Identification of Novel Genes in BRCA1-Regulated Pathways

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Shweta Kotian, M.S.

Graduate Program in Molecular, Cellular and Developmental Biology

The Ohio State University

2013

Dissertation Committee:

Professor Jeffrey D. Parvin, Advisor

Professor Joanna L. Groden

Professor Denis C. Guttridge

Professor Mark R. Parthun

Shweta Kotian

2013

Abstract

BRCA1 is an important breast- and ovarian-specific tumor suppressor gene. It is important in various cellular functions in the body including transcriptional regulation, cell cycle checkpoint activation, DNA damage response, and maintenance of genomic stability. Approximately 40% of hereditary breast cancers have mutations in BRCA1 or BRCA2, and it is unclear how the remaining cases are caused, presumably by mutations or in the alteration of expression of unknown genes.

We hypothesize that these unknown genes or ‘missing BRCAs’ can be identified by bio-informatics methods. Performing gene co-expression analysis, we have compared the expression profiles of hundreds of genes across publicly available breast cancer microarray datasets, with those of BRCA1 and BRCA2. We propose that the genes, whose expression is highly correlated with BRCA1 and

BRCA2, might function together in the same pathways. They could also potentially interact with each other.

The main aim of this study is to identify these genes, and then biologically validate them in functional BRCA1-regulated pathways including homologous recombination and centrosome duplication. We also hope to find any mechanistic

ii associations between these genes and BRCA1. Eventually, we hope to evaluate if these genes contribute to the process of carcinogenesis.

As a result of the informatics data mining, we narrowed down the list of genes co- expressed with BRCA1 to a few dozen. Of these genes, we decided to choose the histone deacetylases (HDACs) and NUSAP1 for biological validation. We had several criteria for choosing these, such as 1) the presence of enzymatic domains in the case of the HDAC proteins, 2) an unknown association with

BRCA1, 3) a paucity of the literature about this gene/protein; and 4) cancer- associated changes in gene expression.

Biological validation was performed using two main assays: homology-directed repair and centrosome duplication. BRCA1 is important for proper DNA repair, as well as regulation of centrosome numbers. When BRCA1 is depleted from cells in these assays using RNA interference, homologous recombination is impaired, and there is aberrant centrosome duplication causing formation of supernumerary centrosomes. Likewise, we were interested in determining if the

HDACs or NUSAP1 would be important for these processes as well.

The HDAC family catalyzes the deacetylation of histones on chromatin causing chromatin compaction. This is turn, suppresses transcription. In this way, HDACs help regulate the expression of key developmental, cell cycle, and transcription factors. We observed that multiple members of the HDAC family of genes were co-expressed with BRCA1 and BRCA2 in multiple gene expression datasets.

Dysregulation of HDACs has been observed in several cancers, and there is

iii much interest in the development of HDAC inhibitor drugs. These drugs inhibit the growth of cancer cells, but it is not known how. Thus, we decided to test

HDAC function in homologous recombination using these inhibitors. A tissue- culture based homology directed repair (HDR) assay was used in which repair of a double-stranded break, by homologous recombination, results in gene conversion of an inactive GFP allele to an active GFP gene. Homologous recombination events were then readily scored by measuring GFP positive cells by FACS. We found that treatment with HDAC inhibitors significantly reduced homologous recombination levels. Upon carrying out an individual depletion of

HDACs 1 to 11 by RNA interference, we found that depletions of HDAC9 and

HDAC10 specifically reduce homologous recombination levels. This was corroborated as HDAC9 and HDAC10 depletion sensitized cells to the interstrand crosslinker mitomycin C, indicating faulty homologous recombination. We think that HDAC9 and HDAC10 catalyze a crucial deacetylation step at the site of the double strand break that is important for homologous recombination.

NUSAP1 is important for proper mitotic spindle assembly and cytokinesis. It is also aberrantly expressed in several cancers. We performed NUSAP1 depletion in the HDR assay, and saw that homologous recombination was impaired. In addition, depletion of NUSAP1 in the centrosome assay, caused centrosome amplification. Over-expression of BRCA1 reversed the defective phenotypes seen in both these assays upon NUSAP1 depletion. While NUSAP1 depletion reduced BRCA1 protein levels in the soluble extracts, and specifically in the

iv chromatin fractions, there was no change in total BRCA1 protein levels. This indicated that NUSAP1 controls BRCA1 localization. We further confirmed this by determining that NUSAP1 depletion inhibited recruitment of BRCA1 to DNA damage-induced foci upon ionizing irradiation. In addition to BRCA1, RAD51 recruitment to the DNA damage foci was affected. Thus, NUSAP1 impacts the

DNA damage response through control of BRCA1 localization.

Thus, we have successfully identified and validated three novel genes- HDAC9,

HDAC10, NUSAP1 as important in the BRCA1-regulated pathway of homologous recombination. NUSAP1 is also important for centrosome amplification. Since, these genes are implicated in cancer; we hope that our study will help shed light on exactly how they are important for tumorigenesis. Further work will determine whether these proteins may be useful as biomarkers for breast cancer.

Dedication

I dedicate this document to my grandparents, the late Mr. Rajeeva Suvarna and the late Mrs. Vimala Suvarna. Words cannot describe how truly grateful I am for

their loving care and support in my crucial, formative years.

Acknowledgments

I would like to offer my humble gratitude to my advisor Dr. Jeffrey D. Parvin. Not only is he an immense source of knowledge, but he also has immense patience. I am thankful to him for his invaluable guidance, but most of all for not giving up on me during a very difficult phase in my life.

Thanks are due to the members of my lab who have always offered help willingly when needed. I would also like to make a special mention of my colleague, Ms.

Mansi Arora, for offering both professional and personal support.

I would like to sincerely thank my committee members, Dr. Joanna L. Groden,

Dr. Mark R. Parthun, and Dr. Denis C. Guttridge, for offering their valuable time and expertise to help guide me with my dissertation.

I would also like to thank my program chair, Dr. David Bisaro, for always being willing to lend a listening ear to even a graduate student’s petty problems, and offer practical advice.

This dissertation would not be complete without acknowledging some key people in my life. I will always be grateful to my best friend and partner, Mr. Daniel

Stanojevic, who has provided tremendous emotional support to me during this journey. Last, but not least, I am eternally indebted to my mother, Ms. Shashikala

vii

Kotian, whose fortitude and personal sacrifices, have made it possible for me to reach this personal milestone in my life.

viii

Vita

2004 ...... B.S. Biochemistry, University of Mumbai

2006 ...... M.S. Biochemistry, University of Mumbai

2006 to present ...... Graduate Research Associate,

Department of Biomedical Informatics,

The Ohio State University

Publications

Kotian S, Parvin JD. NUSAP1 Influences the DNA damage response by controlling BRCA1 localization. Manuscript in preparation

Zhang J, Lu K, Xiang Y, Islam M, Kotian S, Kais Z, Lee C, Arora A, Liu H, Parvin

JD, Huang K (2012). Weighted frequent gene co-expression network mining to identify genes involved in genome stability. PLoS Comput Biol. 8(8):e1002656.

Epub 2012 Aug 30.

Kotian S, Liyanarachchi S, Zelent A, Parvin JD (2011). Histone deacetylases 9 and 10 are required for homologous recombination. J Biol Chem, 286(10):7722-6

Fields of Study

Major Field: Molecular, Cellular and Developmental Biology

Table of Contents

Abstract ...... ii

Dedication ...... vi

Acknowledgments ...... vii

Vita ...... ix

Publications ...... ix

Fields of Study ...... x

Table of Contents ...... xi

List of Tables ...... xvi

List of Figures ...... xvii

Chapter 1: Introduction ...... 1

1.1 BRCA1 – General Biology and Function: ...... 1

1.2 The role of BRCA1 in DNA Damage Response (DDR): ...... 3

1.3 BRCA1 and Centrosome Regulation: ...... 7

1.4 BRCA1 and Cancer Association:...... 9

1.5 Co-expression Profiling to Identify Missing ‘BRCAs’: ...... 9

1.6 Histone Deacetylases (HDACs): ...... 11

1.7 NUSAP1: ...... 13

Chapter 2: Thesis Rationale ...... 23

Chapter 3: Co-Expression Profiling of Breast Cancer Microarrays to Identify

Novel ‘BRCA’ Genes ...... 28

3.1 Introduction ...... 28

3.2 Methods: ...... 30

3.2.1 Downloading Microarray Datasets: ...... 30

3.2.3 Intersect Analysis: ...... 32

3.2.4 Criteria for Selection of Gene Candidates: ...... 33

3.3 Results: ...... 33

3.4 Discussion: ...... 35

Chapter 4: Histone Deacetylases 9 and 10 Are Required For Homologous

Recombination ...... 42

4.1 Introduction: ...... 42

4.2 Methods and Materials: ...... 44

4.2.1 Cell Culture, Plasmids, Antibodies and Reagents: ...... 44

4.2.2 Homology Directed Repair Assay: ...... 44

4.2.3 Western Analysis: ...... 45

xii

4.2.4 Mitomycin C Assay: ...... 45

4.2.5 Cell Cycle Assay: ...... 46

4.2.6 Statistical Analyses: ...... 46

4.3 Results: ...... 46

4.3.1 Treatment with HDACi Inhibits Homologous Recombination: ...... 47

4.3.2 Depletion of either HDAC9 or HDAC10 Inhibits Homologous

Recombination: ...... 48

4.3.3 Both HDACi TreATMent and HDAC9 or HDAC10 Depletion Sensitize

Cells to Mitomycin C: ...... 50

4.3.4 Effect of HDAC Inhibition on Homologous Recombination is not due to

Cell Cycle Perturbation: ...... 52

4.4 Discussion: ...... 53

Chapter 5: NUSAP1 Influences the DNA Damage Response by Controlling

BRCA1 Localization to DNA Damage Foci ...... 73

5.1 Introduction: ...... 73

5.2 Materials and Methods: ...... 77

5.2.1 Cell culture and reagents: ...... 77

5.2.2 Transfection: ...... 77

5.2.3 Western Analysis: ...... 78

5.2.4 Homology-directed Repair Assay: ...... 78 xiii

5.2.5 Single Strand Annealing Assay: ...... 79

5.2.6 Centrosome Duplication Assay: ...... 79

5.2.7 Total RNA Extraction and Reverse Transcription: ...... 80

5.2.8 Quantitative Real-time PCR (Q-PCR) Analysis: ...... 80

5.2.10 Chromatin Fractionation: ...... 81

5.2.11 Immunocytochemistry: ...... 82

5.2.12 Statistical Analyses: ...... 82

5.3 Results: ...... 82

5.3.1 Depletion of NUSAP1 Suppresses Homologous Recombination: ...... 83

5.3.2 NUSAP1 depletion Also Impairs Single Strand Annealing: ...... 84

5.3.3 NUSAP1 depletion Leads to Formation of Supernumerary

Centrosomes: ...... 85

5.3.4 NUSAP1 and BRCA1 are Expressed at Similar Cell Cycle Stages ..... 85

5.3.5 NUSAP1 concentration increased following DNA damage: ...... 87

5.3.6 BRCA1 over-expression suppresses defects caused by NUSAP1

depletion in homologous recombination and centrosome duplication: ...... 87

5.3.7 NUSAP1 depletion does not affect BRCA1 mRNA levels, transcription,

or protein stability: ...... 89

5.3.8 NUSAP1 depletion affects BRCA1 localization: ...... 90

xiv

5.3.9 NUSAP1 depletion affects recruitment of BRCA1 to DNA damage foci:

...... 92

5.3.10 NUSAP1 and BRCA1 do not physically Interact: ...... 93

5.4 Discussion: ...... 94

Chapter 6: Discussion ...... 133

Bibliography ...... 148

Appendix: Bioinformatic Data Analysis ...... 164

List of Tables

Table 1: List of Geo-datasets Downloaded from the GEO database...... 38

Table 2: List of Interesting Candidate Genes ...... 40

Table 3: HDAC siRNA sequences ...... 72

Table 4 NUSAP1 siRNA sequences...... 131

Table 5 RT-PCR primer sequences ...... 132

xvi

List of Figures

Figure 1: BRCA1 Structure...... 14

Figure 2: BRCA1 Cellular Functions...... 15

Figure 3: Three Double Strand Break Repair Pathways...... 16

Figure 4: The Single Strand Annealing (SSA) Pathway...... 17

Figure 5: Schematic of Protein Cascade during Homologous Recombination.

Adapted from (Caestecker and Van de Walle, 2013) ...... 18

Figure 6: Mechanism of BRCA1 Regulation of Centrosome Number...... 19

Figure 7: Illustration of Co-expression Profiling Using the Pearson Correlation

Coefficient (PCC) parameter...... 20

Figure 8: Overview of Method to Identify Missing 'BRCAs'...... 21

Figure 9: Structure and Function of HATs-HDACs...... 22

Figure 10: HDAC9 and HDAC10 are required for the Homologous Recombination

Process ...... 57

Figure 11: HDACi cause Hyperacetylation of Histones...... 59

Figure 12: Coomassie Staining to Observe Equal Loading of Histones...... 60

Figure 13: Triton Acid Urea gel to show that HDACi treatment caused histone hyperacetylation...... 61

Figure 14: Western analysis to Confirm Depletion of HDACs by RNAi...... 62 xvii

Figure 15: Taqman analysis to Confirm Depletion of HDACs ...... 64

Figure 16: HDAC9 and HDAC10 Depletions Sensitize Cells to Mitomycin C. .... 65

Figure 17: HDAC9 and HDAC10 Depletions Do Not Perturb the Cell Cycle...... 68

Figure 18: Localization of HDAC9 relative to γ-H2AX...... 70

Figure 19: Effects of HDAC9 and HDAC10 Depletions on RAD51 Levels...... 71

Figure 20: NUSAP1 depletion Impairs Homologous Recombination ...... 101

Figure 21: NUSAP1 depletion does not perturb the cell cycle...... 103

Figure 22: NUSAP1 depletion Reduces Single Strand Annealing Levels...... 104

Figure 23: NUSAP1 depletion Causes Centrosome Amplification...... 105

Figure 24: Confirmation of NUSAP1 depletion in Hs578T cells...... 107

Figure 25: NUSAP1 shows Similar Cell Cycle Progression Patterns as BRCA1.

...... 108

Figure 26: NUSAP1 Expression Increases in Response to IR...... 109

Figure 27: BRCA1 Over-expression Reverses Defects in HR seen upon NUSAP1 depletion...... 110

Figure 28: BRCA1 Over-expression Rescues the Centrosome Amplification

Phenotype seen upon NUSAP1 depletion...... 111

Figure 29: IF panels of cells treated as described in Figure 29, depicting % centrosome amplification ...... 112

Figure 30: NUSAP1 siRNA specifically targets NUSAP1 mRNA...... 113

Figure 31: NUSAP1 depletion Does Not Affect BRCA1 Transcription...... 114

Figure 32: NUSAP1 depletion has a modest effect on BRCA1 protein levels . 115

xviii

Figure 33: NUSAP1 depletion Reduces Soluble BRCA1 levels...... 116

Figure 34: NUSAP1 depletion does not Affect Total BRCA1 levels...... 117

Figure 35: NUSAP1 depletion Decreases BRCA1 levels in Chromatin Fraction.

...... 118

Figure 36: Effect of NUSAP1 depletion on BRCA1 IRIF...... 119

Figure 37: Effect of NUSAP1 depletion on γ-H2AX Depletion...... 121

Figure 38: Western Analysis of γ-H2AX Levels upon NUSAP1 depletion...... 123

Figure 39: Effect of NUSAP1 depletion on RAD51 Levels...... 124

Figure 40: Effect of NUSAP1 depletion on RAD51 Levels...... 126

Figure 41: Effect of NUSAP1 depletion on 53BP1 Foci...... 127

Figure 42: Effect of NUSAP1 depletion on 53BP1 Levels...... 129

Figure 43: Localization of BRCA1 and NUSAP1...... 130

Figure 44: TCGA database mutation analysis of HDAC9 and HDAC10 ...... 146

Figure 45: TCGA database mutation analysis of NUSAP1 ...... 147

xix

Chapter 1: Introduction

1.1 BRCA1 – General Biology and Function:

BRCA1 (Breast Cancer Associated Gene 1) is an important breast and ovarian specific tumor suppressor (Miki et al., 1994). It consists of 24 exons, and encodes a large protein of 1,863 amino acids. The protein structure consists of a

RING (Really Interesting New Gene) finger domain at the amino terminus, and a

BRCT (BRCA1 C terminus) domain at the carboxyl terminus, consisting of two repeats (Bork et al., 1997; Lorick et al., 1999) (Figure 1).This BRCT motif is capable of binding phosphorylated proteins during the DNA damage response.

BRCA1 also possesses two NLS (nuclear localization sequences).

BRCA1 is indispensable for a host of biological functions including regulation of transcription, replication, recombination, cell cycle, DNA repair, chromatin structural integrity, and genomic stability (Parvin, 2004) (Figure 2).

The first evidence for BRCA1 as a transcription factor came from reporter gene experiments in which a GAL4 domain fused to the carboxyl terminus of BRCA1, activated transcription of the downstream reporter gene (Chapman and Verma,

1996; Monteiro et al., 1996). Importantly, BRCA1 associates with RNA polymerase II in the large holoenzyme complex (Scully et al., 1997). It was

1 further discovered that BRCA1 was important for RNA polymerase-II-dependent transcription activation (Haile and Parvin, 1999).

BRCA1 is also important for cell cycle checkpoint activation (Yarden and Papa,

2006). It is phosphorylated in response to DNA damage signals, by the protein kinases ATM (ataxia telangiectasia mutated), ATR (ATM and Rad3 related), and

Chk2 (Checkpoint 2) (Chen, 2000; Cortez et al., 1999; Tibbetts et al., 2000). In this manner, BRCA1 can trigger cell cycle arrest of damaged cells at G1/S or

G2/M boundaries (Yarden and Papa, 2006). The G2/M checkpoint is especially important, as this is how BRCA1 keeps cancer cells in check. Cancer cells, whichaccumulate various DNA insults, bypass it to ensure survival.

Proteins that have a RING-finger domain are presumed to have E3 ubiquitin ligase activity (Lorick et al., 1999). This ubiquitin ligase activity catalyzes the transfer of one or more ubiquitin moieties onto a target protein (Jackson et al.,

2000). This ubiquitinated target is then marked for degradation by the proteasome. BRCA1 forms a heterodimer with BARD1 (BRCA1 Associated RING

Domain 1) via this RING-finger domain (Brzovic et al., 2001; Wu et al., 1996).

This interaction is important in order for the complex to exhibit ubiquitin ligase activity (Hashizume et al., 2001). This complex recruits UbcH5c which then monoubiquitinates unphosphorylated H2AX, one of the first steps in the DNA

Damage Signaling (DDR) pathway. This complex is also responsible for ubiquitinating γ-tubulin, important in the regulation of the centrosome duplication pathway (Starita et al., 2004).

1.2 The role of BRCA1 in DNA Damage Response (DDR):

Cells acquire thousands of DNA insults every day that challenge genomic integrity (Lindahl and Barnes, 2000). Some of these arise as a result of normal replication and recombination in cells, but some arise as result of exposure to toxic chemicals or ionizing radiation (Ciccia and Elledge, 2010; Friedberg, 2006;

Jackson and Bartek, 2009). Some of these lesions arise from mistakes during genomic rearrangements in immune and germ cells. DNA damage has adverse effects on the cell’s normal transcriptional and replication machinery. These lesions can accumulate in the form of mutations or chromosomal abnormalities.

The various DNA repair processes function to fix such lesions and prevent them from being passed to offspring (Ciccia and Elledge, 2010; Hoeijmakers, 2001). If the cell fails to repair DNA damage, it can lead to aging, neurodegenerative disorders, cancer etc (Jackson and Bartek, 2009). Thus, DNA damage is important for the maintenance of genomic stability.

There are different kinds of DNA lesions, but the repair mechanisms can be broadly divided into two- those that fix single strand breaks (SSBs), and those that fix double strand breaks (DSBs) (Ciccia and Elledge, 2010; Hoeijmakers,

2001). DSBs are more deleterious because they compromise two strands of DNA at once whereas SSBs maintain a strand to template the accurate repair. The primary mechanisms used to fix DSBs are homologous recombination (HR) and non-homologous end-joining repair (NHEJ), with single strand annealing (SSA)

3 being a less prominent pathway (Moynahan et al., 1999; Stark et al., 2004;

Zhong et al., 2002). BRCA1 is required for all these processes, and depletion of

BRCA1 causes defects in each. Homologous recombination is the DSB repair mechanism with the highest fidelity (Hoeijmakers, 2001). It uses a sister chromatid as a template to repair the breaks. Sister chromatids are present only in S- and G2-phases of the cell cycle, and this is why HR only occurs in S and

G2 (Ciccia and Elledge, 2010; Hoeijmakers, 2001).

When a DSB occurs, the 5’ to 3’ ends undergo resectioning via exonucleases.

The single stranded filament invades the intact sister chromatid; the other exposed strand of the sister chromatid gets displaced, forming a D –loop. HR can occur by two separate pathways: classical HR and alternate synthesis dependent strand annealing (SDSA) model. In classical HR, the displaced strand anneals to the 3’ overhang of the DSB, acting as a template for the 5’ end. Double Holliday junctions are formed between the 4 strands, which then are resolved. Cross-over or non cross-over of genetic material can take place at this point, and finally the

DSB is repaired (Johnson and Jasin, 2000). During SDSA, the D-loop is not formed, and the invading strand is displaced due to new complementary sequences to the other DSB end being synthesized.This elongates the invasive strand. This new strand can now act as a template for the other, and the DSB is repaired.. (Figure 3)

NHEJ is an error-prone method because it does not use sister chromatids. The two DSB ends are processed without a template. The Ku 70/80 heterodimer is

4 recruited to the site of the DSB. It in turn recruits DNA-PKCs, the MRE11-

RAD50-NBS1 complex, and XRCC4/DNA ligase IV to the DSB (Weterings and

Chen, 2008). The DNA-PKCs binds forming a synaptic complex which holds both the DSB ends and the repair proteins (DeFazio et al., 2002). Some chewing back of damaged bases might take place by Artemis (Ma et al., 2002). Ultimately, DNA ligase IV re-ligates the broken ends of DNA (Grawunder et al., 1998) (Figure 3).

NHEJ occurs mainly in GO-, G1-, and early S-phases of the cell (Delacote and

Lopez, 2008).

Single Strand Annealing (SSA) is a pathway of DSB repair which repairs DSBs between two repeat sequences. It does not require identical molecules of DNA, as does classical HR or SDSA. It only requires a single DNA duplex, and uses repeat sequences flanking either end of the DSB for repair (Figure 4).

Resectioning of the DSB ends occur. The resulting 3’ overhangs align and anneal, restoring the DNA strand (Helleday et al., 2007; Ivanov et al., 1996). RPA coats itself to the DSB ends to prevent them from sticking, and RAD52 is recruited to anneal the ends (Lyndaker and Alani, 2009). Any unwanted non- homologous sections in the overhangs get chewed by RAD1/RAD10. SSA is considered as a mutagenic process, as it results in the deletion of DNA.

BRCA1 requires to be activated by phosphorylation in order to function in the

DDR. Following DNA damage, the MRN complex binds to the DSBs, and recruits

ATM (Ciccia and Elledge, 2010; Hoeijmakers, 2001). ATM phosphorylates H2AX at Ser139 (Krum et al., 2010). This phosphorylated H2AX is now called γ-H2AX,

5 and it recruits BRCA1 and RAD51 to nuclear foci (Smith et al., 2010). ATM also phosphorylates Chk2 which phosphorylates BRCA1 (Ciccia and Elledge, 2010;

Hoeijmakers, 2001), or ATM/ATR phosphorylate Chk1 (Joughin et al., 2005).

Phosphorylated Chk1 now interacts with BRCA1 via its BRCT motif, and this interaction is important for G2/M checkpoint activation (Yarden et al., 2002).ATM and ATR kinases also recruit and associate with BLM, another HR repair protein

(Beamish et al., 2002; Davies et al., 2004). ATM recruits MDC1 (Mediator of DNA damage protein 1) (Ciccia and Elledge, 2010; Hoeijmakers, 2001). MDC1 gets phosphorylated, which in turn recruits more ATM, forming more γ -H2AX. γ-

H2AX also gets acetylated by the HAT (acetyl transferase) TIP60 and ubiquitylated by the ubiquitin conjugating enzyme UBC13. Acetylated and ubiquitylated γ-H2AX detaches from the chromatin, binding to RNF168 (RING

Finger Containing Nuclear Factor), forming conjugates. This conjugate recruits and binds to BRCA1 via Abraxas/RAP80 (van Attikum and Gasser, 2009). This in turn sets off a whole cascade of repair proteins being activated/recruited to the

DNA damage foci/ IRIF (Irradiation Induced Foci), including TOPBP1, RAD51,

PALB2, FANCD2, 53BP1, BRCA2, CtIP etc (Ciccia and Elledge, 2010;

Hoeijmakers, 2001) (Figure 1.6). IRIF indicate the extent of DNA damage, and are seen in several cancers (Bartek et al., 2007) (Figure 5).

1.3 BRCA1 and Centrosome Regulation:

Centrosomes are major microtubule re-organizing centers (Doxsey, 2001). They are responsible for controlling the number, polarity, localization, shape etc of microtubules. Centrosomes normally undergo one cycle of duplication in S phase, and by end of mitosis, the cell has two mature centrosomes that localize to the mitotic spindle, and help ensure proper chromosomal segregation takes place. Cells in many cancer cell lines display centrosomal aberrations, including supernumerary centrioles (D'Assoro et al., 2002). Breast tumor cells also have abnormal centrosome numbers that in turn show increased microtubule nucleation (Lingle et al., 1998).

BRCA1 is important for proper centrosomal regulation. Studies have shown that mutant Brca1 mice exhibit centrosome amplification and aneuploidy (Deng, 2001;

Xu et al., 1999). Inhibition of BRCA1 activity by binding of a specific protein fragment to its carboxyl terminus also caused centrosome amplification in human mammary cell lines (Schlegel et al., 2003). Depletion of BRCA1 by RNAi resulted in centrosome amplification in human breast cells, but not in non-breast cells

(Starita et al., 2004). This suggests that breast cells growing in culture are dependent on BRCA1 for centrosome regulation. The BRCA1-BARD1 complex localizes to the centrosomes (Hsu and White, 1998; Sankaran et al., 2006).

BRCA1 E3 ubiquitin ligase activity regulates centrosome function, as BRCA1 ubiquitinates γ-tubulin, and full length BRCA1-BARD1 are required for this

(Starita et al., 2004). The ubiquitination occurs on γ-tubulin on Lysine-48.

