Investigation of Ring Finger Protein 8 in Mammalian Physiology and Pathogenesis

Li Li

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Medical Biophysics University of Toronto

Physiology and Pathogenesis Li Li, Doctor of Philosophy, Department of Medical Biophysics, University of Toronto 2015

Abstract

While genomic stability is often viewed as a cellular shield against malignant transformation, DNA double-strand breaks (DSBs) are probably the most dangerous threat to the maintenance of a stable cellular genome. Mammalian cells have evolved a sophisticated DNA damage response (DDR) network that senses the physical existence of DNA DSBs, followed by signal propagation to repair these genetic lesions. Interestingly, DSBs are not only hazardous, but are also programmed to occur during normal physiological processes such as immunoglobulin heavy chain (IgH) class switch recombination (CSR), an important mechanism for antibody diversification and specification of immunoglobulin effector function during a humoral immune response in mammals. Through catalyzing ubiquitylation of the H2A-type histones flanking DNA DSBs, the

E3 ligase Ring finger protein 8 (Rnf8) orchestrates the assembly of components of homologous recombination (HR) and nonhomologous end joining (NHEJ) repair machineries into DSB- induced foci. To examine the in vivo functions of Rnf8 during mammalian development and determine the effects of its deficiency on the development of various diseases, we generated Rnf8 knockout mouse models and elucidated the pleiotropic in vivo functions of Rnf8. Overall, this thesis highlights the physiological significance of Rnf8 in shielding against DNA DSB repair defects, genomic instability, male infertility, immunodeficiency and cancer development at the organism level.

Acknowledgments

I would like to sincerely acknowledge everyone who helped me to make this thesis possible here.

First of all, a huge thanks goes to my mentors Dr. Razqallah Hakem and Dr. Anne Hakem, who inspired me through their passion and dedication for scientific discoveries, provided me with almost daily guidance throughout these years of my PhD study, and supported me unconditionally both inside and outside the lab. Furthermore, I would like to express my gratitude for my committee members Dr. Robert Bristow, Dr. Rama Khokha and Dr. Vuk Stambolic, who provided valuable feedbacks and boundless supports on my research, nurtured my budding interests for diverse scientific research areas apart from my previous education background, and guided me in the right direction during my PhD journey. To all the Hakem lab members both current and former,

I thank you for your friendship, advice and support over the last six and a half years, and enjoyed working with all of you. I also want to thank my friend Hartland Jackson for his support as well as having lunch and coffee chats about mammary gland biology during these years in MBP.

Many thanks to my wife for her love, support, and encouragement during this PhD experience. I could not say enough thank you to express my appreciation for your understanding of the life and timetable of a graduate student, thank you for waiting for me with a light on at home when I was working late nights in the lab so that I did not feel so lonely, and for understanding that I did not show up late for our appointments on purpose, and for all the sacrifices you made. I would not have endured this without your support, and feel that we have earned this degree together. I rarely have the opportunity to express my gratitude in writing to my Mama and Baba, so, here, I dedicate this thesis to you for all the spiritual and financial support, for loving me and believing in my potentials, for critically guiding me over the last three decades of my life experiences, for always being there to love and encourage me through rough times unconditionally.

iii

This thesis is dedicated to my Mama and Baba

Table of Contents

Abstract………………………………………………….………………………...………….….ii

Acknowledgments...... iii

Table of Contents...... v

List of Figures and Tables...... x

List of Abbreviations...... xiii

List of Publications………….…………………………………………………………...…...... xvi

Chapter 1 Introduction...... 1

1.1 Origins of cellular DNA double-strand breaks...... 2

1.1.1 Endogenous and exogenous sources of cellular DSBs...... 3

1.1.2 Genetically programmed generation of cellular DSBs...... 6

1.2 Signaling and repair of DNA DSBs and consequences of defective DSB

repair…...... 11

1.2.1 Intracellular sensing of DNA damage and signal propagation...... 11

1.2.2 Choice of cellular machineries to repair DSBs...... 20

1.2.3 Negative regulation and post-repair termination of DNA damage signaling...... 24

1.2.4 The central role of p53 in DDR and cellular fate…………………………...……….27

1.2.5 Human syndromes associated with defects in signaling and repair of DSBs...... 30

1.3 Mammary gland biology and breast cancer...... 35

1.3.1 Mammary gland development and differentiation...... 35

1.3.2 Notch signaling in mammary gland development and breast cancer...... 38

1.3.3 DNA damage response and breast cancer susceptibility genes...... 42

1.4 Thesis objectives ...... 45

Chapter 2 Rnf8 deficiency impairs class switch recombination, spermatogenesis, and genomic integrity and predisposes for cancer...... 46

2.1 Abstract ...... 47

2.2 Introduction…………………………………………………………………………...…48

2.3 Results...... 51

2.3.1 Generation of Rnf8 knockout mouse models………………………………….……52

2.3.2 Hypoxic conditions ameliorate defective growth of Rnf8-deficient MEFs………...53

2.3.3 Rnf8 deficiency leads to growth retardation and lymphopenia………………….….56

2.3.4 Rnf8 loss impairs spermatogenesis but not ovogenesis ……………………………59

2.3.5 53bp1 is partially recruited to DNA DSB sites via Rnf8-independent mechanisms in

activated B cells…………………………………………………………………………..63

2.3.6 Rnf8 deficiency leads to increased radiosensitivity both in vitro and in vivo……….67

2.3.7 Rnf8-deficient B cells are intrinsically defective for CSR…………………………72

2.3.8 Rnf8 suppresses spontaneous and IR-induced genomic instability………………...80

2.3.9 Rnf8 is a novel tumor suppressor…………………………………………….……..83

2.4 Discussion ...... 88

2.5 Materials and Methods………………………………………………………………..…92

2.5.1 Mouse genotyping and genetic analysis…………………………………………….92

2.5.2 Flow cytometry……………………………………………………………………..93

2.5.3 In vitro MEF proliferation assay under oxic and hypoxia conditions………………93

2.5.4 In vitro activation of B cells and lymphocyte analysis………………………….….94

2.5.5 Immunofluorescence……………………………………………………………….94

2.5.6 Digestion-circularization PCR……………………………………………….…...... 95

2.5.7 Antibody detection by ELISA……………………………………………….....…...96

2.5.8 Protein analysis and antibodies……………………………………………….…….96

2.5.9 BM colony-forming assay……………………………………………….…………97

2.5.10 In vitro and in vivo sensitivity to IR……………………………………………….97

2.5.11 Chromosomal aberration analysis………………………………………………....97

2.5.12 Southern blot analysis ………………………………………………………..…...98

2.5.13 Histology analysis……………………………………………………………..…..98

Chapter 3 Rnf8 coordinates Notch signaling and DNA repair to suppress breast cancer...... 99

3.1 Abstract ...... 100

3.2 Introduction…………………………………………………………………………….101

3.3 Results&Discussion...... 104

3.3.1 Rnf8 is expressed in mammary epithelial subpopulations...... 104

3.3.2 Rnf8 deficiency leads to an aberrant luminal expansion...... 107

3.3.3 Rnf8 loss increases DNA damage susceptibility and leads to focal hyperplasia in

mammary glands...... 111

3.3.4 p53 loss exacerbates Rnf8-deficiency associated aberrant luminal expansion...117

3.3.5 p53 inactivation accelerates Rnf8-mutation associated breast cancer

incidences...... 124

3.3.6 Rnf8 loss increases mitotic index and metastatic frequencies of p53 mutation

associated mammary tumors...... 125

3.3.7 Rnf8 complementation into Rnf8-/- p53∆/∆ mammary tumor cells restores their

vii

DSB signaling and repair efficiency...... 129

3.3.8 Rnf8-deficiency confers mammary tumors with hypersensitivity to PARP

inhibitor and IR treatments...... 134

3.3.9 Notch signaling is hyperactivated in Rnf8-/- luminal progenitors...... 137

3.3.10 Genetic restoration of Rnf8 into Rnf8-/- p53∆/∆ mammary tumor cells

suppresses their growth via dampening Notch signaling………………………….…141

3.3.11 Rnf8 promotes polyubiquitination and turnover of NICD …...………………142

3.3.12 Notch inhibitor treatment preferentially suppresses the growth of Rnf8-deficient

mammary tumors……………………………………………………………………146

3.4 Materials and Methods…………………………………………………………………149

3.4.1 Mouse genotyping and genetic analysis.…………………………………………..149

3.4.2 Vaginal smear cytology…………………………………………………………...150

3.4.3 Flow cytometry analysis and sorting of mammary epithelial cells……………….150

3.4.4 Whole mount staining……………………………………………………………..151

3.4.5 Histological analysis and immunohistochemistry………………………………...151

3.4.6 Realtime quantitative RT-PCR……………………………………………………152

3.4.7 Immunofluorescence……………………………………………………………...153

3.4.8 3-Dimensional Matrigel mammary colony forming……………………………....154

3.4.9 Protein analysis and antibodies……………………………………………………154

3.4.10 Mammary tumor cell lines, constructs and retrovirus infection………………...154

3.4.11 Clonogenic survival assay……………………………………………………….155

3.4.12 Transfections…………………………………………………………………….155

3.4.13 Southern blot analysis……………………………………………………………156

3.4.14 Cumulative growth analysis of tumor cells………………………………………156

viii

3.4.15 Notch/CBF1 dual luciferase reporter assay……….…….………………………..157

3.4.16 Nested step-down PCR analysis of WapCre-mediated p53 deletion……………..157

3.4.17 Cycloheximide chase analysis…………………………………………………...157

3.4.18 Intracellular ubiquitylation assay………………………………………………...158

Chapter 4 Summary and future directions...... 159

References ...... 167

List of Figures and Tables

Chapter 1: Introduction

Figure 1.1. Generation of DSBs from exogenous and endogenous causes.…………………….…..5

Figure 1.2. Antibody diversification in primary and secondary immune organs………………….10

Figure 1.3. The assembly of the MRN complex to sense the existence of a cellular DSB………13

Figure 1.4. A schematic depicting our current model of DNA DSB signaling network..………...18

Figure 1.5. Cellular choices to repair DSBs through NHEJ or HR repair machineries……………23

Table 1. Human syndromes associated with mutations in DDR genes and their clinical features...32

Figure 1.6. Cell types inside the mammary epithelial hierarchy…………………………………36

Figure 1.7. Schematics of three proteolytic events leading to the activation of Notch signaling….40

Chapter 2: Rnf8 deficiency impairs class switch recombination, spermatogenesis, and genomic integrity and predisposes for cancer

Figure 2.1 Generation of Rnf8-deficient mice……………………………………………………52

Figure 2.2 Rnf8-/- mice are growth retarded and display lymphopenia………………………..…54

Figure 2.3 Body weight measurements of Rnf8-/- females …….………………………………….56

Figure 2.4 Effect of Rnf8 deficiency on hematopoeitic cell populations………………………....58

Figure 2.5 Impaired spermatogenesis in Rnf8-/- mice……………………………………………..60

Figure 2.6 Partial recruitment of 53bp1 to the sites of DNA damage in the absence of Rnf8……..65

Figure 2.7 Rnf8-/- thymocytes and BM cells display increased radiosensitivity in vitro………...... 69

Figure 2.8 Rnf8-deficiency increases radiosensitivity in vivo….………………………………...71

Figure 2.9. A schematic depicting the in vitro class switch recombination functional assay…….74

Figure 2.10 Rnf8-/- B cells are intrinsically defective for IgH CSR……………………………...75

Figure 2.11 Additional inactivation of p53 does not rescue Rnf8-deficiency associated CSR defect. ……………………………………………………………………………………………77

Figure 2.12 IgH CSR in Rnf8+/- B cells is impaired and reflects a gene dose effect ………………79

Figure 2.13 Rnf8 deficiency leads to increased genomic instability..………………………….…82

Figure 2.14 Rnf8 mutant mice suffer a higher mortality rate..……………………………………84

Figure 2.15 Rnf8 is a novel tumor suppressor………………………………………………….…87

Chapter 3: Rnf8 coordinates Notch signaling and DNA repair to suppress breast cancer

Figure 3.1 Rnf8 is expressed in mammary epithelial subpopulations………………………..….105

Figure 3.2 Rnf8-deficiency promotes the luminal compartment of mammary glands to aberrantly expand through enhanced proliferation……………………………………………………...... 109

Figure 3.3 Rnf8-/- primary mammary epithelial cells display elevated levels of DNA breaks…..113

Figure 3.4 Loss of Rnf8 results in focal regional hyperplasia in mammary epithelium, and increases cancer risks……………………………………………………………………………………...115

Figure 3.5 Deletion of p53 exacerbates the aberrant luminal expansion and breast cancer risks associated with Rnf8 mutation.………………………………………………………………….120

Figure 3.6 The aberrantly expanded luminal compartment in Rnf8-/- mammary epithelium is exacerbated in the absence of p53…………………………………………………………….…122

Figure 3.7 Rnf8 inactivation increases the frequency of mitosis in p53 mutation associated mammary tumors. ………………………………………………………………………………126

Figure 3.8 Rnf8 loss exacerbates the incidences of metastasis of p53 mutation associated mammary

xi tumors ………………………………………………………...... 128

Figure 3.9 Rnf8-deficient mammary tumors display defective DSB signaling and elevated levels of spontaneous DNA breaks. …………………………………………………………………...131

Figure 3.10 Complementation of exogenous Rnf8 into Rnf8-/-;p53Δ/Δ mammary tumor cells restored their ability to repair H2ax-marked DNA breaks……………………………………...133

Figure 3.11 Rnf8-deficient mammary tumors display defective DSB repair pathways and hypersensitivity to PARP inhibitors and -radiation…………………………………………….135

Figure 3.12 Rnf8 suppresses mammary tumor growth via negative regulation of Notch signaling……………………………………………………………………………………...…138

Figure 3.13 Hyperproliferation of Rnf8-/- luminal differentiated cells is probably Notch- independent……………………………………………………………………………………..140

Figure 3.14 Rnf8 promotes polyubiquitylation and turnover of NICD, while its loss triggers hypersensitivity of mammary tumor cells to Notch inhibitors………………………………….144

Figure 3.15 A proposed model of Rnf8-mediated mammary tumor suppressor activity………………………………………………………………………………………..…148

xii

List of Abbreviations

AID………………………………………………………..Activation induced cytidine deaminase AP…………………………………………………………………...... Apurinic/apyrimidinic site AR………………………………………………………...………………………….Amphiregulin AT…………………………………………………………………………….Ataxia telangiectasia ATM………………………………………………………………..Ataxia Telangiectasia Mutated ATR………………………………………………………Ataxia telangiectasia and RAD3-related BER………………………………………………………...………………….Base excision repair BLM…………………………………………………………………………...…Bloom syndrome BM……………………………………………………………………………………Bone marrow bp………………………………………………………………………………………...Base pairs BRCA1…………………………………………………………………Breast cancer 1, early onset BRCA2…………………………………………………………………Breast cancer 2, early onset BRCT………………………………………………………………….BRCA1 C-terminus domain BSA…………………………………………………………………………Bovine serum albumin CDKs…………………………………………………………………...Cyclin Dependent Kinases cDNA……………………………………………………...complementary Deoxyribonucleic acid CHK2………………………………………………………………………….Checkpoint kinase 2 CFU…………………………………………………………………………...Colony forming unit CSR……………………………………………………………………Class switch recombination CTIP………………………………………………...C-terminal binding protein interacting protein CS…………………………………………………………………………………Catalytic subunit DAPI……………………………………………………………….4',6-diamidino-2-phenylindole DDR……………………………………………...………………………...DNA damage response DMEM……………………………………………………….Dulbecco’s modified eagles medium DMSO…………………………………………………………………………..Dimethylsulfoxide DNA………………………………………………………………………...Deoxyribonucleic acid DNA-PKcs……………………………………….DNA-dependent protein kinase catalytic subunit DNA-PK……………………………………………...DNA-dependent protein kinase holoenzyme DSB………………………………………………...……………………DNA double strand break dsDNA……………………………………………………………………….double stranded DNA DNase I………………………………………………………………………..Deoxyribonuclease I EDTA……………………………………………………………..Ethylenediaminetetraacetic acid ELISA………………………………………………………..enzyme linked immunosorbent assay FA...... Fanconi’s anaemia FACS…………………………………………………………..Fluorescence-activated cell sorting FCS………………………………………………………………………………...Fetal calf serum GFP…………………………………………………………………….Green fluorescence protein Gy……………………………………………………………………………………………...Gray HR………………………………………...…………………………..Homologous recombination hr…………………………………………………………………….………………………...Hour H2A……………………………………………………………………………………..Histone 2A H&E………………………………………...……………………………...Hematoxylin and eosin

xiii

HCl………………………………………………………………………………Hydrochloric acid IL-4…………………………………………………………………………………....Interleukin 4 IP……………………………………………………………………………..Immunoprecipitation IRIF…………………………………………………….Ionizing radiation-induced subnuclear foci IR………………………………………………………………………………...Ionizing radiation IHC…………………………………………………………………………Immunohistochemistry IF……………………………………………………………………………..Immunofluorescence IB………………………………………………………………………………………Immunoblot kDa……………………………………………………………………………………..Kilo Dalton kb………………………………………………………………………………………...Kilo bases LFS………………………………………………………………………..Li Faumeni’s Syndrome LN………………………………………………………………..……………………Lymph node MDC1……………………………………………………………...Mediator of DNA checkpoint 1 MEF………………………………………………...………………...Mouse embryonic fibroblast min…………………………………………………………………………………………Minutes MRE11…………………………………………………………..Meiotic recombination 11 protein MRN……………………………………………………………..MRE11-NBS1-RAD50 complex MMR………………………………………………………………………………Mismatch repair MgCl2…………………………………………………………………………Magnesium chloride mRNA…………………………………………...…………………….Messenger ribonucleic acid M……………………………………………………………………..Moles per litre; marker (lane) MRN……………………………………...…………………………...MRE11, RAD50 and NBS1 nAChR…………………………………………………………....nicotinic Acetylcholine receptor NBS1……………………...………………………………Nijmegen breakage syndrome protein 1 NBS…………………………………………………………………Nijmegen breakage syndrome NER……………………………………………………………………..Nucleotide excision repair NICD………………………………………………………………..…Notch intra-cellular domain NHEJ………………………………………………………………...Non-homologous end joining O2…………………………………………………………………………………………...Oxygen ORF……………………………………………………………………………Open reading frame PAGE………………………………………………………….Polyacrylamide gel electrophoresis PARP……………………………………………………………….Poly(ADP)-ribose polymerase PARPi………………………………………………………………………………PARP inhibitor PBS………………………………………………………………………Phosphate buffered saline PCR……………………………………………………………………..Polymerase chain reaction PCNA…………………………………………………………….Proliferating cell nuclear antigen PI…………………………………………………………………………………Propidium iodide PTM………………………………………………………………..Posttranslational modifications PTIP………………………………………………..Pax transactivation domain-interacting protein p53………………………………………………………………………………Tumour protein 53 RAD51……..………………………………………………………………………….Radiation 51 RB………………………………………………………………………………….Retinoblastoma RIDDLE……...Radiosensitivity, immunodeficiency, dysmorphic features and learning difficulties RFP...... Red fluorescence protein RNA………………………………………………………………………………Ribonucleic acid RNF8…………………………………………Ring (Really interesting new gene) Finger Protein 8 RNF168……………..……………………………………………………...Ring finger protein 168 ROS……………………………………………………………………….Reactive oxygen species

xiv

RT-PCR………………………………………….Reverse transcription-polymerase chain reaction SD……………………………………………………………………………….Standard deviation SDS………………………………………………………………………...Sodium dodecyl sulfate SEM………………………………………...…………………………..Standard error of the mean ssDNA………………………………………………………………………..single stranded DNA TCR……………………………………………………………………………….....T cell receptor UBC13…………………………………………………………...ubiquitin conjugating enzyme 13 UV……………………………………………………………………………………….ultraviolet WB…………………………………………………………………………………….Western blot WT………………………………………………………………………………………..Wild type H2AX…………………………..…………Histone variant H2A.X phosphorylated at serine-139 μg……………………………………………………………………………………….Microgram μM……………………………………………………………………………………...Micromolar 53BP1………………………………………………………………………...p53 binding protein 1 7AAD…………………………………………………………...…………7-Aminoactinomycin D

List of Publications

Li L., Guturi K.K.N., Bissey P.A., Jackson H.W., Khokha R., Sanchez O., Hakem A., Hakem R. (2015). Rnf8 coordinates Notch signaling and DNA damage repair to suppress breast cancer. (In submission).

Halaby M., Hakem A., Li L., Ghamrasni S.E., Venkatesan S., Hande M. P., Sanchez O., Hakem R. (2013). Synergistic interaction of Rnf8 and p53 in the protection against genomic instability and tumorigenesis. PLoS Genetics. 9(1):e1003259.

Bohgaki T., Bohgaki M., Cardoso R., Panier S., Zeegers D., Li L., Stewart G.S., Sanchez O., Hande M.P., Durocher D., Hakem A., Hakem R. (2011). Genomic instability, defective spermatogenesis, immunodeficiency and cancer in a mouse model of the RIDDLE syndrome. PLoS Genetics. 7(4): e1001381.

Ghamrasni S.E., Pamidi A., Halaby M., Bohgaki M., Cardoso R., Li L., Venkatesan S., Sethu S., Hirao A., Mak T.W., Hande M.P., Hakem A., Hakem R. (2011). Inactivation of Chk2 and Mus81 leads to impaired lymphocytes development, reduced genomic instability and suppression of cancer. PLoS Genetics. 7(5): e1001385.

Li L., Halaby M., Hakem A., Cardoso R., Ghamrasni S.E., Harding S., Chan N., Bristow R., Sanchez O., Durocher D., Hakem R. (2010). Rnf8 deficiency impairs class switch recombination, spermatogenesis, and genomic integrity and predisposes for cancer. Journal of Experimental Medicine. 207(5): 983-97.

xvi

Chapter 1

Introduction

1.1 Origins of DNA double-strand breaks

More than half a century ago, Watson and Crick unraveled the DNA double-helix structure and provided a structural foundation to our understanding of how genetic information has been passing on from generation to generation. This structural insight also contributed to the birth of a fundamental scientific question: how do cells maintain the fidelity of all the genetic information encoded inside their DNA double-helices, in the constant presence of challenges imposed by a broad spectrum of DNA damage sources. Among the different types of DNA lesions, DNA double stand breaks (DSBs) are probably the most devastating lesions to mammalian cells, for they can not only trigger cell death but also introduce mutagenic lesions. Although numerous breakthroughs have advanced our knowledge and appreciation of the existence of a highly sophisticated and regulated network of mechanisms for safeguarding the fidelity of our hereditary “recipe” of life, new players have been constantly revealed and integrated into the complex landscape of DNA damage surveillance network. Interestingly, evolution has been refining and optimizing the signaling cascades inside this network, in an effort to tailor specific signaling pathways against specific types of DNA damages under specific cellular states. The sources of undesired DNA double stand breaks can be broadly classified into two major categories, endogenous and exogenous DNA damage agents. Given the numerous sources of DNA damages and the complexity of the mechanisms for repairing these genetic lesions, I will focus on describing the most common DNA damage types, their major mechanisms of repair, and the most frequent consequences following the improper repair of these harmful genetic lesions at both the cellular and organism levels.

1.1.1 Endogenous and exogenous sources of cellular DSBs

Among the broad spectrum of different types of DNA lesions, DSBs are particularly dangerous because they result in physical breakage of both strands of the DNA duplex (Price and D'Andrea,

2013). The generation of DSBs involves the breakage of the two complementary strands of the

DNA double helix at sites that are in close proximity such that base-pairing and chromatin structure become insufficient to hold the two DNA ends juxtaposed (Mehta and Haber, 2014). DSBs can occur both endogenously and exogenously from various sources. For example, DSBs can arise through DNA replication-fork collapse, during the processing of DNA interstrand crosslinks (ICLs) or following exposure to ionizing radiation (IR) (Figure 1.1).

IR can induce DNA single strand breaks (SSBs) by generating radiolysis radicals that cleave the sugar-phosphate backbone, and can also lead to extensive base modifications including oxidation, deamination, alkylation and loss of base residues (Ciccia and Elledge, 2010; Hoeijmakers, 2001).

Interestingly, IR may also, to a less extent, directly generate DSBs especially at high doses of irradiation when two SSBs are generated simultaneously in complementary DNA strands within a few helical turns (Mehta and Haber, 2014). Furthermore, it has been shown that IR can generate clustered oxidative base damages and SSBs on complementary strands of the same DNA molecule within a few helical turns (Sutherland et al., 2000; Sutherland et al., 2002). These clustered DNA lesions may subsequently result in DSBs through replication fork collapse (Sutherland et al., 2000;

Sutherland et al., 2002). DSBs are considered to be a form of complex DNA damage, and are among the most dangerous DNA lesions to cells in terms of lethality and risk of mutagenesis

(Ciccia and Elledge, 2010; Hoeijmakers, 2001). Interestingly, different types of ionizing radiation sources tend to generate DNA lesions via different mechanisms. For instance, exposure to gamma-

3 radiation can result in cellular DNA lesions by direct damaging DNA helices or indirectly through the generation of reactive oxygen species (Barcellos-Hoff et al., 2005; Bertram and Hass, 2008;

Klaunig et al., 2010). In the case of X-rays, however, the majority of DSBs caused by exposure to this radiation source results indirectly through the damaging effects of reactive oxygen species such as hydroxyl radicals, which are generated from the ionization of other molecules including water residing in close proximity to the DNA helices (Barcellos-Hoff et al., 2005; Bertram and

Hass, 2008; Klaunig et al., 2010).

On the other hand, normal endogenous cellular processes including oxidative phosphorylation by mitochondria, breakdown of long chain fatty acids by peroxisomes, and the inflammatory response can also give birth to free radical species such as reactive oxygen species, which can lead to the generation of SSBs (Berquist and Wilson, 2012; Hussain et al., 2003). If cells do not repair these

SSBs properly in a timely fashion, DSBs can be created when replication forks encounter these

SSBs and collapse, leading to the generation of mostly one-ended DSBs (Alexandrov et al., 2013).

Notably, besides IR, chemical compounds also represent an exogenous source that is capable of inducing DSBs. For example, anti-cancer chemotherapeutic drugs, such as cross-linking agents cisplatin and mitomycin C, can lead to the generation of DSBs through replication fork stalling

(Mehta and Haber, 2014). Additionally, topoisomerase inhibitors such as etoposide can trap covalently linked topoisomerase-DNA cleavage complexes, resulting in the formation of DSBs

(Koster et al., 2007). In summary, cellular DSBs can be generated through both exogenous and endogenous causes.

Figure 1.1. Generation of DSBs from exogenous and endogenous causes. Exposure to exogenous agents such as ionizing radiation leads predominantly to the generation of SSBs but can also, to a less extent, directly generate DSBs. On the other hand, DSBs can also be generated endogenously. For example, reactive oxygen species (ROS) produced by normal cellular metabolism can lead to the generation of SSBs, which can then result in the formation of one- ended DSBs when replication forks encounter template strand nicks and collapse. Furthermore, one-ended DSBs can also be generated when a replication fork stalls at an interstrand crosslink

(ICL) in S phase of the cell cycle. Moreover, two-ended DSBs can be generated directly following

IR exposure or indirectly from processing of converging replication forks following stalling at an

ICL. Endonucleolytic incision and translesion synthesis contribute to the generation of DSBs during the processing of stalled replication forks. DSBs can be structurally categorized into one ended DSBs (navy blue boxes) or two ended DSBs (yellow boxes). Green arrows: replication directions. Green color: newly synthesized DNA. Blue color: template DNA.

1.1.2 Genetically programmed generation of cellular DSBs

Although a single DNA double strand break is sufficient to kill a cell and can give rise to dangerous mutagenic lesions, evolution has weighed the benefits over the risks, and purposely employed

DNA double strand breaks as intermediates of several important physiological mechanisms for the host individual survival.

Meiosis: Meiotic crossover utilizes DSB-induced recombination in order to generate genetic diversity. This induction of DNA DSBs is mediated through the transesterase SPO11 during meiotic prophase, and is prior to stable pairing and synapsis of homologous chromosomes

(Kleckner, 1996; Mahadevaiah et al., 2001). Laterally along the chromosomal axes, a multi- module protein complex comprised of SYCP2, SYCP3 (Offenberg et al., 1998; Schalk et al., 1998), components of the cohesion complex (Eijpe et al., 2000; Eijpe et al., 2003), and HORMAD1 and

HORMAD2 (Fukuda et al., 2010; Wojtasz et al., 2009) is assembled. These lateral elements are further linked by a central complex comprised of SYCP1, SYCE1, SYCE2, and TEX12 (Costa et al., 2005; Hamer et al., 2006; Yang and Wang, 2009). In the meantime of generation of synaptonemal complex, SPO11-induced meiotic DSBs are poised for recombination. The X and

Y chromosomes of male mammals represent a unique pair that is capable of synapse solely in their short pseudoautosomal regions at the periphery of the nucleus. Notably, the histone variant H2ax is phosphorylated by the checkpoint kinase ATR at S139, leading to -H2ax formation that symbolizes the initiation of meiotic sex chromosome inactivation (MSCI). MSCI silences the XY chromatin and gives rise to the XY body (Fernandez-Capetillo et al., 2003; Turner et al., 2004).

Interestingly, the generation of -H2ax is not only the earliest known marker of MSCI, but also represents the earliest histone modification of somatic cells in response to DNA damage from various sources (Rogakou et al., 1998; Solovjeva et al., 2007). Although the stepwise events that promotes this transcriptional silencing is not clearly understood, there exists a strong correlation among several key observations including the coating of unsynapsed chromosomal axes by

HORMAD1 and HORMAD2, recruitment of ATR, phosphorylation of H2AX at S139, as well as the transcriptional silencing. Notably, recent studies reveal that the HORMAD1 and HORMAD2 proteins are a prerequisite to the efficient accumulation of ATR on the XY body (Daniel et al.,

2011; Wojtasz et al., 2012). While DSB repair proteins have also been found to be recruited to the

XY body, other histone modifications including ubiquitylation, sumoylation, and methylation are evidently present at the XY body (Inagaki et al., 2010). Possibly owing to stalled DSB repair in regions of synapse, the homologous recombination proteins RAD51 and DMC1, that generate filaments on single-stranded resected DNA ends of DSB via DNA-dependent ATPase activity, are persistently present on meiotic DSBs of the X but not the Y chromosome (Moens et al., 1997;

Moens et al., 2002; Pittman et al., 1998; Plug et al., 1996; Tarsounas et al., 1999). Overall, meiosis employs programed generation of DSBs as an efficient means to generate genetic variation in offspring and thus genetic diversity in human populations.

VDJ recombination: Besides meiotic recombination of germline cells, somatic cells specialized for building up adaptive immunity, have also evolved to utilize DSBs as intermediates to achieve optimized host immune defense system against foreign pathogenic invasions. Inside the primary immune organs bone marrow and thymus, a process termed V(D)J recombination occurs in the immunoglobulin(Ig) H and IgL (consisted of Ig and Ig) chains of immature B cells as well as T cell receptor (TCR) loci ( and ;  and ) of thymocytes, respectively. Recombination activating

7 genes 1 and 2 (RAG1 and RAG2) specialize in recognizing recombination signal sequences (RSSs) that flank each V, D, and J segment by binding to 12RSS or 23RSS, which are non-conserved gap sequences of either 12 or 23 base pairs located in between conserved heptamer and nonamer sequences (Nishana and Raghavan, 2012; Swanson, 2004). To achieve efficient recombination, synapsis takes place exclusively on paired complexes of alternate 12RSS/RAG or 23RSS/RAG complex. The RAG proteins then nick and cleave 5’ of the RSS, generating a hairpin-closed coding end as well as a blunt signal end that triggers non-homologous end joining (NHEJ) for the recombination of a D to a J segment, a V to a DJ, or a V to a J segment (Nishana and Raghavan,

2012). Upon completion of V(D)J recombination, different combinations of numerous V(variable),

D (diversity), and J (joining) gene segments dramatically increases the diversity of lymphocyte antigen receptor repertoires (Figure 1.2).

Class switch recombination: Following V(D)J recombination in the bone marrow, immature

IgM+ B lymphocytes migrate to secondary organs including spleen and Peyer’s patches. Upon encountering foreign pathogenic antigens, germinal center B lymphocytes inside these secondary immune organs undergo two important mechanisms of DNA alterations, namely Class switch recombination (CSR) and somatic hypermutation (SHM) in an effort to switch the effector function of immunoglobulin specifically against the specific type of pathogen as well as increase the binding affinity of the antibodies produced to the cognate epitopes on the pathogen (Alt et al.,

2013). The immunoglobulin heavy chain locus harbors multiple constant region genes, each possessing effector function optimized against a specific class of foreign pathogens. These IgH constant region exons are preceded by G-rich repetitive switch sequences (RSSs), which are target sites for AID. AID is programmed to generate DNA DSBs at the donor S sequence and downstream acceptor switch sequence in a transcription-dependent manner, leading to a

8 recombination-deletion event that results in the expression of specific IgH isotypes and the deletion of the intermediate sequences (Keim et al., 2013; Xu et al., 2012). There are eight IgH constant region exons organized in IgH locus in the sequence of 5’-V(D)J-C-C-C3-C1-C2b-C2a-C-

C-3’ that encodes IgM, IgD, IgG3, IgG1, IgG2b, IgG2a, IgE and IgA, respectively (Alt et al.,

2013). For instance, the switch from the pentameric low affinity antibody IgM to IgE, effectively for combating parasites; to IgA, protecting the mucosal sites; to high affinity IgG1, for eradication of pathogenic bacteria and viruses (Figure 1.2).

Figure 1.2. Antibody diversification in primary and secondary immune organs. RSSs:

Recombination signal sequences for VDJ recombination, or repetitive switch sequences for class switch recombination. C: constant region; IgH: immunoglobulin heavy chain; variable (V), diversity (D) and joining (J) gene segments.

1.2 Signaling and repair of DNA DSBs and consequences of defective DSB repair

1.2.1 Intracellular sensing of DNA damage and signal propagation

To face the challenges imposed by cellular DSBs, mammalian cells have evolved sophisticate d mechanisms to sense the physical existence of such dangerous genetic lesions as well as to remodel chromatins adjacent to the DSB sites within seconds in order to efficiently signal the recruitments of proper repair machineries. The MRN complex, composed of two MRE11, two RAD50, and two

NBS1 molecules, senses the presence of cellular DSBs (Figure 1.3) (Kobayashi et al., 2002;

Matsuura et al., 2004; Stracker and Petrini, 2011). Together, MRE11, NBS1 and the Walker A and

B domains of RAD50 form the large central globular domain of the MRN complex, which mediates the binding to the broken DNA ends (Stracker and Petrini, 2011). This MRN complex- mediated DNA binding activity usually occurs in the context of a higher-order assembly, and is mainly dependent on the activities of MRE11 and RAD50, though some studies indicate that NBS1 may also participate in DNA binding and have an impact on the DNA binding properties of MRE11 and RAD50 (Stracker and Petrini, 2011). Interestingly, X-ray crystallography studies suggest that the hook domains of RAD50 structurally contribute to the assembly of the MRN complex through zinc-dependent homodimerization (Stracker and Petrini, 2011). Further studies of the crystal structure of MRE11 bound to DNA indicate that MRE11 dimerization plays a critical role in DNA binding through the conserved domains in the N-terminus of MRE11, and this DNA binding activity by MRE11 is mediated by six DNA recognition loops and is dependent on contacting sugar-phosphate in the minor groove of DNA without any DNA sequence preference (Stracker 11 and Petrini, 2011). Besides DNA binding activity, MRE11 has also been suggested to prime the broken DNA ends for repair via its 3’-5’ double-strand DNA exonuclease, single-strand DNA endonuclease, and DNA unwinding activity (Shim et al., 2010; Stracker and Petrini, 2011).

Subsequently, the MRN complex directly recruits the ataxia-telangiectasia mutated (ATM) kinase to DSB lesions, and activates the kinase activity of ATM through the presentation of ssDNA to

ATM, leading to the dissociation of ATM dimer and their rapid autophosphorylation (Liu et al.,

2014; Polo and Jackson, 2011a; Shim et al., 2010; Stracker and Petrini, 2011).

Figure 1.3. The assembly of the MRN complex to sense the existence of a cellular DSB.

Following the generation of a cellular DSB, the MRN complex, composed of two NBS1, two

MRE11 and two RAD50 molecules, assemblies at the DSB site. This MRN complex senses the existence of this dangerous DNA lesion inside the host cell, and holds the broken DNA ends together. The amino-terminal and carboxyl-terminal regions of the RAD50 extended coiled-coil domains associate in an antiparallel fashion and fold back on themselves, leading to the generation of the RAD50 hook domain that mediates the assembly of the MRN complex at the DSB site.

Activated ATM phosphorylates histone H2AX on Ser139 (termed H2AX), followed by dephosphorylation of the constitutively phosphorylated neighboring residue Tyr142 (Ciccia and

Elledge, 2010; Cook et al., 2009). These modifications license the recruitment of the adaptor protein MDC1 to DSB sites via binding to H2AX (Jungmichel and Stucki, 2010). MDC1 is constitutively phosphorylated by casein kinase 2 (CK2), and this promotes its binding to NBS1 which also binds activated ATM, leading to transient anchoring of activated ATM at the DSB sites

(Ciccia and Elledge, 2010; Polo and Jackson, 2011b). As a result, these anchored activated ATM molecules further spread H2AX phosphorylation to neighboring nucleosomes, expanding the magnitude of the DNA damage signal. Additionally, a portion of soluble non-anchored activated

ATM contributes to H2AX spreading to distant chromatin regions and the capability of H2AX to “jump over” highly condensed heterochromatic regions, though they progressively lose their activity (Kim et al., 2007; Savic et al., 2009). Interestingly, activated ATM also phosphorylates

MDC1, and this leads to the recruitment of the adaptor protein MDC1 to the sites of DSBs (Bekker-

Jensen and Mailand, 2011; Huen et al., 2007; Kolas et al., 2007a; Mailand et al., 2007; Panier et al., 2012b). Furthermore, H2AX bound MDC1 initiates a positive ATM feedback loop that leads to the recruitment of more MRN complexes to the DSB sites via binding to NBS1 (Polo and

Jackson, 2011a; Stucki et al., 2005).

