DEK Is a Homologous Recombination DNA Repair Protein and Prognostic Marker for a Subset of Oropharyngeal Carcinomas

DEK is a homologous recombination DNA repair protein and prognostic marker for a subset of oropharyngeal carcinomas

A dissertation submitted to the Graduate School Of the University of Cincinnati In partial fulfillment of the requirements to the degree of

Doctor of Philosophy (Ph.D.)

in the Department of Cancer and Cell Biology of the College of Medicine March 29, 2017

ERIC ALAN SMITH

B.S Rose-Hulman Institute of Technology, 2010

Dissertation Committee: Susanne I. Wells, Ph.D. (Chair) Paul R. Andreassen, Ph.D. Nancy Ratner, Ph.D. Peter J. Stambrook , Ph.D. Kathryn A. Wikenheiser-Brokamp, M.D., Ph.D.

Abstract

The DEK oncogene is currently under investigation as a therapeutic target and clinical biomarker for tumor progression, chemotherapy resistance and poor outcomes for multiple types of malignancies. With regard to most cancer types, the degree of DEK overexpression correlates with higher stage tumors and worse patient survival, marking this molecule as a promising prognostic factor. Prior to this work, the utility of DEK as a biomarker had not been assessed in oropharyngeal squamous cell carcinoma

(OPSCC), an aggressive disease characterized by poor survival and high rates of treatment comorbidities.

As discussed in chapter 1, OPSCC is comprised of two subtypes based on the presence or absence of human papillomavirus (HPV) infection. In general, HPV+

OPSCCs have improved therapy response and an overall better prognosis than their

HPV- counterparts. This treatment sensitivity may be due in part to the HPV oncogenes, which inactivate and degrade tumor suppressors as discussed in chapter 2, removing the need for the development of mutations that promote radiation and chemotherapy resistance. Due to improved therapy response, clinical trials are now testing de- intensified regimens for the HPV+ OPSCC subset to prevent unnecessary morbidity from aggressive treatment. However, there is a population of HPV+ tumors that have poor outcomes from standard treatment and may not respond to de-intensified therapy.

Since current staging tools do not fully account for therapy response, it is unclear which

HPV+ OPSCC patients will not benefit from de-intensified treatment, but biomarkers, such as DEK, may play a critical role in identifying such patients.

Chapter 3 describes the biochemical and cellular functions of the DEK oncogene and focuses on how DEK can promote chemotherapy/radiation resistance through DNA repair. Chapter 4 provides new experimental evidence demonstrating that DEK is strictly required for homologous recombination (HR) repair of DNA breaks. DEK was found to be required for γH2AX activation after treatment with ionizing radiation and formed a complex with RAD51, the essential HR recombinase. Based on the lack of association with BRCA1, it is hypothesized that DEK may assist RAD51 in strand invasion. While a previous study has linked the chemotherapy resistance of DEK to NHEJ function, it was instead determined that the loss of DEK is synthetic lethal with NHEJ inactivation via

DNA-PK inhibitors. This suggests that cells rely on remaining NHEJ for survival in the absence of DEK-mediated HR repair.

In Chapter 5, the prognostic value of DEK in predicting outcomes in OPSCC was determined using a retrospective case study with tumor resections from 194 patients.

DEK expression was correlated with higher volume tumors and an increased risk of death in HPV+ OPSCC, but no association was found in HPV- disease. These findings indicate that DEK can be further developed as a biomarker for poor survival in HPV+

OPSCC and mark HPV- OPSCC as one of the few solid tumors where DEK is not predictive of survival outcomes.

In summary, this dissertation identified that DEK is required for HR repair of

DSBs and may be clinically useful as a prognostic marker for HPV+ OPSCC.

iii

Preface

The following work in this dissertation has been published in the following peer- reviewed journals:

Smith E.A., Kumar, B., Kakajan, K., Smith, S.M., Brown N.V., Zhao, S., Kumar, P., Teknos, T.N., Wells, S.I. “DEK associates with tumor stage and outcome in HPV16 positive oropharyngeal squamous cell carcinoma.” Published in Oncotarget

Smith E.A., E.A., Gole, G., Willis, N.A., Soria, R., Starnes L.M., Krumpelbeck, E.F., Jegga, A.G., Abdullah M.A,, Guo, H., Meetei, A.R., Andreassen, P.R., Kappes, F., Privette Vinnedge, L.M., Daniel, J.A., Scully, R., Wiesmüller, L. Wells, S.I. “DEK is required for homologous recombination repair of DNA breaks.” Published in Scientific Reports

Smith E.A., Matrka, M.C.; Wells, S.I. “HPV Virology: Cellular Targets of HPV Oncogenes and Transformation” in “HPV Virology: Cellular Targets of HPV Oncogenes and Transformation” in Human Papillomavirus (HPV)-Associated Oropharyngeal Cancer; Miller D.L.; Stack, S.M., Eds.; Springer International Publishing AG Switzerland, 2015; pp 69-101.

Acknowledgements

Out of all of the sections of this dissertation, this is probably the hardest for me to write. It is not because of lack of content, but rather there are too many people to thank, and words hardly suffice to state their importance in my PhD journey. Still, here I am giving it a go anyway!

First, I must thank my labmates and all of the support they have given me through the years. We are a rather unique group: eager, rambunctious, loud, jocular, passionate, and fiercely loyal. When you join The Wells lab, it is not just a work group.

You join a family. We play pranks on each other, joke alongside each other, and, like any family, argue amongst each other. Despite a whole turnover of the lab besides

Susa (indispensable) and Marie (wishes she was not indispensable), I am happy to say that the spirit that first attracted me to the lab still remains, although we may actually be somewhat quieter now! One of the aspects that I will always cherish about the lab is that it is a lot like Hotel California: you can always join the lab, but you can never leave. I am always seeing and reconnecting with old alumni, even the original personnel in Susa’s laboratory. Our comradery is what defines the lab, and knowing that you always have their support is a feeling I cannot put into words.

Even among a group as awesome as The Wells lab, there are special people that stand out. I want to specifically mention three here. The first is Beth Hoskins. She was the first person who truly believed in me, showing that belief by humbling me until I realized how little I knew and supporting me when I came to doubt myself. I especially want to thank Allie Adams for giving me a chance to dust myself off, train hard, and start anew. I do not know where I would be without her support and guidance. Of course, I

vi should probably also thank my boss, Susa. If I am lucky enough to become a PI in the future, I sincerely hope that I am half as thoughtful and devoted to my work and my people as Susa. She embodies that type of leadership that I value most:

Transformational Leadership. Where authoritarian leaders bark orders and servant leaders lead through example, transformational leaders inspire their compatriots to push themselves beyond their normal limits and accomplish great things (or if you are me, to push myself beyond my limits and end up in the ER, but that’s a story for a different time!). One of the things I admire most about Susa and may have the hardest difficulty replicating is her steadfast belief in her students and post-docs. She never gives up on people and believes in their potential, even when others would have long dismissed them. Like many of her students, I needed a mentor who defines her success by the lifelong accomplishments of her protégés, and I was fortunate to have been able to work for her. I will always be in her debt. Hopefully, I’ll get my just deserts when I also have a student who knocks on my door every day to give me a status update.

One of the things I have learned from Susa is that collaboration is the life-blood of science, and that is a skill I have learned to hone during my time in her lab. From our peers at Children’s and UC, to my dear friends Ferdi, Haihong, and Chris in Aachen

Uniklinic and everyone in-between, I am thankful to have been able to work with and study alongside some of the brightest minds in academia. Special thanks go to Emma

Lou and Dr. Robert Cardell for their fellowship award. Without your support, I could not have travelled to study DEK biochemistry in Ferdi’s lab, attend two international meetings, or have met two wonderful and enthusiastic emeritus professors.

vii

To my thesis committee, I thank you for keeping me on track and my feet to the fire. Your advice has always been well-appreciated. In particular, I want to thank Paul for all of his guidance in the past several years. He was a second mentor to me, and we would spend hours after committee meetings and randomly in his office discussing hypotheses, politics, and random ideas about how in the world DEK actually functions.

One day, I hope to visit Seattle long enough for me to see everything that he has recommended during our talks.

Most of all, I would like to thank my wonderful family for all of their love and support. My mom was the trailblazer, obtaining the first PhD in my family’s history during my first year of graduate school, and she has served as a constant source of guidance and understanding. I also greatly appreciate my sister for her unwavering faith in my abilities and my dad for keeping me humble and grounded. I am also truly blessed to have awesome in-laws, whose doors are always open to me like a port in the storm.

Lastly and most importantly, I want to thank the love of my life, Leah Howard. She is my best friend, the person I rely on and trust the most, and the stabilizing force in this wild ride of graduate school. She has shared with me the crescendos and lows of this period of my life, and her honesty and frank advice has helped keep the myriad of lab challenges in perspective. I have been incredibly fortunate to have someone like her by my side during the journey, and I cherish her dearly.

viii

Table of Contents Abstract ...... ii Preface ...... v Acknowledgements ...... vi List of Tables ...... xii List of Figures ...... xiii Abbreviations ...... xiv Chapter 1: Oropharyngeal squamous cell carcinoma ...... 1 Introduction ...... 2 Anatomy of the oropharynx ...... 2 The epidemiological shift in OPSCC caused by HPV ...... 3 Redefining the staging system for HPV+ OPSCC ...... 5 Current treatment strategies for OPSCC ...... 6 Conclusions...... 9 Chapter 2: HPV Virology: Cellular Targets of HPV Oncogenes and Transformation ...... 13 Abstract ...... 14 Introduction ...... 15 Targets of HPV Viral Oncogenes and the HPV E2 Oncogene Repressor ...... 17 E6 Oncogene ...... 17 E7 Oncogene ...... 19 E5 Oncogene ...... 22 E2 Transcription Factor ...... 24 HPV Oncogene Induced Malignant Transformation ...... 27 Overview ...... 27 Cellular Immortalization and other Malignant Properties Inherent to HPV Oncogenes ...... 27 Immune System Evasion ...... 30 Driving Mutagenesis by Corrupting DNA Repair ...... 31 Mitosis and Chromosomal Instability ...... 36 Conclusions and Future Considerations ...... 43 Chapter 3: The biochemistry of the DEK oncoprotein and its roles in DNA repair and chemotherapy resistance ...... 51 Introduction ...... 52

The biochemistry and post-translational modifications of DEK ...... 53 DEK regulates gene expression through chromatin dynamics and RNA processing .... 55 DEK drives oncogenic pathways and chromosome instability ...... 57 DEK is implicated in DNA double strand break repair ...... 58 DEK is necessary for chemotherapy resistance ...... 58 DEK loss attenuates non-homologous end joining DSB repair (NHEJ) ...... 60 Homologous recombination repair can mediate chemotherapy resistance in cancer ...... 61 Competition and compensation between NHEJ and HR ...... 63 Conclusions ...... 64 Chapter 4: DEK is required for homologous recombination repair of DNA breaks ...... 72 ABSTRACT...... 73 INTRODUCTION ...... 74 RESULTS ...... 77 DEK loss causes apoptosis in conjunction with DNA-PK inhibitors ...... 77 DEK is required for HR DSB repair ...... 78 DEK is necessary for the activation of γH2AX after ionizing radiation ...... 79 Loading of RAD51 onto RPA-protected DNA is significantly reduced ...... 80 DEK interacts with RAD51 ...... 81 DISCUSSION ...... 83 MATERIALS AND METHODS ...... 87 Cell culture, adenoviral infections, and viral transductions ...... 87 Serum Immunoglobulin ELISA Assays ...... 87 Cleaved caspase 3 flow cytometry ...... 88 DNA repair reporter assays ...... 88 Western blot analysis ...... 89 Immunofluorescence microscopy ...... 90 Chromatin fractionation ...... 90 Immunoprecipitation (IP) ...... 91 Statistical Methods ...... 92 Chapter 5: DEK associates with tumor stage and outcome in HPV16 positive oropharyngeal squamous cell carcinoma ...... 107 ABSTRACT...... 108

INTRODUCTION ...... 109 RESULTS ...... 114 DEK is most highly expressed in HPV16+/p16+ OPSCC tumors...... 114 DEK expression in tumors predicts advanced disease and poorer survival in HPV16+, but not in HPV16- OPSCC...... 115 DEK expression is associated with p16+ status in both HPV16+ and HPV16- disease. .... 116 DEK correlates with IL6 expression in HPV16+ OPSCC...... 117 DISCUSSION ...... 119 MATERIALS AND METHODS ...... 122 Study population ...... 122 Tissue microarray (TMA), immunohistochemistry (IHC), and HPV in-situ hybridization ..... 122 IHC scoring ...... 123 Statistical Methods ...... 124 ACKNOWLEDGEMENTS ...... 125 DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST ...... 125 GRANT SUPPORT ...... 125 Chapter 6: Discussion and future directions ...... 141 REFERENCES ...... 148

List of Tables

1 Table 1.1: AJCC TNM characteristics21...... 11 2 Table 1.2: TMN staging breakdown as described by the 2017 AJCC Cancer Staging Manual21...... 12 1 Table 2.1: Molecular Targets of the E6 HPV Oncogene at a Glance ...... 47 2 Table 2.2: Molecular Targets of the E7 HPV Oncogene at a Glance ...... 48 3 Table 2.3: Molecular Targets of the E5 HPV Oncogene at a Glance ...... 49 4 Table 2.4: Molecular Targets of the E2 Viral Oncogene Suppressor at a Glance ...... 50 5 Table 3.1: Cancers in which DEK overexpression contributes to worse prognosis or chemotherapy resistance ...... 69 6 Table 3.2: Known DEK functions and molecular targets ...... 70 7 Table 5.1: Patient and tumor sample characteristics ...... 131 8 Table 5.2: DEK expression is associated with an increased hazard of death in HPV16+/p16+ but not in HPV16- disease (survival univariate models)...... 132 9 Table 5.3: High DEK expression is associated with higher tumor stage in HPV+/p16+ OPSCC, and reduced perineural invasion in HPV- OPSCC. (Statistics for other clinical characteristics listed in Table S4)...... 133 10Supplementary Table 5.1: DEK survival summaries based on individual HPV and p16 status ...... 135 11Supplementary Table 5.2: DEK staining intensity, staining proportion, and quick score correlation with HPV and p16 status in all tumors...... 136 12Supplementary Table 5.3: DEK expression is associated with an increased hazard of death in HPV16+ but not HPV16- disease (survival univariate models)...... 137 13Supplementary Table 5.4: Clinical characteristics associated with DEK staining, based on HPV/p16 status (DEK stain proportion)...... 139 14Supplementary Table 5.5: DEK is not correlated with survival in p16 mono-labeled tumors (survival univariate models)...... 140

xii

List of Figures

1 Figure 3.1: Structure of the DEK oncogene ...... 66 2 Figure 3.2: The non-homologous end joining pathway ...... 67 3 Figure 3.3: The homologous recombination pathway ...... 68 4 Figure 4.1: DEK loss causes apoptosis in conjunction with DNA-PK inhibitors...... 94 5 Figure 4.2: DEK is required for HR DSB repair...... 95 6 Figure 4.3: DEK is necessary for the activation of γH2AX after ionizing radiation...... 98 7 Figure 4.4: Loading of RAD51 onto RPA-protected DNA is significantly reduced...... 99 8 Figure 4.5: DEK interacts with RAD51 and is essential for multiple functions in HR DSB repair...... 101 9 Supplementary Figure 4.1 ...... 102 10Supplementary Figure 4.2 ...... 103 11Supplementary Figure 4.3 ...... 104 12Supplementary Figure 4.4 ...... 105 13Supplementary Figure 4.5 ...... 106 14Figure 5.1: DEK is most highly expressed in HPV16+/p16+ OPSCC ...... 126 15Figure 5.2: DEK expression correlates with p16+ status in both HPV+ and HPV- OPSCC. ... 128 16Figure 5.3: High DEK expression was associated with IL6 expression in HPV16+ tumors .... 130 17Supplementary Figure 5.1: High DEK expression may be associated with IL6- status in HPV16- disease...... 134

xiii

Abbreviations

AML Acute myeloid leukemia AJCC American Joint Committee on Cancer ATM Ataxia telagiectasia mutated CRT Chemoradiotherapy, either CCRT or induction therapy CCRT Concurrent chemotherapy and radiation therapy CKII Casein Kinase II CtIP CtBP-interacting protein DEK-C C-terminal DNA binding domain of DEK DMEM Dulbecco’s modified Eagle Medium DNA Deoxyribonucleic acid DNA-PK DNA dependent protein kinase DNA-PKcs DNA-PK catalytic subunit DSB DNA double strand breaks EGFP Emerald green fluorescence protein FBS Fetal bovine serum HDAC Histone de-acetylase HNSCC Head and neck squamous cell carcinoma HPV Human Papillomavirus HR Homologous recombination HU Hydroxyurea IMRT Intensity-modulated radiation therapy IR Ionizing radiation / γ-irradiation kDa Kilodalton, 1000 g/mol MEF Mouse embryonic fibroblast MRN MRE11-RAD50-NBS1 complex NHEJ Non-homologous end joining OPSCC Oropharyngeal squamous cell carcinoma PTM Post-translational modification RB Retinoblastoma RNA Ribonucleic acid RPA Replication Protein A RT Radiation therapy TLM Transoral laser surgery TRS Transoral robotic surgery

xiv

Chapter 1: Oropharyngeal squamous cell carcinoma

Introduction

Head and neck cancers encompass tumors originating from the hypopharynx, oropharynx, nasopharynx, and oral cavity. Squamous cell carcinomas are the most common malignancy in these anatomical regions, accounting for >90% of cases, and arise from epithelial cells called keratinocytes1. Together, head and neck squamous cell carcinomas (HNSCCs) make up a substantial tumor burden and are the 6th most common malignancy worldwide2. Currently, the overall incidence of HNSCC is decreasing, but oropharyngeal squamous cell carcinoma (OPSCC) is the notable exception, increasing by 225% over the last 30 years. This upsurge in OPSCC is largely due to the 58% increase in human papillomavirus positive (HPV+) OPSCC frequency, which now comprises 50-80% of OPSCC burden in the United States3-6. Compared to

HPV negative (HPV-) disease, oral HPV infection carries significantly different epidemiological, prognostic, and, potentially, therapeutic implications, which are discussed throughout this chapter.

Anatomy of the oropharynx

The oropharynx consists of all structures located within the boundaries of the soft palate (superiorly), hyoid bone (inferiorly), frontal border of the posterior tongue (frontal), and the posterior pharyngeal wall. This includes structures such as the base of the tongue, posterior tongue, palatine tonsils, lingual tonsils, pharyngeal arches, uvula, soft palate, epiglottis, and the oropharynx wall. Squamous cell carcinomas that develop on any of these structures are considered OPSCC.

The anatomy of the oropharynx can make it difficult or impossible for patients to identify early-stage tumors, and OPSCC are usually not symptomatic until they reach a sufficiently advanced size. Due in part to these factors, about half of patients first present with locally advanced (stage IVA/IVB for HPV- and III for HPV+ disease, see

Tables 1.1 and 1.2) or metastatic (stage IVC or IV) disease7.

The epidemiological shift in OPSCC caused by HPV

The traditional HPV- OPSCC patient usually presents in their late 50’s to 60’s, is predominantly male (3-5:1 male to female ratio), and has a significant history of alcohol and tobacco use. While smoking and drinking individually contribute to OPSCC risk, daily consumption of >2 packs of cigarettes with >4 alcoholic drinks results in a dramatic

35-fold increased risk of developing HPV- OPSCC, marking these habits as major risk factors for HPV- OPSCC malignancies8,9. Since most of these tumors present at an advanced stage, treatment options consist of aggressive chemotherapy and radiation treatment (CRT) and surgery. Ideal surgical candidates are tumors that do not involve important anatomical structures and have minimal risk of damaging anatomical function.

However, surgery is often attempted in cases where negative tumor margins are predicted, tumor debulking is required prior to CRT, or in salvage therapy for recurrent tumors. In these cases, surgical tumor resection can damage the function of oropharyngeal structures, causing dysphagia and other symptoms such as chronic pain.

These morbidities may then be compounded by long-term side-effects of CRT which include xerostomia, trismus, dysphagia, ototoxicity, and nephrotoxicity. While the frequency of these adverse events have decreased with improved techniques, they are

3 still major issues in post-treatment care and patient quality of life, especially dysphagia10-15. In short, HPV- OPSCC patients present with advanced disease, a poor

5-year survival rate of 35%16, and life-long morbidities from aggressive treatment regimens.

There are significant differences in the patient population, risk factors, and outcomes in HPV+ OPSCC disease. Foremost, HPV+ OPSCC requires chronic infection with a high-risk HPV virus, over 90% of which is the HPV16 serotype17,18. Since most high-risk HPV serotypes, including HPV16, are sexually transmitted, the major risk factor is a history of six or more orogenital sexual partners19. While male patients still outnumber women (3-5:1 ratio men to women), they are usually younger, in their 40’s and 50’s, have a higher socioeconomic status, and experience better therapeutic outcomes than their HPV- counterparts (5-year survival of 82%)16,19. Because of this improved response to treatment, many clinical trials are currently testing if de-intensified

CRT can still provide remission with decreased risks of complications and improved quality of life after treatment10. Until these trials are concluded, treatment regimens will likely remain the same for both HPV+ and HPV- OPSCC. Unfortunately, it is already known that not all HPV+ patients respond well to current treatment: those with T4 or N3 stage HPV+ OPSCCs generally have a much poorer 54% 5-year survival rate, as is now reflected in the updated AJCC staging system (Tables 1.1 and 1.2)20,21.

Since the discovery of HPV as a risk factor in OPSCC 17 years ago22, studies have consistently shown an increase in HPV+ disease burden while the prevalence of

HPV- OPSCC decreases. According to recent estimates, HPV+ OPSCC incidence has dramatically increased around the world and is now the predominant subtype in the

United States and many parts of Europe (up to 80% of cases in these regions)6,23.

Considering this trend, the proportion of HPV+ OPSCC cases is likely to continue to grow. Even in the event that the majority of eligible people receive vaccination against high risk HPVs it will take 30-40 years before its protective effects against OPSCC development can be properly assessed6. In the meantime, efforts are needed to develop new systems of classification, prognostic markers, and de-intensified treatment strategies to optimize HPV+ OPSCC patient care.

Redefining the staging system for HPV+ OPSCC

Currently, all OPSCCs are graded based on clinical assessment of tumor size, the number and size of involved lymph nodes, and distant metastases according to the

TNM classification designed by the American Joint Committee on Cancer (AJCC)21. The details of staging are presented in Table 1.1 and Table 1.2.

Earlier this year, the AJCC system accounted for the therapy response differences in HPV+ OPSCC. Using the results of recent clinical studies, this staging system now recognizes that HPV+ high volume tumors, either T4 or N3, have significantly worse prognosis and have been re-categorized as stage III disease21,24,25.

Other reports suggest that anatomical staging alone is insufficient to properly risk- stratify patients. One important and modifiable risk factor for HPV+ OPSCC is a history of smoking, which confers a 4-fold mortality risk increase26,27. A history of >20 pack- years greatly reduces survival for T1/T2 disease20. For larger volume HPV+ tumors, patient age of >70 years becomes an indicator of poor outcomes20. There also remains the possibility that other important prognostic factors have yet to be identified, especially

5 considering the potential of the upcoming biomarker era. While the 2017 AJCC reclassification was a step forward in improving risk-stratification of HPV+ OPSCC patients, the significant epidemiological factors mentioned above and promising biomarkers should also be incorporated into future staging algorithms.