Therefore, when a mutant version of tubulin with Lys48 substituted by Arg, was transfected, centrosome amplification resulted (Sankaran et al., 2006; Starita et al., 2004). Thus, as both RNAi and utilization of mutant γ-tubulin, result in centrosome amplification, a strong enzyme-substrate relationship between

BRCA1 and γ-tubulin is indicated. BRCA1-dependent ubiquitin ligase activity controls γ-tubulin localization to the centrosomes (Sankaran et al., 2007).

Thus, experimental evidence points toward an important role for BRCA1 in regulation of centrosome numbers. While BRCA1 loss causes centrosome amplification, BRCA1 enzymatic activity inhibits microtubule nucleation (Parvin,

2009), as this implies that microtubule nucleation activity of centrosomes promotes more centrosome formation One study demonstrated that BRCA1 did not prevent primary duplication of centrosomes, and that its loss blocked reduplication of aberrant centrosomes (Ko et al., 2006). It has been suggested that the mechanism by which BRCA1 regulates centrosomes is cell cycle-dependent.

BRCA1 is primarily expressed in S- and G2-phases of the cell cycle, when microtubule nucleation is low, AURKA (aurora kinase A) levels are low, and its

E3 ubiquitin ligase activity prevents centrosome duplication (Parvin, 2009).

AURKA phosphorylates BRCA1 and inhibits its E3 ubiquitin ligase activity (Ouchi et al., 2004; Sankaran et al., 2007). When BRCA1 activity is lost, microtubule nucleation is increased, AURKA remains low but potent in its inhibitory action; centrosomes undergo multiple cycles of reduplication, causing amplification and aneuploidy (Figure 6). AURKA is over-expressed in 62% of breast cancers,

8 increasing its levels inappropriately in S and G2, and suppressing BRCA1

(Miyoshi et al., 2001). This is why increased AURKA correlates with loss of

BRCA1 and centrosome amplification.

1.4 BRCA1 and Cancer Association:

About a decade ago, researchers discovered that loci on chromosomes 17q21 and 13q12.3 were linked to the inheritance of familial breast and ovarian cancer.

Also, in familial tumors, a loss of heterozygosity was observed at these loci, to which BRCA1 and BRCA2 were positionally cloned. (Miki et al., 1994; Wooster et al., 1995). Later, it was demonstrated that BRCA1 and BRCA2 are tumor suppressor genes (Collins et al., 1995; Cornelis et al., 1995). Statistics show that one out of nine women in the U.S.A. will develop breast cancer in her lifetime.

Upto 10% of all breast cancers are hereditary. Mutations in BRCA1 or BRCA2 are autosomal dominant. Carriers of mutations have an increased risk as high as

80%, of developing breast cancers and a risk for ovarian cancer if they live to be

80 years old. (King et al., 2003; Struewing et al., 1997). Thus, mutations in

BRCA1/2 account for approximately 40% of all inherited breast cancers and increase the risk of developing breast cancers by 80% (Martin et al., 2001).

1.5 Co-expression Profiling to Identify Missing ‘BRCAs’:

As mentioned above, mutations in BRCA1/2 account for nearly 40% of all hereditary cancers. The remaining cases are caused by mutations in unknown genes. They are referred to as the missing ‘BRCAs’ as so far they have not been identified. This may be due to low resolution mutational screening techniques

9 such as SNP (Single Nucleotide Polymorphism) analysis (Mangia et al., 2008;

Oldenburg et al., 2007), comparative genomic hybridization (cGH) and linkage analysis. The rate of success in identification of these genes remains as low as

35% in familial cancers, most likely due to moderate or low penetrance (Gracia-

Aznarez et al., 2013). BRCA1 and BRCA2 on the other hand are high penetrance genes, therefore easy to identify and link to cancer risk (Ford et al., 1998).

Recent advances in sequencing and microarray technology have led to the creation of comprehensive databases including the GEO and Stanford databases

(Barrett et al., 2007; Demeter et al., 2007). Analyses of microarray datasets in these databases have determined that genes and proteins function together as part of vast networks in biological processes. Genes that share such connections could potentially share similar gene expression patterns with one another

(Agrawal, 2002; Jansen et al., 2002; Jansen et al., 2003). These functions can be studied using the gene expression data from such databases (Tornow and

Mewes, 2003). Correlation between gene expression patterns can be observed by applying statistical means like the Pearson Correlation Coefficient (PCC)

(Figure 7). In this manner, the expression patterns of multiple genes can be used to obtain networks and compare a reference transcript such as BRCA1 that is important in cancer. These genes would then undergo biological validation

(Figure 8).

A study in plants used gene expression analysis to identify functional interrelationships between multiple genes (Aoki et al., 2007). More importantly, a

10 study was done with a single microarray dataset of 104 healthy human tissues

(Pujana et al., 2007). The reference genes used were BRCA1, BRCA2, ATM and

CHEK2, all known for their association with cancers. One candidate identified in this manner was HMMR. This gene was experimentally validated theough linkage to the biological pathways in which BRCA1 plays a role. Another similar study identified the candidate PAF15, then shown to interact with BRCA1 (Kais et al., 2011).

Such methods of gene expression analysis can be very successful in identifying genes tin collaborating networks.

1.6 Histone Deacetylases (HDACs):

Acetylation of lysine residues on histones is a key post-translational modification, regulated by two classes of enzymes: histone acetyl transferases (HATs) and histone deacetylases (HDACs). Hyperacetylation of histones catalyzed by HATs opens up the chromatin and stimulates transcription (Hebbes et al., 1994; Ikeda et al., 1999; Kadosh and Struhl, 1998). HDACs carry out hypoacetylation of histones which leads to a closed chromatin configuration, and the limitation of access to transcription factors. This suppresses transcription (Braunstein et al.,

1996; Kuo et al., 1998) (Figure 9). Thus, these two classes of enzymes maintain cellular equilibrium by regulating genes important for transcription, cell proliferation and differentiation, cell cycle dynamics and angiogenesis (Glozak and Seto, 2007; Luo et al., 1998).

HDACs are classified into the following classes based on catalytic domain and cellular localization: Class I (HDAC1, HDAC2, HDAC3, HDAC8), Class IIa

(HDAC4, HDAC5, HDA7, HDAC9), Class IIb (HDAC6, HDAC10), Class III

(SIRT1-7), and Class IV (HDAC11) (Marks and Xu, 2009). Of these, we are especially interested in HDAC9 and HDAC10. HDAC9 is a co-repressor of MEF2

(myocyte enhancer binding factor 2) (Zhang et al., 2002). HDAC9 also interacts with theNcoR repressor complex (Fischer et al., 2002). HDAC10 has two putative binding sites for Rb (Retinoblastoma), indicating it might be important for cell cycle regulation (de Ruijter et al., 2003). HDAC10 can also bind to SMRT repressor complex.

If these HATs and HDACs are aberrantly expressed, cancer may develop

(Marks and Xu, 2009). Epigenetic alterations are observed in many cancers

(Baylin and Ohm, 2006; Pan et al., 2007). Recently, a group of drugs called

HDAC inhibitors (HDACi) have attracted the interest of researchers for their potential as cancer therapeutics (Bolden et al., 2006). Several of these drugs including sodium butyrate (NaB), valproic acid (VPA), apicidin and trichostatin A

(TSA) are cytostatic by decreasing the proliferation of cancer cells by apoptosis or by inducing differentiation (Ginsburg et al., 1973; Han et al., 2000; Vigushin et al., 2001; Zhu et al., 2004). However, besides this, there are few insights into the mechanism of action exerted by these HDACi (Garber, 2007b).

1.7 NUSAP1:

NUSAP1 or NuSAP (Nucleolar and Spindle Associated Protein 1) is a 55 kDa protein that associates with microtubules (Raemaekers et al., 2003). Depletion of

NUSAP1 causes improper mitotic spindle assembly, chromosome segregation and cytokinesis. Over-expression of NUSAP1 causes microtubule bundling.

Thus, it seems that NUSAP1 is important for regaulation of microtubule dynamics, and thus, genomic stability.

NUSAP1 is important for G2/M cell cycle checkpoint activation (Vanden Bosch et al., 2010). NUSAP1 associates with microtubules directly by chromatin binding via a SAP (SAF, Acinus, PIAS) domain (Ribbeck et al., 2007; Verbakel et al.,

2011). Lastly, NUSAP1 is important for embryogenesis as Nusap1 null mice are embryonic lethal (Raemaekers et al., 2003). Also, Nusap1 null zebrafish embryos do not develop properly (Nie et al., 2010).

Recently, several studies have linked NUSAP1 to cancer. NUSAP1 is shown to be over-expressed in several cancers such as prostate, melanoma, gliobalstoma, hepatocellular carcinoma (Marie et al., 2008; Ryu et al., 2007; Satow et al.,

2010).Conversely, NUSAP1 is decreased in ALL (Cario et al., 2008).Its 15q15.1 locus is frequently deleted in cancers (Emanuele et al., 2011)

Figure 1: BRCA1 Structure.

Adapted from (Starita and Parvin, 2003)

Figure 2: BRCA1 Cellular Functions.

Adapted from (Parvin, 2004)

Figure 3: Three Double Strand Break Repair Pathways.

(A) Classical homologous recombination (HR); (B) Alternative HR (SDSA); and

2013)

Figure 4: The Single Strand Annealing (SSA) Pathway.

Adapted from (Valerie and Povirk, 2003)

Figure 5: Schematic of Protein Cascade during Homologous Recombination.

Adapted from (Caestecker and Van de Walle, 2013)

Figure 6: Mechanism of BRCA1 Regulation of Centrosome Number.

Adapted from (Parvin, 2009) 19

Figure 7: Illustration of Co-expression Profiling Using the Pearson Correlation

Coefficient (PCC) parameter.

Figure 8: Overview of Method to Identify Missing 'BRCAs'.

Figure 9: Structure and Function of HATs-HDACs.

(A) Structure of Chromatin; (B) Mechanism of HAT-HDAC function. Adapted from

(de Ruijter et al., 2003)

Chapter 2: Thesis Rationale

Breast cancer is not a single disease, but it is a complex, heterogeneous group of cancer subtypes that are classified as per their molecular signatures (Perou and Borresen-Dale, 2011). Based on if the cancer cells express ER (estrogen receptor), PR (progesterone receptor), and HER-2 (herceptin receptor), breast cancers are divided into five subtypes: Luminal A (ER+ and/or PR+, HER2-);

Luminal B (ER+ and/or PR+, HER2+); HER-2 enriched (ER-, PR-,

HER2+/ErbB2+), basal like/triple negative (ER-, PR-, HER2-), and normal breast- like group. The identification of these subtypes was made possible by pathological analysis and DNA microarray profiling. However, these two techniques have failed to capture the broad spectrum of breast cancers, and there are other subtypes (Weigelt et al., 2008). Of these, BRCA1 is primarily associated with the basal subtype while BRCA2 shows a signature that is more similar to that of luminal B tumors (Joosse, 2012; Roy et al., 2012). Both show mutations in hereditary and sporadic breast tumors However, BRCA1 deficient familial and sporadic breast cancers show similar characteristics (Joosse, 2012)

An inherited mutation in either BRCA1 or BRCA2 highly predisposes an individual to develop breast cancer (Roy et al., 2012). Disease is caused in an autosomal dominant manner, and accounts for 5-7% of all breast cancers. 23

Carriers of these mutations have an increased risk of developing breast cancer as high as 50-80% in their lifetimes. Approximately 40% of hereditary breast cancers show mutations in BRCA1/2 (Martin et al., 2001). The remaining cases of inherited breast cancers that show no evidence of mutations in BRCA1/2, can be caused by mutations in other members of BRCA1-BRCA2 pathways such as

PALB2 (partner and localizer of BRCA2), BRIP1 (BRCA1-interacting protein C- terminal helicase 1), ATM or CHK2 (Roy et al., 2012). Currently, the frequency of mutations in these genes in hereditary breast cancers is low. There are still several unidentified genes.

This could be explained by the fact that BRCA1 and BRCA2 are high penetrance genes, and thus, are easily identified (Ford et al., 1998). The unknown genes that account for hereditary breast cancers in which BRCA1 and BRCA2 are normal, could be of low or moderate penetrance (Gracia-Aznarez et al., 2013). Thus, as these genes are associated with breast cancers, and so far have not been successfully identified, are called ‘missing BRCAs’. Lack of success in discovery is due to the use of low resolution techniques such a genome wide association studies (GWAS) which have identified millions of single nucleotide polymorphisms (SNPs) with no biological relevance (Mangia et al., 2008;

Oldenburg et al., 2007). Exon sequencing approaches reveal the noise of many mutations that may or may not affect the disease. Linkage analysis has failed to identify other high penetrance genes like BRCA1 and BRCA2 that account for a significant number of hereditary breast cancer cases. This is due to the polygenic

24 nature of breast cancer, suggesting that low penetrance genes are involved

(Pharoah et al., 2002). Other techniques like comparative genomic hybridization

(cGH), only look at DNA copy number alterations. Thus, all these conventional techniques have failed because they do not address the multi-factorial nature of cancer development.

The BRCA-regulated pathway(s) is deficient in breast cancer. However, not all of these breast cancer cases are caused by deficient BRCA1/2 function. We hypothesize that there are ‘missing BRCA’ genes that have hitherto, been unidentified in breast cancer that impact the BRCA-regulated pathways in cases where BRCA1/2 are normal. We propose that these BRCA-interacting proteins might be involved in the same BRCA-regulated biological pathways and, when mutant or misregulated, cause breast cancer.

Recently, a number of databases of microarray studies have been created that have enabled researchers to use gene expression data to identify candidate genes (Barrett et al., 2007; Demeter et al., 2007). Analyses of these have shown that genes/proteins do not function independently, but as complex networks in biological pathways (Jansen et al., 2002). The functions of these genes/proteins can be studied using gene expression data from these datasets (Tornow and

Mewes, 2003). This technique is called ‘gene co-expression analysis and it can be applied to identify the missing BRCAs by comparing BRCA1 and BRCA2 gene expression with other genes in the genome. BRCA1 and BRCA2 are aberrantly expressed in different subsets of breast cancer, however, they still

25 function together in the same pathways of homologous recombination and maintenance of genomic stability (Joosse, 2012). Thus, it is possible that although these missing BRCAs, might be expressed in different subsets of breast cancer than BRCA1 and BRCA2, they might be functioning in the same BRCA pathways. We can identify these using a bioinformatics approach i.e. gene co- expression analysis.

Our technique takes into account functional inter-relationships based on similar expression patterns between genes, rather than just genotypic alterations. A common problem with identifying mutations in breast cancer patients or carriers by sequencing analyses, is not knowing if those mutations would be deleterious.

We would not only be identifying potential candidate genes associated with breast cancer risk, but we would also be identifying phenotype based on biological validation in BRCA-regulated pathways. These candidates could be potential biomarkers, and could help give accurate information to patients.

Recent technological advances in whole genome, exome, miRNome, transcriptome sequencing has made it possible to carry out large scale analyses of mutations in genes (Cancer Genome Atlas, 2012). A collaborative effort by an international consortium has combined multiple high thoroughput methods to create a comprehensive database of mutations of hundreds of genes relevant in breast cancer. This technology was inaccessible and cost-prohibitive to us; however, we can use this newly acquired data to determine whether candidate

26 genes obtained through our method are mutated or otherwise disrupted in cancer cases.

The main aim of this study is to identify these genes using co-expression profiling, and then to experimentally validate them in biological assays. These assays test functional significance in the BRCA1-regulated pathways such as homology directed repair and centrosome duplication. Our secondary aim is to identify any mechanistic associations between these genes. Eventually, we plan to test if these genes are linked to cancer development by clinical analysis of breast tumor tissue microarrays.

Chapter 3: Co-Expression Profiling of Breast Cancer Microarrays to Identify

Novel ‘BRCA’ Genes

3.1 Introduction

BRCA1 (Breast Cancer- Associated Gene 1) is an important breast- and ovarian- specific tumor suppressor gene (Huen et al., 2010; Miki et al., 1994). It is important for a number of different cellular functions including transcription, cell cycle regulation, ubiquitination, DNA repair, and centrosome regulation (Huen et al., 2010; Starita and Parvin, 2003). BRCA1 or BRCA2 mutations occur in approximately 40% of all patients with hereditary breast cancer (Martin et al.,

2001). The remaining breast cancer cases are caused by mutations in unknown genes. These genes can be called the ‘missing BRCAs’ because, so far, there has been a failure to identify them by conventional techniques as SNP (single nucleotide polymorphism) analysis (Mangia et al., 2008; Oldenburg et al., 2007).

This could be explained by the fact that these missing ‘BRCAs’ are low penetrance genes, and therefore, hard to identify (Gracia-Aznarez et al., 2013).

BRCA1 and BRCA2, on the other hand, are high penetrance genes, and display a higher association with especially familial breast cancers (Ford et al., 1998).

Large scale genome sequencing and microarray technology has led to the creation of multiple databases such as the GEO database and the Stanford Array

Database (Barrett et al., 2007; Demeter et al., 2007). Analyses of datasets in these publicly available databases have revealed that extensive networks exist between genes and proteins, and that these are potentially responsible for several biological functions. Proteins that function together as subunits in a complex often have similar patterns of gene expression (Agrawal, 2002; Jansen et al., 2002; Jansen et al., 2003). These functional relationships between genes/proteins can be studied using gene expression data from the databases

(Tornow and Mewes, 2003). They can be measured using the statistical parameter Pearson Correlation Coefficient (PCC). This would identify multiple genes that could potentially function together in biological pathways relevant to cancer, by a single approach, using a gene such as BRCA1, which is known to be important for carcinogenesis. Thus, it is imperative that we understand the complex molecular events in cancer development by using a systems approach that would help define the functional interactions between genes/proteins (Khalil and Hill, 2005). Specifically, these networks can be studied by applying informatics approaches including transcriptional profiling which is what we have used in our study, protein-protein interactomes, and phenotypic profiling (Vidal,

2001).

In this study, we have used the strategy of comparing gene expression profiles from several breast cancer microarray datasets in the GEO database, with those of our reference genes - BRCA1, BRCA2 and BARD1. We hypothesize that those genes, whose expression profiles correlate with our reference genes, might

29 function in the same pathways. We were interested in genes that had a similar profile in comparison with our reference genes, implying a potential synergistic associations. Conversely, in chapter 4 we test whether genes that are oppositely expressed (PCC approaches -1.0) have function that opposes our reference genes. In that instance, we found that genes with high correlation but in the opposite direction are not antagonists. Our data-mining has resulted in a list of candidate genes, some of which were tested in the laboratory using functional

BRCA-dependent assays. This list of genes/proteins includes some noteworthy candidates including several members of the HDAC (histone deacetylase) family and NUSAP1 (nucleolar spindle associated protein 1).

3.2 Methods:

3.2.1 Downloading Microarray Datasets:

We scanned the GEO database available on the NCBI website: http://www.ncbi.nlm.nih.gov/pubmed/ for publicly available microarray datasets of

30 breast cancer studies, listed in Table 1. These are called ‘geodatasets’ (GDS).

The dataset of interest was viewed as the ‘geo expression profile’. The data downloaded was provided in the form of compressed “.soft” files. These text files can be opened, and all the genes that the probe-sets hybridized to on a specific microarray chip, can be viewed. In the Mac OS, the .soft format can be opened only by using ‘Text Edit’. Since we are interested in BRCA1, it acts as the main

‘reference/anchor’ gene. We also used other genes including BRCA2 and

BARD1 as anchors. So our anchor set consisted of BRCA1, BARD1, and

BRCA2. Using MATLAB as the platform, a software algorithm called ‘PCC testing’ scanned the various datasets for genes that co-express with each of the anchor genes. The results open up as a separate file with a list of genes, for every GDS run, for each of these anchors. After repeating this coexpression analysis for multiple datasets, a second software algorithm called ‘Intersect’ was applied to obtain an intersection of all the genes that co-express with two or multiple anchor genes in multiple datasets. These results are also in a Microsoft

Excel spreadsheet format. We can select candidates from these ‘hits’ based on

PCC values (higher the better), and other criteria specified below.

3.2.2 PCC Testing:

The Pearson’s Correlation Coefficient (PCC) was used to determine the co- expression profile of genes by comparing the expression of their mRNA across multiple samples in a dataset with that of the anchor genes. PCC values are assigned to every gene by a certain software algorithm, depending on how similar or dissimilar their mRNA levels are to that of the anchors. These values vary from -1 thru 0 to +1 with -1 being a perfect negative correlation, and +1 being a perfect positive correlation. A PCC value < +1, implies a synergistic association between the ‘test’ gene and the anchor gene, whereas a PCC value

> -1, indicates that the two genes being compared are co-regulated, but possibly antagonistic. In both cases, a higher positive or negative value could indicate a greater chance that the two genes function in the same biological pathways.

Different criteria were entered in the PCC testing algorithm, such as names of anchor genes, GDS numbers and most importantly ‘filepath’ or the location where the soft files are located. For PCC testing, one can process multiple datasets at once for multiple anchor genes. However, it is important to ensure that each of the anchor genes is present in the datasets, by clicking the soft file and using the ‘Find’ command. Once everything is properly specified, and this has to be emphasized because MATLAB is very sensitive to even slight changes in syntax, the ‘Run’ tab can be clicked. A window opens up showing ‘Change

Directory’ as one of the options. This should be selected. MATLAB shows a

‘Busy’ status as it scans all the datasets. In the code one should specify Multiple

GDS=1 and Multiple Gene=1 if both are being run in multiples.

3.2.3 Intersect Analysis:

The ‘Intersect’ software code was opened in MATLAB. File path, GDS number,

Anchors and PCC threshold were specified. Usually, +0.6 is preferred from previous networking studies. However, we set a stringent PCC threshold for our analysis i.e. +0.7. Intersect will run multiple anchors to gain an intersection of common genes that are co-expressed with all of the anchor genes, but it can only run one GDS at a time. The ‘Run’ tab was clicked. The ‘change directory’ option was selected upon being prompted. At the end of the analysis, the read out was obtained in the form of multiple excel sheets. A third code called ‘Gene script’ is also required, but this runs independently in the background.

3.2.4 Criteria for Selection of Gene Candidates:

The results of the computational analysis were listed in excel sheets. We scanned these, and ranked genes using the PCC values as the main criteria.

Genes that showed high PCC values, either positive or negative, and that correlated with all three reference genes in multiple datasets were selected.

Ranking of Genes was also based on the following criteria: Gene Ontology, gene orthologs, use of breast cancer lines since BRCA1 is mutated in breast cancer patients. This generated a substantial list of dozens of genes, which were further narrowed for biological validation. The following criteria were used for the selection of candidate genes to be biologically validated in functional assays:

1) High +PCC values.

2) Genes with unknown function or those that are not known to interact with

BRCA1.

3) Genes encoding proteins having enzymatic domains that could act as

potential drug targets.

4) Genes that map to loci that show chromosomal aberrations in cancer.

3.3 Results:

The excel sheets listed hundreds of potential candidate genes that were co- expressed with BRCA1, BRCA2 and BARD1. We narrowed these down to a few dozen based on the gene ranking criteria mentioned above. The results are listed

33 in Table 2 and in the appendix. Our list contained many interesting genes known to play important biological roles. Among these were three members of the

HDAC family i.e. HDAC4, HDAC6, HDAC9. We were especially interested in these because they had a strong negative correlation with BRCA1, BRCA2 and

BARD1. Their PCC value was < -0.7 that was our set threshold. Also, these

HDACs possess enzymatic activity i.e. they hypoacetylate chromatin, compacting chromatin, and causing global transcriptional repression (Luo et al., 1998). Our next interesting candidate was NUSAP1, which demonstrated a strong positive correlation with all three reference genes with a PCC value of > 0.7. NUSAP1 was chosen because it was a relatively unknown gene that had no known connection with BRCA1. Also, the chromosomal region associated with NUSAP1 that is 15q.15 is associated with deletions (Cancer Genome Atlas, 2012;

Emanuele et al., 2011). There were some other candidates such as

KIAA0101/PAF15 on this list that were successfully validated in previous study in our lab (Kais et al., 2011). NOTCH3 was another candidate gene that showed high negative correlation with BRCA1. NOTCH3 belongs to the NOTCH family of signal transduction factors (Guruharsha et al., 2012). NOTCH signaling regulates several biological processes like cell growth and differentiation, apoptosis etc.

Aberrations in NOTCH signaling cause predisposition to cancer and other genetic abnormalities. Given the broad scope of action of NOTCH3, we thought that it would be a good candidate to pursue for biological validation. We used

Compound E, a gamma-secretase inhibitor, which is a crucial enzyme that

34 cleaves NOTCH receptors upon ligand binding, activating the pathway (Ferrari-

Toninelli et al., 2010). We tested the effect of NOTCH3 depletion using compound E in our homologous recombination assay. However, we did not see any significant changes in repair efficiency, and dismissed NOTCH3 as a candidate. This shows that while our method of co-expression profiling yields several candidates that are successfully validated, it does not always corroborate with biological analyses.

3.4 Discussion:

As a result of our bioinformatics analysis, we have identified two sets of genes, the negative correlators including three members of the HDAC family, and the positive correlators like KIAA0101/PAF15 and NUSAP1. We were interested in the HDAC members as their negative correlation could imply an antagonistic association with BRCA1. Multiple different HDAC genes/proteins were repeatedly observed in this analysis, and we reasoned that all of them could simultaneously be blocked using chemical inhibitors. This experiment is done in Chapter 4. This is important because it was our reasoning that inhibiting these HDACs could potentially reverse the mutant phenotype of BRCA1. HDACs play an important role in regulating transcription, thereby, maintaining the dynamic equilibrium of the cell. Aberrant expression of some HDACs has been associated with some cancers (Marks and Xu, 2009). Recently, there has been a lot of interest in

HDAC inhibitors as potential cancer therapeutics (Xu et al., 2007). However, not

35 much is known about their mechanism of action (Garber, 2007b). As for

NUSAP1, as mentioned, it is a little known protein that is important for proper cytokinesis in mitosis (Raemaekers et al., 2003). Aberrant expression of

NUSAP1 is also seen in some cancers (Iyer et al., 2011). This protein is analyzed in Chapter 5.

Our study is along the same lines as another study done in plants, in which gene expression analysis was done using multiple reference genes (Aoki et al., 2007).