Subsequently, the E3 ubiquitin ligase RNF8, together with the E2 conjugating enzyme UBC13, via its FHA domain interacts with the TQXF (ATM-phosphorylated Thr-Gln-Xaa-Phe; Xaa denotes any amino acid) motifs on the BRCT domain of MDC1, and is thereby recruited to the sites of DSBs (Panier et al., 2012b). Through the E3 ligase activity inside its RING domain, RNF8 initiates a cascade of ubiquitylation on the histones H2AX and H2A surrounding the sites of DSBs

(Ciccia and Elledge, 2010; Huen et al., 2007; Kolas et al., 2007b; Mailand et al., 2007; Wang and

Elledge, 2007). Another ubiquitin ligase Ring Finger Protein 168 (RNF168) is then recruited to the DSB sites by recognizing ubiquitylation products of RNF8 via its N-terminal ubiquitin binding motifs MIU (motif interacting with ubiquitin)1 and MIU2. Subsequently, RNF168 in complex with UBC13 catalyzes a cascade of regulatory ubiquitylation events at the sites of DSBs, leading to an amplification of Rnf8-initiated DNA damage signals (Panier et al., 2012b); (Jackson and

Durocher, 2013; Panier and Durocher, 2009a). As a result, the ubiquitin conjugates initiated by

RNF8 are amplified at the DSB sites (Doil et al., 2009); (Bartocci and Denchi, 2013; Stewart et al., 2009). The formation of these polyubiquitin chains at the DSB sites leads to the recruitment of chromatin-associated DNA repair factors including BRCA1-RAP80, RAD18, PTIP and 53BP1

(Panier and Durocher, 2013; Polo and Jackson, 2011b). Additionally, recent studies reveal that a heterodimer of the E3 ligases BMI1 and RING1B also plays a role in DDR, and mediates mono- ubiquitylation of H2AX at K119 (Ismail et al., 2010; Pan et al., 2011). Interestingly, unlike

Rnf168, the recruitment of BMI1 to the sites of DSBs has been suggested to be independent of

Rnf8, implying a non-epistatic relationship between BMI1 and Rnf8 in the hierarchical recruitments of DDR molecules (Ismail et al., 2010). Despite its independence of Rnf8, accumulation of BMI1 into DNA-damage-induced foci co-localizes with H2AX foci and is needed for efficient recruitments of 53BP1 and BRCA1 (Ismail et al., 2010).

In addition to phosphorylation and ubiquitylation, other posttranslational modifications, including sumoylation, methylation, acetylation and neddylation, also play an important role in DDR and coordination of hierarchical recruitments of proteins to the sites of DSBs (Dou et al., 2010; Galanty et al., 2009; Morris et al., 2009). The SUMO E3 ligases PIAS4 and PIAS1 are recruited to DSB

15 sites, and lead to the accumulation of SUMO1, SUMO2, and SUMO3, facilitating the efficient recruitments of RNF168, BRCA1 and 53BP1 as well as the efficient formation of ubiquitin chains mediated by RNF8 (Galanty et al., 2009). Also, RNF168 and HERC2 have been recently reported to be modified with SUMO1 through the catalytic activity of the SUMO E3 ligase PIAS4 in response to DSBs (Danielsen et al., 2012). Furthermore, sumoylation of HERC2 has been suggested to be important for stabilization of RNF8-UBC13 complex in response to DSBs

(Danielsen et al., 2012). Moreover, while sumoylation of BRCA1 by PIAS SUMO E3 ligases has been shown to be required for BRCA1 ubiquitin ligase activity (Morris et al., 2009), RPA70 sumoylation promotes the loading of RAD51 to DSB sites and thereby HR-mediated repair (Dou et al., 2010). Besides sumoylation, RNF111-catalyzed histone neddylation has been recently shown to promote the binding of RNF168 to damaged chromatin, and thereby provides another activation mechanism for the propagation of Rnf8-initiated ubiquitination signaling cascade in response to DSBs (Ma et al., 2013). Additionally, acetylation of histone H4 by TIP60 together with the p400 ATPase relaxes the chromatin adjacent to DSBs sites and thereby facilitates RNF8- dependent ubiquitination (Xu et al., 2010). Recent studies suggest that acetylation of H2AX facilitates its polyubiquitination at K119 (Ikura et al., 2007), while acetylation of ATM promotes its activation so that indirectly contributes to ubiquitination at DSBs sites through recruiting

RNF8(Kaidi and Jackson, 2013; Sun et al., 2010).

Interestingly, 53BP1 accumulates at DSB sites through the binding of its tandem Tudor domains to dimethylated H4 Lys20 (H4K20me2) and ubiquitylated H2A Lys15 (Botuyan et al., 2006;

Fradet-Turcotte et al., 2013; Zgheib et al., 2009). Recent studies have further unraveled that the loading of 53BP1 onto H4K20me2 on the damaged chromatin is regulated by competitive binding

16 via RNF8-dependent ubiquitination to promote turnover of chromatin components. In the absence of DSBs, L3MBTL1 and JMJD2A constitutively bind to H4K20me2 and physically block the binding of 53BP1 to H4K20me2. The AAA+ ATPase, VCP (valosin-containing protein) removes

L3MBTL1 that is Lys48 ubiquitylated by RNF8 post DNA damage (Acs et al., 2011; Meerang et al., 2011). Likewise, following cellular DNA damage, the removal of JMJD2A occurs through

RNF8-dependent ubiquitination and the 26S proteasome (Butler et al., 2012b; Mallette et al.,

2012). Collectively, various types of post-translational modifications contribute to the orchestration of the proper signaling propagation in response to DSBs. Rnf8-initiated ubiquitination at the sites of DSBs plays a central role in the propagation of DNA damage signals and the orchestration of a hierarchical recruitment of downstream DDR molecules, leading to the proper assembly of cellular DNA repair machineries at the broken DNA ends.

Figure 1.4. A schematic depicting our current model of the DNA DSB signaling network. The

MRN complex senses the physical existence of DSBs inside cells and recruits autophosphorylated activated ATM to the sites of DSBs. Activated ATM phosphorylates histone H2AX on Ser139

(H2AX), marking the sites of DSBs. This phosphorylation event leads to the recruitment of the adaptor protein MDC1, which then results in the recruitment of more MRN complexes and thereby amplifies the H2AX phosphorylation surrounding the DSBs. Subsequently, RNF8 is recruited to the sites of DSBs via the interaction of its FHA domain with activated MDC1, phosphorylated on the TQXF (Thr-Gln-Xaa-Phe) motifs by activated ATM. It has been recently proposed that RNF8 ubiquitylates a yet unknown protein (denoted as “X”), leading to the subsequent recruitment of another E3 ligase RNF168. Together, RNF8 and RNF168 catalyze a cascade of regulatory chromatin ubiquitylation events on substrates including H2AX, H2A, JMJD2A and L3MBTL1, leading to DSB signaling propagation and the efficient assembly of components of NHEJ and HR repair pathways including 53BP1 and BRCA1, respectively, to repair DSBs. The sequential recruitments of all these DDR molecules to the sites of DSBs can be experimentally visualized using specific antibodies against these DDR proteins and immunofluorescence, in which a dot-like fluorescent signal (called a “focus”; green color here) inside the nucleus (stained with DAPI; blue color) indicates the accumulation of the DDR protein of interest (that the antibody used recognizes) at a DSB site. P: phosphorylation; Ub: ubiquitination; Me: methylation.

1.2.2 Choice of cellular machineries to repair DSBs

In order to correctly repair DSBs and maintain genomic stability, an Rnf8-initiated ubiquitylation cascade signals the recruitment of cellular DNA repair machineries to assemble at the DSB sites.

To avoid deleterious genome rearrangements, cell death, senescence or malignant transformatio n, mammalian cells have evolved two main repair pathways, non-homologous end-joining (NHEJ) and homologous recombination (HR), to repair DSBs and restore genome integrity (Figure 1.4).

NHEJ is an error-prone repair pathway to simply rejoin the broken ends of a DSB, and does not require a template and thereby operates throughout the cell cycle phases. Notably, because the ends of DSBs frequently contain abasic sites and end-blocking adducts such as 5’ hydroxyl groups,

3’ phosphates or 3’ phosphoglycolate groups (Lieber, 2010), DNA end processing becomes a prerequisite to rejoin two broken DNA ends and inevitably leads to small deletions or insertions at the DSB sites (McVey and Lee, 2008). Interestingly, canonical NHEJ (cNHEJ) machinery does not employ any form of DNA sequence homology, and is initiated by binding of the KU70/80 heterodimer to the ends of DSBs. Remarkably, RNF8 has been shown to catalyze Lys48 linked polyubiquitin chains not only on itself and the checkpoint kinase 2 (CHK2), but also on KU80, the core component of NHEJ, leading to degradation and turnover of these proteins (Feng and Chen,

2012). RNF8 depletion leads to aberrantly prolonged retention of KU80 at DSB sites and thereby impairs the implementation of later steps of NHEJ (Feng and Chen, 2012). Subsequent to KU protein recruitment, DNA-PKcs, a PI3-kinase-like kinase, is recruited and activated to hold the broken ends of DSBs in close proximity. It also leads to the recruitment of end processing factors including Artemis, PNKP, APE1, TDP1 and WRN, that prime the two broken DNA ends of a DSB for re-ligation through the activity of the XRCC4-XLF-DNA ligase IV complex (Lieber, 2010).

Besides canonical NHEJ, recent studies have unraveled alternative end-joining pathways that function in the absence of one or more core components of the canonical NHEJ (cNHEJ) machinery (Boboila et al., 2012). For instance, while microhomology-mediated end-joining

(MMEJ), initiated by short-range end-resection, operate independently of KU70/80 and XRCC4-

DNA ligase IV complexes, it employs 5-25 nucleotides of homologous sequence between the two broken DNA ends of a DSB (Boboila et al., 2012; Grabarz et al., 2012). MMEJ pathway plays an important role in mediating the repair of CSR-associated DSBs (Boboila et al., 2012). Collectively, the biological importance of NHEJ is demonstrated by the various human syndromes associated with mutations in components of this repair pathway (see Table 1).

In contrast to NHEJ, HR-mediated repair of DSBs can be error-free, because it employs large segments of identical DNA sequences to serve as templates for DNA repair. Owing to the template requirement, HR-mediated DNA repair employs sister chromatids as repair templates, and is operational in the late S- and G2 phases of cell cycle (Heyer et al., 2010; Jasin and Rothstein,

2013). Since NHEJ is functional throughout all phases of the cell cycle, the critical determinant of homology-dependent repair is the initiation of 5’-3’ resection of DNA ends, which commits cells to HR-mediated repair and prevents the participation of NHEJ. The end-resection process through the coordinated actions of helicases and nucleases such as BLM helicase and Dna2 nuclease generates long stretches of 3’ single stranded DNA coated by RPA (Heyer et al., 2010; Symington and Gautier, 2011). Subsequently, RAD51, with the help from mediator proteins including RAD52 and BRCA2, displaces RPA to establish a RAD51/ssDNA nucleofilament (Heyer et al., 2010).

The RAD51 nucleofilament then searches for a homologous sequence in the genome, and undergoes synapsis to create a displacement (D)-loop. Within the D-loop, DNA synthesis restores the DNA surrounding the DSB sites. The D-loop is resolved either by synthesis-dependent strand

21 annealing or migrating Holliday junction intermediates cleaved by resolvases such as Mus81–

Eme1 (Heyer et al., 2010; Jasin and Rothstein, 2013). Similar to NHEJ, the biological importance of HR is also highlighted by the various human syndromes associated with mutations of components of this repair pathway (see Table 1).

The decision of repair pathway choice is highly regulated, and the initiation of 5’-3’ DNA end resection is of paramount importance. The extensive resection of the DNA surrounding the DSB is a prerequisite to initiate HR-mediated DNA repair, and, at the same time, completely inhibits

NHEJ-mediated DNA repair by impeding the ability of KU70/80 to bind to DSB sites (Dynan and

Yoo, 1998; Heyer et al., 2010; Symington and Gautier, 2011). Interestingly, the activity of 5’-3’

DNA end-resection nucleases is dependent on S- and G2-specific CDKs (Huertas et al., 2008;

Huertas and Jackson, 2009; Ira et al., 2004; Jazayeri et al., 2006; Yun and Hiom, 2009). Therefore,

HR is favored over NHEJ during S and G2 phase of the cell cycle, probably owing to the upregulation of 5’-3’ DNA end-resection during these phases of the cell cycle. On the other hand, due to the absence of a template, components of HR DNA repair machinery are blocked from being recruited to the DNA DSBs during G1 phase of the cell cycle by 53BP1, RIF1 and KU70/80

(Bothmer et al., 2010; Chapman et al., 2013b; Clerici et al., 2008; Daley and Sung, 2013; Feng et al., 2013; Lazzaro et al., 2008; Mimitou and Symington, 2010; Shim et al., 2010; Symington and

Gautier, 2011; Zhang et al., 2006). Remarkably, recent studies revealed that 53BP1 deficiency is able to rescue HR repair efficiency in BRCA1-deficient cells by increasing DNA end-resection, though the molecular mechanisms governing 53BP1 and BRCA1 regulated DSB repair choice is only partially understood and is currently under active investigation in the field (Chapman et al.,

2013a).

Figure 1.4. Cellular choices to repair DSBs through NHEJ or HR repair machineries.

Following DSB signal propagation, the cell has to choose the proper DNA repair machineries to

23 employ to rejoin the broken DNA ends. This choice of DNA repair pathways is of paramount importance, since NHEJ-mediated repair is error-prone but HR-mediated repair tends to be error- free. HR-mediated repair machinery requires the use of the sister chromatid as a template to repair the broken DNA ends, and therefore it is favored in S and G2 phases of the cell cycle. The cellular decision to perform 5’-3’ DNA end resection on the broken DNA ends facilitates the assembly of

HR-repair components at the sites of DSBs, but inhibits the recruitment of NHEJ-repair components during S and G2 phases of the cell cycle.

1.2.3 Negative regulation and post-repair termination of DNA damage signaling

At the DNA DSB sites, post-translational modifications and recruitment of DNA repair complexes play a central role to hold the broken DNA ends in close proximity, rejoin the genetic lesion, and restore stability to cellular genome. Just like other cellular processes, negative regulation plays an important role in fine-tuning and turning off DNA damage signals. The first limitation to the intensity of the DSB signaling response, including Rnf8-initiated ubiquitination signals, is the natural barrier formed by chromatin compaction (Murga et al., 2007). Interestingly, the formation of open chromatin simply by reducing of levels of the linker histone H1 to half of the wild-type results in an open conformation of chromatin, capable of inducing a stronger H2AX-marked DSB signaling response (Murga et al., 2007). Likewise, the short isoform of BRD4, the bromodomain- containing protein, restricts the amount of H2AX phosphorylation, and thus the downstream

24 ubiquitination signals initiated by Rnf8, via compacting the chromatin (Floyd et al., 2013).

Secondly, to further regulate the initiation of Rnf8-dependent ubiquitination cascade, the phosphorylation mark on Ser139 of H2AX can be reversed by multiple phosphatases including

PP2A, PP4, PP6 and PPM1D, though why so many tools to dephosphorylate one histone mark exist remains an open question, and probably contributes to the complexity of switching on and off of DNA damage signals in different cellular contexts (Cha et al., 2010; Chowdhury et al., 2005;

Chowdhury et al., 2008; Douglas et al., 2010; Macurek et al., 2010; Moon et al., 2010; Nakada et al., 2008). Thirdly, MDC1, the key scaffold for docking Rnf8 and its downstream DSB response proteins to the broken DNA ends of the chromatin, has been shown to be ubiquitylated by RNF4 for degradation and clearance from repaired chromatin, providing a mechanism to turn off Rnf8 - dependent ubiquitination and thereby reverse DSB response (Galanty et al., 2012; Luo et al., 2012;

Shi et al., 2008; Vyas et al., 2013; Yin et al., 2012).

Notably, this RNF8-initiated ubiquitination signal cascade has been recently revealed to be fine- tuned through various mechanisms, to directly regulate DSB-induced ubiquitination. Firstly, recent studies suggest that the deubiquitylating enzyme OTUB1 inhibits UBC13 and UBCH5, which are E2 conjugating enzymes of RNF8, in a non-catalytic fashion, suppressing the formation of RNF8 initiated polyubiquitin chains (Mattiroli et al., 2012). Interestingly, X-ray crystallographic studies have revealed that OTUB1, via binding to both ubiquitin and the E2 conjugating enzyme UBC13, prevents the entrance of ubiquitin into the E2 conjugating enzyme

UBC13’s active site as well as inhibits binding of the E2 conjugating enzyme UBC13 with the E3 ubiquitin ligase RNF8 (Wiener et al., 2012). Secondly, OTUB2 removes ubiquitin chains catalyzed by Rnf8 at the sites of DSBs, and promotes DNA end resection and thereby HR-mediated repair, fine-tuning the magnitude and speed of DSB-induced ubiquitination to allow a sufficient time-

25 window for choosing the proper DNA repair pathway (Kato et al., 2014). Thirdly, another deubiquitinating enzyme (DUB), POH1, inside the 19S proteasome antagonizes Rnf8-catalyzed ubiquitination and Rnf8-mediated degradation of JMJD2A at the DSB sites, and therefore blocking the binding of 53BP1 to H4K20me2 and limiting the magnitude of Rnf8-dependent DSB signaling response (Butler et al., 2012a). Fourthly, the ubiquitin hydrolase DUB3 interacts with H2AX and removes the Rnf8-catalyzed mono-ubiquitin on H2AX in response to DSBs, and thereby restrains the magnitude of DNA damage signals marked on histones (Delgado-Diaz et al., 2014).

Furthermore, USP44 and UPS3, two more DUBs, have been shown to promote removal of Rnf8- catalyzed ubiquitin chains on histone H2A (Mosbech et al., 2013; Sharma et al., 2014). Also, a recent study suggests that activated ATM in response to DSBs phosphorylates BCL10, which in turn facilitates the assembly of the Rnf8-UBC13 complex, providing another layer of regulation for the initiation of Rnf8-dependent ubiquitination at the sites of DSBs (Zhao et al., 2014). Another recent study indicates that the WD40 domain-containing protein WRAP53 promotes the docking of Rnf8 to MDC1 via its WD40 scaffold domain’s simultaneous binding to both Rnf8 and MDC1 in response to DSBs, and therefore further fine-tunes Rnf8-dependent ubiquitination signals

(Henriksson et al., 2014). Interestingly, RNF168, the ubiquitin signal booster of Rnf8-initiated ubiquitination cascade, has been shown to be targeted by two E3 ubiquitin ligases TRIP12 and

UBR5 for proteasomal degradation and thereby fine-tuning RNF8-dependent spreading of the

DNA damage signal, though potential negative regulators of Rnf8 protein abundance have not been discovered yet (Gudjonsson et al., 2012; Mattiroli et al., 2012; Wiener et al., 2012). Finally, the paralog of RNF168, RNF169, represents another mechanism to restrict the magnitude of Rnf8 - initiated ubiquitination in response to DSB, in which RNF169 binds strictly to RNF168- ubiquitylated histone H2A via its MIU-LRM domain and competes with other signaling and repair

26 proteins such as 53BP1 downstream of RNF8 to bind to DSB sites (Chen et al., 2012; Panier et al.,

2012a; Poulsen et al., 2012).

In brief, the assembly, operation, and disassembly of a DNA DSB induced focus employs an Rnf8- initiated ubiquitination signaling cascade that is tightly regulated and fine-tuned by a complex array of molecular mechanisms, in an effort to process the DSB at specific sites in a highly coordinated sequence of time events in order to properly repair the broken DNA ends.

1.2.4 The central role of p53 in DDR and cellular fate

If cellular DSBs are not properly repaired in a timely manner, they represent a significant risk for the survival of the host organism. Cancer can arise from improperly repaired DSBs (Hanahan and

Weinberg, 2011). For example, cellular DSBs may lead to chromosomal rearrangements, and potentially lead to the activation and/or deregulation of oncogenes or inactivation of tumor suppressors, providing driving forces for malignant transformation (Bieging et al., 2014b;

Hanahan and Weinberg, 2011; Meek, 2009). To minimize the cancer risk, cellular stress induced by DSBs can result in the stabilization and activation of p53, a short-lived transcription factor with capability to mediate innate tumor suppression (Bieging et al., 2014b; Hanahan and Weinberg,

2011).

The importance of inactivating the p53 protein during cancer development is probably best

27 demonstrated by the abundance of p53 mutations found in human tumors. Notably, more than half of all sporadic human cancers harbor mutations of p53 and/or alterations in genes encoding critical regulators of p53 (Bieging et al., 2014a; Meek, 2009). Consistent with the biochemical role of p53 as a transcription factor, approximately 80% of the p53 mutations detected in human tumors are missense mutations that reside within the DNA-binding domain of p53 (Bieging et al., 2014a;

Meek, 2009). These mutations can be categorized into contact mutations that disrupt p53 interaction with DNA and structural mutations that affect the 3-dimensional folding of the DNA- binding domain of p53 (Bieging et al., 2014a). Besides sporadic human cancers, germline mutations in p53 result in Li-Fraumeni syndrome. This autosomal dominant inherited disorder is well known to be associated with significantly increased cancer risks for a broad spectrum of tissues/organs including breast and muscle (Bieging et al., 2014a; Meek, 2009). In accordance with the human genome guardian role of p53, p53 knockout mice develop early-onset spontaneous tumors, primarily CD4+ CD8+ T cell lymphomas, with 100% penetrance (Freed-Pastor and Prives,

2012). Interestingly, genetic restoration of wild-type p53 expression in p53-deficient tumor cells significantly suppresses their tumor growth and promotes tumor regression, and thereby directly demonstrating the potency of p53-mediated tumor suppressive activity as well as the persistent addiction of p53-deficent cancer cells for the absence of p53 in spite of the presence of other secondary oncogenic mutations (Ciccia and Elledge, 2010; Ventura et al., 2007; Xue et al., 2007).

Therefore, p53 inactivation plays an important role in promoting cancer development in both human and mouse.

Given the importance of its biological function, p53 is tightly regulated by post-translational modifications including phosphorylation and ubiquitylation (Meek, 2009). Protein levels of p53 within unstressed cells are negatively regulated by the E3 ubiquitin protein ligase MDM2 (Bieging

28 et al., 2014b). Notably, MDM2 binds to the amino terminus of p53, and catalyzes ubiquitination of p53 to promote its degradation (Bieging et al., 2014b). Interestingly, there exists an autoregulatory feedback loop between MDM2 and p53, for MDM2 is a target gene of p53 (Khoo et al., 2014). In response to cellular DSBs, p53 can be activated through uncoupling from its negative regulators, leading to p53 stabilization (Bieging et al., 2014b). Previous studies indicate that activated ATM, generated in response to cellular DSBs, can phosphorylate p53 at serine 15, while CHK2, phosphorylated and activated by activated ATM, also has been shown to phosphorylate p53 at serine 20 (Bieging et al., 2014b; Meek, 2009). These key phosphorylation events of p53 play important roles in stabilizing p53 through disrupting the interaction between p53 and its negative regulators such as MDM2 and MDM4 (Bieging et al., 2014b). Besides p53 stabilization, these phosphorylation events can also promote the interaction of p53 with transcriptional cofactors, and thereby can modulate the activation of target genes and downstream responses (Bieging et al., 2014b; Zilfou and Lowe, 2009).

p53 can mediate its tumor suppressor function through various cellular processes including apoptosis, DNA repair, cell cycle arrest, senescence, autophagy, metabolic reprogramming, or tumor microenvironment signaling (Bieging et al., 2014a; Meek, 2009; Zilfou and Lowe, 2009).

For example, DSB-induced activation of p53 can result in a transient G1 cell cycle arrest by promoting transcription of p21, which inhibits CDK2 and CDK4 (Bieging et al., 2014b). Also, p53 can promote transcription of p21 and 14-3-3, leading to the activation of the G2/M checkpoint. In this case, p21 initiates the G2/M cell cycle arrest by preventing the phosphorylation of CDC2

(Hakem, 2008). 14-3-3 promotes the activation of Wee1, a negative regulator of CDC2, and thus blocks cell cycle entry into mitosis (Hakem, 2008). As a result, this p53-dependent activation of

DNA damage checkpoints can promote cell survival by allowing enough time to properly repair

29 the damaged DNA and thus prevent oncogenic genetic lesions (Meek, 2009). Furthermore, activation of p53 can also lead to apoptosis through upregulating expression of pro-apoptotic Bcl-

2 family members including Bax, Puma and Noxa (Bieging et al., 2014b; Meek, 2009). This p53- mediated apoptosis can eliminate cells harboring DSBs, and thereby can reduce the risk of accumulating oncogenic genetic lesions in the organism (Bieging et al., 2014b; Meek, 2009).

Additionally, p53 induction may also lead to senescence characterized by an irreversible permanent cell cycle arrest state, as well as autophagy, in which cellular components undergo controlled lysosomal degradation (Bieging et al., 2014b; Meek, 2009). Given the large number of genes regulated by p53, the biological outcomes of p53 induction are highly context dependent and can be influenced by factors including the cell type, the type and intensity of cellular stress and the genetic background (Bieging et al., 2014b; Meek, 2009). Collectively, p53 safeguards human and mouse against cancer development through various tumor suppressive mechanisms and serves as the guardian of the genome.

1.2.5 Human syndromes associated with defects in signaling and repair of

DSBs

Cellular DSBs are considered to be the most hazardous DNA lesions not only because of their capability to kill cells, but, more significantly, mis-repaired DSBs can also result in deleterious mutations that can serve as driving forces for a broad spectrum of human syndromes (Hakem,

2008; Jackson and Bartek, 2009). The spectrum of human diseases, caused by mutations of

30 molecules involved in signaling and repair of DSBs, is probably best exemplified by deficiency in the initiator of DSB signaling, ATM. Human ATM mutations lead to Ataxia telangiectasia (A-T), an autosomal recessive complex syndrome characterized by progressive neurological impairment, infertility, immunodeficiency, growth defects, premature ageing and predisposition to leukemia, lymphomas and breast cancer (Byrnes et al., 2008; Ghosal and Chen, 2013; Turnbull and Rahman,

2008).

Although ATM is a rare breast cancer susceptibility allele whose mutation has been observed in less than 1% sporadic breast cancer patients (Byrnes et al., 2008), A-T patients harboring mutated

ATM have an about 30-40% life-time risk for developing leukemia and lymphomas (Gumy-Pause et al., 2008; Gumy-Pause et al., 2004). Interestingly, cells from A-T patients display genomic instability and hypersensitivity to IR treatments (Lavin and Shiloh, 1997). Probably because cells harboring genetic alterations in DDR molecules all suffer from some extent of genomic instability, which is widely believed to be one of the major driving forces promoting cancer development, research results from the past decade have been highlighting the importance of the integrity of this dynamic signaling landscape to safeguard various human tissues against syndromes including cancer and immunodeficiency, as summarized in Table 1 below.

Table 1. Human syndromes associated with mutations in DDR genes and their clinical

features.

DDR Genetic Human Cancer Radio- Immuno- Infertility alteration syndrome predispo sensitivity deficiency sition Sense& NBS1 Nijmegen B cell Yes Yes Yes, females

Signaling breakage lymphom

syndrome a

MRE11 Ataxia N/A Yes Yes Yes

telangiectasi

a-like

disorder (A-

TLD)

RAD50 Nijmegen N/A Yes No N/A

breakage

syndrome-

like disorder

(NBSLD)

ATM Ataxia Breast Yes Yes Yes

telangiectasi cancer,

a (A-T) leukemia,

lymphom

RNF168 Riddle N/A Yes Yes N/A

Syndrome

TP53 Li-Fraumeni Breast No No No

syndrome cancer,

(LFS) brain

tumor,

sarcomas

NHEJ ARTEMIS Radiosensiti Lympho Yes Yes N/A

ve severe mas

combined

immunodefi

ciency (RS-

SCID)

RAG1/2 Severe N/A N/A Yes No

combined

immunodefi

ciency

(SCID)

XLF Immunodefi N/A Yes Yes No

ciency with

microcephal

LIG4 Ligase IV Lympho Yes Yes No

syndrome mas

Class AID, UNG Hyper-IgM N/A No Yes No switch syndrome recombinat ion (CSR)

HR BRCA1 Familial Breast Yes N/A No

breast and and

ovarian ovarian

cancer cancer

BRCA2 Familial Breast Yes N/A No

breast and and

ovarian ovarian

cancer cancer;

prostate

PALB2 Familial Breast N/A N/A N/A

breast and

cancer ovarian

cancer

RAD51C Familial Breast N/A N/A N/A

breast and

cancer ovarian

cancer

(N/A denotes conclusive data not yet available)

1.3 Mammary gland biology and breast cancer

1.3.1 Mammary gland development and differentiation

The mammary gland is a unique organ that mostly develops after birth (Visvader, 2009).

Increasing evidence suggest that mammary epithelial cells are organized in a hierarchical manner, in which a pool of pluripotent mammary stem cells give rise to committed progenitor cells of the two major lineages of the mammary epithelia: luminal and basal (also called myoepithelial) cells

(Figure 1.5). The luminal cells line the inner lumen of the mammary ductal-tree and constitute the alveolar units that secrete milk during pregnancy. On the other hand, basal myoepithelial cells are specialized contractile cells adjacent to the basement membrane, surrounded by fatty stromal matrix (Lim et al., 2009a). These mammary epithelial cells exhibit remarkable cellular plasticity, especially during puberty, pregnancy, lactation and involution (Hennighausen and Robinson,

2005; Kordon and Smith, 1998; Shackleton et al., 2006).

Notably, these different mammary epithelial cells all have the potential to become the cell of origin for breast cancer. Hispathologically, breast cancer can be subcategorized into two major types: luminal or basal-like breast cancers. About 70% to 80% of breast cancer cases are luminal, frequently estrogen receptor (ER) positive, whereas the remaining cancer incidences belong to basal-like breast cancers characterized by high histological grades and mitotic indices, negative status of ER and associated with a poor prognosis (Ades et al., 2014; Althuis et al., 2004; Visvader,

2011). Currently, our understanding of breast cancer etiology is still very limited, and requires a better understanding of mammary gland biology.

Figure 1.6. Cell types inside the mammary epithelial hierarchy. (A), a schematics depicting a hierarchical organization of discrete mammary epithelial cell types. (B), a schematics of a longitudinal cross-section of a mammary duct composed of the indicated different cell types. (This figure is modified from (Visvader, 2009))

Interestingly, a greater number of ovarian hormone-dependent reproductive cycles are associated with increased breast cancer risks (Apter et al., 1989; Bernstein, 2002; Trichopoulos et al., 1972).

This predisposition has been recently proposed to be due to progesterone which propels the mammary stem cell pool to expand by about 14-fold during its maximal levels at the luteal dioestrus phase of the mouse, probably through paracrine signaling from mature progesterone receptor-positive luminal cells (Asselin-Labat et al., 2010; Joshi et al., 2010). Interestingly, these studies also revealed that, accompanying the drastic expansion of the mammary stem cells, progesterone also promotes mammary luminal cells to expand, though to a less magnitude

(Asselin-Labat et al., 2010; Joshi et al., 2010). The maintenance of fidelity and stability of genetic information, under these continuous cycles of rapid cellular proliferation, differentiation, and apoptosis, present a significant challenge to the DNA damage surveillance network within mammary epithelial cells. Hence, genetic mutations of DDR molecules are often associated with breast cancer and are believed to contribute to drive malignant transformation of breast tissue. For instance, BRCA1 and BRCA2, two important DDR molecules, are well known to be breast tumor suppressors, and their mutations dramatically increase breast cancer risks for humans.

1.3.2 Notch signaling in mammary gland development and breast cancer

Notch signaling is critical for both homeostasis and oncogenic transformation of breast tissue

(Andersson and Lendahl, 2014; Bolos et al., 2013; Li et al., 2010b; Mittal et al., 2009; Pece et al.,

2004; Rizzo et al., 2008a; Stylianou et al., 2006; Yao et al., 2011a; Zardawi et al., 2010). The classical model of the activation of Notch receptor involves a series of proteolytic events upon binding to the Delta-Serrate-LAG2 (DSL) ligands (Figure 1.6). The first cleavage event by furin- like convertases acts in the extracellular/luminal domain near the transmembrane domain of nascent Notch receptor, leading to the generation of N- and C-terminal fragments (NTF and CTF, respectively). The generation of the mature Notch heterodimer is accomplished through a noncovalent linkage formation between NTF and CTF (Blaumueller et al., 1997); (Logeat et al.,

1998; Rand et al., 2000). Notably, the site of the second proteolytic cleavage in the extracellular region of the Notch CTF becomes available or exposed via DSL ligand-triggered conformational changes of Notch at the cell surface (Mumm et al., 2000). Metalloproteases, including ADAM10 and ADAM17/TACE, recognize and cleave the exposed site in the extracellular region of Notch, resulting in the removal of the Notch ectodomain (Brou et al., 2000; Lieber et al., 2002). After ectodomain removal, a membrane-anchored Notch CTF, named as Notch extracellular truncation

(NEXT), is generated. The final critical cleavage event occurs within the transmembrane domain of NEXT through the proteolytic activity of the intramembrane aspartyl protease -secretase (De

Strooper et al., 1999; Struhl and Greenwald, 1999), leading to the release of the intracellular domain of Notch receptor (NICD, activated Notch) from the cell membrane. Subsequently, NICD is translocated into the nucleus, though this translocation process is poorly understood. Once inside the nucleus, NICD participates with the DNA-binding protein Suppressor of Hairless (SU(H)) and

38 the nuclear effector Mastermind (MAM), activating transcription of target genes including Hes1/5,

Hey1/2, Gata3 and Cyclin D1 (Guruharsha et al., 2012). Interestingly, post-translational modifications including ubiquitination play a central role in regulating Notch signaling. The turnover and activity of NICD is controlled by a number of E3 ligases such as Fbw7, and deregulation of these ubiquitylation events have been associated with breast cancer development

(Andersson and Lendahl, 2014).

Figure 1.7. Schematics of three proteolytic events leading to the activation of Notch signaling.

After three proteolytic cleavage processes, the Notch intracellular domain (NICD or activated

Notch) is released from the plasma membrane and translocates into the nucleus via a poorly understood process. Once in the nucleus, the NICD interacts with CBF1-SU(H)-LAG1(CSL) proteins, and converts CSL into a transcription activator through replacing co-repressors with co- activators, leading to the transcriptional activation of Notch target genes.

In the murine mammary epithelia, hyperactivation of the Notch signaling pathway results in an aberrant expansion of luminal epithelial cells that progressively develop into hyperplasia and ultimately mammary adenocarcinoma (Bouras et al., 2008; Dievart et al., 1999; Imatani and

Callahan, 2000; Jhappan et al., 1992; Kiaris et al., 2004; Raafat et al., 2004; Robbins et al., 1992;

Smith et al., 1995; Visvader and Stingl, 2014). In accordance with all these mouse model studies, high Notch1 levels have been found in basal-like or luminal progenitor breast cancers, and correlate with poor clinical outcomes for human breast cancer patients (Lee et al., 2008a; Visvader,

2009). Interestingly, several studies have reported that early non-invasive stages of human breast cancer, such as usual ductal hyperplasia and ductal carcinoma in situ, harbor aberrant upregulation of components of Notch signaling pathway, suggesting hyperactivation of Notch signaling contributes to the initiation of human breast cancers (Farnie et al., 2007; Mittal et al., 2009;

Zardawi et al., 2010). Likewise, upregulation of components of Notch signaling was observed in human breast cancers, suggesting them as prognostic markers in breast cancer (Lee et al., 2008b;

Li et al., 2010b; Reedijk et al., 2007; Reedijk et al., 2005; Reedijk et al., 2008; Rizzo et al., 2008b;

Robinson et al., 2011; Yao et al., 2011b).

Furthermore, accumulation of the intracellular domain of Notch1 has been detected in a large variety of human breast carcinomas (Stylianou et al., 2006). Notably, while transformation of normal human breast epithelial cells can be achieved through solely increasing RBP-Jk-dependent

Notch signaling, inhibition of Notch signaling in human breast cancer cell lines restrains their aggressiveness and transformed phenotypes, implying inhibition of Notch signaling may be useful for treating breast cancer (Stylianou et al., 2006). Moreover, Numb, the negative regulator of NICD or Notch signaling, has been found to be degraded via ubiquitination and proteasomal degradation

41 and thus lost in over 50% human mammary carcinomas (Pece et al., 2004). Interestingly, pharmacological inhibition of Notch signaling suppresses the growth of Numb-deficient, but not

Numb-positive, breast tumors (Pece et al., 2004). Given that -secretase releases NICD from the membrane-tethered Notch receptor and triggers the activation of Notch signaling, -secretase inhibitors such as RO4929097 are currently being actively tested in clinical trials to determine their potential benefits of restraining Notch signaling to treat patients with breast cancer and other solid tumors (Wu et al., 2014; Tolcher et al., 2012). Collectively, aberrant hyperactivation of Notch signaling drives the development of breast cancer and represents a promising therapeutic target.

1.3.3 DNA damage response and breast cancer susceptibility genes

Despite a few decades of research effort to elucidate breast cancer etiology, our current knowledge of the genetic alterations driving breast cancer development remained limited and mostly centers on a few tumor suppressors and oncogenes. For instance, BRCA1 and BRCA2 mutations only account for about 25 percent of hereditary breast cancer incidences and about 10 percent of all breast cancer cases (Campeau et al., 2008; Easton, 1999). Similarly, about 30% of breast cancer incidences are associated with p53 mutations, with a higher frequency in some tumor subtypes and carriers of germline BRCA1 mutations (Lamlum et al., 1999; Olivier et al., 2006; Pharoah et al.,

1999; Rahman et al., 2007; Smith et al., 1999). Although mutations in other genes including ATM,

PTEN, CHEK2, PALB2, BRIP1 and CDH1, confer increased risks of breast cancer, their frequencies are rare and account altogether for less than 25% of familial breast cancer and less 42 than 1% of all breast cancer cases (Bell et al., 1999; Easton, 1999; Economopoulou et al., 2015;

Ghoussaini et al., 2013; Guilford et al., 1998; Lamlum et al., 1999; Maxwell and Nathanson, 2013;

Meijers-Heijboer et al., 2002; Nelen et al., 1996; Nichols et al., 1999; Noffsinger et al., 1999;

Rahman et al., 2007). Collectively, though decades of intense research effort to elucidate breast cancer susceptibility genes have advanced our knowledge of breast cancer development, future research effort is needed to elucidate all the breast cancer oncogenic driving mechanisms in order to rationally design therapeutic strategies to improve the outcomes of breast cancer patients.

To gain a better understanding of breast cancer, recent research effort has also attempted to determine whether 53BP1 another important DDR molecule is altered in specific subtypes of breast cancers as well as to elucidate the etiology of breast cancers harboring mutated BRCA1, the major breast cancer susceptibility gene. Notably, the protein expression of 53BP1 has recently been shown to be significantly reduced in subsets of sporadic triple-negative and BRCA1 mutation associated human breast cancers (Bouwman et al., 2010a; Grotsky et al., 2013). On the other hand, besides its critical role in the DNA damage response and suppressing breast cancer, BRCA1 has been shown to play essential roles in mammary gland development and differentiation (Bouras et al., 2008; Lim et al., 2009a; Liu et al., 2008; Proia et al., 2011; Xu et al., 1999). Conditional deletion of Brca1 in luminal epithelial cells in p53 deficient mice suggest luminal progenitors as the cell of origin for Brca1 deficient basal-like mammary tumors with the most resemblance to human BRCA1 mutation associated breast cancer (Drost and Jonkers, 2009; Molyneux et al.,

2010). More importantly, the knowledge that we have gained from studying the DDR network and breast cancer have led to the advent of the concept of synthetic lethality for treating BRCA1- deficient breast cancers. Remarkably, PARP inhibitors have been implicated in preferentially killing breast cancer cells deficient for BRCA1 or homologous recombination repair (Farmer et

43 al., 2005; Gilardini Montani et al., 2013; Lee et al., 2014; Lord and Ashworth, 2013). Given that the mutation spectrum of breast tumors frequently includes mutations of molecules important for repairing damaged DNA (Ghosal and Chen, 2013; Han et al., 2009; McCullough et al., 2014;

Smith et al., 2003), genomic instability triggered by dampened cellular DNA repair ability has been attributed as a hallmark for malignant transformation, and thus is often viewed as a mechanistic explanation for breast cancers harboring mutated DNA damage signaling and repair molecules (Ghosal and Chen, 2013). Because RNF8 has been recently demonstrated to be essential for the efficient recruitments of BRCA1 and 53BP1 to the sites of DSBs in response to DNA damage, it presents a potential candidate for novel breast cancer susceptibility genes and therefore needs further investigation in vivo.