Current treatment strategies for OPSCC

While HPV status does not currently influence clinical decisions, treatment for

OPSCC does depend on many factors: TMN characteristics, AJCC staging, anatomical location, amenability of tumor to surgical resection, and patient preference. Because of the complexity and individuality of OPSCC presentation, personalized patient care is usually delivered by a multidisciplinary team of physicians, and clinical experience is a major factor in deciding on treatment strategies. However, there are general considerations for therapy based on tumor stage, recurrence, and minimizing treatment- based morbidities that are discussed below.

Early stage I or II tumors (see Table 1.2) can be successfully treated with radiation therapy (RT) or surgery without other interventions. Overall, RT is preferred over open surgery as it is more likely to preserve function, results in less facial deformity, and has manageable side-effects, especially with intensity-modulated radiation therapy (IMRT) techniques that can spare the salivary glands and mandible in many cases. This translates to a significantly decreased risk of osteoradionecrosis and xerostomia compared to conventional RT28-32. With the rise of new surgical techniques, such as transoral laser microsurgery (TLM) and transoral robotic surgery (TRS), tumors can be removed with minimal impact on the surrounding tissue. In general these new

6 techniques have similar outcomes to RT but without the associated long-term radiation comorbidities33,34. However, studies directly comparing efficacy between TLM, TLR, and

RT have not been completed35.

For advanced stage III-IV disease, a combination of surgery, chemotherapy, and

RT are utilized. Surgical resection remains an option for select patients, particularly those with smaller T1/T2 tumors or minimal lymph node involvement36,37. Surgery is also considered if the tumor has invaded the mandible or if there is need for tumor debulking prior to concurrent chemotherapy and radiation treatment (CCRT). CCRT has been demonstrated to provide superior survival rates compared to RT alone, and is the preferred method of treatment for advanced OPSCC38. Cisplatin and carboplatin are the most common agents used in CCRT, but the RTOG 1016 trial is currently validating results from a smaller study which indicated cetuximab was a potential substitute for platinum agents39. An alternative to CCRT is induction therapy, where a patient receives three cycles of cisplatin, fluorouracil, and docetaxel prior to RT40. While CCRT and induction therapy have not been compared in sufficiently powerful studies, combining induction therapy and CCRT, while reducing risk of metastasis, does not appear to provide additional survival benefit and increases the risks of neutropenia and leukopenia41,42.

Despite aggressive treatment, up to 30% of advanced OPSCC tumors do not respond well to therapy, and in the case of a complete response, there is a significant

17-31% risk of recurrence or metastasis43-45. Recurrent and metastatic disease is challenging to manage. For recurrent OPSCC, surgery is often the first-line therapy. As with primary interventions, TRS and TLM are more conservative approaches for tumor

7 removal, but may not be applicable to all cases33,35. In general, patients who have a local recurrence in an irradiated field within the first six months after therapy are usually ineligible for further RT unless a radiosensitizing agent is included or palliation is desired. In relapsed patients treated CRT, the response to standard treatment is exceptionally poor with median survival being 6-12 months, and many experimental therapeutics have not had success as second-line agents45. Still, there is some second- line therapy available for recurrent and intractable metastatic disease. These include cetuximab monotherapy (11-13% response rate and median survival of 7.5 months)46, the cetuximab-taxane recurrency scheme (20-43% response, mean survival of 6.1 months)47, and experimental inhibitors for PD-1 (16% response with median survival of

8 months for Pembrolizumab, 13.3% response with median survival of 7.5 months with

Nivolumab). In general, patient response to these second line therapeutics is very modest, especially since classic palliative chemotherapy of a platinum agent with either fluorouracil or a taxane provides 5 to 6.6 month mean survival and a response rate of

21-55%48,49.

The treatment of OPSCC often introduces life-long complications that can severely affect quality of life, even with treatment of early stage tumors. Surgical complications depend on the tumor site and most often arise from removing/compromising structures. For instance, procedures that necessitate a hemiglossectomy, soft-palate excision, or resection of the mandible can introduce symptoms of chronic pain12, dysphagia50, impaired speech51,52, and depression/anxiety from disfiguration12,53, among others. Surgical removal or irradiation of the lymph nodes can induce lymphedema: a swelling of the neck that can limit mobility and

8 worsen/induce symptoms of dysphagia54,55. There are also several toxicities associated with CRT. The most common RT side effect is xerostomia secondary to damage to salivary glands (~60% of cases), which can lead to dysphagia and can lead to dental deterioration56,57. This is followed by trismus secondary to fibrosis (35% of patients)58, and a more serious but less common complication of osteoradionecrosis (5.1-7.4% of patients), which can produce symptoms of dysphagia, difficulty with mastication, trismus, speech difficulties, jaw fractures, and infection32,59,60. Both radiation and chemotherapy can cause long-term damage to hearing, and cisplatin is often the culprit for treatment-associated neuropathy and nephrotoxicity13,14,61.

While most of these complications can be managed to varying degrees by physical therapy, dentistry, and pharmacology, it is ideal to design treatment modalities that reduce risks of complication development. Advances in surgical techniques, hyperfractionated RT schedules, intensity-modulated radiation therapy (IMRT), and conformal RT have made great strides in reducing the amount of tissue disrupted in surgery and in sparing tissue such as the parotid gland during CCRT11,29,31-33,62,63. The risk of complications may be further reduced in the future for HPV+ OPSCC if treatment de-escalation strategies are shown to be effective in clinical trials.

Conclusions

It is generally accepted that HPV+ OPSCC is a distinct disease which carries a better prognosis than HPV- OPSCC. This is now reflected in the 2017 AJCC staging system and may soon be considered in standard of care treatment strategies,

9 depending on the outcomes of ongoing clinical trials that aim to de-intensify therapy. As the number of HPV+ cases continues to grow and overshadow the traditional HPV- subtype, continued refinement of our prognostic methods will be essential to identify the patients who will benefit from treatment de-escalation.

TNM Definitions Tis Tumor in-situ T0 No primary tumor found T1 ≤2cm in largest dimension T2 >2cm and ≤4cm in largest dimension >4cm in largest dimension, or involvement of T3 lingual epiglottis Invasion of larynx, extrinsic muscle of tongue, T4a medial pterygoid, hard palate or mandible Invasion of lateral pterygoid muscle, pterygoid T4b plates, lateral nasopharynx, skull base, or encases carotid artery

N0 No regional lymph node metastases Metastasis in 1 ipsilateral lymph node ≤3cm in N1 largest dimension Metastasis in 1 ipsilateral lymph node >3cm N2a and ≤6cm in largest dimension Metastasis in multiple ipsilateral lymph nodes N2b ≤6cm in largest dimension Metastasis in bilateral/contralateral ipsilateral N2c lymph nodes ≤6cm in largest dimension Metastasis in lymph node >6cm in largest N3 dimension

M0 No distant metastases M1 Distant metastases

1Table 1.1: AJCC TNM characteristics21.

HPV- OPSCC Staging HPV+ OPSCC Staging Stage T N M Stage T N M 0 Tis N0 M0 T0 N1 M0 I T1 N0 M0 I T1 N0/1 M0 II T2 N0 M0 T2 N0/1 M0 T3 N0 M0 T0 N2 M0 T1 N1 M0 T1 N2 M0 III T2 N1 M0 T2 N2 M0 II T3 N1 M0 T3 N0 M0 T4a N0 M0 T3 N1 M0 T4a N1 M0 T3 N2 M0 T1 N2 M0 T0 N3 M0 IVA T2 N2 M0 T1 N3 M0 T3 N2 M0 III T2 N3 M0 T4a N2 M0 T3 N3 M0 T4b Any N M0 T4 Any N M0 IVB Any T N3 M0 IV Any T Any N M1 IVC Any T Any N M1

2Table 1.2: TNM staging breakdown as described by the 2017 AJCC Cancer Staging

Manual21.

Chapter 2: HPV Virology: Cellular Targets of HPV Oncogenes and Transformation

Eric A Smith1, Marie C Matrka1, and Susanne I Wells1

1Division of Oncology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, 45229, USA

Abstract

A subtype of head and neck cancer (HNC) characterized by infection with human papillomavirus (HPV) is clinically and molecularly distinct from classical HNC. In general these tumors arise in younger patients and respond better to therapy, while also expressing wild-type tumor suppressor proteins. Owing to these observations, efforts are currently being made to understand HPV+ HNC biology to improve clinical treatment regimens and develop novel therapeutic agents. To best predict which treatment strategies and novel therapeutics will have the most impact in patient care, a working understanding of how the HPV arsenal of oncogenes manipulates and transforms cells is crucial. This chapter will highlight the most important and clinically interesting cellular targets of the HPV oncogenes; the mechanisms by which the viral oncogenes subvert cell cycle checkpoints, DNA repair, immune surveillance, and cell death to promote malignant transformation; and promising future therapeutic options to target the functions of HPV oncogenes.

Introduction

As the clinical etiology, treatment strategy, and prognosis for human papillomavirus (HPV) positive head and neck cancers (HNC) continues to be defined, it will be important to understand how the HPV arsenal of oncogenes can produce this distinct classification of HNC. Compared to HPV negative HNCs, HPV positive (HPV+) tumors generally manifest at younger ages64 and are more responsive to therapy65. In addition, these malignancies frequently express wild-type tumor suppressors such p53, setting them apart from most other HPV negative cancer types. From an understanding of HPV biology in cervical cancer, this paradox was predicted and subsequently shown to occur from the p53 inactivation and destruction by the HPV oncogene E666. This phenomenon renders the tumor cells functionally p53 deficient. Explanations for the other phenotypes unique to HPV+ HNC, such as the early age of onset and enhanced therapy response compared to HPV negative tumors, may be developed through an understanding of the biology of the papillomaviruses.

There are over 100 different serotypes of HPV, and while the high-risk HPV types

16, 18, 31, 33, and 45 are known to cause cervical, anal, and head and neck cancers67,

HPV16 accounts for between 90 and 95% of HNC22,68. Considering the overwhelming prevalence of HPV16 in HPV+ HNC and the wealth of literature on this specific serotype’s oncogenes, a discussion of how the HPV16 oncogenes drives cells to malignancy will be the focus of this two-part chapter. Part one will broadly describe the functions of the HPV E5, E6, and E7 oncogenes in addition to the E2 transcription factor, and part two will highlight potential mechanisms and functions of the HPV gene products in cellular transformation. For quick reference, tables listing the cellular

15 pathways and molecular targets of the HPV genes are included in the text. These tables reflect the chapter’s focus on the most well studied and/or clinically significant targets of the oncogenes and are not intended to be all-inclusive.

Targets of HPV Viral Oncogenes and the HPV E2 Oncogene Repressor

E6 Oncogene

The central function of the E6 oncogene is to disrupt the apoptosis and senescence pathways in normal cells. The tiny 18 kilodalton protein accomplishes these feats through a combination of direct protein-protein interactions and targeted destruction of pathways regulators through polyubiquitination. E6 carries out targeted polyubiquitination by interacting with and stimulating the function of E6 Associated

Protein (E6AP), an E3 ubiquitin ligase encoded by the host cell genome69. Cellular proteins that become polyubiquitinated by E6AP are then rapidly degraded by the proteasome.

The most well studied and prototypical target of E6 is the p53 tumor suppressor, a potent inhibitor of cell cycle progression and inducer of apoptosis that can respond to a variety of cellular stressors70. To overcome p53-mediated cell death, the E6-E6AP complex polyubiquitinates and degrades p5369. In addition to destroying p53, E6 can independently prevent p53-mediated gene transcription by binding and inactivating the p53 co-activators p300 and CBP71,72. This double inhibition of p53 ensures that the tumor suppressor cannot signal cell cycle arrest and apoptosis. Interestingly, a recent report demonstrates that disrupting the p300-E6 interaction with a small molecule inhibitor is sufficient to restore p53 function and improve cisplatin response in HPV+

HNC73.

The E6 oncogene further disrupts the intrinsic apoptosis cascade downstream of p53 and the p53-independent extrinsic immune cell-mediated apoptosis pathway.

Through binding the tumor necrosis factor receptor 174, interaction with Fas-associated

17 death domain75, and enhanced degradation of caspase 876, E6 can halt apoptosis signaling through the death complex and immune cell death ligands. Additionally, targeted degradation of c-myc77 and, at least in the context of HPV18, the anti-apoptotic bcl-2 family protein BAK78, prevents intrinsic pathway function. Taken together, E6 is a potent inhibitor of apoptosis due to the various and seemingly redundant points at which the protein can destabilize this pathway.

Although less well-studied than apoptosis, several other cellular processes are altered by E6. This oncogene inhibits senescence by enhancing telomerase function79.

Telomerase expression is increased by activating TERT transcription through c-Myc80, and by degradation of the transcriptional repressor NFX1, which controls expression and RNA transcript stability of the hTERT telomerase subunit81,82. E6 can additionally increase hTERT activity by forming complex with telomeric DNA sequences and directly binding hTERT83. In addition to positively regulating telomerase, there is evidence that

E6 promotes the expression of multiple proteins associated with the epithelial to mesenchymal transition (EMT) phenotype, cell motility, and anchorage-independent growth84. E6 can also introduce genomic instability by de-regulating mitosis and the repair of DNA single strand breaks. Improper chromosomal segregation during mitosis, as a result of increased centrosome numbers, has been described in cells expressing low levels of E685. Additionally, some functions of the BRCA1 tumor suppressor86,

XRCC187, and ATR88 can be deactivated by E6, resulting in improper repair of single strand DNA breaks88. While no functional interacting partners or targets have yet been defined, the E6 oncogene can also prevent DNA double-strand break (DSB) repair89, despite HPV-induced DSB signaling for stimulating viral replication90.

E7 Oncogene

Similar to E6, the E7 oncogene is also a small protein that can interact with and alter the function of multiple host cell pathways. A major difference between these oncogenes are their targets, with E7 subverting the function of several endogenous transcription factors, chromatin remodeling proteins, and cell cycle checkpoint proteins.

The most famous targets of E7 are the retinoblastoma family of transcriptional repressors, which include the retinoblastoma tumor suppressor (RB/p105), p107, and p13091. The RB family regulates the G1/S cell cycle checkpoint by binding to and inhibiting the function of E2F transcription factors92. In the absence of RB, E2F proteins recruit histone deacetylases (HDAC) and other epigenetic modifying proteins to promote transcription of S-phase genes92. In healthy cells, E2F is released from RB control by the phosphorylation of RB by cyclin dependent kinases (CDK)93. E7 subverts this relationship by binding, sequestering, and targeting RB for ubiquitination and destruction94, removing CDK regulation of the G1/S checkpoint. The E7 oncoprotein also redundantly targets several other members of the RB-E2F axis to ensure checkpoint deactivation and chromatin remodeling. The E2F6 protein, which functions as an RB-independent antagonist of E2F signaling, is bound by E7 and deactivated95.

E7 can also bypass RB to directly stimulate E2F1 function96 and E2F-targeted histone modifiers, such as HDACs and the p300/pCAF acetyltransferases, to promote and repress gene expression97-99. Some of these upregulated E2F target genes, such as the chromatin-modifying and HDAC-interacting DEK oncogene, can further alter the transcriptional landscape of cells100. E7 also functions upstream of RB-E2F to inactivate

CDK inhibiting kinases and to stimulate CDK2 activity, the activation of which drives the cell into S-phase (as reviewed in Deshpande et al., 2005). Two of the main CDK2 inhibitors in epithelial cells are p27 and p21101, both of which are inactivated by E7102,103.

Stabilization of the active form of CDK2 is accomplished through direct104 and indirect binding through RB. Additionally, E7 increased the level of the CDC25A phosphatase which removes p21/p27 modifications on CDK2 and prolongs signaling105,106. These relationships were reviewed extensively in107.

Beyond the RB-E2F transcription axis and epigenetic remodeling, several other cellular processes are altered by E7. The oncoprotein can prevent cellular senescence by telomere maintenance through a telomerase independent mechanism. This process, known as alternative lengthening of telomeres (ALT), utilizes members of the homologous recombination (HR) DNA repair pathway to maintain telomere length108.

The HR factors implicated in E7-expressing cells are BRCA2, MUS81, and FANCD2109.

FANCD2 appears to be indispensable for this process as knockdown leads to telomere degradation109, and FANCD2 expression is independently upregulated by E7 through

E2F1109. Interestingly, E7-driven ALT also depends on DNA replication stress and the subsequent activation of ATR, a signaling kinase that recruits HR factors for DNA repair109. Under normal cellular conditions, ATR stimulation would activate the checkpoint kinases through claspin, initiating an S-phase or G2/M cell cycle block (as reviewed in Freire et al., 2006). To overcome this cell cycle blockade, E7 promotes the degradation and turnover of claspin, preventing robust cell arrest through the checkpoint kinases111.

The relationship between E7 and DNA repair extends beyond ALT and DNA replication stress. It has been long established that E7 expression is associated with increased DNA damage112, and recent work has indicated that HR repair mediated by

DNA double strand break (DSB) signaling though ATM is necessary for HPV genome replication90,113. E7 appears to promote the ATM response by directly binding to ATM and by indirect activation through the expression of Stat-5 target genes114.

Recent studies have indicated that E7 affects pathways involved in cancer invasion and metastasis. These include cytoskeleton regulation and expression of traditional EMT factors. With regard to the cytoskeleton, actin filament organization is important for regulating cell motility and adherence (as reviewed in Kaverina et al.,

2002). Both E6 and E7 can downregulate the N-cadherin adhesion molecule and increase the levels of Vimentin, E-cadherin, and fibronectin EMT factors116. This initial report suggested that the oncogenes post-translationally regulate these molecules as there was no change in their RNA transcript levels116. However, a more recent report suggests that E6 and E7 can independently increase the RNA transcript levels of four important EMT promoting transcription factors: SLUG, TWIST, ZEB1, and ZEB2. As a result, there was a loss of E-cadherin mRNA, increased Vimentin transcription, and promotion of cell migration and invasion phenotypes84. E7 may further modulate the

EMT phenotype by independently targeting actin metabolism through p190RhoGAP 117 and Gelsolin118. The p190Rho GTPase activating protein catalyzes the inactivation of

RhoA, a promoter of actin stress fiber formation and cell motility. Inhibition by E7 may paradoxically suppress the cell migration phenotype through over-activation of RhoA117.

The other E7 target in actin metabolism, Gelsolin, plays an important role in stabilizing

21 actin filaments and regulating apoptosis118. However, the relationship of the E7-Gelsolin interaction on EMT and migration remains unexplored.

E5 Oncogene

Despite being a weaker oncogene than E6 and E7, the E5 protein likely plays an essential supportive role in tumorigenesis119. While E5 can stimulate transformed and tumorigenic phenotypes in mouse fibroblasts120,121, its oncogenic potential in human keratinocytes appears to be limited to enhancing E6/E7 immortalization122 and enhancing EMT-like cell motility123,124. There is not yet a consensus on a prototypical function and target of E5, but several of the oncogenic phenotypes are associated with

E5-mediated enhancement of epidermal growth factor receptor (EGFR) signaling125.

EGFR activating may be promoted through several mechanisms, most of which are attributed to E5’s role as a transmembrane protein. Owing to this unique biology, E5 can also localize to the ER membrane and, depending on the level of expression, the plasma, nuclear, and vesicle system membranes126. While direct interaction with EGFR remains uncertain127,128, E5 has been shown to block ubiquitination and degradation of

EGFR by Cbl129 and inhibit lysosomal degradation of internalized EGFR130. There is a consensus that E5 blocks lysosomal degradation of EGFR, but the mechanism by which this occurs is debated. Reports suggest that EGFR is spared by direct blockade of lysosome-endosome fusion by E5 or from indirect blockade by altering cytoskeleton rearrangements131,132. Alternatively, E5 may prevent lysosomal acidification, a necessary step to activate lysosomal enzymes. This is reported to occur via the

22 endogenous ion-channel function of E5133, and the binding and subsequent disruption of the lysosomal V-ATPase proton pump function130,134.

As a result of enhanced EGFR signaling, cells harboring E5 excrete pro- inflammatory and angiogenic factors from constitutively activate ERK1/2, AKT, and NF-

Kβ135,136. These pathways up-regulate COX-2, an enzyme that initiates prostaglandin synthesis and inflammation (reviewed in Ricciotti and FitzGerald, 2011), in an EGFR- dependent manner136. In E5-expressing epithelial cells, COX-2 drives the production of the PGE2 prostaglandin, a potent inflammatory molecule, and its corresponding receptor, Ep4. This establishes a feed-forward loop that amplifies PGE2 signaling and drives the expression of vascular endothelial growth factor (VEGF)138, a powerful inducer of angiogenesis and anchorage independent growth that can stimulate tumor development and progression139.

There are several other EGFR-dependent and independent cellular processes that E5 can alter to promote immortalization and tumorigenesis. Expression of the cell cycle checkpoint proteins p21 and p27 is greatly diminished by E5 specifically in the context of EGFR signaling140,141. Independently of EGFR, the mitotic spindle checkpoint markers, BUB1 and MAD2, are disrupted by E5142, and degradation of pro-apoptotic

BAX factor is enhanced143. E5 can also stimulate activation of PLC-1144, ERK1/2, and p38 MAPK145 secondary messenger pathways independent of EGFR function.

Additionally, the oncogene may also protect the tumor from immune surveillance by disabling two important antigen presenting receptors that all cells express. The first receptor is CD-1d, which presents antigens to natural killer T-cells (reviewed in

Rossjohn et al., 2012), and is targeted for degradation by E5147. The second target is

23 the classical major histocompatibility complex I (MHC I), which presents antigens to

CD8 cytotoxic T-cells (see review by Gao and Jakobsen, 2000). MHC I requires several chaperone proteins for proper folding during its assembly in the ER and Golgi (see review by Paulsson and Wang, 2003). The transmembranous E5 oncogene exploits this weakness by binding the BAP31 chaperone150 and the Calnexin-MHC1 heavy chain complex151. As a result, MHC I does not mature and is not expressed on the cell surface, preventing the recognition of viral peptides by cytotoxic CD8 T-cells152.

For comprehensive insights into E5 activities, we refer to a recent review by

DiMaio and Petti153.

E2 Transcription Factor

The E2 HPV gene product is a transcriptional repressor that controls E6 and E7 expression during the course of normal HPV infection and throughout the early stages of tumor development154-156. In later stage tumors, expression of the E2 protein can be lost, allowing an increase in E6 and E7 expression, and enhanced cellular growth157,158.

Reintroduction of E2 in these cells results in growth arrest159, apoptosis160,161, and senescence155,162, suggesting cancer cell dependence on excessive E6/E7 expression.