This enabled identification of genes that interact in the same pathway. We also closely mimicked a study done in which the analysis was performed using a single microarray dataset of 104 samples from healthy human tissue (Pujana et al., 2007). Four reference genes were used i.e. BRCA1, BRCA2, ATM and

CHEK2. All four genes are associated with cancers. The results obtained included HMMR, which ended up being successfully validated in biological assays. HMMR was found to co-immunoprecipitate with BRCA1. Also, its role was tested in the centrosome duplication assay in which BRCA1 is known to play a role. Inhibition of BRCA1 function causes centrosome amplification (Starita et al., 2004), and likewise, depletion of HMMR by RNAi, too, caused the formation of centrosome amplification. Plus, as mentioned above, PAF15 was successfully shown to be associated with BRCA1 (Kais et al., 2011). PAF15 was found in complex with BRCA1, and its depletion and over-expression caused centrosome amplification. Importantly, tissue microarray analysis showed that PAF15 was over-expressed in breast cancer.

This tells us that we have used a robust method for identification of novel genes that might be involved in the same pathways as BRCA1, and could potentially interact with BRCA1. The above mentioned genes ended up being successfully validated in functional assays, and were also shown to be aberrantly expressed in cancer. Thus, in this manner, these genes could fulfill the role of missing

BRCAs.

We plan to validate our candidates i.e. the HDACs and the NUSAP1 by depleting them by RNAi, and testing their function in two biological assays: Homology directed repair and centrosome duplication, in which BRCA1 is known to play an important role (Huen et al., 2010; Starita and Parvin, 2003). If they end up being important for these processes, we plan to test them for any potential interactions with BRCA1 using co-immunoprecipitation analysis. Lastly, we are curious to know if these genes are important for the breast carcinogenesis process by examining their expression in breast tissue microarrays.

Accession No # Description Platform Type Samples GDS 2761 Hypoxia Inducible factor Sentrix Human 6 12 depletion (MCF-7) Beadchip GDS 2760 Hypoxia Inducible factor HG-U133 2.0 12 depletion (MCF-7) GDS 2635 Invasive ductal and lobular Affymetrix gene chip 30 breast carcinomas HG U133 2.0 GDS 2626 Effect of EGF and HRG on Affymetrix gene chip MCF-7 cells HG U133A 2.0 57 GDS 2617 Tumorigenic breast cancer Affymetrix gene chip 12 Cells HG U133A GDS 2415 Breast carcinomas and NKI-AVL Homo Sapiens 59 local recurrence 18k cDNA microaaray GDS 2367 Tamoxifen effect on MCF-7 Affymetrix gene chip 17 expr. ER-a and ER-b HG U133A GDS 2324 MCF-7/BUS cells treated Affymetrix gene chip 25 with 17 B- estradiol HG U133A GDS 2250 Basal like breast tumors Affymetrix gene chip 47 HG U133 2.0 GDS 2046 Ductal carcinoma insitu to Affymetrix gene chip 14 invasive ductal carcinoma HG U133 2.0 Progression GDS 1925 ER-a exressing MCF-7 in Affymetrix gene chip 18 response to hyperactivatn HG U133A of MAPK GDS 1873 Antiestrogen and Affymetrix gene chip 18 aromatase inhibitor effect HG U133A on MCF-7 GDS 1627 Breast cancer cell lines UNC Compugen oligo 83 response to chemotherapy array GDS 1329 Molecular apocrine breast Affymetrix gene chip 49 Cancer HG U133A Continued Table 1: List of Geo -datasets Downloaded from the GEO database.

Table 1 Continued

Accession # No Description Platform Type Samples GDS 881 Breast cancer and ER Affymetrix gene chip 18 Modulators HG-U95A GDS 847 MCF-7 cells in response UNC Compugen oligo 19 to chemotherapeutics array Arcturus 22k human GDS 807 Estrogen positive breast oligo 60 cancer recurrence during microarray tamoxifen therapy Arcturus 22k human GDS 806 Estrogen positive breast oligo 60 cancer recurrence during microarray tamoxifen therapy GDS 360 Breast cancer and Affymetrix gene chip 24 docetaxel treATMent HG-U95A GDS 84 Breast tumor SVC 84 Characterization GDS 2759 Hypoxia inducible 2- Human 6 bead chip 6 oxo dependent dioxygenase Inhibition GDS 1664 PTH knockdown on breast Affymetrix gene chip 6 cancer cells HG U133A GDS 1549 Estrogen effect on ER-a Affymetrix gene chip 12 positive MCF7 HG U133A GDS 1326 Breast cancer cells reexp Affymetrix gene chip 12 ER-a effect to 17-B estra- HG U133A Diol GDS 1250 Atypical ductal hyperplasia Affymetrix gene chip 8 and breast cancer HG U133A GDS 992 ER membrane associated Affymetrix gene chip 6 genes in MCF-7 HG U133A GDS 901 ER-a effect on gene Affymetrix gene chip 12 induction by estradiol HG U133A GDS 846 MCF-7 cells in response UNC Compugen oligo 19 to chemotherapeutics array GDS 483 DACH1 responsive genes Affymetrix gene chip 9 HG U133A

CANDIDATE GENES ANCHOR SET : BRCA1, BRCA2, BARD1 Gene # Datasets # Counts Correlation ~ PCC value ASPM 9 3 P H ZWINT 10 3 P H ZWILCH 7 3 P H MCM6 7 3 P H TOP2A 11 3 P H PRC1 8 3 P H NUSAP1 10 3 P H PBK 6 3 P H CENPE 9 3 P H LSM5 6 3 P L TPX2 5 3 P H KIAA0101 9 3 P H NCAPG 9 3 P H GINS2 8 3 P H DTL 10 3 P H CHEK1 9 3 P H CEP55 7 3 P H CDC25C 6 3 P H CDC2 9 3 P H CCNA2 10 3 P H CCNB1 6 3 P H CCNB2 6 3 P H BUB1 4 3 P H ATAD2 9 3 P H EZHZ 9 3 P H SMC2 8 3 P H KIF4A 3 3 P H KIF11 8 3 P H Continued Table 2: List of Interesting Candidate Genes: These were selected based on the criteria mentioned above (lists in Appendix). This table gives information regarding the number of datasets candidates show correlation in; the number

(counts) of reference genes correlated with; whether correlation was negative (N) or positive (P); and if the correlation was high (H) i.e. >+0.7, or low (L) i.e. < +0.7.

Table 2 Continued

CANDIDATE GENES ANCHOR SET : BRCA1, BRCA2, BARD1 Gene # Datasets # Counts Correlation ~ PCC value KIF14 7 3 P H KIF23 7 3 P H TXN 6 3 P H DNMT1 6 3 P H IL8 5 3 N H VDAC2 2 3 N H VDAC3 2 3 N H VCAM1 2 3 N H VAV3 2 3 N H TSPAN31 3 3 N H SAT1 3 3 N H LDLR 3 3 N H BMP1 3 3 N H HDAC6 2 3 N H HDAC9 2 3 N H TOPBP1 7 3 P H TIPARP 4 3 P H TIPIN 8 3 P H SPAG5 5 3 P H SMC4 4 3 P H CDC20 4 3 P H FAS 3 3 P H NOTCH3 3 3 N H HIST1H2BD 3 3 N H GATA2 3 3 N H BAMBI 3 3 N H CASP3 4 3 P H WDR45L 3 3 N H LAMP1 3 3 N H HDAC4 2 3 N H

Chapter 4: Histone Deacetylases 9 and 10 Are Required For Homologous

Recombination

4.1 Introduction:

Acetylation of lysine residues on histones in chromatin is a key, reversible modification and is regulated by two classes of enzymes: Histone Acetyl

Transferases (HATs) and Histone Deacetylases (HDACs). It has been shown that hyperacetylation of histones, driven by HATs (Ikeda et al., 1999), causes open chromatin structure, thereby stimulating transcription (Hebbes et al., 1994;

Kadosh and Struhl, 1998; Kuo et al., 1998). On the other hand, HDACs catalyze histone hypoacetylation and subsequent chromatin compaction that is transcriptionally repressive (Braunstein et al., 1996; Luo et al., 1998). This form of epigenetic management controls multiple processes like transcription, cell proliferation and differentiation, cell cycle progression, and angiogenesis (Glozak and Seto, 2007; Lane and Chabner, 2009).

While epigenetic regulation is important for normal cell function, epigenetic deregulation is detected in some cancers (Baylin and Ohm, 2006; Pan et al.,

2007). Aberrant expression of some HATS and HDACs have been observed in some cancers (Marks and Xu, 2009). As examples, HDAC2 and HDAC3 are

42 overexpressed in gastrointestinal tumors (Dokmanovic et al., 2007; Wilson et al.,

2006).

A group of cytostatic agents called HDAC inhibitors (HDACi) have attracted significant research interest (Bolden et al., 2006). The field was pioneered by the discovery that sodium butyrate (NaB) can cause a transformed cell to revert to its original state (Boffa et al., 1978; Ginsburg et al., 1973). In addition, valproic acid

(VPA), which is also an aliphatic acid like NaB, was found to decrease adenoma formation in Adenomatous Polyposis Coli (Zhu et al., 2004), as well as, induce differentiation of transformed cells and leukemic blastocysts from acute myeloid leukemia (Gottlicher et al., 2001). Apicidin, a cyclic tetrapeptide, has been shown, not only to induce apoptosis in acute promyelocytic leukemia cells (Kwon et al., 2002) but also inhibits the growth of tumor cells (Han et al., 2000).

Trichostatin A (TSA), a hydroxamate, inhibited the proliferation of breast carcinoma cells in a study and had decreased tumorigenesis in a rat model

(Vigushin et al., 2001). Thus, HDACi can exert their action across a wide range of cancers (Xu et al., 2007) and therefore, show great promise as cancer therapeutics. Despite this information-boom in the last two decades, little is known about the exact biological function of individual HDACs (Garber, 2007a).

In this study, we find that HDAC9 and HDAC10 activity is required in HeLa cells for homology directed repair of DNA double strand breaks. This result was surprising since we had anticipated that HDAC activity would compress chromatin and impede the repair process.

4.2 Methods and Materials:

4.2.1 Cell Culture, Plasmids, Antibodies and Reagents:

The HeLa-DR-13-9 (HeLa-DR) cells had been described before (Pierce et al.,

1999; Ransburgh et al., 2010) and the standard culturing methods were used.

The plasmid expressing the recombination substrate (pDR-GFP) and I-SceI

(pCBASceI) were a kind gift of Drs. K. Nakanishi and M. Jasin (Memorial Sloan-

Kettering Cancer Center, NY). The HDAC inhibitors were procured as follows: trichostatin A (TSA) and apicidin and valproic Acid (2-propylpentanoic acid) and

Sodium Butyrate (NaB) from Sigma-Aldrich. The siRNA sequences for the

HDACs are listed in the Table 3. Transfection reagents Oligofectamine and

Lipofectamine 2000 were obtained from Invitrogen and used according to the manufacturer’s instructions.

4.2.2 Homology Directed Repair Assay:

On Day 1, we plated HeLa-DR cells (~2 X 105 in a 10 cm2 well). The next day, the 50% confluent wells were transfected with 3 µg of pCBASceI plasmid while simultaneously adding the HDAC inhibitors. Lipofectamine 2000 was used as the transfection agent and Optimem as the dilution medium. The inhibitors were added to obtain the following final concentrations: TSA (300 nM), Apicidin (10

µM), Valproic Acid (2 mM) and NaB (10 mM). The transfection medium was replaced 6 hours post transfection and the inhibitors were added again for a total of 24 hours. On Day 5 (96 hours post transfection), the cells were trypsinized and

44 those among 10,000 total cells that expressed green fluorescence were measured using a Becton Dickinson FACSCalibur instrument in the Ohio State

Comprehensive Cancer Center’s Analytical Flow Cytometry core lab. For the experiment with HDAC depletion by RNA interference, we performed two rounds of transfection. On Day 1, we plated HeLa-DR cells (4 X 104 in a 2 cm2 well). On

Day 2, we performed the first transfection with 60 pmoles of siRNA with 1.5 µL of

Oligofectamine. On Day 3, the cells were transferred to 10 cm2 well dishes. On

Day 4, we co-transfected 100 pmoles of siRNA with 3 µg of pCBASCeI expression vector. On Day 7, the cells were analyzed by FACs as above.

4.2.3 Western Analysis:

In parallel with the HDR Assay, protein samples were obtained by siRNA transfected cells using SDS-Lysis Buffer. These were then sonicated, boiled, and analyzed using SDS PAGE gel electrophoresis. The sources of antibodies used to stain the HDACs are as follows: HDAC1 and HDAC2 (Millipore), HDAC3

(Bethyl Labs), HDAC4, HDAC6, HDAC8 (Santa Cruz), HDAC5 (Cell Signaling

Technology), HDAC7 and HDAC11 (Abcam), HDAC10 (Biovision). The HDAC9 antibody has been described previously (Petrie et al., 2003).

4.2.4 Mitomycin C Assay:

For the assay with HDAC inhibitors, the cells were pre-treated with the inhibitors at the following final concentrations: TSA (300 nM), apicidin (10 µM), VPA (2 mM) and NaB (10 mM) for a total of 24 hours. A total of 2000 cells were plated in triplicate per treatment and subjected to 0 ng, 50 ng and 100 ng of Mitomycin C

45 each for 4 hours. Media was then replaced and cells were allowed to proliferate into colonies for two weeks. At the end of that period, the colonies were stained with 0.01% crystal violet solution and then counted.

4.2.5 Cell Cycle Assay:

HeLa-DR cells were pre-treated with the inhibitors at the following final concentrations: TSA (300 nM), apicidin (10 µM), VPA (2 mM) and NaB (10 mM) or cells were depleted of HDAC9 or HDAC10 by RNAi. At time intervals of 24, 48,

72 and 96 hours, cells were harvested and fixed with 70% chilled ethanol for 1 hour. Cells were treated with RNase A (0.1 mg/mL) and stained with propidium iodide (PI) (0.1mg/mL) for 1 hour. Cell cycle distribution was then measured by

FACs which measures DNA content.

4.2.6 Statistical Analyses:

Comparison of each treatment with control was tested by applying a pairwise t- test. Linear trend analysis across the dose level was performed by applying a linear model. For the MMC assay, at the 50 ng/mL and 100 ng/mL dose levels, comparison of each treatment with control is tested by applying a pairwise t-test.

All the data were log2 transformed before applying t-tests.

4.3 Results:

We began this study by searching for genes with expression levels in microarray results that would reflect coordinate regulation with BRCA1, BRCA2, and BARD1

(Pujana et al., 2007). We found that many HDACs were among those genes co-

46 regulated in microarray datasets with these three genes important in breast cancer and in DNA double strand break repair (data not shown). We tested the role of HDACs in homology directed repair (HDR) in a tissue culture based assay and were surprised to find that one or more HDACs were required for the homologous recombination process.

4.3.1 Treatment with HDACi Inhibits Homologous Recombination:

We have cloned a HeLa derived cell line (Ransburgh et al., 2010) that includes the homologous recombination substrate previously developed (Nakanishi et al.,

2005; Pierce et al., 1999). The cell line, called HeLa-DR (Ransburgh et al.,

2010), contains in its genome two inactive alleles of GFP. One of the alleles has an 18-base pair I-SceI restriction endonuclease site. Expression in these cells of the I-SceI endonuclease results in a single double strand break in one inactive

GFP allele, and homologous recombination using sequences in the second allele will result in gene conversion to produce a functional GFP gene. Homologous recombination events are readily scored by flow cytometry of cells.

We tested the HDAC inhibitors trichostatin A (TSA), apicidin, valproic acid (VPA) and sodium butyrate (NaB). We anticipated that these chemicals would result in hyperacetylation and opening of the chromatin facilitating the repair. To test this, we transfected the endonuclease I-SceI expressing plasmid in the presence of the HDACi, and 6 hours later we washed out the transfection mix and added back the HDACi for another 18 hours. Cells were incubated in the absence of

HDACi for an additional two days. Contrary to our expectation, we found that

47 treatment of cells with three of these HDACi compounds resulted in potent inhibition of homologous recombination. With the exception of VPA, significant decreases in levels of homologous recombination were detected on treatment with the inhibitors (Figure 10A). This inhibitory effect of the HDACi was independent of BRCA1 depletion (data not shown). The inhibition of HDR by the

HDACi was dosage dependent and was up to 11-fold decreased in homologous recombination activity. This level of decrease in the HDR assay was comparable to the loss of activity observed after depletion of BRCA1 by RNAi (Figure 10B).

This result clearly shows that HDACs may play a role in homologous recombination and inhibiting them consequently blocks homology directed repair.

It was striking that VPA was not effective in inhibiting homologous recombination whereas the other three HDACi tested were. The VPA was effective in other assays since treatment of cells with VPA did result in hyperacetylation of histones H3 and H4 as detected by immunoblotting and by TAU gel electrophoresis of purified histones (Figures 11, 12, and 13). We inferred from this specificity of different HDACi in inhibiting the HDR assay that a subset of the different HDACs regulated this process.

4.3.2 Depletion of either HDAC9 or HDAC10 Inhibits Homologous

Recombination:

The next step was to determine which of the HDACs were required for homologous recombination activity. The HDAC inhibitors used broadly inhibit

Class I, II and IV HDACs (28). These classes include HDACs 1 through 11 but do

48 not include the sirtuins (Blander and Guarente, 2004). Thus, HDAC1 through -11 were each depleted by RNA interference. The siRNAs for each of these HDACs were culled from the literature and the sequences and references are listed in supplementary table S1. The HDACs most commonly associated with transcriptional corepressor complexes, HDAC1 and HDAC2 (Laherty et al.,

1997), had a modest inhibitory effect on homologous recombination (Figure 10B).

Depletion of HDAC3, -4, -5, -6, -7, -8, and -11 had no significant effect on homologous recombination activity. Of all the HDACs, depletion of HDAC9 and

HDAC10 significantly decreased homologous recombination (Figure 10B). On noting an initial effect, we tested three siRNAs for each HDAC9 and HDAC10.

HDAC9 depletion had the stronger effect, with siRNA-1 decreasing homology directed repair almost 3-fold and HDAC10 siRNA-2 having a 2.5-fold effect.

Since, we used three siRNAs targeting distinct sequences for each HDAC9 and

HDAC10 and all of these resulted in significantly decreased homologous recombination activity, the inhibition is unlikely to be due to any off-target effects of the siRNA.

We tested for depletion of the HDAC proteins by immunoblot analysis (Figure

14). In those cases in which the commercial antibody did not stain a specific band on immunoblot analysis we quantified the HDAC mRNA by RT-PCR (Figure

15). We detected decreased HDAC protein or mRNA abundance in each case. In some cases, the depletion in the HDAC was modest, and this included both

HDAC9 and HDAC10, both of which had the largest effect on the HDR assay.

Since the level of protein depletion was not complete in each case, we posit that the inhibition of the HDR assay by the HDACi compounds was more complete and thus inhibited the process more than 10-fold, but the siRNA treatment yielded more modest effects.

We also tested whether simultaneous depletion of HDAC9 and HDAC10, or a variety of simultaneous depletions, would additively inhibit homologous recombination. However, in each case there was no change in the levels of homologous recombination (data not shown). When testing these double siRNA experiments, we found that the depletion of each HDAC protein was less effective (data not shown), thus providing a possible explanation for the lack of additive effects. Alternatively, HDAC9 and HDAC10 could function at separate steps in the homologous recombination process, but the difficulty in depletion by

RNAi prevented our testing these two HDACs for redundant versus epistatic activities. These results, therefore, implicate HDAC9 and HDAC10 as key players in the homologous recombination process.

4.3.3 Both HDACi TreATMent and HDAC9 or HDAC10 Depletion Sensitize

Cells to Mitomycin C:

To corroborate the results from the HDR assay, we tested whether inhibition of

HDACs or depletion of HDAC9 or HDAC10 decreased the resistance of cells to the interstrand crosslinker, mitomycin C (MMC). Sensitivity to MMC often correlates with inhibition of the HDR assay (Moynahan et al., 2001; Westermark et al., 2003). We performed a clonogenic assay upon treatment of HeLa-DR cells

50 with the HDAC inhibitors. Again as previously described, we used, TSA (300 nM), apicidin (10 µM), VPA (2 mM) and NaB (10 mM). Cells were pretreated with the HDACi for a period of 24 hours and then plated. Cells were subjected to

MMC at 50 ng/ml or 100 ng/ml for 4 hours in the presence of the HDACi, and then both the MMC and HDACi were removed, cells were washed in PBS, and incubated for two weeks to allow colony formation. Pretreatment with the HDAC inhibitors, apicidin and NaB, significantly sensitized cells compared to the control.

Though TSA strongly inhibited the HDR assay, its effect on sensitizing cells to

MMC was modest, and not statistically significant. VPA had no effect, and this was consistent with the results obtained in the HDR assay (Figure 16A).

We tested depletion of HDAC9 and HDAC10 on resistance to MMC. After two rounds of siRNA transfections targeting HDAC9 or HDAC10, cells were plated and MMC was added to the cells for 4 hours. Compared to the control siRNA transfection, HDAC9 and also HDAC10 depleted cells were more sensitive to

MMC as seen by decreased colony survival. Similar effects on cell survival were seen with all three different siRNAs for both HDACs 9 and 10 (Figures 16B and

16C). In each case, the standard error of the mean did not overlap with the control sample, though when applying the pairwise t-test only the results from

HDAC9 siRNA-1 at 50 ng/ml MMC were significant. For HDAC10, the results from HDAC10 siRNA-1 and siRNA-3 were significant at 50 ng/mL MMC. As with the HDR assay, the effects of the apicidin and NaB had greater magnitude than did the depletion of the HDAC9 or HDAC10, but the results were consistent.

4.3.4 Effect of HDAC Inhibition on Homologous Recombination is not due to Cell Cycle Perturbation:

It has been shown in some studies that HDAC inhibitors block the cell cycle in different phases (Yuan and Seto, 2007). Therefore, we tested whether the effect of the inhibitors and HDAC9 or HDAC10 depletion on homology directed repair is indirect via a cell cycle block. We treated HeLa-DR cells with the HDAC inhibitors- TSA (300 nM), apicidin (10 µM), VPA (2 mM) and NaB (10 mM) for varying intervals of time- 24 hrs, 48 hrs, 72 hrs and 96 hrs. At each of these time points, cells were collected, fixed and then stained with propidium iodide. We note that the 24 hour time point correlates with the time at which the HDR assay was done. Flow cytometry results revealed that TSA and VPA treated cells had a normal cell cycle progression at all time intervals as compared to control cells.

Figure 3A shows the results from 24 hours of HDACi treatment. Apicidin and NaB treated cells, at 24 hours, had a decrease in S- phase cells as evidenced by a decrease in the numbers of cells with DNA content between the 2n and 4n peaks. Since the TSA treatment did not affect the cell cycle but did inhibit the

HDR assay, then its effect on the homologous recombination process is not secondary to a cell cycle block.

Whether HDAC9 or HDAC10 depletion impacted the progression through the cell cycle was tested. Using the same timing as used in the HDR assay, siRNAs specific to HDAC9 and HDAC10 were transfected into HeLa-DR cells, after 48

52 hours the transfection was repeated, and at 24 h, 48 h, 72 h, and 96 h after the second transfection, cells were collected for cell cycle analysis by flow cytometry.

HDAC9 as well as HDAC10 depletion had no discernible effect on cell cycle progression at 24 hours post transfection (Figure 17), and at 48 h, 72 h, and 96 h

(data not shown). Therefore, the effect of HDAC9 and HDAC10 depletions on the double-strand break repair process is not attributed to a cell cycle block but to a direct activity required for the repair process.

4.4 Discussion:

In this study, we have established a role for HDAC9 and HDAC10 in the homologous recombination process. We used a tissue-culture based homology directed repair (HDR) assay, in which repair of a specific double stranded break by homologous recombination converts an inactive GFP allele to active GFP.

Contrary to our expectation, three of the HDACi reduced homologous recombination levels significantly.

Next, we found that only HDAC9 and HDAC10 depletions resulted in a large defect in homologous recombination activity. This effect was observed with multiple siRNAs for each HDAC9 and HDAC10. The requirement for HDAC9 and

HDAC10 in the homologous recombination process was not indirect via blocking the cell cycle progression in an inappropriate point in the cycle. Rather, our results suggest that HDAC9 and HDAC10 are direct participants in the process.

Initial attempts to test whether HDAC9 localizes at the site of damage were

53 inconclusive (Figure 18). Immunofluorescence microscopy revealed that HDAC9 protein is diffusely nuclear, and following ionizing radiation it remained diffusely nuclear and did not relocalize to foci containing γ-H2AX. The HDAC9 was present at the DNA repair foci, but it did not concentrate at the foci after DNA damage. Thus, we cannot rule out that it could be acting indirectly, perhaps via transcriptional regulation. We were unable to carry out localization studies on

HDAC10 due to lack of an effective antibody.

HDAC9 has been described as a co-repressor of myocyte enhancer binding factor 2 (MEF2), which represses the expression of myocyte-specific genes

(Zhang et al., 2002; Zhou et al., 2001). HDAC9 also interacts with the NCoR corepressor complex (Guidez and Zelent, 2001). HDAC10 has also been found in complex with transcriptional repressors (Fischer et al., 2002; Lai et al.).

Since we would expect a transcriptional response required for DNA repair to be stimulatory, as opposed to the transcriptional repressive activity of HDAC9 or -

10, it is more likely that HDAC9 or -10 is a direct participant at the site of DNA damage in the homologous recombination process rather than indirect via transcriptional control. Since these histone deacetylases are stimulatory to the homologous recombination process rather than repressive, we feel the most likely scenario is that these histone deacetylases execute an important deacetylation reaction at the site of damage. Current experiments are aimed at finding the critical substrate of this reaction.

A study had shown that treatment with the HDACi PCI-24781 could inhibit homologous recombination via decreased levels of RAD51 protein (Adimoolam et al., 2007). We tested whether the activities of HDAC9 and HDAC10 affect

RAD51 levels. We observed modest increases, not decreases, in levels of

RAD51 protein upon HDAC9 or HDAC10 depletion (Figure 19). This result is inconsistent with an anticipated indirect affect on homologous recombination of the HDAC9 or HDAC10 proteins via regulation of the expression of other known

DNA repair factors.