1.4 Objectives

Global objective: To elucidate the physiological functions of Rnf8.

Specific objective 1: To characterize the potential impacts of Rnf8 inactivation on adaptive immune response, mammalian fertility, DNA damage response, genomic stability and cancer predisposition at the organism-level.

Specific objective 2: To elucidate the potential roles of Rnf8 in mammary gland biology and breast cancer development, and its potential collaboration with p53 in vivo.

Chapter 2

Rnf8 deficiency impairs class switch recombination,

spermatogenesis, and genomic integrity and predisposes for

cancer

This chapter is composed of the following published article, of which I performed all the experiments except Figure 2.1D, 2.2B and 2.7B.

Additionally, Figure 2.10 of this chapter, in which I performed the experiments, is from the following publication:

2.1 Abstract

Signaling and repair of DNA double-strand breaks (DSBs) are critical for preventing immunodeficiency and cancer. These DNA breaks result from exogenous and endogenous DNA insults but are also programmed to occur during physiological processes such as meiosis and immunoglobulin heavy chain (IgH) class switch recombination (CSR). Recent studies reported that the E3 ligase RNF8 plays important roles in propagating DNA DSB signals and thereby facilitating the recruitment of various DNA damage response proteins, such as 53BP1 and BRCA1, to sites of DNA damage. Using mouse models for Rnf8 mutation, we report that Rnf8 deficiency leads to impaired spermatogenesis and increased sensitivity to ionizing radiation both in vitro and in vivo. Interestingly, we also demonstrate the existence of alternative Rnf8-independent mechanisms that respond to irradiation and accounts for the partial recruitment of 53bp1 to sites of DNA damage in activated Rnf8-/- B cells. Remarkably, Rnf8 inactivation in B lymphocytes results in an intrinsic deficiency in IgH CSR, an important mechanism for antibody diversification and specification of immunoglobulin effector function during a mammalian humoral immune response. Furthermore, loss of a single allele of Rnf8 is sufficient to cause defective IgH CSR, and thereby rendering these mutant mice immunodeficient. Most importantly, Rnf8-/- mice exhibit increased genomic instability and elevated risks for tumorigenesis, indicating that Rnf8 is a novel tumor suppressor. These data unravel the in vivo pleiotropic effects of Rnf8.

2.2 Introduction

Mammalian cells have evolved sophisticated DNA damage signaling and repair mechanisms to prevent the accumulation or transmission of damaged DNA during cell divisions (Bartek and

Lukas, 2007; Harper and Elledge, 2007; Hoeijmakers, 2009; O'Driscoll and Jeggo, 2006). Among the various types of DNA damage, DNA double-strand breaks (DSBs) are the most detrimental to our cells. The importance of DSB signaling and repair mechanisms is demonstrated by the association of their defects with various human syndromes characterized by developmental defects, neurodegenerative disorders, immunodeficiency, and increased cancer predisposition

(Hakem, 2008; Hoeijmakers, 2009; O'Driscoll and Jeggo, 2006). In addition to DSBs generated by endogenous and exogenous DNA insults, DSBs are also programmed to occur in vivo during normal physiological processes, such as meiosis and during VDJ recombination in T and B cell development, in which the recombination process is essential to amplify the diversity for T and B cell receptor repertoires (Soulas-Sprauel et al., 2007). More importantly, Ig heavy chain (IgH) class switch recombination (CSR), which is one of the most critical mechanisms for antibody diversification in mammals, also involves programmed generation of DSBs initiated by activation- induced cytidine deaminase (Chaudhuri et al., 2007; Soulas-Sprauel et al., 2007; Stavnezer and

Schrader, 2014). The subsequent signaling and repair of these DSBs is required for peripheral B cells to successfully synapse and join these breaks and switch from expressing low-affinity IgM to various high affinity Ig isotypes, such as IgG1, IgE, and IgA, during an immune response. The joining of DSBs generated during the CSR process involves both classical and alternative nonhomologous end-joining pathways (Kotnis et al., 2009; Robert et al., 2009; Yan et al., 2007).

Interestingly, defects in the signaling or the repair of the CSR-associated DSBs inevitably result in immunodeficiency (Durandy et al., 2007; Kotnis et al., 2009).

The signaling of DSBs employs various DNA damage response (DDR) proteins and elaborate posttranslational modifications (PTM) including ubiquitylation, phosphorylation, methylation, and acetylation of chromatin and DDR proteins (Harper and Elledge, 2007; Panier and Durocher,

2009b). Within a few minutes after the generation of DSBs, subnuclear foci known as ionizing radiation (IR)–induced foci (IRIF) are assembled at the break sites (Wood and Chen, 2008). These

IRIF arise from chromatin remodeling and orchestrated recruitment of various DDR proteins, which are important for mediating the signaling and repair of the damaged DNA as well as cell cycle checkpoint activation or apoptosis. Phosphorylation of the histone variant H2AX at Ser139

(H2AX) is among the earliest PTMs required for the signaling of DSBs (Su, 2006). These early recruitment and PTM events at the damage sites provide important docking platforms to further enlist downstream DDR proteins. In addition to H2AX, several other DDR proteins, including

NBS1, MDC1, 53BP1, and BRCA1, are also phosphorylated by kinases such as ATM, ATR,

DNA-PK, Chk2, and Chk1. These phosphorylations provide important mechanisms for these DDR proteins to interact with each other at damage sites and to mediate the signaling and repair of

DSBs. The recent demonstration of the roles of the E3 ligases RNF8 and RNF168 in DSB signaling has highlighted the importance of regulatory ubiquitylation in the DNA damage signaling and repair processes (Doil et al., 2009; Huen et al., 2007; Kolas et al., 2007a; Mailand et al., 2007;

Panier and Durocher, 2009a; Pinato et al., 2009; Stewart et al., 2009; Doil et al., 2009; Huen et al.,

2007; Kolas et al., 2007a; Mailand et al., 2007; Panier and Durocher, 2009b; Pinato et al., 2009;

Stewart et al., 2009; Wang and Elledge, 2007). Both E3 ligases are required for the recruitment of

DDR proteins such as 53BP1, BRCA1, Rap80, Abraxas, and Brcc36 to DSBs.

After the initial recognition of H2AX by MDC1 at the site of DSBs, RNF8 is recruited to ATM- phosphorylated MDC1 through its FHA domain, and functions in a complex with the E2 ubiquitin conjugating enzyme Ubc13 to mediate mono-ubiquitylation of the histones H2AX and H2A at the chromatin-flanking DSBs (Huen et al., 2007; Kolas et al., 2007a; Mailand et al., 2007; Wang and

Elledge, 2007). These RNF8-dependent histone ubiquitylation events serve to recruit the RNF168–

Ubc13 complex to DSBs, leading to further K63 ubiquitylation of H2AX and H2A and thereby amplifying the DSBs signals (Doil et al., 2009; Pinato et al., 2009; Stewart et al., 2009).

Interestingly, knockdown cell lines for RNF8 or RNF168 and mouse embryonic fibroblasts

(MEFs) deficient for Rnf8 have been reported to abolish IRIF for 53BP1, BRCA1, and other downstream DDR proteins (Doil et al., 2009; Huen et al., 2007; Huen et al., 2008; Kolas et al.,

2007a; Mailand et al., 2007; Panier and Durocher, 2009a; Pinato et al., 2009; Stewart et al., 2009;

Wang and Elledge, 2007). These cells also exhibit increased radiosensitivity and impaired cell cycle checkpoints. More importantly, mouse models for mutations of Atm, H2ax, Mdc1, 53bp1, and Brca1 have demonstrated the roles of these signaling proteins in various processes including development, meiosis, CSR, genomic integrity, and cancer. Interestingly, germline mutations of the RNF168 gene have been associated with the human RIDDLE syndrome, which is characterized by defective 53BP1 foci formation, radiosensitivity, immunodeficiency, dysmorphic features, and learning difficulties (Stewart et al., 2009; Stewart et al., 2007).

In this study, we demonstrate that Rnf8-deficient mice are viable but show a growth retardation phenotype, increased sensitivity to IR both in vitro and in vivo, and impaired spermatogenesis.

Rnf8-/- MEFs show a growth defect that is partially rescued under hypoxic conditions. In contrast to Rnf8-/- MEFs, activated Rnf8-/- B cells show a partial recruitment and/or retention of 53bp1 to form IRIF, suggesting the existence of alternative signaling pathways that compensate for the absence of Rnf8 in activated B cells. We also show that CSR is significantly impaired in Rnf8 mutant mice, revealing that these mice are immunodeficient. Remarkably, loss of Rnf8 leads to increased genomic instability and predisposes the Rnf8 mutant mice for tumorigenesis, indicating that Rnf8 is a novel tumor suppressor. Thus, Rnf8 emerges as an important player in DSBs signaling through its essential roles in spermatogenesis, CSR, maintaining genomic integrity, and suppressing cancer.

2.3 Results

2.3.1 Generation of Rnf8 knockout mouse models

In order to examine the in vivo roles of Rnf8, two strains of Rnf8 mutant mice have been generated using independent gene trap embryonic stem cell (ES) clones (Figure 2.1A). Southern blot analysis confirmed the germline transmission of the two Rnf8 mutations (Figure 2.1B and C). Western blot analysis of serial two-fold dilutions of cell lysates indicated about a two-fold decrease of expression level of Rnf8 protein in heterozygote compared to Wild type (WT) background (Figure

2.1D). Further western blot analyses demonstrated loss of the full length Rnf8 protein in Rnf8-/-

51 thymocytes derived from the two different strains (Figure 2.1E). Rnf8 mutants derived from the two clones will be referred to as Rnf8 mutants.

Figure 2.1. Generation of Rnf8-deficient mice. (A) Schematic diagram of the mouse Rnf8 locus and its disruption by the gene trapping cassettes. For the AS0574 clone, the trapping cassette was inserted between exons 2 and 3 of the Rnf8 locus. For the RRR260 clone, the trapping cassette was inserted between exons 4 and 5. En2, Engrail homologue 2; SA, splice acceptor; PA,

52 polyadenylation signal; -geo, -galactosidase-neomycin resistance gene fusion. (B) Southern blot analysis of Bcl1-digested genomic DNA derived from the tails of the offspring of Rnf8+/- intercrossing for the AS0574 clone. (C) Southern blot analysis of BamH1-digested genomic DNA derived from the tails of the offspring of Rnf8+/- intercrossing for the RRR260 clone. (D)

Examination of Rnf8 protein level in Rnf8+/- mice. Protein samples from WT and Rnf8+/- thymocytes were serially diluted as indicated. Actin was used as the loading control. (E) Western blot analysis of Rnf8 protein expression in WT, Rnf8-/- (AS0574), and Rnf8-/- (RRR260) thymocytes. *, nonspecific band. Three independent experiments were performed.

2.3.2 Hypoxic conditions ameliorate defective growth of Rnf8-deficient

MEFs

To examine the effect of Rnf8 deficiency on in vitro cellular growth, we derived MEFs from day-

14.5 embryos generated from intercrossing of heterozygous mutant mice. Rnf8-/- MEFs of passage

2–5 proliferated at slower rates than WT littermate controls (Figure 2.2A). Furthermore, Rnf8-/-

MEFs of passage 6 and 7 displayed lower saturation densities than WT littermate controls.

Because it has been well documented that reactive oxygen species (ROS) play a role in generating endogenous DNA lesions (Bristow and Hill, 2008), hypoxic environment would offer a reduction in ROS-mediated assaults to chromatin. Therefore, we hypothesize that under hypoxic conditions,

53 the proliferation defect of Rnf8-/- MEFs would be alleviated owing to lesser DNA damage. As

-/- shown in Figure 2.2B, culturing of passage 4 Rnf8 MEFs for 12 d at 5% O2 concentration resulted in a greater proliferation rate than when they were cultured under normoxic conditions.

Interestingly, this increase in proliferation became increasingly evident from days 6 to 12 of in vitro culture. At day 12 of culture under hypoxic conditions, one of the independently derived lines of Rnf8-/- MEFs even showed a saturation density close to those of the WT littermate MEFs. These data indicate that Rnf8 deficiency in MEFs leads to defective proliferation. Furthermore, hypoxic conditions significantly rescue the impaired proliferation of Rnf8-/- MEFs, suggesting that ROS- induced DNA damage in Rnf8-/- MEFs contributes to their in vitro proliferative defects.

Figure 2.2. Rnf8-/- mice are growth retarded and display lymphopenia. (A) Rnf8-/- MEFs

54 exhibited reduced ability to proliferate in vitro. Passage 2 WT and Rnf8-/- MEFs were seeded in 6- cm dishes at a density of 3 × 105. After every 3 days, the MEFs were trypsinized, counted, and reseeded at the same density. The cumulative growth curves are shown. Three independent experiments were performed, and a Student’s t test was used for statistical analysis. *, P < 0.05.

Error bars represent SD. (B) Passage 4 WT and Rnf8-/- MEFs were cultured as in (A), either under normoxic (21% O2) or hypoxic (5% O2) conditions. Data were from two independent experiments.

(C) Body weights of age-matched WT (n = 10), Rnf8+/- (n = 14), and Rnf8-/- (n = 9) males. Statistical significance was established at each time point using Student’s t-test. *, P < 0.05. Error bars represent SD. (D) Total number of cells in BM (one femur) of WT and Rnf8-/- mice. (E) Absolute numbers of pro-B cells (IgM- CD43+ B220+), pre-B cells (IgM- CD43- B220+), and immature B cells (B220+ IgM+ IgD-) in the BM (one femur) of WT and Rnf8-/- mice. (F) Absolute number of cells in thymus and spleen of WT and Rnf8-/- mice. (G) Absolute numbers of double-negative (CD4-

CD8-), double-positive (CD4+ CD8+), CD4+, and CD8+ cells in the thymus of WT and Rnf8-/- mice.

(H) Absolute numbers of B and T cells in the spleen of WT and Rnf8-/- mice. Data in (D–H) were generated from seven pairs of WT and Rnf8-/- littermates at the age of 6–10 wk. Student’s t test was used for statistical analysis. *, P < 0.05; **, P < 0.0007. Error bars represent SD.

2.3.3 Rnf8 deficiency leads to growth retardation and lymphopenia

Rnf8-/- mice generated by intercrossing of heterozygous mutant mice were born at a Mendelian ratio, indicating that Rnf8 is dispensable for embryonic survival. However, both Rnf8-/- female and male mice showed a runted phenotype and were underdeveloped compared with WT female and male littermates as determined by their body weight measurements at various time points throughout development (Figure 2.2C and Figure 2.3).

Figure 2.3. Body weight measurements of Rnf8-/- females. Body weights of WT, Rnf8+/-, and Rnf8-/- female littermates were measured from weeks 3 to 26. Student’s t-test was used for statistical analysis. *, P < 0.05. Error bars represent SD.

Programmed DSBs play essential roles in development and differentiation of hematopoietic cells, and failure to signal or repair these DSBs significantly impairs hematopoietic lineages and can lead to immunodeficiency (Soulas-Sprauel et al., 2007). To assess the effect of Rnf8 deficiency on 56 the immune system development, we examined BM, thymus, and spleen from 6–10-wk-old Rnf8-

/- mice. FACS analysis and determination of absolute cell numbers of hematopoietic cells indicated significant hypocellularity of the immune organs of Rnf8-/- mice, and the Rnf8-/- hematopoietic cell subpopulations seemed to suffer this defect to a similar extent, though there were no significant differences in proportions of Rnf8-/- hematopoietic cell subpopulations compared with WT littermates (Figure 2.2, D–H; and Figure 2.4). The total number of BM cells (12.7 ± 3.3 × 106) in a femur from Rnf8-/- mice was significantly reduced compared with WT littermates (20.6 ± 1.3 ×

106; P < 0.0003; Figure 2.2D). More importantly, the numbers of Pro-B (IgM- B220+ CD43+) and

Pre-B (IgM- B220+ CD43-) progenitors, as well as the number of immature B cells (IgM+ B220+

IgD-), were significantly reduced in BM from Rnf8-/- mice compared with WT littermates (P < 0.01;

Figure 2.2E).

The absolute number of thymocytes was also significantly reduced in Rnf8-/- mice (101.3 ± 30.4 ×

106) compared with WT littermates (178.9 ± 31.5 × 106; P < 0.0006; Figure 2.2F). Further characterization of these thymocytes indicated significant reductions in the absolute numbers of

CD4- CD8- (double negative), CD4+ CD8+ (double positive), and CD4+ and CD8+ (single positive) thymocyte populations (P < 0.05; Figure 2.2G). As 53bp1 is required for distal V–DJ joining and

TCR-β expression (Difilippantonio et al., 2008), we examined the effect of Rnf8 deficiency on

TCR-β expression. Our data indicated a mild decrease of the level of TCR-β expression in Rnf8-/- thymocytes compared with WT controls (Figure 2.4B), which contrasts with the drastic reduction of TCR-β expression in 53bp1-/- thymocytes (Difilippantonio et al., 2008).

Figure 2.4. Effect of Rnf8 deficiency on hematopoeitic cell populations. Flow cytometric

58 analysis of expression of various surface markers on hematopoietic cells from WT and Rnf8-/- mice.

(A) CD4, CD8, and TCR- expression levels in WT and Rnf8-/- thymocytes. (B) Levels of TCR- in freshly isolated thymocytes. (C) Expression of B220, CD43, and IgM in WT and Rnf8-/- BM cells. (D) B220, Thy1.2, IgM, IgD, CD4, and CD8 expression levels in WT and Rnf8-/- splenocytes.

Analyses were performed on seven pairs of WT and Rnf8-/- littermates at the age of 6–10 wk.

Next, we examined the effect of Rnf8 deficiency on splenocytes and observed a significant reduction in total number of splenocytes from Rnf8-/- mice (55.8 ± 12.7 × 106) compared with WT littermates (89.1 ± 14.2 × 106; P < 0.0006; Figure 2.2F). FACS analysis was further used to determine the absolute numbers of B and T cell populations in the spleen of Rnf8-/- mice and WT littermates. Both Rnf8-/- splenic B and T cell populations were found to be significantly reduced compared with WT controls (P < 0.01; Figure 2.2H). Collectively, these data indicate that Rnf8 plays important roles in maintenance of normal mammalian growth and preventing lymphopenia in both primary and secondary immune organs.

2.3.4 Rnf8 loss impairs spermatogenesis but not ovogenesis

Programmed DSBs initiated by the meiosis-specific protein Spo11 are essential for meiotic recombination (Longhese et al., 2009). The repair of these breaks is mediated by homologous

59 recombination. Defects in signaling or repair of these programmed DSBs during spermatogenesis and ovogenesis can potentially lead to infertility.

Figure 2.5. Impaired spermatogenesis in Rnf8-/- mice. (A and B) Histological analysis of ovaries of WT (A) and Rnf8-/- (B) females showed no significant differences in ovarian architecture.

Bars, 50 μm. (C) Testes from 12-mo-old Rnf8-/- male exhibited a significant reduction in size compared with those of WT littermates. Bar, 0.25 cm. (D–K) Images of H&E-stained seminiferous 60 tubules from 1.5-mo-old WT (D and H) and Rnf8-/- (E, F, I, and J) mice and a 3.5-mo-old Rnf8-/- male (G and K). H&E staining of testes from Rnf8-/- males showed a spectrum of alterations in spermatogenesis. Rnf8-/- testis showed seminiferous tubules with well-populated layers of spermatogonia, spermatocytes, and mature spermatozoa (E and I) and other seminiferous tubules with only few or absent mature spermatozoa in the lumen and an unusually elevated number of immature spermatids and mitotic cells (E and I). Rnf8-/- testis showed disorganized epithelial architecture of seminiferous tubules, few spermatogonia and spermatocytes mixed together with apoptotic bodies, and the absence of mature sperm cells (F and J). Rnf8-/- testis showed focal vacuolar degeneration of most seminiferous tubules (G and K). In contrast to WT (L), H&E staining of epididymis of Rnf8-/- mice (M, 20X) show no histological abnormalities; however, they have a significant reduction of mature sperms. Bars: (D–G) 100 μm; (H–K) 50 μm; (L and M)

100 μm. A minimum of five mice per genotype have been analyzed.

Because Rnf8 plays an essential role in the signaling of DSBs, we examined the effect of its loss on male and female fertility. Ovaries from 9-wk-old Rnf8-/- mice and their WT littermates were harvested, fixed, and stained with hematoxylin and eosin (H&E). No overt difference in the size of ovaries from Rnf8-/- females was observed compared with WT littermates. In addition, H&E staining also indicated that Rnf8-/- ovaries are similar to WT ovaries (Figure 2.5A and B).

However, young (9 weeks of age) and old (1 year of age) Rnf8-/- males showed a reduction in testicular size compared with WT controls (Figure 2.5C). Examination of H&E-stained testis sections from 1.5–3.5-mo-old Rnf8-/- males showed a spectrum of alterations in spermatogenesis compared with their WT male littermates (Figure 2.5D–M). Within the same Rnf8-/- testis, we observed seminiferous tubules with well-populated layers of spermatogonia, spermatocytes, and mature spermatozoa, whereas other seminiferous tubules showed only few or absent mature spermatozoa in the lumen, as well as an unusually high number of cells in mitosis and immature spermatids (Figure 2.5E–I), suggesting an arrest in the spermatogenesis process. We also observed

Rnf8-/- testes with seminiferous tubules that had disorganized epithelial architecture but still showed few spermatogonia and spermatocytes interspersed with apoptotic bodies and the absence of lumen and mature sperm cells (Figure 2.5F–J). Furthermore, more severe phenotypes were also observed, which were characterized by focal vacuolar degeneration of most seminiferous tubules of Rnf8-/- testes, without inflammation and with a normal appearance of interstitial Leydig cells

(Figure 2.5G–K). The epididymis of Rnf8-/- mice, although histologically normal, was either poorly populated with or devoid of spermatozoa compared with WT controls (Figure 2.5L and M).

Interestingly, despite their observed defects in spermatogenesis, and in contrast to the complete infertility of Atm-/-, H2ax-/-, and Mdc1-/- males, Rnf8-/- males derived from AS0574 ES clone successfully produced progenies, although the size of their litters (5.16 ± 0.9) was significantly reduced compared with Rnf8+/- males (9.3 ± 0.6; P = 0.0018). However, attempts to breed Rnf8-/- males from RRR260 ES clone failed to yield any pregnancies, possibly because of the difference of the Rnf8 mutation. We therefore conclude that Rnf8 deficiency impairs spermatogenesis.

2.3.5 53bp1 is partially recruited to DNA DSB sites via Rnf8-independent mechanisms in activated B cells

RNF8 plays an essential role in the signaling of DSBs. It mediates ubiquitylation of H2A and

H2AX, facilitating the sequential recruitment of downstream DDR proteins to DSBs (Huen et al.,

2007; Kolas et al., 2007a; Mailand et al., 2007; Panier and Durocher, 2009b; Wu et al., 2009).

Previous studies indicate the abolition of recruitment of certain DDR proteins, especially 53bp1,

to IR-induced DSBs in Rnf8-/- MEFs and human RNF8 knockdown cell lines (Huen et al., 2007;

Kolas et al., 2007a; Mailand et al., 2007; Panier and Durocher, 2009b; Wu et al., 2009). To further

study the hierarchy of DDR proteins in immune cells, we examined time course recruitment of

H2ax and 53bp1 to DSBs in Rnf8-/- and WT B cells. B cells from Rnf8-/- and WT mice were

activated for 60 h in the presence of LPS. Cells were either left untreated or irradiated with 3 Gy

of radiation and were then harvested at various time points after IR for immunofluoresence to

examine H2ax and 53bp1 IRIF formation. A significantly higher number of H2ax foci was

observed in untreated Rnf8-/- cells compared with WT control cells, indicating an elevated number

of endogenous breaks in the activated Rnf8-/- B cells (Figure 2.6A and C). However, no significant

difference in the number or size of H2ax IRIF was observed between Rnf8-/- and WT cells at 0.5

h after IR, although the number of H2ax IRIF in Rnf8-/- B cells was slightly greater at 3.5 h after

IR. Interestingly, despite the increased number of IRIF for H2ax in nonirradiated Rnf8-/- activated

B cells, a significantly lower number of 53bp1 foci was observed in these cells compared with WT

controls (Figure 2.6B and D). About 38% of activated Rnf8-/- B cells was able to form 53bp1 foci

63 at 0.5 h after radiation (3 Gy), and these foci appeared to be smaller in size than WT controls.

Moreover, in contrast to WT controls, the number of Rnf8-/- B cells showing 53bp1 IRIF dropped by ~50% at 3.5 h after IR treatment, indicating the unstable nature of 53bp1 foci in the absence of

Rnf8. Collectively, these data indicate the existence of Rnf8-independent mechanisms for the recruitment of 53bp1 to sites of DNA breaks in activated B cells, although Rnf8-dependent mechanisms are the major driving force for the formation and retention of 53bp1 IRIF.

Figure 2.6. Partial recruitment of 53bp1 to the sites of DNA damage in the absence of

Rnf8. (A and B) LPS-activated B cells either untreated (UT) or harvested 0.5 and 3.5 h after 3

Gy irradiation were stained with anti-H2ax (A) and anti-53bp1 (B) antibodies. Images were taken with a confocal microscope. Bars, 10 μm. (C and D) Quantitative analysis of H2ax and

53bp1 subnuclear foci. Activated B cells harboring two or more than two foci were counted as foci-positive cells. Three independent experiments were performed, and a minimum of 300 cells was counted for each condition and genotype. A Student’s t-test was used for statistical analysis.

*, P < 0.01. Error bars represent SD.

2.3.6 Rnf8 deficiency leads to increased radiosensitivity both in vitro and in vivo

Failure to repair DSBs leads to cell cycle arrest and apoptosis, thus preventing transmission of damaged DNA during cell division. Based on the role of Rnf8 on initiation and maintenance of

IRIF for DDR proteins such as 53bp1, we examined the effect of its loss, both in vitro and in vivo, with respect to radiosensitivity. Rnf8-/- and WT thymocytes were either left untreated or subjected to various dosages of IR (0.5–6 Gy) and were harvested 18 h after IR. The level of cell death was then examined by flow cytometry using 7AAD, as well as Annexin V and propidium iodide assays.

A significant increase in radiosensitivity was observed with Rnf8-/- thymocytes compared with WT controls (Figure 2.7A).

The Atm–Chk2–p53 signaling pathway plays essential roles in response to DSBs (Meek, 2009).

Therefore, we examined the effect of Rnf8 loss on the activation of this pathway. Rnf8-/- and WT thymocytes were either left untreated or irradiated with 2 Gy of radiation, followed by harvesting at 2 and 5 hours after treatment. Western blotting analyses indicated a significantly higher level of

IR-induced phosphorylation of Chk2 and p53 (Ser15), as well as an increase in total p53 level in

Rnf8-/- thymocytes compared with WT controls (Figure 2.7 B).

To examine whether Rnf8 deficiency affects radiosensitivity of other cell types, we performed colony-forming assays using Rnf8-/- , Rnf8+/-, and WT BM cells either untreated or subjected to

67 various dosages of radiation (0.5–6 Gy). Colonies were counted at day 10 and pictures of the dishes were taken (Figure 2.7C and D). After 0.5 Gy of irradiation, the total number of colonies for WT controls was, on average, 1.5-fold greater than that of Rnf8-/-. Remarkably, the fold differences between WT and Rnf8-/- colony numbers displayed an increasing trend with increased IR dosages

(2 Gy, 3.3-fold difference; 4 Gy, 4.3-fold difference; 6 Gy, 13.2-fold difference). In addition,

Rnf8+/- BM cells only displayed a very modest sensitivity to 4 Gy and 6 Gy of radiation compared with WT controls (P < 0.05), whereas they were significantly less sensitive to irradiation compared with Rnf8-/- BM cells at all doses (P < 0.02; Figure 2.7D).

Figure 2.7. Rnf8-/- thymocytes and BM cells display increased radiosensitivity in vitro. (A)

Thymocytes from WT and Rnf8-/- littermates were either left untreated or treated with various dosages of radiation. Cells were then stained with 7-AAD 18 h after treatment and analyzed by

69 flow cytometry. Data from all treated samples were normalized to their respective untreated samples, and percentages of viable cells were plotted. Three independent experiments were performed. Student’s t-test was used for statistical analysis. *, P < 0.05. Error bars represent SD.

(B) Thymocytes from Rnf8-/- mice and WT littermates were either left untreated (0) or irradiated with 2 Gy of radiation and harvested at 2 and 5 h after treatment. Western blot analysis was performed to examine protein levels of Chk2, Ser15 phosphorylated p53 (S15-p53), and total p53.

Actin was used as a loading control. (C) Representative pictures of dishes showing colonies derived from WT and Rnf8-/- BM cells either left untreated or irradiated with the indicated doses

(0.5–6 Gy). All the pictures were taken at day 10 of culture. (D) Quantitative analysis showing the percentages of surviving colonies. Three independent experiments were performed. Data from all treated samples were normalized to their respective untreated samples, and Student’s t-test was used for statistical analysis. *, P < 0.05 (statistically significant difference between WT and Rnf8+/- colony survivals); **, P < 0.02 (statistically significant difference of Rnf8-/- compared with WT and Rnf8+/- colony survivals). Error bars represent SD.

To determine the effect of Rnf8 inactivation regarding radiosensitivity in vivo, cohorts of age- matched Rnf8-/- and WT littermates (n = 10 per genotype) were subjected to 8 Gy of radiation, followed by daily monitoring for 51 days (Figure 2.8). Although 60% of irradiated WT mice survived and recovered from the IR treatment, only 10% of irradiated Rnf8-/- mice remained viable at the end of the monitoring period. Therefore, Rnf8-/- mice were significantly more sensitive to

radiation compared with WT littermates (Kaplan Meier analysis, Log-rank test, P < 0.003).

Collectively, these data indicate the important role of Rnf8 in the response to DNA damage and that its loss increases radiosensitivity.

Figure 2.8. Rnf8-deficiency increases radiosensitivity in vivo. Kaplan-Meier survival analysis on cohorts of age-matched WT (n = 10) and Rnf8-/- (n = 10) mice, subjected to a single dose of 8

Gy of radiation. The cohorts of mice were monitored daily for survival up to 51 days after irradiation. The log-rank test indicates significant difference between WT and Rnf8-/- survival curves (P< 0.003).

2.3.7 Rnf8-deficient B cells are intrinsically defective for CSR

IgH CSR is one of the three important mechanisms for antibody diversification in peripheral B lymphocytes (Soulas-Sprauel et al., 2007; Stavnezer and Schrader, 2014). It is initiated by activation-induced cytidine deaminase, which induces DSBs into the repetitive switch (S) regions upstream of the heavy chain constant (C) region exons of the Ig locus. The subsequent repair is mainly accomplished by the nonhomologous end-joining repair pathway. Interestingly, CSR is reduced in B cells deficient for several DDR proteins that act upstream of Rnf8 including Atm,

H2ax, and Mdc1 (Kotnis et al., 2009). Furthermore, defective CSR has also been observed with

53bp1-/- mice (Difilippantonio et al., 2008; Manis et al., 2004) and with the RIDDLE syndrome, which is associated with RNF168 mutations (Stewart et al., 2009; Stewart et al., 2007).

To determine whether Rnf8 plays a role in CSR, purified Rnf8-/- and WT B cells labeled intracellularly with Carboxyfluorescein Diacetate Succinimidyl Ester (CFSE) were stimulated with anti-CD40 plus IL4 for 4.5 days, and their ability to switch into IgG1 isotype was examined by FACS analysis (Figure 2.9). Interestingly, the percentage of IgG1+ cells in activated Rnf8-/- B

72 cells (8.2 ± 0.8%) was significantly reduced compared with WT cells (28.4 ± 1.1%; P < 0.0001;

Figure 2.10D). These activated IgG1+ cells were B220+ and had similar cell division profiles, as shown by CFSE tracking (Figure 2.10A–C), thus ruling out a role for impaired proliferation in the reduced CSR of Rnf8-/- B cells. Furthermore, to examine whether p53 inactivation can rescue this

Rnf8-deficiency associated CSR defect, B cells were purified from the spleens of WT, Rnf8-/-, p53-

/-, and Rnf8-/- p53-/- mice, followed by stimulation to class switch to IgG1. Notably, while CFSE tracking of cell division profiles were similar among all four genotypes, the ability of Rnf8-/- p53-

/- B cells to undergo IgG1 CSR was not significantly different from that of Rnf8-/- B cells (P>0.05), but was significantly dampened compared with that of both WT and p53-/- B cells (P<0.01; Figure

2.11). Thus, additional deletion of p53 does not rescue nor exacerbate this Rnf8-deficiency associated CSR defect.

Figure 2.9. A schematic depicting the in vitro class switch recombination functional assay.

Splenocytes were freshly collected and treated with NH4Cl to lyse red blood cells. Cells were then labeled with CFSE, followed by stimulation with anti-CD40 and IL4 to class switch towards IgG1 isotype, a high affinity antibody for combating against bacterial and viral infections. At day 4 of in vitro culture, about 30% of WT B cells (B220 positive) would be expected to express IgG1 isotype on their cell surface as assessed by flow cytometry analysis.

Figure 2.10. Rnf8-/- B cells are intrinsically defective for IgH CSR. (A and B) Flow cytometric analysis of IgG1 expression was performed on CFSE-labeled B cells 4.5 d after anti-CD40 and

IL4 stimulation. Populations of IgG1+ switched cells were outlined by the boxes with percentages indicated. Results are representative of four independent experiments. B220 was used as a marker for B cells. (C) Cell divisions tracked by CFSE, after 4.5 d of stimulation with anti-CD40 plus

IL4, were analyzed by flow cytometry. Each peak indicates a cell division. (D) Percentages of switched cells under anti-CD40 and IL4 stimulation for 4.5 d from four independent experiments

(left) and percentages of switched cells under LPS stimulation for 4.5 d from three independent experiments (right). Statistical significance was established by Student’s t-test. *, P < 0.05; **, P

< 0.0002. Error bars represent SD. (E and F) Flow cytometric analysis of IgG3 expression was performed on CFSE-labeled B cells 4.5 d after LPS stimulation. Populations of IgG3+ switched cells were outlined by the boxes with percentages indicated. Three independent experiments were performed. B220 was used as a marker for B cells. (G) Cell division profiles of WT and Rnf8-/- B cells tracked by CFSE at day 4.5 after LPS stimulation. (H) Basal serum levels of various isotypes of Ig’s quantified in 14-month-old Rnf8-/- (n = 8) and WT (n = 8) mice using ELISA. Horizontal bars represent the median values of the indicated serum Ig isotype levels from their respective genotypes. The Mann-Whitney test for two independent samples was employed for statistical analysis. *, P < 0.05. (I) S-S1 recombination products were quantified by digestion- circularization PCR analysis. Genomic DNA was obtained from B cells stimulated for 4 d with anti-CD40 and IL4. Fivefold dilutions of the genomic DNA were used as templates in the PCR reactions. nAChR was used as the control DNA. H2O indicates no input DNA. Results are representative of three independent experiments.

Figure 2.11. Additional inactivation of p53 does not rescue Rnf8-deficiency associated CSR defect. (A) Purified B-cells from 6-week-old WT, Rnf8-/-, p53-/-, and Rnf8-/- p53-/- mice were labeled with CFSE and stimulated with anti-CD40 and IL-4 for 4 days, followed by flow cytometric analysis for IgG1 expressing class switched B cells. Results are representative of four independent experiments. B220 was used as a marker for B cells. (B) Percentages of switched cells under anti-

CD40 and IL4 stimulation for 4 days from four independent experiments, which are separate experiments from Figure 2.9A–D. Statistical significance was established by Student’s t-test. *, P

< 0.01. (C) Cell division profiles of WT, Rnf8-/-, p53-/-, and Rnf8-/- p53-/- B cells tracked by CFSE

77 at day 4 after anti-CD40 and IL-4 stimulation.

Similar analysis of the expression level of IgG3 on B cells stimulated with LPS for 4.5 days indicated a mild decrease in the percentage of IgG3+ B cells from Rnf8-/- mice (7.3 ± 2.2%) compared with WT controls (13.3 ± 3.0%; P < 0.05; Figure 2.10D–G). To further examine the effect of Rnf8 deficiency on CSR in vivo, we measured basal immunoglobulin (Ig) levels in serums of 14-month-old Rnf8-/- mice (n = 8) and WT littermates (n = 8). ELISA analysis indicated no significant differences in the levels of serum IgM, IgG2a, and IgA in Rnf8-/- mice compared with

WT littermates (Figure 2.10H). However, serum levels of IgG1, IgG3, and IgG2b were significantly reduced in Rnf8-/- mice compared with WT littermates (P < 0.05). Therefore, these data indicate that Rnf8 deficiency leads to defective CSR that results in dampened secretion of various classes of Ig’s in vivo.

The CSR defect associated with the loss of Rnf8 was further investigated using digestion- circularization PCR assay. Genomic DNA, extracted from Rnf8-/- and WT B cells stimulated for

4.5 days in the presence of anti-CD40 plus IL4, was used in this assay to examine the efficacy of

S-S recombination. The nicotinic acetylcholine receptor (nAChR) locus was used as a control.

Digestion-circularization PCR assay, using serial dilutions of the genomic DNA template, confirmed the impaired CSR in Rnf8-/- B cells and demonstrated that the CSR defect occurs at the genomic level and is intrinsic to Rnf8-/- B cells (Figure 2.10I), thereby directly ruling out the possibility of the defect being Ig exportation to cell surface.

Interestingly, Rnf8+/- B cells also indicated a significantly impaired CSR to IgG1 compared with

WT controls (Figure 2.12A and B; P < 0.03); however, the defect of IgG1 CSR observed with

Rnf8+/- B cells was not as marked as with Rnf8-/- B cells (P < 0.02), suggesting a gene dosage effect.

Overall, these data demonstrate a significant role for Rnf8 in CSR, and indicate that loss of a single allele of Rnf8 is sufficient to incapacitate a humoral immune response.

