin but variable ability to activate signaling pathways downstream of the receptor. This was addressed by more recent findings that T-47D expressed four (Meng et al., 2004; Ormandy et al., 1997) to eight (Peirce et al., 2001) fold more of the transcript encoding the fully functional long form of PRLR compared to MCF-7 cells. This difference in expression was even greater when compared to MCF-10A (~800X) (Meng et al., 2004). There are few

117 studies that have compared the PRLR expression observed in T-47D to the normal breast. Peirce et al. (2001) noted that the mRNA expression of long form PRLR in normal non- lactating human breast was most similar to that of MDA-MB-468 cells, which was approximately 3% of T-47D cells. Perhaps an even more important comparison is whether PRLR expression in T-47D is abnormally high compared to the lactating mammary gland. Unfortunately, to my knowledge there are no data to address this question. In rodents, it is well recognized that the expression of PRLR is upregulated during lactation (Camarillo et al., 2001; Jahn et al., 1991; Mizoguchi et al., 1996; Nishikawa et al., 1994; Watkin et al., 2008), and is further regulated (maintained) by suckling (Kim et al., 1997). Therefore, although only speculative, it is probable that the lactating human mammary gland expresses PRLR at levels comparable to T-47D. It should be mentioned that despite expressing less PRLR, MCF-7 is generally recognized as a prolactin-responsive cell line. In particular, the mitogenic/proliferative effects of prolactin on MCF-7 cells are well-documented (Acosta et al., 2003; Favy et al., 1999; Lemus-Wilson et al., 1995). Furthermore, prolactin has been shown to induce the phosphorylation and activation of FAK, and/or ERK1/2, and AKT in MCF-7 cells (Acosta et al., 2003; Canbay et al., 1997). Others have shown that prolactin induced tyrosine phosphorylation of STAT5A (Utama et al., 2006) and the phosphorylation of JAK2 in MCF-7 cells, albeit at levels lower than T-47D cells (Canbay et al., 1997), demonstrating an intact JAK2/STAT5 pathway. However, few studies, if any, have shown the direct involvement of JAK2/STAT5 in prolactin-modulated gene expression in MCF-7 cells. Here, I showed that the STAT5-regulated CISH gene was not induced in MCF-7 cells (or MCF10A and HMEC) after prolactin treatment. Therefore, ABCG2 was only induced in cells (i.e. T-47D) that showed responsiveness to prolactin treatment, defined by the induction of CISH. This implies that CISH and ABCG2 may share similar regulatory pathways that are functional in T-47D cells but are deficient in MCF-7 cells.

4.1.2 JAK2/STAT5 dependence in the prolactin-ABCG2 response In general, JAK2 is the main tyrosine kinase associated with prolactin receptor signalling but historically, JAK1, another member of the Janus Kinase family, has also been shown to be weakly activated in some cell systems treated with prolactin (Dusanter-Fourt et al., 1994; Han et al., 1997). More recently, Nielson et al. (2007) demonstrated that prolactin

118 induced tyrosine phosphorylation of not only JAK2 but also JAK1 in various human breast cancer cell lines including T-47D and MCF-7. JAK1 and JAK2 were not phosphorylated in MCF-10A cells, which provides further evidence that these cells do not respond to prolactin treatment, possibly due to low PRLR expression. The tyrosine phosophorylation of JAK1 and JAK2 showed comparable kinetics and dose-responses in T-47D cells (Neilson et al., 2007). Upon further examination, it was revealed that JAK1 activation depended on JAK2. However, because a loss of JAK1 expression reduced prolactin-induced activation of ERK1/2, AKT, and STAT5, Nielson et al. (2007) concluded that JAK1 may serve as an enhancer of JAK2- mediated signaling. Nonetheless, JAK2 is the more proximal kinase in this hierarchical pathway. In addition to JAK2, other kinases such as the SRC family of kinases can also be activated by PRL to induce JAK2-independent signalling cascades (please see ‘Prolactin Receptor Signaling’). I showed here that siRNA mediated knockdown of JAK2 expression significantly attenuated the induction of ABCG2 by prolactin. This demonstrates that prolactin induces ABCG2 in T-47D cells through a JAK2-dependent mechanism. While it is well recognized that STAT5 is the major transcription factor directly activated by JAK2, other members of the STAT family, namely STAT1 and STAT3 are also activated by prolactin in T-47D cells (Barclay et al., 2009; Neilson et al., 2007; Schaber et al., 1998). Using a siRNA that simultaneously targets both STAT5A and STAT5B, I showed that knockdown of STAT5A/B expression significantly attenuated the induction of ABCG2 by PRL. This siRNA was designed by Gutzman et al. (2007) and using Western analysis, these authors demonstrated that this siRNA does not target STAT3; however, the effect on STAT1 expression was not assessed. Despite a lack of experimental evidence, this siRNA only targeted STAT5A and STAT5B when it was analyzed in silico for target specificity using the SpliceCentre siRNA-Check tool (allowing 2 mismatches) (Ryan et al., 2008). Similar results were obtained using the NCBI Blast tool for short sequence alignment to the human genome and transcript database (database updated February 04, 2014; accessed April 05, 2014). Therefore, the attenuated response to PRL observed with the siRNA targeting STAT5 is unlikely due to non-specific targeting of STAT1 or STAT3. Additionally, pharmacological inhibition of STAT5 using a novel STAT5 inhibitor at concentrations which others have shown does not affect interferon-gamma induced STAT1 or STAT3 phosphorylation (Müller et al., 2008), also attenuated the prolactin-ABCG2 response. Collectively, these experiments

119 demonstrate that STAT5 plays an important role in the induction of ABCG2 by PRL in T-47D cells.

4.1.3 STAT5 recruitment to a proximal GAS element Most well-characterized STAT5-regulated genes have a proximal GAS site upstream of the transcription start site (list of genes referenced by Ehret et al., 2001). For this reason, I limited my search for GAS sites to a 10-kbp region upstream of the human ABCG2 transcription start site defined in silico by Bailey-Dell et al. (2001). This transcription start site is within a region containing the promoters of the E1A and E1B/C isoforms (Fig. 1A) and is widely considered a reference point for the majority of previously characterized regulatory elements of the human ABCG2 gene (Fig. 2). My in silico analysis for GAS sites that matched the consensus motif TTCNNNGAA identified two GAS sites at positions -4459 and -434 upstream of the transcription start site. However, only the proximal GAS element (-434) recruited STAT5 after prolactin treatment (Fig. 12). The inability of the distal GAS site to bind STAT5 agrees with the more stringent consensus STAT5 (5’-TTCC/TNG/AGAA-3’) binding motif that was identified by in vitro DNA binding site selection using pools of double stranded oligonucleotides (Soldaini et al., 2000). It is important to note that my search for GAS elements was restricted to sequences that matched the consensus motif TTCNNNGAA. There is evidence that instead of binding DNA as a STAT5 dimer, STAT5 tetramers may also bind to DNA sequences that deviate considerably from the consensus motif (Soldaini et al., 2000). Although my criterion for STAT5 binding sites would have excluded these other potential tetrameric binding sites, my ChIP experiments robustly demonstrate that STAT5 was recruited to at least one region in the human ABCG2 gene.

4.1.4 Functional significance of the proximal GAS element It was important to examine the functional significance of STAT5 recruitment to the proximal GAS element. This is because binding of STAT5 to DNA may not necessarily mean transcriptional activation. The luciferase reporter assay was used to assess the function of the proximal GAS site. Under the conditions of my luciferase assay, PRL only induced a modest 2-fold induction in reporter activity driven by the human ABCG2 promoter and 5’-flanking region that contains the proximal GAS element (Fig. 14). In contrast, a much more pronounced induction was observed with the pGL4-CISH reporter that is driven by the human

120 CISH promoter. This difference in fold induction between the two reporter systems may simply reflect differences in the number of GAS elements between the two genes (1 site in ABCG2 vs 4 sites in CISH). Many luciferase reporter systems under the control of the native gene promoter that contain only one STAT5 binding site are also only induced by 2- to 3-fold with prolactin treatment (Brockman et al., 2002; Fang et al., 2008). Therefore, the modest induction observed with the pGL3-ABCG2 reporter is not uncommon. Importantly, a single mutation to the proximal GAS element was sufficient to significantly reduce PRL-induced pGL3-ABCG2 reporter activity, demonstrating that the proximal GAS element was critical to the prolactin response. The function of the proximal GAS element was further studied using reporter constructs driven by different numbers of copies of the proximal GAS element ligated to a minimal promoter. These reporters responded to PRL in a dose- and copy (GAS)-dependent manner (Fig. 15B). Interestingly, when removed from the ABCG2 promoter, one copy of the proximal GAS element was not sufficient to confer PRL responsiveness. A brief survey of the literature revealed that most STAT5 responsive reporters driven by an artificial (i.e. not native) promoter contain at least three STAT5 binding sites (Goffin et al., 1996b; Metón et al., 1999; Ooi et al., 1998; Rani et al., 2014). Therefore, when taken out of the native context, where additional flanking nucleotides may be essential for stabilizing the transcription initiation complex, a larger number of STAT5 binding sites is required to confer responsiveness to PRL. In addition to dose- and copy-dependence, overexpression of rat STAT5A further enhanced GAS6v2.4 reporter activity (Fig. 15C), demonstrating that the proximal GAS site functions as a STAT5 responsive element. Interestingly, PRL-induced reporter activity was enhanced by overexpressing both mouse STAT5A and STAT5B (Fig. 15D), which is consistent with previous reports that both isoforms have similar DNA binding specificity (Ehret et al., 2001). Although speculative, the presence of a STAT5 binding GAS element near the promoter region may serve as an important site for crosstalk between STAT5 and other nuclear receptors that bind the ABCG2 gene.

4.1.5 Growth hormone and regulation of ABCG2 expression JAK2/STAT5 is not only activated by prolactin but by a variety of hormones and cytokines including erythropoietin, interleukins 2, 3, 5, and 7, granulocyte-macrophage-colony stimulating factor, and growth hormone (Tan and Nevalainen, 2008). Here, I further showed

121 that growth hormone also induced ABCG2 and stimulated STAT5 recruitment to the proximal GAS element in T-47D cells (Fig. 18). This demonstrates that JAK2/STAT5 pathway activators are potential inducers of ABCG2. It is important to mention that a caveat of this finding is that primate and human growth hormone can also bind to the PRLR, and that this interaction (affinity) increases by ~8000 fold in the presence of 50 µM ZnCl2 (Cunningham et al., 1990). In my experiments, cells were treated in serum-free phenol red-free RPMI-1640 supplemented with bovine serum albumin. Based on the manufacturer’s formulation, there should not be appreciable contamination of Zn2+ unless by carry-over in the bovine serum albumin. Therefore, it would seem unlikely that the effects observed are completely due to GH binding to the PRLR. However, there is accumulating evidence that GH-induced activation of JAK2/STAT5 in T-47D cells may be mediated by a GHR / PRLR complex (Xu et al., 2011). This is based on the observations that GHR or PRLR antagonists had limited inhibitory effect on the activation of JAK2/STAT5 by GH and that GH further enhanced the co-immuno- precipitation of GHR with PRLR. Therefore, it is not clear whether the induction of ABCG2 by GH in T-47D cells is independent of PRLR-signalling. Interestingly, GH has many galactopoietic properties. Human and mouse GH stimulated milk protein production in mammary gland explants isolated from rhesus monkey (Kleinberg and Todd, 1980) and pregnant mice (Markoff and Talamantes, 1980), respectively. Bovine GH is widely used in the cattle industry to increase milk yield and to extend the lactation period (Bauman, 1999). GH treatment has also been shown to enhance milk yield in normal lactating women (Milsom et al., 1992) and in lactating women with insufficient milk production (Gunn et al., 1996). This raises the question whether GH may instead be responsible for the induction of ABCG2 during lactation. The answer: it is possible, but unlikely. In human, the serum concentration of GH during lactation is less than 10% of that for prolactin and is not affected by suckling (Noel et al., 1974; Widström et al., 1984). In mice, serum prolactin concentration increases at the end of pregnancy and is mostly maintained throughout lactation but GH levels decrease after parturition to approximately one fifth of that observed during pregnancy (Sinha et al., 1974). Based on the serum profile of both of these hormones, GH may contribute to ABCG2 expression during pregnancy but is unlikely to play a significant role in the upregulation of ABCG2 during established lactation.

122 4.1.6 Regulation of ABCG2 by MAPK and PI3K signalling Meyer zu Schwabedissen et al. (2006) were the first to demonstrate a role for MAPK signalling in the regulation of ABCG2 by showing that the MEK inhibitor PD98059 abolished epidermal growth factor (EGF)-induced ABCG2 expression. In the present study, I showed that co-treatment with two different MEK inhibitors, U0126 and PD98059, attenuated the induction of ABCG2 by prolactin. This further supports the involvement of the MAPK pathway in ABCG2 gene regulation. Unlike the JAK2/STAT5 and MAPK pathway, regulation of ABCG2 by PI3K/AKT signalling has been extensively studied. However, as reviewed in the introduction, its exact role is a matter of debate. This is because some studies report that the PI3K/AKT pathway only influences ABCG2 protein localization and possibly expression but has no role in the transcriptional regulation of ABCG2 (Bleau et al., 2009; Mogi et al., 2003; Takada et al., 2005). In contrast, others report the direct involvement of the PI3K/AKT pathway in ABCG2 gene regulation (Huang et al., 2011; Wang et al., 2010d; Zhang et al., 2011). My observation that PI3K inhibitors reduced PRL-induced ABCG2 expression provides further support for the involvement of the PI3K pathway in the transcriptional regulation of ABCG2. Notably, since I had interrogated this pathway only with PI3K specific inhibitors, it is not clear whether AKT is the main effector downstream of PI3K that is mediating the induction of ABCG2 by prolactin. In this study, PRL-stimulated STAT5 recruitment to the ABCG2 proximal GAS element was not affected by MAPK and PI3K pathway inhibitors (Fig. 17), suggesting that STAT5 recruitment to the GAS element is independent of MAPK and PI3K activation. One interesting question that emerges from these results is how does MAPK and PI3K activation contribute to the prolactin-ABCG2 response? One hypothesis is that STAT5 recruits co- regulators that are activated by MAPK and PI3K pathways, and these are necessary for optimal ABCG2 gene expression. For example, co-activators such as NCoA-1/SRC-1 (Litterst et al., 2003) and p300/CBP (Pfitzner et al., 1998), which are known to interact with and enhance STAT5 mediated transcription, have also been shown to be activated by the MAPK and PI3K/AKT pathways (Chen et al., 2004; Huang and Chen, 2005; Janknecht and Nordheim, 1996; Rowan et al., 2000). This is speculative and the identity of these putative effector proteins remains to be determined. Intriguingly, in addition to JAK2/STAT5 that mediates many effects of prolactin, MAPK and PI3K pathway components are over-represented in the transcriptome of lactating mammary gland (Lemay et al., 2007; Maningat et al., 2009). It

123 would be very interesting to see how these pathways interact at the level of the ABCG2 gene. Notably, these interactions are likely gene-specific since under the conditions of my experiments, the induction of CISH expression by prolactin was not affected by MAPK and PI3K pathway inhibitors. An interesting observation from the pathway inhibitor study is the near complete loss of the prolactin-ABCG2 response when T-47D cells were co-treated with LY294002, but not with any other inhibitors. This is possibly due to residual activity as indicated by the presence of phosphorylated ERK1/2 and AKT even after inhibition with PD98059 and wortmannin, respectively (Fig. 16B-D). However, this was not observed when cells were treated with U0126, which completely blocked PRL-induced ERK1/2 phosphorylation (Fig. 16A). A possible explanation is that although by convention these pathways are considered independent, there is in fact a large degree of cross-talk. Aksamitiene et al. (2011) recently showed that LY294002 blocked prolactin-induced ERK1/2 phosphorylation. Through a complex set of experiments, it was further shown that PRL-induced ERK1/2 phosphorylation in T-47D cells is more dependent on the PI3K/Rac/PAK pathway than the canonical (MAPK) GRB2/SOS/RAS pathway (Aksamitiene et al., 2011). Therefore, the inhibition of PI3K by LY294002 likely impaired signalling from both pathways, which may have contributed to the dramatic loss of the prolactin-ABCG2 response. I did not assess ERK1/2 phosphorylation after inhibiting PI3K with LY294002; therefore, it is not clear if this mechanism applies to the conditions of my experiments.

4.2 Role of STAT5 in Abcg2 expression in the lactating mammary gland in vivo

4.2.1 Expression profiling of Abcg2 mRNA isoforms during lactation My novel finding that E1b was the predominant isoform expressed in the mammary gland (Fig. 19) demonstrates that the mammary gland shows a pattern of expression similar to other mouse tissues such as intestine (Natarajan et al., 2011) and liver (Cui et al., 2012), but differs from mouse haematopoietic stem cells that preferentially express the E1a isoform (Zong et al., 2006). While E1b was predominantly expressed, both the E1b and E1c isoforms were induced during lactation. It is interesting that this induction in transcript expression during lactation was tissue specific since no apparent changes were observed in the liver. This agrees with previous reports that ABCG2 protein expression does not change in mouse liver, kidney and intestine during lactation (Merino et al., 2005b), which further suggests that the regulatory

124 mechanisms governing Abcg2 expression during lactation are relatively specific to the mammary gland. Indeed, different tissues may have varying degrees of sensitivity (or measurable responsiveness) to PRL. For example, Jahn et al., (1997) showed that while treatment with ovine PRL induced activation of the JAK2/STAT5 pathway in the mammary gland of ovariectomized-hysterectomized rats, no activation of this pathway was observed in liver. The authors attributed this to the predominant expression of the short form of PRLR in the liver compared to the mammary gland where the long and fully functional form predominates (Jahn et al., 1997; Nagano and Kelly, 1994). Unfortunately, these data are lacking in mice; therefore, it is not clear if such a mechanism applies. It is interesting to note that the mRNAs of milk protein genes Csn2 and Wap were variably induced in the liver of lactating C57BL/6 mice (Fig. A8). However, the level of induction and expression (based on raw Ct values, data not shown) is many orders of magnitude lower than that in the mammary gland. This implies that the liver is relatively insensitive to PRL. Alternatively, since epigenetic mechanisms appear to regulate the the tissue-specific expression of milk protein genes, the lack of induction observed during lactation may not be indicative of the activity of the JAK2/STAT5 pathway. Ultimately, multiple mechanisms are likely involved.

4.2.2 HC11 and EpH4 mouse mammary epithelial cells as models of lactation Given that the PRL responsive genes Cish and Csn2 were induced by prolactin in HC11 and EpH4 cells, it was suprising that Abcg2 was not induced in these cells (Fig. 21). It is well recognized that human and mouse ABCG2/Abcg2 may be regulated quite differently, as shown in the induction of human, but not mouse ABCG2 by aryl hydrocarbon receptor agonists in vitro and in vivo (Han and Sugiyama, 2006; Tan et al., 2010). Given the data presented in this thesis and discussed below, species-specific regulation of ABCG2/Abcg2 is an unlikely explanation. Instead, this may be an additional example that mouse mammary epithelial cell lines may not adequately model all aspects of lactation biology. For example, Jäger et al. showed that the in vivo downregulation of AP-2 in the mouse mammary gland during lactation (Zhang et al., 2003a) was not observed in HC11 cells after differentiation with lactogenic hormones (Jäger et al., 2008).

125 4.2.3 In vivo models to study STAT5 activation and recruitment during lactation There are very limited numbers of models that can be used to study PRL/JAK2/STAT5 activation during established lactation. This is because mammary glands deficient in any one of the above proteins, either through germ-line mutations or tissue-specific conditional knockouts, have severely impaired alveolar development and as such are unable to lactate (Brisken et al., 1999; Cui et al., 2004; Liu et al., 1997; Wagner et al., 2004). A strategy to overcome these deficits is to conditionally overexpress these proteins. However, these conditional transgenic mouse models often rely on the use of the milk protein promoters to drive transgene expression in the mammary epithelium. As a result, these transgenes are only overexpressed during late pregnancy and lactation when the endogenous proteins in this pathway are already at their highest activity. Therefore overexpressing these components during lactation has limited functional effects. For example, mammary glands from lactating transgenic mice that conditionally express PRL under the control of the Wap gene promoter do not overexpress milk proteins but instead also have disorganized alveolar development that result in a high rate of lactation failure (Manhès et al., 2006). The expression of β-casein is only induced by a maximum of ~2.5-fold in lactating mouse mammary glands conditionally expressing STAT5 or a constitutively active STAT5 under the control of the β-lactoglobulin gene (Iavnilovitch et al., 2002). These transgenic models, which have mainly been used to study how components of the PRL/PRLR/JAK2/STAT5 pathway are important in the context of involution and tumorgenesis, have limited utility in this context. The forced weaning model is therefore a convenient method to rapidly and completely turn-off STAT5 activity in the mammary gland during established lactation. As I have demonstrated, and consistent with this model, within 24 h after the removal of the pups there was a near complete loss of phosphorylated STAT5 but only minimal effect on total STAT5A expression (Fig. 22), which is the predominant isoform (~70%) expressed in the mouse mammary gland. Particularly important is that the epithelial compartment is still relatively intact, which allowed me to compare mammary gland tissues with similar tissue architecture that have significantly different STAT5 activity. However, this model has the limitation that is inherent to any form of comparisons between different stages of mammary gland development. That is, the biology of the mammary gland during involution is quite different from that during lactation. For example, during involution there is increased apoptosis, activation of STAT3, tissue remodelling due to activation of tissue proteases, and infiltration by immune cells. This

126 process is largely initiated by the loss of active STAT5 since transgenic mammary epithelia overexpressing constitutively active STAT5 have a delayed onset of involution (Creamer et al., 2010; Iavnilovitch et al., 2002).

4.2.4 STAT5 recruitment to the mouse Abcg2 gene during lactation The mouse Abcg2 gene contains many DNA sequences that match the consensus GAS motif. However, as I have shown using ChIP, STAT5 was only bound to specific regions identified from the STAT5-ChIP seq dataset of Yamaji et al. (2013) (Fig. 26). Furthermore, STAT5 recruitment to these regions was specific to the lactation period since much lower recruitment was observed at involution and in the virgin mammary gland. It is interesting to note that Abcg2 was not initially identified by Yamaji and colleagues as a gene that was bound by STAT5 during lactation (Yamaji et al., 2013). I speculate that in the original analysis, the +1 kb to -50 kb region relative to the transcription start site of E1a was analyzed and therefore, STAT5 recruitment to regions downstream from this site, which includes the regulatory regions of E1b and E1c, the predominant isoforms in the mammary gland, were not accounted for. In my experiments, I focused on five regions that showed the highest affinity for STAT5A. These regions were also bound by STAT5B, and as such should be robust STAT5 binding sites. With the exception of Peak2, most regions bound by STAT5 in the Abcg2 gene contain classical GAS motifs (Table XI). It has been reported that in mouse embryonic fibroblasts overexpressing STAT5, up to 40% of DNA regions bound by STAT5 do not contain a GAS motif (Zhu et al., 2012). It has yet to be determined how STAT5 binds to regions that lack classical GAS motifs but perhaps neighboring transcription factors can help stabilize the STAT5-DNA interaction. Among five regions of the Abcg2 gene that bound STAT5, three were located more than 30 kb from the E1b promoter. This is interesting because in general, gene expression level in the mammary gland immediately following parturition is correlated with STAT5 binding to the proximal region (~+1 kb to -10 kb) of the gene promoter (Yamaji et al., 2013), which may imply that the proximal GAS elements (i.e. GAS4) may be more important. However, recruitment of STAT5 to regions greater than 10 kb from the transcription start site can still have functional consequences as in the case for the human perforin gene, which is upregulated by interleukin-2 via STAT5 activation and recruitment to enhancer elements located 15 kb and 1 kb upstream of the promoter (Zhang et al., 1999). It is not clear if these

127 STAT5 binding regions distal from the E1b promoter are an exception, and to what degree, if any, they contribute to Abcg2 expression in the mammary gland during lactation. More detailed analysis of the Yamaji et al. (2013) dataset show that RNA polyermase was not only localized to the GAS3 and GAS4 region but also localized to GAS5, which is located ~33 kb upstream from the E1b promoter (Fig. 25). Therefore, it seems plausible that the GAS5 site may function as a distal enhancer. Taken together, multiple STAT5 binding sites likely act together to induce Abcg2 during lactation, but additional experiments will be required to test this hypothesis.

4.2.5 Functional GAS motifs in the intron region It is interesting that under the conditions of my luciferase reporter assay, the STAT5 binding region that contains GAS4, which showed the highest enhancer activity after prolactin treatment, is in intron 1 of the E1b isoform. While upstream enhancer elements have traditionally received more attention, the presence of functional STAT5-binding GAS motifs within intron regions is a relatively well-documented phenomenon. For example, the first intron of the human neural cell adhesion molecule-2 (NCAM2) gene contains two functional GAS motifs (Nelson et al., 2006). Additional examples of genes with a functional GAS motif in the intron region includes the human and mouse FOXP3 gene (Yao et al., 2007; Zorn et al., 2006), rat Igf1 gene (Woelfle et al., 2003), and a variant of the mouse Akt1 gene (Creamer et al., 2010).