While we were preparing this manuscript a recent study was published that found a role for HDAC1 and HDAC2 in nonhomologous end-joining (NHEJ) via specific deacetylations on H3K56 and H4K16 (Miller et al., 2010). HDAC1 and HDAC2 depletion increases levels of H3K56Ac at sites of damage and increased DNA damage response (DDR) signaling. Another study, led to the discovery that inhibition of HDAC3 impairs both homologous recombination and NHEJ

(Bhaskara et al.). However, in our study, depletion of HDAC3 did not affect homologous recombination, and this could be attributed to differences in cell lines. Yet another investigation, highlighted the importance of SIRT1, a class III

HDAC, for homologous recombination, via WRN helicase (Uhl et al.). We did not pursue this since our HDACi do not affect Class III HDACs (29). In light of all this recent evidence, it is now becoming increasingly apparent that HDACs regulate

DNA repair on chromatin. Different HDACs regulate different repair processes:

HDAC1 and -2 regulate NHEJ, and HDAC9 and HDAC10 regulate homologous recombination.

Figure 10: HDAC9 and HDAC10 are required for the Homologous Recombination

Process

A. HDR assay with HeLa-DR cells treated with the following final concentrations of HDACi: TSA (150 nM and 300 nM); apicidin (1 µM and 10 µM); VPA (1 mM and 2 mM); NaB (2 mM and 10 mM). The measure of the cells fluorescing green in the control sample was normalized to 100, and results obtained for the HDACi- treated samples were calculated relative to the normalized control (+/- SEM). A paired t-test was done in which each of the treatments was compared to the control, and statistically significant results were: TSA (p=0.0001), apicidin

(p=0.03), and NaB (p=0.003). B. HDR assay was performed with transfected control siRNA or siRNAs specific for each HDAC1 through HDAC11. The percentage of GFP-positive cells in the control transfected sample was normalized to 100, and the measures for rest of the samples were calculated relative to the control (+/- SEM). By comparing the HDAC-depleted results to the control results using a paired t-test, the following results were found to be statistically significant: HDAC1 (p=0.01); HDAC2 (p=0.04); HDAC9 si1 (p=0.008);

HDAC9 si2 (p=0.00009); HDAC9 si3 (p=0.005); HDAC10 si1 (p=0.007); and

HDAC10 si2 (p=0.003).

Figure 10

Figure 11: HDACi cause Hyperacetylation of Histones.

HeLa-DR cells were treated with the HDAC inhibitors at the following concentrations: VPA (2 mM), TSA (300 nM), NaB (10 mM), and Apicidin (10 uM) for a period of 24 hours. Cell lysates were prepared using NP-40, and Western analysis was performed. Immunoblots were stained ith antibodies specific for acetyl-H3 and acetyl-H4.

Figure 12: Coomassie Staining to Observe Equal Loading of Histones.

Histones were extracted from HeLa-DR cells that had been treated with the

HDAC inhibitors at the indicated concentrations for a period of 24 hours.

Histones were extracted, and loaded onto the gel, after standardizing the volume.

The gel was then Coomassie stained.

Figure 13: Triton Acid Urea gel to show that HDACi treatment caused histone hyperacetylation.

Figure 14: Western analysis to Confirm Depletion of HDACs by RNAi.

A. Immunoblots are shown for protein lysates from cells following two rounds of siRNA transfection as in the HDR assay. Top panel: HDAC9 specific siRNAs

(lanes 2, 3 and 5) were compared with control siRNA that targets luciferase

(lanes 1 and 4). Each immunoblot is paired with a recovery control analyzed using the same filter, RPB1 for HDAC9. Middle panel: HDAC1 (lane 2) and

HDAC2 (lane 3) specific siRNAs were compared to control siRNA (lane 1). In each case, the recovery control was GAPDH. Bottom panel: depletion by HDAC specific siRNAs (each in lane 2) compared to the control siRNA (lane 1) are shown for HDAC3 paired with GAPDH loading control, HDAC4 paired with RHA loading control, HDAC5 paired with RPB1 loading control, HDAC6 paired with

GAPDH loading control, and HDAC8 paired with GAPDH loading control. All

HDACs indicated had bands that migrated at the appropriate position on SDS-

PAGE for the predicted molecular mass and were depleted by the specific siRNA.

Figure 14

Figure 15: Taqman analysis to Confirm Depletion of HDACs

Histogram plots are shown for mRNA levels determined by qRT-PCR, using

Taqman probes, of HDACs 7, 10 and 11 on treatment with the specific siRNAs,

48 hours post transfection, Results were normalized to housekeeping control

OAZ1 and in each set normalized relative to the control RNA.

Figure 16: HDAC9 and HDAC10 Depletions Sensitize Cells to Mitomycin C.

A. Clonogenic survival on exposure to MMC was assessed on treatment of

HeLa-DR cells with the HDAC inhibitors: TSA (300 nM); Apicidin (10 µM); VPA (2 mM); NaB (10 mM). The cells were pre-treated with the HDACi for a period of 24 hours before plating 2000 cells/dish and then adding MMC in doses of 0 ng/mL,

50 ng/mL and 100 ng/mL for 4 hours. Cells were then cultured in the absence of

HDACi and without MMC for two weeks, and surviving colonies were stained and counted. The colony count was normalized to 100 for the 0 ng/mL sample within each treatment, including control (+/- SEM). Results were analyzed using a paired t-test for each treatment using the 0 ng MMC value as the control.

Statistically significant differences from the control were observed with apicidin

(50 ng/mL, p=0.09; and 100 ng/mL, p=0.06) and with NaB (50 ng/mL, p=0.009; and 100 ng/mL, p=0.03). B. The same clonogenic assay as in panel A was repeated but with HDAC9 depletion. The transfections to deplete HDAC9 were performed as for the HDR assay, and sensitivity to MMC was assayed 48 hours after the second transfection. Colony counts were normalized to 100 using the 0 ng/mL MMC results and shown (+/- SEM). Using the paired t-test, comparing the

HDAC9 siRNA to the control siRNA, it was found that HDAC9 si1 significantly sensitized cells to MMC for the 50 ng dose (p=0.01). C. Again the same clonogenic assay as in panel A was repeated but with HDAC10 depletion. The transfections to deplete HDAC10 were performed as for the HDR assay, and 65 sensitivity to MMC was assayed 48 hours after the second transfection. Colony counts were normalized to 100 using the 0 ng/mL MMC results and shown (+/-

SEM). Using the paired t-test, comparing the HDAC10 siRNA to the control siRNA, it was found that HDAC10 si1 and si2 significantly sensitized cells to

MMC for the 50 ng dose (p=0.008) and (p=0.04) respectively.

Figure 16

Figure 17: HDAC9 and HDAC10 Depletions Do Not Perturb the Cell Cycle.

Histogram plots are depicted showing cell cycle progression for the following treatments: A. HeLa-DR cells were pre-treated with the inhibitors at the following final concentrations: TSA (300 nM), apicidin (10 µM), VPA (2 mM) and NaB (10 mM) and harvested at 24 hours. The position in the cell cycle was determined by measuring DNA content of cells stained with PI (0.1 mg/mL) using flow cytometry. B. HDAC9 depleted HeLa cells were analyzed for effects on the cell cycle. C. Cells depleted of HDAC10 were analyzed for effects on the cell cycle.

Figure 17

Figure 18: Localization of HDAC9 relative to γ-H2AX.

Immunofluorescence images are depicted showing localization of HDAC9 relative with γ-H2AX in the absence of irradiation (top) or 6 hrs following 10 Gy of x-ray irradiation (bottom).

Figure 19: Effects of HDAC9 and HDAC10 Depletions on RAD51 Levels.

Immunoblots are shown depicting RAD51 protein levels following HDAC9 and

HDAC10 depletion in both non-irradiated and irradiated cells. The cells were irradiated with 10 Gy X-ray 48 hours post second transfection and then harvested using SDS lysis buffer at intervals of 0.5, 1, 6, 12 and 24 hours. The top panel shows RAD51 levels at different time intervals in control versus HDAC9 and

HDAC10 depleted samples. The lower panel depicts the loading control for

GAPDH.

GL2 Sense : 5’- CGU ACG CGG AAU ACU UCG ATT - 3’ Antisense : 5’- UCG AAG UAU UCC GCG UAC GTT - 3’ BRCA1 Sense : 5’ - ACU GAA GAG UGA GAG GAG CTT - 3’ Antisense : 5’- GCU CCU CUC ACU CUU CAG UTT - 3’ HDAC1(Marchio Sense : 5’ - UCA UCG CUG UGG UAC UUG GTT -3’ n et al., 2009) Antisense : 5’ - CCA AGU ACC ACA GCG AUG ATT -3’ HDAC2(Marchio Sense : 5’ - CAG CCA CAU UUC UUC GAC CTT - 3’ n et al., 2009) Antisense : 5’- GGU CGA AGA AAU GUG GCU GTT - 3’

HDAC3(Ishii et Sense : 5’ - AAA GCG AUG UGG AGA UUU ATT -3’ al., 2008) Antisense : 5’ - UAA AUC UCC ACA UCG CUU UTT - 3’ HDAC4(Chauche Sense : 5’- AUC UUU GGC GUC GUA CAU UTT - 3’ reau et al., 2004) Antisense : 5’ - AAU GUA CGA CGC CAA AGA UTT - 3’

HDAC5(Yoshida Sense : 5’- GGA UGG CAC UGU UAU UAG UTT - 3’ et al., 2007) Antisense : 5’ - ACU AAU AAC AGU GCC AUC CTT - 3’ HDAC6(Rao et Sense : 5’ - GGA UGG AUC UGA ACC UUG AGA TT - 3’ al., 2008) Antisense : 5’- UCU CAA GGU UCA GAU CCA UCC TT - 3’

HDAC7(Chang Sense : 5’ - UGG GUU UCU GUU UCC AGC CTT - 3’ et al., 2006) Antisense : 5’ - GGC UGG AAA CAG AAA CCC ATT - 3’

HDAC8(Waltreg Sense : 5’ - UUG GAU UCG GUG GGG CUC ATT - 3’ ny et al., 2005) Antisense : 5’ - UGA GCC CCA CCG AAU CCA ATT - 3’ HDAC9 siRNA- Sense : 5’ - AUC AUC CUG AGG UCU GUC CTT - 3’ 1(Urbich et al., Antisense : 5’ - GGA CAG ACC UCA GGA UGA UTT - 3’ 2009) HDAC9 siRNA-2 Sense : 5’ - GAA CAA ACU GCU UUC GAA AUC UAU U - 3’ Antisense : 5’ - AAU AGA UUU CGA AAG CAG UUU GUU C - 3’ HDAC9 siRNA-3 Sense : 5’ - UGG GCC AAC UGG AAG UGU UAC UGA A - 3’ Antisense : 5’ - UUC AGU AAC ACU UCC AGU UGG CCC A - 3’ HDAC10 siRNA- Sense : 5’ - CGG AGU CAG UGU GCA UGA CAG UAC A - 3’ 1 Antisense : 5’- UGU ACU GUC AUG CAC ACU GAC UCC G - 3’ HDAC10 siRNA- Sense : 5’- UCA CUG CAC UUG GGA AGC UCC UGU A - 3’ 2 Antisense : 5’ - UAC AGG AGC UUC CCA AGU GCA GUG A - 3’ HDAC10 siRNA- Sense : 5’ - GGU GGU UUC CUG AGC UGC AUC UUG G - 3’ 3 Antisense : 5’ - CCA AGA UGC AGC UCA GGA AAC CAC C - 3’ HDAC11(Liu et Sense : 5’ - AAG CGU GUA UAU AUC AUG GAU TT - 3’ al., 2009) Antisense :5’ - AUC CAU GAU AUA UAC ACG CUU TT - 3’ Table 3: HDAC siRNA sequences

Chapter 5: NUSAP1 Influences the DNA Damage Response by Controlling

BRCA1 Localization to DNA Damage Foci

5.1 Introduction:

The breast cancer associated gene 1 (BRCA1) is a well-documented breast- and ovarian-specific tumor suppressor gene (Huen et al., 2010; Miki et al., 1994). It is indispensable for of a number of important biological functions including maintenance of genomic stability, transcription, and cell cycle checkpoint activation (Chapman and Verma, 1996; Haile and Parvin, 1999; Kim et al., 2007;

Scully et al., 1997; Starita and Parvin, 2003; Wang et al., 2007; Xu et al., 2002).

The BRCA1-BARD1 heterodimer has E3 ubiquitin ligase activity (Hashizume et al., 2001). This complex recruits UbcH5c which then monoubiquitinates unphosphorylated H2AX , and co-localizes with γ-H2AX at DNA damage foci, recruiting other repair factors like RAD51 to the site of double strand breaks

(DSBs) (Chen et al., 2002; Paull et al., 2000). In this manner, BRCA1 regulates a number of different repair processes, namely homologous recombination (HR), non-homologous end joining (NHEJ), and single strand annealing (SSA), and defects in BRCA1 severely impairs these processes (Moynahan et al., 1999;

Starita and Parvin, 2003; Stark et al., 2004; Zhong et al., 2002).The DNA

Damage Response (DDR) is activated to fix DSBs in the DNA by triggering cell

73 cycle arrest, and marking cells for DNA repair by one of these processes (Harper and Elledge, 2007; Huen et al., 2010). If repair fails, the cells either enter into a senescent state, or are fated to die by apoptosis. Defects in the DDR causes predisposition to cancer, as cells acquire mutations, enabling tumor cell survival

(Jackson and Bartek, 2009). Lastly, the BRCA1-BARD1duo is also responsible for the ubiquitination of γ-tubulin (Starita et al., 2004). BRCA1-BARD1 localize to the centrosomes, where they are important for the regulation of centrosome duplication, whereby BRCA1 depletion causes supernumerary centrosome formation, eventually leading to tumorigenesis and aneuploidy (Deng, 2001; Hsu and White, 1998; Schlegel et al., 2003; Xu et al., 1999).

As mentioned above, deficiency in BRCA1- regulated functions somehow causes cancer, although the exact mechanism remains to be elucidated. A known fact, however, is that 40% of familial breast cancer patients havemutations in BRCA1 and BRCA2, and the chances of their offspring who have inherited the mutations, developing the disease is as high as 80% (Ford et al., 1994; Martin et al., 2001;

Struewing et al., 1997). We hypothesize that the remaining cases are caused by mutations in as yet unidentified genes that are probably low penetrance. It is highly likely that these genes might interact with BRCA1 in similar functional pathways.

Nucleolar Spindle Associated Protein 1 (NUSAP1) is a 55 kDa protein that was found to be highly expressed in proliferating cells, that interacts with microtubules

(Raemaekers et al., 2003). Depletion of NUSAP1 caused faulty mitotic spindles,

74 aberrant chromosome segregation, and defective cytokinesis. Over-expression of

NUSAP1 caused microtubule bundling. All this suggests that NUSAP1 is very important for mitotic spindle assembly. NUSAP1 also causes cell cycle arrest at the G2/M checkpoint (Vanden Bosch et al., 2010). NUSAP1 was identified as a mitotic Ran-GTP target, and is phosphorylated by Cdk1 during mitosis (Chou et al., 2011; Ribbeck et al., 2006). Both of these interactions are important for

NUSAP1 association with mitotic microtubules, and as a result of impaired checkpoint activity, the cells are unable to progress through mitosis, ultimately leading to apoptosis (Vanden Bosch et al., 2010). In this way, NUSAP1 is important for normal cell cycle progression. A mechanism by which NUSAP1 associates with microtubules has been suggested, wherein NUSAP1 can bind to chromatin directly via a SAP (after SAF, Acinus, PIAS) domain (Ribbeck et al.,

2007; Verbakel et al., 2011). Lastly, NUSAP1 is important for the process of embryogenesis, as mice embryos null for NUSAP1 are embryonic lethal (Vanden

Bosch et al., 2010). Also, in vivo depletion of NUSAP1 by transfecting in morpholinos showed defective morphogenesis in zebrafish embryos (Nie et al.,

2010).

Most importantly, recent evidence has come forth that shows a potential role for

NUSAP1 in DDR. Studies have shown that NUSAP1 is phosphorylated by

ATM/ATR kinases, inducing a mitotic arrest (Matsuoka et al., 2007; Xie et al.,

2011). Also, NUSAP1 has been shown to be degraded in response to UV damage (Emanuele et al., 2011).

Lastly, there have been numerous studies linking elevated expression of

NUSAP1 to several cancers, including that of prostate, melanoma, glioblastoma, and hepatocellular carcinoma (Bogunovic et al., 2009; Huang et al., 2012; Marie et al., 2008; Ryu et al., 2007; Satow et al., 2010). Higher levels of NUSAP1 antibodies were seen in AML patients (Wadia et al., 2010). NUSAP1 has been associated with increased disease aggression in meningiomas, poor prognosis and higher risk in breast cancers, and increased resistance to chemotherapy in pancreatic ductal adenocarcinoma patients (Chen et al., 2010; Kokkinakis et al.,

2005; Lauss et al., 2008; Stuart et al., 2011). NUSAP1 has also been identified as potential marker for breast ductal carcinoma in situ, and causes resistance to chemotherapy (Kretschmer et al., 2011). Conversely, decreased NUSAP1 levels were seen in childhood ALL patients (Cario et al., 2008). Also, the 15q15.1 chromosomal region associated with NUSAP1 is often decreased in cancers

(Emanuele et al., 2011). Thus, NUSAP1 seems to be an important participant in the process of carcinogenesis, but it remains to be determined exactly how.

Based on initial computational analysis, we have identified NUSAP1 expression to be strongly correlated with BRCA1 expression in breast cancer microarrays. In this study, we have a performed a biological validation of this finding. We found that NUSAP1 depletion impaired key functions that are also regulated by BRCA1 such as homologous recombination and centrosome amplification. In addition, when BRCA1 was over-expressed, it rescued the defective phenotypes seen upon NUSAP1 depletion. This indicated that NUSAP1 and BRCA1 are implicated

76 in the same pathways of DNA repair and centrosome duplication. Upon further investigation, we found that NUSAP1 controls BRCA1 localization and NUSAP1 depletion suppressed recruitment of BRCA1 to DNA damage foci. This is a novel finding that sheds light upon how BRCA1 mediates its repair related functions.

5.2 Materials and Methods:

5.2.1 Cell culture and reagents:

The HeLa-DR, SA-GFP and Hs578T cell lines have been described previously

(Bennardo et al., 2008; Hackett et al., 1977; Pierce et al., 1999; Ransburgh et al.,

2010), and were cultured as per standard American Type Culture Collection

(ATCC) conditions. The siRNA sequences for BRCA1 and NUSAP1 are listed in supplementary table S4. Antibodies for BRCA1 (sc-642), RAD51 (H-92) and

53BP1 (H-300) were procured from Santa Cruz Biotechnology, and γ-H2AX (ser

139, JBW301) and BRCA1 (OP-92) from EMD-Millipore. The antibody for

NUSAP1 was prepared in collaboration with Cocalico Biologicals. Secondary antibodies Alexa Fluor 488 and Alexa Fluor 568 were purchased from Invitrogen.

Transfection reagents Oligofectamine and Lipofectamine 2000 were obtained from Invitrogen and used in accordance with manufacturer’s protocols.

5.2.2 Transfection:

We routinely performed two rounds of transfection. On day 1, we seed cells (∼2

× 105 in a 10-cm2 well) such that they are 50% confluent on day 2. On day 2, 60 pmoles of siRNA with 0.8 ug plasmid were transected into the cells using 1.5 uL 77 oligofectamine, and followed procedures according to the manufacturer. On day

4 100 pmoles of siRNA with 3 ug of plasmid were transfected in using 2.5 uL lipofectamine. Optimem was used as the dilution medium, and tranfection medium was always changed 6 hours post transfection.

5.2.3 Western Analysis:

Cell lysates were harvested using a lysis buffer containing the following: 50 mM

TRIS, 300 mM NaCl, 0.5% NP40, 1 mM EDTA, 5 % glycerol. Alternately, lysates were also prepared using a SDS lysis buffer containing the following: 50 mM

TRIS pH 6.8, 2 % SDS, 10 % glycerol, 5 % β-mercaptoethanol and 0.25 % bromophenol blue dye to color. Cell lysates were centrifuged at 15,000 rpm for

20 minutes, and supernatants were collected. Protein concentration of the supernatant was analyzed using Bradford reagent (Biorad). 50 ug of protein samples were run on SDS-PAGE gels, and then transferred onto polyvinylidene fluoride membranes. These were immune-blotted using specific antibodies.

5.2.4 Homology-directed Repair Assay:

The effect of NUSAP1 depletion on homologous recombination was tested using three different siRNAs specific to NUSAP1, and the HeLa-DR cell line. We performed two rounds of transfection as previously described (Kotian et al.,

2011). Recombination events were scored by flow cytometry using a BD

Biosciences FACS Calibur instrument located at the Ohio State Comprehensive

Cancer Center (OSUCCC) Analytical Flow Cytometry core laboratory.

5.2.5 Single Strand Annealing Assay:

The SA-GFP recombinant cell line was created using stable integration of an hprt-SA-GFP vector into HeLa cells. To test the effect of NUSAP1 depletion on single strand annealing, we performed RNA interference (RNAi) using two siRNAs specific to NUSAP1. Two rounds of transfection were performed. On day

1, SA-GFP cells were plated at a cell density of 4 X 104 cells in a 2 cm2 well. On day 2, the first transfection was performed on cells that were 50% confluent, using 60 pmol of siRNA and 1.5 uL of Oligofectamine in Optimem. The cells were transferred to 10 cm2 dishes on day 3. On day 4, transfection was performed using 100 pmol of siRNA, 2.5 uL of Lipofectamine and 3 ug of pCBASCeI plasmid. On day 7, we analyzed the cells for green fluorescence by FACS analysis.

5.2.6 Centrosome Duplication Assay:

This assay was performed in Hs578T cells using an established technique

(Lingle and Salisbury, 2001). NUSAP1 siRNA and a pJLS148-GFP-centrin vector were simultaneously transfected using Lipofectamine 2000. 48 hours post- transfection, cells were washed with phosphate buffered saline (PBS), fixed with methanol (at -20° C), washed with PBS again, and then treated with DAPI (4', 6- diamidino-2-phenylindole) which is a nucleic acid stain. Cells were analyzed, using a Zeiss Axiovert 200 M immunofluorescence microscope, for the number of centrioles per cell, which are labeled by the GFP-centrin vector.

5.2.7 Total RNA Extraction and Reverse Transcription:

HeLa cells were depleted of BRCA1 and NUSAP1 using RNAi as described above. 48 hours post-transfection, total RNA was extracted and purified from the cells using Trizol reagent (Invitrogen), and as per manufacturer’s instructions.

RNA samples were quantified using Nanodrop ND-1000 Spectrophotometer, and the A260/280 ratio was determined to be less than 1.6. Following this 1 ug/sample of RNA template was reverse transcribed using the iScript cDNA Synthesis kit

(Bio-Rad).

5.2.8 Quantitative Real-time PCR (Q-PCR) Analysis:

The primers used are listed in Supplementary table 5. RPS14 was used as the endogenous control and samples treated with control siRNA acted as the reference. SYBR Green was used as the fluorescent dye using the iQ SYBR

Green Supermix kit (Bio-Rad). The reaction mix was prepared using the kit protocol, and standard Q-PCR cycling conditions were used in the Applied

Biosystems StepOne Plus Real-Time PCR System. Melt curve analysis of the amplicons was performed to ensure that only a single PCR product was being amplified. Samples were run in duplicate and standard deviation was calculated.

Results were calculated using the 2-ΔΔCt formula.

5.2.9 Dual Luciferase Assay:

HeLa cells were subjected to two rounds of transfection as described above with siRNAs specific to NUSAP1, using Lipofectamine 2000. On Day 3, in addition to

80 the siRNAs, HeLa cells were co-transfected with a firefly luciferase reporter gene construct encompassing the promoter region of BRCA1 (pBRC-FF) which was the kind gift of Dr. Peter M. Glazer (Bindra et al., 2005) and a renilla luciferase reporter gene construct containing the thymidine kinase promoter region (pRL-

TK) at the ratio of 10:1. Lysates were harvested 48 hours post second transfection, and the luciferase assay was performed using the Dual-Luciferase

Reporter Assay System (Promega). Luciferase activity was measured using the

Turner Biosystems Veritas Microplate Luminometer. Relative luciferase activity was calculated using the ratio of renilla luciferase activity to firefly luciferase activity.

5.2.10 Chromatin Fractionation:

Cells were washed in cold PBS, and harvested using Buffer N ( 15 mM TRIS-HCl pH 7.5, 60 mM KCl, 15 mM NaCl, 5 mM MgCl2, 1 mM CaCl2, 1 mM DTT, 2 mM sodium vanadate, 250 mM sucrose, 1 mM PMSF, and 1X protease inhibitor cocktail) + 0.6% NP-40. The cells were then pelleted by centrifugation, and the supernatant was collected as the cytoplasmic fraction. Next, the nuclei were lysed in a buffer containing 10 mM PIPES pH 6.5, 10 mM EDTA, 1 mM PMSF, and 1X protease inhibitor cocktail. This was then centrifuged, wherein the supernatant was collected as the nucleoplasmic fraction. The pellet was extracted using 200 ug DNaseI/mL in Buffer B (100 mM KCL, 300 mM sucrose,

10 mM PIPES pH 6.8, 3 mM MgCl2, 1 mM EGTA, 1 mM PMSF, and 1X protease inhibitor cocktail). This was once again subjected to centrifugation, and the

81 supernatant obtained was the chromatin fraction. For more details, refer to the protocols which have been previously described (Citterio et al., 2004;

Remboutsika et al., 1999).

5.2.11 Immunocytochemistry:

Cells were washed with PBS and fixed with 4% paraformaldehyde in PBS. Next, cells were permeabilized using 0.25% Triton-X-100 in PBS, and blocked in buffer containing 1 % goat serum. Cells were then probed using specific primary antibodies, followed by Alexa Fluor secondary antibodies in buffer containing 0.1

% BSA and 6 % Triton X-100. Lastly, cells were stained with DAPI, and then treated with Prolong Gold anti-fade reagent (Invitrogen). Cells were viewed using the Zeiss Axiovert 200 M microscope.

5.2.12 Statistical Analyses:

Comparison of treatments with the controls was done utilizing a pairwise

Student’s t-test. All the data were log 2-transformed before applying the t-test.

5.3 Results:

We performed a computational analysis of publicly available breast cancer microarray datasets to identify genes that are co-regulated with BRCA1, BRCA2 and BARD1 (Hackett et al., 1977; Pujana et al., 2007). NUSAP1 was one of the top-ranked genes that were co-expressed with these three genes across multiple datasets (data not shown). NUSAP1 was also found to be in a cell cycle and

DNA repair network identified in a separate study (Zhang et al., 2012). The three

82 reference genes/proteins are known to play a role in double strand DNA break repair and in the regulation of the centrosome duplication cycle. It was our hypothesis that NUSAP1 would be important in these processes as well.