While the exact mechanism by which E2 is lost in advanced tumors remains uncertain, one explanation correlates the integration of the viral episome into the host cell genome with the subsequent destruction of the E2 open reading frame (ORF)157,163. Indeed, the

E2 gene is frequently silenced in cells with HPV genomes integrated into the host cell genome164. However, recent reports challenge the ORF destruction model. It has been reported that some cell lines with integrated HPV genomes still robustly expressed E2

24 transcripts, but with minor or no E2 protein expression. In these cell lines, E2 is likely post-transcriptionally regulated156. The idea that the E2 ORF is occasionally spared from destruction is also supported by a genome-wide sequencing of HPV integration events. In this report, it was found that HPV integration does not preferentially occur in the E2 gene165, contrary to the traditional model163. Lastly, it is possible that mutations occur in the transactivation region of the E2 gene which impair the ability to recruit cellular factors for E6/E7 repression, without compromising the expression of E2166.

These mutations, however, have not yet been demonstrated to arise naturally. Taken together, these observations indicate that E2 protein expression is compromised after integration, and this could occur through ORF destruction, an unknown post- transcriptional regulation mechanism, or hypothetical mutations in E2 domains which impair E6/E7 repression.

In order for E2 to function optimally as an E6/E7 repressor, the transcription factor requires the recruitment of at least three cellular proteins. The most well studied interacting partner is Brd4, a bromodomain chromatin binding protein that recognizes specific types of acetylated histones167. The E2:Brd4 complex was first described as an anchor that tethered viral episomes to chromosomes during mitosis168. Subsequent work demonstrated that E2-mediated gene repression is dramatically reduced, but not eliminated, following the loss of Brd4169,170. It was later determined that Brd4 enhances the binding of E2 to target sequences and enhanced E2 protein stability171. More recently, the EP400 component of the NuA4/TIP60 histone acetyltransferase complex and the SMCX demethylase were identified to necessary for E2 function172. All three of these factors were found to be necessary for E2-mediated repression of E6/E7, as

25 knockdown of any single factor led to an increase in HPV18 E7 expression172.

Additionally, recent work has identified a potential mechanism to regulate repression of

E6/E7 in HPV16 E2-expressing cells. CCHCR1, a cellular factor that may play a role in psoriasis development173 and keratinocyte differentiation174, can compete with Brd4 for binding of E2175. Once bound, the E2:CCHCR1 complex moves to the cytoplasm with a resulting loss of E2-mediated differentiation175.

Lastly, E2 has at least two functions that are distinct from its role as a transcription factor. The E2:Brd4 complex can tether viral episomes to mitotic chromosomes, ensuring distribution between daughter cells176. Additionally, cellular metabolism may be affected through the interaction of E2 with UQCRC2 at the mitochondrial membrane. As a result of this interaction, the levels of mitochondrial ROS increase, together with HIF-1α stabilization, and elevated glycolysis177.

HPV Oncogene Induced Malignant Transformation

Overview

HPV infection and expression of the HPV oncogenes do not induce tumorigenesis per se but rather provide an ideal environment for progression towards malignant transformation. Evidence supporting this view include the long clinical latency from initial infection to malignancy, the low penetrance of malignancy in HPV infected patients178, and the inability of HPV-infected human keratinocytes to develop tumors in xenograft mouse models179. While HPV does not directly transform cells, the viral oncogenes nevertheless provide several oncogenic phenotypes and create a scenario that supports genomic structural aberrations and mutations, and evasion of the immune system. Within this section the contributions of each HPV oncogene will be discussed in terms of their inherent oncogenic phenotypes and three mechanisms of tumorigenesis which they promote.

Cellular Immortalization and other Malignant Properties Inherent to HPV Oncogenes

In culture, healthy somatic primary cells can only replicate a defined number of times before entering senescence or apoptosis from telomere depletion. Immortalization occurs when cells overcome this replication limit and replicate indefinitely. To achieve immortality, cells must be able to maintain their telomeres and evade apoptosis. With regard to oncogenesis, cellular immortalization is an important precursor event for malignant transformation180.

When the viral oncogenes are expressed individually, only E7 can reliably immortalize human primary keratinocytes181. However, the addition of E6 dramatically

27 improves the efficiency of cellular immortalization, which can be further enhanced another 4-10 fold by E5122,182-184. This synergy is most likely a result of the complementary functions of the oncogenes. E7 initially drives hyperproliferation of cells by disabling RB-dependent G1/S checkpoint activation. However, these cells do not easily overcome senescence182, at least in part as a result of p53 checkpoint arrest and senescence signaling. This p53 mediated cell death can be overcome by E6-directed p53 inhibition and destruction (Table 2.1). Conversely, E6 can disrupt p53 inhibition of cell cycle progression, but, without a mechanism to force replication, E6-expressing cells are stalled in G1 and undergo senescence or quiescence182. It is less clear, however, how E5 augments the complementary E6/E7 immortalization process. E5 may provide a growth and replication stimulus through EGFR signaling125, or further enhance evasion of apoptosis and cell cycle checkpoint by targeting BAX, p21, and p27140,141,143.

In the context of HPV infection, the three oncogenes work cooperatively to immortalize cells by driving cellular replication, evading senescence, and averting apoptosis.

Uncontrolled cellular growth and immortalization is just one oncogenic phenotype linked to HPV infection, but it has an important counterpart in vivo. Epithelial dysplasia is an outcome of HPV infection in humans, particularly with regard to the cervix. High- grade dysplasia and cervical cancer phenotypes are recapitulated in E7 transgenic mouse models, and this effect is enhanced by concurrent E6 expression185. In the same report, the authors showed no dysplasia phenotype associated with E6 expression, mirroring the cell immortalization phenotypes in vitro181,185. Additionally, mice expressing the HPV genome in their epithelia formed squamous cell carcinomas following estrogen therapy, suggesting a hormonal component in tumor development186,187. While HPV16

E5 can transform immortalized murine fibroblasts, enabling the production of tumors in nude mice120,121, this same activity is not observed in human cells. However, E5 appears to have in vivo functions where the oncogene can induce epidermal hyperplasia and increase the rate of spontaneous tumor formation in E5 transgenic mice188.

Concurrent expression of all three oncogenes can also promote adhesion free- survival, migration, and invasion in trophoblasts184. The enhancement of these in vitro attributes, in addition to the epithelial-mesenchymal transition (EMT), is correlates with motility phenotypes in different ways, but there are some overlapping targets. E6 promotes EMT, in part, by binding and degrading cytoskeleton modifying proteins containing a PDZ domain189-192, and upregulating vimentin, a mesenchymal marker 193.

In addition, both E6 and E7 can upregulate classical EMT transcription factors including

TWIST and SLUG84. E7 can independently downregulate the epithelial cell marker fibronectin194, and potentially promote actin reorganization117,118. Lastly, E5 can drive inflammation, angiogenesis, and anchorage independent growth through EGFR, NF-κB, and other signaling pathways125,135,136,195.

The cellular growth phenotype imparted by E6 and E7 can be enhanced following virus integration and the loss of E2158, but genomic integration is not a required event for malignant transformation in cervical cancer or HNC22,196. In cells with viral integration and E2 silencing, the effects of increased E6 and E7 levels on the EMT phenotypes are not well characterized. However, the E5 ORF is frequently lost after viral integration

197,198.

Immune System Evasion

An important aspect of tumor development is the ability to escape detection by the immune system. If nascent oncogenic cells are unable to accomplish this, odds are that immune surveillance will destroy the corrupted cells. Immune evasion is even more important for oncogenic cells expressing viral proteins as they must overcome both an intracellular antiviral response and the intercellular immune system180. The full scope of the HPV oncogenes to repress immune activity is not established, but further work on this topic will increase our understanding of key mechanisms by which HPV promotes tumorigenesis.

E5 has a central role in evading HPV-infected cell death by natural killer cells and tumor-specific cytotoxic T-cells. As discussed earlier, the oncogene degrades CD-1d, the natural killer cell antigen presenting molecule147 and thoroughly disrupts maturation of MHC I, the cytotoxic T-cell antigen presenting molecule, through direct interaction and by binding critical chaperone molecules150,151,199. The loss of these antigen presenting molecules prevents the recognition of (pre-)malignant HPV+ cells

In contrast to E5, the other HPV oncogenes cooperate to disrupt the cytokine, interferon, and apoptosis responses that arise from viral infection. From the studies that have been performed, it is clear that E6 and E7 can prevent cytokine and interferon signaling through different mechanisms depending on the target. A classic example of

E6/E7 cooperation is in the repression of IL-8 expression, where both oncogenes compete against NF-κB for transcriptional co-factors. In this scenario, NF-κB targeted gene regulation is suppressed by the competitive binding of CBP by E6 and pCAF by

E7. While each oncogene can individually reduce IL-8 upregulation by NF-κB, this

30 downregulation is further enhanced when both oncoproteins are present200. Synergistic effects of E6 and E7 can also regulate cytokines at the protein level. In response to infection, keratinocyte release the IL-18 interleukin, a pro-inflammatory and natural killer

T-cell stimulating agent, which can be blocked by direct E6 binding201 and E7-mediated upregulation of the inhibitory IL-18 binding protein202. The oncogenes can also work individually, with E6 binding IRF3 to prevent transcription of IRF3-target interferon genes203. In some cases the mechanism of regulation by the HPV oncogenes is not fully understood. These target proteins include the intracellar antiviral response protein

PKR204; the monocyte chemoattractant protein, MCP-1205; IL-1β206; and the interferon regulatory factor, IRF-9207. Our understanding of how these oncogenes interact with immune signaling and regulation is still in its infancy, as new targets are continually discovered and characterized.

Driving Mutagenesis by Corrupting DNA Repair

Host cell DNA repair pathways play an important role in viral replication.

Depending on the virus, these pathways can either be toxic to or essential for replication and propagation. Not surprisingly, viruses such as HPV have evolved mechanisms to exploit DNA repair to best suit their needs. HPV requires both a strong DNA double strand break (DSB) repair response and rapid cell cycle progression for productive virion production208. At first glance these needs appear to be mutually exclusive, but

HPV can accomplish both by sacrificing host cell genomic integrity. As a result, the virus promotes a cellular environment that is tolerant to DNA mutagenesis and tumorigenesis.

Maintaining DNA fidelity is a priority for regulated cell division, and it is thus not surprising that DNA repair, checkpoint signaling, and cellular replication are intimately linked. When a DSB is recognized by ataxia telangiectasia mutated (ATM), it signals for immediate G1/S checkpoint arrest through checkpoint kinase 2 (Chk2) and p53 signaling. As a result, CDK2 activation is inhibited and the cell does not proceed past the G1/S checkpoint until the damage is repaired. Likewise, defects in DNA replication can recruit DNA repair proteins and halt replication at the G2/M checkpoint. In this situation, exposed single-stranded DNA (ssDNA) and DNA replication errors recruit and activate ATM-related (ATR), which then drives p53 and checkpoint kinase 1 (Chk1) signaling to inhibit CDK1 activation (see review by Smith et al., 2010a) .

To uncouple DNA repair from cell cycle arrest, E6 and E7 promote CDK1 and

CDK2 activation through several mechanisms. E7 stimulates CDK2 through direct interaction, Chk2 inhibition, CDC25A upregulation, and de-regulation of p53 effector proteins. These properties of E7 were discussed above in the HPV oncogene section of this chapter. In addition to stabilizing CDK2 activation, E7 can also prevent CDK1 inactivation by Chk1. This is accomplished by promoting the degradation of claspin, which is essential for Chk1 activation by ATR. With a reduction of available claspin,

Chk1 stimulation is attenuated and G2/M arrest is much less likely to occur111. The loss of checkpoint arrest signaling enables the unfettered HPV-infected cell to undergo rapid cell division without the need to ensure complete DNA damage repair. In this way, E7 enables HPV to capitalize on frequent S-phase conditions to amplify the viral genome210. The expression of E6, which promotes the degradation of p5369, further

32 ensures that checkpoint signaling, apoptosis, or senescence will not result from the accumulating burden of DNA damage.

Cells expressing either E6 or E7 carry an excessive level of DNA damage, reflected by both an increase in chromosomal instability and upregulation in the DSB marker, γH2AX 112. This basal level of damage stimulates the ATM DSB repair pathway, which is necessary for virus replication90. While the origin of the damage remains unclear, it is thought that a combination of incomplete DNA repair during cell division and mitotic defects are accountable for part of the DNA damage burden211. Another portion of the DNA damage might be due to reported interactions between HPV oncogenes and DNA repair proteins. For instance, ATM is indirectly stimulated by

STAT-5 signaling in HPV31 infections114 and might be directly activated by interacting with HPV16 E790. Another DSB repair related protein, DEK, is also upregulated by E7

E2F signaling100. DEK is an oncogene with pro-survival and chromatin modifier functions that regulates ATM-driven DSB repair212.

However, ATM activation and DEK upregulation by E7 are likely unique as the common theme for E6 and E7 is to inactivate or subvert DNA repair protein functions. In the context of HPV5 and HPV8, ATR is degraded by E6, preventing the repair of thymidine dimer DNA lesions88. Likewise, HPV16 E6 hinders the function of XRCC1 in ssDNA repair 87. While E7 promotes ATM signaling, overall end-joining DSB repair is decreased in E6 expressing cells89. Functional DSB repair by homologous recombination may also be stymied as both E6 and E7 can directly bind and alter the function of BRCA186, and E6 can do the same with the BRCA1 stabilizing protein,

BARD1213. The non-destructive targeting of BRCA1 is of particular interest due to the

33 important role the protein has in homologous recombination repair of DSB and the ATR- based Fanconi Anemia (FA) repair pathway of interstrand crosslinks [for a review on

FA, see reference:214]. Some functions of BRCA1 are silenced by the oncogenes86 while

E7 co-opts both BRCA and FANCD2, a critical FA pathway component, to drive telomere elongation109.

Disrupting and hijacking FA repair components is advantageous for the virus as a functional FA pathway opposes cell growth and DNA damage accumulation in HPV16 infected cells. FA repair of DNA interstrand cross links (ICL) is a complex, but necessary, endeavor to preserve genomic integrity in dividing cells. In order to repair

ICLs, ATR recruits the FA core complex to phosphorylate and activate the

FANCD2/FANCI ID complex. The ID complex then proceeds to orchestrate translesion synthesis to continue DNA replication, nucleotide excision to remove the ICL, and homologous recombination to repair the DSB byproduct of ICL excision214,215.

Individuals who suffer from Fanconi Anemia have recessive mutations in a protein of the

FA core or ID complexes, or in downstream associated breast cancer susceptibility genes, resulting in a non-functional FA repair pathway. One characteristic of Fanconi

Anemia is early onset HNC, occurring prematurely before or during the 4th decade of life. Greater than >80% of these malignancies have been described as HPV positive216, while other reports have disputed these findings217-219. Regardless of HPV presence in

FA tumors, the relationship between HPV and FA repair has been described by several independent studies. First, an FA deficient mouse model of E7 HNC formation had a higher incidence of tumor development and a greater burden of DSB when compared to

FA proficient control mice220. Second, an in vitro study demonstrated that HPV16 E7

34 induced FA pathway activation. The addition of E6 further enhanced FA signaling and accelerated chromosomal instability in FA deficient cells221. Furthermore, a defective FA pathway permits enhanced hypertrophy through proliferation in HPV+ raft cultures, suggesting an inhibitory role of FA repair in HPV stimulated cellular growth222. In retrospect, the tumor suppressor function of the FA pathway is not surprising. Both ATR and FANCD2 are involved in initiating cell cycle arrest223, and ICLs must be correctly repaired by FA to prevent radial chromosome formation in mitosis224,225. Thus, the dramatic increase in FANCD2 activation in FA proficient HPV+ HNC could be a response to limit E7-driven hyperproliferation and DNA damage accumulation226.

Correspondingly, the absence of this repair pathway in Fanconi Anemia patients may accelerate HPV replication, resulting in characteristically early head and neck as well as anogenital tumorigenesis.

In summary, HPV E6 and E7 promote a pro-mutagenic cellular environment through the de-regulation of the DNA repair pathways. Mechanisms through which this occurs include the uncoupling of DNA repair from cell cycle checkpoint arrest, mitotic errors from incomplete or altered DNA repair, and the inactivation or hijacking of DNA repair proteins. The HPV oncogenes cultivate DNA damage to nurture ATM signaling for viral replication90,113 but also compromise the fidelity of ATM-dependent DSB repair, although the relevant underlying mechanisms are currently unclear. An intact FA DNA repair pathway opposes the uncontrolled cell cycling and DNA accumulation in cells expressing E6 and E7, but HPV oncogenes may overcome this suppression by subverting the function of critical DNA repair proteins. These include the central FA pathway member FANCD2, and a homologous recombination repair factor BRCA1 that

35 is activated by FA signaling. How the DNA repair functions of BRCA1 and FANCD2 are altered by E6 and E7 expression remains unknown, as do the identities and functions of other DNA repair proteins that are likely targeted by E6 and E7. However, the reduction of DNA repair fidelity may be, at least in part, responsible for one beneficial phenotype in HPV+ HNC: enhanced chemoradiation sensitivity compared to HPV negative tumors227. This enhanced sensitivity was mechanistically linked to inhibition of p53 activity by E6. In human tonsillar epithelial cells expressing exogenous E6, knockdown of p53 increased radiation resistance228. Thus, wild-type p53 may be a critical prognostic marker for HPV+ HNC response to therapy.

Lastly, integration of the viral genome is linked to tumor progression in cervical cancer158,164, and it is believed that integration also plays a role in HPV+ HNC progression. The mechanism by which the HPV genome inserts itself into the host cell’s genome is unclear, but other viruses commonly rely on DSB repair proteins to accomplish this feat (see review by Chaurushiya and Weitzman, 2009). Very recent work involving whole-genome sequencing has identified a potential cataclysmic mechanism by which HPV-driven genomic instability fractures chromosomes in a manner distinct from chromothripsis165,230.

Mitosis and Chromosomal Instability

Most HPV-associated malignancies have numerous chromosomal imbalances, including gains or losses of whole chromosomes (aneuploidy) and chromosomal rearrangements231. These chromosomal aberrations are often a result of mitotic defects including spindle polarity defects, chromosomal missegregation, and chromosomal

36 breaks. The accumulation of these chromosome abnormalities happens over time, a process which explains the long latency between initial HPV infection and the development of cancer232. Importantly, aneuploidy is already detectable in pre- malignant cells suggesting the HPV oncogenes play a causal role in the genetic instability that results in oncogenesis107.

Some of these chromosomal defects are the result of abnormal multipolar mitoses, with tripolar spindles being characteristic of HPV16 oncogene expression.

Such events are associated with abnormal centrosome numbers233,234. E6 and E7 cooperate to induce centrosome abnormalities resulting in multipolar spindles, chromosomal missegregation, and aneuploidy107. Centrosomes are composed of two centrioles that generate the spindle poles during mitosis, and, considering the importance of this organelle to proper cell division, centrosomal duplication is highly controlled and occurs only once prior to each cell division in healthy cells211.

The process starts in S phase and is completed in G2211. During prophase in mitosis, the centrosomes migrate to opposite poles of the cell and the mitotic spindle forms between them. After cell division, each daughter cell receives one centrosome, containing a pair of centrioles. Centrioles, cylindrical organelles involved in the organization of the mitotic spindle, can erroneously be duplicated more than once before mitosis. It has been shown that the number of extraneous centrioles correlates with the risk of mitotic defects85. Centriole amplification can result in two ways. One way is through centrioles being over-duplicated, the other is centriole accumulation in cells that do not properly progress through mitosis. These two defects have different implications. Centrioles are considered over-duplicated when there are one or two

37 maternal centrioles and multiple immature daughter centrioles. Over-duplication is more likely to result in cell division with abnormal chromosomal segregation and aneuploidy.

Centriole accumulation is characterized by multiple maternal centrioles and a normal ratio of daughter centrioles. Centriole accumulation is likely the result of aborted mitoses or problems during cytokinesis that are unlikely to result in viable daughter cells or propagate aneuploidy. E6 expression can cause centriole accumulation in cells that are already genomically unstable; however these cells are unlikely to remain in the proliferative pool due to catastrophic DNA damage211,235. On the other hand, E7 can cause over-duplication and therefore increase instances of multipolar mitosis and chromosomal abnormalities in daughter cells. E7 is thought to directly cause genomic instability through centriole duplication control as is evidenced by E7 expression introducing abnormal centriole numbers in otherwise normal cells prior to the onset of genomic instability236. E7 is also capable of causing centrioles to appear rapidly within a single cell division cycle suggesting a direct role237. These resulting genomic abnormalities and aneuploidy could potentially result in a selective growth advantage that eventually results in carcinogenic transformation.

The mechanism of E7 mediated over-duplication of centrioles enables multiple daughter centrioles to be produced from one maternal centriole. This is partially dependent on high levels of CDK2/cyclin E activity237, and is thought to happen through both RB dependent and independent mechanisms. Highlighting the role of RB in E7 mediated centriole amplification, an E7 mutant protein that cannot degrade RB is also unable to cause centriole over-duplication. Interestingly, full length HPV-16 E7 can cause centriole abnormalities in RB/p107/p130-deficient mouse embryo fibroblasts

38 albeit at a much lower incidence. This suggests there is also an RB independent role in

E7 induced centriole amplification238. Several RB independent mechanisms have been identified including the aberrant recruitment of polo-like kinase 4 (PLK4) to maternal centrioles by high levels of CDK2 activity to promote centriole duplication. PLK4 protein levels are rate limiting in centriole multiplication and its overexpression is sufficient to induce centriole multiplication. Along with aberrantly recruiting PLK4 to maternal centrioles, E7 can activate the PLK4 promoter, up-regulating PLK4 mRNA which correlates with the ability of E7 to induce centriole multiplication239. Typically the PLK4 promoter is bound by the DREAM (DP, RB-like, E2F4 and MuvB) complex which represses PLK4 expression (for a review on DREAM, please refer to Sadasivam and

DeCaprio, 2013). E7 however can disrupt the DREAM complex, thus preventing PLK4 repression and causing deregulated centriole duplication241. During DNA damage, p53 can activate the DREAM complex to further support the repression of PLK4 and centriole duplication. However, E6 dependent degradation of p53 can further support centriole duplication. Lastly, E7 can directly bind to γ-tubulin, resulting in the removal of

γ-tubulin from the mitotic spindles and potentially leading to abnormal centrosome synthesis242,243. Together, these studies suggest that aberrant centrosome duplication is an early event caused by E6 and E7 that may drive chromosomal instability.

In addition to centrosome abnormalities, E6 and E7 have been shown to independently bypass mitotic checkpoints resulting in the accumulation of polyploid cells and aneuploidy244,245. The HPV oncogenes function individually and in concert to overcome the spindle assembly checkpoint (SAC) in mitosis. The SAC ensures proper alignment and segregation of chromosomes and is activated in response to aberrant

39 microtubule-kinetochore attachments but not necessarily due to the presence of supernumerary centrosome or multiple spindle poles246,247. Data suggest that E7 expressing cells exhibiting multipolar spindle poles in metaphase are less likely to complete mitosis. However, the fate of these cells is unknown. Cell fate could include proceeding through mitosis normally, mitotic dysfunction resulting in apoptosis or mitotic catastrophe, or decondensing of the chromosomes and re-entry into a G1-like state with

4n DNA content. The later can happen upon DNA damage, when E7-expressing cells arrest at the G2 checkpoint and then undergo re-replication. Re-replication is a process of successive rounds of host DNA replication without entering mitosis. This process is thought be a result of Cdt1 up-regulation by E7, Cdt1 being the only known mammalian gene to efficiently trigger re-replication when overexpressed in cancer cells248-250. E7 can stabilize Cdt1 at the post translational level potentially inducing DNA re-replication and polyploidy causing further genomic instability251.