4.2.6 Abcg2 expression after forced weaning Within 48 h after forced weaning, loss of STAT5 activity was accompanied by a reduction in all Abcg2 mRNA isoforms. Although the link between STAT5 binding and Abcg2 expression in the mammary gland is at most only associative, there is indeed a growing body of evidence that STAT5 plays a role in Abcg2 expression in mice in vivo. In the experimental system used by Yamaji et al. (2013), following parturition, Abcg2 transcript expression was dramatically lower in donor mammary glands from Stat5a knockout mice compared to wild-type mice. In liver, male Stat5b knockout mice similarly show lower expression of Abcg2 compared to the wild-type counterparts (Clodfelter et al., 2006). While my results clearly show that there is a loss of STAT5 occupancy at Abcg2 after forced weaning, it is not clear if this is the only reason for loss of Abcg2 transcriptional activity. This is

128 because as discussed above, during the early phase of involution, in addition to STAT5, the activity of other signalling cascades such as STAT3 are altered. Therefore in addition to passive loss of transcriptional activity, it is also possible that there is active gene repression. Speculatively, one mechanism could be through differential and competing binding of STAT5 vs STAT3 to the Abcg2 gene. There is evidence that STAT5 and STAT3 can have opposing effects on transcription in a gene-specific manner, with the effects of STAT5 dominating when both STATs are active (Walker et al., 2009, 2013). During lactation when STAT5 activity predominates over STAT3, the GAS motifs at the Abcg2 gene may be exclusively occupied by STAT5. With the loss of active STAT5 and increased activation of STAT3 during involution, one would speculate that instead, the majority of GAS motifs may be occupied by STAT3. Although not well characterized, STAT3 recruitment to certain gene promoters has been shown to inhibit gene transcription (Niu et al., 2005; Ramadoss et al., 2009; Saura et al., 2006). Therefore, STAT3 recruitment to the Abcg2 gene could be a mechanism for active loss of Abcg2 expression during involution. This is only my hypothesis but it would be an interesting question to address in the future. It is surprising that although there was a loss of Abcg2 transcript after forced weaning, there were no observable changes to the expression of ABCG2 protein, which suggest that ABCG2-expressing luminal epithelial cells remained relatively intact. Others have shown through experiments using pulse-chase labeling or cycloheximide blocking that the protein half-life of ABCG2 varies widely across cell lines and can range from 4 h to more than 60 h (Diop and Hrycyna, 2005; Hu et al., 2013; Imai et al., 2005; Peng et al., 2010). Therefore, it is not clear whether this disconnect between transcript and protein expression during the first 48 h of forced weaning is simply an inherent lag in the time required for protein levels to reflect the lower transcript expression or that there is activation of post-translational mechanisms that alter the stability of ABCG2 protein. While the specific physiological role for Abcg2 in the lactating mammary gland is unclear, Abcg2 serves as an important transporter of nutrients such as riboflavin into breast milk (van Herwaarden et al., 2007). It seems probable that during the early and reversible phase of involution, there may be a need to preserve Abcg2 function in case the suckling stimulus is restored. However, by one week of involution, Jonker et al., (2005) showed that the expression of Abcg2 protein returns to levels comparable to the virgin gland, which is expected since there would have already been significant remodelling of the mammary gland.

129 4.2.7 Sexually dimorphic expression of the E1b/c isoforms There is ongoing interest to understand the mechanism(s) that is responsible for sexually dimorphic expression of ABCG2 in liver. GH, which activates similar pathways to prolactin, plays an important role in the sexually dimorphic expression of many pharmacologically important genes in the liver (Waxman and Holloway, 2009). Careful review of supplemental data embedded in the literature reveals that GH, through the actions of STAT5B (the predominant isoform in the liver), may have a role in the male-predominant expression of ABCG2 in the liver. After surgical removal of the GH producing pituitary gland (hypophysectomy), hypophysectomised male mice express approximately 40% less Abcg2 mRNA in the liver compared to sham operated male mice (Wauthier et al., 2010). This loss of of Abcg2 expression was not recovered by acute doses of GH, but it may simply be a result of short term treatment. Most interesting is that hypophysectomised female mice expressed more Abcg2 in the liver compared to sham operated female mice, such that hypophysectomy essentially eliminated differences in liver Abcg2 expression between sexes (Wauthier et al., 2010). Male Stat5b knockout mice also express significantly less Abcg2 mRNA in the liver compared to their wildtype counterparts (Clodfelter et al., 2006). This expression was still higher than that in female Stat5b knockout mice, suggesting that STAT5B alone cannot account for all of the differences. Indeed, others suggest that testosterone may be involved in the sexually dimorphic expression of ABCG2 (Tanaka et al., 2005). Note that these studies described above only assessed global/total Abcg2 mRNA expression and therefore information regarding individual mRNA isoforms was lacking. In this study, I show that E1b, which is the predominant isoform expressed in the liver (Fig. 19 and described by Cui et al., 2012), and E1c were both more highly expressed in the liver of male mice compared to female mice (Fig. 20). In contrast, no significant difference was observed with the E1a isoform. This trend is interesting because it parallels the observations in the mammary gland. Female liver is like non-lactating mammary gland whereas male liver resembles the mammary gland during lactation. Using CistromeFinder (Sun et al., 2013) to review the processed ChIP-seq data of Zhang et al. (2012), two regions along mouse Abcg2 were bound by STAT5B in the mouse liver. In mouse liver samples with high STAT5B binding activity (pre-determined by gel-shift assays), STAT5B was bound to a region between peak 1 and 2 (chromosome6: 58579312- 58579513) in the male but not female liver. In contrast, in liver samples with low STAT5B binding activity, STAT5B was bound to the E1b promoter region in female liver but not in the

130 male liver. Surprisingly, these STAT5 binding sites are distinct from the ones characterized in this thesis. Therefore, whether these regions along the Abcg2 gene can serve as STAT5 binding sites depends on cellular context.

4.2.8 E1b isoform is dynamically expressed It is now clear that the three different Abcg2 mRNA isoforms show distinct patterns of expression and vary in their relative expression. Cui et al. (2012) showed that E1b was predominantly expressed in mouse liver and was the only isoform that showed a pattern of decreasing expression immediately following birth. E1b was also the only isoform that showed a circadian-dependent pattern of expression in mouse kidney, liver, and small intestine (Hamdan et al., 2012). Here I further showed that E1b was predominantly expressed and induced in the mammary gland during lactation and was also more highly expressed in male liver. Based on these observations, there is now an emerging view that under normal physiology, E1b is the more quantitatively important and dynamically expressed transcript in differentiated mouse tissues. Few studies that have investigated the regulatory control of mouse Abcg2 have quantified expression at the level of individual mRNA isoforms. It is therefore important for future studies to determine whether E1b is always the isoform that is preferentially expressed or if there is a switch in promoter usage under certain pathophysiological conditions or after pharmacological intervention. This information is crucial because unless we know which promoter is preferentially used under different conditions, it will be impossible to define the regulatory region(s) that should be compared across species.

4.2.9 Activation of STAT5 by mechanisms other than the PRLR It is important to acknowledge that while my in vivo experiments have clearly shown that STAT5 was recruited to mouse Abcg2 during lactation, it is not clear if this is only induced by PRL. This is because STAT5 activity in the lactating mammary gland is not only regulated by prolactin. For example, Long et al., (2003) showed that conditional knockout mice that lack EGFR-like receptor tyrosine kinase Erbb4/Her4 in the mammary epithelium have dramatically retarded late pregnancy/lactation-associated mammary gland development and are further characterized by a loss of activated STAT5 and deficient milk protein gene expression in lobuloalveolar structures. The authors noted that this phenotype resembled that of Stat5a

131 knockout mice. A series of experiments using immortalized cell lines subsequently showed that ERBB4 alone can induce STAT5A transcriptional activity through a not yet fully characterized mechansim involving the direct binding of ERBB4 to STAT5 (Clark et al., 2005; Williams et al., 2004). However, more recent evidence suggests that there is signficiant cross- talk or overlap between PRLR and ERBB4. It was demonstrated that knockdown of ERBB4 in mouse HC11 cells abolished PRL-induced reporter activity driven by the regulatory region of β-casein (Csn2) and resulted in a loss of STAT5 recruitment to the endogenous Csn2 gene (Muraoka-Cook et al., 2008). Most interesting was that PRL also induced the tyrosine phosphorylation of ERBB4 receptor and that this was dependent on JAK2. Together with the observation that prolactin treatment induced formation of a PRLR-JAK2-ERBB4 complex in HC11 cells (Muraoka-Cook et al., 2008), it suggests that ERBB4 may function as a mediator of PRLR signalling. If this is true, then it is plausible that STAT5 activation may be mediated by at least three different pathways: PRL/PRLR/JAK2/STAT5, PRL/PRLR/JAK2/ERBB4/STAT5, or through prolactin-independent signalling by ERBB4- ligand/ERBB4/STAT5.

4.3 Epigenetic profile of the Abcg2 gene in the virgin and lactating mammary gland

4.3.1 Modulation of E1b promoter activity via a CpG island In addition to regulation by PRLR/JAK2/STAT5, there is accumulating evidence that epigenetics may also play a modulatory role in the tissue- and development-specific expression of milk protein genes. I reasoned that since both Abcg2 and milk proteins are highly induced in the mammary gland during lactation, Abcg2 may also be under epigenetic control. It was particularly interesting that the E1b promoter was the only region along mouse Abcg2 that contains a CpG island (defined in silico). To my knowledge, no studies have explored the role of this CpG island on Abcg2/E1b expression. In order to study the effect of CpG methylation on promoter activity, I first had to define the minimal region that contains E1b promoter activity. Deletion analysis using luciferase reporter assays in HC11 cells, which express E1b under normal growth conditions, showed that a minimal promoter is localized within -71/+199 of the E1b transcription start site (Fig. 28). Promoter activity was still observed even with a shorter construct driven by the - 41/+199 E1b region. This is quite surprising given previous reports by Natarajan et al. (2011)

132 that the core/minimal promoter is within the -231/-42 E1b region. It is important to note that in my study, the deletion analysis was performed in HC11 cells using reporter constructs that end at +199 to include the entire CpG island (-231/+178) where as Natarajan et al. (2011) used mouse small intestinal epithelial (MSIE) cells and constructs that terminated at +60. As such, these results may not be directly comparable. After defining the minimal region that conferred maximal E1b promoter activitiy, the - 377/+199 E1b region, which contains the entire CpG island, and the -71/+199 (minimal promoter) region were subcloned into a novel CpG-free lucia luciferase reporter plasmid. By subcloning these inserts into a novel CpG-free reporter plasmid, it was possible to treat the entire plasmid with SssI methylase to methylate CpG sites that are only present within the insert (E1b promoter region). This removed confounding effects that may be associated with methylation of CpG sites that are present in the vector backbone of certain plasmids such as pGL4.10 that was used in the deletion analysis. Methylation of the E1b promoter almost completely abolished reporter activity (Fig. 29B), suggesting that under certain biological conditions, CpG methylation could be a mechanism for regulating E1b promoter activity. However, the E1b promoter was already hypomethylated in mammary epithelial cells and mammary glands isolated from virgin mice and remained hyomethylated in the lactating mammary gland. The liver was also hypomethylated at both states, which is expected since both virgin and lactating mice express comparable levels of E1b in the liver. These results suggest that since the E1b promoter is already hypomethylated in the virgin gland, CpG methylation does not play a role in the upregulation of Abcg2 in the mammary gland during lactation. Furthermore, since both liver and mammary gland were similarly hypomethylated, CpG methylation also cannot explain why Abcg2 was only induced in the mammary gland but not the liver during lactation. Unexpectedly, the E1b promoter was partially methylated in HC11 and EpH4 cells. The biological significance of this partial methylation is not known. In human myeloma cells, ABCG2 mRNA expression is inversely correlated with promoter methylation (Turner et al., 2006). Interestingly, HC11 and EpH4 cells express less E1b isoform (raw ct values ~24 and ~29, respectively) compared to mammary epithelial cells (raw ct values ~21-22) isolated from virgin mice. Perhaps an important question to address is whether this difference in E1b promoter methylation can account for the lack of effect of prolactin on Abcg2 expression in vitro. Several attempts to demethylate this promoter using the DNA methyltransferase

133 inhibitor 5-azacytidine were unsuccessful (unpublished observation). Nonetheless, HC11 and EpH4 cells may not be appropriate models for Abcg2 regulation during lactation.

4.3.2 Methylation at CpG sites as a marker for CpG island/promoter methylation It is important to note that due to high CpG density and large genomic coverage, it was not possible to design primers for bisulfite pyrosequencing that could detect all CpG sites within the E1b CpG island. Instead, the methylation status of the CpG island was based on average methylation of 5 CpG sites that were reliably detected by pyrosequencing (Fig. A9). This is a valid and common approach for determining CpG island methylation status because there is a very good correlation (r>0.9) between methylation status at one CpG site within a CpG island and the methylation levels of the CpG island as as whole (Barrera and Peinado, 2012). Although this is how we currently understand it, it is important to employ a certain level of caution when interpreting these results.

4.3.3 Histone modifications at the E1b promoter Given the well documented role of histone modifications in many aspects of developmental biology, it is quite surprising that few studies have explored its role in gene expression during lactation. Although the regulation of specific milk protein genes by histone modifications has been documented previously (Jolivet et al., 2005; Kabotyanski et al., 2006; Xu et al., 2007, 2009), Rijnkels et al. (2013) was the first to characterize this phenomenon on a genomic scale using ChIP-seq and in vivo samples derived from the mouse mammary gland. Analysis of this ChIP-seq dataset for enrichment of open chromatin histone mark H3K4me2 at the mouse Abcg2 gene revealed that the E1a, E1b, and E1c promoters in virgin mammary epithelial cells are already marked with H3K4me2, suggesting that these regions were already poised for expression (or transcriptionally active) prior to lactation. Indeed, these isoforms are expressed in virgin MEC, albeit at lower levels compared to lactation (Fig. 19). Consistent with the view that H3K4me2 is associated with transcriptionally active chromatin, the promoter region of the E1b and E1c isoforms was further enriched with this histone mark during lactation. It is not clear whether this histone modification is the cause or consequence of transcriptional upregulation of Abcg2. Particularly, because only two time points (virgin vs lactation) were assessed, there is insufficient information to determine when the E1b/c promoters acquired more of this open chromatin histone mark and whether this was preceded by STAT5 recruitment. In a very recent comprehensive study that utilized ChIP-seq and RNA-

134 seq technologies, Kang et al., (2014) showed that genes that were differentially expressed in the mouse mammary gland between early-mid pregnancy to lactation day 1 (parturition) can be stratified into two classes: genes bound by STAT5 at pregnancy day 6 (class I) and genes bound by STAT5 only at lactation (class II). Abcg2 was a class II gene and was induced ~37- fold from early-mid pregnancy to lactation day 1. Note that whole mammary gland tissues were used; therefore, this study did not account for differences in epithelial content. A particularly interesting finding from this study is that through integration of RNA-seq and H3K4me3 ChIP-seq data, Kang and colleagues (2014) demonstrated that the Abcg2 gene was transcribed from a novel promoter. This promoter corresponds to the E1b promoter. Consistent with Rijnkels et al. (2013), the E1b promoter was already enriched with H3K4me3 (another well-characterized open chromatin histone mark) at mid-pregnancy when Abcg2 was expressed at low levels and was not bound by STAT5. At lactation day 1 when STAT5 was bound to Abcg2, there was a 2-fold increase in H3K4me3 enrichment at the E1b promoter. This further suggests that the E1b promoter region is already poised for expression prior to recruitment of STAT5 and that the E1b/c promoter is further enriched with open chromatin histone marks as they become more transcriptionally active. It is important to acknowledge that it is not clear whether STAT5 is responsible for inducing these histone modifications since to date, the role of STAT5 in the modulation of histone methylation remains unclear. Interestingly, revisiting the Yamaji et al. (2013) study, there was no clear difference in H3K4me3 enrichment at the alternative gene promoters between donor mammary glands of wildtype mice and Stat5a knockout mice even though Abcg2 mRNA was expressed at significantly higher levels in the wild-type mammary gland (Fig. A10). This may imply a greater role for mechanisms other than H3K4me3, possibly STAT5, in the differential expression of Abcg2 between these two strains.

4.4 Summary of Findings and Conclusions It has been nearly one decade since Jonker et al. (2005) demonstrated that the breast cancer resistance protein ABCG2 is upregulated in the mammary gland during lactation. This thesis sought to uncover molecular mechanisms responsible for regulating ABCG2 expression during lactation, with the ultimate goal being that a better understanding of physiological ABCG2 regulation might result in better strategies for inhibition of ABCG2-mediated drug efflux in certain pathophysiological conditions.

135 In this thesis, I first showed that prolactin induced ABCG2 mRNA and protein expression in T-47D human breast cancer cells in a dose- and time-dependent manner. Pharmacological inhibiton or siRNA-mediated knockdown of JAK2 and STAT5 further showed that the JAK2/STAT5 pathway is involved in the induction of ABCG2 by prolactin. In particular, STAT5 was recruited to a functional proximal GAS motif in the ABCG2 gene. A series of experiments using small molecule inhibitors directed against PI3K and MAPK signalling pathway components further demonstrated that these pathways are also important in the prolactin-ABCG2 response. However, inhibition of the PI3K or MAPK pathways did not affect STAT5 recruitment to ABCG2, which suggests that an alternative mechanism other than STAT5 recruitment is responsible for cooperatively inducing ABCG2 expression. Given that STAT5 mediated, in part, the induction of ABCG2 by prolactin in vitro, I sought to explore whether STAT5 also plays a role in the regulation of ABCG2 expression in the mammary gland in vivo. Using the mouse model, I first showed that the E1b transcript was predominantly expressed and induced in the lactating mammary gland. However, this phenomenon was not observed in HC11 or EpH4 cell culture models of lactation. Next, a forced-weaning model was used to obtain post-lactating mammary glands that were depleted of active STAT5 while retaining relatively intact tissue architecture. Acutely (24 and 48 h) after forced weaning, there was a significant loss of Abcg2 mRNA expression without detectable changes to protein expression. STAT5 was recruited to multiple regions along the mouse Abcg2 gene in the lactating mammary gland and STAT5 occupancy at these regions was significantly reduced in non-lactating mammary gland. Using luciferase reporter assays, I further showed that at least one of these STAT5 binding regions that contain a GAS motif functioned as an enhancer after prolactin treatment. Together, these results suggest that STAT5 also regulates ABCG2 expression in the lactating mammary gland in vivo. To further understand whether the expression of ABCG2, like the milk proteins, is also regulated by epigenetic mechanisms, I explored the epigenetic profile of the Abcg2 gene in mammary gland of virgin and lactating mice. Despite the presence of a CpG island at the E1b promoter that could function as a repressor of promoter activity in luciferase assays when methylated in vitro, this CpG island was already hypomethylated in vivo in mammary gland samples from virgin mice and remained hypomethylated at lactation. Analysis of published ChIP-seq data further showed that the alternative Abcg2 gene promoters in virgin mammary

136 gland were already enriched with H3K4me2. However, the promoters of E1b and E1c were further enriched with H3K4me2 during lactation. Integrating the findings of this thesis summarized above, the emerging view (depicted schematically in Fig. 32) is that ABCG2 is already poised for expression in the virgin (and possibly non-lactating) mammary gland. During lactation, hormones such as prolactin induce STAT5 recruitment to the ABCG2 gene. Together with other signalling pathways (MAPK and PI3K) that are also activated by prolactin, STAT5 transactivates expression of the ABCG2 gene.

Figure 32. Framework of the molecular mechanisms governing ABCG2 expression in the lactating mammary gland.

137 4.5 Significance and broader implications This thesis has finally addressed the molecular mechanisms involved in the upregulation of ABCG2 in the mammary gland during lactation. It is tempting to explore how these findings may have broader implications beyond understanding how ABCG2 is regulated during lactation. A few interesting “big questions” are discussed below.

Is there a role for ABCG2 in prolactin-induced drug resistance in breast cancer cells? The PRLR is overexpressed in 60-95% of human breast cancers (Gill et al., 2001; Ormandy et al., 1997; Reynolds et al., 1997; Touraine et al., 1998). This is attributed to reduced phosphorylation of Ser349 within the phospho-degron of the PRLR, resulting in impaired degradation of the PRLR in human breast cancers (Li et al., 2006). Prolactin is also produced by the mammary epithelium (Clevenger et al., 1995a; Reynolds et al., 1997) and has been shown to be upregulated in breast cancers compared to normal/hyperplastic epithelium (McHale et al., 2008). There is strong evidence that PRL plays an important role in breast cancer progression and for this reason, there have been efforts to find chemotherapies that target the PRL pathway (Clevenger et al., 2008; Goffin et al., 2005). In addition to its role in the biology of breast cancer progression, there is evidence that PRL can modulate the cytotoxicity of chemotherapeutic agents. Prolactin antagonists have been shown to enhance the cytotoxic effect of cisplatin, paclitaxel, and doxorubicin in breast cancer cell lines (Howell et al., 2008; Ramamoorthy et al., 2001) whereas pre-treatment with PRL can attenuate the cytotoxicity of chemotherapeutic agents taxol, vinblastine, doxorubicin, and cisplatin (LaPensee et al., 2009). My results suggest that prolactin may confer resistance to chemotherapeutics such as doxorubicin, a substrate for ABCG2, by induction of ABCG2. The significance of this induction will require further study.

Is JAK2/STAT5 a major regulator of ABCG2 expression in normo- and pathophysiology? As was discussed earlier, in addition to PRL, various other hormones and cytokines can also activate the JAK2/STAT5 pathway. These includes granulocyte-macrophage colony- stimulating-factor, erythropoietin, thrombopoietin, interleukins 2, 3, 5, 7, and growth hormone (GH) (Tan and Nevalainen, 2008). I have shown here that GH can also induce ABCG2 in T- 47D cells, which may be a reason to further explore whether other JAK2/STAT5 pathway activators are inducers of ABCG2. Aside from the regulated activation of JAK2/STAT5 under

138 normal physiological conditions, this pathway is also activated aberrantly in certain pathophysiological conditions, most notably in certain haemato-lymphoid diorders (Cook et al., 2014; Ikezoe et al., 2011; Lee et al., 2013; Quintás-Cardama and Verstovsek, 2013; Warsch et al., 2013). It would be extremely interesting to determine whether ABCG2 expression is overexpressed in these disorders and what effects it may have on treatment strategy and outcome.

Is there a role for JAK2/STAT5 and epigenetics in the regulation of other transporters that are differentially expressed during lactation? It is important to note that Abcg2 is not the only transporter that is differentially expressed in the mouse mammary gland between virgin and lactation. The mRNA of drug transporters Abcb1a/b (i.e. P-gp/Mdr1), Abcc1 (Mrp1) and Abcc2 (Mrp2) were downregulated whereas the riboflavin transporter Slc52a2 (Rft1) was upregulated during lactation (Fig. A8). While STAT5 is mainly regarded as an activator of gene transcription, there is evidence that it can also function as a transcriptional repressor at certain genes such as BCL6 (Walker et al., 2007). Therefore, STAT5 could potentially act as a positive or negative regulator of these differentially expressed transporters. Based on ChIP-seq data from Yamaji et al. (2013), STAT5 was recruited to genomic regions of Abcc1, Slc52a2 and Slc52a3 but not to Abcb1a/b and Abcc2 in the mouse mammary gland at lactation day 1 (Fig. A11 to A15). It should be noted that my mRNA results were from mammary glands isolated at 1 week of lactation and as such these results should be very cautiously interpreted. Interestingly, ChIP-seq data from Rijnkels et al. (2013) revealed that there was a reduction in the open chromatin histone mark H3K4me2 at the promoters of Abcb1a, Abcb1b, and Abcc1 during lactation (Fig. A16; other genes A17 and A18), which is consistent with the loss of mRNA expression of these genes. Surprisingly, although Slc52a2 was induced during lactation, unlike the E1b promoter, there was no increase in H3K4me2 enrichment at the Slc52a2 promoter (Fig. A18 panel A). From examining a limited number of transporters, it is becoming apparent, although not entirely unexpected, that mechanisms that regulate ABCG2/Abcg2 expression may also regulate other transporter genes such as those described above. Surely additional work is required to fully elucidate the significance of these molecular mechanisms on transporter gene expression.

139 4.6 Recommendations for future study Before addressing some of the “big questions” discussed above, several interesting questions related to the work described in this thesis remain unanswered and could be the focus of future work. These questions along with the proposed experimental approach are listed below in order of importance.

Is there more than one STAT5 binding site in the human ABCG2 gene? My initial in silico assessment for putative STAT5 binding regions in human ABCG2 focused only on the 10-kbp region upstream of the major E1B/C ABCG2 transcription start site and was limited to sequences that matched the consensus GAS motif. However, since STAT5 was recruited to multiple sites along the mouse Abcg2 gene, many more STAT5 binding sites likely exist in the human ABCG2 gene. In consideration of more recent data that 20-40% of STAT5 binding sites do not contain consensus GAS-motifs (Yamaji et al., 2013; Zhu et al., 2012), an unbiased approach must be used to identify STAT5 binding sites along the human ABCG2 gene. One proposed experiment is to perform genome wide ChIP-seq to identify STAT5 binding sites in untreated and prolactin-stimulated T-47D cells. It is important to identify these additional STAT5 binding regions since it may help us better understand how similar the regulatory region is between the mouse and human ABCG2 gene, particularly since there is currently insufficient information to conclude otherwise. Furthermore, by identifying all STAT5 binding sites, it will be possible to focus on these sites by ChIP quantitative-PCR when using samples that are limited such as human mammary epithelial cells isolated from expressed breast milk (see below).

Are these findings reproducible in more biologically relevant samples? The findings discussed in this thesis were derived from either a prolactin responsive human breast cancer cell line in vitro or from the mouse mammary gland in vivo. Given that we now have a much better understanding of how ABCG2 is potentially regulated during lactation, it is important to determine whether some of these findings can be reproduced in more biologically relevant samples such as differentiated milk producing human mammary epithelial cells (milk MEC). Expressed human breast milk contains viable mammary epithelial cells that can be isolated by various techniques and cultured briefly for downstream applications (Boutinaud and Jammes, 2002; Hassiotou et al., 2013). A logical next-step to the

140 experiments described in this thesis and proposed above is to determine whether prolactin induces ABCG2 expression in isolated human milk MEC and whether this is mediated through recruitment of STAT5 to the ABCG2 gene.