Therefore, biological validation was done by testing the role of NUSAP1 in the following tissue culture based assays: homology-directed repair (HDR), DNA break repair by single strand annealing, and centrosome duplication.

5.3.1 Depletion of NUSAP1 Suppresses Homologous Recombination:

The HDR strategy is based on a recombination substrate that has been previously described (Pierce et al., 1999). The HeLa-DR cell line contains a single copy of this recombination substrate stably integrated in the genome, which encodes two inactive GFP alleles (Ransburgh et al., 2010). One of these alleles contains a specific 18 base pair I-SceI restriction endonuclease site.

When these cells are transfected with a vector expressing the I-SceI enzyme, it generates a double strand break at the specific site. If in a given cell homologous recombination repairs the dsDNA break, using homologous sequences in the second defective allele, then the recombination event converts the GFP gene to be active. The number of green fluorescent cells created as a result of successful homologous recombination, can be measured using flow cytometry.

We tested the effect of NUSAP1 depletion in the HeLa-DR cell line using RNAi.

We found that NUSAP1 depletion using any of three different siRNAs significantly decreased homologous recombination levels relative to the control

(Figure 20A). NUSAP1 siRNA-1 and siRNA-3 had the strongest effect, reducing

83 homologous recombination levels 2-fold. As all three siRNAs tested decreased homologous recombination significantly, this effect is not due to off-target effects.

All three NUSAP1-specific siRNAs efficiently depleted NUSAP1 protein by immuno-blot analysis (Figure 20B).

The effect of NUSAP1 depletion on homologous repair was unlikely to be secondary to a cell cycle defect. We depleted NUSAP, using two different siRNAs, in the HeLa-DR cell line, following the same protocol, and timeline as the HDR assay. Cells were harvested at 24, 48 and 72 hours post second transfection, for measuring cell cycle progression using flow cytometry.

Treatment with NUSAP1 siRNA-2 did not show abnormal cell cycle progression as compared to the control (Figure 21). . Thus, the suppression of homologous recombination caused by NUSAP1 depletion is not secondary to a cell cycle defect.

5.3.2 NUSAP1 depletion Also Impairs Single Strand Annealing:

We tested the effect of NUSAP1 depletion on the process of DSB repair by single strand annealing (SSA) using another stable, recombinant HeLa cell line, generated using a plasmid (Towler et al., 2013). This cell line contains a GFP reporter system that is analogous to that of the HeLa-DR cell line (Bennardo et al., 2008; Stark et al., 2004), except significant resection is needed to expose sections of microhomology flanking the ISceI site that can anneal to bridge the gap, resulting in a 2.7 kb deletion that is GFP-positive. Upon depletion of

NUSAP1 in the SSA cell line, using RNAi, we found that transfection of either of

84 two siRNAs specific for NUSAP1 significantly inhibited the single strand annealing process (Figure 22).

5.3.3 NUSAP1 depletion Leads to Formation of Supernumerary

Centrosomes:

As it is known that depletion of BRCA1 in mammary tumor cell lines causes centrosome amplification (Kais et al., 2011; Sankaran et al., 2006; Starita et al.,

2004), we investigated whether NUSAP1 depletion would similarly result in centrosome amplification. NUSAP1 siRNA was co-transfected along with a plasmid pJLS148, expressing GFP-tagged centrin, into Hs578T breast cancer cells. GFP-centrin localizes to the centrioles, which can be visualized by fluorescence microscopy. In control siRNA treated cells, 4.5% of the cells had supernumerary centrosomes, whereas NUSAP1 depletion resulted in 16.3% of cells with supernumerary centrosomes (Figure 23A and 23B). A similar phenotype with centrosome amplification in 16% of cells was observed after

BRCA1 depletion, consistent with prior observations. Transfection of two other siRNAs targeting NUSAP1 similarly resulted in cells with multiple centrosomes

(data not shown), indicating that this phenotype was not an off-target effect.

Depletion of NUSAP1 protein by these siRNAs was confirmed in the Hs578T cell line by Western analysis (Figure 24).

5.3.4 NUSAP1 and BRCA1 are Expressed at Similar Cell Cycle Stages

As mentioned earlier, studies have shown that NUSAP1 and BRCA1 are expressed at similar stages in the cell cycle.NUSAP1 levels increase as cells

85 progress through S-phase, peaking around the G2-M transition, and promptly decreasing in G1 (Raemaekers et al., 2003). BRCA1 levels increase through S phase, peaking at late S- early G2 in the cell cycle (Ruffner and Verma, 1997).

We sought to confirm if NUSAP1 and BRCA1 follow the same patterns of cell cycle dependent expression in our HeLa cells. We synchronized HeLa cells at the G1-S boundary using a double thymidine block. Similarly, we synchronized another set of cells in early mitosis using a thymidine – nocodazole block. The former set of cells was released at the following different time-points: S0-early S- phase, S3- mid-S phase, S6- late S-phase, and S9- late S-early G2. The second set of cells was released at the following time-points: M0- mitosis, M3-early G1 phase, M6-mid G1 phase, M9-late G1-phase. Cell lysates were harvested at these various timepoints using NP40 lysis buffer. These lysates were then examined by immune-blotting, for NUSAP1 and BRCA1 levels, using specific antibodies. We also checked levels of cyclin A and cyclin B1 levels. These act as markers of cell cycle progression. Their levels increase in S phase and peak at

S-G2 for both, with cyclin A disappearing in mitosis, and cyclin B1 decreasing in the same phase (Pagano et al., 1992; Pines and Hunter, 1989). As expected, our results showed that BRCA1 protein levels increase steadily in S-phase, peaking at S-G2, and greatly decreasing in mitosis (Fig. 25). NUSAP1 levels, also increase during S-phase, but peak at G2-M, begin to decrease slightly in mitosis, and sharply drop off in G1. These patterns of cell cycle progression are corroborated by levels of cyclin A and cyclin B1.

5.3.5 NUSAP1 concentration increased following DNA damage:

Studies have shown that NUSAP1 is phosphorylated by ATM/ATR in response to

UV induced damage (Matsuoka et al., 2007; Xie et al., 2011). Another study, found that NUSAP1 protein levels decrease upon exposure to UV radiation and the UV mimetic 4NQO due to degradation (Emanuele et al., 2011). We thus tested whether ionizing radiation affected NUSAP1 protein levels. NUSAP1 levels greatly decreased after cytokinesis, and were negligible during G1 (data not shown), suggesting NUSAP1 decay after mitosis. We tested NUSAP1 levels in response to ionizing radiation in G1, when its expression is at its lowest. We synchronized cells using a thymidine-nocodazole sequential block, and 5 hours post release (mid-G1 phase) the cells were subjected to ionizing radiation. The cells were fixed and stained 30 minutes post irradiation, using an antibody specific for NUSAP1. We found that the unirradiated G1 cells showed low, basal levels of NUSAP1 (Figure 26), as expected. By contrast, there was a marked increase in NUSAP1 protein detected in cells 30 minutes post-irradiation. Thus,

NUSAP1 protein expression was induced by ionizing radiation.

5.3.6 BRCA1 over-expression suppresses defects caused by NUSAP1 depletion in homologous recombination and centrosome duplication:

NUSAP1 depletion disrupted the processes of homologous recombination and centrosome duplication. Since BRCA1 is important for both these processes, we tested whether BRCA1 depletion or over-expression could affect the phenotype of NUSAP1 depletion. When depleting BRCA1 together with NUSAP1, the

87 phenotype of the HDR assay or centrosome assay was the same as when only

BRCA1 was depleted (data not shown). By contrast, over-expression of BRCA1 in the same cells transfected with the NUSAP1 siRNA resulted in a partial reversal of the effects of NUSAP1 depletion. Two rounds of transfection were performed in the HDR assay as previously described. We harvested the cells 72 hours post transfection, and analyzed them for GFP-positive cells by FACs analysis. NUSAP1 depletion with add-back of empty pcDNA3 decreased homologous recombination levels 2-fold (Figure 27), consistent with results in

Figure 20. NUSAP1 depletion along with add-back of pcDNA3-HA-BRCA1 restored homologous recombination levels to 85% of full activity. Western analysis of this sample showed that BRCA1 levels were restored back to normal levels as the control (data not shown).

We also performed concurrent NUSAP1 depletion and BRCA1 over-expression in the centrosome assay. Hs578T cells were co-transfected with NUSAP1 siRNA, pcDNA3-HA-BRCA1, and plasmid encoding GFP-centrin. 48 hours post transfection; cells were fixed and viewed by immunofluorescence. While

NUSAP1 depletion with empty vector resulted in the formation of supernumerary centrosomes in 26% of cells (Figure 28 and 29), NUSAP1 depletion in conjunction with transfection of pcDNA3-HA-BRCA1 reduced the percentage of cells with centrosome amplification to 10%, as compared to the control of 5% in this experiment.

5.3.7 NUSAP1 depletion does not affect BRCA1 mRNA levels, transcription, or protein stability:

Based on the previous result, we tested whether NUSAP1 depletion was affecting BRCA1 mRNA levels, either directly or indirectly. We performed a quantitative real-time PCR analysis of RNA from transfected cells. HeLa cells were depleted of NUSAP1, and 48 hours post transfection total RNA was harvested from the cells and reverse transcribed into DNA. Real-time PCR of the resultant cDNAs were conducted using probes targeting the coding sequences of

BRCA1 and NUSAP1. Transcript levels of BRCA1 and NUSAP1 were determined in samples subjected to siRNA depletion, and calculated in comparison to the reference sample treated with control siRNA. The resultant values were then normalized, relative to the transcript levels of the control gene,

RPS14. It was found that BRCA1 and NUSAP1 depletions resulted in down- regulation of the respective genes, but siRNA specific to NUSAP1 did not affect

BRCA1 mRNA levels, nor did BRCA1 siRNA affect NUSAP1 mRNA levels

(Figure 30). Thus, NUSAP1 siRNA specifically targets NUSAP1 mRNA only, and has no effect on BRCA1 mRNA abundance.

Depletion of NUSAP1 by RNAi did not affect BRCA1 promoter activity in a reporter assay. There were no significant differences in the relative luciferase activity of cells transfect with NUSAP1 siRNA (Figure 31), and treating the cells with DNA-damaging ionizing radiation also had no effect. Since NUSAP1

89 depletion had no effect on BRCA1 mRNA levels nor on its promoter activity, we conclude that NUSAP1 depletion has no influence BRCA1 transcription.

We tested whether NUSAP1 affected BRCA1 protein stability. We depleted cells of NUSAP1 in two rounds of transfection, and 48 hours after the second transfection, we irradiated the cells. We then added MG132, a proteasome inhibitor, at a concentration of 20 uM, for 4 hours. At 6 hours post-irradiation, cell lysates were harvested, and BRCA1 levels were analyzed by Western blotting.

Upon NUSAP1 depletion, a very slight decrease was noted in BRCA1 levels, which increased very slightly upon addition of MG132 (Figure 32). We interpret these changes in BRCA1 concentration to be insignificant; suggesting that

NUSAP1 does not affect BRCA1 protein stability.

5.3.8 NUSAP1 depletion affects BRCA1 localization:

We checked if NUSAP1 depletion affects BRCA1 protein levels. We, therefore, performed NUSAP1 depletion in HeLa cells as described above. NUSAP1 levels increase in late S phase, peaking at G2-M, and decreasing in G1 phase

(Raemaekers et al., 2003). Similarly, BRCA1 levels peak in late S phase, decrease in G2, and are lowest in mitosis (Ruffner and Verma, 1997). Both proteins are expressed at the same stages in the cell cycle, and in order to see any cell cycle dependent fluctuations, we used a synchronous cell population.

Cells were synchronized in early S phase, using a double thymidine block and harvested at multiple time-points after thymidine release, using a standard NP-40 lysis buffer. This technique extracts only soluble proteins in the cytoplasm and

90 nucleoplasm, but not proteins from other sub-cellular fractions (Staufenbiel and

Deppert, 1983). Lysates were analyzed for BRCA1 levels by Western blotting.

The different time-points – 0, 3, 6 and 9 hours post thymidine release, correspond to early S-phase, mid S-phase, late S-phase and G2-phase respectively. It was seen that NUSAP1 depletion also caused a depletion of

BRCA1 protein levels (Figure 33). This was seen at all the time-points, and also upon irradiation of the synchronized cells (data not shown). We also looked at

SDS lysates made of cells of NUSAP-depleted cells, and there was no change in

BRCA1 levels as compared to the control (Figure 34). Thus, NUSAP1 depletion affects soluble BRCA1 protein levels, but the total BRCA1 levels remain unchanged. This suggests that NUSAP1 somehow controls BRCA1 localization in the cell.

In order to test this hypothesis, we depleted NUSAP1 by RNAi. 48 hours post transfection we extracted different sub-cellular fractions including cytoplasmic, nucleoplasmic and chromatin-bound proteins. We then tested these for BRCA1 protein levels by Western analysis. It was found that NUSAP1 depletion caused a decrease in the levels of chromatin-bound BRCA1 (chromatin bound fraction shown in Figure 35). This decrease in chromatin bound BRCA1 was more significant upon irradiation. Little or no BRCA1 was detected in the cytoplasmic fraction, and a slight decrease in BRCA1 was also seen in the nucleoplasmic fraction (data not shown). These results are consistent with NUSAP1 regulation

91 of BRCA1 association of chromatin, and this localization is consistent with function in DNA damage repair.

5.3.9 NUSAP1 depletion affects recruitment of BRCA1 to DNA damage foci:

Our results so far are consistent with NUSAP1 depletion blocking homologous recombination, and it affects the chromatin association of BRCA1. We tested whether recruitment of BRCA1 to foci induced upon DNA damage (Paull et al.,

2000) would also be affected by NUSAP1 depletion. Cells were irradiated and then stained for BRCA1 at 0 hours or 6 hours post irradiation. Without irradiation, both control and NUSAP1 depleted cells had some discrete BRCA1 S-phase foci

(Jin et al., 1997; Scully et al., 1997) plus diffusely nuclear BRCA1 content (Figure

36, top). However, 6 hours post irradiation, most of the control cells had irradiation induced foci (IRIF) containing BRCA1, but few NUSAP1 depleted cells had detectable BRCA1 foci (Figure 36, bottom). Counting of cells revealed that the percentage of cells with BRCA1-IRIF dropped two-fold in NUSAP1 depleted cells as compared to the control (Figure 37). The magnitude of this effect of

NUSAP1 depletion on BRCA1-IRIF was similar to the magnitude of the HDR decrease after NUSAP1 depletion. We repeated the same experiment with

NUSAP1 depletion, but with additional transfection of a vector expressing

BRCA1, and this add-back restored the BRCA1 IRIF (data not shown).

We then analyzed other proteins at IRIF subsequent to NUSAP1 depletion.

There was no difference in the levels of -H2AX IRIF between the control and

NUSAP1 depleted cells (Figure 38). Western blot analysis also showed that -

H2AX protein levels were not affected by NUSAP1 depletion upon irradiation

(Figure. 39). The phosphorylation of the H2AX is an early sensor protein at the site of the double strand break upon DNA damage (Rogakou et al., 1998).

BRCA1 is downstream of -H2AX in the DNA damage response pathway (Paull et al., 2000), and is required for assembly of RAD51 at IRIF, following DNA damage (Bhattacharyya et al., 2000). We thus tested whether RAD51 binding to

IRIF was affected by NUSAP1 depletion. Depletion of NUSAP1 diminished, but did not eliminate, RAD51 recruitment to the IRIF (Figure 40). While there was no change in the percentage of cells expressing RAD51 foci, the intensity of the

RAD51 foci was significantly decreased when compared to those of the control cells. In a similar manner, we looked at 53BP1 levels at IRIF upon NUSAP1 depletion. We observed that NUSAP1 depletion caused no reduction in 53BP1 recruitment to double strand breaks and were perhaps even more prominent than in control cells (Figure 41). Western analysis did not show any changes in 53BP1 protein levels in NUSAP1 depleted cells versus control cells (Figure 42). This result suggested that NUSAP1 depletion affected the HR, but not the NHEJ pathway.

5.3.10 NUSAP1 and BRCA1 do not physically Interact:

We wanted to see if BRCA1 and NUSAP1 physically interact with each other. A co-immunoprecipitation experiment was performed, where proteins that complex with BRCA1 were pulled down. Lysates were prepared from these HeLa cells

(synchronized at late S-phase), and immune-blotted for presence of NUSAP1.

Similarly, we pulled down binding partners of NUSAP1, and the immune- precipitated lysates were probed for BRCA1. In either case, we did not see any evidence of physical binding of BRCA1 and NUSAP1 (data not shown). Since,

NP40 lysates were used, this tells us that BRCA1 and NUSAP1 do not physically interact at least in the cytoplasmic or nuclear compartments of the cell. However, this does not rule out the possibility that they might directly interact in some other sub-cellular compartment, for instance, the chromatin.

Lastly, we checked if there was any co-localization of expression of BRCA1 and

NUSAP1 proteins in the cell. We synchronized cells at G1/S boundary with the double thymidine block, and released and fixed cells such that they were in mid S phase. These cells were co-stained for BRCA1 and NUSAP1. We saw that

BRCA1 expression was primarily nuclear, and also in the nucleoli (Figure 43).

NUSAP1, though nuclear, was enriched in the nucleoli. Thus, both proteins showed abundant expression in similar sub-cellular locales. However, although we couldn’t establish very clear patterns of co-localization between these two proteins, we believe it is very likely that they are physically abundant and close together, that they could physically interact.

5.4 Discussion:

In this study, we have successfully validated the involvement of NUSAP1, a candidate gene that shows co-expression with BRCA1, in BRCA1-regulated pathways. We have shown that NUSAP1 depletion suppresses homologous

94 recombination and single strand annealing. Both of these are mechanisms of

DSB repair, and it is possible that NUSAP1 depletion may also affect the process of non-homologous end joining, which also fixes DSBs caused by ionizing radiation (Thompson, 2012). Other studies have shown that NUSAP1 is phosphorylated by ATM/ATR kinases, and also gets degraded in response to UV damage (Emanuele et al., 2011; Matsuoka et al., 2007; Xie et al., 2011).

NUSAP1 depletion also causes centrosome amplification, which fits with the role of NUSAP1 being involved in proper mitotic spindle assembly, and chromosome segregation (Raemaekers et al., 2003). Interestingly, when BRCA1 is over- expressed in both the homologous recombination and centrosome duplication assays, the defective phenotypes seen upon NUSAP1 depletion are reversed.

NUSAP1 depletion followed by BRCA1 add-back restored levels of homologous recombination back to normal, and decreased the percentage of cells showing supernumerary centrosomes. This suggests that NUSAP1 depletion is in turn causing BRCA1 deficiency. Thus, NUSAP1 and BRCA1 are implicated together in these common pathways.

The influence of NUSAP1 on BRCA1 is not due to transcriptional repression as the NUSAP1 siRNAs affect neither BRCA1 mRNA levels, nor BRCA1 transcriptional activity. NUSAP1 also does not affect BRCA1 stability, as evidenced by the lack of change in BRCA1 levels upon NUSAP1 depletion, with and without the proteasome inhibitor MG-132. The first hint we had of NUSAP1 affecting BRCA1 localization was when we found that NUSAP1 depletion affects

95 only soluble BRCA1 protein levels, but not total BRCA1 levels. Our hypothesis was further strengthened when we discovered that NUSAP1 depletion affects the levels of BRCA1 protein in chromatin fractions. Final confirmation was obtained when we observed that NUSAP1 depletion affected the recruitment of BRCA1 to

IRIF as the number of cells showing BRCA1 foci decreased by 2-fold. Once again, simultaneous NUSAP1 depletion and BRCA1 over-expression restored the number of cells with BRCA1 foci. More convincingly, the recruitment of

RAD51, a DSB repair protein downstream of BRCA1, to the IRIF was decreased by NUSAP1 depletion (Paull et al., 2000). Protein levels of RAD51 were unchanged, indicating that this concerns localization of repair proteins upon damage. We checked if NUSAP1 affects 53BP1 foci formation in response to ionizing radiation. We found that while 53BP1 protein levels remained unchanged upon NUSAP1 depletion, there was an increase in the intensity of 53BP1 at the

IRIF. While many studies have established that BRCA1 promotes non- homologous end joining, some show the contrary result (Deng and Wang, 2003).

It has been shown before that BRCA1 knockout in murine embryonic stem cells, while decreasing the frequency of homologous recombination events, increased those of non-homologous end joining (Moynahan et al., 1999; Snouwaert et al.,

1999). These disparities could be accounted for by the fact that homologous recombination and non-homologous end joining occur at different stages of the cell cycle, with the former primarily taking place during late S and G2, and the latter predominantly occurring during G0-, G1- and early S-phase (Delacote and

Lopez, 2008). Also, not all types of non-homologous end joining are dependent upon BRCA1 function (Mak et al., 2000).

Putting the data that we have collected in perspective, it seems that NUSAP1 is indirectly involved in the DDR via its control of BRCA1 localization. Based on the facts that NUSAP1 was phosphorylated by ATM/ATR, and also degraded in response to UV damage, we decided to investigate if NUSAP1 might play a more direct role in DNA damage signaling. We saw that this might be the case, as

NUSAP1 expression surged in response to ionizing radiation even in G1-phase cells. Again, NUSAP1 protein levels did not change, indicating that NUSAP1 protein was undergoing a shift in localization in response to DNA damage. Thus, our data points to a NUSAP1 playing a more direct role in the DDR by itself.

Lastly, we were interested in pursuing evidence for an interaction between

NUSAP1 and BRCA1. We saw that there were similarities in the pattern of cell cycle expression between BRCA1 and NUSAP1, with levels of both being increased in S and G2. This makes it more feasible that there might be an interaction since both proteins express at the same phases in the cell cycle.

Efforts to prove a direct physical interaction, however, did not prove to be successful. Neither NUSAP1 nor BRCA1 were present in the protein complexes pulled down with the other. However, it must be noted that the IP was performed with lysates that were prepared with NP-40 lysis buffer, meaning that we only acquired cytoplasmic and nucleoplasmic proteins. Throughout interphase, while

NUSAP1 shows a nuclear presence, it is highly enriched in the nucleolus

(Raemaekers et al., 2003). It becomes chromatin bound during microtubule nucleation in mitosis. It is quite possible that NUSAP1 and BRCA1 might interact in a different sub-cellular compartment, for instance the nucleolus. Recently, evidence linking several DNA repair factors, including BRCA1, to the nucleolus, has come forth. The nucleolus plays an important role in sensing and responding to various cellular stresses including DNA damage (Boulon et al., 2010). The nucleolus is disrupted in response to damage induced by ionizing radiation, stabilizing p53 (Rubbi and Milner, 2003). Also, in response to damage, there is

ATM-dependent inhibition of RNA Pol I activity in the nucleolus (Kruhlak et al.,

2007). Large scale proteomics analysis of the nucleolus, revealed a host of proteins to be associated with it, including numerous DNA repair proteins like

RAD50, RPA40, TOP2A, TOP2B, BLM, ATR (Andersen et al., 2005; Andersen et al., 2002).

NUSAP1 has a strong association with the nucleolus, and it shuttles from the nucleolus to the nuclear chromatin in order to carry out proper cytokinesis. We now know that BRCA1, too, is associated with the nucleolus. One study showed that BRCA1localizes to nucleoli in frozen mammary tumors (Tulchin et al., 1998).

Another immunohistological study showed that not only was BRCA1 present in the nucleoli of breast cancer tissue, but also in HeLa and MCF7 cell lines

(Tulchin et al., 2010). In this study, they also showed that BRCA1 co-localizes with nucleolin in the above mentioned cell types, and that this is cell-cycle

98 dependent i.e. these patterns of expression are seen starting at the G1/S transition to the G2/M boundary of the cell cycle.

Most recently, it was found that BRCA1 and RNF8 are present in the nucleolus, and upon DNA damage caused by γ-radiation, they translocate to DNA damage foci (Guerra-Rebollo et al., 2012). RNF8 binds to a protein called RPSA in the nucleolus, and when this RPSA is depleted by RNAi, it decreases levels of both

BRCA1 and RNF8. BRCA1-RNF8 levels also decrease in response to γ- radiation. This study gives credibility to our finding that BRCA1 localization to

DNA damage foci is impaired upon NUSAP1 depletion. A number of DNA repair proteins shuttle from the nucleolus to the nucleus in response to DNA damage including MUS81, FEN1, WRN, BLM, and PARP1 (Gao et al., 2003; Guo et al.,

2008; Karmakar and Bohr, 2005; Marciniak et al., 1998; Meder et al., 2005;

Rancourt and Satoh, 2009; Yankiwski et al., 2000). Also, the evidence points toward NUSAP1 being important for maintaining normal BRCA1 protein levels.

Since NUSAP1 depletion does not affect BRCA1 mRNA levels or BRCA1 protein degradation, it could be affecting BRCA1 protein translation. This may be why we see a decrease in BRCA1 protein levels upon NUSAP1 depletion. We have already expounded on the importance of both BRCA1 and NUSAP1 in the tumorigenesis process. A break-through, comprehensive analysis was performed on hundreds of breast patient tumor samples, combining five powerful technological platforms (Cancer Genome Atlas, 2012). We would like to highlight that in this study, an association was found between BRCA1 and NUSAP1, our

99 gene of interest. The genomic loci of both these genes show deletional mutations in luminal breast cancers. To some extent, both genes also are associated with basal breast cancers. While BRCA1 shows deleterious variants in germ-line basal cancers, the NUSAP1 locus shows deletions in basal tumors. In this way,

BRCA1 and NUSAP1 seem to have a crucial role together in cancer development, and NUSAP1 could act as a potential bio-marker for breast cancer.

A detailed study needs to be done in order to gain mechanistic insights into this novel partnership.

100

Figure 20: NUSAP1 depletion Impairs Homologous Recombination

(A) HDR assay was performed where control siRNA and three siRNAs specific for NUSAP1 were transfected into HeLa-DR cells. The fraction of GFP-positive cells in the control sample was normalized to 100, and those in the NUSAP1 depleted samples were calculated relative to the control (+/-SEM). Upon applying a two-tail, paired t-test, treatment with the following siRNAs showed a statistically significant decrease: NUSAP1 si-1 (p=0.01) and NUSAP1 si-4 (p=0.0006). (B)

Western analysis of NUSAP1 depletion was performed. Immunoblots are shown for protein lysates from cells following two rounds of siRNA transfection as in the

HDR assay. NUSAP1 specific siRNAs (lanes 2, 3 and 4) were compared with control siRNA that targets luciferase (lane 1). Each immunoblot is paired with a recovery control GAPDH, analyzed using the same filter.