E6 and E7 have independently been shown to overcome the SAC and promote the accumulation of polyploidy cell populations244,245. HPV-16 E6 and E5 cooperate to allow cells to activate the APC/C to promote the metaphase to anaphase transition. The anaphase-promoting complex/cyclosome (APC/C) is a multi-subunit ubiquitin ligase that must be activated for cells to transition from metaphase to anaphase. The APC/C complex is activated by the binding of Cdc20 in metaphase. The SAC proteins BubR1 and Mad2 can bind to Cdc20 to prevent APC/C activation until all chromosomes are aligned correctly on the mitotic spindle. E6 dependent degradation of p53 can lead to

Cdc20 overexpression resulting in a portion of Cdc20 bound to BubR1 and another portion free of BubR1 indicative of an active checkpoint in some cells and not in others.

Free Cdc20 can activate APC/C allowing slippage through the mitotic checkpoint252. E6 can also increase the expression of Ubch10, an E2 ubiquitin-conjugating enzyme that leads to uncontrolled APC/C activity and degradation of cyclin B. Degradation of cyclin

B allows cells to exit mitosis despite the presence of unrepaired DNA damage252

Furthermore, E5 expression correlates with the gradual decreases in expression of

Bub1 and Mad2 during cervical cancer progression142. Exogenous expression of

HPV16/18 E5 can decrease both BubR1 and Mad1 mRNA and protein levels potentially though direct interaction with E5 at least for BubR1. Therefore E5 might also contribute to the weakening of mitotic checkpoints142.

E5/E6/E7 genes cooperate to promote mitotic progression in the presence of

DNA damage, thus further promoting genomic instability and aneuploidy. As malignant progression occurs over the course of many years, it is likely that these mitotic defects occur infrequently and do not often lead to viable progeny. The accumulation of subtle chromosomal alterations may provide a growth advantage to a subclone of HPV- positive cells, resulting in the outgrowth of a cellular population that contributes to viral persistence and, ultimately, malignant progression.

A recent study showed that HPV integration sites in human cancer genomes directly flank genomic structural variations, including focal amplifications, rearrangements, deletions, and/or translocations. These genomic alterations have frequently disrupted the expression and structure of neighboring genes involved in oncogenesis, and correlate with amplification and increased expression of E6 and E7165. HPV integration can lead to oncogene amplification and loss of heterozygosity of tumor suppressor genes253. Homozygous deletion and

41 rearrangements of DIAPH2, a gene whose loss of function promotes chromosomal instability via misalignment of sister chromatids during metaphase, was observed in a HNC cell line after HPV integration254. This is not surprising as microarray data reveals that the number one cellular process transcriptionally altered in cervical cancer compared to healthy cervical epithelium is the cell cycle and specifically M- phase processes255.

Conclusions and Future Considerations

The three HPV16 oncogenes are powerful agents of cellular immortalization, particularly when co-expressed. While the E5, E6, and E7 oncogenes reprogram the cells to exhibit aspects of tumorigenesis such as EMT and immortalization phenotypes, they are not sufficient to induce tumor formation. In an uncertain percentage of infections, HPV can establish a long-standing infection by evading immune detection.

With the functional silencing of the quintessential p53 and RB tumor suppressors, HPV- immortalized cells are primed to progressively acquire mutations that can eventually result in tumorigenesis. How these errors accumulate remains incompletely understood, but the model presented in this chapter incorporates uncontrolled cell division, improper

DNA repair, and mitotic errors as overarching mechanisms. In short, infection with HPV eliminates the necessity of acquiring mutations in key tumor suppressors (e.g. RB, p53, hTERT), reducing the required number of mutations required for tumorigenesis. As a result, HPV+ HNC require fewer mutation events than their HPV negative counterparts, which could promote earlier tumor formation in HPV+ disease256,257. Additionally, the viral oncogenes create a fertile field that promotes genome instability, providing the secondary mutations required to initiate tumor development.

Understanding the molecular underpinnings of HPV-mediated transformation will allow the development of novel prognostic and therapeutic approaches in HNC. HPV tumorigenesis is intimately connected with DNA repair pathways, and, in general, HPV hyperactivates the upstream components of these pathways but suppresses the actual repair of DNA lesions. However, there is a critical gap in our understanding of DNA repair in HPV+ HNC: it is unknown if DNA repair is generally suppressed or if specific

43 repair pathways are targeted. This is an important dual mechanism to understand as it potentially explains why HPV+ HNC are more sensitive to chemoradiation therapy and could be exploited to further sensitize HPV+ HNC to conventional or novel therapies.

Additionally, a conceptual framework of DNA repair in the context of HPV+ HNC would enable future work to probe the mystifying insensitivity of HPV negative HNC to current therapy.

Novel therapeutics for the prevention of chromosomal instability due to mitotic errors may be worth considering. A possible new drug target is PLK4 which plays a role in centrosome duplication as described above. New small molecule inhibitors of PLK4 have been identified and it will be interesting to determine if they can prevent or dampen genomic instability through decreasing the number of multipolar cells. However, this must be balanced against the consideration that mitotic de-regulation by HPV might be responsible for elevated tumor responsiveness to chemotherapies that target mitosis, such as cisplatin and 5-FU.

Integration of the viral genome is an important event in cervical cancer progression, but the necessity of integration is much less certain in HNC. As our understanding of HPV+ HNC comes of age, the progression of these tumors will likely be classified and treated differently when compared to their HPV negative counterparts.

As such, it will be critical to understand at what point during tumor development integration occurs, the frequency of integration in HNC, and whether such integration enhances the aggressiveness of HNC as is believed to be the case for cervical cancer.

If there is a subset of HPV+ HNC found to maintain episomal viral genomes, then these

44 tumors could targeted with a Brd4 inhibitor, such as JQ1, to prevent viral genome replication and mitotic segregation142,258.

Several studies have shown that E6 and E7 are capable of regulating cytokine release, and E5 can dramatically alter the plasma membrane architecture and associated signaling pathways. E6/E7 work cooperatively to prevent immune cell stimulation by hindering the secretion of IL-8, IL18, and other factors (Tables 2.1 and

2.2). While not yet demonstrated in literature, it is possible that these oncogenes also upregulate immunosuppressive factors which could be useful as biomarkers of HPV infection. Additionally, while the E5 oncogene has received less attention in the literature, its targets may be the most druggable. E5 changes the plasma membrane composition by increasing the number of lipid rafts259, and changing the membrane receptor population. Most of the known functions of E5 occur through enhancing EGFR signaling (Table 2.3), and current clinical trials show that EGFR inhibitors may have potential as a substitute for cisplatin in HNC therapy (see meta-analysis of clinical trials by Riaz et al., 2013260). However, there are three mechanistic questions regarding cetuximab that could identify the most optimal patient population to receive the drug: how frequently does HPV integrate in HNC, is E5 predominantly lost in cells with integrated HPV16 genomes, and does E5-mediated protection of internalized EGFR protect cells from therapy? Another potential target of downstream E5 signaling is COX-

2, which mediates and magnifies inflammation and angiogenesis phenotypes through

PGE2138. This pathway would be straightforward to bring to the clinic as NSAIDS are

FDA approved and have a manageable side-effect profile. Lastly, E5 is a transmembrane protein ca.pable of hijacking the entire intracellular membrane complex,

45 simultaneously preventing the degradation of proteins like EGFR and halting the maturation of other receptors (Table 2.3). It is highly probable that other membrane proteins are affected and, once their identities are determined, these could be targeted in future clinical trials.

As the HNC field continues to investigate the HPV+ subset of HNC, it is important to remain mindful of the limitations in our understanding of the HPV oncogenes. The majority of the functional discoveries for these oncogenes were uncovered in mouse fibroblasts, cervical cancer cell lines, and foreskin-derived keratinocytes. While the best- described functions of E6 and E7 are conserved between cervical cancer and HNC cells, their unique developmental origin, environment, and stressors of the oropharyngeal mucosa may permit the development of new HPV-oncogene specific phenotypes not seen in healthy integument or cervical cancer.

Acknowledgements This work was funded by two NIH awards: RO1 CA116316 and

CA102357. Special thanks are given to Timothy Chlon for his critical input and discussion of the manuscript.

Oncogene Pathway Molecular Targets Functional Outcomes Prevents apoptosis by targeting p53 for destruction via E6AP ubiquitin ligase. Binds CBP and p300 to prevent p53, E6AP 69,72; IRF3 203; CBP, p300 71; IL-18 cotransactivation of p53 target genes. Can both sensitize Apoptosis and 201; IL-8 200; MCP-1205; Survivin, IAP2, FADD, and protect cells from extrinsic immune-mediated Immunity Evasion caspase 8, c-Myc, BAK, BAX, TNRF1 261; IL-1β apoptosis and intrinsic mitochondrial apoptosis by 206 targeting pathway members. Degrades IL-1β produced by infected cells through E6AP and represses IRF3 transcriptional activity. Increased hTERT expression as a result of degradation Telomere 79 81,82 of the hTERT transcriptional repressor, NFX1. May also hTERT ; NFX1, Maintenance increase telomerase activity by direct binding with hTERT 83. E6 MUPP1 189, DLG 190; Fibulin 1 262; SCRIBBLE, Anchorage independent growth, loss of cell polarity, and Cytoskeleton and 191 192 193 MAGI1, MAGI3 ; PTPN13 ; Vimentin ; N- hyperplasia as a result of E6 mediated degradation of Cell Mobility 116 84 cadherin ; SLUG, TWIST, ZEB1, ZEB2 most of these targets. Induces N-cadherin expression. Disables BRCA1 inhibition of telomerase and c-Myc 87 86 213 88 function. Binds ATR and XRCC1 and may interfere with DNA Damage Repair XRCC1 ; BRCA1 ; BARD1 ; ATR ; base excision repair and other single strand break repair pathways. Centrosome accumulation, potentially causing 236 chromosome breaks and failed mitosis during cell Mitosis Centrosome ; division. Polyploidy from failed cell division has been described 263. Cell Signaling 14-3-3ζ, PKA 264 PKA phosphorylated form of E6 can bind to 13-3-3ζ.

3Table 2.1: Molecular Targets of the E6 HPV Oncogene at a Glance

Oncogene Pathway Molecular Targets Functional Outcomes Increased proliferation by suppressing inhibitory G1/S 91 102 103,265 checkpoint and CDK2 regulation. RB family proteins are pRB/p105, p107, p130 ; p27 ; p21 ; 105,106 97,98 266 inhibited and targeted for destruction, promoting E2F Cell Cycle CDC25A ; p300, pCAF ; p600 ; cyclin 104 267 target gene transcription. CDC25A levels increased to A/CDK2, cyclin E/CDK2 ; miR-205, miR-24 activate CDK2. E7 binds p21, p27, and the cyclin A/CDK2 and cyclin E/CDK2 complexes to prevent CDK2 inhibition. Modulation of histone acetyltransferases and histone Chromatin 99 97,268 100 HDAC-1, HDAC-2 ; p300, pCAF ; DEK deacetyleaces to remodel chromatin and alter gene Modification expression. Altered transcription through direct, pRB independent 96 97,98 95 interaction with E2F1, binding and inactivation of E2F6, Transcription E2F1 ; p300, pCAF ; E2F6 and binding/disruption of the p53 transcriptional coactivators p300 and pCAF. 194 118 Disrupts regulation of actin polymerization, upregulates Fibronectin ; Gelsolin ; N-cadherin, E- Cytoskeleton and 116 classical EMT factors, and downregulates the E-cadherin cadherin, Vimentin ; SLUG, TWIST, ZEB1, Cell Mobility 84 117 adhesion factor. Inhibition of p190RhoGAP by E7 may ZEB2 ; p190RhoGAP limit cell migration. Alternative E7 109 Evasion of senescence by replicating telomeres with Lengthening of FANCD2, MUS81, BRCA2 homologous recombination DNA repair machinery. Telomeres Increase in γH2AX and DNA repair checkpoint signaling; 112 86 100 213 Attenuates DNA repair checkpoint response and γH2AX ; BRCA1 ; DEK ; BARD1 ; DNA Damage Repair 109 111 90 114 promotes mitotic entry by promoting claspin turnover; FANCD2 ; Claspin ; ATM ; Stat-5 Binds to and may directly activate ATM; Activates ATM through Stat-5; Increased expression of FANCD2. 200 205 207 269 Prevents release of pro-inflammatory cytokines and IL-8 ; MCP-1 ; IRF-9 ; NF-κB ; IL- Immunity Evasion 202 interferons by upregulating IL-18BP, attenuating NF-κB, 18BP ; and inactivating the IRF-9 transcription factor. Centrosome synthesis, potentially causing chromosome breaks and failed mitosis during cell division. Polyploidy 271 221,236 270 251 from failed cell division has been described . Binds Mitosis Centrosome ; CENP-C ; Cdt1 centrosome protein C, potentially disrupting kinetochore function. Cdt1 can induce 4N polyploidy through bypassing mitosis. 272 Increased stability of Src and Yes by posttranslational Cell Signaling Src, Yes modifications. 4Table 2.2: Molecular Targets of the E7 HPV Oncogene at a Glance

Gene Pathway Molecular Targets Functional Outcomes Promotes proliferation, anchorage independent growth, immortalization, inflammation, and angiogenesis by 125 195 prolonging EGFR signaling through the ERK1/2 and AKT EGFR ; NF-1, c-fos, c-jun ; AKT, Epidermal Growth Factor 135 273 pathways. Interacts with NF-κB to drive COX-2 expression ERK1/2, VEGF ; Erb4 ; NF-κB, COX- Receptor Signaling 136 and subsequent prostaglandin (PGE2) production. Promotes 2, PGE2 ; transcription of c-fos, c-jun, and other genes carrying NF-1 responsive elements. Binds and disrupts function of Erb4, a family member of EGFR. EGFR independent signaling PLC-1144; ERK1/2, MAPK family145; Enhances ERK1/2 and MAPK signaling with attenuated 274 E5 cascades TGFβ1 TGFβ1 signaling. Interrupts MHC I maturation by binding BAP31 and HLA-1 in 275 151 150 the ER/Golgi, preventing antigen expression by affected cells. Connexin 43 ; Calnexin ; BAP31 ; Vesicle Transport and 147 132 Degrades CD-1d, reducing CD-1d based cytokine production CD-1d ; V-ATPase , endosome Immune Evasion 133 and natural killer T-cell stimulation. Can alter the pH of pH ; endosomes by disrupting V-ATPase function and through E5’s inherent ion channel properties. Represses p21 and p27 gene expression and promotes BAX Cell Cycle and Apoptosis p21140; p27141, BAX143; MAD2, BUB1142 degradation. Disrupts mitotic spindle checkpoint by decreasing expression of BUB1 and MAD2.

5Table 2.3: Molecular Targets of the E5 HPV Oncogene at a Glance

Gene Pathway Molecular Targets Functional Outcomes Represses transcription of E6 and E7 by binding with Brd4, E6/E7 Transcriptional E6/E7155,156; Brd4168,169; EP400, and association with EP400 and SMCX; E2 function is lost Repression SMCX172; after viral integration into the genome, and E6/E7 repression is removed.

Viral Episome to 176 Complexes with E2 to tether the viral episome to mitotic Brd4 Chromosome Tethering chromosomes during cell division. E2 CCHCR1 competes with Brd4 for binding E2 and can limit Proliferation and Antagonism 175 CCHCR1 Brd4/E2-based repression of E6/E7. This interaction can of E6/E7 Repression increase proliferation and loss of differentiation

276 Downregulates expression of the stimulator of interferon Immune Evasion IFN-kappa, STING genes (STING) and interferon kappa Cell Metabolism Mitochondrial Membrane, UQCRC2177 ROS release, HIF-1α stabilization, and promotes glycolysis.

6Table 2.4: Molecular Targets of the E2 Viral Oncogene Suppressor at a Glance

Chapter 3: The biochemistry of the DEK oncoprotein and its roles in

DNA repair and chemotherapy resistance

Introduction

The human DEK gene, located on chromosome 6p22.3, encodes the 42kDa DEK oncoprotein. This factor was first identified as a t(6;9) DEK-NUP214 fusion protein in

AML277, and the endogenous DEK protein has since been found upregulated in most other forms of cancer, where it is linked to higher stage tumors, chemotherapy resistance, and worse patient outcomes (citations in Table 3.1). Due to the near- ubiquity of DEK overexpression in cancer, many efforts have been made to understand how the biochemical functions of DEK drive cancer progression and to develop it as a prognostic biomarker and therapeutic target. This chapter and chapter 4 focus on the

DNA repair functions of the oncoprotein, which likely contributes to chemotherapy resistance seen in high DEK-expressing tumors278,279. Considering that chemotherapy resistance impacts therapy response and prognosis, understanding DEK biochemistry is especially important for OPSCC as DEK is an E2F-driven factor that is highly upregulated by the HPV E7100 oncogene.

DEK is a highly conserved protein that first appeared in multicellular eukaryotic life and has stringent evolutionary pressure in mammals. To illustrate this, mice and humans share approximately 83% of genetic sequence identity, including introns.

However, a single mechanism that satisfactorily describes this selection pressure remains elusive. One potential explanation is that DEK operates as a multifunctional protein to facilitate a wide array of cellular processes, which is supported by our current understanding of the protein’s activities, as illustrated in Table 3.2. Currently, DEK has no known paralogs or enzymatic activity, making it a challenge to predict the protein’s functions or to develop therapeutic inhibitors280.

The biochemistry and post-translational modifications of DEK

Previous studies have determined that 90% of the protein is localized to the nucleus281, and, in acellular experiments, purified DEK can directly bind and supercoil plasmid DNA282,283. This DNA-binding ability was localized to two DEK regions: a central region containing a SAP domain, spanning amino acids 92-187, and a novel C-terminal domain (DEK-C), which spans amino acids 309-375 (Figure 3.1)284-286. Partial NMR structural analysis revealed that the first region contains tandem ψSAP and SAP DNA binding domains286. While the SAP domain is found in many chromatin architectural proteins287, the pseudo SAP (ψ-SAP) domain bears structural but not sequence homology to SAP, and the combination of the ψ-SAP-SAP domains appears to be unique to DEK. Based on the tertiary structure, ψ-SAP-SAP is also hypothesized to expose hydrophobic, protein-interacting residues after binding to DNA286. The other

DEK DNA binding domain, DEK-C, has a structure similar to a winged helix motif.

Structural data suggests that DEK-C also reveals hydrophobic residues after binding

DNA, which may describe the DEK self-multimerization activity associated with DEK-

C284,285. In the absence of any known kinase or enzymatic function, DNA supercoiling by

DEK most likely occurs through DEK multimerization284.

With the exception of an HIV viral sequence288, DEK does not appear to bind chromatin in a sequence-specific manner. Instead, the protein strongly interacts with histones289,290 and, in acellular studies, avidly binds cruciform DNA structures291.

Cruciform DNA arises from inverted genomic sequence repeats, which have roles in transcription regulation. This DNA structure can form during stalled replication fork

53 repair and between sister chromatids in homologous recombination (HR) DNA repair292.

It is possible that some cellular DEK functions are related to its ability to bind cruciform

DNA, but this has not yet been demonstrated in cells.

The affinity of DEK for nucleic acids is regulated by multiple types of posttranslational modifications (PTMs). Phosphorylation by casein kinase II (CKII) and

Protein Kinase C (PKC) decreases DNA affinity293, and this is reversed by protein phosphatase 2A294. DEK can also be extensively acetylated, which can reduce its affinity for promoters and sequester it in interchromatin granule clusters, regions where

RNA splicing occurs280,295. While small amounts of ADP-ribosylation by PARP1 can weaken the DNA binding affinity of DEK296, extensive poly ADP-ribosylation and phosphorylation removes DEK from the chromatin and marks the protein for cellular excretion. However, this latter process has only been observed in apoptotic cells297.

Interestingly, macrophages have been shown to secrete a form of DEK which acts as a chemotactive factor for immune cells298. This form of DEK can be internalized via heparin sulfate and be functionally imported into the nucleus by DEK-deficient HeLa cells299. However, the PTMs of this DEK species and their importance to internalization are unknown. Unfortunately, our understanding of DEK PTMs and the responsible enzymes remain in its infancy as many types of PTMs, such as sumoylation, have yet to be examined. Furthermore, the essential residues for each type of modification are unknown, necessitating future mutagenic studies. Considering that the 375 amino acid oncoprotein is replete in modifiable residues (67 lysines, 42 serines, 19 threonines, 15 arginines, and 5 tyrosines), PTMs likely have an important role in dictating DEK function

54 and have the potential to explain how DEK is involved in such a diverse number of cellular functions.

DEK regulates gene expression through chromatin dynamics and RNA processing

One of these cellular functions is chromatin and epigenetic regulation. Multiple studies have identified that DEK is a central regulator of chromatin dynamics. This is best exemplified in cell models where DEK loss results in dramatic expansion or shrinkage of heterochromatin, depending on the system289,300,301. Mechanistically, DEK may accomplish this extensive chromatin remodeling through its interactions and functions with histones, histone-modifying proteins, and pro-heterochromatin factors.

DEK has been found to interact with H2A, H2B, H3, and H3.3289,290,302,303; is a histone chaperone for H3.3289,302; restricts H3.3 incorporation into PML bodies and telomeres302; and promotes heterochromatin stability through interactions with HP1α and H3K9me3300.

Gene transcription is also positively and negatively regulated by DEK and depends on the interplay between DEK, epigenetic modification, and recruitment of transcription factors. For instance, DAXX and DEK interact and collaborate to recruit HDAC II to repress transcription via histone deacetylation303, and, alternatively, DEK can interact with and stimulate transactivation of the AP-2α transcription factor304. Conversely, transcription initiation can be blocked simply by DEK being bound to chromatin, from which it must be removed by SET and PARP1 to provide access for transcriptional machinery296. Together, these chromatin-modifying functions likely play an important

55 role in gene transcription and heterochromatin maintenance as DEK-deficiency profoundly changes these landscapes in cells289,300,301,305.

Gene expression for many molecular pathways is influenced by the transcriptional functions of DEK (Table 3.2). In addition to the epigenetic regulation described above, the oncoprotein also recruits or inhibits transcription factors such as

C/EBPα306, p300290,295, HIF-1α307, PCAF290, and the human T-cell virus leukemia Tax protein308. In general, interaction of DEK with these factors can promote oncogenic phenotypes. For instance, VEGF-mediated tumor angiogenesis was found to be promoted, in part, by DEK-mediated recruitment of p300 and HIF-1α to the VEGF promoter307. In some instances the interaction between DEK and transcription factors may be more complex, as is the case for p300/PCAF. In one scenario DEK inhibits the acetyltransferase activity of PCAF and p300290. This is a well-characterized complex that acetylates and activates p53290, and, although this has not been directly established, it is possible that this is one mechanism by which DEK can disrupt p53- mediated apoptosis and senescence309,310. However, DEK is also acetylated by p300/PCAF, which sequesters DEK to interchromatin granule clusters295. It is still unclear if acetylation affects the ability of DEK to inhibit p300/PCAF activity or if other factors are involved in determining the outcome of a DEK/p300/PCAF interaction.