Are these STAT5 binding site(s) in the ABCG2 gene crucial for the induction of ABCG2 by prolactin? I have shown here using luciferase reporter assays that prolactin-induced promoter activity was abolished by mutations to the GAS element. However, these luciferase reporter assays contain only a very short fragment of the ABCG2 gene and as such are not truly representative of what may occur if the mutation was made to the genome. In light of recent developments in the CRISPR/Cas9 gene editing technology (Cho et al., 2013; Mali et al., 2013), it is now possible to efficiently mutate specific regions in the genome. Using this technology, it would be very interesting to determine whether genetic mutations to the identified GAS element (or other STAT5 binding sites identified through ChIP-seq) will attenuate prolactin-induced expression of ABCG2.

Could a more physiologically relevant prolactin treatment schedule result in a greater induction of ABCG2? Here, prolactin treatment only induced a 4- to 5-fold increase in ABCG2 expression in T-47D cells, which is modest compared to that in vivo. It is not clear if this is an inherent property of T-47D cells but it does raise the question whether the treatment protocol was optimal. T-47D cells were treated continuously with the same prolactin concentration for 6-24 h but in vivo prolactin concentrations follow a much more phasic/pulsatile pattern. Basal serum prolactin concentrations during lactation are approximately 30-50 ng/mL but this concentration increases by more than 10-fold after nursing/suckling. The biological relevance of this fluctuating prolactin concentration is unclear. GH concentration in males also follows a pulsatile pattern, characterized by even more dramatic fluctuations compared to prolactin. It has been demonstrated that continuous GH treatment leads to an eventual loss of STAT5b activity (Gebert et al., 1999). In contrast, pulsatile treatment allows STAT5b activity to ‘recover’ in between treatments such that an equivalent increase in STAT5b activity can be observed after each successive pulse of GH (Gebert et al., 1997). If the prolactin response resembles that of GH, then continuous prolactin treatment may result in the downregulation of

141 certain components in the PRLR/JAK2/STAT5 pathway. Therefore, a treatment schedule that more resembles the in vivo ‘pulsatile’ pattern of serum prolactin concentration may produce a greater induction of ABCG2.

142 5. APPENDICES

Figure A1. Effect of JAK2 inhibitor AG490 on prolactin-induced expression of ABCG2 and CISH mRNA in T-47D cells. T-47D cells were serum starved overnight in the presence or absence of AG490 (Jak2 inhibitor) followed by a 6 h co-treatment with PBS (0.1% v/v) or recombinant human prolactin (PRL, 500 ng/mL). Data shown are mean ± SEM of three independent experiments.

143

Figure A2. AG490 induced CYP1A1 mRNA expression in T-47D and HepG2 cells. A. T-47D cells were incubated with indicated concentrations of AG490 for 24 h in serum-free medium. Results shown are independent replicates 1 and 2 (of the PBS treatment group for experiments described in Figure A1). B. HepG2 cells were treated with the indicated concentrations of AG490 or 10 nM TCDD for 24 h (mean ± SEM, n=3). CYP1A1 mRNA expression was quantitated by real-time RT-PCR, normalized to GAPDH, and presented as fold change over vehicle (DMSO) control.

144

Figure A3. STAT5 inhibitor reduced prolactin-induced STAT5 phosphorylation. Serum-starved T-47D cells were incubated with DMSO (0.2% v/v) or STAT5 inhibitor (200 µM) for 1 h, then treated with PBS (0.1% v/v) or 100 ng/mL recombinant human prolactin (PRL) for 10 min or 60 min in the presence or absence of inhibitor. Cells were lysed and 10 µg of whole cell lysate was resolved by SDS-PAGE and immunoblotted for phospho-STAT5 (p- STAT5), STAT5A and STAT5B. A non-specific (n.s.) band was observed at ~110 kDa regardless of prolactin stimulation using the p-STAT5 antibody.

145

Figure A4. A single mutation to the proximal ABCG2 GAS element significantly reduced prolactin-induced reporter activity. T-47D cells were transfected with pGL3-ABCG2 or ABCG2/GASmut2 that contains a single mutation (bold italics) to the conserved/consensus putative proximal GAS element and treated with PBS (0.1% v/v) or 100 ng/mL PRL. Mean ±SEM, n=4. * p < 0.05, Student’s t-test.

146

Figure A5. Prolactin does not induce Abcg2 mRNA in prolactin-responsive HC11 mouse mammary epithelial cells. At confluence, HC11 cells were treated with the indicated concentrations of recombinant mouse prolactin for 24 h (panel A) or 3 days (panel B). Alternatively, cells were grown for 3 days post-confluent and treated with prolactin for 24 h (panel C). All treatments were conducted in RPMI 1640 supplemented with 8% FBS, 5 µg/mL insulin, and 1 uM Dexamethasone. Relative transcript expression was quantitated by real-time RT-PCR using Gapdh as housekeeping gene and presented as fold change over untreated cells. Results shown are mean ± SEM, n =3.

147

Figure A6. Prolactin treatment failed to induce mammary gland gene expression in virgin FVB mice. Virgin FVB mice were given daily subcutaneous injections of 1 IU ovine prolactin (PRL, n = 4) or PBS control (n = 8) for four days. On the fourth day, two hours after receiving the last injection, mice were euthanized and mammary gland tissues were collected. Relative mRNA expression was quantitated by real-time RT-PCR, normalized to the expression of cytokeratin-18, and presented as fold change over the PBS treatment group (set as 1). Results are mean ± SD. n.d. not detected.

148

Figure A7. Effect of differentiation on Abcg2 mRNA isoform and Csn2 mRNA expression in HC11 cells. HC11 cells grown in growth medium for 2-3 days until 100% confluence (0d, day 0) or grown for an additional 3 days (3d) in maintenance medium were lysed and the mRNA expression of (A) total Abcg2 and Csn2, and (B) Abcg2 mRNA isoforms, was assessed by real-time RT-PCR using Gapdh as housekeeping gene. Results shown are fold change relative to mRNA expression in HC11 cells at confluence (0d), mean ± SEM, n=3. Significance was tested using the student’s T-test. n.s., not significant; * p < 0.05. Growth medium: RPMI-1640 supplemented with 10% FBS, 10 ng/mL recombinant human EGF, 5 ug/mL recombinant human insulin. Maintenance medium: RPMI-1640 supplemented with 8% FBS, 5 µg/mL recombinant human insulin and 1 µM dexamethasone.

149

Figure A8. Relative mRNA expression of drug and nutrient transporters in the mammary gland and liver of virgin and lactating C57BL/6 mice. Transcript expression was assessed by real-time RT-PCR and presented as fold change from virgin (V, set as 1) to lactation (L). Relative expression was determined by normalizing mammary gland and liver samples to the expression of cytokeratin-18 (epithelial marker) and Gapdh, respectively. Virgin (n=4) and Lactation (n=6). *p<0.05, **p<0.01, ***p<0.001, Student’s T-test.

150

Figure A9. Representative pyrogram of methylation levels at CpG sites within the mouse Abcg2 CpG island. Shown here is the pyrogram for one sample of virgin MEC. Highlighted in blue from left to right are CpG positions +4, +3, +2, +1, -1, and -2 corresponding to nucleotides number +21, +17, +13, +4, -5, -10 relative to the transcription start site defined by Natarajan et al. (2011). Note that CpG position +1 consistently did not pass the quality check of the pyrosequencing reaction (due to high standard deviation) and as such was excluded from the analysis. For all samples, the remaining 5 CpG sites had very consistent % methylation. Therefore, an average % methylation was obtained for each sample by averaging the % methylation of 5 CpG sites.

151

Figure A10. Histone mark H3K4me3 in the region of the Abcg2 gene, based on ChIP-seq data from Yamaji et al., (2013). The distribution of sequencing reads is shown in the mammary tissues of the wild type mice (WT) and the Stat5a knockout mice (KO). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) of H3K4me3 enrichment compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). There was no significant H3K4me2 enrichment in the wild type compared to the knockout mice. The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

152

Figure A11. Transcription factor (TF) binding sites in the region of the Abcb1a/Abcb1b genes, based on ChIP-seq data from Yamaji et al. (2013). The distribution of sequencing reads are shown in the mammary tissues of the wild type mice (WT) and the Stat5a knockout mice (KO). The Y-axis indicates the read count at every 10 bps. There was no statistically significant intervals (p-value < 1e-05) for each TF compared to control data. The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

153

Figure A12. Transcription factor (TF) binding sites in the region of the Abcc1 gene, based on ChIP-seq data from Yamaji et al. (2013). The distribution of sequencing reads are shown in the mammary tissues of the wild type mice (WT) and the Stat5a knockout mice (KO). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) for each TF compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). Similarly, a direct comparison of the wild type to the knockout mice is indicated by brown-scale bars, labeled "WT-KO" (from light beige for less significant, to dark brown for p-values < 1e-95). The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

154

Figure A13. Transcription factor (TF) binding sites in the region of the Abcc2 gene, based on ChIP-seq data from Yamaji et al. (2013). The distribution of sequencing reads are shown in the mammary tissues of the wild type mice (WT) and the Stat5a knockout mice (KO). The Y-axis indicates the read count at every 10 bps. There was no statistically significant intervals (p-value < 1e-05) of each TF compared to control data. The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

155

Figure A14. Transcription factor (TF) binding sites in the region of the Slc52a2 gene, based on ChIP-seq data from Yamaji et al. (2013). The distribution of sequencing reads are shown in the mammary tissues of the wild type mice (WT) and the Stat5a knockout mice (KO). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) of each TF compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). Similarly, a direct comparison of the wild type to the knockout mice is indicated by brown-scale bars, labeled "WT-KO" (from light beige for less significant, to dark brown for p-values < 1e-95). The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

156

Figure A15. Transcription factor (TF) binding sites in the region of the Slc52a3 gene, based on ChIP-seq data from Yamaji et al. (2013). The distribution of sequencing reads are shown in the mammary tissues of the wild type mice (WT) and the Stat5a knockout mice (KO). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) of each TF compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). Similarly, a direct comparison of the wild type to the knockout mice is indicated by brown-scale bars, labeled "WT-KO" (from light beige for less significant, to dark brown for p-values < 1e-95). The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

157

Figure A16. Histone mark H3K4me2 in the region of the Abcb1a/b genes, based on ChIP-seq data from Rijnkels et al. (2013). The distribution of sequencing reads is shown in lactating mammary gland (Mg, top), mammary epithelial cells isolated from 12 week virgin mammary glands (Mec, middle), and liver tissue (Lvr, bottom). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) of H3K4me2 presence compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). A direct comparison of H3K4me2 in the lactating mammary gland to the virgin MEC is indicated by brown-scale bars, labeled "Mg-Mec" (from light beige for less significant, to dark brown for p-values < 1e-95). Similarly, a comparison between virgin MEC to lactating mammary gland is labeled “Mec-Mg”. The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

158 A

B

Figure A17. Histone mark H3K4me2 in the region of the Abcc1 and Abcc2 genes, based on ChIP-seq data from Rijnkels et al. (2013). The distribution of sequencing reads along the Abcc1 (panel A) and Abcc2 (panel B) genes are shown in lactating mammary gland (Mg, top), mammary epithelial cells isolated from 12 week virgin mammary glands (Mec, middle), and liver tissue (Lvr, bottom). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) of H3K4me2 presence compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). A direct comparison of H3K4me2 in the lactating mammary gland to the virgin MEC is indicated by brown-scale bars, labeled "Mg-Mec" (from light beige for less significant, to dark brown for p-values < 1e-95). Similarly, a comparison between virgin MEC to lactating mammary gland is labeled “Mec-Mg”. The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

159 A

B

Figure A18. Histone mark H3K4me2 in the region of the Slc52a2 and Slc52a3 genes, based on ChIP-seq data from Rijnkels et al. (2013). The distribution of sequencing reads along the Slc52a2 (panel A) and Slc52a3 (panel B) genes are shown in lactating mammary gland (Mg, top), mammary epithelial cells isolated from 12 week virgin mammary glands (Mec, middle), and liver tissue (Lvr, bottom). The Y-axis indicates the read count at every 10 bps. Statistically significant intervals (p-value < 1e-05) of H3K4me2 presence compared to control data are indicated by grey-scale bars (from light grey for less significant, to black for p-values < 1e-95). A direct comparison of H3K4me2 in the lactating mammary gland to the virgin MEC is indicated by brown-scale bars, labeled "Mg-Mec" (from light beige for less significant, to dark brown for p-values < 1e-95). Similarly, a comparison between virgin MEC to lactating mammary gland is labeled “Mec-Mg”. The visual representation is created using the UCSC Genome Browser and Mouse mm9 genome assembly.

160

Table AI. Regions of the mouse Abcg2 gene (chr. 6) with significant STAT5A recruitment Peak& start end length summit tags "-10*LOG10(pvalue)" fold_enrichment 1 58556190 58557800 1611 679 492 1234.54 24.33 2 58559119 58560783 1665 825 526 1691.22 34.69 - 58564144 58565120 977 417 98 117.38 9.11 3 58591238 58593011 1774 536 245 706.98 33 4 58593802 58595427 1626 995 400 1634.44 50 - 58596321 58597327 1007 657 77 100.98 6.39 - 58599830 58601208 1379 463 104 145.74 9.43 - 58622581 58623532 952 403 71 103.87 9.29 5 58623747 58625123 1377 641 359 1686.48 36.42 &Dataset from Yamaji et al., (2013). The top 5 regions with the highest affinity for STAT5A (which also bound STAT5B) were denoted as Peaks and were the subject of further analyses.

161 6. REFERENCES

Acosta, J.J., Muñoz, R.M., González, L., Subtil-Rodríguez, A., Domínguez-Cáceres, M.A., García-Martínez, J.M., Calcabrini, A., Lazaro-Trueba, I., and Martín-Pérez, 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, 2268–2282.

Adachi, T., Nakagawa, H., Chung, I., Hagiya, Y., Hoshijima, K., Noguchi, N., Kuo, M.T., and Ishikawa, T. (2007). Nrf2-dependent and -independent induction of ABC transporters ABCC1, ABCC2, and ABCG2 in HepG2 cells under oxidative stress. J. Exp. Ther. Oncol. 6, 335–348.

Adachi, Y., Suzuki, H., Schinkel, A.H., and Sugiyama, Y. (2005). Role of breast cancer resistance protein (Bcrp1/Abcg2) in the extrusion of glucuronide and sulfate conjugates from enterocytes to intestinal lumen. Mol. Pharmacol. 67, 923–928.

Agarwal, S., Sane, R., Gallardo, J.L., Ohlfest, J.R., and Elmquist, W.F. (2010). Distribution of gefitinib to the brain is limited by P-glycoprotein (ABCB1) and breast cancer resistance protein (ABCG2)-mediated active efflux. J. Pharmacol. Exp. Ther. 334, 147–155.

Agarwal, S., Sane, R., Ohlfest, J.R., and Elmquist, W.F. (2011). The role of the breast cancer resistance protein (ABCG2) in the distribution of sorafenib to the brain. J. Pharmacol. Exp. Ther. 336, 223–233.

Aksamitiene, E., Achanta, S., Kolch, W., Kholodenko, B.N., Hoek, J.B., and Kiyatkin, A. (2011). Prolactin-stimulated activation of ERK1/2 mitogen-activated protein kinases is controlled by PI3-kinase/Rac/PAK signaling pathway in breast cancer cells. Cell. Signal. 23, 1794–1805. al-Sakkaf, K.A., Dobson, P.R., and Brown, B.L. (1996). Activation of phosphatidylinositol 3- kinase by prolactin in Nb2 cells. Biochem. Biophys. Res. Commun. 221, 779–784. al-Sakkaf, K.A., Dobson, P.R., and Brown, B.L. (1997). Prolactin induced tyrosine phosphorylation of p59fyn may mediate phosphatidylinositol 3-kinase activation in Nb2 cells. J. Mol. Endocrinol. 19, 347–350. al-Sakkaf, K.A., Mooney, L.M., Dobson, P.R., and Brown, B.L. (2000). Possible role for protein kinase B in the anti-apoptotic effect of prolactin in rat Nb2 lymphoma cells. J. Endocrinol. 167, 85–92.

Allen, J.D., Brinkhuis, R.F., Wijnholds, J., and Schinkel, A.H. (1999). The mouse Bcrp1/Mxr/Abcp gene: amplification and overexpression in cell lines selected for resistance to topotecan, mitoxantrone, or doxorubicin. Cancer Res. 59, 4237–4241.

Allen, J.D., Jackson, S.C., and Schinkel, A.H. (2002a). A mutation hot spot in the Bcrp1 (Abcg2) multidrug transporter in mouse cell lines selected for Doxorubicin resistance. Cancer Res. 62, 2294–2299.

162 Allen, J.D., van Loevezijn, A., Lakhai, J.M., van der Valk, M., van Tellingen, O., Reid, G., Schellens, J.H.M., Koomen, G.-J., and Schinkel, A.H. (2002b). Potent and specific inhibition of the breast cancer resistance protein multidrug transporter in vitro and in mouse intestine by a novel analogue of fumitremorgin C. Mol. Cancer Ther. 1, 417–425.

Allikmets, R., Gerrard, B., Hutchinson, A., and Dean, M. (1996). Characterization of the human ABC superfamily: isolation and mapping of 21 new genes using the expressed sequence tags database. Hum. Mol. Genet. 5, 1649–1655.

Allikmets, R., Schriml, L.M., Hutchinson, A., Romano-Spica, V., and Dean, M. (1998). A human placenta-specific ATP-binding cassette gene (ABCP) on chromosome 4q22 that is involved in multidrug resistance. Cancer Res. 58, 5337–5339.

Ambrosio, R., Fimiani, G., Monfregola, J., Sanzari, E., De Felice, N., Salerno, M.C., Pignata, C., D’Urso, M., and Ursini, M.V. (2002). The structure of human STAT5A and B genes reveals two regions of nearly identical sequence and an alternative tissue specific STAT5B promoter. Gene 285, 311–318.

Anapolsky, A., Teng, S., Dixit, S., and Piquette-Miller, M. (2006). The role of pregnane X receptor in 2-acetylaminofluorene-mediated induction of drug transport and -metabolizing enzymes in mice. Drug Metab. Dispos. 34, 405–409.

Ando, T., Kusuhara, H., Merino, G., Alvarez, A.I., Schinkel, A.H., and Sugiyama, Y. (2007). Involvement of breast cancer resistance protein (ABCG2) in the biliary excretion mechanism of fluoroquinolones. Drug Metab. Dispos. 35, 1873–1879.

Aronica, E., Gorter, J.A., Redeker, S., van Vliet, E.A., Ramkema, M., Scheffer, G.L., Scheper, R.J., van der Valk, P., Leenstra, S., Baayen, J.C., et al. (2005). Localization of breast cancer resistance protein (BCRP) in microvessel endothelium of human control and epileptic brain. Epilepsia 46, 849–857.

Bailey-Dell, K.J., Hassel, B., Doyle, L.A., and Ross, D.D. (2001). Promoter characterization and genomic organization of the human breast cancer resistance protein (ATP-binding cassette transporter G2) gene. Biochim. Biophys. Acta 1520, 234–241.

Bakhsheshian, J., Hall, M.D., Robey, R.W., Herrmann, M.A., Chen, J.-Q., Bates, S.E., and Gottesman, M.M. (2013). Overlapping substrate and inhibitor specificity of human and murine ABCG2. Drug Metab. Dispos. 41, 1805–1812.

Ball, R.K., Friis, R.R., Schoenenberger, C.A., Doppler, W., and 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, 2089–2095.

Banerjee, R., and Vonderhaar, B.K. (1992). Prolactin-induced protein kinase C activity in a mouse mammary cell line (NOG-8). Mol. Cell. Endocrinol. 90, 61–67.

Barclay, J.L., Anderson, S.T., Waters, M.J., and Curlewis, J.D. (2009). SOCS3 as a tumor suppressor in breast cancer cells, and its regulation by PRL. Int. J. Cancer 124, 1756–1766.

163 Barnes, S.N., Aleksunes, L.M., Augustine, L., Scheffer, G.L., Goedken, M.J., Jakowski, A.B., Pruimboom-Brees, I.M., Cherrington, N.J., and Manautou, J.E. (2007). Induction of Hepatobiliary Efflux Transporters in Acetaminophen-Induced Acute Liver Failure Cases. Drug Metab. Dispos. 35, 1963–1969.

Barrera, V., and Peinado, M.A. (2012). Evaluation of single CpG sites as proxies of CpG island methylation states at the genome scale. Nucleic Acids Res. 40, 11490–11498.

Bart, J., Hollema, H., Groen, H.J.M., de Vries, E.G.E., Hendrikse, N.H., Sleijfer, D.T., Wegman, T.D., Vaalburg, W., and van der Graaf, W.T.A. (2004). The distribution of drug- efflux pumps, P-gp, BCRP, MRP1 and MRP2, in the normal blood-testis barrier and in primary testicular tumours. Eur. J. Cancer 40, 2064–2070.

Basseville, A., Tamaki, A., Ierano, C., Trostel, S., Ward, Y., Robey, R.W., Hegde, R.S., and Bates, S.E. (2012). Histone deacetylase inhibitors influence chemotherapy transport by modulating expression and trafficking of a common polymorphic variant of the ABCG2 efflux transporter. Cancer Res. 72, 3642–3651.

Bauman, D.E. (1999). Bovine somatotropin and lactation: from basic science to commercial application. Domest. Anim. Endocrinol. 17, 101–116.

Ben-Jonathan, N., LaPensee, C.R., and LaPensee, E.W. (2008). What can we learn from rodents about prolactin in humans? Endocr. Rev. 29, 1–41.

Benderra, Z., Faussat, A.-M., Sayada, L., Perrot, J.-Y., Chaoui, D., Marie, J.-P., and Legrand, O. (2004). Breast cancer resistance protein and P-glycoprotein in 149 adult acute myeloid leukemias. Clin. Cancer Res. 10, 7896–7902.

Benoki, S., Yoshinari, K., Chikada, T., Imai, J., and Yamazoe, Y. (2012). Transactivation of ABCG2 through a novel cis-element in the distal promoter by constitutive androstane receptor but not pregnane X receptor in human hepatocytes. Arch. Biochem. Biophys. 517, 123–130.

Berlanga, J.J., Gualillo, O., Buteau, H., Applanat, M., Kelly, P.A., and Edery, M. (1997). Prolactin activates tyrosyl phosphorylation of insulin receptor substrate 1 and phosphatidylinositol-3-OH kinase. J. Biol. Chem. 272, 2050–2052.

Bhattacharya, S., Das, A., Mallya, K., and Ahmad, I. (2007). Maintenance of retinal stem cells by Abcg2 is regulated by notch signaling. J. Cell. Sci. 120, 2652–2662.

Bionaz, M., and Loor, J.J. (2008). Gene networks driving bovine milk fat synthesis during the lactation cycle. BMC Genomics 9, 366.

Black, J.C., and Whetstine, J.R. (2011). Chromatin landscape: methylation beyond transcription. Epigenetics 6, 9–15.

Blankenberg, D., Von Kuster, G., Coraor, N., Ananda, G., Lazarus, R., Mangan, M., Nekrutenko, A., and Taylor, J. (2010). Galaxy: a web-based genome analysis tool for experimentalists. Curr. Protoc. Mol. Biol. Chapter 19, Unit 19.10.1–21.

164 Blazquez, A.G., Briz, O., Romero, M.R., Rosales, R., Monte, M.J., Vaquero, J., Macias, R.I.R., Cassio, D., and Marin, J.J.G. (2012). Characterization of the role of ABCG2 as a bile acid transporter in liver and placenta. Mol. Pharmacol. 81, 273–283.

Bleau, A.-M., Hambardzumyan, D., Ozawa, T., Fomchenko, E.I., Huse, J.T., Brennan, C.W., and Holland, E.C. (2009). PTEN/PI3K/Akt pathway regulates the side population phenotype and ABCG2 activity in glioma tumor stem-like cells. Cell Stem Cell 4, 226–235.

De Boussac, H., Orbán, T.I., Várady, G., Tihanyi, B., Bacquet, C., Brózik, A., Váradi, A., Sarkadi, B., and Arányi, T. (2012). Stimulus-induced expression of the ABCG2 multidrug transporter in HepG2 hepatocarcinoma model cells involves the ERK1/2 cascade and alternative promoters. Biochem. Biophys. Res. Commun. 426, 172–176.

Boutin, J.M., Jolicoeur, C., Okamura, H., Gagnon, J., Edery, M., Shirota, M., Banville, D., Dusanter-Fourt, I., Djiane, J., and Kelly, P.A. (1988). Cloning and expression of the rat prolactin receptor, a member of the growth hormone/prolactin receptor gene family. Cell 53, 69–77.

Boutin, J.M., Edery, M., Shirota, M., Jolicoeur, C., Lesueur, L., Ali, S., Gould, D., Djiane, J., and Kelly, P.A. (1989). Identification of a cDNA encoding a long form of prolactin receptor in human hepatoma and breast cancer cells. Mol. Endocrinol. 3, 1455–1461.

Boutinaud, M., and Jammes, H. (2002). Potential uses of milk epithelial cells: a review. Reprod. Nutr. Dev. 42, 133–147.

Bram, E.E., Stark, M., Raz, S., and Assaraf, Y.G. (2009). Chemotherapeutic drug-induced ABCG2 promoter demethylation as a novel mechanism of acquired multidrug resistance. Neoplasia 11, 1359–1370.

Breedveld, P., Pluim, D., Cipriani, G., Wielinga, P., van Tellingen, O., Schinkel, A.H., and Schellens, J.H.M. (2005). The effect of Bcrp1 (Abcg2) on the in vivo pharmacokinetics and brain penetration of imatinib mesylate (Gleevec): implications for the use of breast cancer resistance protein and P-glycoprotein inhibitors to enable the brain penetration of imatinib in patients. Cancer Res. 65, 2577–2582.