101

Figure 20

102

si-1 si-2

Figure 21: NUSAP1 depletion does not perturb the cell cycle.

HeLa cells were transfected with either control siRNA or two siRNAs specific for

NUSAP1 in two rounds. Cells were harvested at 24, 48 and 72 hours post second transfection, fixed, and stained with propidium iodide. Analysis of cell cycle progression was done by measuring DNA content by FACS.

103

Figure 22: NUSAP1 depletion Reduces Single Strand Annealing Levels.

SSA assay was performed in which the SA-GFP cells were depleted of control siRNA and two siRNAs specific for NUSAP1 in two rounds of transfection. The fraction of GFP positive cells was normalized to 100, and the measures of the

NUSAP1 depleted samples were calculated relative to the control (+/-SEM). A two-tail, paired t-test was applied, and the following treatments were found to significant: NUSAP1 si-1 (p=0.01) and NUSAP1 si-2 (p=0.02).

104

Figure 23: NUSAP1 depletion Causes Centrosome Amplification.

(A) Centrosome assay was performed in which Hs578T cells were co-transfected with a control siRNA, and siRNAs specific for BRCA1 (positive control) and

NUSAP1, along with pJLS-148-GFP-centrin. 48 hours post-transfection, cells were fixed and stained, and visualized by IF. Centrosome amplification was measured by counting the number of cells with abnormal number of centrioles per 100 cells. The percentage of centrosome amplification in the control cells was about 4.5%, while in BRCA1 and NUSAP1 depleted cells, it was comparable at 16.6% and 16.3% respectively. (B) IF panels showing control cells with normal number of centrioles (<4), and BRCA1 and NUSAP1 depleted cells displaying supernumerary centrosomes (>4). A two-tail, paired t-test was applied wherein

NUSAP si-2 treatment was significant (p=0.001).

105

Figure 23

106

si-1 si-3

Figure 24: Confirmation of NUSAP1 depletion in Hs578T cells.

Cells were transfected with control siRNA and NUSAP1 si-1. 48 hours post- transfection, cell lysates were prepared and western blotting done. Immunoblots compare NUSAP1 levels in the NUSAP1 depleted samples (lanes 2 and 3) to that in the control sample (lane 1).

107

Figure 25: NUSAP1 shows Similar Cell Cycle Progression Patterns as BRCA1.

HeLa cells were carefully synchronized using a double thymidine block at G1/S, and a thymidine nocodazole block in early mitosis. Cells were released at the following timepoints after each block: 0 hrs, 3 hrs, 6 hrs and 9 hrs, and NP-40 cell lysates were prepared. Immunoblot analysis was carried out, and levels of

BRCA1, NUSAP1, cyclin A, cyclin B1 were analyzed at these timepoints. The same filters were probed for the following loading controls- RHA and GAPDH.

108

Figure 26: NUSAP1 Expression Increases in Response to IR.

HeLa cells were synchronized with thymidine-nocodazole block to be synchronized in early mitosis. Cells were released, and 5 hours later in G1- phase, irradiated (10 Gy). Cells were fixed and stained with NUSAP1 specific antibody at 30 mins post-IR.

109

Figure 27: BRCA1 Over-expression Reverses Defects in HR seen upon NUSAP1 depletion.

HDR assay was performed where cells were transfected with control siRNA and

NUSAP1 si-1. Add-backs of empty pCDNA3 or pCDNA3-HA-BRCA1 were performed along with each siRNA treatment. The number of cells expressing

GFP in the control siRNA with empty pCDNA3 add-back was normalized to 100.

The measures of the other samples were calculated relative to the control (+/-

SEM). A two-tail, paired t-test was applied to determine significance of NUSAP1 si-2 treatment and pCDNA3-HA-BRCA1 addback (p=0.06).

110

Figure 28: BRCA1 Over-expression Rescues the Centrosome Amplification

Phenotype seen upon NUSAP1 depletion.

Centrosome assay was performed where Hs578T cells were transfected with control siRNA and NUSAP1 si-1. Add-backs of empty pCDNA3 or pCDNA3-HA-

BRCA1 were performed along with each siRNA treatment. The percentage of centrosome amplification was calculated y counting the number of cells with >4 centrioles per 100 cells. A two-tail, paired t-test was applied to the values obtained in the sample with NUSAP1 si-2 treatment and pCDNA3-HA-BRCA1 add-back, which was found to be significant (p=0.04).

111

Figure 29: IF panels of cells treated as described in Figure 29, depicting % centrosome amplification

112

si-1 si-2 si-3

Figure 30: NUSAP1 siRNA specifically targets NUSAP1 mRNA.

RT-PCR analysis was performed upon samples treated either with control siRNA,

BRCA1 siRNA and NUSAP1 si-1. Total RNA was extracted, reverse transcribed, and the cDNA was amplified and analyzed using primers specific for BRCA1,

NUSAP1 and RPS14. The calculations were done using the 2-ΔΔCT method, relative to the control sample which acted as the reference, and RPS14 which acted as the endogenous control. NUSAP1 siRNA did not cause any significant changes in BRCA1 mRNA levels (p=0.5).

113

Figure 31: NUSAP1 depletion Does Not Affect BRCA1 Transcription.

HeLa cells were co-transfected with two siRNAs specific for NUSAP1 along with vectors pRL-TK and pBRC-FF in 1:10 ratio (0.2 ug: 2 ug). 48 hours post- transfection, cells lysates were harvested and processed using the Promega lysis buffer. 5 ug protein was loaded into wells of a 96 well plated, which was then subjected to analysis by a luminometer. Ratios of firefly luciferase to renilla luciferase were measured, and it was determined as compared to a control siRNA treatment, NUSAP1 siRNA did not cause any changes in BRCA1 transcriptional activity as reflected by the expression of the pBRC-FF vector.

Subjecting cells to IR before cell lysis, also did not cause any significant changes.

114

Figure 32: NUSAP1 depletion has a modest effect on BRCA1 protein levels .

HeLa cells were treated with either control siRNA or NUSAP1 siRNA, and each was subjected to the following conditions: +IR (10 Gy) and +MG132 (20 uM).

Immunoblot analysis was done comparing BRCA1 protein levels in NUSAP1 depleted samples (lanes 2, 4, 6 and 8) with control siRNA depleted samples

(lanes 1, 3, 5, and 7) in that specific order. The same filter was analyzed for loading using anti-RHA.

115

Figure 33: NUSAP1 depletion Reduces Soluble BRCA1 levels.

HeLa cells were transfected with control siRNA, BRCA1 siRNA and NUSAP1 si-1 in two rounds of transfection. 36 hours post transfection; cells were synchronized at the G1/S boundary with a double thymidine block. Cell lysates were harvested at the following timepoints post release: 0 hrs (early S-phase), 3 hrs (mid S-phase), 6 hrs (late S-phase), 9 hrs (mid to late G2-phase).

Immunoblots were stained for BRCA1 protein levels and NUSAP1 levels. The lanes in comparison are of the NUSAP depleted samples (lanes 2, 5, 8 and 11) versus the control sample (lanes 1, 4, 7 and 10) respectively. Loading controls probed were RHA and GAPDH.

116

Figure 34: NUSAP1 depletion does not Affect Total BRCA1 levels.

NUSAP1 depletion was performed and SDS lysates were obtained.

Immunoblots were stained for BRCA1 and NUSAP1, and the levels of BRCA1 protein were compared between lane 3 (NUSAP1 siRNA) and lane 1 (control siRNA). Loading controls probed were RHA and GAPDH

117

Figure 35: NUSAP1 depletion Decreases BRCA1 levels in Chromatin Fraction.

HeLa cells depleted of NUSAP1 were subjected to fractionation, obtaining cytoplasmic, nucleoplasmic and chromatin fractions of the cell. The above immunoblot represents BRCA1 protein levels in the chromatin fraction, analyzed in NUSAP1 depleted cells in comparison with the control cells. This experiment was performed +IR (10 Gy). The same filter was also probed for histone H4, to demonstrate equal loading.

118

Figure 36: Effect of NUSAP1 depletion on BRCA1 IRIF.

HeLa cells were subjected to a control and NUSAP1 depletion. 48 hours post transfection, and under conditions of +IR (10 Gy), the cells were fixed and stained for BRCA1 to analyze foci formation.

119

Figure 36

120

Figure 37: Effect of NUSAP1 depletion on γ-H2AX Depletion.

HeLa cells were subjected to a control and NUSAP1 depletion. 48 hours post transfection, and under conditions of +IR, the cells were fixed and stained for γ-

H2AX to analyze foci formation.

121

Figure 37

122 siRNA: CTL NUS si-1 si-2 si-3 si-1 si-2 si-3

Figure 38: Western Analysis of γ-H2AX Levels upon NUSAP1 depletion.

HeLa cells were subjected to a control and NUSAP1 depletion. 48 hours post

transfection, and under conditions of +IR, the cells were fixed and stained for γ-

H2AX to analyze foci formation.

123

Figure 39: Effect of NUSAP1 depletion on RAD51 Levels.

HeLa cells were subjected to a control and NUSAP1 depletion. 48 hours post transfection, and under conditions of +IR, the cells were fixed and stained for

RAD51 to analyze foci formation.

124

Figure 39

125

si-1 si-2 si-3 si-1 si-2 si-3

Figure 40: Effect of NUSAP1 depletion on RAD51 Levels.

HeLa cells were depleted of NUSAP1, and 48 hours post-transfection were subjected to 10 Gy IR. NP-40 lysates were prepared and analyzed for RAD51 levels.

126

Figure 41: Effect of NUSAP1 depletion on 53BP1 Foci.

HeLa cells were subjected to a control and NUSAP1 depletion. 48 hours post transfection, and under conditions of +IR, the cells were fixed and stained for

53BP1 to analyze foci formation.

127

Figure 41

128

si-1 si-2 si-3 si-1 si-2 si-3

Figure 42: Effect of NUSAP1 depletion on 53BP1 Levels.

HeLa cells were depleted of NUSAP1, and 48 hours post-transfection were subjected to 10 Gy IR. NP-40 lysates were prepared from asynchronous cells, and analyzed for 53BP1 levels.

129

Figure 43: Localization of BRCA1 and NUSAP1.

Co-localization of BRCA1 and NUSAP1: Cells were synchronized in late S phase using double thymidine block, fixed, and stained for BRCA1 and NUSAP1. In addition, cells were subjected to conditions of +IR (10 Gy). IFpanels depict cells stained for BRCA1 and NUSAP, as single-color images, and as composite, merged images.

130

GL2 siRNA Sense : 5’- CGU ACG CGG AAU ACU UCG ATT - 3’ Antisense : 5’- UCG AAG UAU UCC GCG UAC GTT - 3’ BRCA1 siRNA Sense : 5’ - ACU GAA GAG UGA GAG GAG CTT - 3’ Antisense : 5’- GCU CCU CUC ACU CUU CAG UTT - 3’ NUSAP1 siRNA-1 Sense : 5’ - UAUUGGAGACUGGAGUCUGCGUUGC -3’ (Marchion et al., Antisense : 5’ - GCAACGCAGACUCCAAGUCUCCAAUA -3’ 2009) NUSAP1 siRNA-2 Sense : 5’ - UUUCGUUCUUGCUCGCGUUUCUUCC - 3’ (Marchion et al., Antisense : 5’- GGAAGAAACGCGAGCAAGAACGAAA - 3’ 2009)

NUSAP1 siRNA-3 Sense : 5’ -GAGCACCAAGAAGCUGAGAAU -3’ (Ishii et al., 2008) Antisense : 5’ - AUUCUCAGCUUCUUGGUGCUC - 3’ Table 4 NUSAP1 siRNA sequences.

131

BRCA1 Forward primer 5’- GTGTCCCATCTGTCTGGAGT–3’ BRCA1 Reverse primer 5’–TAAAGGACACTGTGAAGGCCC–3’ NUSAP1 Forward primer 5’–CAACCTGAGGGCAACCAAGT–3’ NUSAP1 Reverse primer 5’–CTGTGAGTCAGGGTCCACAC–3’ RPS14 Forward primer 5'-GGTCGGGTTCCGGAAGACGC-3' RPS14 Reverse primer 5'-ACCTCACCCCGCGGGTGTAG-3’ Table 5 RT-PCR primer sequences

132

Chapter 6: Discussion

BRCA1 is an important breast- and ovarian-specific cancer tumor suppressor gene (Miki et al., 1994). It plays a key role in several important processes such as transcription, cell cycle, DNA repair, and centrosome duplication (Parvin,

2004). Breast cancers are usually deficient in BRCA1. However, breast cancer is not a single disease, but a heterogenous disease consisting of multiple subtypes.

These include luminal A and B, basal, HER-2 enriched, and normal breast like group (Perou and Borresen-Dale, 2011). BRCA1 is predominantly mutated in basal or triple negative breast cancers while BRCA2 is mutated in luminal B tumors (Joosse, 2012; Roy et al., 2012). Both genes are mutated in hereditary and sporadic cases of breast cancer, although the latter displays similar characteristics to the former.

About 40% of hereditary breast cancers are characterized by mutations in

BRCA1/2 (Martin et al., 2001). The remaining cases of hereditary breast cancer that show normal BRCA1 and BRCA2 function are caused by mutations in low or moderate penetrance genes, in comparison with the high penetrance BRCA1 and BRCA2. (Ford et al., 1998; Gracia-Aznarez et al., 2013). These genes might involve genes known to play a role in the same processes as BRCA1 and

BRCA2 such as PALB2, BRIP1, ATM, CHK2 etc (Roy et al., 2012). These

133 account for only a small percentage of mutations in breast cancer, and there is still a large proportion of genes that remain unidentified, presumably due to this issue of low penetrance, and also because of low resolution methods of discovery. Therefore, these genes are called ‘missing BRCAs’. We hypothesized that these genes might be implicated in BRCA-regulated pathways. We set about identifying these genes using a bioinformatics approach. Advances in sequencing and array technology have generated vast databases of gene information such as the GEO and Stanford data (Barrett et al., 2007; Demeter et al., 2007). Analyses of these microarray datasets have revealed that genes/proteins exist as part of complex networks and co-ordinated functions in biological pathways (Agrawal, 2002; Jansen et al., 2002). Such genes may also share similar gene expression patterns across the datasets, making it possible to use gene expression data from these databases to identify the genes in such inter-relationships (Tornow and Mewes, 2003). Importantly, in spite of BRCA1 and BRCA2 being mutated in different subsets of breast cancer, they function together in the same pathways of DNA repair and genomic stability maintenance

(Joosse, 2012).

Similarly, we hypothesized that these missing ‘BRCAs’ can be identified using gene co-expression profiling. Also, such genes that share similar expression profiles with BRCA1 could function in the same pathway. Their proteins may also physically interact with each other, or BRCA1.

134

The primary aim of this study was to identify and validate these genes in BRCA1

–regulated pathways including DNA repair and centrosome duplication. Our second aim was to determine any direct/indirect interactions between them. In the near future, we plan to take the study to the next level and identify expression levels of these candidate genes in tumor tissue microarrays (TMAs).

Our co-expression analyses obtained read-outs of dozens of candidate genes that could be co-regulated with BRCA1. Of these, we were especially interested in the HDACs, which showed strong negative correlation with BRCA1, and

NUSAP1, which showed strong positive correlation with BRCA1. We were interested in the HDACs because we hypothesized that they might show opposite function to BRCA1, in the same pathways. Therefore, inhibiting them could offset the deleterious effects of BRCA1 mutations. Now, HDACs cause chromatin compaction, limiting access for transcription factors (Braunstein et al.,

1996), and potentially DNA repair factors to bind. This is why we expected to see that HDAC inhibition would increase homologous recombination levels. Other criteria we used for our selection also included if the genes were relatively unknown, had enzymatic function that could be targeted for drug design, or if their expression was deregulated in cancers.

The biological assays to be used to test the candidates included the homology directed repair (HDR) assay, and the centrosome duplication assay. BRCA1 is important for both of these processes. When BRCA1 is inhibited; normal function is impaired (Parvin, 2004). We wanted to test whether inhibition of our candidate

135 genes i.e. the HDACs would affect homologous recombination. A tissue-culture based homology-directed repair (HDR) assay was used in which repair of a double-stranded break, by homologous recombination, results in gene conversion of an inactive GFP allele to an active GFP gene (Kotian et al., 2011).

Homologous recombination events were then readily scored by measuring GFP positive cells by FACS. Treatment with HDAC inhibitors significantly reduced homologous recombination levels. Upon individual depletion of the HDACs from

1-11 using RNAi, we found that HDAC9 and HDAC10 significantly reduce homologous recombination levels. We also had corroborating evidence from a clonogenic repair assay utilizing mitomycin C as an interstrand cross linker. We found that treatment with the HDACi, as well HDAC9 and HDAC10 depletions sensitized cells to mitomycin C, and significantly reduced colony survival. We also checked to see if the effect of HDAC9/HDAC10 on homologous recombination was secondary to a cell cycle defect. We found that neither

HDAC9 nor HDAC10 depletions affected cell cycle progression in comparison to the control.

As HDAC depletion leads to open chromatin structure by hyperacetylating histones, we sought to identify the specific lysine residue being acetylated due to depletion of HDAC9 and HDAC10. Another study that was published while we were working on this, demonstrated that HDAC1 and HDAC2 depletions suppress DSB repair by hyperacetylating the lysine residue H3K56 (Miller et al.,

2010). They proposed that doing so, causes unfolding of chromatin, and the DSB

136 ends to be thrown apart from each other. This is what they claimed inhibited repair, and this is how HDAC1 and HDAC2 are important for NHEJ. Based on this evidence, we tried to identify if HDAC9 or HDAC10 depletions caused hyperacetylation of H3K56. We performed a western analysis on cell lysates from

HDAC9 and HDAC10 depleted cells, and tested them for levels of H3K56Ac.

However, we did not see any change in the levels of H3K56Ac as compared to the control.

We also performed LC/MS on histones purified from cells treated with the HDAC inhibitor drugs TSA and VPA. We wanted to test if we would see any specific acetylations. In comparison to the control, treatment with TSA caused very low levels of hyperacetylation. VPA caused noticeable hyperacetylation, but we were unable to pin-point any specific acetylations on lysine residues. It could be that we were only looking at purified histones, and that HDAC9 and HDAC10 might have non-histone targets. Thus, it would be interesting to perform the mass spectrometric analysis on HDAC9 and HDAC10 depleted cells again, but this time, it would be better to use whole cell protein extracts.

We also tested if HDAC9 co-localized with an important DSB repair protein i.e. γ-

H2AX, and saw that while it was abundantly present around the nuclear foci, it didn’t actually co-localize with them. Thus, we think it is possible that HDAC9 and

HDAC10 might catalyze an important deacetylation step at the site of the DSB that is important for normal repair to take place. In the future, we plan to perform a chromatin immunoprecipitation experiment using primers designed to flank the

137

I-SceI restriction endonuclease site in one of the GFP alleles in our recombinant

HeLa-DR cell line (Kotian et al., 2011). If either HDAC9 or HDAC10 or both bind to the DSB site generated by I-SceI enzyme, we will be able to confirm whether these HDACs act locally at DSB sites to exert their catalytic actions. We are also interested to perform immunoprecipitation experiments where we will attempt to pull down and identify proteins that bind to the HDACs.

It has been seen that HDAC expression is deregulated in cancers (Marks and

Xu, 2009). Also, there has been a lot of interest generated by the HDAC inhibitor drugs which exert cytostatic action, and repress the growth of cancer cells

(Garber, 2007a; Ginsburg et al., 1973; Han et al., 2000; Vigushin et al., 2001;

Zhu et al., 2004). However, not much is known about their exact mechanism of action, and we hope that our study will help shed light on that. Eventually, we plan to look at HDAC9 and HDAC10 expression levels in tumor tissue microarrays.

Our next candidate, NUSAP1, was also tested for a potential role in homologous recombination. We performed NUSAP1 depletion in our HDR assay, and found that it caused significant reduction in homologous recombination levels. We also tested the effect of NUSAP1 depletion on another DSB repair assay which tests efficacy of single strand annealing (Towler et al., 2013). We found that NUSAP1 depletion impaired single strand annealing levels as well. This tentatively indicates that NUSAP1 is important for DSB repair overall. We are in the process of confirming that by testing if NUSAP1 depletion affects non-homologous end-

138 joining as well. We are testing this using a repair assay that uses HeLa cells, stably integrated with a GFP reporter system analogous to the HDR assay.

We also tested the effects of NUSAP1 inhibition on another BRCA1-regulated pathway of centrosome duplication. We discovered that NUSAP1 depletion caused the formation of supernumerary centrosomes, which means that it is important for maintenance proper centrosomal numbers.

Interestingly, when BRCA1 was over-expressed in both the HDR and the centrosomal duplication assay, this reversed the defective phenotype seen upon

NUSAP1 depletion. This pointed toward an interaction between BRCA1 and

NUSAP1. When we performed NUSAP1 depletion, we saw that it decreased

BRCA1 protein levels in soluble extracts, and in chromatin fractions of the cell.

However, NUSAP1 depletion had no impact whatsoever on total SDS lysates, which meant that somehow NUSAP1 was affecting BRCA1 localization. We confirmed this when we discovered that NUSAP1 depletion also impaired recruitment of BRCA1 to DNA damage foci in response to ionizing radiation.

NUSAP1 depletion also affected RAD51 localization to the foci in response to IR.

The data strongly suggest that NUSAP1 is important for normal BRCA1 protein expression. Since, NUSAP1 depletion does not affect BRCA1 transcription or degradation; it might suppress BRCA1 protein translation. This can be tested by doing a cycloheximide chase experiment with NUSAP1 depleted cells, and analyzing protein levels of BRCA1, with and without damage.

139

We performed a western analysis, where we tracked patterns of BRCA1 and

NUSAP1 across the cell cycle using carefully synchronized HeLa cells. On comparing the levels, we saw that levels of both BRCA1 and NUSAP1 follow a similar progression through the cell cycle, increasing in S-phase gradually up to the G2/M-phase. At M-phase, BRCA1 levels begin to drop steadily, gradually disappearing by late mitosis. NUSAP1 levels lag behind BRCA1, and remain high in mitosis, before decreasing. This means that NUSAP1 could potentially have an effect upon BRCA1 levels as it lags behind through G2/M. Immunofluorescence studies were used to determine if BRCA1 and NUSAP co-localize. Co- localization was not observed, although both proteins were abundant enough in the cell not to exclude the possibility of an interaction.

Co-immunoprecipitation analysis also was unable to demonstrate that BRCA1 and NUSAP1 physically interact with each other. However, soluble cell lysates were used, which only include cytoplasmic and nucleoplasmic proteins. It is very possible that BRCA1 and NUSAP1 do interact, but perhaps they do so in a subcellular compartment like the nucleolus.

Recent evidence suggests that the nucleolus plays an important role in DDR i.e.it gets disrupted in response to IR (Rubbi and Milner, 2003). A comprehensive proteomics analysis revealed several repair proteins like TOP2A, BLM, ATR,

RAD50 etc to be a part of the nucleolus (Andersen et al., 2002). BRCA1, too, localizes to nucleoli and nucleolin in breast tumors, HeLa and MCF7 cell lines

(Tulchin et al., 2010; Tulchin et al., 1998). One study showed that RNF8 and

140

BRCA1 together translocate from the nucleolus to the nucleus in response to

DNA damage (Guerra-Rebollo et al., 2012). When a protein RPSA that binds

RNF8 was depleted, it prevented this translocation. Several repair proteins like

WRN, BLM, MUS81, and FEN1 have been shown to localize to the nucleus to form damage foci from PML bodies and the nucleolus in response to DNA damage (Gao et al., 2003; Guo et al., 2008; Karmakar and Bohr, 2005;

Yankiwski et al., 2000). Therefore, it might be possible that NUSAP1 controls localization of BRCA1 in response to DNA damage. It might be worthwhile to investigate if this is true, by carefully studying BRCA1 localization in the cell normally, and in response to damage signals.

This study has successfully identified and validated three novel genes: HDAC9,

HDAC10 and NUSAP1 in BRCA1-regulated pathways of homologous recombination and centrosomal duplication. Recent evidence has shown that both HDACs and NUSAP1 could be involved in the DNA damage response

(Emanuele et al., 2011; Miller et al., 2010). Also, both of these have been shown to be aberrantly expressed in cancers.

Cancers including breast cancer are heterogeneous diseases consisting of multiple subtypes, and new evidence has shown that a single tumor can consist of sub-populations of cells showing different genomic alterations (Fisher et al.,

2013). So it is possible that the same tumor may comprise BRCA1, BRCA2 as well as mutations in one or more of our candidate genes- HDAC9, HDAC10 and

NUSAP1. There have been several genomic and proteomics approaches toward

141 achieving biomarker discovery in the field of oncology. Some of the genomics approaches include use of next generation deep sequencing technologies. These have shown that tumors consist of heterogeneous sub-populations of cells

(Russnes et al., 2011). Such techniques have revealed various genetic aberrations in tumors, including copy number variations. Some other approaches include cGH, SAGE (serial analysis of gene expression), GWAS, DNA microarray profiling, and a combination of techniques that provide mRNA quantification. These include RT-PCR, RNA-seq, Northern blotting. Proteomics approaches to biomarker discovery include 2D-PAGE (polyacrylamide gel electrophoresis), various Mass Spectrometric techniques such as LC-MS (liquid chromatography mass spectrometry) and MALDITOF (matrix associated laser desorption/ionization), antibody microarrays, and cancer tissue microarrays.

However, both these sets of approaches are limited in nature because they each only provide information about alterations in either DNA, mRNA or protein. They absolutely do not provide any information about the functional phenotypes of such alterations. Therefore, a systems biology approach is more powerful because it integrates information about DNA, mRNA and protein alterations generated from multiple platforms (Trapé and Gonzalez-Angulo, 2012).

Downstream functional analyses can then be performed confirming the phenotypic role for genotypic alterations.

One such systems biology approach was performed recently, involving recent large sequencing efforts by an international consortium called TCGA (Cancer

142

Genome Atlas, 2012). These have brought forth mutation data from hundreds of breast cancer tumor samples. The study combined multiple, powerful platforms like gene expression analyses (mRNA and miRNA), DNA methylation arrays,

SNP arrays, reverse protein phase arrays to analyze DNA, mRNA and protein alterations genome wide. They then used multiplatform subtype discovery to apply this to the different subtypes of breast cancer. All of this data was compiled in a vast, comprehensive online database that is easily accessible to researchers.