Figure 2.12. IgH CSR in Rnf8+/- B cells is impaired and reflects a gene dose effect. (A) Flow

cytometric analysis of IgG1 expression was performed on WT, Rnf8+/-, and Rnf8-/- resting splenic

B cells 4 days after anti-CD40 and IL4 stimulation. B220 was used as a marker for B cells. Results

are representative of four independent experiments, which are separate experiments from Figure

2.10A–D and Figure 2.11. (B) Bar chart depicting percentages of WT, Rnf8+/-, and Rnf8-/- class

switched IgG1 positive B cells at day 4 post anti-CD40 and IL4 stimulation. Means of percentages

of WT, Rnf8+/-, and Rnf8-/- IgG1 positive B cells were obtained from four independent experiments,

which are separate experiments from Figure 2.10A-D and Figure 2.11. Student’s t-test was used

for statistical analysis. *, P < 0.03. Error bars represent SD.

2.3.8 Rnf8 suppresses spontaneous and IR-induced genomic instability

Mutations in the DDR components often result in failure to signal or repair DSBs and,

consequently, lead to increased genomic instability, which serves as a driving force for cancer

development (Bartek and Lukas, 2007). To examine the potential role of Rnf8 in maintaining

genomic integrity, metaphase spreads of LPS-activated Rnf8-/- and WT B cells were analyzed for

chromosomal aberrations (Figure 2.13A and B). Under untreated conditions, Rnf8-/- cells displayed

an elevated frequency of cells harboring total chromosomal aberrations (10.8 ± 2.7%) compared

with WT controls (1.7 ± 1%; P < 0.05). Furthermore, we subcategorized cells harboring

chromosomal aberrations into cells harboring chromosomal breaks as well as those that exhibit

structural chromosomal aberrations. Notably, the level of breaks (8.3 ± 1.7%) in untreated Rnf8-/-

80 activated B cells was ~4.9-fold greater than that of WT controls (1.7 ± 1%; P < 0.05). Moreover, structural chromosomal aberrations (2.5 ± 1.4%) were observed in untreated Rnf8-/- activated B cells, whereas such aberrations were completely absent in WT controls. In response to 2 Gy of radiation, the frequency of metaphases with chromosomal aberrations (45.8 ± 3.8%) in Rnf8-/- cells was significantly increased compared with WT controls (34 ± 1.7%; P < 0.05). The level of IR- induced breaks appeared to be higher in Rnf8-/- cells than WT controls, although this difference did not reach statistical significance (P > 0.05). Interestingly, the frequency of IR-induced structural chromosomal aberrations was significantly greater in Rnf8-/- cells (25 ± 3.4%) compared with WT controls (17.5 ± 3.8%; P < 0.05; Figure 2.13A and B). Therefore, these data demonstrate that Rnf8 is required for maintaining genomic integrity, as its loss results in elevation of both spontaneous and IR-induced genomic instability.

Figure 2.13. Rnf8 deficiency leads to increased genomic instability. (A) Representative metaphases of WT and Rnf8-/- LPS-activated B cells. Cells were either left untreated or irradiated with 2 Gy of radiation and harvested 12 h after treatment. (B) The graph shows the incidence of total spontaneous and IR-induced chromosomal aberrations. A minimum of 40 metaphase spreads

of untreated or irradiated Rnf8-/- and WT cells were analyzed for each genotype and treatment in

three independent experiments. Data are presented as mean ± SD. *, P < 0.05 compared with WT

controls. r, ring; b, break; f, fragment; g, gap.

2.3.9 Rnf8 is a novel tumor suppressor

The fundamental and most important role of DDR proteins is to safeguard the integrity of the

genome, thus minimizing the risk for disorders including cancer. Because Rnf8-/- cells exhibit

elevated levels of genomic instability, we examined the potential role of Rnf8 in cancer. Cohorts

of Rnf8-/- (n = 27), Rnf8+/- (n = 35), and WT (n = 29) mice were monitored daily for survival for

465 days. Only 56% of Rnf8-/- (15 out of 27) mice were viable at the end of the monitoring period,

whereas 77% of Rnf8+/- mice and 96.6% of WT littermates remained alive. Kaplan-Meier survival

analysis (Figure 2.14A) indicated that the life spans of Rnf8-/- and Rnf8+/- mice were significantly

reduced compared with those of WT mice (Log-rank test, P < 0.0005 and 0.04, respectively).

Figure 2.14. Rnf8 mutant mice suffer a higher mortality rate. (A) Kaplan-Meier survival analysis of overall survivals for cohorts of WT (n = 29), Rnf8+/- (n = 35), and Rnf8-/- (n = 27) mice.

Mice were monitored for survival for 465 days. A log-rank test was used for statistical analysis.

There was a statistically significant difference between WT and Rnf8-/- curves (P < 0.0005) and between WT and Rnf8+/- curves (P < 0.05). (B) Picture of a 25-day-old female Rnf8-/- mouse (right) showing runted phenotype. A WT female (left) from the same litter is shown as a control. (C) Two different Rnf8-/- thymomas were analyzed by flow cytometry, and WT thymocytes were used as controls. (D) A lymphoma from Rnf8-/- mouse was analyzed by flow cytometry, and WT lymph node cells were used as controls. (E and F) One Rnf8+/- mouse developed brain cancer. (E) Tumor cells were of a similar size to adjacent normal cells and exhibited limited infiltration capability into adjacent tissues. Bar, 100 μm. (F) A higher magnification of the brain tumor revealed that the tumor cells were under active mitosis. Bar, 25 μm.

Kaplan-Meier tumor-free survival curves were established (Figure 2.15A), and biopsies of the moribund mice were examined using H&E staining in an effort to determine their cause of death

(Figure 2.15B–G). 36.4% of the Rnf8-/- mice cohort developed cancer, and Kaplan-Meier tumor- free survival analysis indicated statistically significant differences of the Rnf8-/- cohort of mice from both the WT and Rnf8+/- cohorts of mice regarding cancer development (Log-rank test, P <

0.001 and 0.02, respectively). Characterization of Rnf8-/- tumors by H&E staining indicated that four were lymphomas (50%), three were thymomas (38%), one was mammary carcinoma (13%), one was a skin tumor (13%), and one was a sarcoma (13%). FACS analysis of the Rnf8-/- thymomas indicated that they were of CD4+CD8+ or CD4+ in origin, whereas the lymphomas were

B220+ or B220- (Figure 2.14C and D). In addition, 6.9% of Rnf8+/- mice also developed tumors

(one thymoma and one brain tumor; Figure 2.14E and F), although the difference between the

Rnf8+/- and WT cohorts of mice did not reach statistical significance based on Kaplan Meier analysis (Log-rank test, P = 0.16). The tumor cells from the Rnf8 mutants were frequently found to infiltrate other organs such as liver (Figure 2.15G).

In addition to the Rnf8 mutant mice that developed malignancies, ~11% of Rnf8-/- and Rnf8+/- cohorts of mice were extremely runted and displayed premature death with no sign of tumorigenesis (Figure 2.14A and B). Although the exact cause of death for these mice remains unknown, based on the impaired CSR in homozygous and heterozygous Rnf8 mutant backgrounds

(Figure 2.11 and Figure 2.13) we speculate that immunodeficiency might account for a subset of these premature deaths. Collectively, these data demonstrate that Rnf8 is a novel tumor suppressor and that its loss also results in non-cancer-related death, likely as a result of immunodeficiency.

Figure 2.15. Rnf8 is a novel tumor suppressor. (A) Kaplan-Meier tumor-free survival analysis for cohorts of WT (n = 28), Rnf8+/- (n = 29), and Rnf8-/- (n = 22) mice. Mice were monitored for

87 survival for 465 days. A log-rank test was used for statistical analysis. There was a statistically significant difference between WT and Rnf8-/- curves (P < 0.0005) and between Rnf8+/- and Rnf8-/- curves (P < 0.01). (B and C) Sarcoma invading soft tissues, including skeletal muscle, adipose tissue, and a peripheral nerve, is shown (B). High magnification shows that the tumor is poorly differentiated and has a high proliferative index and atypical mitosis (C). (D) Malignant transformation of the mammary gland with a multinodular growth pattern and invasion into adjacent connective tissues. The tumor is well vascularized and shows no necrosis. (E) A higher magnification shows that tumor cells are growing in small groups, with prominent mitosis and invasion. (F) A tumor composed of small cells with round, sometimes eccentric nuclei and small cytoplasm, is compatible with a lymphoma of B cell origin. (G) Despite its well differentiated appearance, this tumor aggressively infiltrates the periportal areas of the liver. Bars: (B) 100 μm;

(C, E, and F) 25 μm; (D) 250 μm; (G) 50 μm.

2.4 Discussion

The fundamental roles of signaling and repair mechanisms of DSBs are best demonstrated by mutations that impair these mechanisms and that associate with several human syndromes characterized by developmental defects, neurodegenerative disorders, immunodeficiency, and increased cancer predisposition (Hakem, 2008; Hoeijmakers, 2009; O'Driscoll and Jeggo, 2006).

For example, mutations of ATM lead to the human autosomal recessive disorder ataxia– telangiectasia, which is characterized by cerebellar ataxia, progressive mental retardation,

88 neurological defects, impaired immune functions, and increased cancer risk. Furthermore, germline mutations of the BRCA1 gene, which is involved in the signaling and repair of DSBs, predispose to hereditary breast and ovarian cancer.

The E3 ligases RNF8 and RNF168 have been recently demonstrated to play critical roles in DSBs signaling. Interestingly, mutations of RNF168 have been shown to be associated with the human

RIDDLE syndrome, which is characterized by features including radiosensitivity and immunodeficiency (Stewart et al., 2009; Stewart et al., 2007). However, the in vivo roles of RNF8 and whether its inactivation results in disease remained unknown.

Our data indicate that, similar to Atm (Barlow et al., 1996), H2ax (Celeste et al., 2002), Mdc1

(Lou et al., 2006), and 53bp1 (Ward et al., 2003) but in contrast to Brca1 (Hakem et al., 1996),

Rnf8 is dispensable for embryonic development. However, in line with the effect of null or hypomorphic mutations of genes involved in early DSB signaling, such as Atm, H2ax, Mdc1,

53bp1, and Brca1 (Celeste et al., 2002; Lou et al., 2006; McPherson et al., 2004; Ward et al., 2003),

Rnf8-/- mice are growth retarded and display lymphopenia. Interestingly, RNF8 mediates the recruitment of RNF168 to the break sites, and the growth retardation of Rnf8-/- mice is reminiscent of the growth defects associated with RNF168 mutation in the human RIDDLE syndrome.

The sequential recruitment of proteins, which is involved in DNA damage signaling and repair as well as cell cycle checkpoint activation, to DNA break sites allows cells to efficiently repair their

DSBs through elegantly orchestrated mechanisms. Knockdown experiments of RNF8 in human 89 cell lines and MEFs deficient for Rnf8 have demonstrated the requirement for RNF8 for the recruitment and maintenance of several DDR proteins, including 53BP1, at the site of damage

(Huen et al., 2007; Huen et al., 2008; Kolas et al., 2007a; Mailand et al., 2007; Wang and Elledge,

2007). Remarkably, loss of Rnf8 in activated Rnf8-/- B cells failed to completely mitigate IR- induced formation of 53bp1 foci. This finding indicates the coexistence of Rnf8-dependent and

Rnf8-independent signaling pathways involved in the recruitment and retention of 53bp1, and likely other downstream DDR proteins, to IR-induced DSBs.

The Spo11-mediated generation of DSBs and their repair by homologous recombination are required for spermatogenesis and ovogenesis (Longhese et al., 2009). Interestingly, although no ovogenesis problems were observed in Rnf8-/- females, increased cell death and a reduced number to a complete absence of spermatocytes was observed in seminiferous tubules from Rnf8-/- males.

The spermatogenesis defect of Rnf8-/- males is reminiscent of loss of the upstream DDR proteins including Atm, H2ax, and Mdc1. Interestingly, the spermatogenesis defect observed with Rnf8-/- males contrasts with the lack of any apparent role for 53bp1 in spermatogenesis, suggesting that the role of Rnf8 in spermatogenesis is independent of 53bp1.

Besides spermatogenesis, programmed DSBs are also required for CSR. The generation of these

DSBs within the Ig locus results in exchange of IgH constant region exons, providing an effective mechanism for resting B cells expressing low affinity IgM to switch to produce specific Ig isotypes such as IgG1 and IgA (Soulas-Sprauel et al., 2007; Stavnezer and Schrader, 2014). This ultimately enhances the effector function of the Ig’s produced during an immune response to protect against

90 foreign pathogens in a highly effective and specific manner. Although CSR efficiency is defective in Atm-/-, H2ax-/-, Mdc1-/-, and 53bp1-/- mice, it is the loss of 53bp1 that most significantly impairs this process (Kotnis et al., 2009). Interestingly, CSR to IgG1 in Rnf8-/- B cells is reduced by >66%, although this defect did not reach the severity of 53bp1-deficient B cells. This finding indicates that the role of 53bp1 in CSR is only partially affected by the loss of Rnf8, and this is consistent with our observation of the incomplete abolition of 53bp1 recruitment to DNA DSBs in irradiated

Rnf8-/- activated B cells, suggesting the coexistence of Rnf8-dependent and Rnf8-independent mechanisms for the recruitment and retention of 53bp1 to DSBs in activated B cells. Exactly how

Rnf8 regulates CSR requires further investigation. Interestingly, the CSR defect observed in Rnf8- deficient mice indicates their immunodeficient state and is reminiscent of the immunodeficiency associated with the human RIDDLE syndrome.

A major role of DDR proteins in mammalian cells is to maintain genomic integrity. The impairment of this process increases cancer risk, as it can lead to inactivation of tumor suppressor genes or deregulation of the expression of oncoproteins. Interestingly, the increased spontaneous and IR induced genomic instability in Rnf8-deficient cells further demonstrates the importance of

Rnf8 in the signaling of DSBs and in the maintenance of genomic integrity, and it is in line with the increased radiosensitivity and IR-induced activation of the Chk2–p53 response pathway in the absence of Rnf8. The role of Rnf8 in maintaining genomic integrity is also consistent with the involvement of several other DDR proteins in this process, and it further highlights the importance of the integrity of the DSB signaling network for the maintenance of genomic integrity.

Consistent with their genomic instability, Rnf8-/- mice suffered a statistically significant increase in the incidences of cancer compared with their WT littermates, thus demonstrating that Rnf8 is a novel tumor suppressor. The increased cancer risk is associated, to a various extent, with inactivation of individual components of the DDR pathway that function upstream or downstream from Rnf8. This highlights the need for DDR proteins to function in close collaborations to prevent cancer development.

Interestingly, a gene dose effect for Rnf8 has also been identified in this study. Loss of one allele of Rnf8 reduced the level of Rnf8 protein by approximately twofold, impaired CSR in B cells, and resulted in a mild increase in radiosensitivity of BM cells. Although the increased cancer predisposition of Rnf8+/- mice did not reach a statistical significance, their increased lethality further supports a gene dose effect for Rnf8. Although, no data are yet available to link RNF8 to human pathologies, the pleiotropic defects associated with Rnf8 inactivation, including increased radiosensitivity, impaired spermatogenesis, and defective CSR together with growth retardation, immunodeficiency, and increased cancer predisposition, all highlight the critical physiological functions of RNF8.

2.5 Materials and Methods

Mouse genotyping and genetic analysis. Two Rnf8 ES gene trap clones (AS0574 and RRR260) were obtained from MMRRC and used to generate Rnf8 chimeras. Germline transmission of the

92 two Rnf8 mutations was successfully obtained, and Rnf8-/- mice were generated by intercrossing

Rnf8+/- mice. All mice in this study were in a mixed 129/J × C57BL/6 genetic background, maintained in a specific pathogen-free environment, and genotyped by Southern blotting and PCR

(Primer sequences for AS0574 clone: mutant forward 5’-TCAAAGGTTTGCCCTCTGAT-3’, mutant reverse 5’-CGGAGCGGATCTCAAACTCT-3’, WT forward 5’-

TGATGACACCTGGGCATGT-3’, and WT reverse 5’-TCTTTGAGACAGCGCCTGG-3’; primer sequences for RRR260 clone: mutant forward 5’-GAGCCTGAAAGGCATCTTGG-3’, mutant reverse 5’-CCGGCTAAAACTTGAGACCTT-3’; WT forward 5’-

GAGCCTGAAAGGCATCTTGG-3’, and WT reverse 5’-TGTAAGCCGCTCACTGTGCT-3’).

All experiments were performed in compliance with the Ontario Cancer Institute animal care committee guidelines. Animal protocols were approved by the Animal Resource Center of Ontario

Cancer Institute.

Flow cytometry. Single-cell suspensions from BM (femurs), thymus, spleen, and lymph node of

6–10-wk-old mice were stained with monoclonal antibodies (eBioscience) against B220, IgM,

IgD, Thy-1.2, TCR-β, CD4, CD8α, and CD43. FACS analyses were performed using a

FACSCalibur (BD) and data were subsequently analyzed by FlowJo software (Tree Star, Inc.).

In vitro MEF proliferation assay under oxic and hypoxia conditions. Rnf8-/- and WT MEFs were generated according to standard procedures. For the in vitro proliferation assay, 3 × 105 passage 2 MEFs were seeded onto each 6-cm dish containing culture media and then incubated at

37°C under 5% CO2. After 3 days, the cells were trypsinized, counted, and reseeded at the same

93 density of 3 × 105 cells/6-cm dish. This process was repeated until day 15. For the hypoxia experiments, the same procedure was performed using passage 4 MEFs. Hypoxia (0.2 or 5% O2 with 5% CO2 and balanced N2) was achieved using Whitley H35 hypoxystations (Don Whitley

Scientific).

In vitro activation of B cells and lymphocyte analysis. Single-cell suspensions from spleens of

6 to 10-wk-old mice were prepared according to standard procedures. Splenic B cells were subsequently enriched using the Mouse B Cell Negative Isolation kit (Invitrogen). Purified B cells were labeled with 5 μM CFSE for 10 min at 37°C (Invitrogen) and then cultured at a density of 1

× 106 cells/ml in 6-well flat-bottom plates in RPMI medium supplemented with 10% FCS, 5 × 10-

5 M 2-ME, and 500 ng/ml anti-CD40 (HM40-3; BD) plus 1,000 U/ml IL4 (eBioscience) or 25

μg/ml LPS alone (Sigma-Aldrich). Anti-CD40– and IL4-stimulated cells were harvested from cultures at day 4.5 to assess surface expression of IgG1, whereas B cells activated by LPS to class switch to IgG3 isotype were harvested from cultures at day 4.5. Cells were then stained, respectively, with PE-conjugated anti–mouse IgG1 (BD) or PE-conjugated anti–mouse IgG3

(SouthernBiotech) and APC-conjugated B220 (eBioscience), followed by analysis on a

FACSCalibur. Data were subsequently interpreted using FlowJo software.

Immunofluorescence. Passage 3 MEFs derived from WT and Rnf8-/- littermates were seeded onto coverslips and either left untreated or subjected to 3 Gy of radiation. At various time points after

IR, the MEFs were fixed with 2% paraformaldehyde, permeabilized in 0.5% NP-40, and blocked in 2% BSA/1% donkey serum, followed by staining with rabbit anti-53bp1 (Bethyl Laboratories,

Inc.) and mouse monoclonal anti-H2ax (Ser139; Millipore). Donkey anti–rabbit Alexa Fluor 568 and donkey anti–mouse Alexa Fluor 488 (Invitrogen) were used, respectively for the secondary staining, followed by DAPI (Invitrogen) counterstaining. Images were taken on a microscope

(IX81; Olympus) under 60× magnification. Image quantification was performed on maximum intensity projections using Image Pro Plus software (Media Cybernetics), and ImageJ software

(National Institutes of Health) was used for processing the raw data. Furthermore, splenocytes from WT and Rnf8-/- mice were activated with 10 ng/μl LPS for 60 h. Cells were either left untreated or subjected to 3 Gy of radiation. At various time points after IR, the cells were cytospun onto glass slides (Thermo Fisher Scientific), fixed with 2% paraformaldehyde, blocked with antibody dilution buffer (10% normal goat serum, 3% BSA, and 0.05% Triton X-100 in PBS), and then stained with rabbit anti-53bp1(Bethyl) and rabbit anti-H2ax (Ser139; Millipore). Goat anti–rabbit Alexa Fluor 488 (Invitrogen) was used for the secondary staining. Subsequently, slides were counterstained with DAPI (Invitrogen) and mounted with Mowiol (Sigma-Aldrich). Images were taken on a laser confocal microscope (LSM510; Carl Zeiss, Inc.) under 63× magnification.

ImageJ software was used for processing the raw data, and foci-positive cells were quantified by blind manual counting.

Digestion-circularization PCR. 1 μg of genomic DNA from cultured B cells under anti-CD40 and IL4 stimulation for 4.5 days was digested with EcoRI overnight, and 100 ng of the digested product was ligated with T4 DNA ligase (New England Biolabs, Inc.). Two rounds of nested step- down PCR were then performed on the ligated DNA using nested primer pairs for Sμ-S and nAChR. Primer sequences for the first round of PCR are as follows: Sμ-S, 5’-

GAGCAGCTACCAAGGATCAGGGA-3’ and 5’-CTTCACGCCACTGACTGACTGAG-3’; and

95 nAChR, 5’-GCAAACAGGGCTGGATGAGGCTG-3’ and 5’-

GTCCCATACTTAGAACCCCAGCG-3’. Primer sequences for the second round are as follows:

Sμ-S1, 5’-GGAGACCAATAATCAGAGGGAAG-3’ and 5’-

GAGAGCAGGGTCTCCTGGGTAGG-3’; and nAChR, 5’-

GGACTGCTGTGGGTTTCACCCAG-3’and 5’-GCCTTGCTTGCTTAAGACCCTGG-3’.

Antibody detection by ELISA. Total IgG1, IgG2a, IgG2b, IgG3, IgA, and IgM levels were determined in serum from eight pairs of 14-month-old WT and Rnf8-/- littermates using Mouse

Immunoglobulin Isotype Panel kits (Southern Biotech). 96-well plates (MaxiSorp; Thermo Fisher

Scientific) were precoated with 9 μg/ml of goat anti–mouse Ig capture antibody overnight and were then blocked with 1% BSA/PBS for 1 hour. The various isotypes of serum Ig were diluted and captured onto the plates overnight at 4°C. Detections were made, respectively, through diluted

HRP-labeled goat anti–mouse IgG1, IgG2a, IgG2b, IgG3, IgA, and IgM, and ABTS substrate.

Standard curves were generated using purified mouse IgG1, IgG2a, IgG2b, IgG3, IgA, and IgM isotype controls (Southern Biotech), and absorbance was measured at 405 nm.

Protein analysis and antibodies. Western blot analysis was performed using rabbit anti-p53

(DO1; Santa Cruz Biotechnology, Inc.), anti-phospho-p53 (S15; Cell Signaling Technology) and

-Actin (C4; Santa Cruz Biotechnology, Inc.) antibodies. Rabbit anti-Chk2 antibody (Home- made) was raised against aa 81–95 of mouse Chk2. Rabbit polyclonal anti-Rnf8 antibody was a kind gift from J. Chen (University of Texas, Houston, TX).

BM colony-forming assay. BM cells from femurs of 6–10-wk-old mice were seeded on 35-mm culture dishes (1 × 105 cells/ml in MethoCult GF M3434 media; STEM CELL Technologies Inc.).

Cultures were either left untreated or subjected to various doses of radiation and incubated at

37°C under 5% CO2. Pictures of the dishes were taken, and numbers of colonies were quantified at day 10.

In vitro and in vivo sensitivity to IR. Thymocytes (1 × 106/ml of RPMI medium supplemented with 10% FCS and 5 × 10-5 M 2-ME) were either left untreated or subjected to various doses of

radiation, and then incubated at 37°C under 5% CO2. After 18 hours, the cells were stained with

7AAD or Annexin V and PI and then analyzed on a FACSCalibur. To assess radiosensitivity in vivo, 10 pairs of age-matched WT and Rnf8-/- littermates were subjected to IR (8 Gy). Mice were monitored daily for survival and Kaplan-Meier survival curves were established. A log-rank test was used for statistical analysis.

Chromosomal aberration analysis. Splenocytes from mice of 6–10 wk of age were cultured in the presence of 10 μg/ml LPS for 48 hours. Cells were either left untreated or subjected to 2 Gy

IR. 0.1 μg/ml colcemid was added to each sample 24 hours after irradiation. Cells were then harvested 3 hours after colcemid treatment, followed by hypotonic lysis (0.075 M KCl, 37°C, 15 min) and fixation (methanol 3:1 acetic acid, -20°C, overnight). Fixed cells were dropped onto glass slides (Thermo Fisher Scientific) and were then stained with 0.5 mg/ml DAPI for 10 min (Sigma -

Aldrich), followed by mounting with Mowiol. Chromosome number and chromosomal aberration types were examined under an epifluorescence microscope (DMIRB; Leica). Images were

97 acquired and processed using a digital camera (DC 300RF; Leica) and Image Manager Software

(Leica).

Southern blot analysis. Mouse tail DNA’s were purified by phenol-chloroform extraction. 15g of genomic DNA’s was digested with either Bcl1 for AS0574 clone or BamH1 for RRR260 clone.

Subsequently, standard blotting and hybridization procedures were performed as described in

Molecular Cloning a Laboratory Manual (Orkin, 1990).

Histology analysis. Paraffin sections of tumor and normal tissue were stained with hematoxylin and eosin (H&E) for histological analysis as described previously (Pamidi et al., 2007).

Chapter 3

Rnf8 coordinates Notch signaling and DNA damage repair to

suppress breast cancer

This chapter is composed of the following manuscript, of which I performed all the experiments except Figure 3.5A, 3.12H and 3.14B-D.

Li L., Guturi K.K.N., Bissey P.A., Jackson H.W., Khokha R., Sanchez O., Hakem A., Hakem R. (2015). Rnf8 coordinates Notch signaling and DNA damage repair to suppress breast cancer. (In submission).

3.1 Abstract

Coordination of crosstalk between signaling pathways is essential for maintaining tissue homeostasis as well as for shielding against diseases including breast cancer (Martin-Belmonte and Perez-Moreno, 2012; Visvader, 2009; Visvader and Stingl, 2014). Through catalyzing ubiquitylation of the H2A-type histones flanking DNA double strand breaks (DSBs), the E3 ligase

Rnf8 orchestrates the assembly of components of homologous recombination (HR) and nonhomologous end joining (NHEJ) repair into DSB-induced foci of molecular machineries (Huen et al., 2007; Kolas et al., 2007b; Mailand et al., 2007). Rnf8-initiated ubiquitination cascade is indispensable for the efficient recruitment of Brca1 and 53bp1, key molecules for HR and NHEJ respectively, to repair damaged DNA and maintain genomic stability (Goodarzi and Jeggo, 2013;

Li et al., 2010a; Lukas et al., 2011; Papamichos-Chronakis and Peterson, 2013; Smeenk and van

Attikum, 2013). Here, we report that Rnf8 deficiency in mammary epithelium results in an aberrant luminal expansion that is at least in part mediated via enhanced proliferation. Additional inactivation of p53 exacerbated Rnf8-mutation associated aberrant luminal expansion, and significantly accelerated Rnf8-mutation associated breast cancer incidences. Notably, defective

HR and NHEJ repair systems caused by Rnf8-deficiency render mammary tumor cells hypersensitive to PARP inhibitor- and IR-induced killing. Interestingly, we provide evidence that mechanistically Rnf8 suppresses breast cancer development, probably not only through its role in

DNA repair, but also via its novel function in mediating the regulation of Notch signaling, whose hyperactivation promotes aberrant luminal expansion, hyperplasia and malignant transformation of breast tissue (Andersson and Lendahl, 2014; Bolos et al., 2013; Li et al., 2010b; Mittal et al.,

2009; Pece et al., 2004; Rizzo et al., 2008a; Stylianou et al., 2006; Yao et al., 2011a; Zardawi et al., 2010). Our data indicate that Rnf8 promotes polyubiquitylation and turnover of the Notch1 intracellular domain (NICD, or activated form of Notch1), and thus fine-tunes Notch signaling. 100

Remarkably, genetic restoration of Rnf8 into poorly differentiated Rnf8-deficient mammary tumor cells appreciably retarded tumor growth probably through restoration of their DNA repair capability as well as suppression of Notch signaling, and thereby directly demonstrates the potency of Rnf8-mediated mammary tumor suppressive activity as well as the persistent addiction of Rnf8- deficent mammary tumor cells for the absence of Rnf8 in spite of the presence of other secondary oncogenic mutations. Additionally, pharmacological inhibition of Notch signaling by -secretase inhibitors preferentially restrained the growth of Rnf8-deficient mammary tumors. Overall, these results highlight a novel mechanism for Rnf8-mediated co-regulation of DSB repair and Notch signaling to suppress breast cancer development.

3.2 Introduction

Breast cancer is the most common cancer among women, and also accounts for nearly 400,000 cancer related death per year worldwide (Mego et al., 2010; Rehman et al., 2010). It is a highly heterogeneous disease that can originate from distinct cell types within the mammary epithelium.

(Polyak, 2011). Dissection of the breast cancer subtypes, based on histology, gene expression profiles, and genetic alterations, that constitute this heterogeneity is a major prerequisite to eradicate this leading cause of death in women ( Wood et al., 2007; Chan et al., 2008). Reflective of different mutation profiles and cell of origins, gene expression profiling analyses have further stratified human breast tumors into six distinct subtypes, including the luminal A or B,

HER2/ERBB2 overexpressing, basal-like, normal breast-like, and claudin-low breast tumors

(Sorlie et al., 2003; Sotiriou et al., 2003; Herschkowitz et al., 2007). Classical clinical parameters

101 have also been employed to divide breast cancer into three distinct grades (well differentiated, moderately differentiated, and poorly differentiated) as well as into different subtypes based on status of hormone receptors and amplification of HER2 (Bocker, 2002; Alexandrov et al., 2013).

Although our current knowledge of these subtypes has been somewhat useful in forecasting patient outcome, there exists great variability in patient response to chemotherapy or targeted therapy, implying a better understanding of the etiology of breast cancer as well as mammary gland biology is needed.

The rudimentary mammary gland structure is established during embryogenesis, but much of the mammary gland morphogenesis occurs postnatally. During puberty, mammary stem cells contribute to the formation of terminal end buds which invade into the empty mammary fat pad.

This ductal branching and elongation process continues until week 10 in mouse, at which the entire mammary ductal tree is established and terminal end buds disappear (Visvader, 2009). Besides the importance of Brca1, Palb2, Gata3, and Bmi1 as breast cancer susceptibility genes, their deficiencies all impair mammary gland development and differentiation and their inactivation frequently predisposes to specific subtypes of breast cancer (Antoniou and Easton, 2006;

Apostolou and Fostira, 2013; Campeau et al., 2008; Ghoussaini et al., 2013; Mavaddat et al., 2010;

Mavaddat et al., 2009; Rahman et al., 2007), implying genetic alterations in breast cancer susceptibility genes probably trigger aberrations very early on in specific types of mammary epithelial cells. While increasing evidence suggests mammary epithelium is organized in a hierarchical manner, currently we are able to separate mammary basal cells, luminal progenitors, and luminal differentiated cells based on expression of surface markers including CD49f, CD24 and CD61(Visvader and Stingl, 2014). Notably, the fluctuations of female hormone progesterone

102 during estrus cycle have been reported to play a critical role in mammary epithelial cell homeostasis. During diestrus, progesterone has been shown to promote the expansion of mammary stem cells dramatically by about 14 fold, and to a less extent, concomitantly drive the expansion of luminal cells, compared with other phases of estrus cycle (Asselin-Labat et al., 2010; Joshi et al., 2010). Thus, to rule out the confounding effect of female hormone changes during estrus cycle, it becomes necessary to perform experiments, using estrus phase matched female mice at 10-14 weeks of age when their mammary glands are fully developed, to study mammary gland biology and breast cancer etiology (Joshi et al., 2010; Visvader and Stingl, 2014).

Through ubiquitylation of chromatins flanking DNA double strand breaks (DSBs), Ring Finger

Protein 8 (Rnf8) orchestrates the two major machineries of DNA repair via recruiting repair factors

53BP1 central for nonhomologous end joining (NHEJ) as well as the RAP80-BRCA1 complex essential for homologous recombination (Huen et al., 2007; Kolas et al., 2007b; Mailand et al.,

2007). We reported previously that Rnf8 knockout mice, though develop a broad spectrum of cancers, mainly die of lymphoma and thymoma around one year of age (Li et al., 2010a), while

Rnf8-/- p53-/- double mutant mice all die owing to thymoma before 12 weeks of age with 100% penetrance (Halaby et al., 2013). These very early deaths of Rnf8 mutants, owing to non-solid tumors, precluded the determination of Rnf8’s role in breast cancer. In this study, we found that deletion of Rnf8 leads to an aberrant luminal expansion in mammary glands. Interestingly, this expanded luminal population is maintained to an old age, and is exacerbated in the absence of p53.

To circumvent the very early death of Rnf8-/- p53-/- double mutant mice, we have generated Rnf8-

/-;WapCre;p53fl/fl conditional double mutant females, and discovered that Rnf8 inactivation is a major genetic factor promoting p53 mutation associated breast cancer development.

103

Mechanistically, Rnf8 mediates its breast tumor suppressor function through maintaining genomic stability as well as fine-tuning of Notch signaling. Interestingly, Rnf8 deficiency confers mammary tumors with hypersensitivity to PARP inhibitor- and IR-induced lethality as well as with an increased susceptibility to Notch inhibitor-mediated mammary tumor growth suppression.

3.3 Results & Discussion

3.3.1 Rnf8 is expressed in mammary epithelial subpopulations

Breast cancer is a highly complex disease that develops progressively from early aberrations inside specific mammary epithelial cell of origin (Visvader, 2011). Given that we previously uncovered that Rnf8 knockout mice suffer an elevated mammary tumor risk (Li et al., 2010a), we sought to investigate early cellular aberrations and oncogenic driving mechanisms inside Rnf8-deficient mammary glands. To circumvent progesterone-propelled dynamic shifts of mammary epithelial cell subpopulations during estrus cycle, all experiments in this study were performed using female mice at the estrus phase characterized by a basal or minimum level of progesterone (Asselin-Labat et al., 2010; Joshi et al., 2010) (Figure 3.1A). To study the potential roles of Rnf8 in mammary gland biology, we first asked whether Rnf8 is expressed in mammary epithelial subpopulations.

Real-time RT-PCR analysis revealed that Rnf8 is expressed in cell sorted wild-type (WT) Lin-

(CD45- CD31- TER119-) luminal progenitors (CD49flow CD24+ CD61+), luminal differentiated cells (CD49flow CD24+ CD61-) and basal cells (CD49fhigh CD24+), and thereby warrants us to further investigate the potential role of Rnf8 in mammary epithelial cell biology (Figure 3.1B).

104

Figure 3.1. Rnf8 is expressed in mammary epithelial subpopulations. (A) Tracking of estrus cycle by vaginal smear cytology. Representative images of hematoxylin stained vaginal smears taken at the indicated phases of the estrus cycle. The mouse estrus cycle typically encompasses

105 about 5 days. The period of estrus is represented by the presence of large anucleated cornified squamous-type epithelial cells devoid of leukocytes, and lasts about one day. In contrast, proestrus is populated by well-formed nucleated epithelial cells and leukocytes and lasts about one day, while metestrus is indicated by about equal numbers of large cornified epithelial cells and leukocytes and last one day. Diestrus is predominated by leukocytes and typically spans about 1 to 2 days. (B) Expression levels of Rnf8 in different mammary epithelial subpopulations at estrus phase (mean ± s.e.m.). Bar chart depicting realtime RT-PCR analysis of cell sorting purified Lin-

CD49flow CD24+ CD61- luminal differentiated cells, Lin- CD49flow CD24+ CD61+ luminal progenitor cells, and Lin- CD49fhigh CD24+ basal cells from 12 week-old WT females (n=6) at estrus phase.

106

3.3.2 Rnf8 deficiency leads to an aberrant luminal expansion

Interestingly, flow cytometry (FACS) analysis of mammary epithelial subpopulations revealed that 10-14 week-old Rnf8-/- females harbored an aberrantly expanded luminal compartment compared with their WT littermates, while there is no statistically significant difference in their basal compartments (Figure 3.2A-C). Further fractionation of Lin- CD49flow CD24+ luminal compartment into luminal progenitor cells and luminal differentiated cells revealed significant increases in both CD61+ luminal progenitor and CD61- luminal differentiated epithelial subpopulations (Figure 3.2A-C), implying either an increased proliferation ability of luminal progenitor cells or skewed commitment to luminal lineage from mammary stem/bi-potent progenitor cells.

The 3-dimensional (3D) Matrigel mammary colony forming cell (CFC) assay provides an in vitro readout of the clonogenic potentials of progenitor cells (Shackleton et al., 2006; Joshi et al., 2010).

To preserve in vivo luminal-basal crosstalk, we used total mammary cells from 10-14 weeks old

WT and Rnf8-/- females at estrus phase to examine the potential impact of Rnf8 loss on mammary colony forming potentials of progenitor cells. Interestingly, the number of 3D mammary colonies formed increased about 2-fold in the absence of Rnf8, and we consistently observed that Rnf8-/- mammary cells gave rise to colonies that were much larger in size compared with WT controls

(Figure 3.2D and E), suggesting an increase in colony forming and proliferation capabilities of

Rnf8-/- progenitors.

107

We hypothesized that Rnf8 deficiency expands the luminal compartment of mammary epithelium at least in part through increased cellular proliferation. To address this, we performed Ki67 immunohistochemistry on WT and Rnf8-/- mammary glands at estrus phase. In accordance with an increased clonogenic potential in the absence of Rnf8, Rnf8-/- mammary glands harbored an about

4-fold increase in proliferating (Ki67+) cells compared to WT littermates (Figure 3.2F and G). As determined microscopically by their cuboidal-shape and inner localization adjacent to the lumen of mammary ducts, a majority of proliferating cells in Rnf8-/- mammary glands were luminal cells.

Collectively, these data indicate that Rnf8 deficiency promotes an aberrant expansion of mammary luminal epithelial cells at least in part through increased cellular proliferation.

108

Figure 3.2. Rnf8-deficiency promotes the luminal compartment of mammary glands to aberrantly expand through enhanced proliferation. (A) Representative FACS plots showing

Lin- (CD45- CD31- TER119-) luminal cells (CD49flow CD24+) and basal cells (CD49fhigh CD24+) of 4th inguinal mammary glands from WT and Rnf8-/- littermate females at estrus phase. (B) CD24

109 positive mammary epithelial cells from (A) were further analyzed for CD61 expression to enrich for CD61+ luminal progenitor cells (CD49flow CD24+ CD61+). (C) Bar chart depicting the absolute numbers of basal cells, luminal progenitor cells, and luminal differentiated cells per 4th inguinal mammary gland from 10-14 week-old WT (n=7) and Rnf8-/- (n=7) females at estrus phase (mean

± s.e.m.). (D) Representative composite tiled images of Matrigel droplets (50l in volume) harboring WT and Rnf8-/- mammary colonies at day 12 of culture. (E) Bar chart depicting the clonogenic potentials of 5,000 mammary cells derived from WT and Rnf8-/- littermate females at estrus phase in 50l Matrigel (mean ± s.e.m.). Four independent experiments were performed. (F)

Representative images of Ki67 immunohistochemistry of mammary glands from WT and Rnf8-/- females at estrus phase. Bar: 50m. (G) Bar chart depicting the quantification of Ki67 positive cells from WT and Rnf8-/- mammary glands (mean ± s.e.m.). Three independent experiments were performed, and a minimum of 1000 cells were counted per gland. * P<0.05; n.s., not significant, obtained from two-sided Student’s t-test.