Brisken, C., and O’Malley, B. (2010). Hormone Action in the Mammary Gland. Cold Spring Harb. Perspect. Biol. 2, a003178.

Brisken, C., Park, S., Vass, T., Lydon, J.P., O’Malley, B.W., and Weinberg, R.A. (1998). A paracrine role for the epithelial progesterone receptor in mammary gland development. Proc. Natl. Acad. Sci. U.S.A. 95, 5076–5081.

Brisken, C., Kaur, S., Chavarria, T.E., Binart, N., Sutherland, R.L., Weinberg, R.A., Kelly, P.A., and Ormandy, C.J. (1999). Prolactin controls mammary gland development via direct and indirect mechanisms. Dev. Biol. 210, 96–106.

Brockman, J.L., Schroeder, M.D., and Schuler, L.A. (2002). PRL activates the cyclin D1 promoter via the Jak2/Stat pathway. Mol. Endocrinol. 16, 774–784.

165 De Bruin, M., Miyake, K., Litman, T., Robey, R., and Bates, S.E. (1999). Reversal of resistance by GF120918 in cell lines expressing the ABC half-transporter, MXR. Cancer Lett. 146, 117–126.

Bunting, K.D., Zhou, S., Lu, T., and Sorrentino, B.P. (2000). Enforced P-glycoprotein pump function in murine bone marrow cells results in expansion of side population stem cells in vitro and repopulating cells in vivo. Blood 96, 902–909.

Burdon, T.G., Maitland, K.A., Clark, A.J., Wallace, R., and Watson, C.J. (1994). Regulation of the sheep beta-lactoglobulin gene by lactogenic hormones is mediated by a transcription factor that binds an interferon-gamma activation site-related element. Mol. Endocrinol. 8, 1528–1536.

Burger, H., Foekens, J.A., Look, M.P., Meijer-van Gelder, M.E., Klijn, J.G.M., Wiemer, E.A.C., Stoter, G., and Nooter, K. (2003). RNA expression of breast cancer resistance protein, lung resistance-related protein, multidrug resistance-associated proteins 1 and 2, and multidrug resistance gene 1 in breast cancer: correlation with chemotherapeutic response. Clin. Cancer Res. 9, 827–836.

Burger, H., van Tol, H., Boersma, A.W.M., Brok, M., Wiemer, E.A.C., Stoter, G., and Nooter, K. (2004). Imatinib mesylate (STI571) is a substrate for the breast cancer resistance protein (BCRP)/ABCG2 drug pump. Blood 104, 2940–2942.

Caballero, J., Frenette, G., D’Amours, O., Dufour, M., Oko, R., and Sullivan, R. (2012). ATP- binding cassette transporter G2 activity in the bovine spermatozoa is modulated along the epididymal duct and at ejaculation. Biol. Reprod. 86, 181.

Calcagno, A.M., Fostel, J.M., To, K.K.W., Salcido, C.D., Martin, S.E., Chewning, K.J., Wu, C.-P., Varticovski, L., Bates, S.E., Caplen, N.J., et al. (2008). Single-step doxorubicin-selected cancer cells overexpress the ABCG2 drug transporter through epigenetic changes. Br. J. Cancer 98, 1515–1524.

Calo, E., and Wysocka, J. (2013). Modification of enhancer chromatin: what, how, and why? Mol. Cell 49, 825–837.

Camarillo, I.G., Thordarson, G., Moffat, J.G., Van Horn, K.M., Binart, N., Kelly, P.A., and Talamantes, F. (2001). Prolactin receptor expression in the epithelia and stroma of the rat mammary gland. J. Endocrinol. 171, 85–95.

Campbell, P.K., Zong, Y., Yang, S., Zhou, S., Rubnitz, J.E., and Sorrentino, B.P. (2011). Identification of a novel, tissue-specific ABCG2 promoter expressed in pediatric acute megakaryoblastic leukemia. Leuk. Res. 35, 1321–1329.

Campos, C.R., Schröter, C., Wang, X., and Miller, D.S. (2012). ABC transporter function and regulation at the blood-spinal cord barrier. J. Cereb. Blood Flow Metab. 32, 1559–1566.

Canbay, E., Norman, M., Kilic, E., Goffin, V., and Zachary, I. (1997). Prolactin stimulates the JAK2 and focal adhesion kinase pathways in human breast carcinoma T47-D cells. Biochem. J. 324, 231–236.

166 Cargnello, M., and Roux, P.P. (2011). Activation and Function of the MAPKs and Their Substrates, the MAPK-Activated Protein Kinases. Microbiol. Mol. Biol. Rev. 75, 50–83.

Chen, J., Halappanavar, S.S., St-Germain, J.R., Tsang, B.K., and Li, Q. (2004). Role of Akt/protein kinase B in the activity of transcriptional coactivator p300. Cell. Mol. Life Sci. 61, 1675–1683.

Chen, Y., Agarwal, S., Shaik, N.M., Chen, C., Yang, Z., and Elmquist, W.F. (2009). P- glycoprotein and breast cancer resistance protein influence brain distribution of dasatinib. J. Pharmacol. Exp. Ther. 330, 956–963.

Chen, Y., Bieber, M.M., and Teng, N.N.H. (2014). Hedgehog signaling regulates drug sensitivity by targeting ABC transporters ABCB1 and ABCG2 in epithelial ovarian cancer. Mol. Carcinog. 53, 625-534.

Chen, Z.-S., Robey, R.W., Belinsky, M.G., Shchaveleva, I., Ren, X.-Q., Sugimoto, Y., Ross, D.D., Bates, S.E., and Kruh, G.D. (2003). Transport of Methotrexate, Methotrexate Polyglutamates, and 17β-Estradiol 17-(β-d-glucuronide) by ABCG2: Effects of Acquired Mutations at R482 on Methotrexate Transport. Cancer Res. 63, 4048–4054.

Cheng, Y., Zhizhin, I., Perlman, R.L., and Mangoura, D. (2000). Prolactin-induced cell proliferation in PC12 cells depends on JNK but not ERK activation. J. Biol. Chem. 275, 23326–23332.

Chikamatsu, K., Ishii, H., Murata, T., Sakakura, K., Shino, M., Toyoda, M., Takahashi, K., and Masuyama, K. (2013). Alteration of cancer stem cell-like phenotype by histone deacetylase inhibitors in squamous cell carcinoma of the head and neck. Cancer Sci. 104, 1468-1475.

Cho, S.W., Kim, S., Kim, J.M., and Kim, J.-S. (2013). Targeted genome engineering in human cells with the Cas9 RNA-guided endonuclease. Nat. Biotechnol. 31, 230–232.

Clark, D.E., Williams, C.C., Duplessis, T.T., Moring, K.L., Notwick, A.R., Long, W., Lane, W.S., Beuvink, I., Hynes, N.E., and Jones, F.E. (2005). ERBB4/HER4 potentiates STAT5A transcriptional activity by regulating novel STAT5A serine phosphorylation events. J. Biol. Chem. 280, 24175–24180.

Clevenger, C.V., and Medaglia, M.V. (1994). The protein tyrosine kinase P59fyn is associated with prolactin (PRL) receptor and is activated by PRL stimulation of T-lymphocytes. Mol. Endocrinol. 8, 674–681.

Clevenger, C.V., Torigoe, T., and Reed, J.C. (1994). Prolactin induces rapid phosphorylation and activation of prolactin receptor-associated RAF-1 kinase in a T-cell line. J. Biol. Chem. 269, 5559–5565.

Clevenger, C.V., Chang, W.P., Ngo, W., Pasha, T.L., Montone, K.T., and Tomaszewski, J.E. (1995a). Expression of prolactin and prolactin receptor in human breast carcinoma. Evidence for an autocrine/paracrine loop. Am. J. Pathol. 146, 695–705.

167 Clevenger, C.V., Ngo, W., Sokol, D.L., Luger, S.M., and Gewirtz, A.M. (1995b). Vav is necessary for prolactin-stimulated proliferation and is translocated into the nucleus of a T-cell line. J. Biol. Chem. 270, 13246–13253.

Clevenger, C.V., Zheng, J., Jablonski, E.M., Galbaugh, T.L., and Fang, F. (2008). From bench to bedside: future potential for the translation of prolactin inhibitors as breast cancer therapeutics. J Mammary Gland Biol. Neoplasia 13, 147–156.

Clodfelter, K.H., Holloway, M.G., Hodor, P., Park, S.-H., Ray, W.J., and Waxman, D.J. (2006). Sex-dependent liver gene expression is extensive and largely dependent upon signal transducer and activator of transcription 5b (STAT5b): STAT5b-dependent activation of male genes and repression of female genes revealed by microarray analysis. Mol. Endocrinol. 20, 1333–1351.

Cole, S.P., Downes, H.F., and Slovak, M.L. (1989). Effect of calcium antagonists on the chemosensitivity of two multidrug-resistant human tumour cell lines which do not overexpress P-glycoprotein. Br. J. Cancer 59, 42–46.

Cole, S.P., Bhardwaj, G., Gerlach, J.H., Mackie, J.E., Grant, C.E., Almquist, K.C., Stewart, A.J., Kurz, E.U., Duncan, A.M., and Deeley, R.G. (1992). Overexpression of a transporter gene in a multidrug-resistant human lung cancer cell line. Science 258, 1650–1654.

Cook, A.M., Li, L., Ho, Y., Lin, A., Li, L., Stein, A., Forman, S., Perrotti, D., Jove, R., and Bhatia, R. (2014). Role of altered growth factor receptor mediated JAK2 signaling in growth and maintenance of human acute myeloid leukemia stem cells. Blood. 123, 2826-2837.

Cooray, H.C., Blackmore, C.G., Maskell, L., and Barrand, M.A. (2002). Localisation of breast cancer resistance protein in microvessel endothelium of human brain. Neuroreport 13, 2059– 2063.

Cowin, P., and Wysolmerski, J. (2010). Molecular mechanisms guiding embryonic mammary gland development. Cold Spring Harb. Perspect. Biol. 2, a003251.

Creamer, B.A., Sakamoto, K., Schmidt, J.W., Triplett, A.A., Moriggl, R., and Wagner, K.-U. (2010). Stat5 promotes survival of mammary epithelial cells through transcriptional activation of a distinct promoter in Akt1. Mol. Cell. Biol. 30, 2957–2970.

Cressman, A.M., McDonald, C.R., Silver, K.L., Kain, K.C., and Piquette-Miller, M. (2014). Malaria Infection Alters the Expression of Hepatobiliary and Placental Drug Transporters in Pregnant Mice. Drug Metab. Dispos. 42, 603-610.

Cui, J.Y., Gunewardena, S.S., Yoo, B., Liu, J., Renaud, H.J., Lu, H., Zhong, X., and Klaassen, C.D. (2012). RNA-Seq reveals different mRNA abundance of transporters and their alternative transcript isoforms during liver development. Toxicol. Sci. 127, 592–608.

Cui, Y., Riedlinger, G., Miyoshi, K., Tang, W., Li, C., Deng, C.-X., Robinson, G.W., and 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, 8037–8047.

168 Cunningham, B.C., Bass, S., Fuh, G., and Wells, J.A. (1990). Zinc mediation of the binding of human growth hormone to the human prolactin receptor. Science 250, 1709–1712.

Cygalova, L., Ceckova, M., Pavek, P., and Staud, F. (2008). Role of breast cancer resistance protein (Bcrp/Abcg2) in fetal protection during gestation in rat. Toxicol. Lett. 178, 176–180.

Damiani, D., Tiribelli, M., Michelutti, A., Geromin, A., Cavallin, M., Fabbro, D., Pianta, A., Malagola, M., Damante, G., Russo, D., et al. (2010). Fludarabine-based induction therapy does not overcome the negative effect of ABCG2 (BCRP) over-expression in adult acute myeloid leukemia patients. Leuk. Res. 34, 942–945.

Danielson, K.G., Oborn, C.J., Durban, E.M., Butel, J.S., and Medina, D. (1984). Epithelial mouse mammary cell line exhibiting normal morphogenesis in vivo and functional differentiation in vitro. Proc. Natl. Acad. Sci. U.S.A. 81, 3756–3760.

Daood, M., Tsai, C., Ahdab-Barmada, M., and Watchko, J.F. (2008). ABC transporter (P- gp/ABCB1, MRP1/ABCC1, BCRP/ABCG2) expression in the developing human CNS. Neuropediatrics 39, 211–218.

Das, R., and Vonderhaar, B.K. (1996a). Involvement of SHC, GRB2, SOS and RAS in prolactin signal transduction in mammary epithelial cells. Oncogene 13, 1139–1145.

Das, R., and Vonderhaar, B.K. (1996b). Activation of raf-1, MEK, and MAP kinase in prolactin responsive mammary cells. Breast Cancer Res. Treat. 40, 141–149.

Dauchy, S., Dutheil, F., Weaver, R.J., Chassoux, F., Daumas-Duport, C., Couraud, P.-O., Scherrmann, J.-M., De Waziers, I., and Declèves, X. (2008). ABC transporters, cytochromes P450 and their main transcription factors: expression at the human blood-brain barrier. J. Neurochem. 107, 1518–1528.

Deaton, A.M., and Bird, A. (2011). CpG islands and the regulation of transcription. Genes Dev. 25, 1010–1022.

De Vries, N.A., Zhao, J., Kroon, E., Buckle, T., Beijnen, J.H., and van Tellingen, O. (2007). P- glycoprotein and breast cancer resistance protein: two dominant transporters working together in limiting the brain penetration of topotecan. Clin. Cancer Res. 13, 6440–6449.

Dehghan, A., Köttgen, A., Yang, Q., Hwang, S.-J., Kao, W.L., Rivadeneira, F., Boerwinkle, E., Levy, D., Hofman, A., Astor, B.C., et al. (2008). Association of three genetic loci with uric acid concentration and risk of gout: a genome-wide association study. Lancet 372, 1953–1961.

Diestra, J.E., Scheffer, G.L., Català, I., Maliepaard, M., Schellens, J.H.M., Scheper, R.J., Germà-Lluch, J.R., and Izquierdo, M.A. (2002). Frequent expression of the multi-drug resistance-associated protein BCRP/MXR/ABCP/ABCG2 in human tumours detected by the BXP-21 monoclonal antibody in paraffin-embedded material. J. Pathol. 198, 213–219.

Diop, N.K., and Hrycyna, C.A. (2005). N-Linked glycosylation of the human ABC transporter ABCG2 on asparagine 596 is not essential for expression, transport activity, or trafficking to the plasma membrane. Biochemistry 44, 5420–5429.

169 Do, T.M., Noel-Hudson, M.-S., Ribes, S., Besengez, C., Smirnova, M., Cisternino, S., Buyse, M., Calon, F., Chimini, G., Chacun, H., et al. (2012). ABCG2- and ABCG4-mediated efflux of amyloid-β peptide 1-40 at the mouse blood-brain barrier. J. Alzheimers Dis. 30, 155–166.

Doyle, L.A., Yang, W., Abruzzo, L.V., Krogmann, T., Gao, Y., Rishi, A.K., and Ross, D.D. (1998). A multidrug resistance transporter from human MCF-7 breast cancer cells. Proc. Natl. Acad. Sci. U.S.A. 95, 15665–15670.

Dusanter-Fourt, I., Muller, O., Ziemiecki, A., Mayeux, P., Drucker, B., Djiane, J., Wilks, A., Harpur, A.G., Fischer, S., and Gisselbrecht, S. (1994). Identification of JAK protein tyrosine kinases as signaling molecules for prolactin. Functional analysis of prolactin receptor and prolactin-erythropoietin receptor chimera expressed in lymphoid cells. EMBO J. 13, 2583– 2591.

Ebert, B., Seidel, A., and Lampen, A. (2005). Identification of BCRP as transporter of benzo[a]pyrene conjugates metabolically formed in Caco-2 cells and its induction by Ah- receptor agonists. Carcinogenesis 26, 1754–1763.

Ebert, B., Seidel, A., and Lampen, A. (2007). Phytochemicals induce breast cancer resistance protein in Caco-2 cells and enhance the transport of benzo[a]pyrene-3-sulfate. Toxicol. Sci. 96, 227–236.

Ee, P.L.R., Kamalakaran, S., Tonetti, D., He, X., Ross, D.D., and Beck, W.T. (2004a). Identification of a novel estrogen response element in the breast cancer resistance protein (ABCG2) gene. Cancer Res. 64, 1247–1251.

Ee, P.L.R., He, X., Ross, D.D., and Beck, W.T. (2004b). Modulation of breast cancer resistance protein (BCRP/ABCG2) gene expression using RNA interference. Mol. Cancer Ther. 3, 1577–1583.

Ehata, S., Johansson, E., Katayama, R., Koike, S., Watanabe, A., Hoshino, Y., Katsuno, Y., Komuro, A., Koinuma, D., Kano, M.R., et al. (2011). Transforming growth factor-β decreases the cancer-initiating cell population within diffuse-type gastric carcinoma cells. Oncogene 30, 1693–1705.

Ehret, G.B., Reichenbach, P., Schindler, U., Horvath, C.M., Fritz, S., Nabholz, M., and Bucher, P. (2001). DNA binding specificity of different STAT proteins. Comparison of in vitro specificity with natural target sites. J. Biol. Chem. 276, 6675–6688.

ENCODE Project Consortium, Bernstein, B.E., Birney, E., Dunham, I., Green, E.D., Gunter, C., and Snyder, M. (2012). An integrated encyclopedia of DNA elements in the human genome. Nature 489, 57–74.

Enokizono, J., Kusuhara, H., and Sugiyama, Y. (2007a). Effect of breast cancer resistance protein (Bcrp/Abcg2) on the disposition of phytoestrogens. Mol. Pharmacol. 72, 967–975.

Enokizono, J., Kusuhara, H., and Sugiyama, Y. (2007b). Involvement of breast cancer resistance protein (BCRP/ABCG2) in the biliary excretion and intestinal efflux of troglitazone

170 sulfate, the major metabolite of troglitazone with a cholestatic effect. Drug Metab. Dispos. 35, 209–214.

Enokizono, J., Kusuhara, H., Ose, A., Schinkel, A.H., and Sugiyama, Y. (2008). Quantitative investigation of the role of breast cancer resistance protein (Bcrp/Abcg2) in limiting brain and testis penetration of xenobiotic compounds. Drug Metab. Dispos. 36, 995–1002.

Erlichman, C., Boerner, S.A., Hallgren, C.G., Spieker, R., Wang, X.-Y., James, C.D., Scheffer, G.L., Maliepaard, M., Ross, D.D., Bible, K.C., et al. (2001). The HER Tyrosine Kinase Inhibitor CI1033 Enhances Cytotoxicity of 7-Ethyl-10-hydroxycamptothecin and Topotecan by Inhibiting Breast Cancer Resistance Protein-mediated Drug Efflux. Cancer Res. 61, 739–748.

Erwin, R.A., Kirken, R.A., Malabarba, M.G., Farrar, W.L., and Rui, H. (1995). Prolactin activates Ras via signaling proteins SHC, growth factor receptor bound 2, and son of sevenless. Endocrinology 136, 3512–3518.

Evseenko, D.A., Paxton, J.W., and Keelan, J.A. (2007). Independent regulation of apical and basolateral drug transporter expression and function in placental trophoblasts by cytokines, steroids, and growth factors. Drug Metab. Dispos. 35, 595–601.

Faneyte, I.F., Kristel, P.M.P., Maliepaard, M., Scheffer, G.L., Scheper, R.J., Schellens, J.H.M., and van de Vijver, M.J. (2002). Expression of the breast cancer resistance protein in breast cancer. Clin. Cancer Res. 8, 1068–1074.

Fang, F., Antico, G., Zheng, J., and Clevenger, C.V. (2008). Quantification of PRL/Stat5 signaling with a novel pGL4-CISH reporter. BMC Biotechnol. 8, 11.

Fata, J.E., Chaudhary, V., and Khokha, R. (2001). Cellular turnover in the mammary gland is correlated with systemic levels of progesterone and not 17beta-estradiol during the estrous cycle. Biol. Reprod. 65, 680–688.

Fatima, S., Zhou, S., and Sorrentino, B.P. (2012). Abcg2 expression marks tissue-specific stem cells in multiple organs in a mouse progeny tracking model. Stem Cells 30, 210–221.

Favy, D.A., Rio, P., Maurizis, J.C., Hizel, C., Bignon, Y.J., and Bernard-Gallon, D.J. (1999). Prolactin-dependent up-regulation of BRCA1 expression in human breast cancer cell lines. Biochem. Biophys. Res. Commun. 258, 284–291.

Fetsch, P.A., Abati, A., Litman, T., Morisaki, K., Honjo, Y., Mittal, K., and Bates, S.E. (2006). Localization of the ABCG2 mitoxantrone resistance-associated protein in normal tissues. Cancer Lett. 235, 84–92.

Fialka, I., Schwarz, H., Reichmann, E., Oft, M., Busslinger, M., and Beug, H. (1996). The estrogen-dependent c-JunER protein causes a reversible loss of mammary epithelial cell polarity involving a destabilization of adherens junctions. J. Cell Biol. 132, 1115–1132.

Freeman, M.E., Kanyicska, B., Lerant, A., and Nagy, G. (2000). Prolactin: structure, function, and regulation of secretion. Physiol. Rev. 80, 1523–1631.

171 Fresno Vara, J.A., Carretero, M.V., Gerónimo, H., Ballmer-Hofer, K., and Martín-Pérez, J. (2000). Stimulation of c-Src by prolactin is independent of Jak2. Biochem. J. 345 Pt 1, 17–24.

Fresno Vara, J.A., Cáceres, M.A., Silva, A., and Martín-Pérez, J. (2001). Src family kinases are required for prolactin induction of cell proliferation. Mol. Biol. Cell 12, 2171–2183.

Fu, Z.D., Csanaky, I.L., and Klaassen, C.D. (2012). Effects of aging on mRNA profiles for drug-metabolizing enzymes and transporters in livers of male and female mice. Drug Metab. Dispos. 40, 1216–1225.

Gahir, S.S., and Piquette-Miller, M. (2011). Gestational and pregnane X receptor-mediated regulation of placental ATP-binding cassette drug transporters in mice. Drug Metab. Dispos. 39, 465–471.

Gebert, C.A., Park, S.H., and Waxman, D.J. (1997). Regulation of signal transducer and activator of transcription (STAT) 5b activation by the temporal pattern of growth hormone stimulation. Mol. Endocrinol. 11, 400–414.

Gebert, C.A., Park, S.H., and Waxman, D.J. (1999). Down-regulation of liver JAK2-STAT5b signaling by the female plasma pattern of continuous growth hormone stimulation. Mol. Endocrinol. 13, 213–227.

Gilani, R.A., Kazi, A.A., Shah, P., Schech, A.J., Chumsri, S., Sabnis, G., Jaiswal, A.K., and Brodie, A.H. (2012). The importance of HER2 signaling in the tumor-initiating cell population in aromatase inhibitor-resistant breast cancer. Breast Cancer Res. Treat. 135, 681–692.

Gill, S., Peston, D., Vonderhaar, B.K., and Shousha, S. (2001). Expression of prolactin receptors in normal, benign, and malignant breast tissue: an immunohistological study. J. Clin. Pathol. 54, 956–960.

Goffin, V., Shiverick, K.T., Kelly, P.A., and Martial, J.A. (1996a). Sequence-function relationships within the expanding family of prolactin, growth hormone, placental lactogen, and related proteins in mammals. Endocr. Rev. 17, 385–410.

Goffin, V., Kinet, S., Ferrag, F., Binart, N., Martial, J.A., and Kelly, P.A. (1996b). Antagonistic Properties of Human Prolactin Analogs That Show Paradoxical Agonistic Activity in the Nb2 Bioassay. J. Biol. Chem. 271, 16573–16579.

Goffin, V., Bernichtein, S., Touraine, P., and Kelly, P.A. (2005). Development and potential clinical uses of human prolactin receptor antagonists. Endocr. Rev. 26, 400–422.

Gouilleux, F., Wakao, H., Mundt, M., and Groner, B. (1994). Prolactin induces phosphorylation of Tyr694 of Stat5 (MGF), a prerequisite for DNA binding and induction of transcription. EMBO J. 13, 4361–4369.

Gunn, A.J., Gunn, T.R., Rabone, D.L., Breier, B.H., Blum, W.F., and Gluckman, P.D. (1996). Growth hormone increases breast milk volumes in mothers of preterm infants. Pediatrics 98, 279–282.

172 Gunness, P., Aleksa, K., and Koren, G. (2011). Acyclovir is a substrate for the human breast cancer resistance protein (BCRP/ABCG2): implications for renal tubular transport and acyclovir-induced nephrotoxicity. Can. J. Physiol. Pharmacol. 89, 675–680.

Gutmann, H., Hruz, P., Zimmermann, C., Beglinger, C., and Drewe, J. (2005). Distribution of breast cancer resistance protein (BCRP/ABCG2) mRNA expression along the human GI tract. Biochem. Pharmacol. 70, 695–699.

Gutzman, J.H., Rugowski, D.E., Nikolai, S.E., and Schuler, L.A. (2007). Stat5 activation inhibits prolactin-induced AP-1 activity: distinct prolactin-initiated signals in tumorigenesis dependent on cell context. Oncogene 26, 6341–6348.

Guyda, H.J., and Friesen, H.G. (1973). Serum Prolactin Levels in Humans from Birth to Adult Life. Pediatr Res. 7, 534–540.

Hadsell, D.L., Torres, D.T., Lawrence, N.A., George, J., Parlow, A.F., Lee, A.V., and Fiorotto, M.L. (2005). Overexpression of des(1-3) insulin-like growth factor 1 in the mammary glands of transgenic mice delays the loss of milk production with prolonged lactation. Biol. Reprod. 73, 1116–1125.