Our approach uses gene expression analysis to identify co-expressed candidate genes which are then biologically validated in functional assays. Our approach is significant because it gives a functional significance to gene expression signatures. However, our method relies on a microarray based strategy which could be limited. Therefore, we could combine data from our analyses with the data from the TCGA database. This would provide a more reliable method of identifying, as well as, validating potential biomarkers.

We queried mutation data for our candidate genes HDAC9, HDAC10, and

NUSAP1 from this database. All three genes showed alterations, including several homozygous deletions in both breast and ovarian cancers (Figures 44 and 45). Importantly, the TCGA data showed that basal and luminal subsets of breast cancer show deletions in BRCA1 and NUSAP1.This is extremely significant because, while our data support a role for these factors in BRCA- regulated pathways, the TCGA results indicate that these genes are actually

143 mutated in a number of tumors. The prevalence is low, but it is possible that a small subset of tumors is due to these mutations.

Therefore, all these data strengthen a role for these genes as missing BRCAs in

BRCA-regulated pathways. All three- HDAC9, HDAC10 and NUSAP1 could act as biomarkers for specific subsets of breast and ovarian cancers, specifically in those cancers which show normal BRCA1 function. To further confirm a role for these genes as biomarkers, we want to evaluate their expression in breast tumor tissue microarrays.

Cancer is a disease that is driven by two types of mutations; ‘driver’ mutations that drive cancer development and ‘passenger’ mutations that just randomly accumulate in somatic cells, and which have no functional phenotype (Vandin et al., 2011). In order to identify if our candidate genes- HDAC9, HDAC10 and

NUSAP1 are ‘driver’ genes, we would have to perform gene expression analyses on a biological specimens taken from a large cohort of breast cancer patients.

Now, ‘driver’ mutations in a breast cancer subset consist of multiple mutations in genes that are involved in multiple biological pathways (Vandin et al., 2011).

These mutations can together drive cancer development. Once we identify mutations in these genes in these patient samples, we can apply it to various bioinformatics algorithms, and identify those mutations that have functional consequences. We can also identify which of these ‘driver’ genes function together in a biological pathway. Again, this has already been done for us by the

TCGA study, and we can simply integrate data from that and our study.

144

In order to validate a physiologically relevant role for these three candidates as missing BRCAs, further analysis would have to be performed. This would involve checking the expression levels of HDAC9, HDAC10, and NUSAP1 in various breast cancer cell lines as compared to normal breast cell lines, and in patient tumor samples versus control samples. We can also perform over-expression and depletion studies in normal versus breast cancer cell lines, and determine transformation efficiency and colony proliferation. In order investigate their tumorigenic potential; we can also study xenograft mouse models with depletion and over-expression of HDAC9, HDAC10 and NUSAP1, and knockout and/or transgenic mouse models to determine if these demonstrate tumor formation.

Thus, in this manner we can identify if HDAC9, HDAC10 and NUSAP1 are the

‘missing BRCAs’ in breast cancer.

145

Figure 44: TCGA database mutation analysis of HDAC9 and HDAC10

146

Figure 45: TCGA database mutation analysis of NUSAP1

147

Bibliography

Adimoolam, S., Sirisawad, M., Chen, J., Thiemann, P., Ford, J.M., and Buggy, J.J. (2007). HDAC inhibitor PCI-24781 decreases RAD51 expression and inhibits homologous recombination. Proceedings of the National Academy of Sciences of the United States of America 104, 19482-19487. Agrawal, H. (2002). Extreme Self-Organization in Networks Constructed from Gene Expression Data. Physical review letters 89. Andersen, J.S., Lam, Y.W., Leung, A.K.L., Ong, S.-E., Lyon, C.E., Lamond, A.I., and Mann, M. (2005). Nucleolar proteome dynamics. Nature 433, 77-83. Andersen, J.S., Lyon, C.E., Fox, A.H., Leung, A.K.L., Lam, Y.W., Steen, H., Mann, M., and Lamond, A.I. (2002). Directed proteomic analysis of the human nucleolus. Curr Biol 12, 1-11. Aoki, K., Ogata, Y., and Shibata, D. (2007). Approaches for extracting practical information from gene co-expression networks in plant biology. Plant Cell Physiol 48, 381-390. Barrett, T., Troup, D.B., Wilhite, S.E., Ledoux, P., Rudnev, D., Evangelista, C., Kim, I.F., Soboleva, A., Tomashevsky, M., and Edgar, R. (2007). NCBI GEO: mining tens of millions of expression profiles--database and tools update. Nucleic acids research 35, D760-765. Bartek, J., Lukas, J., and Bartkova, J. (2007). DNA damage response as an anticancer barrier: damage threshold and the concept of 'conditional haploinsufficiency'. Cell Cycle 6, 2344-2347. Baylin, S.B., and Ohm, J.E. (2006). Epigenetic gene silencing in cancer - a mechanism for early oncogenic pathway addiction? Nature reviews Cancer 6, 107-116. Beamish, H., Kedar, P., Kaneko, H., Chen, P., Fukao, T., Peng, C., Beresten, S., Gueven, N., Purdie, D., Lees-Miller, S., et al. (2002). Functional Link between BLM Defective in Bloom's Syndrome and the Ataxia-telangiectasia-mutated Protein, ATM. Journal of Biological Chemistry 277, 30515-30523. Bennardo, N., Cheng, A., Huang, N., and Stark, J.M. (2008). Alternative-NHEJ is a mechanistically distinct pathway of mammalian chromosome break repair. PLoS genetics 4, e1000110. Bhaskara, S., Knutson, S.K., Jiang, G., Chandrasekharan, M.B., Wilson, A.J., Zheng, S., Yenamandra, A., Locke, K., Yuan, J.L., Bonine-Summers, A.R., et al. Hdac3 is essential for the maintenance of chromatin structure and genome stability. Cancer Cell 18, 436-447.

148

Bhattacharyya, A., Ear, U.S., Koller, B.H., Weichselbaum, R.R., and Bishop, D.K. (2000). The breast cancer susceptibility gene BRCA1 is required for subnuclear assembly of Rad51 and survival following treatment with the DNA cross-linking agent cisplatin. The Journal of biological chemistry 275, 23899-23903. Bindra, R.S., Gibson, S.L., Meng, A., Westermark, U., Jasin, M., Pierce, A.J., Bristow, R.G., Classon, M.K., and Glazer, P.M. (2005). Hypoxia-induced down- regulation of BRCA1 expression by E2Fs. Cancer research 65, 11597-11604. Blander, G., and Guarente, L. (2004). The Sir2 family of protein deacetylases. Annu Rev Biochem 73, 417-435. Boffa, L.C., Vidali, G., Mann, R.S., and Allfrey, V.G. (1978). Suppression of histone deacetylation in vivo and in vitro by sodium butyrate. The Journal of biological chemistry 253, 3364-3366. Bogunovic, D., O'Neill, D.W., Belitskaya-Levy, I., Vacic, V., Yu, Y.L., Adams, S., Darvishian, F., Berman, R., Shapiro, R., Pavlick, A.C., et al. (2009). Immune profile and mitotic index of metastatic melanoma lesions enhance clinical staging in predicting patient survival. Proceedings of the National Academy of Sciences of the United States of America 106, 20429-20434. Bolden, J.E., Peart, M.J., and Johnstone, R.W. (2006). Anticancer activities of histone deacetylase inhibitors. Nat Rev Drug Discov 5, 769-784. Bork, P., Hofmann, K., Bucher, P., Neuwald, A.F., Altschul, S.F., and Koonin, E.V. (1997). A superfamily of conserved domains in DNA damage-responsive cell cycle checkpoint proteins. FASEB J 11, 68-76. Boulon, S., Westman, B.J., Hutten, S., Boisvert, F.-M., and Lamond, A.I. (2010). The nucleolus under stress. Mol Cell 40, 216-227. Braunstein, M., Sobel, R.E., Allis, C.D., Turner, B.M., and Broach, J.R. (1996). Efficient transcriptional silencing in Saccharomyces cerevisiae requires a heterochromatin histone acetylation pattern. Molecular and cellular biology 16, 4349-4356. Brzovic, P.S., Rajagopal, P., Hoyt, D.W., King, M.C., and Klevit, R.E. (2001). Structure of a BRCA1-BARD1 heterodimeric RING-RING complex. Nat Struct Biol 8, 833-837. Caestecker, K.W., and Van de Walle, G.R. (2013). The role of BRCA1 in DNA double-strand repair: Past and present. Experimental Cell Research 319, 575- 587. Cancer Genome Atlas, N. (2012). Comprehensive molecular portraits of human breast tumours. Nature 490, 61-70. Cario, G., Fetz, A., Bretscher, C., Moricke, A., Schrauder, A., Stanulla, M., and Schrappe, M. (2008). Initial leukemic gene expression profiles of patients with poor in vivo prednisone response are similar to those of blasts persisting under prednisone treatment in childhood acute lymphoblastic leukemia. Ann Hematol 87, 709-716. Chang, S., Young, B.D., Li, S., Qi, X., Richardson, J.A., and Olson, E.N. (2006). Histone deacetylase 7 maintains vascular integrity by repressing matrix metalloproteinase 10. Cell 126, 321-334. 149

Chapman, M.S., and Verma, I.M. (1996). Transcriptional activation by BRCA1. Nature 382, 678-679. Chauchereau, A., Mathieu, M., de Saintignon, J., Ferreira, R., Pritchard, L.L., Mishal, Z., Dejean, A., and Harel-Bellan, A. (2004). HDAC4 mediates transcriptional repression by the acute promyelocytic leukaemia-associated protein PLZF. Oncogene 23, 8777-8784. Chen, A., Kleiman, F.E., Manley, J.L., Ouchi, T., and Pan, Z.Q. (2002). Autoubiquitination of the BRCA1*BARD1 RING ubiquitin ligase. The Journal of biological chemistry 277, 22085-22092. Chen, D.T., Nasir, A., Culhane, A., Venkataramu, C., Fulp, W., Rubio, R., Wang, T., Agrawal, D., McCarthy, S.M., Gruidl, M., et al. (2010). Proliferative genes dominate malignancy-risk gene signature in histologically-normal breast tissue. Breast cancer research and treatment 119, 335-346. Chen, J. (2000). Ataxia telangiectasia-related protein is involved in the phosphorylation of BRCA1 following deoxyribonucleic acid damage. Cancer research 60, 5037-5039. Chou, H.Y., Wang, T.H., Lee, S.C., Hsu, P.H., Tsai, M.D., Chang, C.L., and Jeng, Y.M. (2011). Phosphorylation of NuSAP by Cdk1 regulates its interaction with microtubules in mitosis. Cell Cycle 10, 4083-4089. Ciccia, A., and Elledge, S.J. (2010). The DNA damage response: making it safe to play with knives. Mol Cell 40, 179-204. Citterio, E., Papait, R., Nicassio, F., Vecchi, M., Gomiero, P., Mantovani, R., Di Fiore, P.P., and Bonapace, I.M. (2004). Np95 Is a Histone-Binding Protein Endowed with Ubiquitin Ligase Activity. Molecular and cellular biology 24, 2526- 2535. Collins, N., McManus, R., Wooster, R., Mangion, J., Seal, S., Lakhani, S.R., Ormiston, W., Daly, P.A., Ford, D., Easton, D.F., et al. (1995). Consistent loss of the wild type allele in breast cancers from a family linked to the BRCA2 gene on chromosome 13q12-13. Oncogene 10, 1673-1675. Cornelis, R.S., Neuhausen, S.L., Johansson, O., Arason, A., Kelsell, D., Ponder, B.A., Tonin, P., Hamann, U., Lindblom, A., Lalle, P., et al. (1995). High allele loss rates at 17q12-q21 in breast and ovarian tumors from BRCAl-linked families. The Breast Cancer Linkage Consortium. Genes Chromosomes Cancer 13, 203-210. Cortez, D., Wang, Y., Qin, J., and Elledge, S.J. (1999). Requirement of ATM- dependent phosphorylation of brca1 in the DNA damage response to double- strand breaks. Science 286, 1162-1166. D'Assoro, A.B., Barrett, S.L., Folk, C., Negron, V.C., Boeneman, K., Busby, R., Whitehead, C., Stivala, F., Lingle, W.L., and Salisbury, J.L. (2002). Amplified centrosomes in breast cancer: a potential indicator of tumor aggressiveness. Breast cancer research and treatment 75, 25-34. Davies, S.L., North, P.S., Dart, A., Lakin, N.D., and Hickson, I.D. (2004). Phosphorylation of the Bloom's Syndrome Helicase and Its Role in Recovery from S-Phase Arrest. Molecular and cellular biology 24, 1279-1291.

150 de Ruijter, A.J., van Gennip, A.H., Caron, H.N., Kemp, S., and van Kuilenburg, A.B. (2003). Histone deacetylases (HDACs): characterization of the classical HDAC family. Biochem J 370, 737-749. DeFazio, L.G., Stansel, R.M., Griffith, J.D., and Chu, G. (2002). Synapsis of DNA ends by DNA-dependent protein kinase. EMBO J 21, 3192-3200. Delacote, F., and Lopez, B.S. (2008). Importance of the cell cycle phase for the choice of the appropriate DSB repair pathway, for genome stability maintenance: the trans-S double-strand break repair model. Cell Cycle 7, 33-38. Demeter, J., Beauheim, C., Gollub, J., Hernandez-Boussard, T., Jin, H., Maier, D., Matese, J.C., Nitzberg, M., Wymore, F., Zachariah, Z.K., et al. (2007). The Stanford Microarray Database: implementation of new analysis tools and open source release of software. Nucleic acids research 35, D766-770. Deng, C.X. (2001). Tumorigenesis as a consequence of genetic instability in Brca1 mutant mice. Mutat Res 477, 183-189. Deng, C.X., and Wang, R.H. (2003). Roles of BRCA1 in DNA damage repair: a link between development and cancer. Hum Mol Genet 12 Spec No 1, R113-123. Dokmanovic, M., Clarke, C., and Marks, P.A. (2007). Histone deacetylase inhibitors: overview and perspectives. Mol Cancer Res 5, 981-989. Doxsey, S. (2001). Re-evaluating centrosome function. Nat Rev Mol Cell Biol 2, 688-698. Emanuele, M.J., Elia, A.E.H., Xu, Q., Thoma, C.R., Izhar, L., Leng, Y., Guo, A., Chen, Y.-N., Rush, J., Hsu, P.W.-C., et al. (2011). Global identification of modular cullin-RING ligase substrates. Cell 147, 459-474. Ferrari-Toninelli, G., Bonini, S.A., Uberti, D., Buizza, L., Bettinsoli, P., Poliani, P.L., Facchetti, F., and Memo, M. (2010). Targeting Notch pathway induces growth inhibition and differentiation of neuroblastoma cells. Neuro-Oncology 12, 1231-1243. Fischer, D.D., Cai, R., Bhatia, U., Asselbergs, F.A.M., Song, C., Terry, R., Trogani, N., Widmer, R., Atadja, P., and Cohen, D. (2002). Isolation and characterization of a novel class II histone deacetylase, HDAC10. The Journal of biological chemistry 277, 6656-6666. Fisher, R., Pusztai, L., and Swanton, C. (2013). Cancer heterogeneity: implications for targeted therapeutics. British journal of cancer 108, 479-485. Ford, D., Easton, D.F., Bishop, D.T., Narod, S.A., and Goldgar, D.E. (1994). Risks of cancer in BRCA1-mutation carriers. Breast Cancer Linkage Consortium. Lancet 343, 692-695. Ford, D., Easton, D.F., Stratton, M., Narod, S., Goldgar, D., Devilee, P., Bishop, D.T., Weber, B., Lenoir, G., Chang-Claude, J., et al. (1998). Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. The Breast Cancer Linkage Consortium. Am J Hum Genet 62, 676-689. Friedberg, E.C. (2006). DNA repair and mutagenesis (Washington: ASM).

151

Gao, H., Chen, X.B., and McGowan, C.H. (2003). Mus81 endonuclease localizes to nucleoli and to regions of DNA damage in human S-phase cells. Mol Biol Cell 14, 4826-4834. Garber, K. (2007a). HDAC inhibitors overcome first hurdle. Nat Biotechnol 25, 17-19. Garber, K. (2007b). Hepatitis C: staying the course. Nat Biotechnol 25, 1379- 1381. Ginsburg, E., Salomon, D., Sreevalsan, T., and Freese, E. (1973). Growth inhibition and morphological changes caused by lipophilic acids in mammalian cells. Proceedings of the National Academy of Sciences of the United States of America 70, 2457-2461. Glozak, M.A., and Seto, E. (2007). Histone deacetylases and cancer. Oncogene 26, 5420-5432. Gottlicher, M., Minucci, S., Zhu, P., Kramer, O.H., Schimpf, A., Giavara, S., Sleeman, J.P., Lo Coco, F., Nervi, C., Pelicci, P.G., et al. (2001). Valproic acid defines a novel class of HDAC inhibitors inducing differentiation of transformed cells. EMBO J 20, 6969-6978. Gracia-Aznarez, F.J., Fernandez, V., Pita, G., Peterlongo, P., Dominguez, O., de la Hoya, M., Duran, M., Osorio, A., Moreno, L., Gonzalez-Neira, A., et al. (2013). Whole Exome Sequencing Suggests Much of Non-BRCA1/BRCA2 Familial Breast Cancer Is Due to Moderate and Low Penetrance Susceptibility Alleles. PloS one 8. Grawunder, U., Zimmer, D., Kulesza, P., and Lieber, M.R. (1998). Requirement for an interaction of XRCC4 with DNA ligase IV for wild-type V(D)J recombination and DNA double-strand break repair in vivo. The Journal of biological chemistry 273, 24708-24714. Guerra-Rebollo, M., Mateo, F., Franke, K., Huen, M.S., Lopitz-Otsoa, F., Rodriguez, M.S., Plans, V., and Thomson, T.M. (2012). Nucleolar exit of RNF8 and BRCA1 in response to DNA damage. Exp Cell Res 318, 2365-2376. Guidez, F., and Zelent, A. (2001). Role of nuclear receptor corepressors in leukemogenesis. Curr Top Microbiol Immunol 254, 165-185. Guo, Z., Qian, L., Liu, R., Dai, H., Zhou, M., Zheng, L., and Shen, B. (2008). Nucleolar localization and dynamic roles of flap endonuclease 1 in ribosomal DNA replication and damage repair. Molecular and cellular biology 28, 4310- 4319. Guruharsha, K.G., Kankel, M.W., and Artavanis-Tsakonas, S. (2012). The Notch signalling system: recent insights into the complexity of a conserved pathway. Nat Rev Genet 13, 654-666. Hackett, A.J., Smith, H.S., Springer, E.L., Owens, R.B., Nelson-Rees, W.A., Riggs, J.L., and Gardner, M.B. (1977). Two syngeneic cell lines from human breast tissue: the aneuploid mammary epithelial (Hs578T) and the diploid myoepithelial (Hs578Bst) cell lines. J Natl Cancer Inst 58, 1795-1806.

152

Haile, D.T., and Parvin, J.D. (1999). Activation of transcription in vitro by the BRCA1 carboxyl-terminal domain. The Journal of biological chemistry 274, 2113- 2117. Han, J.W., Ahn, S.H., Park, S.H., Wang, S.Y., Bae, G.U., Seo, D.W., Kwon, H.K., Hong, S., Lee, H.Y., Lee, Y.W., et al. (2000). Apicidin, a histone deacetylase inhibitor, inhibits proliferation of tumor cells via induction of p21WAF1/Cip1 and gelsolin. Cancer research 60, 6068-6074. Harper, J.W., and Elledge, S.J. (2007). The DNA damage response: ten years after. Molecular cell 28, 739-745. Hashizume, R., Fukuda, M., Maeda, I., Nishikawa, H., Oyake, D., Yabuki, Y., Ogata, H., and Ohta, T. (2001). The RING heterodimer BRCA1-BARD1 is a ubiquitin ligase inactivated by a breast cancer-derived mutation. The Journal of biological chemistry 276, 14537-14540. Hebbes, T.R., Clayton, A.L., Thorne, A.W., and Crane-Robinson, C. (1994). Core histone hyperacetylation co-maps with generalized DNase I sensitivity in the chicken beta-globin chromosomal domain. EMBO J 13, 1823-1830. Helleday, T., Lo, J., van Gent, D.C., and Engelward, B.P. (2007). DNA double- strand break repair: from mechanistic understanding to cancer treatment. DNA Repair (Amst) 6, 923-935. Hoeijmakers, J.H. (2001). Genome maintenance mechanisms for preventing cancer. Nature 411, 366-374. Hsu, L.C., and White, R.L. (1998). BRCA1 is associated with the centrosome during mitosis. Proceedings of the National Academy of Sciences of the United States of America 95, 12983-12988. Huang, S., Gulzar, Z.G., Salari, K., Lapointe, J., Brooks, J.D., and Pollack, J.R. (2012). Recurrent deletion of CHD1 in prostate cancer with relevance to cell invasiveness. Oncogene 31, 4164-4170. Huen, M.S.Y., Sy, S.M.H., and Chen, J. (2010). BRCA1 and its toolbox for the maintenance of genome integrity. Nat Rev Mol Cell Biol 11, 138-148. Ikeda, K., Steger, D.J., Eberharter, A., and Workman, J.L. (1999). Activation domain-specific and general transcription stimulation by native histone acetyltransferase complexes. Molecular and cellular biology 19, 855-863. Ishii, S., Kurasawa, Y., Wong, J., and Yu-Lee, L.Y. (2008). Histone deacetylase 3 localizes to the mitotic spindle and is required for kinetochore-microtubule attachment. Proc Natl Acad Sci U S A 105, 4179-4184. Ivanov, E.L., Sugawara, N., Fishman-Lobell, J., and Haber, J.E. (1996). Genetic requirements for the single-strand annealing pathway of double-strand break repair in Saccharomyces cerevisiae. Genetics 142, 693-704. Iyer, J., Moghe, S., Furukawa, M., and Tsai, M.-Y. (2011). What's Nu(SAP) in mitosis and cancer? Cell Signal 23, 991-998. Jackson, P.K., Eldridge, A.G., Freed, E., Furstenthal, L., Hsu, J.Y., Kaiser, B.K., and Reimann, J.D. (2000). The lore of the RINGs: substrate recognition and catalysis by ubiquitin ligases. Trends Cell Biol 10, 429-439.

153

Jackson, S.P., and Bartek, J. (2009). The DNA-damage response in human biology and disease. Nature 461, 1071-1078. Jansen, R., Greenbaum, D., and Gerstein, M. (2002). Relating whole-genome expression data with protein-protein interactions. Genome Res 12, 37-46. Jansen, R., Yu, H., Greenbaum, D., Kluger, Y., Krogan, N.J., Chung, S., Emili, A., Snyder, M., Greenblatt, J.F., and Gerstein, M. (2003). A Bayesian networks approach for predicting protein-protein interactions from genomic data. Science 302, 449-453. Jin, Y., Xu, X.L., Yang, M.C., Wei, F., Ayi, T.C., Bowcock, A.M., and Baer, R. (1997). Cell cycle-dependent colocalization of BARD1 and BRCA1 proteins in discrete nuclear domains. Proceedings of the National Academy of Sciences of the United States of America 94, 12075-12080. Johnson, R.D., and Jasin, M. (2000). Sister chromatid gene conversion is a prominent double-strand break repair pathway in mammalian cells. EMBO J 19, 3398-3407. Joosse, S.A. (2012). BRCA1 and BRCA2: a common pathway of genome protection but different breast cancer subtypes. Nature reviews Cancer 12, 372; author reply 372. Joughin, B.A., Green, D.F., and Tidor, B. (2005). Action-at-a-distance interactions enhance protein binding affinity. Protein science : a publication of the Protein Society 14, 1363-1369. Kadosh, D., and Struhl, K. (1998). Targeted recruitment of the Sin3-Rpd3 histone deacetylase complex generates a highly localized domain of repressed chromatin in vivo. Molecular and cellular biology 18, 5121-5127. Kais, Z., Barsky, S.H., Mathsyaraja, H., Zha, A., Ransburgh, D.J., He, G., Pilarski, R.T., Shapiro, C.L., Huang, K., and Parvin, J.D. (2011). KIAA0101 interacts with BRCA1 and regulates centrosome number. Mol Cancer Res 9, 1091-1099. Karmakar, P., and Bohr, V.A. (2005). Cellular dynamics and modulation of WRN protein is DNA damage specific. Mech Ageing Dev 126, 1146-1158. Khalil, I.G., and Hill, C. (2005). Systems biology for cancer. Curr Opin Oncol 17, 44-48. Kim, H., Huang, J., and Chen, J. (2007). CCDC98 is a BRCA1-BRCT domain- binding protein involved in the DNA damage response. Nat Struct Mol Biol 14, 710-715. King, M.C., Marks, J.H., and Mandell, J.B. (2003). Breast and ovarian cancer risks due to inherited mutations in BRCA1 and BRCA2. Science 302, 643-646. Ko, M.J., Murata, K., Hwang, D.S., and Parvin, J.D. (2006). Inhibition of BRCA1 in breast cell lines causes the centrosome duplication cycle to be disconnected from the cell cycle. Oncogene 25, 298-303. Kokkinakis, D.M., Liu, X., and Neuner, R.D. (2005). Modulation of cell cycle and gene expression in pancreatic tumor cell lines by methionine deprivation (methionine stress): implications to the therapy of pancreatic adenocarcinoma. Mol Cancer Ther 4, 1338-1348. 154

Kotian, S., Liyanarachchi, S., Zelent, A., and Parvin, J.D. (2011). Histone deacetylases 9 and 10 are required for homologous recombination. Journal of Biological Chemistry 286, 7722-7726. Kretschmer, C., Sterner-Kock, A., Siedentopf, F., Schoenegg, W., Schlag, P.M., and Kemmner, W. (2011). Identification of early molecular markers for breast cancer. Mol Cancer 10, 15. Kruhlak, M., Crouch, E.E., Orlov, M., Montano, C., Gorski, S.A., Nussenzweig, A., Misteli, T., Phair, R.D., and Casellas, R. (2007). The ATM repair pathway inhibits RNA polymerase I transcription in response to chromosome breaks. Nature 447, 730-734. Krum, S.A., la Rosa Dalugdugan, E.d., Miranda-Carboni, G.A., and Lane, T.F. (2010). BRCA1 Forms a Functional Complex with γ-H2AX as a Late Response to Genotoxic Stress. J Nucleic Acids 2010. Kuo, M.H., Zhou, J., Jambeck, P., Churchill, M.E., and Allis, C.D. (1998). Histone acetyltransferase activity of yeast Gcn5p is required for the activation of target genes in vivo. Genes Dev 12, 627-639. Kwon, S.H., Ahn, S.H., Kim, Y.K., Bae, G.U., Yoon, J.W., Hong, S., Lee, H.Y., Lee, Y.W., Lee, H.W., and Han, J.W. (2002). Apicidin, a histone deacetylase inhibitor, induces apoptosis and Fas/Fas ligand expression in human acute promyelocytic leukemia cells. The Journal of biological chemistry 277, 2073- 2080. Laherty, C.D., Yang, W.M., Sun, J.M., Davie, J.R., Seto, E., and Eisenman, R.N. (1997). Histone deacetylases associated with the mSin3 corepressor mediate mad transcriptional repression. Cell 89, 349-356. Lai, I.L., Lin, T.P., Yao, Y.L., Lin, C.Y., Hsieh, M.J., and Yang, W.M. Histone deacetylase 10 relieves repression on the melanogenic program by maintaining the deacetylation status of repressors. The Journal of biological chemistry 285, 7187-7196. Lane, A.A., and Chabner, B.A. (2009). Histone deacetylase inhibitors in cancer therapy. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 27, 5459-5468. Lauss, M., Kriegner, A., Vierlinger, K., Visne, I., Yildiz, A., Dilaveroglu, E., and Noehammer, C. (2008). Consensus genes of the literature to predict breast cancer recurrence. Breast cancer research and treatment 110, 235-244. Lindahl, T., and Barnes, D.E. (2000). Repair of endogenous DNA damage. Cold Spring Harb Symp Quant Biol 65, 127-133. Lingle, W.L., Lutz, W.H., Ingle, J.N., Maihle, N.J., and Salisbury, J.L. (1998). Centrosome hypertrophy in human breast tumors: implications for genomic stability and cell polarity. Proceedings of the National Academy of Sciences of the United States of America 95, 2950-2955. Lingle, W.L., and Salisbury, J.L. (2001). Methods for the analysis of centrosome reproduction in cancer cells. Methods Cell Biol 67, 325-336.