110

3.3.3 Rnf8 loss increases DNA damage susceptibility and leads to focal hyperplasia in mammary glands

Mammary gland is an organ that is destined to continuous cycles of proliferation, differentiation and apoptosis during female reproductive cycles, and thereby requires an increased usage of DDR pathways in order to ensure high fidelity of genetic information maintenance and propagation

(Visvader and Stingl, 2014). Since Rnf8 is known to play a central role in propagating DNA damage signals, we hypothesized that Rnf8-/- mammary epithelial cells will be more susceptible to

DNA damage. To examine the levels of DNA breaks in Rnf8-/- mammary epithelium, we depleted lineage negative cells (CD31+ endothelial cells, CD45+ immune cells and TER119+ red blood cells) from total mammary cells of 10-14 weeks old WT and Rnf8-/- females at estrus phase via magnetic beads, and checked their purity by flow cytometry before and post purification (Figure 3.3E). The purified WT and Rnf8-/- mammary epithelial cells were either left untreated or treated with ionizing radiation (IR), and were cytospan onto glass slides and stained for H2ax, a marker of DNA breaks.

In accordance with the critical role of Rnf8 in DSB signaling and repair (Huen et al., 2007; Kolas et al., 2007b; Li et al., 2010a; Mailand et al., 2007), Rnf8-/- mammary epithelial cells harbored a statistically significantly increase in H2ax marked DNA breaks even in the untreated condition compared with WT controls (Figure 3.3A-D), implying that overtime these mutant cells may have an increased probability of acquiring secondary mutagenic lesions required to drive malignant transformation.

111

Since hyperplasia is a premalignant hallmark for breast cancer development (Bombonati and Sgroi,

2011), we examined its occurrence in about 13-month-old WT and Rnf8-/- females. Whole mount and histopathology analyses revealed that, at this age, mammary glands from 12 out of 20 Rnf8-/- females displayed focal hyperplastic lesions, whereas none of the 19 WT control littermates contained such lesions (Figure 3.4A-C). Interestingly, we observed Rnf8-/- mammary glands harboring areas of typical hyperplasia with no adipose tissue present as well as atypical papillary hyperplasia with irregular dark nuclei (Figure 3.4C). Overall, the above data suggest that Rnf8- deficiency in mammary epithelium confers an increase of cellular susceptibility to DNA damage that overtime frequently induces focal hyperplastic lesions.

112

Figure 3.3. Rnf8-/- primary mammary epithelial cells display elevated levels of DNA breaks.

(A) Cells from mammary glands of WT and Rnf8-/- littermate females at estrus phase were negatively depleted for CD45+ immune cells, CD31+ endothelial cells and TER119+ red blood cells by magnetic beads. Purified mammary epithelial cells, either left untreated (UT) or harvested at

113 the indicated time-points post 5Gy IR, were freshly cytospun onto glass slides and stained with anti-H2ax antibody. Images were taken with a fluorescent microscope. UT: untreated. (B-D)

Quantitative analysis of H2ax subnuclear foci in purified WT and Rnf8-/- mammary epithelial cells. Cells harboring ≥ three foci were counted as foci-positive cells. Bar graph shows mean ± s.e.m. of four independent experiments, and a minimum of 100 cells were counted for each condition and genotype. Two-sided Student’s t-test: * P<0.05, n.s.: not significant. (E) FACS plots showing the purity of the isolated mammary epithelial cells, before and after negative depletion of

CD45+ immune cells, CD31+ endothelial cells and TER119+ red blood cells by magnetic beads.

114

Figure 3.4. Loss of Rnf8 results in focal regional hyperplasia in mammary epithelium, and increases cancer risks. (A) Representative carmine-alum stained whole mount of inguinal mammary glands from 13 month-old WT and Rnf8-/- females. Boxed regions are shown at higher

115 magnification. (B) Bar chart depicting the proportion of Rnf8-/- females (n=20) that harbored focal hyperplastic lesions compared with WT (n=19) controls at about 13 months of age. (C)

Representative H&E stained images of thoracic mammary glands from the same cohorts of WT and Rnf8-/- females as in (A and B). Whereas histopathological analysis revealed the absence of hyperplasia in the age-matched WT mammary glands, the Rnf8-/- mammary glands contained areas of typical hyperplasia with no dispose tissue present (middle) as well as atypical papillary hyperplasia with irregular dark nuclei (right). (D) Kaplan Meier total tumor-free survival curves of cohorts of WT (n=23), Rnf8-/- (n=33), WapCre;p53fl/fl (n=28), WapCre;p53fl/wt (n=12), Rnf8-/-

;WapCre;p53fl/wt (n=16), and Rnf8-/-;WapCre;p53fl/fl (n=31) females. The cohorts of female mice were monitored for cancer development twice a week for 618 days. Log rank tests indicate statistically significant differences between Rnf8-/-;WapCre;p53fl/fl and WapCre;p53fl/fl curves

(P<0.00001), between Rnf8-/-;WapCre;p53fl/wt and WapCre;p53fl/wt curves (P<0.003), and between

Rnf8-/- and WT curves (P<0.000005).

116

3.3.4 p53 loss exacerbates Rnf8-deficiency associated aberrant luminal expansion

p53 is essential for the cellular response to DNA damage (Donehower and Lozano, 2009). Its expression as well as the levels of its downstream targets p21 and Bax were elevated in mammary glands of one year old Rnf8-/- females compared to WT littermates (Figure 3.5A). Given that p53 is also a major breast tumor suppressor and Rnf8-/-p53-/- mice all die exclusively of thymomas before 12 weeks of age (Halaby et al., 2013), we hypothesized that p53 activity plays a role in restraining Rnf8-deficiency associated aberrant luminal expansion. Thus, we circumvented the very early death of Rnf8-/- p53-/- double mutants by conditionally deleting p53 using the WapCre system, in which the wap acidic protein (Wap) promoter directs Cre recombinase expression specifically to mammary epithelium.

Previous studies suggest that the expression of WapCre transgene is restricted to differentiating mammary epithelium during the estrous cycle, lobuloalveologenesis and parity-identified stem cells during pregnancy (Li et al., 2002; Kordon et al., 1995; Robinson et al., 1995). The WapCre deleter line was crossed with p53fl mice, which harbor loxP sites in intron 1 and intron 10 of p53 allele (Jonkers et al., 2001), yielding WapCre;p53fl/fl mice. We first determined the cellular subtypes within the mammary epithelium in which p53 is deleted by WapCre. To determine the efficiency of WapCre-mediated p53 deletion in mammary subpopulations, we cell sorted Lin- luminal differentiated cells, luminal progenitors, and basal cells from 14 week-old WapCre;p53fl/fl virgin female mice at estrus phase, 4-month old WapCre;p53fl/fl female mice at estrus phase that

117 had undergone 2 pregnancies and involution, and 6.5-month old WapCre;p53fl/fl female mice at estrus phase that had undergone 4 pregnancies and involution. Subsequently, we extracted DNA from each fraction, and assessed Cre-mediated recombination efficiency by nested step down PCR analysis for p53 deletion exonΔ2-10 product (612bp), non-recombined 5’ loxP product of p53fl allele

(370bp) and germline WT product of p53 (288bp). Notably, three previous studies independently reported that WapCre expression is present only at estrus phase of the estrus cycle in mammary glands of virgin female mice (Kordon et al., 1995; Robinson et al., 1995; Wagner et al., 2002), but there lacks a knowledge on the mammary epithelial subpopulations in which WapCre is expressed at estrus phase in virgin WapCre transgenic female mice. Interestingly, we revealed that, in 14 week-old WapCre;p53fl/fl virgin female mice at estrus phase, the recombined p53 deletion exonΔ2-

10 product was detected only in CD61+ luminal progenitor cells, but was undetectable in other mammary epithelial subpopulations (Figure 3.6C). Furthermore, after 2 pregnancies and involution, 4-month old WapCre;p53fl/fl female mice at estrus phase harbored a significant amount of the recombined p53 deletion exonΔ2-10 product in luminal progenitors and luminal differentiated cells, but had only a low amount of the recombined p53 deletion exonΔ2-10 product in basal cells.

Moreover, after 4 pregnancies and involution, 6.5-month old WapCre;p53fl/fl female mice at estrus phase showed a complete deletion of p53 in both luminal progenitor and luminal differentiated compartments, though the deletion efficiency is much slower in basal cells (Figure 3.6C).

Therefore, we first generated Rnf8-/-;WapCre;p53fl/fl conditional double mutant and investigated the impact of additional deletion of p53 on the aberrantly expanded luminal compartment of Rnf8-

/- mammary epithelium after 4 pregnancies and involution at estrus phase. We harvested both 4th inguinal mammary glands from 6.5-month old WT, Rnf8-/-, WapCre;p53fl/fl, and Rnf8-/-;WapCre;

118 p53fl/fl females at estrus phase that had undergone 4 pregnancies and involution to conditionally delete p53 completely in luminal compartment of mammary epithelium. Interestingly, FACS analysis of mammary epithelial cell subpopulations revealed that the aberrantly expanded luminal compartment in Rnf8-/- females was maintained to this old age, and was remarkably exacerbated by additional loss of p53 (Figure 3.5B and Figure 3.6A-C). Additionally, we examined the thoracic mammary glands from these females by carmine-alum whole mount staining and found no sign of hyperplasia (Figure 3.6D), indicating this exacerbation of Rnf8-deficiency associated aberrant luminal expansion is not owing to and precedes hyperplasia. Collectively, these data suggest that p53 actively restrains Rnf8-deficiency associated aberrant luminal expansion.

119

Figure 3.5. Deletion of p53 exacerbates the aberrant luminal expansion and breast cancer risks associated with Rnf8 mutation. (A) Immunoblot analysis of p53, p21 and bax expression in mammary glands of one year old WT and Rnf8-/- females. (B) Representative FACS plots showing Lin- luminal cells (CD49flow CD24+) and basal cells (CD49fhigh CD24+) of mammary glands from 6.5-month-old WT (n=4), Rnf8-/- (n=4), WapCre;p53fl/fl (n=4), and Rnf8-/-

;WapCre;p53fl/fl (n=4) females at estrus phase that have undergone 4 pregnancies and involution

120 to completely delete p53 in luminal compartment. (C) Kaplan Meier mammary tumor-free survival curves of cohorts of WT (n=23), Rnf8-/- (n=33), WapCre;p53fl/fl (n=28), WapCre;p53fl/wt (n=12),

Rnf8-/-;WapCre;p53fl/wt (n=16), and Rnf8-/-;WapCre;p53fl/fl (n=31) females. The cohorts of female mice were monitored for breast tumor onset by palpation twice a week for 618 days. Log rank tests indicate statistically significant differences between Rnf8-/-;WapCre;p53fl/fl and WapCre;p53fl/fl curves (P<0.00005), between Rnf8-/-;WapCre;p53fl/wt and WapCre;p53fl/wt curves (P<0.003), and between Rnf8-/- and WT curves (P<0.005). (D) Bar chart depicting frequencies of metastasis by

Rnf8-/-, WapCre;p53fl/fl, and Rnf8-/-;WapCre;p53fl/fl mammary tumors. (E) Representative images of CK18 and CK14 immunohistochemistry analysis of poorly differentiated mammary tumors from Rnf8-/-, WapCre;p53fl/fl, and Rnf8-/-;WapCre;p53fl/fl females. Bar: 50m. (F) Representative

FACS plots showing Lin- luminal cells (CD49flow CD24+) and basal cells (CD49fhigh CD24+) of mammary tumors from Rnf8-/-, WapCre;p53fl/fl and Rnf8-/-;WapCre;p53fl/fl survival cohorts, with mammary glands from a one year old WT female as control.

121

Figure 3.6. The aberrantly expanded luminal compartment in Rnf8-/- mammary epithelium is exacerbated in the absence of p53. (A) Representative FACS plots showing expression of

CD61 and CD49f in CD24 positive cells, gated from Figure 3.5B, to enrich for CD61+ luminal progenitors. (B) Bar chart depicting the absolute number of luminal cells per 4th inguinal mammary

122 gland from WT (n=4), Rnf8-/- (n=4), WapCre;p53fl/fl (n=4), and Rnf8-/-;WapCre;p53fl/fl (n=4) female mice at estrus phase (mean ± s.e.m.). All female mice have undergone 4 pregnancies and involution to completely delete p53 in luminal compartment. * P<0.05, two-sided Student’s t-test. (C)

Analysis of WapCre-mediated p53 deletion efficiency in cell sorting purified mammary epithelial subpopulations. Nested step-down PCR analysis of Cre and p53 alleles (deletion exon2-10 (p53Δ2-

10), p53fl, and p53WT) in tail DNA’s from controls (WapCre, WapCre;p53fl/fl, and p53fl/WT female mice), as well as purified DNA’s from cell sorted LD (CD49flow CD24+ CD61- luminal differentiated cells), LP (CD49flow CD24+ CD61+ luminal progenitor cells), and Ba (CD49fhigh

CD24+ basal cells) from WapCre;p53fl/fl virgin female mice at estrus phase, WapCre;p53fl/fl pregnancy 2 and 4 (female mice at estrus phase that have undergone 2 and 4 pregnancies and involution, respectively). (D) Carmine-alum whole mount analysis of the respective thoracic mammary glands from the same female mice revealed the absence of hyperplasia.

123

3.3.5 p53 inactivation accelerates Rnf8-mutation associated breast cancer incidences

For additional deletion of p53 exacerbates the aberrant expansion of Rnf8-/- luminal progenitors

(Figure 3.5B and Figure 3.6), we hypothesized that p53 is a major genetic factor suppressing malignant transformation inside the aberrantly expanded luminal compartment of Rnf8-/- female mice. Through monitoring breast cancer development in cohorts of Rnf8-/-, Rnf8-/-;WapCre;p53fl/fl,

Rnf8-/-;WapCre;p53fl/wt, WapCre;p53fl/fl, WapCre;p53fl/wt, and WT female mice for 618 days by palpation at least twice a week, we observed that while 6 out of 33 Rnf8-/- female mice developed mammary adenocarcinoma though all of them died of thymoma and/or lymphoma, additional deletion of p53 significantly accelerated Rnf8-mutation associated breast cancer incidences

(Figure 3.5C and Figure 3.4D). In accordance with their aberrant expansion of luminal progenitors,

Rnf8-/- and Rnf8-/-;WapCre;p53fl/fl female mice developed mammary tumors that exhibited positivity predominantly for luminal marker cytokeratin (CK) 18, but not basal marker CK14

(Figure 3.5E). Interestingly, these Rnf8-deficient tumors, as well as mammary tumors from

WapCre;p53fl/fl females, displayed luminal progenitor surface expression profiles (Lin- CD49flow

CD24+ CD61+) (Figure 3.5F and Figure 3.7C). Furthermore, the recombined p53 deletion exonΔ2-

10 allele was detected in observed mammary tumors from Rnf8-/-;WapCre;p53fl/fl and

WapCre;p53fl/fl survival cohorts (Figure 3.8D), further confirming the deletion of p53.

Additionally, our data also indicated increased risks for Rnf8-/-;WapCre;p53fl/wt females to develop mammary tumors which showed loss of heterozygosity for WT p53 allele (Figure 3.5C and Figure

3.7D). Overall, these results highlight the importance of Rnf8 in safeguarding the luminal

124 compartment and its collaboration with p53 in suppressing malignant transformation of breast tissue.

3.3.6 Rnf8 loss increases mitotic index and metastatic frequencies of p53 mutation associated mammary tumors

Notably, histopathological analyses revealed that mammary tumors Rnf8-/-;WapCre;p53fl/fl females were undergoing active mitosis and displayed statistically significantly higher mitotic indices compared to mammary tumors from single mutant females (Figure 3.7A and B).

Furthermore, through analyzing all the organs from these females that developed mammary tumors, we observed that mammary tumors from Rnf8-/-;WapCre;p53fl/fl females exhibited a statistically significant increase of frequency of metastasis compared to mammary tumors from single mutant females (Figure 3.5D and Figure 3.8A-C). Interestingly, only lung metastasis by mammary tumors was found in Rnf8-/- and WapCre;p53fl/fl single mutant females (Figure 3.8A and

B), whereas mammary tumors from Rnf8-/-;WapCre;p53fl/fl females were able to infiltrate into not only lung but also brain, dermis, lymph node and salivary gland (Figure 3.8C). Collectively, these data suggest that dual deletions of Rnf8 and p53 not only accelerate breast cancer incidences but also increase the aggressiveness of mammary tumors.

125

Figure 3.7. Rnf8 inactivation increases the frequency of mitosis in p53 mutation associated mammary tumors. (A) Representative images of H&E stained mammary adenocarcinomas, subcategorized into well, moderately, or poorly differentiated mammary tumors, from Rnf8-/-,

WapCre;p53fl/fl, and Rnf8-/-;WapCre;p53fl/fl female mice. Bar: 50m. (B) Bar chart depicting the

126 mitotic indices for the observed mammary tumors from Rnf8-/-, WapCre;p53fl/fl, and Rnf8-/-

;WapCre;p53fl/fl female mice. (C) Representative FACS plots showing expression of CD61 and

CD49f on CD24 positive mammary tumor cells, gated from Figure 3.5F, to enrich for CD61+ luminal progenitor tumor cells. (D) Southern blot analysis of loss of heterozygosity for WT p53 allele from Rnf8-/-;WapCre;p53fl/wt mammary tumors. C: control tail DNA; T: mammary tumor.

127

Figure 3.8. Rnf8 loss exacerbates the incidences of metastasis of p53 mutation associated mammary tumors. (A) Representative H&E stained Rnf8-/- mammary tumor cells infiltrating into the lung. (B) Representative H&E stained WapCre;p53fl/fl mammary tumor cells infiltrating into the lung. (C) Representative H&E images showing Rnf8-/-;WapCre;p53fl/fl mammary tumor

128 cells infiltrating into the lung, dermis, brain, lymph node, and salivary gland. (D) Confirmation of p53 deletion in mammary tumors from Rnf8-/-;WapCre;p53fl/fl and WapCre;p53fl/fl females. PCR analysis of p53-deleted allele (p53Δ2-10) using DNA’s from Rnf8-/-; WapCre;p53fl/fl and

WapCre;p53fl/fl tails (Control: C) and mammary tumors (Tumor: T).

3.3.7 Rnf8 complementation into Rnf8-/- p53∆/∆ mammary tumor cells restores their DSB signaling and repair efficiency

Deregulation of DNA DSB or Notch signaling pathways have been associated with malignant transformation of mammary epithelial cells (O'Driscoll, 2012; Andersson and Lendahl, 2014).

Therefore we examined these signaling pathways in three independent poorly differentiated Rnf8-

/-;WapCre;p53fl/fl (Rnf8-/-;p53Δ/Δ) luminal tumor cell lines, complemented with either empty Flag- vector (mock), or Rnf8-Flag. While recruitment of 53bp1 to DNA damage sites is essential for

NHEJ-mediated repair, accumulations of Brca1 and Rad51 molecules to DNA damage sites are required for HR-mediated repair (Ciccia and Elledge, 2010). Rnf8-initiated ubiquitination at the sites of DNA DSB breaks have been shown to orchestrate the recruitment of these key components of HR and NEHJ repair machineries (Jackson and Durocher, 2013). Thus, we hypothesized that

Rnf8 restoration in these Rnf8-/-;p53Δ/Δ luminal tumor cell lines will rescue their DSB signaling defects and ameliorates their abilities to repair DNA DSBs. Notably, Rnf8 complementation of

129 these mammary tumor cells significantly enhanced their ability to recruit HR repair factors Brca1 and Rad51 as well as NHEJ repair factor 53bp1, to DSB sites in the untreated condition as well as in response to IR (Figure 3.9A-D). In accordance with their enhanced efficiencies of recruitments of DNA repair factors, the levels of H2ax marked DNA breaks in Rnf8-restored mammary tumor cells were significantly ameliorated compared with their isogenic Rnf8-deficient mammary tumor cells in the untreated condition as well as 24 hours post-IR (Figure 3.9 and Figure 3.10). To further confirm their HR and NHEJ repair efficiencies, we employed HR and NHEJ reporter assays. We linearized and introduced DSBs into HR and NHEJ reporter plasmids through I-SceI enzymatic digestion. Notably, in accordance with their enhanced efficiency to repair γ-H2ax marked DNA breaks, Rnf8-restored mammary tumor cells exhibited ameliorated abilities to re-join I-SceI- induced DSBs through HR and NHEJ machineries compared to their isogenic Rnf8-deficient mammary tumor cells (Figure 3.11A). Collectively, these results highlight the importance of Rnf8 in orchestrating HR and NHEJ machineries to repair DNA DSBs in mammary tumors.

130

131

Figure 3.9. Rnf8-deficient mammary tumors display defective DSB signaling and elevated levels of spontaneous DNA breaks. (A) Representative images of spontaneous H2ax, 53bp1,

Brca1, and Rad51 subnuclear foci formation in Rnf8-/-;p53Δ/Δ mammary tumor cells reconstituted with either empty vector (Mock) or Rnf8. (B) Bar charts depicting quantifications of H2ax, 53bp1,

Brca1, and Rad51 subnuclear foci formation observed in (A). (C) Representative images of H2ax,

53bp1, Brca1, and Rad51 subnuclear foci formation 6h post 5Gy IR in Rnf8-/-;p53Δ/Δ mammary tumor cells reconstituted with either empty vector (Mock) or Rnf8. (D) Bar charts depicting quantifications of H2ax, 53bp1, Brca1, and Rad51 subnuclear foci formation observed in (C).

Cells harboring ≥10 foci were counted as foci-positive. Three independent experiments were performed, and a minimum of 100 cells were counted for each condition and genotype (mean ± s.e.m.). * P<0.05, ** P<0.01, *** P<0.001, obtained from two-sided Student’s t-test.

132

Figure 3.10. Complementation of exogenous Rnf8 into Rnf8-/-;p53Δ/Δ mammary tumor cells restored their ability to repair H2ax-marked DNA breaks. (A) Representative images of

H2ax marked DSBs in Rnf8-deficient mammary tumor cells as well as their Rnf8-reconstituted isogenic cells either left untreated, or subjected to 5Gy of -radiation followed by recovery at 6h and 24h post irradiation. (B-D) Bar charts depicting quantifications of H2ax marked foci of DSBs

(mean ± s.e.m.). Tumor cells harboring ≥10 foci were counted as foci-positive cells, and a

133 minimum of 100 cells were counted per condition and genotype. Three independent experiments were performed. * P<0.05; ** P<0.01; n.s., not significant (two-sided Student’s t-test).

3.3.8 Rnf8-deficiency confers mammary tumors with hypersensitivity to

PARP inhibitor and IR treatments

The exploitation of cancer cells’ deficiency in components of DNA damage repair have demonstrated to be useful to preferentially kill tumor cells (Hosoya and Miyagawa, 2014). This is probably best exemplified by PARP inhibitor-induced synthetic lethality of breast cancer cells harboring defects in HR repair machinery (Gilardini Montani et al., 2013; Lee et al., 2014). In accordance with their defective HR-mediated repair of DNA DSBs (Figure 3.9, Figure 3.10 and

Figure 3.11A), Rnf8-/-;p53Δ/Δ mammary tumor cells displayed statistically significant hypersensitivity to PARP inhibitor KU0058948 treatments at various doses compared with their

Rnf8-restored isogenic tumor cells (Figure 3.11B and C). Additionally, in accordance with their defective NHEJ and HR mediated repair of DNA DSBs (Figure 3.9, Figure 3.10 and Figure 3.11A),

Rnf8-/-;p53Δ/Δ mammary tumor cells also exhibited statistically significant hypersensitivity to - radiation treatments at different doses compared with their Rnf8-restored isogenic tumor cells

(Figure 3.11B and D). Collectively, these results suggest that PARP inhibitor and IR treatments can be used to preferentially kill Rnf8-deficient mammary tumors.

134

Figure 3.11. Rnf8-deficient mammary tumors display defective DSB repair pathways and hypersensitivity to PARP inhibitors and -radiation. (A) Bar charts depicting efficiencies of

HR and NHEJ mediated repair of I-SceI-induced DSBs in Rnf8-/-;p53Δ/Δ mammary tumor cells reconstituted with either empty vector (Mock) or Rnf8(mean ± s.e.m.). Tumor cells were co- transfected with I-SceI linearized reporter constructs for HR or NHEJ and DsRed expression vector

135 to normalize for the differences in transfection efficiency. Three independent experiments were performed. * P<0.05, two-sided Student’s t-test. (B) Representative images of clonogenic survival of Rnf8-/-;p53Δ/Δ mammary tumor cells reconstituted with either empty vector (Mock) or Rnf8 at day 12 post M PARP inhibitor (KU0058948) and 4Gy -radiation treatments. (C) Bar chart depicting clonogenic survival of Rnf8-/-;p53Δ/Δ mammary tumor cells reconstituted with mock or

Rnf8 after 12 days of continuous exposure to PARP inhibitors at various concentrations (mean ± s.e.m.). (D) Bar chart depicting clonogenic survival of Rnf8-/-;p53Δ/Δ mammary tumor cells reconstituted with mock or Rnf8 at day 12 after exposure to different doses of IR (mean ± s.e.m.).

Three independent experiments were performed. * P<0.05, ** P<0.01, *** P<0.001, obtained from two-sided Student’s t-test.

136

3.3.9 Notch signaling is hyperactivated in Rnf8-/- luminal progenitors

To gain an insight into the mechanisms for these novel functions of Rnf8 in mammary gland development and malignant transformation, we first cell-sorted luminal differentiated cells, luminal progenitors, and basal cells from WT and Rnf8-/- mammary glands, and investigated the expression of genes whose alterations have been associated with aberrant luminal expansion

(Bouras et al., 2008; Carr et al., 2012; Chakrabarti et al., 2012; Lim et al., 2009b; Pei et al., 2009;

Visvader, 2009). Remarkably, we uncovered that Rnf8 deficiency in CD61+ luminal progenitors significantly increased expression levels of Notch target genes, Hes1, Hes5, Gata3, Slug, and

Cyclin D1 (Figure 3.12A). Furthermore, expression levels of Cyclin D2, a proliferation marker similar to Cyclin D1, were also significantly increased in Rnf8-/- CD61+ luminal progenitors compared with WT controls (Figure 3.12A), further supporting active proliferation of these mutant progenitors (Figure 3.2D-G). Interestingly, expression levels of amphiregulin, a growth factor for driving epithelial cell proliferation in an autocrine and/or paracrine manner, were also increased in Rnf8-/- luminal progenitors and differentiated cells compared with WT controls (Figure 3.12A and Figure 3.13). Although expression levels of Notch target genes were not affected in Rnf8-/- luminal differentiated cells compared with WT controls, levels of Cyclin D1 were significantly elevated in these cells (Figure 3.13), implying their active proliferation was probably Notch- independent and was likely triggered extrinsically by the growth promoting activity of amphiregulin. Consistent with hyperactivation of Notch signaling in Rnf8-/- luminal progenitors, immunoblot (IB) analysis indicated elevated levels of NICD, the activated form of Notch1, and its canonical target Hes1 in purified mammary epithelial cells from Rnf8-/- female mice compared with WT littermates (Figure 3.12B). Overall, these results demonstrate that Rnf8 deficiency promotes hyperactivation of Notch signaling, specifically in CD61+ luminal progenitors; thus providing a mechanism for the aberrant luminal expansion associated with Rnf8 deficiency.

137

Figure 3.12. Rnf8 suppresses mammary tumor growth via negative regulation of Notch signaling. (A) Realtime quantitative RT-PCR analysis of CD61+ luminal progenitors cell-sorting purified from mammary glands of WT (n=3) and Rnf8-/- (n=3) littermate female mice at estrus phase (mean ± s.e.m.). (B) Immunoblot analysis of NICD, and Hes1 on magnetically purified

138 mammary epithelial cells from WT and Rnf8-/- littermate female mice at estrus phase. -Actin was used as the loading control. (C) Top panel showing representative microscopic images of poorly differentiated Rnf8-/-;p53Δ/Δ mammary tumor cells, complemented with either Flag-empty vector

(left) or Rnf8-Flag (right), that were allowed to grow for 48h post seeding 3x105 cells onto 6 cm dishes, followed by fixation and crystal violet staining of the same dishes (bottom panel). (D)

Cumulative growth curves of 3 independent lines of Rnf8-deficient mammary tumor cells and their

Rnf8-restored isogenic cells (mean ± s.e.m.). 3x105 cells were seeded onto 6 cm dishes in triplicates, followed by trypsinization, counting, and re-seeding at the same density after every

72h. (E) Immunoblot analyses were performed to examine expression levels of the indicated proteins in Rnf8-deficient mammary tumor cells and their Rnf8-restored isogenic cells. -Actin was used as a loading control. (F) CBF1/Notch dual luciferase reporter assay was performed using

Rnf8-deficient mammary tumor cells and their Rnf8-restored controls (mean ± s.e.m.). pNICD was co-transfected as the positive control, and pRL-TK-renilla luciferase reporter was employed as the internal control. (G) Immunohistochemistry analysis of NICD on Rnf8-/-;p53Δ/Δ and p53Δ/Δ poorly differentiated mammary luminal tumors. (H) MDA-MB-231 human breast cancer cells were transfected with Flag-empty vector or Rnf8-Flag, followed by immunoblot analysis of their levels of NICD. -Actin was used as a loading control. * P<0.05; ** P<0.01; *** P<0.001; n.s., not significant (two-sided Student’s t-test).

139

Figure 3.13. Hyperproliferation of Rnf8-/- luminal differentiated cells is probably Notch- independent. (A) Cell sorting purified Lin- CD49flow CD24+ CD61- luminal differentiated cells from WT (n=3) and Rnf8-/- (n=3) littermate female mice at estrus phase were examined by realtime quantitative RT-PCR for expression of the indicated genes (mean ± s.e.m.). (B) Cell sorting purified Lin- CD49fhigh CD24+ basal cells from WT (n=3) and Rnf8-/- (n=3) littermate female mice at estrus phase were examined by real-time quantitative RT-PCR for the indicated target genes

(mean ± s.e.m.). Two-sided Student’s t-test: * P<0.05; ** P<0.01; n.s., not significant.

140

3.3.10 Genetic restoration of Rnf8 into Rnf8-/- p53∆/∆ mammary tumor cells suppresses their growth via dampening Notch signaling

Astonishingly, while Rnf8 complementation of Rnf8-/-;p53Δ/Δ mammary tumor cell lines rescued their DSB repair defects (Figure 3.9, Figure 3.10 and Figure 3.11A), this genetic restoration of

Rnf8 into these poorly differentiated Rnf8-deficient mammary tumor cells also significantly suppressed their tumor growth (Figure 3.12C and D), directly demonstrating the potency of Rnf8- mediated mammary tumor suppressive activity as well as the persistent addiction of Rnf8-deficent mammary tumor cells for the absence of Rnf8 in spite of the presence of other secondary oncogenic mutations. Furthermore, this observation implies that the exogenous Rnf8 ameliorates not only

DSB signaling but likely also other signaling pathways that might have been de-regulated to drive transformation of Rnf8-deficient mammary epithelial cells. Since Notch signaling is hyperactivated in Rnf8-deficient luminal progenitors, we hypothesized, that Notch signaling is hyperactivated in these poorly differentiated Rnf8-/-;p53Δ/Δ mammary luminal tumor cell lines, and that aberrant Notch activation promotes their tumor growth. Indeed, our IB analyses revealed that complementation of Rnf8-/-;p53Δ/Δ mammary tumor cells with exogenous Rnf8 significantly dampened their levels of NICD, Hes1, Cyclin D1, and pAkt (Figure 3.12E). Hyperactivation of

Notch signaling in Rnf8-/-;p53Δ/Δ mammary tumor cells and its downregulation by exogenous Rnf8 were further confirmed using the CBF1/Notch dual luciferase reporter assay (Figure 3.12F).

Consistently, immunohistochemistry revealed significantly elevated levels of NICD in primary

Rnf8-/-;WapCre;p53fl/fl mammary tumors compared to WapCre;p53fl/fl mammary tumors (Figure

3.12G). Notably, over-expression of exogenous Rnf8 in the human breast cancer cells MDA-MB-

231 significantly diminished their NICD levels (Figure 3.12H). Collectively, these results unravel

141 a novel function of Rnf8 in the negative regulation of NICD abundance and Notch signaling that contributes to its mammary tumor suppressor role.

3.3.11 Rnf8 promotes polyubiquitination and turnover of NICD

To explore the mechanisms behind Rnf8-mediated downregulation of NICD, we first examined whether Rnf8 interacts with NICD. Immunoprecipitation (IP) of Flag in Rnf8-/-;p53Δ/Δ mammary tumor cells complemented with Rnf8-Flag, but not Flag-empty vector, pulled down NICD (Figure

3.14A), demonstrating Rnf8-NICD interaction. The turnover and activity of NICD is controlled by a number of E3 ligases such as Fbw7, and deregulation of these ubiquitylation events have been associated with breast cancer development (Andersson and Lendahl, 2014). Based on Rnf8’s interaction with NICD, we performed intracellular ubiquitylation assays and examined the ability of Rnf8 to mediate ubiquitylation of NICD. Notably, co-expression of exogenous Rnf8 and

Ubiquitin (Ub) in HEK293T cells produced stronger Ub-smears of NICD compared to controls, indicating that Rnf8 promotes polyubiquitylation of NICD (Figure 3.14B).

Rnf8 ubiquitylates its DDR substrates through either lysine (K) K48 or K63-Ub-linkages (Panier and Durocher, 2013). Examination of the Ub-linkages for NICD polyubiquitylation in the presence of Rnf8 revealed the existence of both K48- and K63-linked polyubiquitin chains; though K63 linkages were more prominent (Figure 3.14B). Given the importance of K48- and K63-Ub linkages for the regulation of protein turnover (Hoeller et al., 2006) and that NICD accumulates in the absence of Rnf8, we performed cycloheximide chase experiments to examine the effect of Rnf8

142 on NICD half-life. Interestingly, half-life of NICD in Rnf8-/-;p53Δ/Δ mammary tumor cells became significantly shorter upon their complementation with Rnf8 (Figure 3.14C). We next examined whether NICD ubiquitylation in the presence of Rnf8 is dependent on its E3 ligase activity. Our data indicated that in contrast to Rnf8WT, the E3 ligase dead Rnf8 mutant (Rnf8C406S) was incompetent in mediating NICD polyubiquitylation (Figure 3.14D). Moreover, whereas Rnf8- mediated degradation of NICD was abolished in the presence of the E3 ligase dead mutant

Rnf8C406S, the FHA mutant Rnf8R42A effectively mediated this degradation (Figure 3.14E), further highlighting the importance of the E3 ligase function of Rnf8 for the control of NICD turnover.

Overall, these results suggest that Rnf8 promotes polyubiquitination of NICD to fine-tune Notch signaling in order to safeguard against mammary tumor development.

143

Figure 3.14. Rnf8 promotes polyubiquitylation and turnover of NICD, while loss of Rnf8 triggers hypersensitivity of mammary tumor cells to Notch inhibitors. (A) Immunoblot analysis of NICD and Rnf8-Flag, following anti-Flag IP from Rnf8-deficient mammary tumor cells and their Rnf8-restored isogenic cells. -Actin was used as the loading control. WCL: whole

144 cell lysate. (B) Intracellular NICD ubiquitylation assay. HEK293T cells, transfected with Flag- empty vector, Rnf8-Flag, Ub-HA, K48-Ub-HA, and K63-Ub-HA as indicated, were subjected to

IP with anti-NICD followed by anti-HA IB. WCL: whole cell lysate. (C) Cycloheximide (CHX) chase analysis of NICD protein half-lives in Rnf8-deficient mammary tumor cells and their Rnf8- restored isogenic controls. Cells were incubated in the presence of CHX for the indicated times.

Cell lysates were subjected to IB analysis with antibodies for NICD, Flag and -Actin. NICD levels were quantified by densitometry using ImageJ and normalized with -Actin level. NICD fold decrease is indicated. s.e.: short exposure; l.e.: long exposure. (D) HEK293T cells, transfected with empty-Flag vector, Rnf8WT-Flag, Rnf8C406S-Flag (RING domain mutant) and Ub-HA as indicated, were subjected to IP with anti-NICD followed by IB analysis for HA and NICD (Top).

IB analysis of NICD, Rnf8-Flag and -Actin on WCL (Bottom). (E) IB analysis of NICD and

Flag-Rnf8 expression levels in Rnf8-deficient mammary tumor cells reconstituted with empty-

Flag vector, Rnf8WT-Flag, Rnf8C406S-Flag, and Flag-Rnf8R42A (FHA domain mutant) as indicated.

-Actin was used as a loading control. (F) IB analysis of NICD levels in Rnf8-deficient mammary tumor cells 24h post-treatment with or without Notch inhibitor RO4929097. -Actin was used as a loading control. (G) Representative images of clonogenic survival of Rnf8-deficient mammary tumor cells as well as their Rnf8WT, Rnf8C406S, and Rnf8R42A reconstituted isogenic counterparts post Notch inhibitor treatments at the indicated doses. (H) Bar chart depicting clonogenic survival of Rnf8-deficient mammary tumor cells and their Rnf8WT, Rnf8C406S, and Rnf8R42A reconstituted isogenic controls either treated with Notch inhibitor alone or in combination with IR (2Gy) as indicated (mean ± s.e.m.). * P<0.05, ** P<0.01, *** P<0.001 (two-sided Student’s t-test).

145

3.3.12 Notch inhibitor treatment preferentially suppresses growth of Rnf8- deficient mammary tumor cells

Given the hyperactivated Notch signaling in Rnf8-deficient mammary tumors, we hypothesized that the suppression of their tumor growth can be achieved through pharmacological inhibition of

Notch signaling. Indeed, -secretase inhibitor RO4929097, a potent inhibitor of Notch signaling

(Munch et al., 2013), significantly diminished NICD levels in Rnf8-/-;p53Δ/Δ mammary tumor cells

(Figure 3.14F). Interestingly, Notch inhibitor treatments at various doses preferentially suppressed the tumor growth of Rnf8-/-;p53Δ/Δ mammary tumor cells compared with their Rnf8WT- or Rnf8R42A- reconstituted isogenic mammary tumor cells (Figure 3.14G, H). Furthermore, consistent with the inability of Rnf8C406S to promote polyubiquitination and turnover of NICD (Figure 3.14D, E), Rnf8-

/-;p53Δ/Δ mammary tumor cells reconstituted with Rnf8C406S remained hypersensitive to Notch inhibitor treatments similar to Rnf8-/-;p53Δ/Δ mammary tumor cells (Figure 3.14G, H). Moreover, in accordance with the role of Rnf8’s FHA domain in mediating its recruitment to DSB sites,

Rnf8R42A-reconstituted isogenic mammary tumor cells displayed significant hypersensitivity to IR treatments similar to Rnf8-/-;p53Δ/Δ mammary tumor cells, but exhibited levels of sensitivity to

Notch inhibitor treatments similar to Rnf8WT-reconstituted isogenic mammary tumor cells (Figure

3.14G, H). In addition, Rnf8-deficient mammary tumor cells exhibited an increased sensitivity to a combination of Notch inhibitor and IR treatments compared with single agent treatments (Figure

3.14H). Collectively, these results suggest that pharmacological inhibition of Notch signaling preferentially restrained the growth of Rnf8-deficient mammary tumors.