Hagiya, Y., Adachi, T., Ogura, S., An, R., Tamura, A., Nakagawa, H., Okura, I., Mochizuki, T., and Ishikawa, T. (2008). Nrf2-dependent induction of human ABC transporter ABCG2 and heme oxygenase-1 in HepG2 cells by photoactivation of porphyrins: biochemical implications for cancer cell response to photodynamic therapy. J. Exp. Ther. Oncol. 7, 153–167.

Halwachs, S., Wassermann, L., Lindner, S., Zizzadoro, C., and Honscha, W. (2013). Fungicide Prochloraz and Environmental Pollutant Dioxin Induce the ABCG2 Transporter in Bovine Mammary Epithelial Cells by the Arylhydrocarbon Receptor Signaling Pathway. Toxicol. Sci. 131, 491–501.

Hamdan, A.M., Koyanagi, S., Wada, E., Kusunose, N., Murakami, Y., Matsunaga, N., and Ohdo, S. (2012). Intestinal expression of mouse Abcg2/breast cancer resistance protein (BCRP) gene is under control of circadian clock-activating transcription factor-4 pathway. J. Biol. Chem. 287, 17224–17231.

Han, Y., and Sugiyama, Y. (2006). Expression and regulation of breast cancer resistance protein and multidrug resistance associated protein 2 in BALB/c mice. Biol. Pharm. Bull. 29, 1032–1035.

Han, Y., Watling, D., Rogers, N.C., and Stark, G.R. (1997). JAK2 and STAT5, but not JAK1 and STAT1, are required for prolactin-induced beta-lactoglobulin transcription. Mol. Endocrinol. 11, 1180–1188.

Hang, D., Dong, H.-C., Ning, T., Dong, B., Hou, D.-L., and Xu, W.-G. (2012). Prognostic value of the stem cell markers CD133 and ABCG2 expression in esophageal squamous cell carcinoma. Dis. Esophagus 25, 638–644.

Harmston, N., and Lenhard, B. (2013). Chromatin and epigenetic features of long-range gene regulation. Nucleic Acids Res. 42, 7185-7199.

173 Hartz, A.M.S., Mahringer, A., Miller, D.S., and Bauer, B. (2010a). 17-β-Estradiol: a powerful modulator of blood-brain barrier BCRP activity. J. Cereb. Blood Flow Metab. 30, 1742–1755.

Hartz, A.M.S., Madole, E.K., Miller, D.S., and Bauer, B. (2010b). Estrogen receptor beta signaling through phosphatase and tensin homolog/phosphoinositide 3-kinase/Akt/glycogen synthase kinase 3 down-regulates blood-brain barrier breast cancer resistance protein. J. Pharmacol. Exp. Ther. 334, 467–476.

Hassiotou, F., Geddes, D.T., and Hartmann, P.E. (2013). Cells in human milk: state of the science. J. Hum. Lact. 29, 171–182.

Hauswald, S., Duque-Afonso, J., Wagner, M.M., Schertl, F.M., Lübbert, M., Peschel, C., Keller, U., and Licht, T. (2009). Histone deacetylase inhibitors induce a very broad, pleiotropic anticancer drug resistance phenotype in acute myeloid leukemia cells by modulation of multiple ABC transporter genes. Clin. Cancer Res. 15, 3705–3715.

Hennighausen, L., and Robinson, G.W. (2005). Information networks in the mammary gland. Nat. Rev. Mol. Cell Biol. 6, 715–725.

Herpel, E., Jensen, K., Muley, T., Warth, A., Schnabel, P.A., Meister, M., Herth, F.J.F., Dienemann, H., Thomas, M., and Gottschling, S. (2011). The cancer stem cell antigens CD133, BCRP1/ABCG2 and CD117/c-KIT are not associated with prognosis in resected early-stage non-small cell lung cancer. Anticancer Res. 31, 4491–4500.

Hessel, S., and Lampen, A. (2010). All-trans retinoic acid enhances the transport of phase II metabolites of benzo[a]pyrene by inducing the Breast Cancer Resistance Protein expression in Caco-2 cells. Toxicol. Lett. 197, 151–155.

Hilgendorf, C., Ahlin, G., Seithel, A., Artursson, P., Ungell, A.-L., and Karlsson, J. (2007). Expression of thirty-six drug transporter genes in human intestine, liver, kidney, and organotypic cell lines. Drug Metab. Dispos. 35, 1333–1340.

Hirano, M., Maeda, K., Matsushima, S., Nozaki, Y., Kusuhara, H., and Sugiyama, Y. (2005). Involvement of BCRP (ABCG2) in the biliary excretion of pitavastatin. Mol. Pharmacol. 68, 800–807.

Ho, M.M., Hogge, D.E., and Ling, V. (2008). MDR1 and BCRP1 expression in leukemic progenitors correlates with chemotherapy response in acute myeloid leukemia. Exp. Hematol. 36, 433–442.

Honjo, Y., Hrycyna, C.A., Yan, Q.W., Medina-Pérez, W.Y., Robey, R.W., van de Laar, A., Litman, T., Dean, M., and Bates, S.E. (2001). Acquired mutations in the MXR/BCRP/ABCP gene alter substrate specificity in MXR/BCRP/ABCP-overexpressing cells. Cancer Res. 61, 6635–6639.

Honorat, M., Mesnier, A., Di Pietro, A., Lin, V., Cohen, P., Dumontet, C., and Payen, L. (2008). Dexamethasone down-regulates ABCG2 expression levels in breast cancer cells. Biochem. Biophys. Res. Commun. 375, 308–314.

174 Hoque, M.T., Robillard, K.R., and Bendayan, R. (2012). Regulation of breast cancer resistant protein by peroxisome proliferator-activated receptor α in human brain microvessel endothelial cells. Mol. Pharmacol. 81, 598–609.

Hori, S., Ohtsuki, S., Tachikawa, M., Kimura, N., Kondo, T., Watanabe, M., Nakashima, E., and Terasaki, T. (2004). Functional expression of rat ABCG2 on the luminal side of brain capillaries and its enhancement by astrocyte-derived soluble factor(s). J. Neurochem. 90, 526– 536.

Hosomi, A., Nakanishi, T., Fujita, T., and Tamai, I. (2012). Extra-renal elimination of uric acid via intestinal efflux transporter BCRP/ABCG2. PLoS ONE 7, e30456.

Hovey, R.C., Harris, J., Hadsell, D.L., Lee, A.V., Ormandy, C.J., and Vonderhaar, B.K. (2003). Local Insulin-Like Growth Factor-II Mediates Prolactin-Induced Mammary Gland Development. Mol. Endocrinol. 17, 460–471.

Howell, S.J., Anderson, E., Hunter, T., Farnie, G., and Clarke, R.B. (2008). Prolactin receptor antagonism reduces the clonogenic capacity of breast cancer cells and potentiates doxorubicin and paclitaxel cytotoxicity. Breast Cancer Res. 10, R68.

Hu, Y., Tao, L., Wang, F., Zhang, J., Liang, Y., and Fu, L. (2013). Secalonic acid D reduced the percentage of side populations by down-regulating the expression of ABCG2. Biochem. Pharmacol. 85, 1619–1625.

Hu, Z.Z., Meng, J., and Dufau, M.L. (2001). Isolation and characterization of two novel forms of the human prolactin receptor generated by alternative splicing of a newly identified exon 11. J. Biol. Chem. 276, 41086–41094.

Huang, W.-C., and Chen, C.-C. (2005). Akt Phosphorylation of p300 at Ser-1834 Is Essential for Its Histone Acetyltransferase and Transcriptional Activity. Mol. Cell. Biol. 25, 6592–6602.

Huang, W.-C., Chen, Y.-J., Li, L.-Y., Wei, Y.-L., Hsu, S.-C., Tsai, S.-L., Chiu, P.-C., Huang, W.-P., Wang, Y.-N., Chen, C.-H., et al. (2011). Nuclear translocation of epidermal growth factor receptor by Akt-dependent phosphorylation enhances breast cancer-resistant protein expression in gefitinib-resistant cells. J. Biol. Chem. 286, 20558–20568.

Huls, M., van den Heuvel, J.J.M.W., Dijkman, H.B.P.M., Russel, F.G.M., and Masereeuw, R. (2006). ABC transporter expression profiling after ischemic reperfusion injury in mouse kidney. Kidney Int. 69, 2186–2193.

Huls, M., Brown, C.D.A., Windass, A.S., Sayer, R., van den Heuvel, J.J.M.W., Heemskerk, S., Russel, F.G.M., and Masereeuw, R. (2008). The breast cancer resistance protein transporter ABCG2 is expressed in the human kidney proximal tubule apical membrane. Kidney Int. 73, 220–225.

Hwang, P., Guyda, H., and Friesen, H. (1971). A radioimmunoassay for human prolactin. Proc. Natl. Acad. Sci. U.S.A. 68, 1902–1906.

175 Iavnilovitch, E., Groner, B., and Barash, I. (2002). Overexpression and Forced Activation of Stat5 in Mammary Gland of Transgenic Mice Promotes Cellular Proliferation, Enhances Differentiation, and Delays Postlactational Apoptosis. Mol. Cancer Res. 1, 32–47.

Ichida, K., Matsuo, H., Takada, T., Nakayama, A., Murakami, K., Shimizu, T., Yamanashi, Y., Kasuga, H., Nakashima, H., Nakamura, T., et al. (2012). Decreased extra-renal urate excretion is a common cause of hyperuricemia. Nat. Commun. 3, 764.

Ikezoe, T., Kojima, S., Furihata, M., Yang, J., Nishioka, C., Takeuchi, A., Isaka, M., Koeffler, H.P., and Yokoyama, A. (2011). Expression of p-JAK2 predicts clinical outcome and is a potential molecular target of acute myelogenous leukemia. Int. J. Cancer 129, 2512–2521.

Imai, Y., Asada, S., Tsukahara, S., Ishikawa, E., Tsuruo, T., and Sugimoto, Y. (2003). Breast cancer resistance protein exports sulfated estrogens but not free estrogens. Mol. Pharmacol. 64, 610–618.

Imai, Y., Tsukahara, S., Asada, S., and Sugimoto, Y. (2004). Phytoestrogens/flavonoids reverse breast cancer resistance protein/ABCG2-mediated multidrug resistance. Cancer Res. 64, 4346–4352.

Imai, Y., Ishikawa, E., Asada, S., and Sugimoto, Y. (2005). Estrogen-mediated post transcriptional down-regulation of breast cancer resistance protein/ABCG2. Cancer Res. 65, 596–604.

Jäger, R., Pappas, L., and Schorle, H. (2008). Lactogenic differentiation of HC11 cells is not accompanied by downregulation of AP-2 transcription factor genes. BMC Research Notes 1, 29.

Jahn, G.A., Edery, M., Belair, L., Kelly, P.A., and Djiane, J. (1991). Prolactin receptor gene expression in rat mammary gland and liver during pregnancy and lactation. Endocrinology 128, 2976–2984.

Jahn, G.A., Daniel, N., Jolivet, G., Belair, L., Bole-Feysot, C., Kelly, P.A., and Djiane, J. (1997). In vivo study of prolactin (PRL) intracellular signalling during lactogenesis in the rat: JAK/STAT pathway is activated by PRL in the mammary gland but not in the liver. Biol. Reprod. 57, 894–900.

Janknecht, R., and Nordheim, A. (1996). MAP Kinase-Dependent Transcriptional Coactivation by Elk-1 and Its Cofactor CBP. Biochem. and Biophys. Res. Commun. 228, 831–837.

Ji, Y., and Morris, M.E. (2005). Membrane transport of dietary phenethyl isothiocyanate by ABCG2 (breast cancer resistance protein). Mol. Pharm. 2, 414–419.

Jiao, X., Zhao, L., Ma, M., Bai, X., He, M., Yan, Y., Wang, Y., Chen, Q., Zhao, X., Zhou, M., et al. (2013). MiR-181a enhances drug sensitivity in mitoxantone-resistant breast cancer cells by targeting breast cancer resistance protein (BCRP/ABCG2). Breast Cancer Res. Treat. 139, 717–730.

176 Jigorel, E., Le Vee, M., Boursier-Neyret, C., Parmentier, Y., and Fardel, O. (2006). Differential regulation of sinusoidal and canalicular hepatic drug transporter expression by xenobiotics activating drug-sensing receptors in primary human hepatocytes. Drug Metab. Dispos. 34, 1756–1763.

Johnson, M.L., Levy, J., Supowit, S.C., Yu-Lee, L.Y., and Rosen, J.M. (1983). Tissue- and cell-specific casein gene expression. II. Relationship to site-specific DNA methylation. J. Biol. Chem. 258, 10805–10811.

Jolivet, G., Pantano, T., and Houdebine, L.M. (2005). Regulation by the extracellular matrix (ECM) of prolactin-induced alpha s1-casein gene expression in rabbit primary mammary cells: role of STAT5, C/EBP, and chromatin structure. J. Cell. Biochem. 95, 313–327.

Jones, P.A. (2012). Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat. Rev. Genet. 13, 484–492.

Jonker, J.W., Smit, J.W., Brinkhuis, R.F., Maliepaard, M., Beijnen, J.H., Schellens, J.H., and Schinkel, A.H. (2000). Role of breast cancer resistance protein in the bioavailability and fetal penetration of topotecan. J. Natl. Cancer Inst. 92, 1651–1656.

Jonker, J.W., Buitelaar, M., Wagenaar, E., Van Der Valk, M.A., Scheffer, G.L., Scheper, R.J., Plosch, T., Kuipers, F., Elferink, R.P.J.O., Rosing, H., et al. (2002). The breast cancer resistance protein protects against a major chlorophyll-derived dietary phototoxin and protoporphyria. Proc. Natl. Acad. Sci. U.S.A. 99, 15649–15654.

Jonker, J.W., Merino, G., Musters, S., van Herwaarden, A.E., Bolscher, E., Wagenaar, E., Mesman, E., Dale, T.C., and Schinkel, A.H. (2005). The breast cancer resistance protein BCRP (ABCG2) concentrates drugs and carcinogenic xenotoxins into milk. Nat. Med. 11, 127–129.

Juliano, R.L., and Ling, V. (1976). A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants. Biochim. Biophys. Acta. 455, 152–162.

Kabotyanski, E.B., Huetter, M., Xian, W., Rijnkels, M., and Rosen, J.M. (2006). Integration of prolactin and glucocorticoid signaling at the beta-casein promoter and enhancer by ordered recruitment of specific transcription factors and chromatin modifiers. Mol. Endocrinol. 20, 2355–2368.

Kalalinia, F., Elahian, F., and Behravan, J. (2011). Potential role of cyclooxygenase-2 on the regulation of the drug efflux transporter ABCG2 in breast cancer cell lines. J. Cancer Res. Clin. Oncol. 137, 321–330.

Kalalinia, F., Elahian, F., Hassani, M., Kasaeeian, J., and Behravan, J. (2012). Phorbol ester TPA modulates chemoresistance in the drug sensitive breast cancer cell line MCF-7 by inducing expression of drug efflux transporter ABCG2. Asian Pac. J. Cancer Prev. 13, 2979– 2984.

Kalgutkar, A.S., Feng, B., Nguyen, H.T., Frederick, K.S., Campbell, S.D., Hatch, H.L., Bi, Y.- A., Kazolias, D.C., Davidson, R.E., Mireles, R.J., et al. (2007). Role of transporters in the disposition of the selective phosphodiesterase-4 inhibitor (+)-2-[4-({[2-(benzo[1,3]dioxol-5-

177 yloxy)-pyridine-3-carbonyl]-amino}-methyl)-3-fluoro-phenoxy]-propionic acid in rat and human. Drug Metab. Dispos. 35, 2111–2118.

Kang, K., Yamaji, D., Yoo, K.H., Robinson, G.W., and Hennighausen, L. (2014). Mammary- specific gene activation is defined by progressive recruitment of STAT5 during pregnancy and the establishment of H3K4me3 marks. Mol. Cell. Biol. 34, 464–473.

Karolchik, D., Barber, G.P., Casper, J., Clawson, H., Cline, M.S., Diekhans, M., Dreszer, T.R., Fujita, P.A., Guruvadoo, L., Haeussler, M., et al. (2014). The UCSC Genome Browser database: 2014 update. Nucleic Acids Res. 42, D764–770.

Kartner, N., Riordan, J.R., and Ling, V. (1983). Cell surface P-glycoprotein associated with multidrug resistance in mammalian cell lines. Science 221, 1285–1288.

Kibbe, W.A. (2007). OligoCalc: an online oligonucleotide properties calculator. Nucleic Acids Res. 35, W43–W46.

Kim, J.E., Singh, R.R., Cho-Vega, J.H., Drakos, E., Davuluri, Y., Khokhar, F.A., Fayad, L., Medeiros, L.J., and Vega, F. (2009a). Sonic hedgehog signaling proteins and ATP-binding cassette G2 are aberrantly expressed in diffuse large B-Cell lymphoma. Mod. Pathol. 22, 1312– 1320.

Kim, J.Y., Mizoguchi, Y., Yamaguchi, H., Enami, J., and Sakai, S. (1997). Removal of milk by suckling acutely increases the prolactin receptor gene expression in the lactating mouse mammary gland. Mol. Cell. Endocrinol. 131, 31–38.

Kim, Y.H., Ishii, G., Goto, K., Ota, S., Kubota, K., Murata, Y., Mishima, M., Saijo, N., Nishiwaki, Y., and Ochiai, A. (2009b). Expression of breast cancer resistance protein is associated with a poor clinical outcome in patients with small-cell lung cancer. Lung Cancer 65, 105–111.

Kitamura, S., Maeda, K., Wang, Y., and Sugiyama, Y. (2008). Involvement of multiple transporters in the hepatobiliary transport of rosuvastatin. Drug Metab. Dispos. 36, 2014–2023.

Kleinberg, D.L., and Todd, J. (1980). Evidence that human growth hormone is a potent lactogen in primates. J. Clin. Endocrinol. Metab. 51, 1009–1013.

Kline, J.B., and Clevenger, C.V. (2001). Identification and characterization of the prolactin- binding protein in human serum and milk. J. Biol. Chem. 276, 24760–24766.

Kline, J.B., Roehrs, H., and Clevenger, C.V. (1999). Functional characterization of the intermediate isoform of the human prolactin receptor. J. Biol. Chem. 274, 35461–35468.

Kline, J.B., Moore, D.J., and Clevenger, C.V. (2001). Activation and association of the Tec tyrosine kinase with the human prolactin receptor: mapping of a Tec/Vav1-receptor binding site. Mol. Endocrinol. 15, 832–841.

Kline, J.B., Rycyzyn, M.A., and Clevenger, C.V. (2002). Characterization of a novel and functional human prolactin receptor isoform (deltaS1PRLr) containing only one extracellular fibronectin-like domain. Mol. Endocrinol. 16, 2310–2322.

178 Knight, C.H., Calvert, D.T., and Flint, D.J. (1986). Inhibitory effects of bromocriptine on mammary development and function in lactating mice. J. Endocrinol. 110, 263–270.

Kobayashi, D., Ieiri, I., Hirota, T., Takane, H., Maegawa, S., Kigawa, J., Suzuki, H., Nanba, E., Oshimura, M., Terakawa, N., et al. (2005). Functional assessment of ABCG2 (BCRP) gene polymorphisms to protein expression in human placenta. Drug Metab. Dispos. 33, 94–101.

Kodaira, H., Kusuhara, H., Ushiki, J., Fuse, E., and Sugiyama, Y. (2010). Kinetic analysis of the cooperation of P-glycoprotein (P-gp/Abcb1) and breast cancer resistance protein (Bcrp/Abcg2) in limiting the brain and testis penetration of erlotinib, flavopiridol, and mitoxantrone. J. Pharmacol. Exp. Ther. 333, 788–796.

Kouzarides, T. (2007). Chromatin Modifications and Their Function. Cell 128, 693–705.

Kress, S., and Greenlee, W.F. (1997). Cell-specific regulation of human CYP1A1 and CYP1B1 genes. Cancer Res. 57, 1264–1269.

Krishnamurthy, P., Ross, D.D., Nakanishi, T., Bailey-Dell, K., Zhou, S., Mercer, K.E., Sarkadi, B., Sorrentino, B.P., and Schuetz, J.D. (2004). The stem cell marker Bcrp/ABCG2 enhances hypoxic cell survival through interactions with heme. J. Biol. Chem. 279, 24218–24225.

Kruijtzer, C.M.F., Beijnen, J.H., and Schellens, J.H.M. (2002). Improvement of oral drug treatment by temporary inhibition of drug transporters and/or cytochrome P450 in the gastrointestinal tract and liver: an overview. Oncologist 7, 516–530.

Lagas, J.S., van Waterschoot, R.A.B., van Tilburg, V.A.C.J., Hillebrand, M.J., Lankheet, N., Rosing, H., Beijnen, J.H., and Schinkel, A.H. (2009). Brain accumulation of dasatinib is restricted by P-glycoprotein (ABCB1) and breast cancer resistance protein (ABCG2) and can be enhanced by elacridar treatment. Clin. Cancer Res. 15, 2344–2351.

Lagas, J.S., van Waterschoot, R.A.B., Sparidans, R.W., Wagenaar, E., Beijnen, J.H., and Schinkel, A.H. (2010). Breast cancer resistance protein and P-glycoprotein limit sorafenib brain accumulation. Mol. Cancer Ther. 9, 319–326.

Langmead, B., Trapnell, C., Pop, M., and Salzberg, S.L. (2009). Ultrafast and memory- efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25.

LaPensee, E.W., Schwemberger, S.J., LaPensee, C.R., Bahassi, E.M., Afton, S.E., and Ben- Jonathan, N. (2009). Prolactin confers resistance against cisplatin in breast cancer cells by activating glutathione-S-transferase. Carcinogenesis 30, 1298–1304.

Le Vee, M., Gripon, P., Stieger, B., and Fardel, O. (2008). Down-regulation of organic anion transporter expression in human hepatocytes exposed to the proinflammatory cytokine interleukin 1beta. Drug Metab. Dispos. 36, 217–222.

Le Vee, M., Lecureur, V., Stieger, B., and Fardel, O. (2009). Regulation of drug transporter expression in human hepatocytes exposed to the proinflammatory cytokines tumor necrosis factor-alpha or interleukin-6. Drug Metab. Dispos. 37, 685–693.

179 Le Vee, M., Jouan, E., Stieger, B., Lecureur, V., and Fardel, O. (2011). Regulation of drug transporter expression by oncostatin M in human hepatocytes. Biochem. Pharmacol. 82, 304– 311.

Le Vee, M., Jouan, E., Stieger, B., and Fardel, O. (2013). Differential regulation of drug transporter expression by all-trans retinoic acid in hepatoma HepaRG cells and human hepatocytes. Eur J Pharm Sci 48, 767–774.

Lebrun, J.J., Ali, S., Sofer, L., Ullrich, A., and Kelly, P.A. (1994). Prolactin-induced proliferation of Nb2 cells involves tyrosine phosphorylation of the prolactin receptor and its associated tyrosine kinase JAK2. J. Biol. Chem. 269, 14021–14026.

Lebrun, J.J., Ali, S., Ullrich, A., and Kelly, P.A. (1995a). Proline-rich sequence-mediated Jak2 association to the prolactin receptor is required but not sufficient for signal transduction. J. Biol. Chem. 270, 10664–10670.

Lebrun, J.J., Ali, S., Goffin, V., Ullrich, A., and Kelly, P.A. (1995b). A single phosphotyrosine residue of the prolactin receptor is responsible for activation of gene transcription. Proc. Natl. Acad. Sci. U.S.A. 92, 4031–4035.

Lechner, A., Leech, C.A., Abraham, E.J., Nolan, A.L., and Habener, J.F. (2002). Nestin- positive progenitor cells derived from adult human pancreatic islets of Langerhans contain side population (SP) cells defined by expression of the ABCG2 (BCRP1) ATP-binding cassette transporter. Biochem. Biophys. Res. Commun. 293, 670–674.

Lee, H.J., Daver, N., Kantarjian, H.M., Verstovsek, S., and Ravandi, F. (2013). The Role of JAK Pathway Dysregulation in the Pathogenesis and Treatment of Acute Myeloid Leukemia. Clin. Cancer Res. 19, 327–335.

Lee, J.K., Abe, K., Bridges, A.S., Patel, N.J., Raub, T.J., Pollack, G.M., and Brouwer, K.L.R. (2009). Sex-dependent disposition of acetaminophen sulfate and glucuronide in the in situ perfused mouse liver. Drug Metab. Dispos. 37, 1916–1921.

Lee, S.H., Kim, H., Hwang, J.-H., Lee, H.S., Cho, J.Y., Yoon, Y.-S., and Han, H.-S. (2012). Breast cancer resistance protein expression is associated with early recurrence and decreased survival in resectable pancreatic cancer patients. Pathol. Int. 62, 167–175.

Lee, Y.-J., Kusuhara, H., Jonker, J.W., Schinkel, A.H., and Sugiyama, Y. (2005). Investigation of efflux transport of dehydroepiandrosterone sulfate and mitoxantrone at the mouse blood- brain barrier: a minor role of breast cancer resistance protein. J. Pharmacol. Exp. Ther. 312, 44–52.

Lefebvre, P., Benomar, Y., and Staels, B. (2010). Retinoid X receptors: common heterodimerization partners with distinct functions. Trends Endocrinol. Metab. 21, 676–683.

Lemay, D.G., Pollard, K.S., Martin, W.F., Freeman Zadrowski, C., Hernandez, J., Korf, I., German, J.B., and Rijnkels, M. (2013). From genes to milk: genomic organization and epigenetic regulation of the mammary transcriptome. PLoS ONE 8, e75030.