In addition to regulating gene expression through chromatin dynamics, DEK also has post-transcriptional regulatory abilities. The first function described was intron removal, where DEK interacts with U2AF to coordinate accurate RNA-splicing311.

Recently, a second method of regulation was discovered, wherein DEK acts as an IRES trans-activating factor and is required for the IRES-dependent expression of LEF1, a

56 pro-invasion factor in breast cancer312. Considering these intimate roles in RNA processing and translation, it is likely that many genes are regulated in this manner by

DEK in addition to or separately from epigenetic regulation.

DEK drives oncogenic pathways and chromosome instability

DEK promotes many oncogenic pathways, although a mechanism between DEK and gene expression in these cases is often not established. For example, apoptosis can be inhibited by DEK-driven expression of BCL-2313 or MCL-1314, depending on the model system, and DEK is required for robust IRAK1 expression in inflammation signaling305. In both cases the mechanism by which DEK mediates gene expression is unclear. Similarly, the oncogene has been shown to influence expression of factors involved in cell-cell signaling. DEK is a downstream target of the Ron receptor and promotes the β-catenin/WNT signaling circuitry. As a consequence of β-catenin activation, nearby cells are induced to replicate through paracrine WNT4, WNT7b, and

WNT10b secretion315, which in turn drives proliferation, migration, invasion, and stem- cell like phenotypes316. The precise mechanism of how DEK stimulates or enhances the

β-catenin/WNT cascade is still under investigation.

One of the most prominent phenotypes of elevated DEK expression is the promotion of cell proliferation, which is lost following DEK knockdown212,300,315,317,318. In addition to coordinating gene expression, DEK also appears to have direct roles in regulating and supporting DNA replication. Interestingly, the biochemical in vitro models of DNA replication show that DEK reduces the efficiency of chromatinized SV40 genome replication282. This may be due to an absence of factors such as PARP1 or

SET that remove DEK from the chromatinized DNA to allow access for replication machinery296. In cultured cells DEK loss impairs stalled replication fork restart and results in accumulation of γH2AX and DSBs after treatment with chemotherapy212,317.

With regard to replication fork restart, it is uncertain if DEK promotes repair of replication-associated damage via a process like HR or transient recruitment of DNA repair signaling factors to promote fork progression. However, this same study did note an increase in γH2AX foci on mitotic chromosomes in DEK-deficient cells, indicating the presence of unrepaired DNA damage317. This suggests that DEK is necessary for reducing the DSB burden in G2/S prior to entering mitosis.

Under normal circumstances, DEK is removed from mitotic chromosomes from prophase through late anaphase, and re-associates with DNA in telophase319.

Association of DEK with telophase chromosomes has also been reported previously, although other stages of mitosis were not examined320. A recent study from our laboratory demonstrated that forced overexpression results in the retention of DEK throughout mitosis and an increased frequency of micronuclei, nuclear buds, and anaphase bridges319. These latter cellular aberrations were seen across multiple immortalized and transformed cell lines where DEK was exogenously overexpressed319, and are considered to be the result of mitotic mis-segregation events and improperly or unrepaired DSBs321. This indicates that DEK overexpression can have a mutagenic role in cancer, where it is most commonly overexpressed.

DEK is implicated in DNA double strand break repair

DEK is necessary for chemotherapy resistance

While DEK overexpression is associated with mitotic non-disjunction abnormalities at the chromosome level, the protein also has clear activity in reducing

DNA double strand break (DSB) burden in cells and in promoting chemotherapy resistance in cancer212,278,279,314,322,323. The association between DEK and poor outcomes across multiple cancer types may be related directly to its necessity for genotoxic therapy resistance. Thus, studies aimed at understanding the function of DEK in DNA repair, such as in Chapter 4, are essential for developing novel CRT augmenting therapeutics that target DEK and for predicting the prognostic utility of a

DEK biomarker.

DEK was first discovered to play a role in DNA-repair when the c-terminal domain was found to partially restore radiation resistance in ataxia-telangiectasia cells324. Since then several studies have shown that DEK loss sensitizes cells to chemotherapeutics across multiple cell models in different species212,314,317,322.

However, the exact mechanism for how DEK promotes chemotherapy resistance is not well understood. Depending on the cell model, DEK loss can result in either an increase in endogenous γH2AX, a histone marker of DSBs212 or no change317. However, most cell models agree that the absence of DEK results in enhanced γH2AX labeling following chemotherapy treatment. Comet assay evidence published by our laboratory demonstrates that an increase in DNA breaks is the likely cause of increased γH2AX212.

The same study suggests that DSB burden in DEK-deficient cells is derived from a mild decrease in non-homologous end joining (NHEJ) DSB repair pathway function.

However, new evidence described in Chapter 4 indicates that DEK-deficient cells rely

59 on the residual NHEJ for survival since HR repair is compromised. Thus, the decrease in NHEJ activity does not explain sufficiently the chemosensitivity effects of DEK loss.

DEK loss attenuates non-homologous end joining DSB repair (NHEJ)

NHEJ is a constitutively active and rapid DSB repair process that corrects the majority of DSBs throughout the cell cycle, especially in G0/G1325. The pathway can be broken down into three individual stages (Figure 3.2). The first stage of repair is DSB recognition: a Ku70/80 heterodimer avidly binds each DSB and keeps the ends in close proximity326. The Ku70/80 complex is then rapidly bound by DNA-PKcs, which undergoes autophosphorylation and recruits NHEJ factors necessary for downstream end-processing327. Most often the DSBs are not clean breaks and require processing before ligation. Factors involved in this step are thoroughly reviewed by Menon and

Povirk328 and include abasic endonucleases (APE1, APE2), endo/exonucleases

(Artemis, APLF, Exo1, Metnase WRN), DNA phosphodiesterases (Aprataxin, PNKP),

DNA polymerases (Pol λ and Pol μ), and helicases (WRN). After compatible blunt ends are produced, a core ligase complex consisting of XRCC4, XLF, and Lig4 is recruited on each side of the DSB and anneals the break.

While NHEJ can often repair DSBs without inducing mutation, it is considered an error-prone repair pathway by the nature of DSB end processing and inability to utilize a homologous template during repair. The most common type of mutation associated with

NHEJ repair arises from exonuclease and polymerase activity that generates blunt ends. Since there is no proofreading mechanism, short nucleotide insertions and deletions can be incorporated into the genome327. While the majority of the genome

60 may not be affected by these mutations, it can create deleterious frameshift or nonsense mutations in coding sequences. This inherent mutation risk with NHEJ repair is exploited by TALEN or CRISPR/Cas9 gene knockout approaches329. In humans but not mice, NHEJ can lead to translocations if two DSBs have their ends aligned improperly330. Despite the risk of mutation, it is estimated that over 80% of G2 phase

DSBs are repaired by NHEJ331, suggesting that fast and efficient repair of a potentially lethal lesion is preferred over sequence integrity.

A previous study by our laboratory determined that DEK was required for robust

NHEJ activity, using reporter constructs that express GFP only after successful NHEJ repair. Loss of DEK resulted in a 40-60% decrease in pathway function compared to control cells. The repair defect was localized to the DNA-recognition phase of repair as

Ku70/80 recruitment to DSBs was delayed and DNA-PK enzymatic activity was decreased212. However, the molecular mechanism by which DEK supports rapid and robust NHEJ function remains unknown. Further, the potential in vivo consequences of attenuated NHEJ, such as deficient immunoglobulin (Ig) class switching in B-cells332, and the impact of DEK on other DSB pathways that could potentially compensate for

NHEJ deficiency were also unexamined until the study in Chapter 4. In that study it was found that DEK-loss did not affect Ig class switching but instead severely attenuated HR repair, the major compensating pathway for NHEJ.

Homologous recombination repair can mediate chemotherapy resistance in cancer

HR is a S/G2 cell cycle-specific DNA maintenance pathway that repairs both

DSBs and replication forks that have stalled or collapsed into DSBs. HR is considered

61 the most error-free DSB repair mechanism as it utilizes the nascent sister chromatid formed during DNA synthesis as a repair template333. An overview of the repair pathway is represented in Figure 3.3. The first stage of HR repair involves a signaling cascade that is initiated by MRE11-RAD50-NBS1 (MRN) binding to both ends of the DSB. MRN then stimulates ATM dimer recruitment334, monomerization and self-activation by autophosphorylation on S1981335,336. Active pATM proceeds to phosphorylate H2AX, producing the γH2AX DSB marker. While not essential for HR progression, γH2AX promotes timely and efficient HR by driving further pATM activation and function337.

Following signal amplification through γH2AX, the end-resection stage of HR is conducted by the MRN-CtIP complex, which performs 5’-3’ resection of DNA on either side of the break. This produces a 3’ single strand DNA (ssDNA) probe that is immediately protected from cellular exonucleases by the RPA heterotrimer. The subsequent stage of HR involves RAD51 loading onto the resected ssDNA and RPA removal. A complex comprised of BRCA1, BRCA2, and RAD51 homologues are required to overcome RPA’s avid affinity for ssDNA and promote efficient RAD51 loading. The ssDNA-RAD51 filament is essential for the final steps of HR, wherein

RAD51 mediates invasion of the sister chromatid to find homology specific to the 3’ ssDNA probe. Strand invasion results in a transient triple-helix D-loop structure that, as

DNA polymerases elongate the probe with the sister chromatid sequence, evolves into a cruciform-like DNA structure. Eventually, the D-loop dissolves prior to final annealing and ligation333,338,339.

While HR has a clear role in preventing the accumulation of pro-oncogenic mutations in healthy cells, previous studies have found that HR’s high-fidelity repair is

62 an important contributor to chemotherapy and radiation resistance in tumors340,341.

Clinical studies and cell models have demonstrated that cancers deficient in HR are more sensitive to chemotherapy and radiation compared to those that either maintain or regain functional HR342-346. Additionally, chemotherapy resistance has been attributed to

RAD51 overexpression in multiple types of tumors347-350, an attribute that is also shared with DEK178,278. Considering that HR is well understood to promote chemo- and radiotherapy resistance to tumor cells, it is reasonable that DEK overexpression can contribute to treatment resistance via its functions in HR, as described in Chapter 4.

Competition and compensation between NHEJ and HR

Understanding the factors and mechanisms involved in DSB repair pathway choice is an active area of research and debate338,351,352, especially with the dawn of

CRISPR/Cas9 gene editing technology that makes use of these pathways353,354. The following is a current model of how DNA repair choice is made at a DSB. In G0/G1 phase of the cell cycle, NHEJ is the predominant repair mechanism, but competition between HR and NHEJ occurs in the S/G2 phases of the cell cycle, when HR is active and has the sister chromatid template available325. If the lesion is amenable to both forms of repair, then repair choice is dictated either by initial DSB end-protection by

Ku70/80, which blocks resection and commits repair to NHEJ, or by MRN-CtIP 5-3’ end resection, which abrogates Ku70/80 affinity and initiates HR repair338. Competition between 53BP1 and BRCA1 helps to determine which type of repair occurs. 53BP1 is a well appreciated HR antagonizing factor that prevents erroneous HR from occurring in

G0/G1. After being phosphorylated by ATM, 53BP1 forms a complex with RIF1 at sites

63 of DSBs to block BRCA1 recruitment and CtIP resection activity355-357. This is offset in

S/G2 by increased expression of BRCA1, which inhibits 53BP1 phosphorylation by ATM and promotes the accumulation of competing BRCA1-CtIP-MRN complexes onto DNA during S-phase358,359.

While DNA repair pathway choice regulation is still not fully understood, it has been well demonstrated that loss of one pathway results in a compensatory increase in the other. For instance, loss or inhibition of 53BP1 or DNA-PKcs can increase HR repair360-363, and the opposite is true for NHEJ when BRCA1 or a pro-HR Fanconi anemia factor are mutated364-367. The results in Chapter 4 are in line with this compensation model: DEK-deficient cells rely upon NHEJ survival due to the near- ablation of HR DSB repair. Further inhibition of NHEJ via a DNA-PK inhibitor resulted in synthetic lethality, a relationship can potentially be therapeutically exploited.

Conclusions

DEK is a highly conserved oncogene that is upregulated across most cancer types

(Table 3.1), and is a dynamic multi-functional chromatin regulator280. While many of the cellular activities of DEK can be relegated to gene expression regulation at the epigenetic and RNA levels, the protein appears to have a more direct function in processes like DNA repair. Previous work found that DEK is required for optimal NHEJ function and is essential for reducing DSB burden212. However, in Chapter 4 the mild impact on NHEJ repair was found to be insufficient to explain the chemotherapy resistance phenotype associated with DEK, and an essential function for the

64 oncoprotein in HR was described. It is possible that DEK overexpression may be promoting CRT resistance in tumors through its activities in HR repair of DNA damage.

1Figure 3.1: Structure of the DEK oncogene

2Figure 3.2: The non-homologous end joining pathway

3Figure 3.3: The homologous recombination pathway

7Table 3.1: Cancers in which DEK overexpression contributes to worse prognosis or chemotherapy resistance

References linking DEK References linking DEK Cancer Type to worse prognosis and to chemotherapy advanced stage tumors resistance t(6;9)/DEK-NUP214 AML Sandahl JD, 2014368 Bladder cancer Datta A, 2011369 278 Ying G, 2015 278 Breast cancer 370 Ying G, 2015 Liu S, 2012 Cervical cancer Wu Q, 2008371 372 Lin L, 2013 279 Colorectal cancer 279 Martinez-Useros, 2014 Martinez-Useros, 2014 Piao J, 2014373 Gastric adenocarcinoma 374 Ou Y, 2016 Hepatocellular carcinoma Lin J, 2013375 Neuroendocrine prostate cancer Lin D, 2015376 377 Kappes, 2011 314 Melanoma 378 Khodadoust, 2009 Riviero-Falkenbach, 2016 Non-small cell lung carcinoma Liu X, 2016379 Pancreatic cancer Sun J, 2017380 Small cell lung cancer Wang X, 2014381 Serous ovarian cancer Han S, 2009382

8Table 3.2: Known DEK functions and molecular targets

DEK Functions Molecular Targets Functional Outcomes Angiogenesis HIF-1α and VEGF307 Activates VEGF expression by recruiting p300 and HIF-1α to VEGF promoter. Reduces BCL-2 expression in drosophila; Drives MCL-1 expression with no effect on BCL-2 in human melanoma; Interferes with p53 apoptosis functions; Is Apoptosis BCL-2313; MCL-1314; p53309 PARP1297 Poly(ADP-riosyl)ated and shuttled out of the nucleus during apoptosis; Is hypophosphorylated in apoptosis 383; Loss induces apoptosis309,384. Is a histone chaperone that assists in nucleosome assembly; Requires H2A and H2B for topology changes in chromatinized plasmid assays; Targets H3.3 to PML bodies and DEK loss results in improper loading of H3.3 onto chromosome arms; H2A, H2B, H3B, H4282,289; H3.3302; HP1α and Promotes heterochromatin by interacting with HP1α and H3K9me3; Assists Chromatin dynamics H3K9me3300; LANA385 LANA in tethering Herpes Simplex Virus 8 genomes to chromosomes; Promotes DNA supercoiling and cruciform DNA structures in aceullular in vitro studies283,291; Uptake of extracellular DEK reconstitutes chromatin in DEK- deficient cells299. Inhibits differentiation of keratinocytes386 and hematopoietic stem cells387, and loss of DEK promotes differentiation in these cell types; Loss may reduce Differentiation C/EBPα306 C/EBPα-mediated G-CSF granulocyte differentiation306 and may have a function in myeloid differentiation388. Is a transcriptional target of AP2-α, estrogen receptor alpha, and E2F; Affinity for AP2-α389; CKIIα289,293; PKC293; ERα390; DNA is weakened by phosphorylation via PKC and CKIIα and/or poly ADP- Regulators of DEK E2F100; PARP296,297; p300/PCAF295 ribosylation via PARP; requires phosphorylated by CKIIα for histone chaperone ability. Is moved to interchromatin clusters after acetylation by p300/PCAF. Mediates resistance to genotoxic agents and IR212,297,314,322,324; Needed for optimal Ku70/80 recruitment to breaks, DNA-PK kinase activity, and non-homologous end joining repair; Loss increases γH2AX levels with chemotherapy212; Promotes DNA repair DNA-PKcs, γH2AX, Ku70, and Ku80212 chemoresistance212,297,314; Overexpression is sufficient for mitotic genome instability319; Uptake of extracellular DEK reduces γH2AX burden in deficient cells299. Loss increases γH2AX levels with DNA repair stress; Inhibits SV40 plasmid DNA replication γH2AX317 replication efficiency282, but promotes stalled DNA replication fork progression in cells317. Histone interactions and H2A, H2B, H3, H4, and HDAC II303; p300 Interacts with histones and HDAC II. Inhibits histone acetylation of H3K9, H4K5, modifications and PCAF290 H3K14, H4K8, H4K12, and H4K16 by interacting with p300/PCAF290,313; Enhances IRAK1 expression and dual loss of DEK and IRAK1 enhance cell Inflammation IRAK1305 death; Is a chemotactic and arthritic factor when secreted298,391. Invasion and migration E-cadherin and β-catenin316,392; RhoA393 Decreases E-cadherin and stimulates β-catenin signaling when overexpressed;

Increases E-cadherin and inhibits β-catenin signaling when lost; Loss reduces RhoA activation. Dissociates from chromatin and may be degraded during prophase, returns to chromatin in telophase; Co-localizes with anaphase bridges, chromosome Mitosis Mitotic chromosomes319 fragments, and micronuclei; Overexpression is sufficient for micronuclei formation and abnormal retention of DEK on chromosomes throughout mitosis. Oncogenesis and tumor Is upregulated by E7; Potentiates E6/E7 transformation of cells 384; DEK-NUP214 HPV E7394; DEK-NUP214395 development fusion protein is sufficient for AML induction Promotes proliferation314,315,318. A potential mechanism is Wnt upregulation in Proliferation Wnt 4, Wnt 7b, and Wnt10b315 Ron-receptor positive cells Senescence Inhibits senescence394, mechanism unknown. Important for loading of H3.3 and ATRX onto telomeres; Co-localizes to Telomere stability ATRX and H3.3302; TAX308 telomeres302; Recruited to hTERT promoter, works with TAX to repress hTERT expression. Interacts with DAXX, HDAC II, and AP-2α to regulate transcription; Functions as Transcription, RNA AP-2α304; DAXX, HDAC II303; Lef1312; SET an IRES-trans-activating factor for LEF1 translation; Removed from chromatin by processing, and translation and PARP1296; U2AF311 SET and PARP1 to allow transcription; Facilitates intron removal by interacting with U2AF;

Chapter 4: DEK is required for homologous recombination repair of DNA breaks

Eric A. Smith1, Boris Gole2, Nicholas A. Willis3, Rebeca Soria4, Linda M. Starnes4, Eric

F. Krumpelbeck1, Anil G. Jegga1, Abdullah M. Ali5, Haihong Guo6, Amom R. Meetei5,

Paul R. Andreassen5, Ferdinand Kappes6, Lisa M. Privette Vinnedge1, Jeremy A.

Daniel4, Ralph Scully3, Lisa Wiesmüller2, Susanne I. Wells1,*

1Division of Oncology; Cincinnati Children’s Hospital Medical Center; Cincinnati, OH, 45229; USA.

2Department of Obstetrics and Gynecology; Ulm University; Ulm, 89075; Germany.

3Department of Medicine, Division of Hematology-Oncology and Cancer Research Institute, Beth Israel

Deaconess Medical Center and Harvard Medical School, Boston, MA 02215.

4Chromatin Structure and Function Group, The Novo Nordisk Foundation Center for Protein Research,

University of Copenhagen, Copenhagen 2200, Denmark.

5Division of Experimental Hematology and Cancer Biology; Cincinnati Children’s Hospital Medical Center;

Cincinnati, OH, 45229; USA.

6Institute of Biochemistry and Molecular Biology; Medical School, RWTH Aachen University; Aachen,

52074; Germany.

ABSTRACT

DEK is a highly conserved chromatin-bound protein whose upregulation across cancer types correlates with genotoxic therapy resistance. Loss of DEK induces genome instability and sensitizes cells to DNA double strand breaks (DSBs), suggesting defects in DNA repair. While these DEK-deficiency phenotypes were thought to arise from a moderate attenuation of non-homologous end joining (NHEJ) repair, the role of DEK in

DNA repair remains incompletely understood. We present new evidence demonstrating the observed decrease in NHEJ is insufficient to impact immunoglobulin class switching in DEK knockout mice. Furthermore, DEK knockout cells were sensitive to apoptosis with NHEJ inhibition. Thus, we hypothesized DEK plays additional roles in homologous recombination (HR). Using episomal and integrated reporters, we demonstrate that HR repair of conventional DSBs is severely compromised in DEK-deficient cells. To define responsible mechanisms, we tested the role of DEK in the HR repair cascade. DEK- deficient cells were impaired for γH2AX phosphorylation and attenuated for RAD51 filament formation. Additionally, DEK formed a complex with RAD51, but not BRCA1, suggesting a potential role regarding RAD51 filament formation, stability, or function.

These findings define DEK as an important and multifunctional mediator of HR, and establish a synthetic lethal relationship between DEK loss and NHEJ inhibition.

INTRODUCTION

The DNA-binding and chromatin-regulating DEK oncogene is expressed across multicellular eukaryotes and is highly conserved in mammals. Strong sequence conservation of the ψSAP-SAP and C-terminal DNA binding domains, as well as the lack of any known DEK paralogs, suggest stringent evolutionary pressure on this gene280,320. However, despite extensive biochemical, cellular, and clinical investigations into the DEK protein, molecular functions that explain this selective pressure remain poorly understood, as does the frequent over-expression of DEK in human cancers278,318,369,372,376,381.

In cultured cells, DEK has functions in chromatin remodeling282,290,300, DNA replication282,317, and mRNA splicing311. Depending on the experimental system chosen,

DEK loss attenuates distinct oncogenic phenotypes such as proliferation314,316,394, survival297,309, and chemoresistance212,297,314. Dek knockout mice are viable, smaller in size, and resistant to macroscopic tumor development318,384. Biochemically, there are no known enzymatic functions associated with DEK, but the protein self-multimerizes, induces positive supercoils in DNA through the ψ-SAP-SAP and C-terminal DNA binding domains, and preferentially binds cruciform DNA structures283,284,291,396. DEK also has roles in maintaining chromatin architecture282,300 and interacts with histones289,290 and chromatin modifiers300. Since the discovery of the DEK gene as a

DEK-NUP214 fusion protein in AML277 and the discovery of elevated DEK expression in breast278,370, colorectal279,372, lung379,381, and several other types of cancer280, many systems have been used to investigate the pathological consequences of DEK overexpression. A prominent phenotype in cell models is the requirement of DEK for

74 chemotherapy and radiation resistance. For example, expression of the DEK C-terminal domain in ataxia-telangiectasia fibroblasts partially restored radiation resistance, and the loss of DEK conferred sensitivity to DNA damaging agents in multiple cell types212,314,324. Mechanistically, our prior report found that DEK was required for optimal kinase activity of DNA-PK. This kinase is a key mediator of canonical non-homologous end joining (NHEJ), which repairs DNA double strand breaks (DSBs)397, and DEK loss correspondingly suppressed NHEJ212.