146

In brief, this study presents a novel mechanism for the regulation of Notch signaling through Rnf8 - mediated ubiquitylation and degradation of NICD. Given that our data also uncovered Rnf8 as a novel mammary tumor suppressor, we provide evidence that Rnf8 mediates its mammary tumor suppressor function through not only its role in DSB repair and maintenance of genomic stability, but also its activity in the fine-tuning of Notch signaling (Figure 3.15). Collectively, these findings suggest that cross-talk between DNA damage repair and Notch signaling pathways may play important roles in breast tissue homeostasis and malignant transformation, and potentially represent novel avenues for targeted therapies.

147

Figure 3.15. A proposed model of Rnf8-mediated mammary tumor suppressor activity.

Besides triggering genomic instability due to defective DNA repair, loss of Rnf8 also leads to hyperactivation of Notch signaling that promotes mammary luminal progenitors to aberrantly expand. Additional loss of p53 further exacerbates Rnf8-deficiency associated aberrant luminal expansion, and leads to the development of hyperplasia and ultimately breast cancer.

148

3.4 Materials and Methods

Mouse genotyping and genetic analysis. Rnf8-/- mice (AS0574 strain) were previously reported (Li et al., 2010a). p53fl/fl conditional mutant mice were intercrossed with WapCre transgenic mice (Jackson Laboratory) to obtain WapCre;p53fl/fl mice. Subsequent crosses of Rnf8-

/- mice with WapCre;p53fl/fl mice generated Rnf8-/-;WapCre;p53fl/fl females. Because we observed no phenotypic difference among mammary glands from WT, WapCre, p53fl/fl, and p53fl/wt females, they were collectively named as WT and were used as experimental controls. All mice in this study were in a mixed 129/J × C57BL/6 genetic background, maintained in a specific pathogen-free environment, genotyped by PCR (Primers available upon request) and all procedures were performed in compliance with the Ontario Cancer Institute animal care committee guidelines, and animal protocols were approved by the Animal Resource Center of Ontario Cancer Institute. To rule out the effect of changes in progesterone levels on mammary epithelial cells (Joshi et al.,

2010), all experimental female mice used in this study were followed by vaginal smear cytology to determine their phases of estrus cycle for at least 4 continuous days, and were subsequently employed for experiments at estrus phase, characterized by a basal level of progesterone and therefore a minimum effect of progesterone-induced expansion of mammary epithelial subpopulations. Survival cohorts of female mice were monitored for mammary tumor onset by palpation at least twice a week for 618 days. Log rank test was employed for statistical analysis of survival curves.

149

Vaginal smear cytology. Vaginal smears were taken from all experimental female mice using

l PBS followed by loading onto hydrophobic marker marked grid of glass slides. After the smears were air-dried, the slides were fixed with ethanol, rinsed with distilled water, stained with hematoxylin and rinsed with distilled water, followed by mounting with coverslips and examination under a light microscope. The mouse estrus cycle typically encompasses about 5 days.

The period of estrus, normally spanning about one day, is represented by the presence of large anucleated cornified squamous-type epithelial cells devoid of leukocytes inside vaginal smears as described previously(Joshi et al., 2010).

Flow cytometry analysis and sorting of mammary epithelial cells. Mammary glands from female mice at estrus phase were processed into mouse Epicult-B medium supplemented with 3% fetal bovine serum (FBS), 750U/ml collagenase and 250U/ml hyaluronidase, and incubated at

37°C for 2.5 hrs. Subsequently, mammary organoids were subjected to red blood cell lysis in

NH4Cl, followed by further dissociation in 0.25% trypsin for 2 min, and then in a mixture of 5 mg/ml dispase and 0.1 mg/ml DNase I for 2 min. After filtering through a 40 m mesh, mammary single cell suspensions were obtained. All the reagents used above were from Stem Cell

Technology, and all antibodies used below were from BD Pharmingen unless otherwise indicated.

Subsequently, mammary cells were blocked with Fc receptor antibody, followed by incubation with biotinylated anti-CD31, anti-CD45, and anti-TER119. CD45+ immune cells, TER119+ red blood cells, and CD31+ endothelial cells were further conjugated with streptavidin-PE-Cy7, and were excluded using flow cytometry. Propidium iodide (Sigma) was employed to stain dead cells, which were excluded from analysis. Mammary epithelial cell subpopulations were identified using anti-CD49f-FITC (clone GoH3), anti-CD24-PE (clone M1/69), anti-CD61-APC (Invitrogen)

150 using a FACSCalibur (BD), and FlowJo software (Tree Star, Inc). Cell sorting of mammary epithelial subpopulations was performed on a FACSAria (BD), and the purity of sorted populations was routinely greater than 97%.

Whole mount staining. Mammary glands from estrus-stage matched females were mounted on glass slides and fixed with Carnoy’s formula 2 fixative (10% glacial acetic acid, 30% chloroform, and 60% absolute ethanol) for 12 hrs. Fixed mammary glands were processed through a hydration series using descending concentrations of ethanol (70%, 50%, and 20%; 15 min each) and distilled water for 10 min, followed by staining in a carmine-alum mix (0.2% carmine (Sigma) and 0.5% aluminum potassium sulphate (Sigma)) for 16h. Subsequently, the stained mammary glands were dehydrated through ascending concentrations of ethanol (25%, 50%, 70%, and 100%; 15 min each), and then the stained mammary glands were mounted with Permount medium (Fisher

Scientific).

Histological analysis and immunohistochemistry. Paraffin sections of mammary glands and mammary tumors were stained with H&E for histological analysis as previously described (Li et al., 2010a). For immunohistochemistry, tissues or tumors were fixed with 4% paraformaldehyde, followed by paraffin-embedding. Then, tissue sections were de-paraffinized in xylene, followed by rehydration through descending concentrations of ethanol. The tissue sections were then treated in Borg Decloaker antigen retrieval solution (pH 9) for 30 min at 121°C and 10 seconds at 90°C using a Decloaking chamber (Biocare Medical), followed by staining according to HRP-AEC tissue staining kit’s instructions (R&D Systems). Antibodies employed were anti-CK18

151

(Fitzgerald), anti-CK14 (Covance), anti-Ki67 (Novus Biological, Inc) and anti-NICD (Val1744;

Cell Signaling).

Realtime quantitative RT-PCR. Total RNA was extracted using Trizol (Invitrogen) from FACS- sorted primary mammary cell subpopulations, and the quality and concentration of RNA was determined using a NanoDrop 2000 Spectrometer (260/280 ratio; Thermo Scientific). The purified

RNA’s were treated with RNAse-free DNA digestion kit (Invitrogen), and were reverse transcribed into cDNA using the Oligo(dT) priming method of the first-strand cDNA synthesis system kit (Invitrogen). Subsequently, the cDNA was added to a quantitative real time PCR mixture of 1x SYBR Green PCR master mix (Applied Biosystems) and 70nM gene-specific primers. Primer sequences (Hes1, forward 5’-AAAGCCTATCATGGAGAAGAGGCG-3’, reverse 5’-GGAATGCCGGGAGCTATCTTTCTT-3’; Hes5, forward 5’-AAAGCCTATCA

TGGAGAAGAGGCG-3’, reverse 5’-GGAATGCCGGGAGCTATCTTTCTT-3’; Cyclin D1, forward 5'-GCGTACCCTGACACCAATCTC-3', reverse 5'-ACTTGAAGTAAGATACGGA

GGGC-3'; Cyclin D2, forward 5'-GAGTGGGAACTGGTAGTG TTG-3', reverse 5'-

CGCACAGAGCGATGAAGGT-3'; Hey1, forward 5’-ACACTGCAGGAGGGAAAGGTT-3’, reverse 5’-CAAACTCCGATAGTCCATAGCCA-3’; Hey2, forward 5’-AAGCGCCCTTGTGA

GGAAAC-3’, reverse 5’-GGTAGTTGTCGGTGAATTGGAC-3’; Gata3, forward 5'-AGCC

ACATCTCTCCCTTCAG-3', reverse 5'-AGGGCTCTGCCTCTCTAACC-3'; Slug, forward 5'-

GATGTGCCCTCAGGTTTGAT-3', reverse 5'-ACACA TTGCCTTGTGTCTGC-3'; p18, forward 5’-TTATGAAGCACACAGCCTGCAATGT-3’, reverse 5’-ACGGACAGCCA

ACCAACTAACGG-3’; FoxM1, forward 5’-GAGGAAAGAGCACCTTCAGC-3’, reverse 5’-

AGGCAATGTCTCCTTGATGG-3’; Notch1, forward 5' – CAATG TTCGAGGACCAGATGG-

152

3', reverse 5'-ACTGCAGGAGGCAATCATGAG-3'; Elf5, forward 5’-GATCTGTTCAGCAA

TGAAG-3’, reverse 5’-GGTCTCTTCAGCATCATTG-3’; Amphiregulin, forward 5’- ACTCAC

AGCGAGGATGACAAGG-3’, reverse 5’-TAACGATGCCGATGCCAATAG-3’; -Actin, forward 5'-CCATACCCAAGAAGGAAGGCT-3', reverse 5'-TATTGGCAACGAGCGGTTC-

3'). Assays were performed in triplicates using ABI PRISM 7900HT Sequence Detection System

(Applied Biosystems), followed by data analysis using Sequence Detection System software

Version 2.1 (Applied Biosystems). The expression level of each target gene was normalized to endogenous -Actin transcripts.

Immunofluorescence. Immunofluorescence staining for subnuclear foci formation was performed as previously described (Li et al., 2010a). Briefly, for immunofluorescence of H2ax on purified mammary epithelial cells, mammary single cells from estrus stage matched females were prepared as described above, and blocked with Fc antibody and stained with biotinylated anti-CD31, anti-CD45, and anti-TER119, followed by negative depletion of endothelial cells, immune cells, and red blood cells using streptavidin magnetic beads (Dynabeads; Invitrogen). The purity of the isolated mammary epithelial cells was determined to be greater than 97% via staining of cells before and after purification with streptavidin-PE-Cy7, followed by FACS analysis.

Primary antibodies used for immunofluorescence staining were anti-H2ax (Ser139; Millipore), anti-53bp1 (Bethyl), anti-Rad51 (Santa Cruz), and anti-Brca1 (home-made) antibodies. Anti-rabbit or anti-mouse IgG conjugated with Alexa Fluor 488 or 568 (Invitrogen) were used for secondary staining. Subsequently, slides were counterstained with DAPI (Invitrogen) and mounted with

Mowiol (Sigma-Aldrich). Images were taken using a fluorescence microscope (Leica) under 100X magnifications. ImageJ software (National Institutes of Health) was used to process the raw data.

153

3-Dimensional Matrigel mammary colony forming. Mammary single cell suspensions were prepared as described above from inguinal mammary glands of 10-12 weeks old females at estrus phase. 5,000 total mammary cells were seeded into 50l of growth factor reduced Matrigel (BD) and cultured as a droplet on the center of 6 cm plates containing mouse Epicult-B medium supplemented with 5% FBS, cytokines EGF and FGF, and Epicult-B proliferation supplements

(Stem Cell Technology ). Colonies were scored after 12 days of culture under a light microscope, and tiled image composites spanning Matrigel droplets were then taken.

Protein analysis and antibodies. IB analyses were performed using antibodies against NICD

(Val1744; Cell Signaling), Hes1 (H140; Santa Cruz), Cyclin D1 (Santa Cruz), phospho-Akt (Cell

Signaling), Akt (Cell Signaling), Rb (Santa Cruz), p53 (FL393; Santa Cruz Biotechnology, Inc.), p21 (M19; Cell Signaling), Bax (Cell Signaling), Flag (M2; Santa Cruz Biotechnology, Inc.) and

β-actin (C4; Santa Cruz Biotechnology, Inc).

Mammary tumor cell lines, constructs and retrovirus infection. Mammary tumors were processed into single cell suspensions following the same procedure for mammary gland single cell suspension described above. Then, mammary tumor cells were analyzed by FACS for Lin- expression of CD49f, CD24 and CD61, and were cultured in vitro into stable cell lines in

Dulbecco’s modified Eagle’s medium supplemented with 10% FBS, 5x10-5M 2-ME, 100 U/ml penicillin, and 100 µg/mL streptomycin (complete DMEM). To complement Rnf8-/-;p53Δ/Δ mammary tumor cell lines with WT and mutant forms of Rnf8, cDNA’s for mouse Rnf8 (WT,

154

C406S and R42A) were cloned into pMSCV-Flag for retroviral expression. Rnf8 retroviral constructs and pMSCV-Flag were transfected into Phoenix cells, followed by collection of virus supernatants 48h post-transfection. Mammary tumor cell lines were transduced with the harvested retroviral supernatants in the presence of 8 μg/mL polybrene, and transduced cells were selected for their resistance to puromycin (5μg/mL) for 3 weeks.

Clonogenic survival assay. Exponentially growing mammary tumor cells were seeded at 500 cells/6cm dishes in complete DMEM. For PARP inhibitor treatment, tumor cells in complete

DMEM medium were either left untreated, or cultured in the presence of 0.1-5M of PARP inhibitor (KU0058948). For ionizing radiation treatment, mammary tumor cells were either left untreated or subjected to radiation (2-16 Gy). For Notch inhibitor treatment, mammary tumor cells were either left untreated, or cultured in the presence of 1-50M of -secretase inhibitor

RO4929097. Cells were allowed to grow in complete DMEM medium at 37°C under 5% CO2 for

12 days before fixation with methanol and staining with crystal violet. Numbers of colonies in each dish were counted, and pictures of the dishes were taken. All experiments were performed in triplicates. Data from all treated samples were normalized to their respective untreated samples.

Transfections. To determine HR and NHEJ efficiency, plasmids containing HR or NHEJ reporter cassettes(Mao et al., 2009) were linearized by I-SceI restriction enzyme, followed by purification using Qiagen Qiaex II purification kit (Qiagen). Mammary tumor cell lines were split two days prior to transfection in order to reach exponential growing phase and 1x106 cells were transfected with 2g of linearized HR reporter construct or 0.5g of the NHEJ reporter construct, together

155 with 0.1g of RFP construct as the internal control, according to GenJet In Vitro DNA Transfection kit (SignaGen Laboratories). Transfected cells were seeded on 6 well plates and incubated at 37°C under 5% CO2 for 48h, followed by flow cytometry analysis of expression of eGFP and RFP.

Repair efficiencies of HR and NHEJ were expressed as eGFP+/RFP+ ratios, and thereby were independent of the transfection efficiency. Furthermore, human breast cancer cells MDA-MB-231 were transfected with Rnf8-Flag construct or Flag-empty vector on 6cm dishes, according to

PloyJet In Vitro DNA Transfection kit (SignaGen Laboratories).

Southern blot analysis. DNA’s from Rnf8-/-;WapCre;p53fl/wt mammary tumors and their respective tails of the same female mice were purified by phenol-chloroform extraction. 15g of genomic DNA’s were digested with BglII, followed by blotting and hybridization procedures as described previously (Li et al., 2010a).

Cumulative growth analysis of tumor cells. Passage 5 Rnf8-deficient mammary tumor cells and their respective Rnf8-reconstituted isogenic cells were seeded at a density of 3x105 cells/6 cm dish in triplicates, and cultured in complete DMEM at 37°C under 5% CO2. After 72h, cells were trypsinized, counted and re-seeded at the same density of 3x105 cells/6 cm dish in triplicates, and this procedure was repeated for 9 days. At day 9, cells were trypsinized, counted and re-seeded at the same density of 3x105 cells/6 cm dish. 48h post-seeding, images of cells were taken under an inverted light microscope (Leica). Subsequently, the dishes were washed with PBS, fixed with methanol, stained with crystal violet and washed with distilled water, after which pictures of the dishes were taken.

156

Notch/CBF1 dual luciferase reporter assay. Passage 5 Rnf8-deficient mammary tumor cells and their respective Rnf8-reconstituted isogenic cells were co-transfected with pCBF1-firefly luciferase reporter and pRL-TK-renilla luciferase reporter (a kind gift from Dr. R. Khokha) plus

0-300ng of pNICD plasmid(McKenzie et al., 2005) using PolyJet DNA In Vitro Tranfection

Reagent (SignaGen Laboratories). 48h later, transfected cells were counted and then plated at

2.5x105 cells/75ul DMEM/well of flat-bottomed opaque 96 well tissue culture plates (Greiner

CellStar). Luciferase activities of the transfected cells were measured using the Dual Glo

Luciferase assay kit (Promega) with Firefly luciferase-based reporter gene activity normalized to the Renilla luciferase internal control reporter activity. Background auto-luminescence from non- transfected cells was subtracted from the transfected cells.

Nested step-down PCR analysis of WapCre-mediated p53 deletion. DNA’s were extracted from cell sorted Lin- mammary epithelial subpopulations (CD49flow CD24+ CD61- luminal differentiated cells; CD49flow CD24+ CD61+ luminal progenitor cells; CD49fhigh CD24+ basal cells) of WapCre;p53fl/fl females at estrus phase that have undergone zero (16 week-old), two (4 month- old) and four pregnancies (6.5 month-old) and involution. Nested step-down PCR were performed on equal amount of genomic DNA’s to detect the presence of p53WT, p53fl and p53Δ2-10 alleles as previously described (Jonkers et al., 2001). Cre PCR was used a control.

Cycloheximide chase analysis. To examine the effect of Rnf8 on NICD turnover, Rnf8-/-;p53Δ/Δ mammary tumor cells reconstituted with empty vector (mock) or Rnf8 were seeded onto 6 cm

157 dishes overnight. Subsequently, cells were either left untreated or treated with 150M of cycloheximide for 0.5-6h, and levels of NICD, Rnf8-flag and -Actin were then examined by IB analysis.

Intracellular ubiquitylation assay. HEK293T cells were seeded on 6 cm dishes overnight, and then transfected with Rnf8-Flag or Rnf8C406S-Flag, and WT or mutated Ub-HA’s [Addgene: pRK5-

HA-Ub-WT (ID 17608), pRK5-HA-Ub-K48 (ID 17605), pRK5-HA-Ub-K63 (ID 17606)] as indicated. 48h later, cells were washed with PBS and lysed on ice. Cell lysates were precleared with Protein A Sepharose beads (Life Technologies) and subjected to IP using anti-NICD

(Val1744; Cell Signaling). Subsequently, the beads were washed thrice with 1ml of RIPA buffer.

The proteins were released from the beads by boiling in 2x SDS-PAGE sample buffer, followed by IB analysis against the indicated antibodies.

158

Chapter 4

Summary and future directions

159

4.1 Summary and future directions regarding Rnf8 in immunodeficiency

Recent studies reported that the E3 ligase Rnf8 plays important roles in propagating DNA DSB signals and thereby facilitating the efficient recruitment of various DNA damage response proteins, such as 53BP1 and BRCA1, to sites of DNA damage (Huen et al., 2007; Kolas et al., 2007a;

Mailand et al., 2007; Wang and Elledge, 2007). Using Rnf8 knockout mouse models, we demonstrated that Rnf8 deficiency significantly reduces the efficiencies of CSR (Li et al., 2010a).

Interestingly, basal serum levels of secreted immunoglobulins or antibodies including IgG1, IgG3 and IgG2b isotypes are significantly diminished in Rnf8-/- mice compared with WT littermates.

Furthermore, additional deletion of p53 does not rescue nor exacerbate CSR defects associated with Rnf8-deficiency. Thus, Rnf8 deficient mice are immunodeficient. Interestingly, quantitative analysis of S-S1 recombined genomic DNA’s, corresponding to class switched IgG1 isotype, in

WT and Rnf8-/- B cells stimulated with anti-CD40 and IL4 indicates that Rnf8-deficiency associated CSR defect occurs at the genomic level, but is not due to disruptions of immunoglobulin exportation to cell surface and secretion, and thus is intrinsic to Rnf8-/- B cells. Additionally, we reveal that loss of a single allele of Rnf8 is sufficient to cause immunodeficiency. These data unravel an important role of Rnf8 in mediating the IgH locus of B cells to class switch into expressing the proper effector function upon pathogen challenges, and thereby demonstrating

Rnf8’s key role in specifying the proper effector function of a humoral immune response.

Interestingly, by integrating knowledge obtained from this thesis and our current knowledge of knockout mice of other related DDR molecules regarding their CSR defects, there exists an unexplained difference that implies the existence of yet to be elucidated DSB signaling pathways

160 for re-joining AID-induced DSBs, and thereby are important for preventing immunodeficiency. It is interesting to note that the capabilities of Rnf8-/-, Atm-/-, H2ax-/-, Rnf168-/-, and Mdc1-/- B cells to recombine AID-generated DSBs (when class switching from IgM expression to IgG1 expression; an endogenous readout of efficiency of NHEJ repair machinery) are all consistently reduced to approximately a third of the capability of WT B cells (Bohgaki et al., 2011; Celeste et al., 2002; Li et al., 2010a; Lou et al., 2006; Lumsden et al., 2004; Santos et al., 2010). In striking contrast, the

CSR efficiency of 53bp1-/- B cells is reduced to about only 1% of WT B cells (Manis et al., 2004;

Ward et al., 2004). Furthermore, similar to 53bp1-/- B cells, as expected, the CSR efficiency of Aid-

/- B cells is reduced to about 1% of WT B cells, indicating that the Aid-initiated DSB signals converge on 53bp1 to mediate the subsequent recombination of the broken chromatins (Muramatsu et al., 2000; Pefanis et al., 2014; Vuong et al., 2009). Based on the current proposed model of the hierarchical recruitments of the above mentioned molecules in DSB signaling in the literature,

53bp1 is recruited to sites of DSBs downstream of Rnf8-initiated ubiquitination events (Huen et al., 2007; Kolas et al., 2007b; Mailand et al., 2007). Collectively, the differences in the severity of the immunodeficiency phenotypes associated with deficiencies of Rnf8, 53bp1, Atm, H2ax, Mdc1,

Rnf168 and Aid in vivo obviously contradict with their respective rankings of their sequential recruitments to repair DSBs in our current model of DSB signaling derived mainly from in vitro cell line data. This discrepancy suggests that for CSR-associated DSBs, there exists alternative mechanisms for 53bp1 recruitment and function that are independent of Rnf8, Atm, H2ax, Mdc1, and Rnf168 and that would account for the difference between the more severe CSR defect in

53bp1-/- mice and the less severe CSR defect in Rnf8-/-/Atm-/-/H2ax-/-/Mdc1-/-/Rnf168-/- mice.

161

In accordance with this future direction, data from this thesis indicate that 53bp1 can be recruited to irradiated activated B cells in the absence of Rnf8 to a significant extent, though much reduced compared with WT B cells (Li et al., 2010a). Interestingly, recent studies suggest that the E3 ligase

BMI1 catalyzes monoubiquitination of H2A types histones in response to DNA damage, and has been suggested to be needed for the efficient recruitment of 53bp1 and Brca1 independent of Rnf8

(Ismail et al., 2010). Furthermore, dual deletions of BMI1 and Rnf8 result in an additive increase in radiation sensitivity, implying their non-epistatic relationships in response to DNA damage

(Ismail et al., 2010). Also, another study indicates that BMI1 co-purifies with components of

NHEJ, such as DNA-PK, and components of DSB signaling, including ATM and H2AX, in brain tumor cells following irradiation (Facchino et al., 2010). The above discoveries raise the possibility that BMI1 could potentially account for this Rnf8-independent mechanism responsible for the remaining CSR efficiency (about a third of WT B cells’ capability) in Rnf8-/- B cells. Hence, it will be interesting for future research work to elucidate whether BMI1 can account fully for that missing puzzle of the remaining CSR efficiency (i.e. NHEJ-repair efficiency) in the absence of

Rnf8. Since BMI1 knockout mice rarely survive to adulthood (Vanderlugt et al., 1994), the generation of CD19-Cre;BMI1floxp/floxp;Rnf8-/- conditional double mutant mice with BMI1 specifically deleted in B cells will be useful to examine this in vivo.

Additionally, besides CSR, mammals have also evolved other mechanisms to diversify their immune repertoires, in order to combat the numerous types of pathogens encountered during their lifetimes. Interestingly, DNA DSBs, though they can become a lethal threat to cellular survival, are programmed to occur during another two important mechanisms of antibody diversification, namely VDJ recombination and IgH gene conversion. Given that 1) the efficient recruitment of

162

53BP1 to DSB sites is largely dependent on RNF8, and 2) 53BP1 promotes long-range joining of

DNA breaks during V(D)J recombination (Difilippantonio et al., 2008; Li et al., 2010a), future research work will be needed to address whether RNF8 participates in joining of RAG1/2-induced

DNA DSBs during V(D)J recombination.

4.2 Summary and future directions regarding Rnf8 in breast cancer

Besides processing of DNA DSBs generated through normal cellular programs, the concerted actions of all DDR molecules create a response network to DNA damage that protects the fidelity of mammalian genome and maintains the genomic stability for mammalian cells. Interestingly, whereas breast cancers all exhibit the hallmark of genomic instability, a well-accepted oncogenic driving mechanism, mutation frequencies of many DDR molecules except BRCA1, BRCA2 and p53, in breast tumors are surprisingly not as frequent as we anticipated. Also, the observed breast cancer phenotypes associated with mutations of various DDR molecules do not directly reflect their rankings in the hierarchy of our current model of the DDR network, since mutations of downstream DDR molecules are known to result in more severe breast cancer phenotypes than mutations of their upstream DDR molecules. For instance, although ATM functions upstream of

BRCA1 in our proposed model of DDR network, BRCA1 mutations undoubtedly confer higher breast cancer risks than ATM mutations both in mouse and human. Thus, genomic instability, induced by defective DNA repair associated with mutated DDR molecules, likely only account for

163 a part of all the driving forces necessary to malignantly transform breast tissues harboring mutated

DDR molecules, and thereby the relative rankings of these DDR molecules in our proposed model of DDR network do not exactly reflect their relative importance in preventing breast cancer development. This implies that other physiological activities of these DDR molecules are likely also perturbed to drive breast cancer development and contribute to breast cancer heterogeneity and the complexity of this devastating disease.

Interestingly, our data unraveled that besides Rnf8-mediated maintenance of genomic stability,

Rnf8 also fine-tunes Notch signaling to safeguard against breast cancer development. Notably, we demonstrate that loss of Rnf8 in mammary glands triggers an aberrant luminal expansion via hyperproliferation. The guardian of the genome p53 actively restrains Rnf8-deficiency associated aberrant luminal expansion, and significantly suppresses Rnf8-mutation associated breast cancer development. Mechanistically, Rnf8 suppresses breast cancer, not only through its role in DNA repair, but also via its novel function in polyubiquitinating NICD to promote its turnover and the negative regulation of Notch signaling. Interestingly, genetic restoration of Rnf8 into poorly differentiated Rnf8-deficient mammary tumor cells appreciably retarded tumor growth through restoration of their DNA repair capability as well as suppression of Notch signaling, and thereby directly demonstrating the potency of Rnf8-mediated mammary tumor suppressive activity as well as the persistent addiction of Rnf8-deficent mammary tumor cells for the absence of Rnf8 in spite of the presence of other secondary oncogenic mutations.

164

Furthermore, based on current publicly available human breast tumor sequencing data, RNF8 mutations in breast tumors are infrequent, just like many other DDR molecules except BRCA1,

BRCA2 and p53. For instance, 53BP1 mutations are also infrequent in the current public available human breast tumor sequencing data, but the protein expression of 53BP1 has been demonstrated to be significantly reduced in specific subsets of sporadic triple-negative and BRCA mutation- associated human breast tumors (Bouwman et al., 2010b). Another recent study suggests that the reduction of 53BP1 protein levels in BRCA1 deficient human breast cancer cells is achieved through degradation of 53BP1 protein by cathepsin L, which is activated by the loss of BRCA1

(Grotsky et al., 2013). Therefore, given the potency of Rnf8’s mammary tumor suppressive activity in our mouse models, it will be interesting to further determine the potential epigenetic regulation of RNF8 expression and the status of RNF8 protein expression in a large cohort of human breast tumors, and examine whether loss of Rnf8 activity is specific to any subtypes of breast tumors.

Interestingly, a recent study reported hypermethylation of RNF8 promoter in human triple - negative breast tumors similar to BRCA1 hypermethylation (Watanabe et al., 2013). Moreover, to add another layer of potential complexity, it will be interesting as a future direction to identify negative or positive regulators of RNF8 and determine if they are overexpressed or silenced in human breast tumors. Also, future research work will be needed to determine whether the functional partners of RNF8 such as the E2 conjugating enzyme UBC13 have reduced expression and/or loss of enzymatic activity in human breast tumors. Collectively, future research effort will be needed to address the above scientific questions, and will provide us with a better understanding of mechanisms of alterations of RNF8 in human breast cancers and thus will potentially identify a specific subtype of breast cancers that associates with RNF8 inactivation.

165

Moreover, data from this thesis indicate that pharmacological inhibition of Notch signaling by gamma-secretase inhibitor treatments preferentially restrains the growth of Rnf8-deficient mammary tumors, suggesting that inhibition of Notch signaling may represent a novel avenue for treating Rnf8-deficient breast cancer. Hence, as a future direction, it will be interesting to examine human breast tumor samples for a potential correlation between high levels of NOTCH signaling and low levels of RNF8 protein expressions, or vice versa. Furthermore, since our data also reveal that Rnf8-deficient mammary tumors reconstituted with Rnf8-E3 ligase dead mutant also display hypersensitivity to NOTCH inhibitors at a similar level to their Rnf8-deficient parental mammary tumors. A more complicated but rational future direction will be to examine human breast tumors for a potential correlation between levels of NOTCH signaling and levels of E3 ligase activity of

RNF8. Also, it will be worth to examine for a potential correlation between high levels of NOTCH signaling and low levels of the E2 conjugating enzyme UBC13 (which functions in complex with

RNF8) and/or abrogation of its enzymatic activity in human breast tumors. Given that NOTCH inhibitors are currently being actively tested in a number of clinical trials for potential therapeutic efficacies against advanced breast cancers (Andersson and Lendahl, 2014), it will be rational to investigate whether patients harboring advanced RNF8-deficient breast tumors will respond better to NOTCH inhibitor treatments than patients harboring advanced RNF8-proficient breast tumors.

Overall, the above future directions will potentially contribute to the advent of novel personalized breast cancer therapeutic strategies.

166

References

Acs, K., M.S. Luijsterburg, L. Ackermann, F.A. Salomons, T. Hoppe, and N.P. Dantuma. 2011. The AAA-ATPase VCP/p97 promotes 53BP1 recruitment by removing L3MBTL1 from DNA double-strand breaks. Nature Structural & Molecular Biology 18:1345-U1355.

Ades, F., D. Zardavas, I. Bozovic-Spasojevic, L. Pugliano, D. Fumagalli, E. de Azambuja, G. Viale, C. Sotiriou, and M. Piccart. 2014. Luminal B breast cancer: molecular characterization, clinical management, and future perspectives. J Clin Oncol 32:2794- 2803.

Alexandrov, L.B., S. Nik-Zainal, D.C. Wedge, S.A. Aparicio, S. Behjati, A.V. Biankin, G.R. Bignell, N. Bolli, A. Borg, A.L. Borresen-Dale, S. Boyault, B. Burkhardt, A.P. Butler, C. Caldas, H.R. Davies, C. Desmedt, R. Eils, J.E. Eyfjord, J.A. Foekens, M. Greaves, F. Hosoda, B. Hutter, T. Ilicic, S. Imbeaud, M. Imielinski, N. Jager, D.T. Jones, D. Jones, S. Knappskog, M. Kool, S.R. Lakhani, C. Lopez-Otin, S. Martin, N.C. Munshi, H. Nakamura, P.A. Northcott, M. Pajic, E. Papaemmanuil, A. Paradiso, J.V. Pearson, X.S. Puente, K. Raine, M. Ramakrishna, A.L. Richardson, J. Richter, P. Rosenstiel, M. Schlesner, T.N. Schumacher, P.N. Span, J.W. Teague, Y. Totoki, A.N. Tutt, R. Valdes- Mas, M.M. van Buuren, L. van 't Veer, A. Vincent-Salomon, N. Waddell, L.R. Yates, J. Zucman-Rossi, P.A. Futreal, U. McDermott, P. Lichter, M. Meyerson, S.M. Grimmond, R. Siebert, E. Campo, T. Shibata, S.M. Pfister, P.J. Campbell, and M.R. Stratton. 2013. Signatures of mutational processes in human cancer. Nature 500:415-421.

Alt, F.W., Y. Zhang, F.L. Meng, C. Guo, and B. Schwer. 2013. Mechanisms of programmed DNA lesions and genomic instability in the immune system. Cell 152:417-429.

Althuis, M.D., J.H. Fergenbaum, M. Garcia-Closas, L.A. Brinton, M.P. Madigan, and M.E. Sherman. 2004. Etiology of hormone receptor-defined breast cancer: a systematic review of the literature. Cancer Epidemiol Biomarkers Prev 13:1558-1568.

Andersson, E.R., and U. Lendahl. 2014. Therapeutic modulation of Notch signalling--are we there yet? Nat Rev Drug Discov 13:357-378.

Antoniou, A.C., and D.F. Easton. 2006. Models of genetic susceptibility to breast cancer. Oncogene 25:5898-5905.

Apostolou, P., and F. Fostira. 2013. Hereditary breast cancer: the era of new susceptibility genes. Biomed Res Int 2013:747318.

Apter, D., M. Reinila, and R. Vihko. 1989. Some Endocrine Characteristics of Early Menarche, a Risk Factor for Breast-Cancer, Are Preserved into Adulthood. International Journal of Cancer 44:783-787.

167

Asselin-Labat, M.L., F. Vaillant, J.M. Sheridan, B. Pal, D. Wu, E.R. Simpson, H. Yasuda, G.K. Smyth, T.J. Martin, G.J. Lindeman, and J.E. Visvader. 2010. Control of mammary stem cell function by steroid hormone signalling. Nature 465:798-802.

Barcellos-Hoff, M.H., C. Park, and E.G. Wright. 2005. Radiation and the microenvironment - tumorigenesis and therapy. Nature Reviews Cancer 5:867-875.

Barlow, C., S. Hirotsune, R. Paylor, M. Liyanage, M. Eckhaus, F. Collins, Y. Shiloh, J.N. Crawley, T. Ried, D. Tagle, and A. Wynshaw-Boris. 1996. Atm-deficient mice: a paradigm of ataxia telangiectasia. Cell 86:159-171.

Bartek, J., and J. Lukas. 2007. DNA damage checkpoints: from initiation to recovery or adaptation. Curr Opin Cell Biol 19:238-245.

Bartocci, C., and E.L. Denchi. 2013. Put a RING on it: regulation and inhibition of RNF8 and RNF168 RING finger E3 ligases at DNA damage sites. Front Genet 4:128.

Bekker-Jensen, S., and N. Mailand. 2011. The ubiquitin- and SUMO-dependent signaling response to DNA double-strand breaks. FEBS Lett 585:2914-2919.

Bell, D.W., J.M. Varley, T.E. Szydlo, D.H. Kang, D.C. Wahrer, K.E. Shannon, M. Lubratovich, S.J. Verselis, K.J. Isselbacher, J.F. Fraumeni, J.M. Birch, F.P. Li, J.E. Garber, and D.A. Haber. 1999. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science 286:2528-2531.

Bernstein, L. 2002. Epidemiology of endocrine-related risk factors for breast cancer. J Mammary Gland Biol Neoplasia 7:3-15.

Berquist, B.R., and D.M. Wilson. 2012. Pathways for repairing and tolerating the spectrum of oxidative DNA lesions. Cancer Lett 327:61-72.

Bertram, C., and R. Hass. 2008. Cellular responses to reactive oxygen species-induced DNA damage and aging. Biol Chem 389:211-220.

Bieging, K.T., S.S. Mello, and L.D. Attardi. 2014a. Unravelling mechanisms of p53-mediated tumour suppression. Nature Reviews Cancer 14:359-370.

Bieging, K.T., S.S. Mello, and L.D. Attardi. 2014b. Unravelling mechanisms of p53-mediated tumour suppression. Nature Reviews Cancer 14:359-370.

Blaumueller, C.M., H. Qi, P. Zagouras, and S. Artavanis-Tsakonas. 1997. Intracellular cleavage of Notch leads to a heterodimeric receptor on the plasma membrane. Cell 90:281-291.

Boboila, C., F.W. Alt, and B. Schwer. 2012. Classical and alternative end-joining pathways for repair of lymphocyte-specific and general DNA double-strand breaks. Adv Immunol 116:1-49.

Bocker, W. 2002. [WHO classification of breast tumors and tumors of the female genital organs: pathology and genetics]. Verh Dtsch Ges Pathol 86:116-119.

168

Bohgaki, T., M. Bohgaki, R. Cardoso, S. Panier, D. Zeegers, L. Li, G.S. Stewart, O. Sanchez, M.P. Hande, D. Durocher, A. Hakem, and R. Hakem. 2011. Genomic instability, defective spermatogenesis, immunodeficiency, and cancer in a mouse model of the RIDDLE syndrome. PLoS Genet 7:e1001381.

Bolos, V., E. Mira, B. Martinez-Poveda, G. Luxan, M. Canamero, A.C. Martinez, S. Manes, and J.L. de la Pompa. 2013. Notch activation stimulates migration of breast cancer cells and promotes tumor growth. Breast Cancer Res 15:R54.

Bombonati, A., and D.C. Sgroi. 2011. The molecular pathology of breast cancer progression. J Pathol 223:307-317.

Bothmer, A., D.F. Robbiani, N. Feldhahn, A. Gazumyan, A. Nussenzweig, and M.C. Nussenzweig. 2010. 53BP1 regulates DNA resection and the choice between classical and alternative end joining during class switch recombination. Journal of Experimental Medicine 207:855-865.

Botuyan, M.V., J. Lee, I.M. Ward, J.E. Kim, J.R. Thompson, J. Chen, and G. Mer. 2006. Structural basis for the methylation state-specific recognition of histone H4-K20 by 53BP1 and Crb2 in DNA repair. Cell 127:1361-1373.

Bouras, T., B. Pal, F. Vaillant, G. Harburg, M.L. Asselin-Labat, S.R. Oakes, G.J. Lindeman, and J.E. Visvader. 2008. Notch signaling regulates mammary stem cell function and luminal cell-fate commitment. Cell Stem Cell 3:429-441.

Bouwman, P., A. Aly, J.M. Escandell, M. Pieterse, J. Bartkova, H. van der Gulden, S. Hiddingh, M. Thanasoula, A. Kulkarni, Q. Yang, B.G. Haffty, J. Tommiska, C. Blomqvist, R. Drapkin, D.J. Adams, H. Nevanlinna, J. Bartek, M. Tarsounas, S. Ganesan, and J. Jonkers. 2010a. 53BP1 loss rescues BRCA1 deficiency and is associated with triple- negative and BRCA-mutated breast cancers. Nat Struct Mol Biol 17:688-695.