155

Liu, H., Hu, Q., D'Ercole A, J., and Ye, P. (2009). Histone deacetylase 11 regulates oligodendrocyte-specific gene expression and cell development in OL- 1 oligodendroglia cells. Glia 57, 1-12. Lorick, K.L., Jensen, J.P., Fang, S., Ong, A.M., Hatakeyama, S., and Weissman, A.M. (1999). RING fingers mediate ubiquitin-conjugating enzyme (E2)-dependent ubiquitination. Proceedings of the National Academy of Sciences of the United States of America 96, 11364-11369. Luo, R.X., Postigo, A.A., and Dean, D.C. (1998). Rb interacts with histone deacetylase to repress transcription. Cell 92, 463-473. Lyndaker, A.M., and Alani, E. (2009). A tale of tails: insights into the coordination of 3' end processing during homologous recombination. Bioessays 31, 315-321. Ma, Y., Pannicke, U., Schwarz, K., and Lieber, M.R. (2002). Hairpin opening and overhang processing by an Artemis/DNA-dependent protein kinase complex in nonhomologous end joining and V(D)J recombination. Cell 108, 781-794. Mak, T.W., Hakem, A., McPherson, J.P., Shehabeldin, A., Zablocki, E., Migon, E., Duncan, G.S., Bouchard, D., Wakeham, A., Cheung, A., et al. (2000). Brca1 required for T cell lineage development but not TCR loci rearrangement. Nat Immunol 1, 77-82. Mangia, A., Chiarappa, P., Tommasi, S., Chiriatti, A., Petroni, S., Schittulli, F., and Paradiso, A. (2008). Genetic heterogeneity by comparative genomic hybridization in BRCAx breast cancers. Cancer Genet Cytogenet 182, 75-83. Marchion, D.C., Bicaku, E., Turner, J.G., Schmitt, M.L., Morelli, D.R., and Munster, P.N. (2009). HDAC2 regulates chromatin plasticity and enhances DNA vulnerability. Mol Cancer Ther 8, 794-801. Marciniak, R.A., Lombard, D.B., Johnson, F.B., and Guarente, L. (1998). Nucleolar localization of the Werner syndrome protein in human cells. Proceedings of the National Academy of Sciences of the United States of America 95, 6887-6892. Marie, S.K.N., Okamoto, O.K., Uno, M., Hasegawa, A.P.G., Oba-Shinjo, S.M., Cohen, T., Camargo, A.A., Kosoy, A., Carlotti, C.G., Toledo, S., et al. (2008). Maternal embryonic leucine zipper kinase transcript abundance correlates with malignancy grade in human astrocytomas. Int J Cancer 122, 807-815. Marks, P.A., and Xu, W.S. (2009). Histone deacetylase inhibitors: Potential in cancer therapy. J Cell Biochem 107, 600-608. Martin, A.M., Blackwood, M.A., Antin-Ozerkis, D., Shih, H.A., Calzone, K., Colligon, T.A., Seal, S., Collins, N., Stratton, M.R., Weber, B.L., et al. (2001). Germline mutations in BRCA1 and BRCA2 in breast-ovarian families from a breast cancer risk evaluation clinic. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 19, 2247-2253. Matsuoka, S., Ballif, B.A., Smogorzewska, A., McDonald, E.R., 3rd, Hurov, K.E., Luo, J., Bakalarski, C.E., Zhao, Z., Solimini, N., Lerenthal, Y., et al. (2007). ATM and ATR substrate analysis reveals extensive protein networks responsive to DNA damage. Science 316, 1160-1166.

156

Meder, V.S., Boeglin, M., de Murcia, G., and Schreiber, V. (2005). PARP-1 and PARP-2 interact with nucleophosmin/B23 and accumulate in transcriptionally active nucleoli. J Cell Sci 118, 211-222. Miki, Y., Swensen, J., Shattuck-Eidens, D., Futreal, P.A., Harshman, K., Tavtigian, S., Liu, Q., Cochran, C., Bennett, L.M., Ding, W., et al. (1994). A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 266, 66-71. Miller, K.M., Tjeertes, J.V., Coates, J., Legube, G., Polo, S.E., Britton, S., and Jackson, S.P. (2010). Human HDAC1 and HDAC2 function in the DNA-damage response to promote DNA nonhomologous end-joining. Nat Struct Mol Biol 17, 1144-1151. Miyoshi, Y., Iwao, K., Egawa, C., and Noguchi, S. (2001). Association of centrosomal kinase STK15/BTAK mRNA expression with chromosomal instability in human breast cancers. Int J Cancer 92, 370-373. Monteiro, A.N., August, A., and Hanafusa, H. (1996). Evidence for a transcriptional activation function of BRCA1 C-terminal region. Proceedings of the National Academy of Sciences of the United States of America 93, 13595- 13599. Moynahan, M.E., Chiu, J.W., Koller, B.H., and Jasin, M. (1999). Brca1 controls homology-directed DNA repair. Molecular cell 4, 511-518. Moynahan, M.E., Cui, T.Y., and Jasin, M. (2001). Homology-directed dna repair, mitomycin-c resistance, and chromosome stability is restored with correction of a Brca1 mutation. Cancer research 61, 4842-4850. Nakanishi, K., Yang, Y.G., Pierce, A.J., Taniguchi, T., Digweed, M., D'Andrea, A.D., Wang, Z.Q., and Jasin, M. (2005). Human Fanconi anemia monoubiquitination pathway promotes homologous DNA repair. Proceedings of the National Academy of Sciences of the United States of America 102, 1110- 1115. Nie, J., Wang, H., He, F., and Huang, H. (2010). Nusap1 is essential for neural crest cell migration in zebrafish. Protein Cell 1, 259-266. Oldenburg, R.A., Meijers-Heijboer, H., Cornelisse, C.J., and Devilee, P. (2007). Genetic susceptibility for breast cancer: how many more genes to be found? Crit Rev Oncol Hematol 63, 125-149. Ouchi, M., Fujiuchi, N., Sasai, K., Katayama, H., Minamishima, Y.A., Ongusaha, P.P., Deng, C., Sen, S., Lee, S.W., and Ouchi, T. (2004). BRCA1 phosphorylation by Aurora-A in the regulation of G2 to M transition. The Journal of biological chemistry 279, 19643-19648. Pagano, M., Pepperkok, R., Verde, F., Ansorge, W., and Draetta, G. (1992). Cyclin A is required at two points in the human cell cycle. EMBO J 11, 961-971. Pan, L.N., Lu, J., and Huang, B. (2007). HDAC inhibitors: a potential new category of anti-tumor agents. Cell Mol Immunol 4, 337-343. Parvin, J.D. (2004). Overview of history and progress in BRCA1 research: the first BRCA1 decade. Cancer Biol Ther 3, 505-508.

157

Parvin, J.D. (2009). The BRCA1-dependent ubiquitin ligase, gamma-tubulin, and centrosomes. Environ Mol Mutagen 50, 649-653. Paull, T.T., Rogakou, E.P., Yamazaki, V., Kirchgessner, C.U., Gellert, M., and Bonner, W.M. (2000). A critical role for histone H2AX in recruitment of repair factors to nuclear foci after DNA damage. Curr Biol 10, 886-895. Perou, C.M., and Borresen-Dale, A.L. (2011). Systems biology and genomics of breast cancer. Cold Spring Harbor perspectives in biology 3. Petrie, K., Guidez, F., Howell, L., Healy, L., Waxman, S., Greaves, M., and Zelent, A. (2003). The histone deacetylase 9 gene encodes multiple protein isoforms. The Journal of biological chemistry 278, 16059-16072. Pharoah, P.D., Antoniou, A., Bobrow, M., Zimmern, R.L., Easton, D.F., and Ponder, B.A. (2002). Polygenic susceptibility to breast cancer and implications for prevention. Nature genetics 31, 33-36. Pierce, A.J., Johnson, R.D., Thompson, L.H., and Jasin, M. (1999). XRCC3 promotes homology-directed repair of DNA damage in mammalian cells. Genes Dev 13, 2633-2638. Pines, J., and Hunter, T. (1989). Isolation of a human cyclin cDNA: evidence for cyclin mRNA and protein regulation in the cell cycle and for interaction with p34cdc2. Cell 58, 833-846. Pujana, M.A., Han, J.D., Starita, L.M., Stevens, K.N., Tewari, M., Ahn, J.S., Rennert, G., Moreno, V., Kirchhoff, T., Gold, B., et al. (2007). Network modeling links breast cancer susceptibility and centrosome dysfunction. Nature genetics 39, 1338-1349. Raemaekers, T., Ribbeck, K., Beaudouin, J., Annaert, W., Van Camp, M., Stockmans, I., Smets, N., Bouillon, R., Ellenberg, J., and Carmeliet, G. (2003). NuSAP, a novel microtubule-associated protein involved in mitotic spindle organization. J Cell Biol 162, 1017-1029. Rancourt, A., and Satoh, M.S. (2009). Delocalization of nucleolar poly(ADP- ribose) polymerase-1 to the nucleoplasm and its novel link to cellular sensitivity to DNA damage. DNA Repair (Amst) 8, 286-297. Ransburgh, D.J., Chiba, N., Ishioka, C., Toland, A.E., and Parvin, J.D. (2010). Identification of breast tumor mutations in BRCA1 that abolish its function in homologous DNA recombination. Cancer research 70, 988-995. Rao, R., Fiskus, W., Yang, Y., Lee, P., Joshi, R., Fernandez, P., Mandawat, A., Atadja, P., Bradner, J.E., and Bhalla, K. (2008). HDAC6 inhibition enhances 17- AAG--mediated abrogation of hsp90 chaperone function in human leukemia cells. Blood 112, 1886-1893. Remboutsika, E., Lutz, Y., Gansmuller, A., Vonesch, J.L., Losson, R., and Chambon, P. (1999). The putative nuclear receptor mediator TIF1alpha is tightly associated with euchromatin. J Cell Sci 112 ( Pt 11), 1671-1683. Ribbeck, K., Groen, A.C., Santarella, R., Bohnsack, M.T., Raemaekers, T., Kocher, T., Gentzel, M., Gorlich, D., Wilm, M., Carmeliet, G., et al. (2006). NuSAP, a mitotic RanGTP target that stabilizes and cross-links microtubules. Mol Biol Cell 17, 2646-2660. 158

Ribbeck, K., Raemaekers, T., Carmeliet, G., and Mattaj, I.W. (2007). A role for NuSAP in linking microtubules to mitotic chromosomes. Curr Biol 17, 230-236. Rogakou, E.P., Pilch, D.R., Orr, A.H., Ivanova, V.S., and Bonner, W.M. (1998). DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139. The Journal of biological chemistry 273, 5858-5868. Roy, R., Chun, J., and Powell, S.N. (2012). BRCA1 and BRCA2: different roles in a common pathway of genome protection. Nature reviews Cancer 12, 68-78. Rubbi, C.P., and Milner, J. (2003). Disruption of the nucleolus mediates stabilization of p53 in response to DNA damage and other stresses. EMBO J 22, 6068-6077. Ruffner, H., and Verma, I.M. (1997). BRCA1 is a cell cycle-regulated nuclear phosphoprotein. Proceedings of the National Academy of Sciences of the United States of America 94, 7138-7143. Russnes, H.G., Navin, N., Hicks, J., and Borresen-Dale, A.-L. (2011). Insight into the heterogeneity of breast cancer through next-generation sequencing. The Journal of Clinical Investigation 121, 3810-3818. Ryu, B., Kim, D.S., Deluca, A.M., and Alani, R.M. (2007). Comprehensive expression profiling of tumor cell lines identifies molecular signatures of melanoma progression. PloS one 2, e594. Sankaran, S., Crone, D.E., Palazzo, R.E., and Parvin, J.D. (2007). BRCA1 regulates gamma-tubulin binding to centrosomes. Cancer Biol Ther 6, 1853- 1857. Sankaran, S., Starita, L.M., Simons, A.M., and Parvin, J.D. (2006). Identification of domains of BRCA1 critical for the ubiquitin-dependent inhibition of centrosome function. Cancer research 66, 4100-4107. Satow, R., Shitashige, M., Kanai, Y., Takeshita, F., Ojima, H., Jigami, T., Honda, K., Kosuge, T., Ochiya, T., Hirohashi, S., et al. (2010). Combined functional genome survey of therapeutic targets for hepatocellular carcinoma. Clinical cancer research : an official journal of the American Association for Cancer Research 16, 2518-2528. Schlegel, B.P., Starita, L.M., and Parvin, J.D. (2003). Overexpression of a protein fragment of RNA helicase A causes inhibition of endogenous BRCA1 function and defects in ploidy and cytokinesis in mammary epithelial cells. Oncogene 22, 983-991. Scully, R., Anderson, S.F., Chao, D.M., Wei, W., Ye, L., Young, R.A., Livingston, D.M., and Parvin, J.D. (1997). BRCA1 is a component of the RNA polymerase II holoenzyme. Proceedings of the National Academy of Sciences of the United States of America 94, 5605-5610. Smith, J., Tho, L.M., Xu, N., and Gillespie, D.A. (2010). The ATM-Chk2 and ATR- Chk1 pathways in DNA damage signaling and cancer. Adv Cancer Res 108, 73- 112. Snouwaert, J.N., Gowen, L.C., Latour, A.M., Mohn, A.R., Xiao, A., DiBiase, L., and Koller, B.H. (1999). BRCA1 deficient embryonic stem cells display a decreased homologous recombination frequency and an increased frequency of 159 non-homologous recombination that is corrected by expression of a brca1 transgene. Oncogene 18, 7900-7907. Starita, L.M., Machida, Y., Sankaran, S., Elias, J.E., Griffin, K., Schlegel, B.P., Gygi, S.P., and Parvin, J.D. (2004). BRCA1-dependent ubiquitination of gamma- tubulin regulates centrosome number. Molecular and cellular biology 24, 8457- 8466. Starita, L.M., and Parvin, J.D. (2003). The multiple nuclear functions of BRCA1: transcription, ubiquitination and DNA repair. Curr Opin Cell Biol 15, 345-350. Stark, J.M., Pierce, A.J., Oh, J., Pastink, A., and Jasin, M. (2004). Genetic steps of mammalian homologous repair with distinct mutagenic consequences. Molecular and cellular biology 24, 9305-9316. Staufenbiel, M., and Deppert, W. (1983). Different structural systems of the nucleus are targets for SV40 large T antigen. Cell 33, 173-181. Struewing, J.P., Hartge, P., Wacholder, S., Baker, S.M., Berlin, M., McAdams, M., Timmerman, M.M., Brody, L.C., and Tucker, M.A. (1997). The risk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N Engl J Med 336, 1401-1408. Stuart, J.E., Lusis, E.A., Scheck, A.C., Coons, S.W., Lal, A., Perry, A., and Gutmann, D.H. (2011). Identification of gene markers associated with aggressive meningioma by filtering across multiple sets of gene expression arrays. J Neuropathol Exp Neurol 70, 1-12. Thompson, L.H. (2012). Recognition, signaling, and repair of DNA double-strand breaks produced by ionizing radiation in mammalian cells: the molecular choreography. Mutat Res 751, 158-246. Tibbetts, R.S., Cortez, D., Brumbaugh, K.M., Scully, R., Livingston, D., Elledge, S.J., and Abraham, R.T. (2000). Functional interactions between BRCA1 and the checkpoint kinase ATR during genotoxic stress. Genes Dev 14, 2989-3002. Tornow, S., and Mewes, H.W. (2003). Functional modules by relating protein interaction networks and gene expression. Nucleic acids research 31, 6283- 6289. Towler, W.I., Zhang, J., Ransburgh, D.J.R., Toland, A.E., Ishioka, C., Chiba, N., and Parvin, J.D. (2013). Analysis of BRCA1 Variants in Double-Strand Break Repair by Homologous Recombination and Single-Strand Annealing. Hum Mutat 34, 439-445. Trapé, A.P., and Gonzalez-Angulo, A.M. (2012). Breast Cancer and Metastasis: On the Way Toward Individualized Therapy. Cancer Genomics - Proteomics 9, 297-310. Tulchin, N., Chambon, M., Juan, G., Dikman, S., Strauchen, J., Ornstein, L., Billack, B., Woods, N.T., and Monteiro, A.N. (2010). BRCA1 protein and nucleolin colocalize in breast carcinoma tissue and cancer cell lines. Am J Pathol 176, 1203-1214. Tulchin, N., Ornstein, L., Bleiweiss, I.J., Mendlowitz, M., Weinstein, E., and Dikman, S.H. (1998). Nucleolar localization of BRCA1 protein in human breast cancer. Int J Oncol 13, 513-518. 160

Uhl, M., Csernok, A., Aydin, S., Kreienberg, R., Wiesmuller, L., and Gatz, S.A. Role of SIRT1 in homologous recombination. DNA Repair (Amst) 9, 383-393. Urbich, C., Rossig, L., Kaluza, D., Potente, M., Boeckel, J.N., Knau, A., Diehl, F., Geng, J.G., Hofmann, W.K., Zeiher, A.M., et al. (2009). HDAC5 is a repressor of angiogenesis and determines the angiogenic gene expression pattern of endothelial cells. Blood 113, 5669-5679. Valerie, K., and Povirk, L.F. (2003). Regulation and mechanisms of mammalian double-strand break repair. Oncogene 22, 5792-5812. van Attikum, H., and Gasser, S.M. (2009). Crosstalk between histone modifications during the DNA damage response. Trends Cell Biol 19, 207-217. Vanden Bosch, A., Raemaekers, T., Denayer, S., Torrekens, S., Smets, N., Moermans, K., Dewerchin, M., Carmeliet, P., and Carmeliet, G. (2010). NuSAP is essential for chromatin-induced spindle formation during early embryogenesis. Journal of cell science 123, 3244-3255. Vandin, F., Upfal, E., and Raphael, B. (2011). Finding Driver Pathways in Cancer: Models and Algorithms. In Algorithms in Bioinformatics, T. Przytycka, and M.-F. Sagot, eds. (Springer Berlin Heidelberg), pp. 314-325. Verbakel, W., Carmeliet, G., and Engelborghs, Y. (2011). SAP-like domain in nucleolar spindle associated protein mediates mitotic chromosome loading as well as interphase chromatin interaction. Biochem Biophys Res Commun 411, 732-737. Vidal, M. (2001). A biological atlas of functional maps. Cell 104, 333-339. Vigushin, D.M., Ali, S., Pace, P.E., Mirsaidi, N., Ito, K., Adcock, I., and Coombes, R.C. (2001). Trichostatin A is a histone deacetylase inhibitor with potent antitumor activity against breast cancer in vivo. Clinical cancer research : an official journal of the American Association for Cancer Research 7, 971-976. Wadia, P.P., Coram, M., Armstrong, R.J., Mindrinos, M., Butte, A.J., and Miklos, D.B. (2010). Antibodies specifically target AML antigen NuSAP1 after allogeneic bone marrow transplantation. Blood 115, 2077-2087. Waltregny, D., Glenisson, W., Tran, S.L., North, B.J., Verdin, E., Colige, A., and Castronovo, V. (2005). Histone deacetylase HDAC8 associates with smooth muscle alpha-actin and is essential for smooth muscle cell contractility. FASEB J 19, 966-968. Wang, B., Matsuoka, S., Ballif, B.A., Zhang, D., Smogorzewska, A., Gygi, S.P., and Elledge, S.J. (2007). Abraxas and RAP80 form a BRCA1 protein complex required for the DNA damage response. Science 316, 1194-1198. Weigelt, B., Horlings, H.M., Kreike, B., Hayes, M.M., Hauptmann, M., Wessels, L.F., de Jong, D., Van de Vijver, M.J., Van't Veer, L.J., and Peterse, J.L. (2008). Refinement of breast cancer classification by molecular characterization of histological special types. The Journal of pathology 216, 141-150. Westermark, U.K., Reyngold, M., Olshen, A.B., Baer, R., Jasin, M., and Moynahan, M.E. (2003). BARD1 participates with BRCA1 in homology-directed repair of chromosome breaks. Molecular and cellular biology 23, 7926-7936.

161

Weterings, E., and Chen, D.J. (2008). The endless tale of non-homologous end- joining. Cell Res 18, 114-124. Wilson, A.J., Byun, D.S., Popova, N., Murray, L.B., L'Italien, K., Sowa, Y., Arango, D., Velcich, A., Augenlicht, L.H., and Mariadason, J.M. (2006). Histone deacetylase 3 (HDAC3) and other class I HDACs regulate colon cell maturation and p21 expression and are deregulated in human colon cancer. The Journal of biological chemistry 281, 13548-13558. Wooster, R., Bignell, G., Lancaster, J., Swift, S., Seal, S., Mangion, J., Collins, N., Gregory, S., Gumbs, C., and Micklem, G. (1995). Identification of the breast cancer susceptibility gene BRCA2. Nature 378, 789-792. Wu, L.C., Wang, Z.W., Tsan, J.T., Spillman, M.A., Phung, A., Xu, X.L., Yang, M.C., Hwang, L.Y., Bowcock, A.M., and Baer, R. (1996). Identification of a RING protein that can interact in vivo with the BRCA1 gene product. Nature genetics 14, 430-440. Xie, P., Li, L., Xing, G., Tian, C., Yin, Y., He, F., and Zhang, L. (2011). ATM- mediated NuSAP phosphorylation induces mitotic arrest. Biochemical and biophysical research communications 404, 413-418. Xu, B., Kim, S.T., Lim, D.S., and Kastan, M.B. (2002). Two molecularly distinct G(2)/M checkpoints are induced by ionizing irradiation. Molecular and cellular biology 22, 1049-1059. Xu, W.S., Parmigiani, R.B., and Marks, P.A. (2007). Histone deacetylase inhibitors: molecular mechanisms of action. Oncogene 26, 5541-5552. Xu, X., Weaver, Z., Linke, S.P., Li, C., Gotay, J., Wang, X.W., Harris, C.C., Ried, T., and Deng, C.X. (1999). Centrosome amplification and a defective G2-M cell cycle checkpoint induce genetic instability in BRCA1 exon 11 isoform-deficient cells. Molecular cell 3, 389-395. Yankiwski, V., Marciniak, R.A., Guarente, L., and Neff, N.F. (2000). Nuclear structure in normal and Bloom syndrome cells. Proceedings of the National Academy of Sciences of the United States of America 97, 5214-5219. Yarden, R.I., and Papa, M.Z. (2006). BRCA1 at the crossroad of multiple cellular pathways: approaches for therapeutic interventions. Mol Cancer Ther 5, 1396- 1404. Yarden, R.I., Pardo-Reoyo, S., Sgagias, M., Cowan, K.H., and Brody, L.C. (2002). BRCA1 regulates the G2/M checkpoint by activating Chk1 kinase upon DNA damage. Nature genetics 30, 285-289. Yoshida, T., Gan, Q., Shang, Y., and Owens, G.K. (2007). Platelet-derived growth factor-BB represses smooth muscle cell marker genes via changes in binding of MKL factors and histone deacetylases to their promoters. Am J Physiol Cell Physiol 292, C886-895. Yuan, Z., and Seto, E. (2007). A functional link between SIRT1 deacetylase and NBS1 in DNA damage response. Cell Cycle 6, 2869-2871. Zhang, C.L., McKinsey, T.A., Chang, S., Antos, C.L., Hill, J.A., and Olson, E.N. (2002). Class II histone deacetylases act as signal-responsive repressors of cardiac hypertrophy. Cell 110, 479-488. 162

Zhang, J., Lu, K., Xiang, Y., Islam, M., Kotian, S., Kais, Z., Lee, C., Arora, M., Liu, H.-W., Parvin, J.D., et al. (2012). Weighted frequent gene co-expression network mining to identify genes involved in genome stability. PLoS Comput Biol 8. Zhong, Q., Boyer, T.G., Chen, P.L., and Lee, W.H. (2002). Deficient nonhomologous end-joining activity in cell-free extracts from Brca1-null fibroblasts. Cancer research 62, 3966-3970. Zhou, X., Marks, P.A., Rifkind, R.A., and Richon, V.M. (2001). Cloning and characterization of a histone deacetylase, HDAC9. Proceedings of the National Academy of Sciences of the United States of America 98, 10572-10577. Zhu, P., Martin, E., Mengwasser, J., Schlag, P., Janssen, K.-P., and Gottlicher, M. (2004). Induction of HDAC2 expression upon loss of APC in colorectal tumorigenesis. Cancer Cell 5, 455-463.

163

Appendix: Bioinformatic Data Analysis

This consists of tables, listing candidate genes for each dataset, obtained after running the PCC testing and Intersect codes. The tables only show the top 30 representative candidate gene hits for each dataset, and show correlation of each candidate with the reference genes – BRCA1, BRCA2 and BARD1.Some datasets did not contain much gene expression data, and so list only a few candidates.

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192