The observed NHEJ defects in DEK-deficient cells are unlikely to fully account for the severe sensitivity to genotoxic agents, especially DNA interstrand cross linkers and topoisomerase inhibitors314,322. This suggests additional roles for DEK in genotoxic drug tolerance and DNA repair. A common mechanism by which genotoxic agents induce cell death is through perturbation of replication fork progression 398. A recent study determined that DEK attenuates DNA replication stress317 in a manner similar to RAD51 and FANCD2, factors well known for their function in both homologous recombination

(HR) DSB repair and activities at arrested replication forks399-402.

HR requires the presence of a homologous template, often the sister-chromatid, to be used for repair, and is therefore largely confined to S/G2 phases of the cell cycle.

By copying a homologous DNA sequence, HR is considered an error-free repair process that preserves genome integrity333. This signal cascade is initiated by the DSB sensor, ATM kinase. After localizing to a DSB, ATM autophosphorylates335,403, pATM catalyzes the phosphorylation of CHK2 to inhibit cell cycle progression404, and the H2AX histone to produce gamma-H2AX (γH2AX) epigenetic marks on both sides of the DSB.

The γH2AX mark supports DSB repair by enhancing the recruitment of BRCA1 and key

75 nucleases including the MRN complex and CtIP333,338,405,406. These factors coordinate

DNA end processing into single strand DNA (ssDNA) 3’ tails333,338,405. The resulting ssDNA is initially coated by RPA, which is then efficiently replaced with a RAD51 filament through the combined activities of BRCA1, BRCA2, the RAD51 paralogs, and other factors333,338. This RAD51 filament catalyzes strand invasion, complementary strand annealing, and the formation of a stable synaptic complex with a homologous sequence on the sister chromatid407.

While the cellular regulation and choice between DSB-repair pathways remains incompletely understood367,408-411, NHEJ and HR factors have been shown to be mutually antagonistic as one pathway tends to compensate when the other is compromised338,361. For example, loss of critical HR gene functions in Fanconi Anemia mutant cells results in enhanced dependence upon NHEJ factors364-366 while HR repair efficiency is increased following the disruption of essential NHEJ factors by chemical inhibition or mutation of DNA-PK360-362. Compromising both the NHEJ and HR repair pathways may be synthetic lethal to cancer cells.

In this report, we demonstrate that DEK is required for the repair of DSBs by HR.

Investigation of potential DEK functions in the HR pathway revealed that the protein was important for γH2AX activation, promoted the co-recruitment of BRCA1 and RAD51 to resected ssDNA, and formed a complex with RAD51 in a BRCA1-independent manner.

The dependence of HR on DEK expression was underscored by the synthetic lethal relationship between DEK loss and NHEJ inhibition. Together, these results reveal a novel, multifunctional role for DEK in HR.

RESULTS

DEK loss causes apoptosis in conjunction with DNA-PK inhibitors

We have previously shown that DEK knockout (DEK-/-) mouse embryonic fibroblasts (MEFs) harbor decreased NHEJ activity212. To determine the contribution of

DEK to NHEJ activity in vivo, we measured the concentrations of immunoglobulins in

DEK-/- mouse serum. Class switch recombination (CSR) from IgM to the IgA and IgG classes of immunoglobulins is an NHEJ-dependent process397, but DEK-/- mice were fully competent in producing all classes of immunoglobulins (Figure 4.1a). This data suggests the residual NHEJ activity in the knockout mice is sufficient for CSR under physiologically normal in vivo conditions, despite the sensitivity of DEK-deficient cells to

DNA damaging agents212,314. To determine if DEK-deficient cells require NHEJ for survival, we examined the need for NHEJ by inhibiting the upstream kinase, DNA-PK.

We utilized two well-established DNA-PK inhibitors, NU7026 and NU7441, to prevent

DNA-PKcs autophosphorylation and activation412-415. DEK-/- MEFs treated with the

DNA-PK inhibitor NU7026 displayed dramatically more cell death than wild-type

(DEK+/+) or untreated samples (Figure 4.1b). Analysis of cleaved caspase 3 positive cells by flow cytometry revealed that the inhibitors were sufficient to induce a significant and specific 2-3 fold increase in apoptosis in the DEK-/- cells in the absence of exogenous DNA damaging agents (Figure 4.1c). Similar experiments in HeLa cells infected with a well-characterized DEK knockdown versus control adenovirus vector

(AdDEKsh versus AdGFP)309 revealed a similar 2-fold increase in apoptosis after treatment with either DNA-PK inhibitor, NU7026 or NU7441 (Figure 4.1d-e). These data

77 suggest that DEK-deficient cells, despite their attenuated NHEJ activity212, rely significantly on the remaining NHEJ for survival.

DEK is required for HR DSB repair

Several reports have described an antagonistic relationship between HR and

NHEJ mechanisms of DSB repair355,360,367, and it is generally thought that loss of one pathway will result in compensation by the remaining repair modality360,362,365. Given the dependence of DEK-deficient cells on NHEJ and their inherent sensitivity to chemotherapeutics, we hypothesized that HR may be compromised in DEK-/- cells. To test this we first utilized two established episomal HR reporter systems 362,416. These vectors harbor two defunct EGFP genes, the first bearing an I-SceI meganuclease cleavage site and the second bearing a truncated gene sequence (Figure 4.2a). Co- transfection with an I-SceI expression vector induces a DSB that, when repaired by HR, generates a functional EGFP gene. Both the HR-EGFP/5’EGFP and pHPRT-DR-GFP reporters demonstrated a remarkable and severe loss of HR efficiency in DEK-/- MEFs

(Figure 4.2b). To compare the effects of DEK loss on chromosomal HR repair following direct or replication-dependent DSB induction, we utilized the recently published 6xTer-

HR reporter technology417. The 11CO/47 mouse embryonic stem (mES) cell line bearing a single copy of the 6xTer-I-SceI-GFP reporter cassette integrated at the Rosa26 locus quantifies HR either at an I-SceI endonuclease-induced DSB or at an adjacent Tus/Ter- induced stalled replication fork. Fork stalling is induced by Tus protein binding to an array of six Ter elements, forming a physical barrier that impedes approaching replication fork progression417 (Figure 4.2c). Using this system, we quantified the

78 impact of DEK depletion on error-free short tract gene conversion (STGC), producing

GFP+RFP- cells, and error-prone long tract gene conversion (LTGC), producing

GFP+RFP+ cells through duplication and thus proper alignment of two synthetic exons of

RFP. After co-transfecting Dek siRNA (Figure 4.2D) and I-SceI, we found that siDek treated cells were significantly compromised for both total HR and STGC, and LTGC followed the same trend (Figure 4.2e). This decrease was similar in magnitude to loss of the BRCA proteins418, but not as severe as RAD51 deficiency417. However, there was no significant decrease in STGC, LTGC, or total HR with Tus-induced DNA replication fork stalling (Figure 4.2f). This suggests that DEK, unlike most HR factors417, is dispensable for replication fork associated repair but required for efficient repair of

DSBs.

DEK is necessary for the activation of γH2AX after ionizing radiation

To understand how DEK functions in HR, we examined the consequences of

DEK loss at successive steps in the repair pathway following irradiation (IR) mediated

DSB induction. For these studies we used DEK-/- MEFs and AdDEKsh-infected HeLa cell systems, whose cell cycle progression was reported comparable to that of their respective DEK-proficient controls212,309. Beginning with the relevant upstream DNA damage sensor kinase, we found that ATM was strongly autophosphorylated in

AdDEKsh-treated HeLa cells (Supplementary Figure 4.1a) and functional as indicated by phosphorylation of pCHK2 T68, an ATM substrate (Supplementary Figure 4.1b).

Surprisingly, γH2AX phosphorylation was not enhanced in the AdDEKsh HeLa cells despite robustly activated pATM (Supplementary Figure 4.1a). This held true and was

79 even more severe in the DEK-/- MEF system where DEK was required to increase

γH2AX phosphorylation above baseline by immunofluorescence (IF) (Figure 4.3a). Both the total number of cells harboring γH2AX foci and the average number of foci per cell were severely attenuated from 3-24 hours after irradiation (Figure 4.3b-c). These results were also confirmed by western blot analysis in MEFs and in DEK knockdown

C33A cancer cells (Figure 4.3d-e), suggesting that DEK is an important general contributor to γH2AX phosphorylation in IR-treated cells.

Loading of RAD51 onto RPA-protected DNA is significantly reduced

DSB end-processing occurs downstream of the pATM-γH2AX signal amplification feedback loop and generates ssDNA overhangs that are bound and protected by RPA. End processing is coordinated by several factors, including the MRN complex and BRCA1405,406. Chromatin fractionation of HeLa cells revealed that chromatin recruitment of MRE11 and NBS1, components of the MRN complex, as well as BRCA1 were not affected by DEK status following IR (Supplementary Figure

4.S2a-b). Following end-processing, BRCA1 both co-localizes with and is necessary for robust RAD51 IF foci development405. Through assessing the quality and quantity of these foci, we determined whether DEK loss disrupts RAD51 loading onto RPA-bound ssDNA405. Since our BRCA1 antibody does not detect murine BRCA1, we performed these experiments in HeLa cells. DEK knockdown in HeLa cells had a mild, but statistically significant delay in BRCA1-RAD51 foci formation at 3 hours following IR

(Figure 4.4a-b) with no difference in individual BRCA1 or RAD51 foci formation at any time point (Supplementary Figure 4.S3). By 6 hours the co-localization of BRCA1 and

RAD51 foci was indistinguishable from control cells. To determine if the delay in

RAD51-BRCA1 co-localization correlated with impaired RAD51 loading onto RPA- protected ssDNA, we also determined the kinetics of RAD51 localization to RPA foci. In line with the delay with BRCA1-RAD51 co-localization, we found a significant decrease of RAD51-RPA2 co-localized foci at 3 and 6 hours post irradiation (Figure 4.4C). A 24- hour time course in DEK+/+ and DEK-/- MEFs revealed that RAD51 loading was similarly attenuated (Figure 4.4d-e). There was no difference in RAD51 foci quantity or quality (Supplementary Figure 4.S4a), but RPA2 foci were generally smaller in DEK-/- cells. DEK-/- cells also retained persistent RPA2 foci at later time points, suggesting a

DNA repair defect (Supplementary Figure 4.S4b). In summary, DEK-deficient cells have a small reduction in the recruitment of RAD51 to BRCA1 foci, which results in a mild attenuation of RAD51 loading on RPA-coated ssDNA.

DEK interacts with RAD51

While RAD51 loading was not dramatically affected by DEK loss, physical DEK interactions with the RAD51 recombinase were possible. To test this, pMIEG His-FLAG-

DEK expressing HeLa cells were left untreated or received 1mM hydroxyurea (HU) for

15hrs to induce replication fork stalling and a low level of DSB generation as published previously419, prior to FLAG immunoprecipitation (IP). Under these conditions, we found that RAD51 complex formation occurred with His-FLAG DEK (Figure 4.5a). Validation of the complex was performed by IP of endogenous DEK and reverse IP of RAD51 in parental HeLa cells (Figure 4.5b-c). No BRCA1 interaction with DEK was observed

(Figure 4.5b), suggesting that DEK forms a separate complex from the standard BRCA-

RAD51 homologue apparatus. The DEK-RAD51 complex was also observed in untreated and irradiated cells. Thus, DEK interacts with the HR recombinase RAD51 in the presence or absence of exogenous DNA damage (Supplementary Figure 4.S5).

DISCUSSION

Herein we demonstrate that DEK is required for the repair of DSBs by HR, and regulates multiple steps in the HR cascade by promoting γH2AX activation, enabling robust RAD51 loading, and forming a complex with RAD51 (Figure 4.5d). This importance of DEK in HR was further underscored by the exquisite sensitivity of DEK- deficient cells to NHEJ loss through DNA-PK inhibition. In summary, this is the first report to describe DEK as a HR factor and to identify a synthetic lethal relationship that exists between DNA-PK inhibition and DEK loss.

According to the 6xTer reporter assays (Figure 4.2c-f), DEK appears to be required for HR only in the context of conventional DSBs, wherein DEK-deficient cells had a similar HR deficiency to BRCA1 loss417, but not in the context of stalled replication forks. However, this selectivity differs when DEK activities are compared to those of

BRCA1, BRCA2, and RAD51, factors which are required for the HR repair triggered by both I-SceI induced DSBs and Tus-stalled replication forks417. Based on these observations, it is likely that DEK is expendable for replication-associated HR repair, but this by no means precludes non-HR functions at stalled replication forks. Indeed, published DNA fiber assays have elegantly shown that DEK can promote stalled replication fork restart and reduce γH2AX response to DNA replication stress317.

Differential γH2AX responses to irradiation (Figure 4.3) versus chemotherapy- induced damage212,317 lends further support to a model where DEK is specifically required for HR repair of DSBs. Previous studies have found that loss of DEK enhances

γH2AX activation following treatment with replication fork stalling agents212,317. While similar pATM activation was observed in response to stalling or irradiation, γH2AX

83 phosphorylation was specifically not engaged after IR exposure. This suggests that

DEK is specifically required for ATM-H2AX signal transduction under conditions of

DSBs. While it is known that γH2AX is not strictly required for HR, its loss can slow the kinetics of the pathway and lead to a moderate decrease in repair efficiency337,406. Thus, the attenuation of BRCA1-RAD51 and RAD51-RPA2 foci co-localization (Figure 4.4) is likely a consequence of failed γH2AX activation. These pulldown results, in which DEK interacted with RAD51 in a BRCA1-free complex (Figure 4.5), further support this hypothesis as they suggest DEK is not part of the BRCA1-Palb2-BRCA2-RAD51 complex that facilitates RAD51 loading. However, we cannot presently rule out weak interactions or a role for BRCA in promoting the DEK-RAD51 interaction. Also, since these interactions were identified and confirmed in cervical cancer cells, it is uncertain if the DEK-RAD51 complex is a component of normal cell biology or specific to cancer cells.

Taken together this data support a model in which DEK has two distinct roles in the repair of DSBs by HR. First, DEK has an important function in γH2AX activation by

ATM, but this is unlikely to be sufficient for the degree of HR deficiency observed in

DEK-deficient cells337,406. Therefore, DEK must have a second function involving the

DEK-RAD51 complex. With regard to a possible mechanism of action, DEK might play an early role in homologous recombination by supporting the initial formation of the

RAD51-DNA filament. However, since BRCA1 and BRCA2 play key roles in this step and, since BRCA1 was not detectable in the DEK-RAD51 complex (Figure 4.5b), we do not favor this scenario. Based on previous structural and cell-free studies, DEK binding to DNA stimulates self-multimerization and potential interactions with other proteins284-

286. Furthermore, DEK preferentially binds cruciform over conventional linear DNA structures291, and harbors three putative DNA binding motifs285,286. Thus, one possible scenario is that DEK interacts with both the invading RAD51 filament and the sister chromatid to stabilize resulting D-loop structures. Such activities can now be readily explored through established biochemical assays420.

Since this is the first report to establish a role for DEK in HR, our work opens up multiple avenues for future study. At the molecular level, mechanisms whereby DEK promotes γH2AX activation specifically at DSBs and the nature and function of the complex with RAD51 remain unclear. Secondly, in the absence of known DEK homologues, detailed studies of the evolutionary history and origin of DEK are lacking despite its strong conservation across mammals and presence in plants. Interestingly,

DEK is absent in bacteria and yeast, wherein mechanisms of HR are well described. It is possible that the early multicellular eukaryotes required a novel chromatin regulator to stabilize their increasingly complex genomes during DNA repair. Considering its role in chromatin modification and ATM-mediated γH2AX activation, its affinity for RAD51 and

DNA structures similar to Holliday junctions, and its ability to stabilize DNA breaks for in vitro DNA ligation, DEK is a likely candidate to fill this niche280,291.

Still, further investigations into the activities of DEK in HR and NHEJ repair will remain important for deepening our knowledge of the evident evolutionary pressure on

DEK and HR, understanding the multiple activities of DEK in HR, and for the development of small molecule inhibitors to target relevant functions of this oncogene.

In support of these future therapeutic efforts, we identified a synthetic lethal relationship between DEK loss and canonical NHEJ inhibition, as well as a cellular phenotype that

85 can be assessed for DEK inhibition in vivo. This is important as DEK is largely dispensable for cell division, and DEK-/- mice are healthy and viable309,317,384, suggesting a high therapeutic index for a potential anti-DEK drug. Thus, it is conceivable that DEK overexpressing tumors may regress without a need for adjuvant radiation or traditional chemotherapy if substituted for tumor-specific co-targeting of

DEK and DNA-PK.

MATERIALS AND METHODS

Cell culture, adenoviral infections, and viral transductions

HeLa and C33A cells were grown in Dulbecco’s Modified Essential Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics. MEFs, generated previously212, were cultured in DMEM with 10% heat inactivated FBS, 100μM MEM non- essential amino acids, 0.055 mM β-mercaptoethanol (BME), 2 mM L-glutamine, and

10μg/ml gentamycin. Mouse embryonic stem (ES) cells were grown in ES media on fibroblast or gelatin substrate as previously described417. To induce DNA damage, cells were treated with γ-IR from a 137Cs source or 1mM hydroxyurea (HU). Both 40 μM

NU7026 and 2 μM NU7441 (Tocris, Bristol, UK) were used to specifically inhibit DNA-

PK as previously published414,421. DEK knockdown was accomplished using the adenoviral AdDEKsh as compared to control AdGFP vectors at 10 infectious units (IU) per cell for 48 hr prior to IR treatment, as described previously309. The pMIEG His-

FLAG-DEK retroviral vector was described recently422, and cells were transduced with virus for 24 hrs prior to sorting for GFP positive cells on a BD-FACSAria II flow cytometer.

Serum Immunoglobulin ELISA Assays

Dek+/+ and Dek-/- littermate mice were born from heterozygous crosses on a mixed

C57Bl6/S129 background as previously described384. Usage and handling of mice was performed with the approval of the Cincinnati Children’s Institutional Animal Care and

Use Committee and complied with institutional, state, and federal guidelines and regulations as well as AAALAC accreditation standards. All mice were housed in

87 specific pathogen free housing with ad libitum access to food and water. Blood was obtained by cardiac puncture of three DEK+/+ and three DEK-/- mice. To measure immunoglobulin (Ig) in the blood serum by ELISA, plates were coated with the following Southern Biotechnology Associates antibodies (Birmingham, AL, USA): anti- mouse IgM (no. 406501), IgA (no.556969), or IgG (no. 1030-01), and Ig was detected with horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG1 (no. 1070-05),

IgG3 (no. 1100-05), IgG2a (no.1080-05), IgG2b (no. 1090-05), IgA (no. 1040-05), or

IgM (no. 1020-05). In all cases, wells were developed with the Ultra TMB peroxidase substrate system (Thermo Scientific) and OD was measured at 450nm using a Fluostar Omega microplate reader (BMG-Labtech, Ortenberg, Germany).

Cleaved caspase 3 flow cytometry

Cells were trypsinized, fixed in 4% PFA for 10 min at 37ºC, permeabilized in 90% ice- cold methanol for 30 min, and incubated with the cleaved caspase 3 cell signaling antibody for 1 hr. Analysis was performed on a BD FACSCanto II instrument.

DNA repair reporter assays

MEFs were co-transfected with 2 μg of HR-EGFP/5’EGFP, pHPRT-DR-GFP, or an

EGFP expressing plasmid (transfection control) and 2 μg pCMV-I-SceI using Fugene

HD transfection reagent (Roche, Penzberg, Germany) prior to a 48 hr incubation416.

Cells were collected and analyzed for EGFP expression by flow cytometry. Analysis of chromosomal HR repair was conducted using mouse ES cells bearing a single copy knock-in of the 6xTer-HR reporter cassette targeted to the Rosa26 locus417. For the

6xTer-HR reporter cells, 1.6 x 105 cells were cotransfected in suspension with 0.35 μg empty vector, pcDNA3-myc NLS-Tus, or pcDNA3-myc NLS-I-SceI, and 20 pmol

ONTargetPlus-smartpool, (Dharmacon, Lafayette, CO), siLuc (p-002099-01-20), or siDek (M-050260-00-0005) as described previously417. Transfected cells were analyzed by flow cytometry 72 hrs post transfection using a Becton Dickinson LSRII. GFP+ frequencies were measured in duplicate samples, 3-6 x 105 total events were scored.

Repair frequencies presented are corrected for background events and for transfection efficiency (60-85%). Transfection efficiency was measured by parallel transfection with

0.05 μg wild type GFP expression vector, 0.30 μg control vector and 20 pmol siRNA.

Western blot analysis

Cell lysates were generated by lysis in NETN buffer (50 mM Tris-HCl pH=7.4, 250 mM

NaCl, 5 mM EDTA, 0.1% NP-40, 10 mM NaF, 200 μM Na2VO3, 0.5 mM PMSF, and 1x protease inhibitor cocktail) on ice for 30min. Non-soluble debris was precipitated by a

5min 12000rpm spin at 4ºC and discarded. Lysate was treated with 4x loading buffer

(0.5 M Tris-HCl pH 6.8, 277 mM SDS, 40% glycerol, bromophenol blue, 4% 2- mercaptoethanol), ran on a 10% SDS-PAGE gel, and transferred onto a PVDF membrane for 1-2 hr at 500 mA. The following Santa Cruz (Dallas, TX, USA) antibodies were used: BRCA1 (sc-6954) and RAD51 (sc-8349 and in-house aliquot B32). The following Cell Signaling (Danvers, MA, USA) antibodies were used: ATM (2873),

GAPDH (5174), Chk2 (2662), pChk2 T68 (2661), Chk1 (2345), and pChk1 S345 (2341).

Other antibodies included DEK (610948 BD Bioscience, San Jose, CA, USA), DEK

(16448-1-AP, Protein Tech, Rosemont, IL, USA), DEK (in-house K-877)293, DNA-PKcs

(ab1832, Abcam, Cambridge, MA, USA), pDNA-PKcs S2056 (ab18192), pATM S1981

(AF1655, R&D Systems, Minneapolis, MN), γH2AX (05-636, Millipore, Darmstadt,

Germany), MRE11 (GTX70212, genetex, San Antonio, TX, USA), and NBS1 (NB100-

143SS, Novus Biologicals, Littleton, CO, USA).

Immunofluorescence microscopy

1-5x105 cells were plated onto poly-d-lysine treated coverslips and allowed to attach prior to adenovirus infection and/or drug treatment. To stain, coverslips were fixed in 4% paraformaldehyde, antigens retrieved with a 0.2% Triton x-100 PBS solution, and cells blocked with PBS plus 5% normal goat serum and 0.3% Triton X-100. Antibody dilutions are as follows: BRCA1 (sc-6954), γH2AX 1:1000 (05-3636, Millipore), RPA2 1:300

(ab2175), RAD51 1:500 (sc-8349), goat anti-mouse Alexa Fluor 568 1:500 (A-11004,

ThermoFischer Scientific, Waltham, MA), and goat anti-rabbit Alexa Fluor 633 1:500 (A-

21070). Nuclei were counterstained with DAPI ProLong Gold (P-36931, Life

Technologies). RPA and RAD51 foci were captured on a Carl Zeiss Apotome instrument, and foci within 100 cells counted per time point per replicate. BRCA1,

RAD51, and RPA foci counts were performed by eye due to background signal from the antibodies. High resolution images were imaged on an inverted Nikon A1R GaAsP confocal microscope. Whole cell γH2AX foci were also collected on the confocal microscope using z-stacks and were quantified using Imaris 6 (Bitplane AG, Zurich,

Switzerland).