180 Lemmen, J., Tozakidis, I.E.P., and Galla, H.-J. (2013a). Pregnane X receptor upregulates ABC-transporter Abcg2 and Abcb1 at the blood-brain barrier. Brain Res. 1491, 1–13.

Lemmen, J., Tozakidis, I.E.P., Bele, P., and Galla, H.-J. (2013b). Constitutive androstane receptor upregulates Abcb1 and Abcg2 at the blood-brain barrier after CITCO activation. Brain Res. 1501, 68–80.

Lemos, C., Kathmann, I., Giovannetti, E., Dekker, H., Scheffer, G.L., Calhau, C., Jansen, G., and Peters, G.J. (2008). Folate deprivation induces BCRP (ABCG2) expression and mitoxantrone resistance in Caco-2 cells. Int. J. Cancer 123, 1712–1720.

Lemos, C., Kathmann, I., Giovannetti, E., Beliën, J.A.M., Scheffer, G.L., Calhau, C., Jansen, G., and Peters, G.J. (2009). Cellular folate status modulates the expression of BCRP and MRP multidrug transporters in cancer cell lines from different origins. Mol. Cancer Ther. 8, 655–664.

Lemus-Wilson, A., Kelly, P.A., and Blask, D.E. (1995). Melatonin blocks the stimulatory effects of prolactin on human breast cancer cell growth in culture. Br. J. Cancer 72, 1435–1440.

Li, S., and Rosen, J.M. (1995). Nuclear factor I and mammary gland factor (STAT5) play a critical role in regulating rat whey acidic protein gene expression in transgenic mice. Mol. Cell. Biol. 15, 2063–2070.

Li, B., Carey, M., and Workman, J.L. (2007). The role of chromatin during transcription. Cell 128, 707–719.

Li, H., Jin, H.-E., Kim, W., Han, Y.-H., Kim, D.-D., Chung, S.-J., and Shim, C.-K. (2008). Involvement of P-glycoprotein, multidrug resistance protein 2 and breast cancer resistance protein in the transport of belotecan and topotecan in Caco-2 and MDCKII cells. Pharm. Res. 25, 2601–2612.

Li, J., Li, Z.-N., Du, Y.-J., Li, X.-Q., Bao, Q.-L., and Chen, P. (2009). Expression of MRP1, BCRP, LRP, and ERCC1 in advanced non-small-cell lung cancer: correlation with response to chemotherapy and survival. Clin. Lung Cancer 10, 414–421.

Li, J., Li, Z.-N., Yu, L.-C., Bao, Q.-L., Wu, J.-R., Shi, S.-B., and Li, X.-Q. (2010). Association of expression of MRP1, BCRP, LRP and ERCC1 with outcome of patients with locally advanced non-small cell lung cancer who received neoadjuvant chemotherapy. Lung Cancer 69, 116–122.

Li, M., Liu, X., Robinson, G., Bar-Peled, U., Wagner, K.U., Young, W.S., Hennighausen, L., and Furth, P.A. (1997). Mammary-derived signals activate programmed cell death during the first stage of mammary gland involution. Proc. Natl. Acad. Sci. U.S.A. 94, 3425–3430.

Li, Y., Clevenger, C.V., Minkovsky, N., Kumar, K.G.S., Raghunath, P.N., Tomaszewski, J.E., Spiegelman, V.S., and Fuchs, S.Y. (2006). Stabilization of prolactin receptor in breast cancer cells. Oncogene 25, 1896–1902.

181 Liao, R., Sun, J., Zhang, L., Lou, G., Chen, M., Zhou, D., Chen, Z., and Zhang, S. (2008). MicroRNAs play a role in the development of human hematopoietic stem cells. J. Cell. Biochem. 104, 805–817.

Litman, T., Brangi, M., Hudson, E., Fetsch, P., Abati, A., Ross, D.D., Miyake, K., Resau, J.H., and Bates, S.E. (2000). The multidrug-resistant phenotype associated with overexpression of the new ABC half-transporter, MXR (ABCG2). J. Cell. Sci. 113, 2011–2021.

Litterst, C.M., Kliem, S., Marilley, D., and Pfitzner, E. (2003). NCoA-1/SRC-1 Is an Essential Coactivator of STAT5 That Binds to the FDL Motif in the α-Helical Region of the STAT5 Transactivation Domain. J. Biol. Chem. 278, 45340–45351.

Liu, T., Ortiz, J.A., Taing, L., Meyer, C.A., Lee, B., Zhang, Y., Shin, H., Wong, S.S., Ma, J., Lei, Y., et al. (2011). Cistrome: an integrative platform for transcriptional regulation studies. Genome Biol. 12, R83.

Liu, W.-H., Liu, H.-B., Gao, D.-K., Ge, G.-Q., Zhang, P., Sun, S.-R., Wang, H.-M., and Liu, S.-B. (2013). ABCG2 protects kidney side population cells from hypoxia/reoxygenation injury through activation of the MEK/ERK pathway. Cell Transplant 22, 1859–1868.

Liu, X., Robinson, G.W., Gouilleux, F., Groner, B., and 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, 8831–8835.

Liu, X., Robinson, G.W., and Hennighausen, L. (1996). Activation of Stat5a and Stat5b by tyrosine phosphorylation is tightly linked to mammary gland differentiation. Mol. Endocrinol. 10, 1496–1506.

Liu, X., Robinson, G.W., Wagner, K.U., Garrett, L., Wynshaw-Boris, A., and Hennighausen, L. (1997). Stat5a is mandatory for adult mammary gland development and lactogenesis. Genes Dev. 11, 179–186.

Livak, K.J., and Schmittgen, T.D. (2001). Analysis of relative gene expression data using real- time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25, 402–408.

Long, W., Wagner, K.-U., Lloyd, K.C.K., Binart, N., Shillingford, J.M., Hennighausen, L., and Jones, F.E. (2003). Impaired differentiation and lactational failure of Erbb4-deficient mammary glands identify ERBB4 as an obligate mediator of STAT5. Development 130, 5257– 5268.

Lydon, J.P., DeMayo, F.J., Funk, C.R., Mani, S.K., Hughes, A.R., Montgomery, C.A., Shyamala, G., Conneely, O.M., and O’Malley, B.W. (1995). Mice lacking progesterone receptor exhibit pleiotropic reproductive abnormalities. Genes Dev. 9, 2266–2278.

Ma, M.-T., He, M., Wang, Y., Jiao, X.-Y., Zhao, L., Bai, X.-F., Yu, Z.-J., Wu, H.-Z., Sun, M.- L., Song, Z.-G., et al. (2013). MiR-487a resensitizes mitoxantrone (MX)-resistant breast cancer cells (MCF-7/MX) to MX by targeting breast cancer resistance protein (BCRP/ABCG2). Cancer Lett. 339, 107–115.

182 Ma, Q., Baldwin, K.T., Renzelli, A.J., McDaniel, A., and Dong, L. (2001). TCDD-inducible poly(ADP-ribose) polymerase: a novel response to 2,3,7,8-tetrachlorodibenzo-p-dioxin. Biochem. Biophys. Res. Commun. 289, 499–506.

MacLean, C., Moenning, U., Reichel, A., and Fricker, G. (2008). Closing the gaps: a full scan of the intestinal expression of p-glycoprotein, breast cancer resistance protein, and multidrug resistance-associated protein 2 in male and female rats. Drug Metab. Dispos. 36, 1249–1254.

Malekshah, O.M., Lage, H., Bahrami, A.R., Afshari, J.T., and Behravan, J. (2012). PXR and NF-κB correlate with the inducing effects of IL-1β and TNF-α on ABCG2 expression in breast cancer cell lines. Eur. J. Pharm. Sci. 47, 474–480.

Mali, P., Yang, L., Esvelt, K.M., Aach, J., Guell, M., DiCarlo, J.E., Norville, J.E., and Church, G.M. (2013). RNA-guided human genome engineering via Cas9. Science 339, 823–826.

Maliepaard, M., Scheffer, G.L., Faneyte, I.F., van Gastelen, M.A., Pijnenborg, A.C., Schinkel, A.H., van De Vijver, M.J., Scheper, R.J., and Schellens, J.H. (2001). Subcellular localization and distribution of the breast cancer resistance protein transporter in normal human tissues. Cancer Res. 61, 3458–3464.

Mallepell, S., Krust, A., Chambon, P., and Brisken, C. (2006). Paracrine signaling through the epithelial estrogen receptor alpha is required for proliferation and morphogenesis in the mammary gland. Proc. Natl. Acad. Sci. U.S.A. 103, 2196–2201.

Manhès, C., Kayser, C., Bertheau, P., Kelder, B., Kopchick, J.J., Kelly, P.A., Touraine, P., and Goffin, V. (2006). Local over-expression of prolactin in differentiating mouse mammary gland induces functional defects and benign lesions, but no carcinoma. J. Endocrinol. 190, 271–285.

Maningat, P.D., Sen, P., Rijnkels, M., Sunehag, A.L., Hadsell, D.L., Bray, M., and Haymond, M.W. (2009). Gene expression in the human mammary epithelium during lactation: the milk fat globule transcriptome. Physiol. Genomics 37, 12–22.

Marchetti, S., Oostendorp, R.L., Pluim, D., van Eijndhoven, M., van Tellingen, O., Schinkel, A.H., Versace, R., Beijnen, J.H., Mazzanti, R., and Schellens, J.H. (2007). In vitro transport of gimatecan (7-t-butoxyiminomethylcamptothecin) by breast cancer resistance protein, P- glycoprotein, and multidrug resistance protein 2. Mol. Cancer Ther. 6, 3307–3313.

Markoff, E., and Talamantes, F. (1980). The lactogenic response of mouse mammary explants to mose prolactin and growth hormone. Endocr. Res. Commun. 7, 269–278.

Martin, C.M., Meeson, A.P., Robertson, S.M., Hawke, T.J., Richardson, J.A., Bates, S., Goetsch, S.C., Gallardo, T.D., and Garry, D.J. (2004). Persistent expression of the ATP- binding cassette transporter, Abcg2, identifies cardiac SP cells in the developing and adult heart. Dev. Biol. 265, 262–275.

Mathias, A.A., Hitti, J., and Unadkat, J.D. (2005). P-glycoprotein and breast cancer resistance protein expression in human placentae of various gestational ages. Am. J. Physiol. Regul. Integr. Comp. Physiol. 289, R963–969.

183 McDevitt, C.A., Collins, R., Kerr, I.D., and Callaghan, R. (2009). Purification and structural analyses of ABCG2. Adv. Drug Deliv. Rev. 61, 57–65.

McHale, K., Tomaszewski, J.E., Puthiyaveettil, R., Livolsi, V.A., and Clevenger, C.V. (2008). Altered expression of prolactin receptor-associated signaling proteins in human breast carcinoma. Mod. Pathol. 21, 565–571.

McNally, S., and Martin, F. (2011). Molecular regulators of pubertal mammary gland development. Ann. Med. 43, 212–234.

Meng, J., Tsai-Morris, C.-H., and Dufau, M.L. (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, 5677–5682.

Menter, D.G., Schilsky, R.L., and DuBois, R.N. (2010). Cyclooxygenase-2 and Cancer Treatment: Understanding the Risk Should Be Worth the Reward. Clin. Cancer Res. 16, 1384– 1390.

Merino, G., Jonker, J.W., Wagenaar, E., van Herwaarden, A.E., and Schinkel, A.H. (2005a). The breast cancer resistance protein (BCRP/ABCG2) affects pharmacokinetics, hepatobiliary excretion, and milk secretion of the antibiotic nitrofurantoin. Mol. Pharmacol. 67, 1758–1764.

Merino, G., van Herwaarden, A.E., Wagenaar, E., Jonker, J.W., and Schinkel, A.H. (2005b). Sex-dependent expression and activity of the ATP-binding cassette transporter breast cancer resistance protein (BCRP/ABCG2) in liver. Mol. Pharmacol. 67, 1765–1771.

Merino, G., Alvarez, A.I., Pulido, M.M., Molina, A.J., Schinkel, A.H., and Prieto, J.G. (2006). Breast cancer resistance protein (BCRP/ABCG2) transports fluoroquinolone antibiotics and affects their oral availability, pharmacokinetics, and milk secretion. Drug Metab. Dispos. 34, 690–695.

Metón, I., Boot, E.P.J., Sussenbach, J.S., and Steenbergh, P.H. (1999). Growth hormone induces insulin-like growth factor-I gene transcription by a synergistic action of STAT5 and HNF-1α. FEBS Letters 444, 155–159.

Meydan, N., Grunberger, T., Dadi, H., Shahar, M., Arpaia, E., Lapidot, Z., Leeder, J.S., Freedman, M., Cohen, A., Gazit, A., et al. (1996). Inhibition of acute lymphoblastic leukaemia by a Jak-2 inhibitor. Nature 379, 645–648.

Meyer zu Schwabedissen, H.E., Grube, M., Dreisbach, A., Jedlitschky, G., Meissner, K., Linnemann, K., Fusch, C., Ritter, C.A., Völker, U., and Kroemer, H.K. (2006). Epidermal growth factor-mediated activation of the map kinase cascade results in altered expression and function of ABCG2 (BCRP). Drug Metab. Dispos. 34, 524–533.

Miller, S.L., DeMaria, J.E., Freier, D.O., Riegel, A.M., and Clevenger, C.V. (2005). Novel association of Vav2 and Nek3 modulates signaling through the human prolactin receptor. Mol. Endocrinol. 19, 939–949.

184 Milsom, S.R., Breier, B.H., Gallaher, B.W., Cox, V.A., Gunn, A.J., and Gluckman, P.D. (1992). Growth hormone stimulates galactopoiesis in healthy lactating women. Acta Endocrinol. 127, 337–343.

Mirski, S.E.L., Gerlach, J.H., and Cole, S.P.C. (1987). Multidrug Resistance in a Human Small Cell Lung Cancer Cell Line Selected in Adriamycin. Cancer Res. 47, 2594–2598.

Mittapalli, R.K., Vaidhyanathan, S., Sane, R., and Elmquist, W.F. (2012). Impact of P- glycoprotein (ABCB1) and breast cancer resistance protein (ABCG2) on the brain distribution of a novel BRAF inhibitor: vemurafenib (PLX4032). J. Pharmacol. Exp. Ther. 342, 33–40.

Mittapalli, R.K., Vaidhyanathan, S., Dudek, A.Z., and Elmquist, W.F. (2013). Mechanisms limiting distribution of the threonine-protein kinase B-RaF(V600E) inhibitor dabrafenib to the brain: implications for the treatment of melanoma brain metastases. J. Pharmacol. Exp. Ther. 344, 655–664.

Miyake, K., Mickley, L., Litman, T., Zhan, Z., Robey, R., Cristensen, B., Brangi, M., Greenberger, L., Dean, M., Fojo, T., et al. (1999). Molecular cloning of cDNAs which are highly overexpressed in mitoxantrone-resistant cells: demonstration of homology to ABC transport genes. Cancer Res. 59, 8–13.

Mizoguchi, Y., Kim, J.Y., Sasaki, T., Hama, T., Sasaki, M., Enami, J., and Sakai, S. (1996). Acute expression of the PRL receptor gene after ovariectomy in midpregnant mouse mammary gland. Endocr. J. 43, 537–544.

Mizuno, N., Suzuki, M., Kusuhara, H., Suzuki, H., Takeuchi, K., Niwa, T., Jonker, J.W., and Sugiyama, Y. (2004). Impaired renal excretion of 6-hydroxy-5,7-dimethyl-2-methylamino-4- (3-pyridylmethyl) benzothiazole (E3040) sulfate in breast cancer resistance protein (BCRP1/ABCG2) knockout mice. Drug Metab. Dispos. 32, 898–901.

Mizuno, N., Takahashi, T., Kusuhara, H., Schuetz, J.D., Niwa, T., and Sugiyama, Y. (2007). Evaluation of the role of breast cancer resistance protein (BCRP/ABCG2) and multidrug resistance-associated protein 4 (MRP4/ABCC4) in the urinary excretion of sulfate and glucuronide metabolites of edaravone (MCI-186; 3-methyl-1-phenyl-2-pyrazolin-5-one). Drug Metab. Dispos. 35, 2045–2052.

Moffit, J.S., Aleksunes, L.M., Maher, J.M., Scheffer, G.L., Klaassen, C.D., and Manautou, J.E. (2006). Induction of hepatic transporters multidrug resistance-associated proteins (Mrp) 3 and 4 by clofibrate is regulated by peroxisome proliferator-activated receptor alpha. J. Pharmacol. Exp. Ther. 317, 537–545.

Mogi, M., Yang, J., Lambert, J.-F., Colvin, G.A., Shiojima, I., Skurk, C., Summer, R., Fine, A., Quesenberry, P.J., and Walsh, K. (2003). Akt signaling regulates side population cell phenotype via Bcrp1 translocation. J. Biol. Chem. 278, 39068–39075.

Mosaffa, F., Kalalinia, F., Lage, H., Afshari, J.T., and Behravan, J. (2012). Pro-inflammatory cytokines interleukin-1 beta, interleukin 6, and tumor necrosis factor-alpha alter the expression and function of ABCG2 in cervix and gastric cancer cells. Mol. Cell. Biochem. 363, 385–393.

185 Müller, J., Sperl, B., Reindl, W., Kiessling, A., and Berg, T. (2008). Discovery of chromone- based inhibitors of the transcription factor STAT5. Chembiochem 9, 723–727.

Muraoka-Cook, R.S., Sandahl, M., Hunter, D., Miraglia, L., and Earp, H.S. (2008). Prolactin and ErbB4/HER4 Signaling Interact via Janus Kinase 2 to Induce Mammary Epithelial Cell Gene Expression Differentiation. Mol. Endocrinol. 22, 2307–2321.

Nagano, M., and Kelly, P.A. (1994). Tissue distribution and regulation of rat prolactin receptor gene expression. Quantitative analysis by polymerase chain reaction. J. Biol. Chem. 269, 13337–13345.

Nakamichi, N., Morii, E., Ikeda, J., Qiu, Y., Mamato, S., Tian, T., Fukuhara, S., and Aozasa, K. (2009). Synergistic effect of interleukin-6 and endoplasmic reticulum stress inducers on the high level of ABCG2 expression in plasma cells. Lab. Invest. 89, 327–336.

Nakanishi, T., Bailey-Dell, K.J., Hassel, B.A., Shiozawa, K., Sullivan, D.M., Turner, J., and Ross, D.D. (2006). Novel 5’ untranslated region variants of BCRP mRNA are differentially expressed in drug-selected cancer cells and in normal human tissues: implications for drug resistance, tissue-specific expression, and alternative promoter usage. Cancer Res. 66, 5007– 5011.

Nakano, H., Nakamura, Y., Soda, H., Kamikatahira, M., Uchida, K., Takasu, M., Kitazaki, T., Yamaguchi, H., Nakatomi, K., Yanagihara, K., et al. (2008). Methylation status of breast cancer resistance protein detected by methylation-specific polymerase chain reaction analysis is correlated inversely with its expression in drug-resistant lung cancer cells. Cancer 112, 1122–1130.

Nakatomi, K., Yoshikawa, M., Oka, M., Ikegami, Y., Hayasaka, S., Sano, K., Shiozawa, K., Kawabata, S., Soda, H., Ishikawa, T., et al. (2001). Transport of 7-ethyl-10- hydroxycamptothecin (SN-38) by breast cancer resistance protein ABCG2 in human lung cancer cells. Biochem. Biophys. Res. Commun. 288, 827–832.

Nakayama, A., Matsuo, H., Takada, T., Ichida, K., Nakamura, T., Ikebuchi, Y., Ito, K., Hosoya, T., Kanai, Y., Suzuki, H., et al. (2011). ABCG2 is a high-capacity urate transporter and its genetic impairment increases serum uric acid levels in humans. Nucleosides Nucleotides Nucleic Acids 30, 1091–1097.

Nandi, S. (1958). Endocrine control of mammarygland development and function in the C3H/ He Crgl mouse. J. Natl. Cancer Inst. 21, 1039–1063.

Narang, V.S., Fraga, C., Kumar, N., Shen, J., Throm, S., Stewart, C.F., and Waters, C.M. (2008). Dexamethasone increases expression and activity of multidrug resistance transporters at the rat blood-brain barrier. Am. J. Physiol., Cell Physiol. 295, C440–450.

Naspinski, C., Gu, X., Zhou, G.-D., Mertens-Talcott, S.U., Donnelly, K.C., and Tian, Y. (2008). Pregnane X receptor protects HepG2 cells from BaP-induced DNA damage. Toxicol. Sci. 104, 67–73.

186 Natarajan, K., Xie, Y., Nakanishi, T., Beck, W.T., Bauer, K.S., and Ross, D.D. (2011). Identification and characterization of the major alternative promoter regulating Bcrp1/Abcg2 expression in the mouse intestine. Biochim. Biophys. Acta 1809, 295–305.

Neilson, L.M., Zhu, J., Xie, J., Malabarba, M.G., Sakamoto, K., Wagner, K.-U., Kirken, R.A., and Rui, H. (2007). Coactivation of janus tyrosine kinase (Jak)1 positively modulates prolactin-Jak2 signaling in breast cancer: recruitment of ERK and signal transducer and activator of transcription (Stat)3 and enhancement of Akt and Stat5a/b pathways. Mol. Endocrinol. 21, 2218–2232.

Nelson, E.A., Walker, S.R., Li, W., Liu, X.S., and Frank, D.A. (2006). Identification of Human STAT5-dependent Gene Regulatory Elements Based on Interspecies Homology. J. Biol. Chem. 281, 26216–26224.

Neville, M.C., and Morton, J. (2001). Physiology and endocrine changes underlying human lactogenesis II. J. Nutr. 131, 3005S–8S.

Neville, M.C., and Walsh, C.T. (1995). Effects of xenobiotics on milk secretion and composition. Am. J. Clin. Nutr. 61, 687S–694S.

Neville, M.C., McFadden, T.B., and Forsyth, I. (2002). Hormonal regulation of mammary differentiation and milk secretion. J. Mammary Gland Biol. Neoplasia 7, 49–66.

Nguyen, D.A., Parlow, A.F., and Neville, M.C. (2001). Hormonal regulation of tight junction closure in the mouse mammary epithelium during the transition from pregnancy to lactation. J. Endocrinol. 170, 347–356.

Nishikawa, S., Moore, R.C., Nonomura, N., and Oka, T. (1994). Progesterone and EGF inhibit mouse mammary gland prolactin receptor and beta-casein gene expression. Am. J. Physiol. 267, C1467–1472.

Niu, G., Wright, K.L., Ma, Y., Wright, G.M., Huang, M., Irby, R., Briggs, J., Karras, J., Cress, W.D., Pardoll, D., et al. (2005). Role of Stat3 in regulating p53 expression and function. Mol. Cell. Biol. 25, 7432–7440.

Noel, G.L., Suh, H.K., and Frantz, A.G. (1974). Prolactin release during nursing and breast stimulation in postpartum and nonpostpartum subjects. J. Clin. Endocrinol. Metab. 38, 413– 423.

Olazabal, I., Muñoz, J., Ogueta, S., Obregón, E., and García-Ruiz, J.P. (2000). Prolactin (PRL)-PRL receptor system increases cell proliferation involving JNK (c-Jun amino terminal kinase) and AP-1 activation: inhibition by glucocorticoids. Mol. Endocrinol. 14, 564–575.

Ooi, G.T., Hurst, K.R., Poy, M.N., Rechler, M.M., and Boisclair, Y.R. (1998). Binding of STAT5a and STAT5b to a Single Element Resembling a γ-Interferon-Activated Sequence Mediates the Growth Hormone Induction of the Mouse Acid-Labile Subunit Promoter in Liver Cells. Mol. Endocrinol. 12, 675–687.

187 Ormandy, C.J., Hall, R.E., Manning, D.L., Robertson, J.F., Blamey, R.W., Kelly, P.A., Nicholson, R.I., and Sutherland, R.L. (1997). Coexpression and cross-regulation of the prolactin receptor and sex steroid hormone receptors in breast cancer. J. Clin. Endocrinol. Metab. 82, 3692–3699.

Ota, S., Ishii, G., Goto, K., Kubota, K., Kim, Y.H., Kojika, M., Murata, Y., Yamazaki, M., Nishiwaki, Y., Eguchi, K., et al. (2009). Immunohistochemical expression of BCRP and ERCC1 in biopsy specimen predicts survival in advanced non-small-cell lung cancer treated with cisplatin-based chemotherapy. Lung Cancer 64, 98–104.

Pan, Y.-Z., Morris, M.E., and Yu, A.-M. (2009). MicroRNA-328 negatively regulates the expression of breast cancer resistance protein (BCRP/ABCG2) in human cancer cells. Mol. Pharmacol. 75, 1374–1379.

Pang, W.W., and Hartmann, P.E. (2007). Initiation of human lactation: secretory differentiation and secretory activation. J. Mammary Gland Biol. Neoplasia 12, 211–221.

Pavek, P., Merino, G., Wagenaar, E., Bolscher, E., Novotna, M., Jonker, J.W., and Schinkel, A.H. (2005). Human Breast Cancer Resistance Protein: Interactions with Steroid Drugs, Hormones, the Dietary Carcinogen 2-Amino-1-methyl-6-phenylimidazo(4,5-b)pyridine, and Transport of Cimetidine. J. Pharmacol. Exp. Ther. 312, 144–152.

Peirce, S.K., Chen, W.Y., and Chen, W.Y. (2001). Quantification of prolactin receptor mRNA in multiple human tissues and cancer cell lines by real time RT-PCR. J. Endocrinol. 171, R1–4.

Peng, H., Qi, J., Dong, Z., and Zhang, J.-T. (2010). Dynamic vs Static ABCG2 Inhibitors to Sensitize Drug Resistant Cancer Cells. PLoS ONE 5, e15276.