Bouwman, P., A. Aly, J.M. Escandell, M. Pieterse, J. Bartkova, H. van der Gulden, S. Hiddingh, M. Thanasoula, A. Kulkarni, Q.F. Yang, B.G. Haffty, J. Tommiska, C. Blomqvist, R. Drapkin, D.J. Adams, H. Nevanlinna, J. Bartek, M. Tarsounas, S. Ganesan, and J. Jonkers. 2010b. 53BP1 loss rescues BRCA1 deficiency and is associated with triple- negative and BRCA-mutated breast cancers. Nature Structural & Molecular Biology 17:688-U656.

Bristow, R.G., and R.P. Hill. 2008. Hypoxia and metabolism. Hypoxia, DNA repair and genetic instability. Nature Reviews Cancer 8:180-192.

Brou, C., F. Logeat, N. Gupta, C. Bessia, O. LeBail, J.R. Doedens, A. Cumano, P. Roux, R.A. Black, and A. Israel. 2000. A novel proteolytic cleavage involved in Notch signaling: the role of the disintegrin-metalloprotease TACE. Mol Cell 5:207-216.

Butler, L.R., R.M. Densham, J. Jia, A.J. Garvin, H.R. Stone, V. Shah, D. Weekes, F. Festy, J. Beesley, and J.R. Morris. 2012a. The proteasomal de-ubiquitinating enzyme POH1 promotes the double-strand DNA break response. EMBO J 31:3918-3934.

169

Butler, L.R., R.M. Densham, J.Y. Jia, A.J. Garvin, H.R. Stone, V. Shah, D. Weekes, F. Festy, J. Beesley, and J.R. Morris. 2012b. The proteasomal de-ubiquitinating enzyme POH1 promotes the double-strand DNA break response. Embo Journal 31:3918-3934.

Byrnes, G.B., M.C. Southey, and J.L. Hopper. 2008. Are the so-called low penetrance breast cancer genes, ATM, BRIP1, PALB2 and CHEK2, high risk for women with strong family histories? Breast Cancer Res 10:208.

Campeau, P.M., W.D. Foulkes, and M.D. Tischkowitz. 2008. Hereditary breast cancer: new genetic developments, new therapeutic avenues. Hum Genet 124:31-42.

Cardiff, R.D., and S.R. Wellings. 1999. The comparative pathology of human and mouse mammary glands. J Mammary Gland Biol Neoplasia 4:105-122.

Carr, J.R., M.M. Kiefer, H.J. Park, J. Li, Z. Wang, J. Fontanarosa, D. DeWaal, D. Kopanja, E.V. Benevolenskaya, G. Guzman, and P. Raychaudhuri. 2012. FoxM1 regulates mammary luminal cell fate. Cell Rep 1:715-729.

Celeste, A., S. Petersen, P.J. Romanienko, O. Fernandez-Capetillo, H.T. Chen, O.A. Sedelnikova, B. Reina-San-Martin, V. Coppola, E. Meffre, M.J. Difilippantonio, C. Redon, D.R. Pilch, A. Olaru, M. Eckhaus, R.D. Camerini-Otero, L. Tessarollo, F. Livak, K. Manova, W.M. Bonner, M.C. Nussenzweig, and A. Nussenzweig. 2002. Genomic instability in mice lacking histone H2AX. Science 296:922-927.

Cha, H., J.M. Lowe, H. Li, J.S. Lee, G.I. Belova, D.V. Bulavin, and A.J. Fornace, Jr. 2010. Wip1 directly dephosphorylates gamma-H2AX and attenuates the DNA damage response. Cancer Res 70:4112-4122.

Chakrabarti, R., Y. Wei, R.A. Romano, C. DeCoste, Y. Kang, and S. Sinha. 2012. Elf5 regulates mammary gland stem/progenitor cell fate by influencing notch signaling. Stem Cells 30:1496-1508.

Chan, T.A., S. Glockner, J.M. Yi, W. Chen, L. Van Neste, L. Cope, J.G. Herman, V. Velculescu, K.E. Schuebel, N. Ahuja, and S.B. Baylin. 2008. Convergence of mutation and epigenetic alterations identifies common genes in cancer that predict for poor prognosis. PLoS Med 5:e114.

Chapman, J.R., P. Barral, J.B. Vannier, V. Borel, M. Steger, A. Tomas-Loba, A.A. Sartori, I.R. Adams, F.D. Batista, and S.J. Boulton. 2013a. RIF1 Is Essential for 53BP1-Dependent Nonhomologous End Joining and Suppression of DNA Double-Strand Break Resection. Molecular Cell 49:858-871.

Chapman, J.R., P. Barral, J.B. Vannier, V. Borel, M. Steger, A. Tomas-Loba, A.A. Sartori, I.R. Adams, F.D. Batista, and S.J. Boulton. 2013b. RIF1 is essential for 53BP1-dependent nonhomologous end joining and suppression of DNA double-strand break resection. Mol Cell 49:858-871.

170

Chaudhuri, J., U. Basu, A. Zarrin, C. Yan, S. Franco, T. Perlot, B. Vuong, J. Wang, R.T. Phan, A. Datta, J. Manis, and F.W. Alt. 2007. Evolution of the immunoglobulin heavy chain class switch recombination mechanism. Adv Immunol 94:157-214.

Chen, J., W. Feng, J. Jiang, Y. Deng, and M.S. Huen. 2012. Ring finger protein RNF169 antagonizes the ubiquitin-dependent signaling cascade at sites of DNA damage. J Biol Chem 287:27715-27722.

Chowdhury, D., M.C. Keogh, H. Ishii, C.L. Peterson, S. Buratowski, and J. Lieberman. 2005. gamma-H2AX dephosphorylation by protein phosphatase 2A facilitates DNA double- strand break repair. Mol Cell 20:801-809.

Chowdhury, D., X. Xu, X. Zhong, F. Ahmed, J. Zhong, J. Liao, D.M. Dykxhoorn, D.M. Weinstock, G.P. Pfeifer, and J. Lieberman. 2008. A PP4-phosphatase complex dephosphorylates gamma-H2AX generated during DNA replication. Mol Cell 31:33-46.

Ciccia, A., and S.J. Elledge. 2010. The DNA damage response: making it safe to play with knives. Mol Cell 40:179-204.

Clerici, M., D. Mantiero, I. Guerini, G. Lucchini, and M.P. Longhese. 2008. The Yku70-Yku80 complex contributes to regulate double-strand break processing and checkpoint activation during the cell cycle. EMBO Rep 9:810-818.

Cook, P.J., B.G. Ju, F. Telese, X. Wang, C.K. Glass, and M.G. Rosenfeld. 2009. Tyrosine dephosphorylation of H2AX modulates apoptosis and survival decisions. Nature 458:591-596.

Costa, Y., R. Speed, R. Ollinger, M. Alsheimer, C.A. Semple, P. Gautier, K. Maratou, I. Novak, C. Hoog, R. Benavente, and H.J. Cooke. 2005. Two novel proteins recruited by synaptonemal complex protein 1 (SYCP1) are at the centre of meiosis. Journal of Cell Science 118:2755-2762.

D'Amours, D., and S.P. Jackson. 2002. The Mre11 complex: at the crossroads of dna repair and checkpoint signalling. Nat Rev Mol Cell Biol 3:317-327.

Daley, J.M., and P. Sung. 2013. RIF1 in DNA break repair pathway choice. Mol Cell 49:840- 841.

Daniel, K., J. Lange, K. Hached, J. Fu, K. Anastassiadis, I. Roig, H.J. Cooke, A.F. Stewart, K. Wassmann, M. Jasin, S. Keeney, and A. Toth. 2011. Meiotic homologue alignment and its quality surveillance are controlled by mouse HORMAD1. Nat Cell Biol 13:599-610.

Danielsen, J.R., L.K. Povlsen, B.H. Villumsen, W. Streicher, J. Nilsson, M. Wikstrom, S. Bekker-Jensen, and N. Mailand. 2012. DNA damage-inducible SUMOylation of HERC2 promotes RNF8 binding via a novel SUMO-binding Zinc finger. J Cell Biol 197:179-187.

De Strooper, B., W. Annaert, P. Cupers, P. Saftig, K. Craessaerts, J.S. Mumm, E.H. Schroeter, V. Schrijvers, M.S. Wolfe, W.J. Ray, A. Goate, and R. Kopan. 1999. A presenilin-1-

171

dependent gamma-secretase-like protease mediates release of Notch intracellular domain. Nature 398:518-522.

Delgado-Diaz, M.R., Y. Martin, A. Berg, R. Freire, and V.A. Smits. 2014. Dub3 controls DNA damage signalling by direct deubiquitination of H2AX. Mol Oncol 8:884-893.

Dievart, A., N. Beaulieu, and P. Jolicoeur. 1999. Involvement of Notch1 in the development of mouse mammary tumors. Oncogene 18:5973-5981.

Difilippantonio, S., E. Gapud, N. Wong, C.Y. Huang, G. Mahowald, H.T. Chen, M.J. Kruhlak, E. Callen, F. Livak, M.C. Nussenzweig, B.P. Sleckman, and A. Nussenzweig. 2008. 53BP1 facilitates long-range DNA end-joining during V(D)J recombination. Nature 456:529-533.

Doil, C., N. Mailand, S. Bekker-Jensen, P. Menard, D.H. Larsen, R. Pepperkok, J. Ellenberg, S. Panier, D. Durocher, J. Bartek, J. Lukas, and C. Lukas. 2009. RNF168 binds and amplifies ubiquitin conjugates on damaged chromosomes to allow accumulation of repair proteins. Cell 136:435-446.

Donehower, L.A., and G. Lozano. 2009. 20 years studying p53 functions in genetically engineered mice. Nat Rev Cancer 9:831-841.

Dou, H., C. Huang, M. Singh, P.B. Carpenter, and E.T.H. Yeh. 2010. Regulation of DNA Repair through DeSUMOylation and SUMOylation of Replication Protein A Complex. Molecular Cell 39:333-345.

Douglas, P., J. Zhong, R. Ye, G.B. Moorhead, X. Xu, and S.P. Lees-Miller. 2010. Protein phosphatase 6 interacts with the DNA-dependent protein kinase catalytic subunit and dephosphorylates gamma-H2AX. Mol Cell Biol 30:1368-1381.

Drost, R.M., and J. Jonkers. 2009. Preclinical mouse models for BRCA1-associated breast cancer. Br J Cancer 101:1651-1657.

Durandy, A., N. Taubenheim, S. Peron, and A. Fischer. 2007. Pathophysiology of B-cell intrinsic immunoglobulin class switch recombination deficiencies. Adv Immunol 94:275-306.

Dynan, W.S., and S. Yoo. 1998. Interaction of Ku protein and DNA-dependent protein kinase catalytic subunit with nucleic acids. Nucleic Acids Res 26:1551-1559.

Easton, D.F. 1999. How many more breast cancer predisposition genes are there? Breast Cancer Res 1:14-17.

Economopoulou, P., G. Dimitriadis, and A. Psyrri. 2015. Beyond BRCA: New hereditary breast cancer susceptibility genes. Cancer Treat Rev 41:1-8.

Eijpe, M., C. Heyting, B. Gross, and R. Jessberger. 2000. Association of mammalian SMC1 and SMC3 proteins with meiotic chromosomes and synaptonemal complexes. Journal of Cell Science 113 ( Pt 4):673-682.

172

Eijpe, M., H. Offenberg, R. Jessberger, E. Revenkova, and C. Heyting. 2003. Meiotic cohesin REC8 marks the axial elements of rat synaptonemal complexes before cohesins SMC1beta and SMC3. J Cell Biol 160:657-670.

Facchino, S., M. Abdouh, W. Chatoo, and G. Bernier. 2010. BMI1 Confers Radioresistance to Normal and Cancerous Neural Stem Cells through Recruitment of the DNA Damage Response Machinery. Journal of Neuroscience 30:10096-10111.

Farmer, H., N. McCabe, C.J. Lord, A.N. Tutt, D.A. Johnson, T.B. Richardson, M. Santarosa, K.J. Dillon, I. Hickson, C. Knights, N.M. Martin, S.P. Jackson, G.C. Smith, and A. Ashworth. 2005. Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature 434:917-921.

Farnie, G., R.B. Clarke, K. Spence, N. Pinnock, K. Brennan, N.G. Anderson, and N.J. Bundred. 2007. Novel cell culture technique for primary ductal carcinoma in situ: role of Notch and epidermal growth factor receptor signaling pathways. J Natl Cancer Inst 99:616-627.

Feng, L., and J. Chen. 2012. The E3 ligase RNF8 regulates KU80 removal and NHEJ repair. Nat Struct Mol Biol 19:201-206.

Feng, L., K.W. Fong, J. Wang, W. Wang, and J. Chen. 2013. RIF1 counteracts BRCA1-mediated end resection during DNA repair. J Biol Chem 288:11135-11143.

Fernandez-Capetillo, O., S.K. Mahadevaiah, A. Celeste, P.J. Romanienko, R.D. Camerini-Otero, W.M. Bonner, K. Manova, P. Burgoyne, and A. Nussenzweig. 2003. H2AX is required for chromatin remodeling and inactivation of sex chromosomes in male mouse meiosis. Developmental Cell 4:497-508.

Floyd, S.R., M.E. Pacold, Q. Huang, S.M. Clarke, F.C. Lam, I.G. Cannell, B.D. Bryson, J. Rameseder, M.J. Lee, E.J. Blake, A. Fydrych, R. Ho, B.A. Greenberger, G.C. Chen, A. Maffa, A.M. Del Rosario, D.E. Root, A.E. Carpenter, W.C. Hahn, D.M. Sabatini, C.C. Chen, F.M. White, J.E. Bradner, and M.B. Yaffe. 2013. The bromodomain protein Brd4 insulates chromatin from DNA damage signalling. Nature 498:246-250.

Fradet-Turcotte, A., M.D. Canny, C. Escribano-Diaz, A. Orthwein, C.C.Y. Leung, H. Huang, M.C. Landry, J. Kitevski-LeBlanc, S.M. Noordermeer, F. Sicheri, and D. Durocher. 2013. 53BP1 is a reader of the DNA-damage-induced H2A Lys 15 ubiquitin mark. Nature 499:50-+.

Freed-Pastor, W.A., and C. Prives. 2012. Mutant p53: one name, many proteins. Genes Dev 26:1268-1286.

Fukuda, T., K. Daniel, L. Wojtasz, A. Toth, and C. Hoog. 2010. A novel mammalian HORMA domain-containing protein, HORMAD1, preferentially associates with unsynapsed meiotic chromosomes. Experimental Cell Research 316:158-171.

Galanty, Y., R. Belotserkovskaya, J. Coates, and S.P. Jackson. 2012. RNF4, a SUMO-targeted ubiquitin E3 ligase, promotes DNA double-strand break repair. Genes & Development 26:1179-1195.

173

Galanty, Y., R. Belotserkovskaya, J. Coates, S. Polo, K.M. Miller, and S.P. Jackson. 2009. Mammalian SUMO E3-ligases PIAS1 and PIAS4 promote responses to DNA double- strand breaks. Nature 462:935-U132.

Ghosal, G., and J. Chen. 2013. DNA damage tolerance: a double-edged sword guarding the genome. Transl Cancer Res 2:107-129.

Ghoussaini, M., P.D. Pharoah, and D.F. Easton. 2013. Inherited genetic susceptibility to breast cancer: the beginning of the end or the end of the beginning? Am J Pathol 183:1038- 1051.

Gilardini Montani, M.S., A. Prodosmo, V. Stagni, D. Merli, L. Monteonofrio, V. Gatti, M.P. Gentileschi, D. Barila, and S. Soddu. 2013. ATM-depletion in breast cancer cells confers sensitivity to PARP inhibition. J Exp Clin Cancer Res 32:95.

Goodarzi, A.A., and P.A. Jeggo. 2013. The Repair and Signaling Responses to DNA Double- Strand Breaks. Advances in Genetics, Vol 82 82:1-45.

Grabarz, A., A. Barascu, J. Guirouilh-Barbat, and B.S. Lopez. 2012. Initiation of DNA double strand break repair: signaling and single-stranded resection dictate the choice between homologous recombination, non-homologous end-joining and alternative end-joining. Am J Cancer Res 2:249-268.

Grotsky, D.A., I. Gonzalez-Suarez, A. Novell, M.A. Neumann, S.C. Yaddanapudi, M. Croke, M. Martinez-Alonso, A.B. Redwood, S. Ortega-Martinez, Z. Feng, E. Lerma, T. Ramon y Cajal, J. Zhang, X. Matias-Guiu, A. Dusso, and S. Gonzalo. 2013. BRCA1 loss activates cathepsin L-mediated degradation of 53BP1 in breast cancer cells. J Cell Biol 200:187- 202.

Gudjonsson, T., M. Altmeyer, V. Savic, L. Toledo, C. Dinant, M. Grofte, J. Bartkova, M. Poulsen, Y. Oka, S. Bekker-Jensen, N. Mailand, B. Neumann, J.K. Heriche, R. Shearer, D. Saunders, J. Bartek, J. Lukas, and C. Lukas. 2012. TRIP12 and UBR5 Suppress Spreading of Chromatin Ubiquitylation at Damaged Chromosomes. Cell 150:697-709.

Guilford, P., J. Hopkins, J. Harraway, M. McLeod, N. McLeod, P. Harawira, H. Taite, R. Scoular, A. Miller, and A.E. Reeve. 1998. E-cadherin germline mutations in familial gastric cancer. Nature 392:402-405.

Gumy-Pause, F., H. Ozsahin, M. Khoshbeen-Boudal, D.R. Betts, P. Maillet, and A.P. Sappino. 2008. Detection of ATM gene deletion/duplication by multiplex ligation-dependant probe amplification in childhood lymphoid malignancies: a report from the Children's Oncology Group. Leuk Res 32:1207-1213.

Gumy-Pause, F., P. Wacker, and A.P. Sappino. 2004. ATM gene and lymphoid malignancies. Leukemia 18:238-242.

Guruharsha, K.G., M.W. Kankel, and S. Artavanis-Tsakonas. 2012. The Notch signalling system: recent insights into the complexity of a conserved pathway. Nat Rev Genet 13:654-666.

174

Hakem, R. 2008. DNA-damage repair; the good, the bad, and the ugly. EMBO J 27:589-605.

Hakem, R., J.L. delaPompa, C. Sirard, R. Mo, M. Woo, A. Hakem, A. Wakeham, J. Potter, A. Reitmair, F. Billia, E. Firpo, C.C. Hui, J. Roberts, J. Rossant, and T.W. Mak. 1996. The tumor suppressor gene Brca1 is required for embryonic cellular proliferation in the mouse. Cell 85:1009-1023.

Halaby, M.J., A. Hakem, L. Li, S. El Ghamrasni, S. Venkatesan, P.M. Hande, O. Sanchez, and R. Hakem. 2013. Synergistic interaction of Rnf8 and p53 in the protection against genomic instability and tumorigenesis. PLoS Genet 9:e1003259.

Hamer, G., K. Gell, A. Kouznetsova, I. Novak, R. Benavente, and C. Hoog. 2006. Characterization of a novel meiosis-specific protein within the central element of the synaptonemal complex. Journal of Cell Science 119:4025-4032.

Han, J., C. Haiman, T. Niu, Q. Guo, D.G. Cox, W.C. Willett, S.E. Hankinson, and D.J. Hunter. 2009. Genetic variation in DNA repair pathway genes and premenopausal breast cancer risk. Breast Cancer Res Treat 115:613-622.

Hanahan, D., and R.A. Weinberg. 2011. Hallmarks of cancer: the next generation. Cell 144:646- 674.

Harper, J.W., and S.J. Elledge. 2007. The DNA damage response: ten years after. Mol Cell 28:739-745.

Hennighausen, L., and G.W. Robinson. 2005. Information networks in the mammary gland. Nature Reviews Molecular Cell Biology 6:715-725.

Henriksson, S., H. Rassoolzadeh, E. Hedstrom, C. Coucoravas, A. Julner, M. Goldstein, G. Imreh, B. Zhivotovsky, M.B. Kastan, T. Helleday, and M. Farnebo. 2014. The scaffold protein WRAP53beta orchestrates the ubiquitin response critical for DNA double-strand break repair. Genes Dev 28:2726-2738.

Herschkowitz, J.I., K. Simin, V.J. Weigman, I. Mikaelian, J. Usary, Z. Hu, K.E. Rasmussen, L.P. Jones, S. Assefnia, S. Chandrasekharan, M.G. Backlund, Y. Yin, A.I. Khramtsov, R. Bastein, J. Quackenbush, R.I. Glazer, P.H. Brown, J.E. Green, L. Kopelovich, P.A. Furth, J.P. Palazzo, O.I. Olopade, P.S. Bernard, G.A. Churchill, T. Van Dyke, and C.M. Perou. 2007. Identification of conserved gene expression features between murine mammary carcinoma models and human breast tumors. Genome Biol 8:R76.

Heyer, W.D., K.T. Ehmsen, and J. Liu. 2010. Regulation of homologous recombination in eukaryotes. Annu Rev Genet 44:113-139.

Hoeijmakers, J.H. 2001. Genome maintenance mechanisms for preventing cancer. Nature 411:366-374.

Hoeijmakers, J.H. 2009. DNA damage, aging, and cancer. N Engl J Med 361:1475-1485.

175

Hoeller, D., C.M. Hecker, and I. Dikic. 2006. Ubiquitin and ubiquitin-like proteins in cancer pathogenesis. Nat Rev Cancer 6:776-788.

Hosoya, N., and K. Miyagawa. 2014. Targeting DNA damage response in cancer therapy. Cancer Science 105:370-388.

Huen, M.S., R. Grant, I. Manke, K. Minn, X. Yu, M.B. Yaffe, and J. Chen. 2007. RNF8 transduces the DNA-damage signal via histone ubiquitylation and checkpoint protein assembly. Cell 131:901-914.

Huen, M.S., J. Huang, J. Yuan, M. Yamamoto, S. Akira, C. Ashley, W. Xiao, and J. Chen. 2008. Noncanonical E2 variant-independent function of UBC13 in promoting checkpoint protein assembly. Mol Cell Biol 28:6104-6112.

Huertas, P., F. Cortes-Ledesma, A.A. Sartori, A. Aguilera, and S.P. Jackson. 2008. CDK targets Sae2 to control DNA-end resection and homologous recombination. Nature 455:689-692.

Huertas, P., and S.P. Jackson. 2009. Human CtIP mediates cell cycle control of DNA end resection and double strand break repair. J Biol Chem 284:9558-9565.

Hussain, S.P., L.J. Hofseth, and C.C. Harris. 2003. Radical causes of cancer. Nat Rev Cancer 3:276-285.

Ikura, T., S. Tashiro, A. Kakino, H. Shima, N. Jacob, R. Amunugama, K. Yoder, S. Izumi, I. Kuraoka, K. Tanaka, H. Kimura, M. Ikura, S. Nishikubo, T. Ito, A. Muto, K. Miyagawa, S. Takeda, R. Fishel, K. Igarashi, and K. Kamiya. 2007. DNA damage-dependent acetylation and ubiquitination of H2AX enhances chromatin dynamics. Mol Cell Biol 27:7028-7040.

Imatani, A., and R. Callahan. 2000. Identification of a novel NOTCH-4/INT-3 RNA species encoding an activated gene product in certain human tumor cell lines. Oncogene 19:223- 231.

Inagaki, A., S. Schoenmakers, and W.M. Baarends. 2010. DNA double strand break repair, chromosome synapsis and transcriptional silencing in meiosis. Epigenetics 5:255-266.

Ira, G., A. Pellicioli, A. Balijja, X. Wang, S. Fiorani, W. Carotenuto, G. Liberi, D. Bressan, L. Wan, N.M. Hollingsworth, J.E. Haber, and M. Foiani. 2004. DNA end resection, homologous recombination and DNA damage checkpoint activation require CDK1. Nature 431:1011-1017.

Ismail, I.H., C. Andrin, D. McDonald, and M.J. Hendzel. 2010. BMI1-mediated histone ubiquitylation promotes DNA double-strand break repair. J Cell Biol 191:45-60.

Jackson, S.P., and J. Bartek. 2009. The DNA-damage response in human biology and disease. Nature 461:1071-1078.

Jackson, S.P., and D. Durocher. 2013. Regulation of DNA Damage Responses by Ubiquitin and SUMO. Molecular Cell 49:795-807.

176

Jasin, M., and R. Rothstein. 2013. Repair of strand breaks by homologous recombination. Cold Spring Harb Perspect Biol 5:a012740.

Jazayeri, A., J. Falck, C. Lukas, J. Bartek, G.C. Smith, J. Lukas, and S.P. Jackson. 2006. ATM- and cell cycle-dependent regulation of ATR in response to DNA double-strand breaks. Nat Cell Biol 8:37-45.

Jhappan, C., D. Gallahan, C. Stahle, E. Chu, G.H. Smith, G. Merlino, and R. Callahan. 1992. Expression of an activated Notch-related int-3 transgene interferes with cell differentiation and induces neoplastic transformation in mammary and salivary glands. Genes Dev 6:345-355.

Jonkers, J., R. Meuwissen, H. van der Gulden, H. Peterse, M. van der Valk, and A. Berns. 2001. Synergistic tumor suppressor activity of BRCA2 and p53 in a conditional mouse model for breast cancer. Nat Genet 29:418-425.

Joshi, P.A., H.W. Jackson, A.G. Beristain, M.A. Di Grappa, P.A. Mote, C.L. Clarke, J. Stingl, P.D. Waterhouse, and R. Khokha. 2010. Progesterone induces adult mammary stem cell expansion. Nature 465:803-807.

Jungmichel, S., and M. Stucki. 2010. MDC1: The art of keeping things in focus. Chromosoma 119:337-349.

Kaidi, A., and S.P. Jackson. 2013. KAT5 tyrosine phosphorylation couples chromatin sensing to ATM signalling. Nature 498:70-74.

Kato, K., K. Nakajima, A. Ui, Y. Muto-Terao, H. Ogiwara, and S. Nakada. 2014. Fine-tuning of DNA damage-dependent ubiquitination by OTUB2 supports the DNA repair pathway choice. Mol Cell 53:617-630.

Keim, C., D. Kazadi, G. Rothschild, and U. Basu. 2013. Regulation of AID, the B-cell genome mutator. Genes Dev 27:1-17.

Khoo, K.H., C.S. Verma, and D.P. Lane. 2014. Drugging the p53 pathway: understanding the route to clinical efficacy. Nat Rev Drug Discov 13:217-236.

Kiaris, H., K. Politi, L.M. Grimm, M. Szabolcs, P. Fisher, A. Efstratiadis, and S. Artavanis- Tsakonas. 2004. Modulation of notch signaling elicits signature tumors and inhibits hras1-induced oncogenesis in the mouse mammary epithelium. Am J Pathol 165:695- 705.

Kim, J.A., M. Kruhlak, F. Dotiwala, A. Nussenzweig, and J.E. Haber. 2007. Heterochromatin is refractory to gamma-H2AX modification in yeast and mammals. J Cell Biol 178:209- 218.

Klaunig, J.E., L.M. Kamendulis, and B.A. Hocevar. 2010. Oxidative stress and oxidative damage in carcinogenesis. Toxicol Pathol 38:96-109.

177

Kleckner, N. 1996. Meiosis: How could it work? Proceedings of the National Academy of Sciences of the United States of America 93:8167-8174.

Kobayashi, J., H. Tauchi, S. Sakamoto, A. Nakamura, K. Morishima, S. Matsuura, T. Kobayashi, K. Tamai, K. Tanimoto, and K. Komatsu. 2002. NBS1 localizes to gamma-H2AX foci through interaction with the FHA/BRCT domain. Curr Biol 12:1846-1851.

Kolas, N.K., J.R. Chapman, S. Nakada, J. Ylanko, R. Chahwan, F.D. Sweeney, S. Panier, M. Mendez, J. Wildenhain, T.M. Thomson, L. Pelletier, S.P. Jackson, and D. Durocher. 2007a. Orchestration of the DNA-damage response by the RNF8 ubiquitin ligase. Science 318:1637-1640.

Kolas, N.K., J.R. Chapman, S. Nakada, J. Ylanko, R. Chahwan, F.D. Sweeney, S. Panier, M. Mendez, J. Wildenhain, T.M. Thomson, L. Pelletier, S.P. Jackson, and D. Durocher. 2007b. Orchestration of the DNA-damage response by the RNF8 ubiquitin ligase. Science 318:1637-1640.

Kordon, E.C., R.A. McKnight, C. Jhappan, L. Hennighausen, G. Merlino, and G.H. Smith. 1995. Ectopic TGF beta 1 expression in the secretory mammary epithelium induces early senescence of the epithelial stem cell population. Dev Biol 168:47-61.

Kordon, E.C., and G.H. Smith. 1998. An entire functional mammary gland may comprise the progeny from a single cell. Development 125:1921-1930.

Koster, D.A., K. Palle, E.S. Bot, M.A. Bjornsti, and N.H. Dekker. 2007. Antitumour drugs impede DNA uncoiling by topoisomerase I. Nature 448:213-217.

Kotnis, A., L. Du, C. Liu, S.W. Popov, and Q. Pan-Hammarstrom. 2009. Non-homologous end joining in class switch recombination: the beginning of the end. Philos Trans R Soc Lond B Biol Sci 364:653-665.

Lamlum, H., M. Ilyas, A. Rowan, S. Clark, V. Johnson, J. Bell, I. Frayling, J. Efstathiou, K. Pack, S. Payne, R. Roylance, P. Gorman, D. Sheer, K. Neale, R. Phillips, I. Talbot, W. Bodmer, and I. Tomlinson. 1999. The type of somatic mutation at APC in familial adenomatous polyposis is determined by the site of the germline mutation: a new facet to Knudson's 'two-hit' hypothesis. Nat Med 5:1071-1075.

Lavin, M.F., and Y. Shiloh. 1997. The genetic defect in ataxia-telangiectasia. Annu Rev Immunol 15:177-202.

Lazzaro, F., V. Sapountzi, M. Granata, A. Pellicioli, M. Vaze, J.E. Haber, P. Plevani, D. Lydall, and M. Muzi-Falconi. 2008. Histone methyltransferase Dot1 and Rad9 inhibit single- stranded DNA accumulation at DSBs and uncapped telomeres. EMBO J 27:1502-1512.

Lee, C.W., K. Simin, Q. Liu, J. Plescia, M. Guha, A. Khan, C.C. Hsieh, and D.C. Altieri. 2008a. A functional Notch-survivin gene signature in basal breast cancer. Breast Cancer Res 10:R97.

178

Lee, C.W., K. Simin, Q. Liu, J. Plescia, M. Guha, A. Khan, C.C. Hsieh, and D.C. Altieri. 2008b. A functional Notch-survivin gene signature in basal breast cancer. Breast Cancer Research 10:

Lee, J.M., J.A. Ledermann, and E.C. Kohn. 2014. PARP Inhibitors for BRCA1/2 mutation- associated and BRCA-like malignancies. Ann Oncol 25:32-40.

Li, G., G.W. Robinson, R. Lesche, H. Martinez-Diaz, Z. Jiang, N. Rozengurt, K.U. Wagner, D.C. Wu, T.F. Lane, X. Liu, L. Hennighausen, and H. Wu. 2002. Conditional loss of PTEN leads to precocious development and neoplasia in the mammary gland. Development 129:4159-4170.

Li, L., M.J. Halaby, A. Hakem, R. Cardoso, S. El Ghamrasni, S. Harding, N. Chan, R. Bristow, O. Sanchez, D. Durocher, and R. Hakem. 2010a. Rnf8 deficiency impairs class switch recombination, spermatogenesis, and genomic integrity and predisposes for cancer. J Exp Med 207:983-997.

Li, Y., J.A. Burns, C.A. Cheney, N. Zhang, S. Vitelli, F. Wang, A. Bett, M. Chastain, L.P. Audoly, and Z.Q. Zhang. 2010b. Distinct expression profiles of Notch-1 protein in human solid tumors: Implications for development of targeted therapeutic monoclonal antibodies. Biologics 4:163-171.

Lieber, M.R. 2010. The mechanism of double-strand DNA break repair by the nonhomologous DNA end-joining pathway. Annu Rev Biochem 79:181-211.

Lieber, T., S. Kidd, and M.W. Young. 2002. kuzbanian-mediated cleavage of Drosophila Notch. Genes Dev 16:209-221.

Lim, E., F. Vaillant, D. Wu, N.C. Forrest, B. Pal, A.H. Hart, M.L. Asselin-Labat, D.E. Gyorki, T. Ward, A. Partanen, F. Feleppa, L.I. Huschtscha, H.J. Thorne, S.B. Fox, M. Yan, J.D. French, M.A. Brown, G.K. Smyth, J.E. Visvader, and G.J. Lindeman. 2009a. Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat Med 15:907-913.

Lim, E., F. Vaillant, D. Wu, N.C. Forrest, B. Pal, A.H. Hart, M.L. Asselin-Labat, D.E. Gyorki, T. Ward, A. Partanen, F. Feleppa, L.I. Huschtscha, H.J. Thorne, kConFab, S.B. Fox, M. Yan, J.D. French, M.A. Brown, G.K. Smyth, J.E. Visvader, and G.J. Lindeman. 2009b. Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat Med 15:907-913.

Liu, C., S. Srihari, K.A.L. Cao, G. Chenevix-Trench, P.T. Simpson, M.A. Ragan, and K.K. Khanna. 2014. A fine-scale dissection of the DNA double-strand break repair machinery and its implications for breast cancer therapy. Nucleic Acids Research 42:6106-6127.

Liu, S., C. Ginestier, E. Charafe-Jauffret, H. Foco, C.G. Kleer, S.D. Merajver, G. Dontu, and M.S. Wicha. 2008. BRCA1 regulates human mammary stem/progenitor cell fate. Proc Natl Acad Sci U S A 105:1680-1685.

179

Logeat, F., C. Bessia, C. Brou, O. LeBail, S. Jarriault, N.G. Seidah, and A. Israel. 1998. The Notch1 receptor is cleaved constitutively by a furin-like convertase. Proceedings of the National Academy of Sciences of the United States of America 95:8108-8112.

Longhese, M.P., D. Bonetti, I. Guerini, N. Manfrini, and M. Clerici. 2009. DNA double-strand breaks in meiosis: checking their formation, processing and repair. DNA Repair (Amst) 8:1127-1138.

Lord, C.J., and A. Ashworth. 2013. Mechanisms of resistance to therapies targeting BRCA- mutant cancers. Nat Med 19:1381-1388.

Lou, Z., K. Minter-Dykhouse, S. Franco, M. Gostissa, M.A. Rivera, A. Celeste, J.P. Manis, J. van Deursen, A. Nussenzweig, T.T. Paull, F.W. Alt, and J. Chen. 2006. MDC1 maintains genomic stability by participating in the amplification of ATM-dependent DNA damage signals. Mol Cell 21:187-200.

Lukas, J., C. Lukas, and J. Bartek. 2011. More than just a focus: The chromatin response to DNA damage and its role in genome integrity maintenance. Nature Cell Biology 13:1161-1169.

Lumsden, J.M., T. McCarty, L.K. Petiniot, R. Shen, C. Barlow, T.A. Wynn, H.C. Morse, P.J. Gearhart, A. Wynshaw-Boris, E.E. Max, and R.J. Hodes. 2004. Immunoglobulin class switch recombination is impaired in Atm-deficient mice. Journal of Experimental Medicine 200:1111-1121.

Luo, K.T., H.X. Zhang, L.W. Wang, J. Yuan, and Z.K. Lou. 2012. Sumoylation of MDC1 is important for proper DNA damage response. Embo Journal 31:3008-3019.

Ma, T., Y. Chen, F. Zhang, C.Y. Yang, S. Wang, and X. Yu. 2013. RNF111-dependent neddylation activates DNA damage-induced ubiquitination. Mol Cell 49:897-907.

Macurek, L., A. Lindqvist, O. Voets, J. Kool, H.R. Vos, and R.H. Medema. 2010. Wip1 phosphatase is associated with chromatin and dephosphorylates gamma H2AX to promote checkpoint inhibition. Oncogene 29:2281-2291.

Mahadevaiah, S.K., J.M. Turner, F. Baudat, E.P. Rogakou, P. de Boer, J. Blanco-Rodriguez, M. Jasin, S. Keeney, W.M. Bonner, and P.S. Burgoyne. 2001. Recombinational DNA double-strand breaks in mice precede synapsis. Nat Genet 27:271-276.

Mailand, N., S. Bekker-Jensen, H. Faustrup, F. Melander, J. Bartek, C. Lukas, and J. Lukas. 2007. RNF8 ubiquitylates histones at DNA double-strand breaks and promotes assembly of repair proteins. Cell 131:887-900.

Mallette, F.A., F. Mattiroli, G.F. Cui, L.C. Young, M.J. Hendzel, G. Mer, T.K. Sixma, and S. Richard. 2012. RNF8-and RNF168-dependent degradation of KDM4A/JMJD2A triggers 53BP1 recruitment to DNA damage sites. Embo Journal 31:1865-1878.

Manis, J.P., J.C. Morales, Z. Xia, J.L. Kutok, F.W. Alt, and P.B. Carpenter. 2004. 53BP1 links DNA damage-response pathways to immunoglobulin heavy chain class-switch recombination. Nat Immunol 5:481-487.

180

Mao, Z., Y. Jiang, X. Liu, A. Seluanov, and V. Gorbunova. 2009. DNA repair by homologous recombination, but not by nonhomologous end joining, is elevated in breast cancer cells. Neoplasia 11:683-691.

Martin-Belmonte, F., and M. Perez-Moreno. 2012. Epithelial cell polarity, stem cells and cancer. Nat Rev Cancer 12:23-38.

Matsuura, S., J. Kobayashi, H. Tauchi, and K. Komatsu. 2004. Nijmegen breakage syndrome and DNA double strand break repair by NBS1 complex. Adv Biophys 38:65-80.

Mattiroli, F., J.H. Vissers, W.J. van Dijk, P. Ikpa, E. Citterio, W. Vermeulen, J.A. Marteijn, and T.K. Sixma. 2012. RNF168 ubiquitinates K13-15 on H2A/H2AX to drive DNA damage signaling. Cell 150:1182-1195.

Mavaddat, N., A.C. Antoniou, D.F. Easton, and M. Garcia-Closas. 2010. Genetic susceptibility to breast cancer. Mol Oncol 4:174-191.

Mavaddat, N., A.M. Dunning, B.A. Ponder, D.F. Easton, and P.D. Pharoah. 2009. Common genetic variation in candidate genes and susceptibility to subtypes of breast cancer. Cancer Epidemiol Biomarkers Prev 18:255-259.

Maxwell, K.N., and K.L. Nathanson. 2013. Common breast cancer risk variants in the post- COGS era: a comprehensive review. Breast Cancer Res 15:212.