Chromatin fractionation

Chromatin fractionation was performed as described previously423 according to the schematic in Fig. 3D. Briefly, HeLa cells were infected with AdGFP or AdDEKsh 48 hr prior to receiving 10Gy of IR. Six hours following IR, cells were counted to ensure equal numbers of cells per sample. Cell lysis and nuclear isolation was performed in a wash buffer (10 mM PIPES pH=7.0, 1 mM EGTA, 0.1 M NaCl, 0.3 M sucrose, 0.5 M NaF, 0.5 mM Na3VO4, and 1x protease inhibitor cocktail) plus 1% Triton X-100. Serial separation of the precipitated nuclear fraction was performed first by treatment with wash buffer +

20 U DNAseI (AM2235, ThermoFischer Scientific), followed by a 5 min incubation with wash buffer + 0.5 M (NH4)SO4 on the resulting pellet. The resulting fractions were treated with 4x loading buffer.

Immunoprecipitation (IP)

Subconfluent cells were treated with or without 1mM HU for 15hrs. Cells were homogenized in lysis buffer (10 mM HEPES, pH=7.9, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM PMSF, 0.5 mM DTT, 1 mM NaVO3, 10 mM NaF, and 1x Sigma P5955 protease inhibitor cocktail), and the nuclear pellet was lysed in a nuclear lysis buffer (20 mM

HEPES pH=7.9, 400 mM NaCl, 1% Triton X-100, 0.1% NP-40, 10% Glycerol, 0.5 mM

PMSF, 0.5 mM DTT, 1 mM NaVO3, 10 mM NaF, and 1x protease inhibitor cocktail) on ice. After ultracentrifugation, pMIEG and pMIEG His-FLAG-DEK supernatants were loaded onto anti-FLAG M2 affinity gel (A2200, Sigma, St. Louis, MO, USA). Parental cell supernatants were pretreated with antibodies for 1 hour prior to loading onto Protein

A sepharose beads. For both IP experiments, lysates were treated with either 50 μg/ml ethidium bromide (EtBr) or 125 U of benzonase (E8263, Sigma) plus 2 mM MgCl2 as

91 controls. After incubating overnight at 4°C, the beads were washed, and the IP products collected by boiling in 2x Laemmli Buffer (12.5 mM Tris-HCl, pH=6.8, 20% Glycerol, 4%

SDS, 0.004% Bromophenol Blue, 10% 2-β-mercaptoethanol).

Statistical Methods

Statistics were performed using Graphpad Prism 6 software (Graphpad Software Inc.,

San Diego, CA). Significance is indicated by asterisks (*p<0.05, **p<0.01, ***p<0.001,

****p<0.0001), and the number of independent biological replicates and type of analysis is indicated in the Fig. legends.

4Figure 4.1: DEK loss causes apoptosis in conjunction with DNA-PK inhibitors. (a) Mouse serum immunoglobulin levels were quantified by ELISA (n=3). (b)

Representative live-cell images of mouse embryonic fibroblasts (MEFs) 72 hrs post treatment with a DNA-PK inhibitor (10x magnification). (c) DEK-/- MEF cells display increased apoptosis following 48 hr treatment with NU7026. (paired t-test, n=6, mean ±

SEM) (d) Hela cells were pre-treated with adenovirus for 24 hr prior to treatment with two different DNA-PK inhibitors. Representative images collected after incubating with drug for 48 hrs. (e) Both drugs demonstrated a significant increase in apoptosis at this time point. (paired t-test, n=3, mean ± SEM)

5Figure 4.2: DEK is required for HR DSB repair. (a) Schematic of two episomal HR reporter plasmids. After co-transfection with an I-

SceI plasmid, the mutant GFP gene (light green) site indicated by double arrows is cleaved. The circle represents the corresponding region of homology to the I-SceI site, and the X represents gene truncations. Expression of functional GFP (dark green) is dependent on HR repair of the cleaved plasmid. (b) GFP+ FACs quantification of repair frequency of the constructs in (A). (Mann-Whitney test, n=12 for HR-EGFP/5’EGFP and n=5 for pHPRT-DRGFP, mean ± SEM) (c) Schematic of the 6xTer vector. Tus-Ter binding induces replication fork stalling at the 6xTer array (red hourglass) while I-SceI

95 creates a DSB at the location indicated by the double arrows. HR repair triggered by I-

SceI or Tus expression results in short tract gene conversion (STGC, GFP+ RFP-), or long-tract gene conversion (LTGC) through the duplication and proper alignment of two synthetic RFP exons allowing RFP expression by alternative mRNA splicing (GFP+

RFP+). (d) Western blot analysis of 6xTer mES cells after siDek treatment. (e)

Quantification of total HR repair (STGC + LTGC) and individual STGC and LTGC repair FACs following I-SceI transfection. (Unpaired Welch's t test, n=6, mean ± SEM)

(f) STGC and LTGC HR repair following Tus transfection. (Unpaired Welch's t test, n=6, mean ± SEM)

6Figure 4.3: DEK is necessary for the activation of γH2AX after ionizing radiation. (a) MEFs were treated with 10 Gy of gamma ionizing radiation (IR) and stained for

γH2AX. Representative cross sections from z-stack confocal images are shown. (b)

Quantification of cells with >10 γH2AX foci in (A). At least 50 cells per time point were counted in each biological replicate, and all foci within the z-stack were quantified.

DEK-/- MEFs were unable to activate γH2AX above baseline after irradiation. (paired t- test, n=3, mean ± SEM). (c) Quantification of the average total number of γH2AX foci in each cell from (A). (paired t-test, n=3, mean ± SEM) The number of foci per cell remained unchanged in DEK-/- MEFs after irradiation. (d) Western blot analysis of

γH2AX in MEFs at 6 hs following irradiation with 10 Gy. (e) Western blot analysis of

C33A human cancer cells infected with adenovirus 48 hr prior to receiving 10 Gy IR

(see also Supplementary Fig. S1).

7Figure 4.4: Loading of RAD51 onto RPA-protected DNA is significantly reduced.

(a) HeLa cells were infected with 10 IU of adenovirus 48hr prior to receiving 10 Gy and collection at 3, 6, and 12hr post irradiation. Cells were subjected to immunofluorescence for BRCA1 (red) and RAD51 (far red, pseudo-colored green). (b)

Quantification of the co-localized foci from (A). Cells with ≥3 overlapping BRCA1 and

RAD51 foci were counted as positive, and >100 cells were counted in three biological replicate experiments. (student’s t-test, n=3, mean ± SEM, see also Supplementary

Fig. S3). (c) HeLa cells were treated, collected, stained for RPA2 and RAD51 and quantified as in (A) and (B). (student’s t-test, n=4, mean ± SEM, see also

Supplementary Fig. S4). (d) To examine time points beyond 12 hrs, MEF cells were treated with 10 Gy, harvested at 3, 6, 12, 18, and 24 hr post treatment, and subjected to immunofluorescence for RPA2 (red) and RAD51 (far red, pseudo-colored green). (e)

Quantification of the co-localized foci from (D) performed as in (C).

100

8Figure 4.5: DEK interacts with RAD51 and is essential for multiple functions in

HR DSB repair.

(a) A FLAG IP performed in HeLa cells expressing His-FLAG tagged DEK, both in untreated and samples treated with 1mM HU for 17 hrs. This identified RAD51 as a

DNA-independent interacting factor. (b) RAD51 also co-immunoprecipitated with DEK in untransduced parental HeLa cells. (c) Reverse pulldown of RAD51 in parental HeLa cells confirmed DEK as an interacting partner. (d) DEK is required for successful HR of

DSBs and functions at multiple levels in the repair cascade.

101

9Supplementary Figure 4.1 DEK knockdown HeLa cells have increased active pATM, but no increase in γH2AX compared to control cells. (a) Nuclear fraction was performed on HeLa cells treated with adenovirus and 10 Gy IR (same experiment as in Supplementary Fig. S2). (b)

ATM activation in (a) resulted in CHK2 phosphorylation in AdDEKsh treated HeLa cells after IR treatment. HeLa cells were treated with adenovirus 48 hr prior to treatment with

10 Gy IR. Whole cell extracts were collected in (b) and used for western blot analysis.

102

10Supplementary Figure 4.2 DEK knockdown does not affect chromatin recruitment of the DSB end-processing

complex. (a) Chromatin fractionation was performed on HeLa cells by serial

centrifugation as outlined in the schematic. Whole cells were lysed with a 0.1% Triton

X-100 solution, and the insoluble nuclear extract was further fractionated by stepwise

DNAse I treatment and high salt conditions. (5% (NH4)2SO4 solution) (b) Chromatin

fractionation was performed as outlined in (A) following a 48 hr incubation with

adenovirus and 6 hr incubation post irradiation (n=2). The nuclear extract in

Supplementary Fig. S1a is from the same experiment.

103

11Supplementary Figure 4.3 DEK loss does not affect the total number of cells positive for BRCA1 foci. HeLa cells

were infected with 10 IU of AdGFP or AdDEKsh 48 hr prior to receiving 10 Gy of

ionizing radiation (IR). Cells were labelled for BRCA1 and Rad51, and cells with (a) ≥6

BRCA1 foci or (b) ≥6 RAD51 foci were counted as positive. (>100 cells per sample,

unpaired t-test, n=3, mean ± SEM).

104

12Supplementary Figure 4.4 DEK loss does not affect the total number of cells positive for RAD51 foci, but RPA

foci positive cells remain elevated in Dek KO cells at 18 and 24 hrs post IR. (a-b)

Quantification of RAD51 (green) and RPA (red) from the experiments in Figure 4B.

Cells with ≥6 individual RPA (B) or RAD51 foci (C) were counted as positive, and 100

cells were counted in four biological replicate experiments (unpaired t-test, n=4, mean

± SEM).

105

13Supplementary Figure 4.5 DEK interacts with RAD51 in untreated cells and in cells that received 10Gy IR.

Pulldown was performed MIEG-His-FLAG-DEK transduced HeLa cells as in Fig. 5a.

106

Chapter 5: DEK associates with tumor stage and outcome in HPV16

positive oropharyngeal squamous cell carcinoma

Eric A. Smith1,‡, Bhavna Kumar2,3,‡, Kakajan Komurov1, Stephen M. Smith4, Nicole V.

Brown5, Songzhu Zhao5, Pawan Kumar2,3, Theodoros N. Teknos2,3,*, Susanne I. Wells1,*

1Cancer and Blood Diseases institute; Cincinnati Children’s Hospital Medical Center;

Cincinnati, OH, 45229; USA.

2Department of Otolaryngology – Head and Neck Surgery; The Ohio State University;

Columbus, OH, 43210, USA.

3The Ohio State University, Comprehensive Cancer Center, Columbus, OH 43210,

USA.

4Department of Pathology; The Ohio State University; Columbus, OH, 43210, USA.

5Center for Biostatistics; The Ohio State University; Columbus, OH, 43210, USA.

‡Co-first author

107

ABSTRACT

Oropharyngeal squamous cell carcinomas (OPSCC) are common, have poor outcomes, and comprise two biologically and clinically distinct diseases. While OPSCC that arise from human papillomavirus infections (HPV+) have better overall survival than their

HPV- counterparts, the incidence of HPV+ OPSCC is increasing dramatically, affecting younger individuals which are often left with life-long co-morbidities from aggressive treatment. To identify patients which do poorly versus those who might benefit from milder regimens, risk-stratifying biomarkers are now needed within this population. One potential marker is the DEK oncoprotein, whose transcriptional upregulation in most malignancies is associated with chemotherapy resistance, advanced tumor stage, and worse outcomes. Herein, a retrospective case study was performed on DEK protein expression in therapy-naïve surgical resections from 194 OPSCC patients. We found that DEK was associated with advanced tumor stage, increased hazard of death, and interleukin IL6 expression in HPV16+ disease. Surprisingly, DEK levels in HPV16-

OPSCC were not associated with advanced tumor stage or increased hazard of death.

Overall, these findings mark HPV16- OPSCC as an exceptional malignancy were DEK expression does not correlate with outcome, and support the potential prognostic utility of DEK to identify aggressive HPV16+ disease.

108

INTRODUCTION

Each year over 500,000 new cases of head and neck cancers (HNCs) are reported worldwide2, 40,000 of which occur in the United States424. There are two biologically distinct subtypes of these malignancies: human papillomavirus positive

(HPV+) and negative (HPV-)22. HPV- disease predominates in older HNC patients with a history of long-term alcohol and tobacco use and is declining in prevalence alongside a decrease in smoking habits7,425. This translates to a decrease in HNC at anatomical sites that predominately harbor HPV- tumors. However, the incidence for HPV+ HNC, particularly in oropharyngeal squamous cell carcinomas (OPSCC), has increased by

58% in the past two decades3-6. Even with the generous assumption that all eligible people receive immunization against HPV, this upward trend is still expected to continue until the vaccinated generation comes of age in 30-40 years6. Meanwhile, developing new innovations to characterize and treat OPSCC, the form of HNC most likely to harbor HPV, will be a necessity.

In general, patients with HPV+ malignancies have a decreased risk of disease progression, respond better to therapy, and have overall better survival compared to

HPV- malignances67,426,427. However, HPV+ disease occurs in younger patients and the morbidities associated with aggressive surgical and chemoradiation therapy can reduce their quality of life dramatically428,429. Because of their more favorable prognosis, there are efforts to de-escalate treatment regimens to avoid these morbidities, but it is unclear which patients will respond optimally427. Additionally, a sub-population of patients with advanced T4 or N3 stage HPV+ OPSCCs do poorly with standard treatment and have a

5-year survival rate of 54%20. In order to identify high-risk HPV+ OPSCCs, further

109 stratification of their clinical and biological characteristics is needed, together with the identification of new biomarkers indicative of more aggressive disease.

HPV is the most prevalent sexually transmitted virus, well-known to induce cervical cancer430, and the high-risk HPV16 serotype is the cause of almost 90% of

HPV+ OPSCC17,18. While just over 50% of all OPSCC are HPV+5, the mechanism for how HPV+ tumors respond more favorably to treatment67,426 is still incompletely understood431. However, inhibition of p53 432 and retinoblastoma (RB) pocket proteins 94 by the HPV E6 and E7 oncoproteins, respectively, are likely at play as the p53 and RB tumor suppressors often remain intact during malignant transformation91. Because of the sustained expression of E6 and E7, HPV+ OPSCC require fewer somatic mutations to develop tumors256,257,433, may not have developed the complement of mutations necessary for chemoradiation therapy resistance, and can still respond to treatment by activating p53, RB family members, and other interacting tumor suppressors 73. In both

HPV+ and HPV- OPSCC, early stage non-invasive tumors can be resected with good outcomes (>80% survival)434. Unfortunately, the majority of cases are diagnosed at advanced, locally invasive stages where chemoradiotherapy response and overall survival decrease435. Generally, patients with HPV+ disease respond better than their

HPV- counterparts even at these advanced stages; however, HPV+ tumors with large primary (T4) and/or large tumor extensions into lymph node (N3) still carry a relatively poor survival prognosis compared to early stage disease24.

One method of determining the HPV status of OPSCC is by immunohistochemical staining for the presence of p16 (CDKN2A). In OPSCC there is a high degree of correlation between p16 and HPV infection436, but there are false

110 positives which can be ruled out with high-sensitivity in-situ hybridization (ISH) or polymerase chain reaction for viral nucleic acids436-439. The p16 gene product is a well- known tumor suppressor capable of inducing senescence in both OPSCC and primary keratinocytes, the normal epidermal cell type from which OPSCC tumors originate440-442.

In these systems, p16 drives senescence by stabilizing and preventing the inactivation of the RB pocket proteins RB1, p107, and p130. Should these factors be inactivated, such as by E7 expression during HPV infection, then p16 expression can rise dramatically in the absence of senescence induction and even acquires oncogenic functions443-445. In HPV- OPSCC disease, p16 is commonly mutated, deleted, or silenced through promoter methylation441,446. In general, p16 is considered a good surrogate marker for HPV infection. However, there is a subset of HPV- OPSCC that are p16 positive, whose clinical and biological characteristics are not well studied436,438,439.

Inactivation of the retinoblastoma proteins by HPV also drives the expression of other oncogenic factors through E2F-mediated transcription91,92. One important oncogene in many tumors upregulated in this manner is DEK100. This oncogene was originally described as a DEK-CAN (NUP214) fusion protein in t(6:9) acute myeloid leukemia 447 and is a highly conserved DNA binding protein in vertebrates with no known paralogs. In normal cells, this protein has functions in DNA replication282, mRNA splicing311, chromatin remodeling282,290,300, and DNA repair212,297. In normal keratinocytes, DEK overexpression has been shown to promote hyperplasia and proliferation318,386, inhibit differentiation386, induce mitotic defects and chromosome abnormalities319, block apoptosis384, and drive transformation in cells expressing HPV

111

E6 and E7384. While the molecular mechanisms whereby this oncogene promotes these phenotypes remain surprisingly unclear, it is thought that DEK functions through chromatin binding/modification280. Due to the importance of DEK for these varied oncogenic phenotypes, the near-ubiquity of high DEK expression in most cancers448, the similarity of phenotypes observed across tumor types448, and the ability of the protein to be secreted by cells298, DEK is currently being evaluated as a biomarker for bladder carcinoma and other malignancies369.

Compared to normal tissue, DEK expression is upregulated in most surveyed tumor types, including breast278,370, hepatocellular carcinoma392, colorectal cancer372, and, recently, OPSCC318. Our previous study found that DEK protein was highly expressed in all of a small subset of 21 HPV+ and HPV- OPSCC samples that were analyzed318. To validate these initial results and establish a more refined relationship between DEK expression and HPV status, we surveyed a large population of OPSCC patients using an established set of primary OPSCC tissue microarrays (TMAs). We first examined the association of DEK with HPV based on p16 (CDKN2A) status and HPV16 genome in-situ hybridization (ISH). Following this analysis, the association of DEK with

IL6 expression was also tested. IL6 is a pro-inflammatory interleukin that is strongly associated with poor overall OPSCC patient survival 449 and increased risk of metastasis 450. Even though IL6 was found to be transcriptionally downregulated following loss of DEK in HNC tissue culture models305, a link between DEK and IL6 expression in patient OPSCC tumors has yet to be elucidated.

In this study, we found that elevated DEK expression associates with IL6 expression, higher stage tumors, and worse prognosis in HPV16+ OPSCC. These

112 findings support further work into developing DEK as a biomarker for HPV16+ disease and are in agreement with the DEK biomarker literature. Surprisingly, however, our data do not support similar conclusions for HPV16 negative OPSCC as there was no association between DEK expression and survival or tumor stage. This potentially marks HPV16- OPSCC as one of few solid tumors where DEK is not useful as a biomarker, and may indicate distinct biological activities for this protein in the development and progression of HPV16+ versus HPV16- disease.

113

RESULTS

For this retrospective case study, 194 patients were enrolled, and the cohort represented the expected demographics of OPSCC disease (Table 5.1). Approximately

57% of this cohort had HPV16+ disease, as determined by genomic HPV in-situ hybridization (ISH), which was expected from a previous epidemiological study 5. In line with the prognostic focus of this work, the collected tumor specimens were treatment naïve and obtained from initial surgery with curative intent. Overall the cohort experienced a 56.0% survival rate with 34.9% and 71.8% in the HPV16- and HPV16+ groups, respectively (Supplementary Table 5.1).

DEK is most highly expressed in HPV16+/p16+ OPSCC tumors.

DEK protein expression levels and HPV16/p16 status were determined using tissue microarrays (TMAs). For each patient sample, three non-adjacent tumor tissue cores and one normal tissue core were used, and DEK staining was consistent across the three tumors cores in each case. Figure 5.1A-D shows representative DEK staining patterns and quantification of stain intensity, proportion of tumor cells stained, and the calculated quick score for each image. All quantified samples were grouped based on

HPV16 and p16 status; representative positive p16 stains are shown in Figure 5.1E-F.

There were significant differences in DEK staining across the HPV16+/p16+ (n=109),

HPV16-/p16+ (n=36), and HPV16-/p16- (n=46) OPSCC subtypes (HPV16+/p16- not shown because n=1). HPV16+/p16+ OPSCC had significantly higher average DEK stain intensity (Figure 5.1G), stain proportion (Figure 5.1H), and quick score (Figure 5.1I) compared to double-negative HPV16-/p16- subjects (p<0.01 in all cases after

114

Bonferroni correction), while the HPV16-/p16+ group only had higher stain proportion

(p=0.01) and quick score (p=0.03) relative to HPV16-/p16- subjects. HPV16+/p16+ and

HPV16-/p16+ groups were not statistically significantly different for any of the three DEK measures after Bonferroni correction.

DEK expression in tumors predicts advanced disease and poorer survival in HPV16+, but not in HPV16- OPSCC.

While our previous report indicated consistently high DEK staining in 21 OPSCC tumors, 17 of these were AJCC stage IV disease318. This current work sought to use a larger sample size, with well-represented stage I-IV disease to further assess the utility of DEK as a biomarker in OPSCC (Table 5.1). In this cohort, a wider range of DEK staining was noted compared to our previous study318, indicating that DEK expression was more dynamic than initially predicted (Figure 5.1, Supplementary Table 5.1), and some of the increased variability may be due to the increased representation of early stage OPSCCs (Table 5.1). Tumors bearing high DEK staining and a larger percentage of DEK positive cells, as indicated by the quick score (p=0.039), had an increase in the hazard of death for patients with HPV16+/p16+ malignancies (p = 0.039, Table 5.2). To better clinically understand the association, Kaplan-Meier curves based on categorizing patients into two groups using the 75th percentile of the DEK quick score (DEK Q > 200 vs. DEK Q ≤ 200) are given (Figure 5.1J, hazard ratio (HR) =2.1, 95% CI = (1.02, 4.3), p = 0.039). When the malignancies where separated based on HPV16 status alone, patients with HPV16+ tumors were much more likely to bear higher DEK quick scores than their negative counterparts (Figure 5.2A, Supplementary Table 5.2, p<0.005) and

115 maintained the increased hazard of death (Supplementary Table 5.3, p=0.03). These results suggest DEK is a marker of worse prognosis within the subset of HPV16+

OPSCC. This hypothesis is supported by the finding that a high burden of DEK positive cells in HPV16+/p16+ specimens correlates with advanced tumor stage (Tables 5.3,

Supplementary 5.4, p=0.02). In stark contrast, HPV16- OPSCCs did not show a similar correlation between high DEK expression and tumor stage (Table 5.3) or hazard of death (Tables 5.2, Supplementary 5.3). DEK instead predicted a lesser degree of perineural invasion, the scientific basis of which is unknown. Importantly, the survival data mark HPV16- OPSCC, both p16+ and p16-, as a minority of solid tumors where high DEK expression does not correlate with disease stage or outcome.

DEK expression is associated with p16+ status in both HPV16+ and HPV16- disease.