Perez, M., Blazquez, A.G., Real, R., Mendoza, G., Prieto, J.G., Merino, G., and Alvarez, A.I. (2009). In vitro and in vivo interaction of moxidectin with BCRP/ABCG2. Chem. Biol. Interact. 180, 106–112.

Petropoulos, S., Gibb, W., and Matthews, S.G. (2011a). Breast cancer-resistance protein (BCRP1) in the fetal mouse brain: development and glucocorticoid regulation. Biol. Reprod. 84, 783–789.

Petropoulos, S., Gibb, W., and Matthews, S.G. (2011b). Glucocorticoid regulation of placental breast cancer resistance protein (Bcrp1) in the mouse. Reprod.Sci. 18, 631–639.

Petrovic, V., Wang, J.-H., and Piquette-Miller, M. (2008). Effect of endotoxin on the expression of placental drug transporters and glyburide disposition in pregnant rats. Drug Metab. Dispos. 36, 1944–1950.

Pezet, A., Ferrag, F., Kelly, P.A., and Edery, M. (1997). Tyrosine docking sites of the rat prolactin receptor required for association and activation of stat5. J. Biol. Chem. 272, 25043– 25050.

188 Pfister, O., Oikonomopoulos, A., Sereti, K.-I., Sohn, R.L., Cullen, D., Fine, G.C., Mouquet, F., Westerman, K., and Liao, R. (2008). Role of the ATP-binding cassette transporter Abcg2 in the phenotype and function of cardiac side population cells. Circ. Res. 103, 825–835.

Pfitzner, E., Jähne, R., Wissler, M., Stoecklin, E., and Groner, B. (1998). p300/CREB-binding protein enhances the prolactin-mediated transcriptional induction through direct interaction with the transactivation domain of Stat5, but does not participate in the Stat5-mediated suppression of the glucocorticoid response. Mol. Endocrinol. 12, 1582–1593.

Poller, B., Drewe, J., Krähenbühl, S., Huwyler, J., and Gutmann, H. (2010). Regulation of BCRP (ABCG2) and P-glycoprotein (ABCB1) by cytokines in a model of the human blood- brain barrier. Cell. Mol. Neurobiol. 30, 63–70.

Pollex, E.K., Anger, G., Hutson, J., Koren, G., and Piquette-Miller, M. (2010). Breast cancer resistance protein (BCRP)-mediated glyburide transport: effect of the C421A/Q141K BCRP single-nucleotide polymorphism. Drug Metab. Dispos. 38, 740–744.

Polli, J.W., Humphreys, J.E., Harmon, K.A., Castellino, S., O’Mara, M.J., Olson, K.L., John- Williams, L.S., Koch, K.M., and Serabjit-Singh, C.J. (2008). The role of efflux and uptake transporters in [N-{3-chloro-4-[(3-fluorobenzyl)oxy]phenyl}-6-[5-({[2- (methylsulfonyl)ethyl]amino}methyl)-2-furyl]-4-quinazolinamine (GW572016, lapatinib) disposition and drug interactions. Drug Metab. Dispos. 36, 695–701.

Polli, J.W., Olson, K.L., Chism, J.P., John-Williams, L.S., Yeager, R.L., Woodard, S.M., Otto, V., Castellino, S., and Demby, V.E. (2009). An unexpected synergist role of P-glycoprotein and breast cancer resistance protein on the central nervous system penetration of the tyrosine kinase inhibitor lapatinib (N-{3-chloro-4-[(3-fluorobenzyl)oxy]phenyl}-6-[5-({[2- (methylsulfonyl)ethyl]amino}methyl)-2-furyl]-4-quinazolinamine; GW572016). Drug Metab. Dispos. 37, 439–442.

Poonkuzhali, B., Lamba, J., Strom, S., Sparreboom, A., Thummel, K., Watkins, P., and Schuetz, E. (2008). Association of Breast Cancer Resistance Protein/ABCG2 Phenotypes and Novel Promoter and Intron 1 Single Nucleotide Polymorphisms. Drug Metab. Dispos. 36, 780– 795.

Porro, A., Iraci, N., Soverini, S., Diolaiti, D., Gherardi, S., Terragna, C., Durante, S., Valli, E., Kalebic, T., Bernardoni, R., et al. (2011). c-MYC oncoprotein dictates transcriptional profiles of ATP-binding cassette transporter genes in chronic myelogenous leukemia CD34+ hematopoietic progenitor cells. Mol. Cancer Res. 9, 1054–1066.

Pradhan, M., Bembinster, L.A., Baumgarten, S.C., and Frasor, J. (2010). Proinflammatory cytokines enhance estrogen-dependent expression of the multidrug transporter gene ABCG2 through estrogen receptor and NF{kappa}B cooperativity at adjacent response elements. J. Biol. Chem. 285, 31100–31106.

Prasad, B., Lai, Y., Lin, Y., and Unadkat, J.D. (2013). Interindividual variability in the hepatic expression of the human breast cancer resistance protein (BCRP/ABCG2): effect of age, sex, and genotype. J Pharm. Sci. 102, 787–793.

189 Qian, X., Mruk, D.D., Wong, E.W.P., and Cheng, C.Y. (2013). Breast cancer resistance protein regulates apical ectoplasmic specialization dynamics stage specifically in the rat testis. Am. J. Physiol. Endocrinol. Metab. 304, E757–769.

Quintás-Cardama, A., and Verstovsek, S. (2013). Molecular pathways: Jak/STAT pathway: mutations, inhibitors, and resistance. Clin. Cancer Res. 19, 1933–1940.

Raaijmakers, M.H.G.P., de Grouw, E.P.L.M., Heuver, L.H.H., van der Reijden, B.A., Jansen, J.H., Scheper, R.J., Scheffer, G.L., de Witte, T.J.M., and Raymakers, R.A.P. (2005). Breast cancer resistance protein in drug resistance of primitive CD34+38- cells in acute myeloid leukemia. Clin. Cancer Res. 11, 2436–2444.

Rabindran, S.K., Ross, D.D., Doyle, L.A., Yang, W., and Greenberger, L.M. (2000). Fumitremorgin C reverses multidrug resistance in cells transfected with the breast cancer resistance protein. Cancer Res. 60, 47–50.

Ramadoss, P., Unger-Smith, N.E., Lam, F.S., and Hollenberg, A.N. (2009). STAT3 targets the regulatory regions of gluconeogenic genes in vivo. Mol. Endocrinol. 23, 827–837.

Ramamoorthy, P., Sticca, R., Wagner, T.E., and Chen, W.Y. (2001). In vitro studies of a prolactin antagonist, hPRL-G129R in human breast cancer cells. Int. J. Oncol. 18, 25–32.

Rani, A., Greenlaw, R., Runglall, M., Jurcevic, S., and John, S. (2014). FRA2 Is a STAT5 Target Gene Regulated by IL-2 in Human CD4 T Cells. PLoS ONE 9, e90370.

Reichmann, E., Ball, R., Groner, B., and Friis, R.R. (1989). New mammary epithelial and fibroblastic cell clones in coculture form structures competent to differentiate functionally. J. Cell Biol. 108, 1127–1138.

Reynolds, C., Montone, K.T., Powell, C.M., Tomaszewski, J.E., and Clevenger, C.V. (1997). Expression of prolactin and its receptor in human breast carcinoma. Endocrinology 138, 5555– 5560.

Richert, M.M., Schwertfeger, K.L., Ryder, J.W., and Anderson, S.M. (2000). An Atlas of Mouse Mammary Gland Development. J. Mammary Gland Biol. Neoplasia 5, 227–241.

Rijnkels, M., Kabotyanski, E., Montazer-Torbati, M.B., Hue Beauvais, C., Vassetzky, Y., Rosen, J.M., and Devinoy, E. (2010). The epigenetic landscape of mammary gland development and functional differentiation. J. Mammary Gland Biol. Neoplasia 15, 85–100.

Rijnkels, M., Freeman-Zadrowski, C., Hernandez, J., Potluri, V., Wang, L., Li, W., and Lemay, D.G. (2013). Epigenetic modifications unlock the milk protein gene loci during mouse mammary gland development and differentiation. PLoS ONE 8, e53270.

Riordan, J.R., Deuchars, K., Kartner, N., Alon, N., Trent, J., and Ling, V. (1985). Amplification of P-glycoprotein genes in multidrug-resistant mammalian cell lines. Nature 316, 817–819.

Robey, R.W., Medina-Pérez, W.Y., Nishiyama, K., Lahusen, T., Miyake, K., Litman, T., Senderowicz, A.M., Ross, D.D., and Bates, S.E. (2001). Overexpression of the ATP-binding

190 cassette half-transporter, ABCG2 (Mxr/BCrp/ABCP1), in flavopiridol-resistant human breast cancer cells. Clin. Cancer Res. 7, 145–152.

Robey, R.W., Honjo, Y., Morisaki, K., Nadjem, T.A., Runge, S., Risbood, M., Poruchynsky, M.S., and Bates, S.E. (2003). Mutations at amino-acid 482 in the ABCG2 gene affect substrate and antagonist specificity. Br. J. Cancer 89, 1971–1978.

Robey, R.W., Steadman, K., Polgar, O., Morisaki, K., Blayney, M., Mistry, P., and Bates, S.E. (2004). Pheophorbide a is a specific probe for ABCG2 function and inhibition. Cancer Res. 64, 1242–1246.

Roninson, I.B., Chin, J.E., Choi, K.G., Gros, P., Housman, D.E., Fojo, A., Shen, D.W., Gottesman, M.M., and Pastan, I. (1986). Isolation of human mdr DNA sequences amplified in multidrug-resistant KB carcinoma cells. Proc. Natl. Acad. Sci. U.S.A. 83, 4538–4542.

Rosales, R., Romero, M.R., Vaquero, J., Monte, M.J., Requena, P., Martinez-Augustin, O., Sanchez de Medina, F., and Marin, J.J.G. (2013). FXR-dependent and -independent interaction of glucocorticoids with the regulatory pathways involved in the control of bile acid handling by the liver. Biochem. Pharmacol. 85, 829–838.

Rowan, B.G., Weigel, N.L., and O’Malley, B.W. (2000). Phosphorylation of steroid receptor coactivator-1. Identification of the phosphorylation sites and phosphorylation through the mitogen-activated protein kinase pathway. J. Biol. Chem. 275, 4475–4483.

Rui, H., Kirken, R.A., and Farrar, W.L. (1994). Activation of receptor-associated tyrosine kinase JAK2 by prolactin. J. Biol. Chem. 269, 5364–5368.

Ryan, M.C., Zeeberg, B.R., Caplen, N.J., Cleland, J.A., Kahn, A.B., Liu, H., and Weinstein, J.N. (2008). SpliceCenter: A suite of web-based bioinformatic applications for evaluating the impact of alternative splicing on RT-PCR, RNAi, microarray, and peptide-based studies. BMC Bioinformatics 9, 313.

Saito, J., Hirota, T., Furuta, S., Kobayashi, D., Takane, H., and Ieiri, I. (2013). Association between DNA methylation in the miR-328 5’-flanking region and inter-individual differences in miR-328 and BCRP expression in human placenta. PLoS ONE 8, e72906.

Salphati, L., Lee, L.B., Pang, J., Plise, E.G., and Zhang, X. (2010). Role of P-glycoprotein and breast cancer resistance protein-1 in the brain penetration and brain pharmacodynamic activity of the novel phosphatidylinositol 3-kinase inhibitor GDC-0941. Drug Metab. Dispos. 38, 1422–1426.

Saura, M., Zaragoza, C., Bao, C., Herranz, B., Rodriguez-Puyol, M., and Lowenstein, C.J. (2006). Stat3 mediates interleukin-6 [correction of interelukin-6] inhibition of human endothelial nitric-oxide synthase expression. J. Biol. Chem. 281, 30057–30062.

Schaber, J.D., Fang, H., Xu, J., Grimley, P.M., and Rui, H. (1998). Prolactin activates Stat1 but does not antagonize Stat1 activation and growth inhibition by type I interferons in human breast cancer cells. Cancer Res. 58, 1914–1919.

191 Scharenberg, C., Mannowetz, N., Robey, R.W., Brendel, C., Repges, P., Sahrhage, T., Jähn, T., and Wennemuth, G. (2009). ABCG2 is expressed in late spermatogenesis and is associated with the acrosome. Biochem. Biophys. Res. Commun. 378, 302–307.

Scharenberg, C.W., Harkey, M.A., and Torok-Storb, B. (2002). The ABCG2 transporter is an efficient Hoechst 33342 efflux pump and is preferentially expressed by immature human hematopoietic progenitors. Blood 99, 507–512.

Schmitt-Ney, M., Doppler, W., Ball, R.K., and Groner, B. (1991). Beta-casein gene promoter activity is regulated by the hormone-mediated relief of transcriptional repression and a mammary-gland-specific nuclear factor. Mol. Cell. Biol. 11, 3745–3755.

Schwertfeger, K.L., Hunter, S., Heasley, L.E., Levresse, V., Leon, R.P., DeGregori, J., and Anderson, S.M. (2000). Prolactin stimulates activation of c-jun N-terminal kinase (JNK). Mol. Endocrinol. 14, 1592–1602.

Sereti, K.-I., Oikonomopoulos, A., Unno, K., Cao, X., Qiu, Y., and Liao, R. (2013). ATP- binding cassette G-subfamily transporter 2 regulates cell cycle progression and asymmetric division in mouse cardiac side population progenitor cells. Circ. Res. 112, 27–34.

Shaw-Bruha, C.M., Pirrucello, S.J., and Shull, J.D. (1997). Expression of the prolactin gene in normal and neoplastic human breast tissues and human mammary cell lines: promoter usage and alternative mRNA splicing. Breast Cancer Res. Treat. 44, 243–253.

Shen, B., Li, D., Dong, P., and Gao, S. (2011). Expression of ABC transporters is an unfavorable prognostic factor in laryngeal squamous cell carcinoma. Ann. Otol. Rhinol. Laryngol. 120, 820–827.

Shiu, R.P. (1979). Prolactin receptors in human breast cancer cells in long-term tissue culture. Cancer Res. 39, 4381–4386.

Shukla, S., Wu, C.-P., Nandigama, K., and Ambudkar, S.V. (2007). The naphthoquinones, vitamin K3 and its structural analogue plumbagin, are substrates of the multidrug resistance linked ATP binding cassette drug transporter ABCG2. Mol. Cancer Ther. 6, 3279–3286.

Sims-Mourtada, J., Izzo, J.G., Ajani, J., and Chao, K.S.C. (2007). Sonic Hedgehog promotes multiple drug resistance by regulation of drug transport. Oncogene 26, 5674–5679.

Singh, A., Wu, H., Zhang, P., Happel, C., Ma, J., and Biswal, S. (2010). Expression of ABCG2 (BCRP) is regulated by Nrf2 in cancer cells that confers side population and chemoresistance phenotype. Mol. Cancer Ther. 9, 2365–2376.

Sinha, Y.N., Selby, F.W., and Vanderlaan, W.P. (1974). Relationship of prolactin and growth hormone to mammary function during pregnancy and lactation in the C3H-ST mouse. J. Endocrinol. 61, 219–229.

Sisodiya, S.M., Martinian, L., Scheffer, G.L., van der Valk, P., Cross, J.H., Scheper, R.J., Harding, B.N., and Thom, M. (2003). Major vault protein, a marker of drug resistance, is upregulated in refractory epilepsy. Epilepsia 44, 1388–1396.

192 Soldaini, E., John, S., Moro, S., Bollenbacher, J., Schindler, U., and Leonard, W.J. (2000). DNA binding site selection of dimeric and tetrameric Stat5 proteins reveals a large repertoire of divergent tetrameric Stat5a binding sites. Mol. Cell. Biol. 20, 389–401.

Spicuglia, S., and Vanhille, L. (2012). Chromatin signatures of active enhancers. Nucleus 3, 126–131.

Steinbach, D., Sell, W., Voigt, A., Hermann, J., Zintl, F., and Sauerbrey, A. (2002). BCRP gene expression is associated with a poor response to remission induction therapy in childhood acute myeloid leukemia. Leukemia 16, 1443–1447.

Summer, R., Kotton, D.N., Sun, X., Ma, B., Fitzsimmons, K., and Fine, A. (2003). Side population cells and Bcrp1 expression in lung. Am. J. Physiol. Lung Cell Mol. Physiol. 285, L97–104.

Sun, H., Qin, B., Liu, T., Wang, Q., Liu, J., Wang, J., Lin, X., Yang, Y., Taing, L., Rao, P.K., et al. (2013). CistromeFinder for ChIP-seq and DNase-seq data reuse. Bioinformatics 29, 1352–1354.

Susanto, J., Lin, Y.-H., Chen, Y.-N., Shen, C.-R., Yan, Y.-T., Tsai, S.-T., Chen, C.-H., and Shen, C.-N. (2008). Porphyrin homeostasis maintained by ABCG2 regulates self-renewal of embryonic stem cells. PLoS ONE 3, e4023.

Suzuki, M., Suzuki, H., Sugimoto, Y., and Sugiyama, Y. (2003). ABCG2 transports sulfated conjugates of steroids and xenobiotics. J. Biol. Chem. 278, 22644–22649.

Szatmari, I., Vámosi, G., Brazda, P., Balint, B.L., Benko, S., Széles, L., Jeney, V., Ozvegy- Laczka, C., Szántó, A., Barta, E., et al. (2006). Peroxisome proliferator-activated receptor gamma-regulated ABCG2 expression confers cytoprotection to human dendritic cells. J. Biol. Chem. 281, 23812–23823.

Szczuraszek, K., Materna, V., Halon, A., Mazur, G., Wróbel, T., Kuliczkowski, K., Maciejczyk, A., Zabel, M., Drag, M., Dietel, M., et al. (2009). Positive correlation between cyclooxygenase-2 and ABC-transporter expression in non-Hodgkin’s lymphomas. Oncol. Rep. 22, 1315–1323.

Tachikawa, M., Watanabe, M., Hori, S., Fukaya, M., Ohtsuki, S., Asashima, T., and Terasaki, T. (2005). Distinct spatio-temporal expression of ABCA and ABCG transporters in the developing and adult mouse brain. J. Neurochem. 95, 294–304.

Tai, L.M., Loughlin, A.J., Male, D.K., and Romero, I.A. (2009). P-glycoprotein and breast cancer resistance protein restrict apical-to-basolateral permeability of human brain endothelium to amyloid-beta. J. Cereb. Blood Flow Metab. 29, 1079–1083.

Takada, T., Suzuki, H., Gotoh, Y., and Sugiyama, Y. (2005). Regulation of the cell surface expression of human BCRP/ABCG2 by the phosphorylation state of Akt in polarized cells. Drug Metab. Dispos. 33, 905–909.

193 Takenaka, K., Morgan, J.A., Scheffer, G.L., Adachi, M., Stewart, C.F., Sun, D., Leggas, M., Ejendal, K.F.K., Hrycyna, C.A., and Schuetz, J.D. (2007). Substrate overlap between Mrp4 and Abcg2/Bcrp affects purine analogue drug cytotoxicity and tissue distribution. Cancer Res. 67, 6965–6972.

Tan, S.-H., and Nevalainen, M.T. (2008). Signal transducer and activator of transcription 5A/B in prostate and breast cancers. Endocr. Relat. Cancer 15, 367–390.

Tan, K.P., Wang, B., Yang, M., Boutros, P.C., Macaulay, J., Xu, H., Chuang, A.I., Kosuge, K., Yamamoto, M., Takahashi, S., et al. (2010). Aryl hydrocarbon receptor is a transcriptional activator of the human breast cancer resistance protein (BCRP/ABCG2). Mol. Pharmacol. 78, 175–185.

Tanaka, K.K., Hall, J.K., Troy, A.A., Cornelison, D.D.W., Majka, S.M., and Olwin, B.B. (2009). Syndecan-4-expressing muscle progenitor cells in the SP engraft as satellite cells during muscle regeneration. Cell Stem Cell 4, 217–225.

Tanaka, Y., Slitt, A.L., Leazer, T.M., Maher, J.M., and Klaassen, C.D. (2005). Tissue distribution and hormonal regulation of the breast cancer resistance protein (Bcrp/Abcg2) in rats and mice. Biochem. Biophys. Res. Commun. 326, 181–187.

Tang, S.C., Lagas, J.S., Lankheet, N.A.G., Poller, B., Hillebrand, M.J., Rosing, H., Beijnen, J.H., and Schinkel, A.H. (2012). Brain accumulation of sunitinib is restricted by P-glycoprotein (ABCB1) and breast cancer resistance protein (ABCG2) and can be enhanced by oral elacridar and sunitinib coadministration. Int. J. Cancer 130, 223–233.

Teng, S., and Piquette-Miller, M. (2005). The involvement of the pregnane X receptor in hepatic gene regulation during inflammation in mice. J. Pharmacol. Exp. Ther. 312, 841–848.

Thompson, M.D., and Nakhasi, H.L. (1985). Methylation and expression of rat kappa-casein gene in normal and neoplastic rat mammary gland. Cancer Res. 45, 1291–1295.

To, K.K.W., Zhan, Z., and Bates, S.E. (2006). Aberrant promoter methylation of the ABCG2 gene in renal carcinoma. Mol. Cell. Biol. 26, 8572–8585.

To, K.K.W., Zhan, Z., Litman, T., and Bates, S.E. (2008a). Regulation of ABCG2 expression at the 3’ untranslated region of its mRNA through modulation of transcript stability and protein translation by a putative microRNA in the S1 colon cancer cell line. Mol. Cell. Biol. 28, 5147– 5161.

To, K.K.W., Polgar, O., Huff, L.M., Morisaki, K., and Bates, S.E. (2008b). Histone modifications at the ABCG2 promoter following treatment with histone deacetylase inhibitor mirror those in multidrug-resistant cells. Mol. Cancer Res. 6, 151–164.

To, K.K.W., Robey, R.W., Knutsen, T., Zhan, Z., Ried, T., and Bates, S.E. (2009). Escape from hsa-miR-519c enables drug-resistant cells to maintain high expression of ABCG2. Mol. Cancer Ther. 8, 2959–2968.

194 To, K.K.W., Robey, R., Zhan, Z., Bangiolo, L., and Bates, S.E. (2011). Upregulation of ABCG2 by Romidepsin via the Aryl Hydrocarbon Receptor Pathway. Mol. Cancer Res. 9, 516–527.

Tompkins, L.M., Li, H., Li, L., Lynch, C., Xie, Y., Nakanishi, T., Ross, D.D., and Wang, H. (2010). A novel xenobiotic responsive element regulated by aryl hydrocarbon receptor is involved in the induction of BCRP/ABCG2 in LS174T cells. Biochem. Pharmacol. 80, 1754– 1761.

Touraine, P., Martini, J.F., Zafrani, B., Durand, J.C., Labaille, F., Malet, C., Nicolas, A., Trivin, C., Postel-Vinay, M.C., Kuttenn, F., et al. (1998). Increased expression of prolactin receptor gene assessed by quantitative polymerase chain reaction in human breast tumors versus normal breast tissues. J. Clin. Endocrinol. Metab. 83, 667–674.

Tsunoda, S., Okumura, T., Ito, T., Kondo, K., Ortiz, C., Tanaka, E., Watanabe, G., Itami, A., Sakai, Y., and Shimada, Y. (2006). ABCG2 expression is an independent unfavorable prognostic factor in esophageal squamous cell carcinoma. Oncology 71, 251–258.

Tucker, T.G.H.A., Milne, A.M., Fournel-Gigleux, S., Fenner, K.S., and Coughtrie, M.W.H. (2012). Absolute immunoquantification of the expression of ABC transporters P-glycoprotein, breast cancer resistance protein and multidrug resistance-associated protein 2 in human liver and duodenum. Biochem. Pharmacol. 83, 279–285.

Turner, J.G., Gump, J.L., Zhang, C., Cook, J.M., Marchion, D., Hazlehurst, L., Munster, P., Schell, M.J., Dalton, W.S., and Sullivan, D.M. (2006). ABCG2 expression, function, and promoter methylation in human multiple myeloma. Blood 108, 3881–3889.

Udy, G.B., Towers, R.P., Snell, R.G., Wilkins, R.J., Park, S.H., Ram, P.A., Waxman, D.J., and Davey, H.W. (1997). Requirement of STAT5b for sexual dimorphism of body growth rates and liver gene expression. Proc. Natl. Acad. Sci. U.S.A. 94, 7239–7244.

Urquhart, B.L., Ware, J.A., Tirona, R.G., Ho, R.H., Leake, B.F., Schwarz, U.I., Zaher, H., Palandra, J., Gregor, J.C., Dresser, G.K., et al. (2008). Breast cancer resistance protein (ABCG2) and drug disposition: intestinal expression, polymorphisms and sulfasalazine as an in vivo probe. Pharmacogenet. Genomics 18, 439–448.

Utama, F.E., LeBaron, M.J., Neilson, L.M., Sultan, A.S., Parlow, A.F., Wagner, K.-U., and Rui, H. (2006). Human prolactin receptors are insensitive to mouse prolactin: implications for xenotransplant modeling of human breast cancer in mice. J. Endocrinol. 188, 589–601. van Herwaarden, A.E., Jonker, J.W., Wagenaar, E., Brinkhuis, R.F., Schellens, J.H.M., Beijnen, J.H., and Schinkel, A.H. (2003). The breast cancer resistance protein (Bcrp1/Abcg2) restricts exposure to the dietary carcinogen 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine. Cancer Res. 63, 6447–6452. van Herwaarden, A.E., Wagenaar, E., Karnekamp, B., Merino, G., Jonker, J.W., and Schinkel, A.H. (2006). Breast cancer resistance protein (Bcrp1/Abcg2) reduces systemic exposure of the dietary carcinogens aflatoxin B1, IQ and Trp-P-1 but also mediates their secretion into breast milk. Carcinogenesis 27, 123–130.