McCullough, L.E., R.M. Santella, R.J. Cleveland, R.C. Millikan, A.F. Olshan, K.E. North, P.T. Bradshaw, S.M. Eng, M.B. Terry, J. Shen, K.D. Crew, P. Rossner, Jr., S.L. Teitelbaum, A.I. Neugut, and M.D. Gammon. 2014. Polymorphisms in DNA repair genes, recreational physical activity and breast cancer risk. Int J Cancer 134:654-663.

McKenzie, G.J., P. Stevenson, G. Ward, S. Papadia, H. Bading, S. Chawla, M. Privalsky, and G.E. Hardingham. 2005. Nuclear Ca2+ and CaM kinase IV specify hormonal- and Notch- responsiveness. J Neurochem 93:171-185.

McPherson, J.P., N. Lemmers, A. Hirao, A. Hakem, J. Abraham, E. Migon, E. Matysiak- Zablocki, L. Tamblyn, O. Sanchez-Sweatman, R. Khokha, J. Squire, M.P. Hande, T.W. Mak, and R. Hakem. 2004. Collaboration of Brca1 and Chk2 in tumorigenesis. Genes & Development 18:1144-1153.

McVey, M., and S.E. Lee. 2008. MMEJ repair of double-strand breaks (director's cut): deleted sequences and alternative endings. Trends Genet 24:529-538.

Meek, D.W. 2009. Tumour suppression by p53: a role for the DNA damage response? Nature Reviews Cancer 9:714-723.

Meerang, M., D. Ritz, S. Paliwal, Z. Garajova, M. Bosshard, N. Mailand, P. Janscak, U. Hubscher, H. Meyer, and K. Ramadan. 2011. The ubiquitin-selective segregase VCP/p97 orchestrates the response to DNA double-strand breaks. Nature Cell Biology 13:1376- U1204.

181

Mego, M., S.A. Mani, and M. Cristofanilli. 2010. Molecular mechanisms of metastasis in breast cancer--clinical applications. Nat Rev Clin Oncol 7:693-701.

Mehta, A., and J.E. Haber. 2014. Sources of DNA double-strand breaks and models of recombinational DNA repair. Cold Spring Harb Perspect Biol 6:a016428.

Meijers-Heijboer, H., A. van den Ouweland, J. Klijn, M. Wasielewski, A. de Snoo, R. Oldenburg, A. Hollestelle, M. Houben, E. Crepin, M. van Veghel-Plandsoen, F. Elstrodt, C. van Duijn, C. Bartels, C. Meijers, M. Schutte, L. McGuffog, D. Thompson, D. Easton, N. Sodha, S. Seal, R. Barfoot, J. Mangion, J. Chang-Claude, D. Eccles, R. Eeles, D.G. Evans, R. Houlston, V. Murday, S. Narod, T. Peretz, J. Peto, C. Phelan, H.X. Zhang, C. Szabo, P. Devilee, D. Goldgar, P.A. Futreal, K.L. Nathanson, B. Weber, N. Rahman, and M.R. Stratton. 2002. Low-penetrance susceptibility to breast cancer due to CHEK2(*)1100delC in noncarriers of BRCA1 or BRCA2 mutations. Nat Genet 31:55- 59.

Mimitou, E.P., and L.S. Symington. 2010. Ku prevents Exo1 and Sgs1-dependent resection of DNA ends in the absence of a functional MRX complex or Sae2. EMBO J 29:3358-3369.

Mittal, S., D. Subramanyam, D. Dey, R.V. Kumar, and A. Rangarajan. 2009. Cooperation of Notch and Ras/MAPK signaling pathways in human breast carcinogenesis. Mol Cancer 8:128.

Moens, P.B., D.J. Chen, Z. Shen, N. Kolas, M. Tarsounas, H.H. Heng, and B. Spyropoulos. 1997. Rad51 immunocytology in rat and mouse spermatocytes and oocytes. Chromosoma 106:207-215.

Moens, P.B., N.K. Kolas, M. Tarsounas, E. Marcon, P.E. Cohen, and B. Spyropoulos. 2002. The time course and chromosomal localization of recombination-related proteins at meiosis in the mouse are compatible with models that can resolve the early DNA-DNA interactions without reciprocal recombination. Journal of Cell Science 115:1611-1622.

Molyneux, G., F.C. Geyer, F.A. Magnay, A. McCarthy, H. Kendrick, R. Natrajan, A. Mackay, A. Grigoriadis, A. Tutt, A. Ashworth, J.S. Reis-Filho, and M.J. Smalley. 2010. BRCA1 basal-like breast cancers originate from luminal epithelial progenitors and not from basal stem cells. Cell Stem Cell 7:403-417.

Moon, S.H., L. Lin, X.N. Zhang, T.A. Nguyen, Y. Darlington, A.S. Waldman, X.B. Lu, and L.A. Donehower. 2010. Wild-type p53-induced Phosphatase 1 Dephosphorylates Histone Variant gamma-H2AX and Suppresses DNA Double Strand Break Repair. Journal of Biological Chemistry 285:12935-12947.

Morris, J.R., C. Boutell, M. Keppler, R. Densham, D. Weekes, A. Alamshah, L. Butler, Y. Galanty, L. Pangon, T. Kiuchi, T. Ng, and E. Solomon. 2009. The SUMO modification pathway is involved in the BRCA1 response to genotoxic stress. Nature 462:886-U877.

Mosbech, A., C. Lukas, S. Bekker-Jensen, and N. Mailand. 2013. The deubiquitylating enzyme USP44 counteracts the DNA double-strand break response mediated by the RNF8 and RNF168 ubiquitin ligases. J Biol Chem 288:16579-16587.

182

Mumm, J.S., E.H. Schroeter, M.T. Saxena, A. Griesemer, X. Tian, D.J. Pan, W.J. Ray, and R. Kopan. 2000. A ligand-induced extracellular cleavage regulates gamma-secretase-like proteolytic activation of Notch1. Mol Cell 5:197-206.

Munch, J., A. Gonzalez-Rajal, and J.L. de la Pompa. 2013. Notch regulates blastema proliferation and prevents differentiation during adult zebrafish fin regeneration. Development 140:1402-1411.

Muramatsu, M., K. Kinoshita, S. Fagarasan, S. Yamada, Y. Shinkai, and T. Honjo. 2000. Class switch recombination and hypermutation require activation-induced cytidine deaminase (AID), a potential RNA editing enzyme. Cell 102:553-563.

Murga, M., I. Jaco, Y. Fan, R. Soria, B. Martinez-Pastor, M. Cuadrado, S.M. Yang, M.A. Blasco, A.I. Skoultchi, and O. Fernandez-Capetillo. 2007. Global chromatin compaction limits the strength of the DNA damage response. J Cell Biol 178:1101-1108.

Nakada, S., G.I. Chen, A.C. Gingras, and D. Durocher. 2008. PP4 is a gamma H2AX phosphatase required for recovery from the DNA damage checkpoint. EMBO Rep 9:1019-1026.

Nelen, M.R., G.W. Padberg, E.A. Peeters, A.Y. Lin, B. van den Helm, R.R. Frants, V. Coulon, A.M. Goldstein, M.M. van Reen, D.F. Easton, R.A. Eeles, S. Hodgsen, J.J. Mulvihill, V.A. Murday, M.A. Tucker, E.C. Mariman, T.M. Starink, B.A. Ponder, H.H. Ropers, H. Kremer, M. Longy, and C. Eng. 1996. Localization of the gene for Cowden disease to chromosome 10q22-23. Nat Genet 13:114-116.

Nichols, K.E., S. Levitz, K.E. Shannon, D.C. Wahrer, D.W. Bell, G. Chang, S. Hegde, D. Neuberg, T. Shafman, N.J. Tarbell, P. Mauch, C. Ishioka, D.A. Haber, and L. Diller. 1999. Heterozygous germline ATM mutations do not contribute to radiation-associated malignancies after Hodgkin's disease. J Clin Oncol 17:1259.

Nishana, M., and S.C. Raghavan. 2012. Role of recombination activating genes in the generation of antigen receptor diversity and beyond. Immunology 137:271-281.

Noffsinger, A.E., J.M. Belli, F. Fogt, J. Fischer, H. Goldman, and C.M. Fenoglio-Preiser. 1999. A germline hMSH2 alteration is unrelated to colonic microsatellite instability in patients with ulcerative colitis. Hum Pathol 30:8-12.

O'Driscoll, M. 2012. Diseases associated with defective responses to DNA damage. Cold Spring Harb Perspect Biol 4:

O'Driscoll, M., and P.A. Jeggo. 2006. The role of double-strand break repair - insights from human genetics. Nat Rev Genet 7:45-54.

Offenberg, H.H., J.A. Schalk, R.L. Meuwissen, M. van Aalderen, H.A. Kester, A.J. Dietrich, and C. Heyting. 1998. SCP2: a major protein component of the axial elements of synaptonemal complexes of the rat. Nucleic Acids Research 26:2572-2579.

183

Olivier, M., A. Langerod, P. Carrieri, J. Bergh, S. Klaar, J. Eyfjord, C. Theillet, C. Rodriguez, R. Lidereau, I. Bieche, J. Varley, Y. Bignon, N. Uhrhammer, R. Winqvist, A. Jukkola- Vuorinen, D. Niederacher, S. Kato, C. Ishioka, P. Hainaut, and A.L. Borresen-Dale. 2006. The clinical value of somatic TP53 gene mutations in 1,794 patients with breast cancer. Clin Cancer Res 12:1157-1167.

Orkin, S. 1990. Molecular-Cloning - a Laboratory Manual, 2nd Edition - Sambrook,J, Fritsch,Ef, Maniatis,T. Nature 343:604-605.

Pamidi, A., R. Cardoso, A. Hakem, E. Matysiak-Zablocki, A. Poonepalli, L. Tamblyn, B. Perez- Ordonez, M.P. Hande, O. Sanchez, and R. Hakem. 2007. Functional interplay of p53 and Mus81 in DNA damage responses and cancer. Cancer Res 67:8527-8535.

Pan, M.R., G. Peng, W.C. Hung, and S.Y. Lin. 2011. Monoubiquitination of H2AX protein regulates DNA damage response signaling. J Biol Chem 286:28599-28607.

Panier, S., and D. Durocher. 2009a. Regulatory ubiquitylation in response to DNA double-strand breaks. DNA Repair 8:436-443.

Panier, S., and D. Durocher. 2009b. Regulatory ubiquitylation in response to DNA double-strand breaks. DNA Repair (Amst) 8:436-443.

Panier, S., and D. Durocher. 2013. Push back to respond better: regulatory inhibition of the DNA double-strand break response. Nat Rev Mol Cell Biol 14:661-672.

Panier, S., Y. Ichijima, A. Fradet-Turcotte, C.C. Leung, L. Kaustov, C.H. Arrowsmith, and D. Durocher. 2012a. Tandem protein interaction modules organize the ubiquitin-dependent response to DNA double-strand breaks. Mol Cell 47:383-395.

Panier, S., Y. Ichijima, A. Fradet-Turcotte, C.C.Y. Leung, L. Kaustov, C.H. Arrowsmith, and D. Durocher. 2012b. Tandem Protein Interaction Modules Organize the Ubiquitin- Dependent Response to DNA Double-Strand Breaks. Molecular Cell 47:383-395.

Papamichos-Chronakis, M., and C.L. Peterson. 2013. Chromatin and the genome integrity network. Nature Reviews Genetics 14:62-75.

Paulovich, A.G., D.P. Toczyski, and L.H. Hartwell. 1997. When checkpoints fail. Cell 88:315- 321.

Pece, S., M. Serresi, E. Santolini, M. Capra, E. Hulleman, V. Galimberti, S. Zurrida, P. Maisonneuve, G. Viale, and P.P. Di Fiore. 2004. Loss of negative regulation by Numb over Notch is relevant to human breast carcinogenesis. J Cell Biol 167:215-221.

Pefanis, E., J. Wang, G. Rothschild, J. Lim, J. Chao, R. Rabadan, A.N. Economides, and U. Basu. 2014. Noncoding RNA transcription targets AID to divergently transcribed loci in B cells. Nature 514:389-393.

Pei, X.H., F. Bai, M.D. Smith, J. Usary, C. Fan, S.Y. Pai, I.C. Ho, C.M. Perou, and Y. Xiong. 2009. CDK inhibitor p18(INK4c) is a downstream target of GATA3 and restrains

184

mammary luminal progenitor cell proliferation and tumorigenesis. Cancer Cell 15:389- 401.

Perou, C.M., T. Sorlie, M.B. Eisen, M. van de Rijn, S.S. Jeffrey, C.A. Rees, J.R. Pollack, D.T. Ross, H. Johnsen, L.A. Akslen, O. Fluge, A. Pergamenschikov, C. Williams, S.X. Zhu, P.E. Lonning, A.L. Borresen-Dale, P.O. Brown, and D. Botstein. 2000. Molecular portraits of human breast tumours. Nature 406:747-752.

Pharoah, P.D., N.E. Day, and C. Caldas. 1999. Somatic mutations in the p53 gene and prognosis in breast cancer: a meta-analysis. Br J Cancer 80:1968-1973.

Pinato, S., C. Scandiuzzi, N. Arnaudo, E. Citterio, G. Gaudino, and L. Penengo. 2009. RNF168, a new RING finger, MIU-containing protein that modifies chromatin by ubiquitination of histones H2A and H2AX. BMC Mol Biol 10:55.

Pittman, D.L., J. Cobb, K.J. Schimenti, L.A. Wilson, D.M. Cooper, E. Brignull, M.A. Handel, and J.C. Schimenti. 1998. Meiotic prophase arrest with failure of chromosome synapsis in mice deficient for Dmc1, a germline-specific RecA homolog. Mol Cell 1:697-705.

Plug, A.W., J.H. Xu, G. Reddy, E.I. Golub, and T. Ashley. 1996. Presynaptic association of Rad51 protein with selected sites in meiotic chromatin. Proceedings of the National Academy of Sciences of the United States of America 93:5920-5924.

Polo, S.E., and S.P. Jackson. 2011a. Dynamics of DNA damage response proteins at DNA breaks: a focus on protein modifications. Genes & Development 25:409-433.

Polo, S.E., and S.P. Jackson. 2011b. Dynamics of DNA damage response proteins at DNA breaks: a focus on protein modifications. Genes Dev 25:409-433.

Polyak, K. 2011. Heterogeneity in breast cancer. J Clin Invest 121:3786-3788.

Poulsen, M., C. Lukas, J. Lukas, S. Bekker-Jensen, and N. Mailand. 2012. Human RNF169 is a negative regulator of the ubiquitin-dependent response to DNA double-strand breaks. Journal of Cell Biology 197:189-199.

Price, B.D., and A.D. D'Andrea. 2013. Chromatin Remodeling at DNA Double-Strand Breaks. Cell 152:1344-1354.

Proia, T.A., P.J. Keller, P.B. Gupta, I. Klebba, A.D. Jones, M. Sedic, H. Gilmore, N. Tung, S.P. Naber, S. Schnitt, E.S. Lander, and C. Kuperwasser. 2011. Genetic predisposition directs breast cancer phenotype by dictating progenitor cell fate. Cell Stem Cell 8:149-163.

Raafat, A., S. Bargo, M.R. Anver, and R. Callahan. 2004. Mammary development and tumorigenesis in mice expressing a truncated human Notch4/Int3 intracellular domain (h- Int3sh). Oncogene 23:9401-9407.

Rahman, N., S. Seal, D. Thompson, P. Kelly, A. Renwick, A. Elliott, S. Reid, K. Spanova, R. Barfoot, T. Chagtai, H. Jayatilake, L. McGuffog, S. Hanks, D.G. Evans, D. Eccles, D.F.

185

Easton, and M.R. Stratton. 2007. PALB2, which encodes a BRCA2-interacting protein, is a breast cancer susceptibility gene. Nat Genet 39:165-167.

Rand, M.D., L.M. Grimm, S. Artavanis-Tsakonas, V. Patriub, S.C. Blacklow, J. Sklar, and J.C. Aster. 2000. Calcium depletion dissociates and activates heterodimeric notch receptors. Molecular and Cellular Biology 20:1825-1835.

Reedijk, M., B. Dickson, D. Pinnaduwage, A.M. Mulligan, H. Zhang, S.B. Bull, F. O'Malley, S.E. Egan, and I.L. Andrulis. 2007. Elevated JAG1 mRNA expression, associated with the basal phenotype, is a poor-prognosis indicator in lymph node-negative breast cancer. Breast Cancer Research and Treatment 106:S55-S55.

Reedijk, M., S. Odorcic, L. Chang, H. Zhang, N. Miller, D.R. McCready, G. Lockwood, and S.E. Egan. 2005. High-level coexpression of JAG1 and NOTCH1 is observed in human breast cancer and is associated with poor overall survival. Cancer Res 65:8530-8537.

Reedijk, M., D. Pinnaduwage, B.C. Dickson, A.M. Mulligan, H. Zhang, S.B. Bull, F.P. O'Malley, S.E. Egan, and I.L. Andrulis. 2008. JAG1 expression is associated with a basal phenotype and recurrence in lymph node-negative breast cancer. Breast Cancer Research and Treatment 111:439-448.

Rehman, F.L., C.J. Lord, and A. Ashworth. 2010. Synthetic lethal approaches to breast cancer therapy. Nat Rev Clin Oncol 7:718-724.

Rizzo, P., H. Miao, G. D'Souza, C. Osipo, L.L. Song, J. Yun, H. Zhao, J. Mascarenhas, D. Wyatt, G. Antico, L. Hao, K. Yao, P. Rajan, C. Hicks, K. Siziopikou, S. Selvaggi, A. Bashir, D. Bhandari, A. Marchese, U. Lendahl, J.Z. Qin, D.A. Tonetti, K. Albain, B.J. Nickoloff, and L. Miele. 2008a. Cross-talk between notch and the estrogen receptor in breast cancer suggests novel therapeutic approaches. Cancer Res 68:5226-5235.

Rizzo, P., H. Miao, G. D'Souza, C. Osipo, J. Yun, H.P. Zhao, J. Mascarenhas, D. Wyatt, G. Antico, L. Hao, K. Yao, P. Rajan, C. Hicks, K. Siziopikou, S. Selvaggi, A. Bashir, D. Bhandari, A. Marchese, U. Lendahl, J.Z. Qin, D.A. Tonetti, K. Albain, B.J. Nickoloff, and L. Miele. 2008b. Cross-talk between Notch and the estrogen receptor in breast cancer suggests novel therapeutic approaches. Cancer Research 68:5226-5235.

Robbins, J., B.J. Blondel, D. Gallahan, and R. Callahan. 1992. Mouse mammary tumor gene int- 3: a member of the notch gene family transforms mammary epithelial cells. J Virol 66:2594-2599.

Robert, I., F. Dantzer, and B. Reina-San-Martin. 2009. Parp1 facilitates alternative NHEJ, whereas Parp2 suppresses IgH/c-myc translocations during immunoglobulin class switch recombination. Journal of Experimental Medicine 206:1047-1056.

Robinson, D.R., S. Kalyana-Sundaram, Y.M. Wu, S. Shankar, X.H. Cao, B. Ateeq, I.A. Asangani, M. Iyer, C.A. Maher, C.S. Grasso, R.J. Lonigro, M. Quist, J. Siddiqui, R. Mehra, X.J. Jing, T.J. Giordano, M.S. Sabel, C.G. Kleer, N. Palanisamy, R. Natrajan, M.B. Lambros, J.S. Reis-Filho, C. Kumar-Sinha, and A.M. Chinnaiyan. 2011.

186

Functionally recurrent rearrangements of the MAST kinase and Notch gene families in breast cancer. Nature Medicine 17:1646-U1163.

Robinson, G.W., R.A. McKnight, G.H. Smith, and L. Hennighausen. 1995. Mammary epithelial cells undergo secretory differentiation in cycling virgins but require pregnancy for the establishment of terminal differentiation. Development 121:2079-2090.

Rogakou, E.P., D.R. Pilch, A.H. Orr, V.S. Ivanova, and W.M. Bonner. 1998. DNA double- stranded breaks induce histone H2AX phosphorylation on serine 139. J Biol Chem 273:5858-5868.

Santos, M.A., M.S. Huen, M. Jankovic, H.T. Chen, A.J. Lopez-Contreras, I.A. Klein, N. Wong, J.L. Barbancho, O. Fernandez-Capetillo, M.C. Nussenzweig, J. Chen, and A. Nussenzweig. 2010. Class switching and meiotic defects in mice lacking the E3 ubiquitin ligase RNF8. J Exp Med 207:973-981.

Savic, V., B. Yin, N.L. Maas, A.L. Bredemeyer, A.C. Carpenter, B.A. Helmink, K.S. Yang-Iott, B.P. Sleckman, and C.H. Bassing. 2009. Formation of dynamic gamma-H2AX domains along broken DNA strands is distinctly regulated by ATM and MDC1 and dependent upon H2AX densities in chromatin. Mol Cell 34:298-310.

Schalk, J.A., A.J. Dietrich, A.C. Vink, H.H. Offenberg, M. van Aalderen, and C. Heyting. 1998. Localization of SCP2 and SCP3 protein molecules within synaptonemal complexes of the rat. Chromosoma 107:540-548.

Shackleton, M., F. Vaillant, K.J. Simpson, J. Stingl, G.K. Smyth, M.L. Asselin-Labat, L. Wu, G.J. Lindeman, and J.E. Visvader. 2006. Generation of a functional mammary gland from a single stem cell. Nature 439:84-88.

Sharma, N., Q. Zhu, G. Wani, J. He, Q.E. Wang, and A.A. Wani. 2014. USP3 counteracts RNF168 via deubiquitinating H2A and gammaH2AX at lysine 13 and 15. Cell Cycle 13:106-114.

Shi, W., Z.F. Ma, H. Willers, K. Akhtar, S.P. Scott, J.Q. Zhang, S. Powell, and J.R. Zhang. 2008. Disassembly of MDC1 Foci Is Controlled by Ubiquitin-Proteasome-dependent Degradation. Journal of Biological Chemistry 283:31608-31616.

Shim, E.Y., W.H. Chung, M.L. Nicolette, Y. Zhang, M. Davis, Z. Zhu, T.T. Paull, G. Ira, and S.E. Lee. 2010. Saccharomyces cerevisiae Mre11/Rad50/Xrs2 and Ku proteins regulate association of Exo1 and Dna2 with DNA breaks. EMBO J 29:3370-3380.

Smeenk, G., and H. van Attikum. 2013. The Chromatin Response to DNA Breaks: Leaving a Mark on Genome Integrity. Annual Review of Biochemistry, Vol 82 82:55-80.

Smith, G.H., D. Gallahan, F. Diella, C. Jhappan, G. Merlino, and R. Callahan. 1995. Constitutive expression of a truncated INT3 gene in mouse mammary epithelium impairs differentiation and functional development. Cell Growth Differ 6:563-577.

187

Smith, P.D., S. Crossland, G. Parker, P. Osin, L. Brooks, J. Waller, E. Philp, M.R. Crompton, B.A. Gusterson, M.J. Allday, and T. Crook. 1999. Novel p53 mutants selected in BRCA- associated tumours which dissociate transformation suppression from other wild-type p53 functions. Oncogene 18:2451-2459.

Smith, T.R., E.A. Levine, N.D. Perrier, M.S. Miller, R.I. Freimanis, K. Lohman, L.D. Case, J. Xu, H.W. Mohrenweiser, and J.J. Hu. 2003. DNA-repair genetic polymorphisms and breast cancer risk. Cancer Epidemiol Biomarkers Prev 12:1200-1204.

Solovjeva, L.V., M.P. Svetlova, V.O. Chagin, and N.V. Tomilin. 2007. Inhibition of transcription at radiation-induced nuclear foci of phosphorylated histone H2AX in mammalian cells. Chromosome Research 15:787-797.

Sorlie, T., C.M. Perou, R. Tibshirani, T. Aas, S. Geisler, H. Johnsen, T. Hastie, M.B. Eisen, M. van de Rijn, S.S. Jeffrey, T. Thorsen, H. Quist, J.C. Matese, P.O. Brown, D. Botstein, P.E. Lonning, and A.L. Borresen-Dale. 2001. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci U S A 98:10869-10874.

Sorlie, T., R. Tibshirani, J. Parker, T. Hastie, J.S. Marron, A. Nobel, S. Deng, H. Johnsen, R. Pesich, S. Geisler, J. Demeter, C.M. Perou, P.E. Lonning, P.O. Brown, A.L. Borresen- Dale, and D. Botstein. 2003. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci U S A 100:8418-8423.

Sotiriou, C., S.Y. Neo, L.M. McShane, E.L. Korn, P.M. Long, A. Jazaeri, P. Martiat, S.B. Fox, A.L. Harris, and E.T. Liu. 2003. Breast cancer classification and prognosis based on gene expression profiles from a population-based study. Proc Natl Acad Sci U S A 100:10393- 10398.

Soulas-Sprauel, P., P. Rivera-Munoz, L. Malivert, G. Le Guyader, V. Abramowski, P. Revy, and J.P. de Villartay. 2007. V(D)J and immunoglobulin class switch recombinations: a paradigm to study the regulation of DNA end-joining. Oncogene 26:7780-7791.

Stavnezer, J., and C.E. Schrader. 2014. IgH Chain Class Switch Recombination: Mechanism and Regulation. Journal of Immunology 193:5370-5378.

Stewart, G.S., S. Panier, K. Townsend, A.K. Al-Hakim, N.K. Kolas, E.S. Miller, S. Nakada, J. Ylanko, S. Olivarius, M. Mendez, C. Oldreive, J. Wildenhain, A. Tagliaferro, L. Pelletier, N. Taubenheim, A. Durandy, P.J. Byrd, T. Stankovic, A.M. Taylor, and D. Durocher. 2009. The RIDDLE syndrome protein mediates a ubiquitin-dependent signaling cascade at sites of DNA damage. Cell 136:420-434.

Stewart, G.S., T. Stankovic, P.J. Byrd, T. Wechsler, E.S. Miller, A. Huissoon, M.T. Drayson, S.C. West, S.J. Elledge, and A.M. Taylor. 2007. RIDDLE immunodeficiency syndrome is linked to defects in 53BP1-mediated DNA damage signaling. Proc Natl Acad Sci U S A 104:16910-16915.

Stracker, T.H., and J.H. Petrini. 2011. The MRE11 complex: starting from the ends. Nat Rev Mol Cell Biol 12:90-103.

188

Struhl, G., and I. Greenwald. 1999. Presenilin is required for activity and nuclear access of Notch in Drosophila. Nature 398:522-525.

Stucki, M., J.A. Clapperton, D. Mohammad, M.B. Yaffe, S.J. Smerdon, and S.P. Jackson. 2005. MDC1 directly binds phosphorylated histone H2AX to regulate cellular responses to DNA double-strand breaks. Cell 123:1213-1226.

Stylianou, S., R.B. Clarke, and K. Brennan. 2006. Aberrant activation of notch signaling in human breast cancer. Cancer Res 66:1517-1525.

Su, T.T. 2006. Cellular responses to DNA damage: one signal, multiple choices. Annu Rev Genet 40:187-208.

Sun, Y., X. Jiang, and B.D. Price. 2010. Tip60: connecting chromatin to DNA damage signaling. Cell Cycle 9:930-936.

Sutherland, B.M., P.V. Bennett, O. Sidorkina, and J. Laval. 2000. Clustered damages and total lesions induced in DNA by ionizing radiation: oxidized bases and strand breaks. Biochemistry 39:8026-8031.

Sutherland, B.M., P.V. Bennett, J.C. Sutherland, and J. Laval. 2002. Clustered DNA damages induced by x rays in human cells. Radiat Res 157:611-616.

Swanson, P.C. 2004. The bounty of RAGs: recombination signal complexes and reaction outcomes. Immunol Rev 200:90-114.

Symington, L.S., and J. Gautier. 2011. Double-strand break end resection and repair pathway choice. Annu Rev Genet 45:247-271.

Tarsounas, M., T. Morita, R.E. Pearlman, and P.B. Moens. 1999. RAD51 and DMC1 form mixed complexes associated with mouse meiotic chromosome cores and synaptonemal complexes. J Cell Biol 147:207-220.

Tauchi, H., S. Matsuura, J. Kobayashi, S. Sakamoto, and K. Komatsu. 2002. Nijmegen breakage syndrome gene, NBS1, and molecular links to factors for genome stability. Oncogene 21:8967-8980.

Tolcher, A.W., W.A. Messersmith, S.M. Mikulski, K.P. Papadopoulos, E.L. Kwak, D.G. Gibbon, A. Patnaik, G.S. Falchook, A. Dasari, G.I. Shapiro, J.F. Boylan, Z.X. Xu, K. Wang, A. Koehler, J. Song, S.A. Middleton, J. Deutsch, M. Demario, R. Kurzrock, and J.J. Wheler. 2012. Phase I study of RO4929097, a gamma secretase inhibitor of Notch signaling, in patients with refractory metastatic or locally advanced solid tumors. J Clin Oncol 30:2348-2353.

Trichopoulos, D., B. MacMahon, and P. Cole. 1972. Menopause and breast cancer risk. J Natl Cancer Inst 48:605-613.

Turnbull, C., and N. Rahman. 2008. Genetic predisposition to breast cancer: past, present, and future. Annu Rev Genomics Hum Genet 9:321-345.

189

Turner, J.M., O. Aprelikova, X. Xu, R. Wang, S. Kim, G.V. Chandramouli, J.C. Barrett, P.S. Burgoyne, and C.X. Deng. 2004. BRCA1, histone H2AX phosphorylation, and male meiotic sex chromosome inactivation. Current Biology 14:2135-2142.

Vanderlugt, N.M.T., J. Domen, K. Linders, M. Vanroon, E. Robanusmaandag, H. Teriele, M. Vandervalk, J. Deschamps, M. Sofroniew, M. Vanlohuizen, and A. Berns. 1994. Posterior Transformation, Neurological Abnormalities, and Severe Hematopoietic Defects in Mice with a Targeted Deletion of the Bmi-1 Protooncogene. Genes & Development 8:757-769.

Ventura, A., D.G. Kirsch, M.E. McLaughlin, D.A. Tuveson, J. Grimm, L. Lintault, J. Newman, E.E. Reczek, R. Weissleder, and T. Jacks. 2007. Restoration of p53 function leads to tumour regression in vivo. Nature 445:661-665.

Visvader, J.E. 2009. Keeping abreast of the mammary epithelial hierarchy and breast tumorigenesis. Genes Dev 23:2563-2577.

Visvader, J.E. 2011. Cells of origin in cancer. Nature 469:314-322.

Visvader, J.E., and J. Stingl. 2014. Mammary stem cells and the differentiation hierarchy: current status and perspectives. Genes Dev 28:1143-1158.

Vuong, B.Q., M. Lee, S. Kabir, C. Irimia, S. Macchiarulo, G.S. McKnight, and J. Chaudhuri. 2009. Specific recruitment of protein kinase A to the immunoglobulin locus regulates class-switch recombination. Nat Immunol 10:420-426.

Vyas, R., R. Kumar, F. Clermont, A. Helfricht, P. Kalev, P. Sotiropoulou, I.A. Hendriks, E. Radaelli, T. Hochepied, C. Blanpain, A. Sablina, H. van Attikum, J.V. Olsen, A.G. Jochemsen, A.C.O. Vertegaal, and J.C. Marine. 2013. RNF4 is required for DNA double- strand break repair in vivo. Cell Death and Differentiation 20:490-502.

Wagner, K.U., C.A. Boulanger, M.D. Henry, M. Sgagias, L. Hennighausen, and G.H. Smith. 2002. An adjunct mammary epithelial cell population in parous females: its role in functional adaptation and tissue renewal. Development 129:1377-1386.

Wang, B., and S.J. Elledge. 2007. Ubc13/Rnf8 ubiquitin ligases control foci formation of the Rap80/Abraxas/Brca1/Brcc36 complex in response to DNA damage. Proc Natl Acad Sci U S A 104:20759-20763.

Ward, I.M., K. Minn, J. van Deursen, and J. Chen. 2003. p53 Binding protein 53BP1 is required for DNA damage responses and tumor suppression in mice. Mol Cell Biol 23:2556-2563.

Ward, I.M., B. Reina-San-Martin, A. Olaru, K. Minn, K. Tamada, J.S. Lau, M. Cascalho, L.P. Chen, A. Nussenzweig, F. Livak, M.C. Nussenzweig, and J.J. Chen. 2004. 53BP1 is required for class switch recombination. Journal of Cell Biology 165:459-464.

Watanabe, Y., I. Maeda, R. Oikawa, W. Wu, K. Tsuchiya, Y. Miyoshi, F. Itoh, K. Tsugawa, and T. Ohta. 2013. Aberrant DNA methylation status of DNA repair genes in breast cancer treated with neoadjuvant chemotherapy. Genes Cells 18:1120-1130.

190

Wiener, R., X. Zhang, T. Wang, and C. Wolberger. 2012. The mechanism of OTUB1-mediated inhibition of ubiquitination. Nature 483:618-622.

Wojtasz, L., J.M. Cloutier, M. Baumann, K. Daniel, J. Varga, J. Fu, K. Anastassiadis, A.F. Stewart, A. Remenyi, J.M. Turner, and A. Toth. 2012. Meiotic DNA double-strand breaks and chromosome asynapsis in mice are monitored by distinct HORMAD2- independent and -dependent mechanisms. Genes Dev 26:958-973.

Wojtasz, L., K. Daniel, I. Roig, E. Bolcun-Filas, H. Xu, V. Boonsanay, C.R. Eckmann, H.J. Cooke, M. Jasin, S. Keeney, M.J. McKay, and A. Toth. 2009. Mouse HORMAD1 and HORMAD2, two conserved meiotic chromosomal proteins, are depleted from synapsed chromosome axes with the help of TRIP13 AAA-ATPase. PLoS Genet 5:e1000702.

Wood, J.L., and J. Chen. 2008. DNA-damage checkpoints: location, location, location. Trends Cell Biol 18:451-455.

Wood, L.D., D.W. Parsons, S. Jones, J. Lin, T. Sjoblom, R.J. Leary, D. Shen, S.M. Boca, T. Barber, J. Ptak, N. Silliman, S. Szabo, Z. Dezso, V. Ustyanksky, T. Nikolskaya, Y. Nikolsky, R. Karchin, P.A. Wilson, J.S. Kaminker, Z. Zhang, R. Croshaw, J. Willis, D. Dawson, M. Shipitsin, J.K. Willson, S. Sukumar, K. Polyak, B.H. Park, C.L. Pethiyagoda, P.V. Pant, D.G. Ballinger, A.B. Sparks, J. Hartigan, D.R. Smith, E. Suh, N. Papadopoulos, P. Buckhaults, S.D. Markowitz, G. Parmigiani, K.W. Kinzler, V.E. Velculescu, and B. Vogelstein. 2007. The genomic landscapes of human breast and colorectal cancers. Science 318:1108-1113.

Wu, G., G. Wilson, J. George, and L. Qiao. 2014. Modulation of Notch signaling as a therapeutic approach for liver cancer. Curr Gene Ther

Wu, J., M.S. Huen, L.Y. Lu, L. Ye, Y. Dou, M. Ljungman, J. Chen, and X. Yu. 2009. Histone ubiquitination associates with BRCA1-dependent DNA damage response. Mol Cell Biol 29:849-860.

Xu, X., K.U. Wagner, D. Larson, Z. Weaver, C. Li, T. Ried, L. Hennighausen, A. Wynshaw- Boris, and C.X. Deng. 1999. Conditional mutation of Brca1 in mammary epithelial cells results in blunted ductal morphogenesis and tumour formation. Nat Genet 22:37-43.

Xu, Y., Y. Sun, X. Jiang, M.K. Ayrapetov, P. Moskwa, S. Yang, D.M. Weinstock, and B.D. Price. 2010. The p400 ATPase regulates nucleosome stability and chromatin ubiquitination during DNA repair. J Cell Biol 191:31-43.

Xu, Z., H. Zan, E.J. Pone, T. Mai, and P. Casali. 2012. Immunoglobulin class-switch DNA recombination: induction, targeting and beyond. Nat Rev Immunol 12:517-531.

Xue, W., L. Zender, C. Miething, R.A. Dickins, E. Hernando, V. Krizhanovsky, C. Cordon- Cardo, and S.W. Lowe. 2007. Senescence and tumour clearance is triggered by p53 restoration in murine liver carcinomas. Nature 445:656-660.

191

Yan, C.T., C. Boboila, E.K. Souza, S. Franco, T.R. Hickernell, M. Murphy, S. Gumaste, M. Geyer, A.A. Zarrin, J.P. Manis, K. Rajewsky, and F.W. Alt. 2007. IgH class switching and translocations use a robust non-classical end-joining pathway. Nature 449:478-482.

Yang, F., and P.J. Wang. 2009. The Mammalian synaptonemal complex: a scaffold and beyond. Genome Dyn 5:69-80.

Yao, K., P. Rizzo, P. Rajan, K. Albain, K. Rychlik, S. Shah, and L. Miele. 2011a. Notch-1 and notch-4 receptors as prognostic markers in breast cancer. Int J Surg Pathol 19:607-613.

Yao, K., P. Rizzo, P. Rajan, K. Albain, K. Rychlik, S. Shah, and L. Miele. 2011b. Notch-1 and Notch-4 Receptors as Prognostic Markers in Breast Cancer. International Journal of Surgical Pathology 19:607-613.

Yin, Y.L., A. Seifert, J.S. Chua, J.F. Maure, F. Golebiowski, and R.T. Hay. 2012. SUMO- targeted ubiquitin E3 ligase RNF4 is required for the response of human cells to DNA damage. Genes & Development 26:1196-1208.

Yun, M.H., and K. Hiom. 2009. CtIP-BRCA1 modulates the choice of DNA double-strand-break repair pathway throughout the cell cycle. Nature 459:460-463.

Zardawi, S.J., I. Zardawi, C.M. McNeil, E.K. Millar, D. McLeod, A.L. Morey, P. Crea, N.C. Murphy, M. Pinese, E. Lopez-Knowles, S.R. Oakes, C.J. Ormandy, M.R. Qiu, A. Hamilton, A. Spillane, C. Soon Lee, R.L. Sutherland, E.A. Musgrove, and S.A. O'Toole. 2010. High Notch1 protein expression is an early event in breast cancer development and is associated with the HER-2 molecular subtype. Histopathology 56:286-296.

Zgheib, O., K. Pataky, J. Brugger, and T.D. Halazonetis. 2009. An Oligomerized 53BP1 Tudor Domain Suffices for Recognition of DNA Double-Strand Breaks. Molecular and Cellular Biology 29:1050-1058.

Zhang, D., K. Zaugg, T.W. Mak, and S.J. Elledge. 2006. A role for the deubiquitinating enzyme USP28 in control of the DNA-damage response. Cell 126:529-542.

Zhao, H., M. Zhu, G. Dou, B. Zhu, J. Li, J. Liao, and X. Xu. 2014. BCL10 regulates RNF8/RNF168-mediated ubiquitination in the DNA damage response. Cell Cycle 13:1777-1787.

Zilfou, J.T., and S.W. Lowe. 2009. Tumor suppressive functions of p53. Cold Spring Harb Perspect Biol 1:a001883.

192