195 van Herwaarden, A.E., Wagenaar, E., Merino, G., Jonker, J.W., Rosing, H., Beijnen, J.H., and Schinkel, A.H. (2007). Multidrug transporter ABCG2/breast cancer resistance protein secretes riboflavin (vitamin B2) into milk. Mol. Cell. Biol. 27, 1247–1253. van den Heuvel-Eibrink, M.M., Wiemer, E.A.C., Prins, A., Meijerink, J.P.P., Vossebeld, P.J.M., van der Holt, B., Pieters, R., and Sonneveld, P. (2002). Increased expression of the breast cancer resistance protein (BCRP) in relapsed or refractory acute myeloid leukemia (AML). Leukemia 16, 833–839.

Vander Borght, S., Libbrecht, L., Katoonizadeh, A., van Pelt, J., Cassiman, D., Nevens, F., Van Lommel, A., Petersen, B.E., Fevery, J., Jansen, P.L., et al. (2006). Breast cancer resistance protein (BCRP/ABCG2) is expressed by progenitor cells/reactive ductules and hepatocytes and its expression pattern is influenced by disease etiology and species type: possible functional consequences. J. Histochem. Cytochem. 54, 1051–1059.

Vanselow, J., Yang, W., Herrmann, J., Zerbe, H., Schuberth, H.-J., Petzl, W., Tomek, W., and Seyfert, H.-M. (2006). DNA-remethylation around a STAT5-binding enhancer in the alphaS1- casein promoter is associated with abrupt shutdown of alphaS1-casein synthesis during acute mastitis. J. Mol. Endocrinol. 37, 463–477.

Volk, E.L., and Schneider, E. (2003). Wild-type breast cancer resistance protein (BCRP/ABCG2) is a methotrexate polyglutamate transporter. Cancer Res. 63, 5538–5543.

Volk, E.L., Farley, K.M., Wu, Y., Li, F., Robey, R.W., and Schneider, E. (2002). Overexpression of wild-type breast cancer resistance protein mediates methotrexate resistance. Cancer Res. 62, 5035–5040.

Wagner, K.-U., Krempler, A., Triplett, A.A., Qi, Y., George, N.M., Zhu, J., and Rui, H. (2004). Impaired Alveologenesis and Maintenance of Secretory Mammary Epithelial Cells in Jak2 Conditional Knockout Mice. Mol. Cell. Biol. 24, 5510–5520.

Wakao, H., Gouilleux, F., and Groner, B. (1994). Mammary gland factor (MGF) is a novel member of the cytokine regulated transcription factor gene family and confers the prolactin response. EMBO J. 13, 2182–2191.

Walker, S.R., Nelson, E.A., and Frank, D.A. (2007). STAT5 represses BCL6 expression by binding to a regulatory region frequently mutated in lymphomas. Oncogene 26, 224–233.

Walker, S.R., Nelson, E.A., Zou, L., Chaudhury, M., Signoretti, S., Richardson, A., and Frank, D.A. (2009). Reciprocal effects of STAT5 and STAT3 in breast cancer. Mol. Cancer Res. 7, 966–976.

Walker, S.R., Nelson, E.A., Yeh, J.E., Pinello, L., Yuan, G.-C., and Frank, D.A. (2013). STAT5 outcompetes STAT3 to regulate the expression of the oncogenic transcriptional modulator BCL6. Mol. Cell. Biol. 33, 2879–2890.

Wang, F., Xue, X., Wei, J., An, Y., Yao, J., Cai, H., Wu, J., Dai, C., Qian, Z., Xu, Z., et al. (2010a). hsa-miR-520h downregulates ABCG2 in pancreatic cancer cells to inhibit migration, invasion, and side populations. Br. J. Cancer 103, 567–574.

196 Wang, H., Wu, X., Hudkins, K., Mikheev, A., Zhang, H., Gupta, A., Unadkat, J.D., and Mao, Q. (2006a). Expression of the breast cancer resistance protein (Bcrp1/Abcg2) in tissues from pregnant mice: effects of pregnancy and correlations with nuclear receptors. Am. J. Physiol. Endocrinol. Metab. 291, E1295–1304.

Wang, H., Zhou, L., Gupta, A., Vethanayagam, R.R., Zhang, Y., Unadkat, J.D., and Mao, Q. (2006b). Regulation of BCRP/ABCG2 expression by progesterone and 17beta-estradiol in human placental BeWo cells. Am. J. Physiol. Endocrinol. Metab. 290, E798–807.

Wang, H., Unadkat, J.D., and Mao, Q. (2008a). Hormonal regulation of BCRP expression in human placental BeWo cells. Pharm. Res. 25, 444–452.

Wang, H., Lee, E.-W., Zhou, L., Leung, P.C.K., Ross, D.D., Unadkat, J.D., and Mao, Q. (2008b). Progesterone receptor (PR) isoforms PRA and PRB differentially regulate expression of the breast cancer resistance protein in human placental choriocarcinoma BeWo cells. Mol. Pharmacol. 73, 845–854.

Wang, T., Agarwal, S., and Elmquist, W.F. (2012). Brain Distribution of Cediranib Is Limited by Active Efflux at the Blood-Brain Barrier. J. Pharmacol. Exp. Ther. 341, 386–395.

Wang, X., Furukawa, T., Nitanda, T., Okamoto, M., Sugimoto, Y., Akiyama, S.-I., and Baba, M. (2003). Breast cancer resistance protein (BCRP/ABCG2) induces cellular resistance to HIV-1 nucleoside reverse transcriptase inhibitors. Mol. Pharmacol. 63, 65–72.

Wang, X., Wu, X., Wang, C., Zhang, W., Ouyang, Y., Yu, Y., and He, Z. (2010b). Transcriptional suppression of breast cancer resistance protein (BCRP) by wild-type p53 through the NF-kappaB pathway in MCF-7 cells. FEBS Lett. 584, 3392–3397.

Wang, X., Sykes, D.B., and Miller, D.S. (2010c). Constitutive androstane receptor-mediated up-regulation of ATP-driven xenobiotic efflux transporters at the blood-brain barrier. Mol. Pharmacol. 78, 376–383.

Wang, X., Xia, B., Liang, Y., Peng, L., Wang, Z., Zhuo, J., Wang, W., and Jiang, B. (2013). Membranous ABCG2 expression in colorectal cancer independently correlates with shortened patient survival. Cancer Biomark 13, 81–88.

Wang, X.Q., Ongkeko, W.M., Chen, L., Yang, Z.F., Lu, P., Chen, K.K., Lopez, J.P., Poon, R.T.P., and Fan, S.T. (2010d). Octamer 4 (Oct4) mediates chemotherapeutic drug resistance in liver cancer cells through a potential Oct4-AKT-ATP-binding cassette G2 pathway. Hepatology 52, 528–539.

Warsch, W., Walz, C., and Sexl, V. (2013). JAK of all trades: JAK2-STAT5 as novel therapeutic targets in BCR-ABL1+ chronic myeloid leukemia. Blood 122, 2167–2175.

Watkin, H., Richert, M.M., Lewis, A., Terrell, K., McManaman, J.P., and Anderson, S.M. (2008). Lactation failure in Src knockout mice is due to impaired secretory activation. BMC Dev. Biol. 8, 6.

197 Watson, C.J. (2006). Involution: apoptosis and tissue remodelling that convert the mammary gland from milk factory to a quiescent organ. Breast Cancer Res. 8, 203.

Watson, C.J., and Khaled, W.T. (2008). Mammary development in the embryo and adult: a journey of morphogenesis and commitment. Development 135, 995–1003.

Watson, C.J., and Kreuzaler, P.A. (2011). Remodeling mechanisms of the mammary gland during involution. Int. J. Dev. Biol. 55, 757–762.

Wauthier, V., Sugathan, A., Meyer, R.D., Dombkowski, A.A., and Waxman, D.J. (2010). Intrinsic Sex Differences in the Early Growth Hormone Responsiveness of Sex-Specific Genes in Mouse Liver. Mol. Endocrinol. 24, 667–678.

Waxman, D.J., and Holloway, M.G. (2009). Sex differences in the expression of hepatic drug metabolizing enzymes. Mol. Pharmacol. 76, 215–228.

Widström, A.M., Winberg, J., Werner, S., Hamberger, B., Eneroth, P., and Uvnäs-Moberg, K. (1984). Suckling in lactating women stimulates the secretion of insulin and prolactin without concomitant effects on gastrin, growth hormone, calcitonin, vasopressin or catecholamines. Early Hum. Dev. 10, 115–122.

Williams, C.C., Allison, J.G., Vidal, G.A., Burow, M.E., Beckman, B.S., Marrero, L., and Jones, F.E. (2004). The ERBB4/HER4 receptor tyrosine kinase regulates gene expression by functioning as a STAT5A nuclear chaperone. J. Cell. Biol. 167, 469–478.

Woelfle, J., Chia, D.J., and Rotwein, P. (2003). Mechanisms of growth hormone (GH) action. Identification of conserved Stat5 binding sites that mediate GH-induced insulin-like growth factor-I gene activation. J. Biol. Chem. 278, 51261–51266.

Woodward, O.M., Köttgen, A., Coresh, J., Boerwinkle, E., Guggino, W.B., and Köttgen, M. (2009). Identification of a urate transporter, ABCG2, with a common functional polymorphism causing gout. Proc. Natl. Acad. Sci. U.S.A. 106, 10338–10342.

Wu, A.M.L., Dalvi, P., Lu, X., Yang, M., Riddick, D.S., Matthews, J., Clevenger, C.V., Ross, D.D., Harper, P.A., and Ito, S. (2013). Induction of multidrug resistance transporter ABCG2 by prolactin in human breast cancer cells. Mol. Pharmacol. 83, 377–388.

Wu, A.M.L., Yang, M., Dalvi, P., Turinsky, A.L., Wang, W., Butcher, D., Egan, S.E., Weksberg, R., Harper, P.A., and Ito, S. (2014). Role of STAT5 and epigenetics in lactation- associated upregulation of multidrug transporter ABCG2 in the mammary gland. Am. J. Physiol. Endocrinol. Metab. - in press

Xie, Y., Natarajan, K., Bauer, K.S., Nakanishi, T., Beck, W.T., Moreci, R.S., Jeyasuria, P., Hussain, A., and Ross, D.D. (2013). Bcrp1 transcription in mouse testis is controlled by a promoter upstream of a novel first exon (E1U) regulated by steroidogenic factor-1. Biochim. Biophys. Acta 1829, 1288–1299.

Xiong, H., Callaghan, D., Jones, A., Bai, J., Rasquinha, I., Smith, C., Pei, K., Walker, D., Lue, L.-F., Stanimirovic, D., et al. (2009). ABCG2 is upregulated in Alzheimer’s brain with cerebral

198 amyloid angiopathy and may act as a gatekeeper at the blood-brain barrier for Abeta(1-40) peptides. J. Neurosci. 29, 5463–5475.

Xu, J., Zhang, Y., Berry, P.A., Jiang, J., Lobie, P.E., Langenheim, J.F., Chen, W.Y., and Frank, S.J. (2011). Growth hormone signaling in human T47D breast cancer cells: potential role for a growth hormone receptor-prolactin receptor complex. Mol. Endocrinol. 25, 597–610.

Xu, R., Spencer, V.A., and Bissell, M.J. (2007). Extracellular matrix-regulated gene expression requires cooperation of SWI/SNF and transcription factors. J. Biol. Chem. 282, 14992–14999.

Xu, R., Nelson, C.M., Muschler, J.L., Veiseh, M., Vonderhaar, B.K., and Bissell, M.J. (2009). Sustained activation of STAT5 is essential for chromatin remodeling and maintenance of mammary-specific function. J. Cell Biol. 184, 57–66.

Yamaji, D., Kang, K., Robinson, G.W., and Hennighausen, L. (2013). Sequential activation of genetic programs in mouse mammary epithelium during pregnancy depends on STAT5A/B concentration. Nucleic Acids Res. 41, 1622–1636.

Yamauchi, T., Kaburagi, Y., Ueki, K., Tsuji, Y., Stark, G.R., Kerr, I.M., Tsushima, T., Akanuma, Y., Komuro, I., Tobe, K., et al. (1998). Growth hormone and prolactin stimulate tyrosine phosphorylation of insulin receptor substrate-1, -2, and -3, their association with p85 phosphatidylinositol 3-kinase (PI3-kinase), and concomitantly PI3-kinase activation via JAK2 kinase. J. Biol. Chem. 273, 15719–15726.

Yang, Q., Köttgen, A., Dehghan, A., Smith, A.V., Glazer, N.L., Chen, M.-H., Chasman, D.I., Aspelund, T., Eiriksdottir, G., Harris, T.B., et al. (2010). Multiple genetic loci influence serum urate levels and their relationship with gout and cardiovascular disease risk factors. Circ. Cardiovasc. Genet. 3, 523–530.

Yao, Z., Kanno, Y., Kerenyi, M., Stephens, G., Durant, L., Watford, W.T., Laurence, A., Robinson, G.W., Shevach, E.M., Moriggl, R., et al. (2007). Nonredundant roles for Stat5a/b in directly regulating Foxp3. Blood 109, 4368–4375.

Yasuda, K., Cline, C., Vogel, P., Onciu, M., Fatima, S., Sorrentino, B.P., Thirumaran, R.K., Ekins, S., Urade, Y., Fujimori, K., et al. (2013). Drug transporters on arachnoid barrier cells contribute to the blood-cerebrospinal fluid barrier. Drug Metab. Dispos. 41, 923–931.

Yasuda, S., Kobayashi, M., Itagaki, S., Hirano, T., and Iseki, K. (2009). Response of the ABCG2 promoter in T47D cells and BeWo cells to sex hormone treatment. Mol. Biol. Rep. 36, 1889–1896.

Yeboah, D., Sun, M., Kingdom, J., Baczyk, D., Lye, S.J., Matthews, S.G., and Gibb, W. (2006). Expression of breast cancer resistance protein (BCRP/ABCG2) in human placenta throughout gestation and at term before and after labor. Can. J. Physiol. Pharmacol. 84, 1251–1258.

Yoh, K., Ishii, G., Yokose, T., Minegishi, Y., Tsuta, K., Goto, K., Nishiwaki, Y., Kodama, T., Suga, M., and Ochiai, A. (2004). Breast cancer resistance protein impacts clinical outcome in platinum-based chemotherapy for advanced non-small cell lung cancer. Clin. Cancer Res. 10, 1691–1697.

199 Yoon, S., and Seger, R. (2006). The extracellular signal-regulated kinase: multiple substrates regulate diverse cellular functions. Growth Factors 24, 21–44.

Zamber, C.P., Lamba, J.K., Yasuda, K., Farnum, J., Thummel, K., Schuetz, J.D., and Schuetz, E.G. (2003). Natural allelic variants of breast cancer resistance protein (BCRP) and their relationship to BCRP expression in human intestine. Pharmacogenetics 13, 19–28.

Zhang, J., Scordi, I., Smyth, M.J., and Lichtenheld, M.G. (1999). Interleukin 2 Receptor Signaling Regulates the Perforin Gene through Signal Transducer and Activator of Transcription (Stat)5 Activation of Two Enhancers. J. Exp. Med 190, 1297–1308.

Zhang, J., Brewer, S., Huang, J., and Williams, T. (2003a). Overexpression of transcription factor AP-2α suppresses mammary gland growth and morphogenesis. Dev. Biol. 256, 128–146.

Zhang, W., Mojsilovic-Petrovic, J., Andrade, M.F., Zhang, H., Ball, M., and Stanimirovic, D.B. (2003b). The expression and functional characterization of ABCG2 in brain endothelial cells and vessels. FASEB J. 17, 2085–2087.

Zhang, W., Ding, W., Chen, Y., Feng, M., Ouyang, Y., Yu, Y., and He, Z. (2011). Up- regulation of breast cancer resistance protein plays a role in HER2-mediated chemoresistance through PI3K/Akt and nuclear factor-kappa B signaling pathways in MCF7 breast cancer cells. Acta Biochim. Biophys. Sin. (Shanghai) 43, 647–653.

Zhang, Y., Zhou, G., Wang, H., Zhang, X., Wei, F., Cai, Y., and Yin, D. (2006). Transcriptional upregulation of breast cancer resistance protein by 17beta-estradiol in ERalpha-positive MCF-7 breast cancer cells. Oncology 71, 446–455.

Zhang, Y., Wang, H., Unadkat, J.D., and Mao, Q. (2007a). Breast cancer resistance protein 1 limits fetal distribution of nitrofurantoin in the pregnant mouse. Drug Metab. Dispos. 35, 2154–2158.

Zhang, Y., Bressler, J.P., Neal, J., Lal, B., Bhang, H.-E.C., Laterra, J., and Pomper, M.G. (2007b). ABCG2/BCRP expression modulates D-Luciferin based bioluminescence imaging. Cancer Res. 67, 9389–9397.

Zhang, Y., Liu, T., Meyer, C.A., Eeckhoute, J., Johnson, D.S., Bernstein, B.E., Nusbaum, C., Myers, R.M., Brown, M., Li, W., et al. (2008). Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137.

Zhang, Y., Laz, E.V., and Waxman, D.J. (2012). Dynamic, sex-differential STAT5 and BCL6 binding to sex-biased, growth hormone-regulated genes in adult mouse liver. Mol. Cell. Biol. 32, 880–896.

Zhou, H.-L., Luo, G., Wise, J.A., and Lou, H. (2014). Regulation of alternative splicing by local histone modifications: potential roles for RNA-guided mechanisms. Nucleic Acids Res. 42, 701–713.

Zhou, L., Naraharisetti, S.B., Wang, H., Unadkat, J.D., Hebert, M.F., and Mao, Q. (2008). The Breast Cancer Resistance Protein (Bcrp1/Abcg2) Limits Fetal Distribution of Glyburide in the

200 Pregnant Mouse: An Obstetric-Fetal Pharmacology Research Unit Network and University of Washington Specialized Center of Research Study. Mol. Pharmacol. 73, 949–959.

Zhou, L., Schmidt, K., Nelson, F.R., Zelesky, V., Troutman, M.D., and Feng, B. (2009). The effect of breast cancer resistance protein and P-glycoprotein on the brain penetration of flavopiridol, imatinib mesylate (Gleevec), prazosin, and 2-methoxy-3-(4-(2-(5-methyl-2- phenyloxazol-4-yl)ethoxy)phenyl)propanoic acid (PF-407288) in mice. Drug Metab. Dispos. 37, 946–955.

Zhou, S., Schuetz, J.D., Bunting, K.D., Colapietro, A.M., Sampath, J., Morris, J.J., Lagutina, I., Grosveld, G.C., Osawa, M., Nakauchi, H., et al. (2001). The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype. Nat. Med. 7, 1028–1034.

Zhou, S., Morris, J.J., Barnes, Y., Lan, L., Schuetz, J.D., and Sorrentino, B.P. (2002). Bcrp1 gene expression is required for normal numbers of side population stem cells in mice, and confers relative protection to mitoxantrone in hematopoietic cells in vivo. Proc. Natl. Acad. Sci. U.S.A. 99, 12339–12344.

Zhou, S., Zong, Y., Ney, P.A., Nair, G., Stewart, C.F., and Sorrentino, B.P. (2005). Increased expression of the Abcg2 transporter during erythroid maturation plays a role in decreasing cellular protoporphyrin IX levels. Blood 105, 2571–2576.

Zhu, B.-M., Kang, K., Yu, J.H., Chen, W., Smith, H.E., Lee, D., Sun, H.-W., Wei, L., and Hennighausen, L. (2012). Genome-wide analyses reveal the extent of opportunistic STAT5 binding that does not yield transcriptional activation of neighboring genes. Nucleic Acids Res. 40, 4461–4472.

Zhu, W., Xu, H., Wang, S.W.J., and Hu, M. (2010). Breast cancer resistance protein (BCRP) and sulfotransferases contribute significantly to the disposition of genistein in mouse intestine. AAPS J. 12, 525–536.

Zhuang, Y., Fraga, C.H., Hubbard, K.E., Hagedorn, N., Panetta, J.C., Waters, C.M., and Stewart, C.F. (2006). Topotecan central nervous system penetration is altered by a tyrosine kinase inhibitor. Cancer Res. 66, 11305–11313.

Zong, Y., Zhou, S., Fatima, S., and Sorrentino, B.P. (2006). Expression of mouse Abcg2 mRNA during hematopoiesis is regulated by alternative use of multiple leader exons and promoters. J. Biol. Chem. 281, 29625–29632.

Zorn, E., Nelson, E.A., Mohseni, M., Porcheray, F., Kim, H., Litsa, D., Bellucci, R., Raderschall, E., Canning, C., Soiffer, R.J., et al. (2006). IL-2 regulates FOXP3 expression in human CD4+CD25+ regulatory T cells through a STAT-dependent mechanism and induces the expansion of these cells in vivo. Blood 108, 1571–1579.

Zrieki, A., Farinotti, R., and Buyse, M. (2008). Cyclooxygenase inhibitors down regulate P- glycoprotein in human colorectal Caco-2 cell line. Pharm. Res. 25, 1991–2001.

201 7. LIST OF PUBLICATIONS, ABSTRACTS, AND COPYRIGHT STATEMENTS

Publications Wu, A.M.L., Dalvi, P., Lu, X., Yang, M., Riddick, D.S., Matthews, J., Clevenger, C.V., Ross, D.D., Harper, P.A., and Ito, S. (2013). Induction of multidrug resistance transporter ABCG2 by prolactin in human breast cancer cells. Mol. Pharmacol. 83, 377–388.

Wu, A.M.L., Yang, M., Dalvi, P., Turinsky, A.L., Wang, W., Butcher, D., Egan, S.E., Weksberg, R., Harper, P.A., and Ito, S. (2014). Role of STAT5 and epigenetics in lactation- associated upregulation of multidrug transporter ABCG2 in the mammary gland. Am. J. Physiol. Endocrinol. Metab. – in press

Abstracts Wu A, Dalvi P, Yang M, Turinsky A, Wang K, Butcher D, Egan SE, Weksberg R, Brudno M, Harper PA, and Ito S (2014) Role of epigenetics and STAT5 in Breast Cancer Resistance Protein (BCRP/ABCG2) expression in the lactating mammary gland. Poster presented at Experimental Biology 2014, San Diego, CA.

Wu A, Yang M, Butcher D, Weksberg R, Harper PA, and Ito S (2013) Abcg2 mRNA isoform expression and DNA methylation profile in the mouse mammary gland. Poster presented at the 10th International ISSX meeting, Toronto, Canada.

Wu A, Dalvi P, Lu X, Yang M, Riddick DS, Matthews J, Clevenger CV, Ross DD, Harper PA, and Ito S (2012) Identification of a prolactin-responsive STAT5 binding element in the 5’- flanking region of the human BCRP/ABCG2 gene. Poster presented at the 19th International Symposium on Microsomes and Drug Oxidation and 12th European ISSX meeting, Noordwijk aan Zee, Netherlands.

Wu A, Dalvi P, Lu X, Yang M, Harper PA, and Ito S (2012) Regulation of multidrug resistance efflux transporter BCRP/ABCG2 by prolactin. Poster presented at Mammary Gland Biology, Gordon Research Conferences, Lucca (Barga), Italy.

Wu A, Dalvi P, Yang M, Lu X, Harper PA, Ito S (2012) Prolactin induces BCRP/ABCG2 expression in T-47D breast cancer cells via activation of multiple signalling cascades. FASEB J 26:1047.2. Poster presented at Experimental Biology 2012, San Diego, CA.

Copyright Statements Please note that certain passages and figures within sections 3.1, 4.1, and 4.5 contain materials from Wu et al. (2013) reprinted with permission from the American Society for Pharmacology and Experimental Therapeutics. See copyright permission letter (page 202).

Please note that certain passages and figures within sections 3.2, 3.3, 4.2, and 4.3 contain materials from Wu et al. (2014) reprinted without a need to request for permission from the American Physiological Society.

202

Council March 21, 2014

Richard R. Neubig President Michigan State University Alex Man Lai Wu Hospital for Sick Children Annette E. Fleckenstein President-Elect 555 University Avenue University of Utah Toronto Canada John S. Lazo Past President University of Virginia Email: [email protected] Sandra P. Welch Secretary/Treasurer Virginia Commonwealth University Dear Alex Man Lai Wu:

Paul A. Insel Secretary/Treasurer-Elect This is to grant you permission to include the following article in your thesis University of California – San Diego entitled “Regulation of ABCG2 expression in the lactating mammary gland” Edward T. Morgan for the University of Toronto: Past Secretary/Treasurer Emory University

Charles P. France Alex Man Lai Wu, Pooja Dalvi, Xiaoli Lu, Mingdong Yang, David S. Councilor University of Texas Health Science Riddick, Jason Matthews, Charles V. Clevenger, Douglas D. Ross, Center – San Antonio Patricia A. Harper, and Shinya Ito, Induction of Multidrug Resistance

John D. Schuetz Transporter ABCG2 by Prolactin in Human Breast Cancer Cells, Mol Councilor Pharmacol February 2013 83:377-388 St. Jude Children’s Research Hospital

Kenneth E. Thummel Councilor On the first page of each copy of this article, please add the following: University of Washington

Mary E. Vore Reprinted with permission of the American Society for Pharmacology Board of Publications Trustees University of Kentucky and Experimental Therapeutics. All rights reserved.

Brian M. Cox FASEB Board Representative In addition, the original copyright line published with the paper must be shown Uniformed Services University of the Health Sciences on the copies included with your thesis.

Scott A. Waldman Program Committee Sincerely yours, Thomas Jefferson University

Judith A. Siuciak Executive Officer

Richard Dodenhoff Journals Director

9650 Rockville Pike | Bethesda | MD | 20814-3995 P: (301) 634-7060 | F: (301) 634-7061 | E: [email protected] | www.aspet.org