While DEK expression did not correlate with survival in either p16+ or p16-

OPSCC patients (Supplementary Table 5.1, Supplementary Table 5.5), DEK expression by quick score was significantly increased in p16+ tumors (p<0.001, Figure

5.2B, Supplementary Table 5.2). Considering that HPV E7 disrupts and inactivates the entire RB pocket protein family, a correlation between the expression of DEK and p16 was expected as both are upregulated upon pocket protein loss of function443,444. This correlation between p16 and DEK quick score was conserved in HPV16-/p16+ OPSCCs

(Figure 5.2C). Since the percentage of HPV16-/p16+ tumors in this study (24.8% of p16+ tumors) was similar to what has been reported previously (25.7%)438,439, we sought to validate the relationship between p16 and DEK using published HPV- HNC

Cancer Genome Atlas data433. HPV status was rigorously validated through a

116 combination of whole genome, whole exome, and RNA sequencing for viral sequences, as well as ISH for HPV16, 18, 33, 35, 39, 45, 51, 52, 56, 58, and 66 serotypes. DEK and HPV status correlated with p16 (CDKN2A) expression (Figure 5.2D-E), and the majority of p16+ tumors were also HPV16+, as expected (Figure 5.2E). However, a sizable subgroup of HPV16- tumors was identified that expressed moderate to high levels of p16 (Figure 5.2E-F), with the top 4% of DEK-expressing HPV- HNCs harboring significant p16 mRNA over-expression (Figure 5.2F). To provide a potential mechanism for how some HPV- OPSCCs may induce p16 overexpression, we re- analyzed the TCGA data, shown in Figure 2E-F, and found that chromosome region

6p22.3 was amplified in tumors where DEK was most highly expressed (Figure 5.2G).

This amplified region encodes both the DEK and E2F3 genes; the latter has been implemented in driving expression of p16451.

DEK correlates with IL6 expression in HPV16+ OPSCC.

In addition to promoting tumor growth and invasion, DEK has potential pro- inflammatory properties. These include autoantigen properties when secreted or present in body fluids298,391 and control of robust expression of inflammatory pathway members such as IRAK1305. In the latter report, RNA-sequencing of HPV+ and HPV-

HNC cells identified shared decreased IL6 expression in response to DEK knockdown.

To determine whether DEK expression correlated with IL6 status in OPSCC, we quantified IL6 expression in the TMAs (Figure 5.3A). HPV16+ and HPV+/p16+ OPSCC tumors showed a correlation between high DEK stain intensity and positive IL6 status

(Figure 5.3B-C, p<0.04 and p=0.05 respectively). While this relationship was inverse in

117 total HPV16- OPSCC (Supplementary Figure 5.1), this study lacked the power to confirm this association in either the HPV-/p16+ or HPV-/p16- subgroup as no significance was observed in DEK stain intensity, stain proportion, or quick score

(Figure 5.3C, data not shown). A role for HPV16 in independently regulating both IL6 and DEK expression in HPV16+ disease cannot be ruled out by this study, and the relationship between DEK and IL6 in HPV16- OPSCC is subtle.

118

DISCUSSION

Most clinical studies of DEK across human tumors have correlated high expression of the oncogene with advanced tumor stages and poor outcomes

279,370,372,375,379. We found that this relationship holds true in HPV16+ OPSCC (Figure

5.1J, Table 5.2 and Supplementary Table 5.3). Considering that HPV16 accounts for approximately 90% of HPV+ OPSCCs17,18, these data indicate that DEK has potential as a prognostic biomarker for the vast majority of these tumors. However, this study was a medium-sized retrospective cohort designed to determine the potential for DEK as a biomarker in different OPSCC groups. To rigorously ascertain the clinical significance and prognostic value of DEK in HPV16+ disease, a large multi-institutional prospective study is now warranted.

Importantly, our data do not support DEK as a biomarker in OPSCC identified as

HPV16-, HPV16-/p16+, or HPV16-/p16- (Table 5.2 and Supplementary Table 5.3).

This is a surprising finding given the extensive literature on DEK expression in other

HPV- solid tumor types, and suggests that HPV16- OPSCC is one of few wherein DEK does not have prognostic utility. Our data thus strongly suggests that DEK is not prognostic in all subgroups of a given malignancy, and must therefore be carefully examined as an appropriate biomarker in each case. Interestingly, DEK was significantly associated with a decreased risk of perineural invasion in the double negative group. The scientific basis of this latter finding is unknown, and should warrant further investigation as high DEK expression is usually associated with invasion and migration phenotypes316. With regard to the HPV16-/p16+ group, the combination of limited sample number (n=36) and heterogeneity of HPV+ and HPV- tumors likely

119 precluded any significant clinical findings. To address these issues and more deeply study HPV16-/p16+ OPSCC, a significantly larger patient population would be required.

DEK mutations cannot explain the observed disparities between HPV16+ versus

HPV16- disease. Such mutations are rare in all tumors, and found in only 0.4% of

OPSCC. Specifically, we analyzed the TCGA provisional head and neck cancer study, and only 2 out of 530 tumors harbored mutations within the DEK gene. The first was an

E20V substitution, and the second was a nonsense frameshift at amino acid 11. No

DEK mutations were identified in the 279 samples previously published by TCGA433.

Alongside survival analysis, this report determined that the average DEK expression was highest in HPV16+ OPSCC by TMA (Figure 5.1G-I) and in published

TCGA data433 (Figure 5.2D). This is likely due to early increases in DEK expression.

DEK is upregulated by E2F transcriptional activator family members, which in turn are activated by HPV E7 disruption of RB family members92,100. The additional increase in

DEK protein levels which correlated with advanced stage HPV16+ tumors (Table 5.3) is most likely a result of HPV E7 gene amplification and increases in E7 transcriptional activation and mRNA stability. These are all common occurrences in HPV+ malignancies165,452. Apart from HPV16+ tumors, the next highest DEK-expressing group was the HPV16-/p16+ cohort. This group is partially comprised of tumors bearing other high-risk HPV serotypes which would over-express DEK in the same manner as

HPV16+ OPSCCs. True HPV-/p16+ tumors likely comprise the remainder of this group, and we have found that a significant fraction of HPV- tumors express p16 in the TCGA cohort (Figure 5.3E-F). Other reports have also described this HPV-/p16+ population436,438,439, but their clinical and biological characteristics are the least studied

120 of OPSCC subtypes. We found that p16+ status was significantly correlated with the highest DEK-expressing HPV- tumors (Figure 5.2F). Biologically, this may be a consequence of chromosome 6p22.3 amplification, a common occurrence in multiple tumor types, but one that has not been reported in HNC until now453. This region contains both the DEK and E2F3 genes, and was amplified in top DEK-expressing

HPV- malignancies (Figure 5.2G). Currently, it is thought that one mechanism of p16 upregulation may occur through activator E2F genes, including E2F3, although the exact mechanism for this relationship is unknown451. It is thus possible that the amplification of E2F3 drives p16 expression.

As the field moves towards incorporating personalized medicine for HNC care, identifying and characterizing distinct subtypes of HPV+ and HPV- OPSCC will become paramount. This study contributes to this endeavor by identifying DEK as a potential prognostic biomarker for further study in HPV16+ OPSCC, and does not support its use for HPV16- disease. Prior to this current study, we and others have argued for the development of therapeutic DEK inhibiting molecules to take advantage of broad solid tumor dependency on high DEK expression for survival279,309,314,322,384. While DEK targeting strategies will undoubtedly be useful in the treatment of most solid tumor types, our study also suggests for the first time that use of these agents will need to be carefully tailored in the care of OPSCC patients.

121

MATERIALS AND METHODS

Study population

This retrospective case study was approved by the Ohio State University Institutional

Review Board and a waiver of HIPAA authorization was obtained. All patient OPSCC specimens were requested from samples obtained by The Ohio State University James

Cancer Hospital and Solove Research Institute from 2002 to 2009, and were treatment naïve at the time of collection. During the enrollment period, all patients were given the option of primary (C)RT or primary surgery with follow-up (C)RT as necessary. The majority of patients opted for initial surgery, and this is the population that was included in this study. All samples were chemo- and radiotherapy naïve, and post-surgical standard of care treatment was followed as necessary. There was no bias in selecting patient samples based on size or stage. The following patient attributes were accessed: age, race, gender, marital status, smoking status, tumor size, nodal status, presence of local metastasis, presence of perineural invasion, survival time and tumor recurrence. In this report, survival time was defined as the time from the patient’s primary surgical resection of OPSCC to death. The date of the last living observation was censored.

Recurrence is defined as any occurrence of a new suspicious head and neck mass that is confirmed by radiology or pathology as squamous cell carcinoma within five years of surgical resection.

Tissue microarray (TMA), immunohistochemistry (IHC), and HPV in-situ hybridization

The Ohio State University Department of Pathology Histology Core Laboratory generated master TMA blocks from selected archived paraffin-embedded tissue. Briefly,

122 the distribution of tumor and normal tissue was determined via hematoxlin and eosin staining by a pathologist, and the TMA master blocks were created from 0.6-mm punch cores of 3 representative tumor tissues and 1 normal tissue for comparison in each specimen. TMA slides were stained for IL6 and p16, by IHC, and for HPV16, by in-situ hybridization (GenPoint, Dako), following previously published methods 449. IHC staining for p16 was performed using the CINtec mtm antibody (E6H4 clone). DEK IHC staining followed standard xylene deparaffination and rehydration in decreasing ethanol concentrations. Antigens were retrieved using a Biocare Medical LLC (Concord, CA,

USA) decloaking chamber with Dako antigen unmasking buffer for 20 minutes at 120ºC.

After returning to room temperature, slides were incubated for 10 minutes at ambient conditions with Dako dual endogenous enzyme block. This was followed with blocking in PBS/serum solution corresponding to species of the secondary antibody. Following this, BD-Pharmingen mouse anti-DEK primary antibody was applied to the TMA and incubated for 1 hour at 37ºC. After washing, a room temperature 30 minute incubation was performed using biotinylated donkey anti-mouse secondary antibody (Vectastain

Elite Kit). After washing, slides were incubated with an avidin-biotin complex for 30 minutes (Vector Laboratories, Burlingame, CA, USA) prior to the addition of 3,3’- diaminobenzidine (Sigma). The reaction was quenched in water, counterstained with

Mayer’s hematoxylin, and coverslipped with Permount.

IHC scoring

A treatment-blinded pathologist interpreted the slides and scored for stain intensity (0: none, 1: low, 2: moderate, 3: high), and stain proportion (0-100%). A descriptive quick

123 score (0-300) was acquired by multiplying these two dimensions. To be considered positive for p16, strong and diffuse nuclear and cytoplasmic staining in ≥50% of the tumor cells was required.

Statistical Methods

Descriptive statistics were used to summarize the study population including means for the continuous variables and frequencies for the categorical variables. Cox proportional hazards models were used to assess univariate associations of DEK expression and the risk of death for the overall study population, and also stratified by HPV status, p16 status, and a combination of HPV and p16 status. Unadjusted hazard ratios and confidence intervals (CI) are reported. Mann-Whitney tests were used to assess associations between biomarkers/demographic/clinical characteristics and DEK expression (quantitative). Analyses were conducted in SAS, version 9.3 (SAS Institute,

Cary, North Carolina).

124

ACKNOWLEDGEMENTS

We would like to thank Drs. Lisa Privette-Vinnedge and Trisha Wise-Draper for their assistance in proof-reading this manuscript.

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors do not report any potential conflicts of interest.

GRANT SUPPORT

This work was supported by the National Institutes of Health [R01-CA116316 to S. Wells, T32

GM063483-13, T32-ES007250, CA178649 to P. Kumar] and The Ohio State University

Comprehensive Cancer Center.

125

14Figure 5.1: DEK is most highly expressed in HPV16+/p16+ OPSCC

126

Tumors were scored based on DEK stain intensity (0: none, 1: low, 2: moderate, 3: high), and the proportion of tumor cells stained for DEK (0-100%). Representative images of DEK staining and quantification are as follows: low DEK staining (A), moderate staining (B), high

DEK staining (C), and high DEK staining with complete tumor coverage (D) (100x magnification). Scores for intensity (I), proportion of tumor cells stained (P), and the quick score (Q, Q=IxP) are shown in the bottom left of each image. Representative images of positive stains for HPV16 ISH (E) and p16 IHC (F) are shown. After quantification, tumors were separated based on HPV16/p16 status and analyzed for differences in DEK stain intensity (G), proportion of cells stained (H), and quick score (I). DEK quick score dichotomized at >200 vs. ≤200 is a predictor of survival in HPV16+/p16+ OPSCC (hazard ratio

(HR) =2.1, 95% CI = (1.02, 4.3), p = 0.039). The number at risk for each group (i.e., the number remaining for each group at a given time point) is given at the bottom of the graph.

(J).

127

15Figure 5.2: DEK expression correlates with p16+ status in both HPV+ and HPV-

OPSCC.

Comparing all 194 oropharyngeal squamous cell carcinomas, the proportion and intensity of

DEK expression was significantly higher in HPV16+ (A) and p16+ (B) tumors. Quartiles and

128 statistical significance of DEK staining intensity and tissue proportion for (A) and (B) are depicted in Table S2. DEK expression correlated with p16 status in HPV16- tumors (C). D-G are an analysis of published Cancer Genome Atlas Network RNA-Seq data for HNCs [22].

DEK mRNA expression was significantly elevated in HPV+ disease (D). While p16 (CDKN2A) message was significantly elevated in HPV16+ disease as expected, there was substantial variability in p16 expression in HPV- tumors (E). The highest 4% of HPV- DEK expressing tumors (9 out of 242) were significantly increased for p16 expression (F). DEK was highly expressed in these tumors in correlation with amplification of DEK and the nearby E2F3 locus on 6p22.3 (G).

129

16Figure 5.3: High DEK expression was associated with IL6 expression in HPV16+

tumors

A representative IL6+ staining section is shown (A). HPV16+ tumors that were also IL6+

stained darker for DEK than tumors not expressing IL6 (B). There was no association

between DEK and IL6 status in HPV16-/p16+ or HPV16-/p16- tumors (C).

130

Patient Characteristics Mean SD

Age (years) 57.6 9.8 n % Gender Male 158 81.4 Female 36 18.6 Race African American/Black 9 4.6 White 185 95.4 Marital Status Single/Divorced/Widowed 80 45.2 Married 97 54.8 HPV16 Status Negative 83 43.0 Positive 110 57.0 Smoking Status: Pack Years 10 pack years or less 47 25.3 More than 10 pack years 139 74.7 Node Stage N0/N1 70 36.1 N2/N3 124 63.9 Tumor Stage T1/T2 122 62.9 T3/T4 72 37.1 AJCC Stage I 4 2.1 II 10 5.1 IIII 47 24.2 IV 133 68.6 Recurrence Status No Recurrence 131 72.0 Recurrence 51 28.0 Recurrence Type Distant 19 37.3 Locoregional 32 62.7 Extranodal Extension No 110 58.2 Yes 79 41.8 Perineural Invasion No 143 74.1 Yes 50 25.9

9Table 5.1: Patient and tumor sample characteristics

131

Predictor Hazard Ratio 95% CI p-value N HPV+/p16+ DEK Stain Intensity 1.518 0.964 2.392 0.0718 109 DEK Stain Proportion 1.013 0.998 1.028 0.0859 109 DEK Quick Score 1.004 1.000 1.008 0.0388 109 HPV-/p16+ DEK Stain Intensity 0.918 0.589 1.430 0.7043 36 DEK Stain Proportion 1.002 0.990 1.015 0.7217 36 DEK Quick Score 0.999 0.995 1.003 0.6319 36 HPV-/p16- DEK Stain Intensity 0.650 0.421 1.006 0.0531 46 DEK Stain Proportion 0.993 0.982 1.004 0.2063 46 DEK Quick Score 0.997 0.993 1.002 0.2015 46

10Table 5.2: DEK expression is associated with an increased hazard of death in

HPV16+/p16+ but not in HPV16- disease (survival univariate models).

132

HPV16+/p16+, DEK (Stain Proportion) Tumor Stage N Minimum 25th Pctl Median 75th Pctl Maximum p-value T1/T2 76 0.00 53.33 75.00 98.33 100.00 0.0227 T3/T4 33 23.33 80.00 93.33 100.00 100.00 Perineural Invasion N Minimum 25th Pctl Median 75th Pctl Maximum p-value No 90 0.00 60.00 80.00 98.33 100.00 0.2786 Yes 19 11.67 83.33 90.00 100.00 100.00 HPV16-/p16+, DEK (Stain Proportion) Tumor Stage N Minimum 25th Pctl Median 75th Pctl Maximum p-value T1/T2 24 0.00 41.67 81.67 98.33 100.00 0.2399 T3/T4 12 5.00 65.00 92.50 100.00 100.00 Perineural Invasion N Minimum 25th Pctl Median 75th Pctl Maximum p-value No 26 0.00 63.33 89.17 100.00 100.00 0.1189 Yes 10 0.00 43.33 78.33 90.00 100.00 HPV16-/p16-, DEK (Stain Proportion) Tumor Stage N Minimum 25th Pctl Median 75th Pctl Maximum p-value T1/T2 19 0.00 23.33 53.33 76.67 98.33 0.5842 T3/T4 27 0.00 20.00 53.33 83.33 100.00 Perineural Invasion N Minimum 25th Pctl Median 75th Pctl Maximum p-value No 25 0.00 50.00 66.67 86.67 100.00 0.0144 Yes 20 0.00 11.67 33.33 51.67 100.00

11Table 5.3: High DEK expression is associated with higher tumor stage in HPV+/p16+

OPSCC, and reduced perineural invasion in HPV- OPSCC. (Statistics for other clinical characteristics listed in Table S4).

133

17Supplementary Figure 5.1: High DEK expression may be associated with IL6-

status in HPV16- disease.

The proportion of cells staining for DEK (A) and the overall DEK staining (B) was significantly

lower in IL6+ tumors.

134

HPV- HPV+ Survival Survival Alive Dead Alive Dead Biomarker n=29 n=54 n=79 n=31 DEK Stain Intensity Median (IQR) 2 (1-2.7) 1.3 (1-2) 2 (1-2.3) 2 (1.5-3) Mean (SD) 1.8 (1.0) 1.4 (0.9) 1.8 (0.8) 2.1 (0.8) Min – Max 0-3 0-3 0-3 0-3 DEK Stain Proportion Median (IQR) 75 (38.3-95) 66.7 (30-90) 80 (53.3-98.3) 90 (73.3-100) Mean (SD) 64.1 (33.7) 58.5 (35.8) 71.4 (29.9) 80.9 (23.5) Min – Max 0-100 0-100 0-100 0-100 DEK Quick Score Median (IQR) 120 (43.3-230) 96.7 (40-183.3) 140 (73.3-200) 183.3 (113.3-253.3) Mean (SD) 137.2 (106.6) 109.4 (89.2) 145.2 (88.8) 181.0 (84.9) Min – Max 0-300 0-300 0-300 0-300 p16- p16+ Survival Survival Alive Dead Alive Dead Biomarker n=14 n=33 n=94 n=52 DEK Stain Intensity Median (IQR) 1.3 (1-2.5) 1 (1-2) 2 (1-2.3) 2 (1.2-2.7) Mean (SD) 1.7 (0.8) 1.3 (0.8) 1.8 (0.9) 1.9 (0.8) Min – Max 1-3 0-3 0-3 0-3 DEK Stain Proportion Median (IQR) 58.3 (36.7-80) 50 (18.3-80) 81.7 (58.3-100) 88.3 (73.3-100) Mean (SD) 58.7 (28.8) 48.1 (34.1) 71.3 (31.3) 78.5 (27.5) Min – Max 16.7-100 0-100 0-100 0-100 DEK Quick Score Median (IQR) 64.2 (38.3-150) 56.7 (18.3-120) 140 (73.3-200) 166.7 (100-246.7) Mean (SD) 111 (94) 84.7 (79.5) 148.9 (94) 167.8 (88.3) Min – Max 23.3-300 0-266.7 0-300 0-300

12Supplementary Table 5.1: DEK survival summaries based on individual HPV and

p16 status

135

DEK (Stain Intensity) HPV16 Status N Minimum 25th Pctl Median 75th Pctl Maximum p-value Negative 83 0 1 1.33 2.33 3 0.0115 Positive 110 0 1.33 2 2.33 3 p16 Status Negative 47 0 1 1 2 3 0.0036 Positive 146 0 1 2 2.67 3 DEK (Stain Proportion) HPV16 Status N Minimum 25th Pctl Median 75th Pctl Maximum p-value Negative 83 0 30 75 90 100 0.0087 Positive 110 0 60 84.17 98.33 100 p16 Status Negative 47 0 23.33 53.33 80 100 <.0001 Positive 146 0 60 85.83 100 100 DEK (Quick Score) HPV16 Status N Minimum 25th Pctl Median 75th Pctl Maximum p-value Negative 83 0 40 100 200 300 0.0046 Positive 110 0 80 160 206.67 300 p16 Status Negative 47 0 33.33 63.33 150 300 <.0001 Positive 146 0 80 153.33 225 300

13Supplementary Table 5.2: DEK staining intensity, staining proportion, and quick

score correlation with HPV and p16 status in all tumors.

136

Predictor Hazard Ratio 95% CI p-value N HPV16 Negative DEK Stain Intensity 0.763 0.568 1.026 0.0738 83 DEK Stain Proportion 0.996 0.989 1.004 0.3372 83 DEK Quick Score 0.998 0.995 1.001 0.1234 83 HPV16 Positive DEK Stain Intensity 1.535 0.975 2.416 0.0642 110 DEK Stain Proportion 1.013 0.999 1.028 0.0765 110 DEK Quick Score 1.004 1.000 1.008 0.0339 110

14Supplementary Table 5.3: DEK expression is associated with an increased hazard of

death in HPV16+ but not HPV16- disease (survival univariate models).

137

DEK (stain proportion) HPV+/p16+ Node Stage N Minimum Q1 Median Q3 Maximum p-value N0/N1 34 6.67 53.33 81.67 98.33 100.00 0.6537 N2/N3 75 0.00 63.33 86.67 100.00 100.00 Tumor Stage N Minimum Q1 Median Q3 Maximum p-value T1/T2 76 0.00 53.33 75.00 98.33 100.00 0.0227 T3/T4 33 23.33 80.00 93.33 100.00 100.00 Recurrence Status N Minimum Q1 Median Q3 Maximum p-value No Recurrence 87 0.00 60.00 86.67 100.00 100.00 0.8417 Recurrence 17 6.67 73.33 83.33 93.33 100.00 Recurrence Type N Minimum Q1 Median Q3 Maximum p-value Distant 6 70.00 83.33 90.83 100.00 100.00 0.1884 Locoregional 11 6.67 66.67 83.33 93.33 93.33 Smoking Status: Pack N Minimum Q1 Median Q3 Maximum p-value Years 10 pack years or less 34 6.67 60.00 82.50 93.33 100.00 0.4653 More than 10 pack years 68 0.00 60.00 86.67 100.00 100.00 Extranodal Extension N Minimum Q1 Median Q3 Maximum p-value No 69 0.00 60.00 83.33 98.33 100.00 0.4561 Yes 39 11.67 66.67 86.67 100.00 100.00 Perineural Invasion N Minimum Q1 Median Q3 Maximum p-value No 90 0.00 60.00 80.00 98.33 100.00 0.2786 Yes 19 11.67 83.33 90.00 100.00 100.00 HPV-/p16+ Node Stage N Minimum Q1 Median Q3 Maximum p-value N0/N1 9 6.67 76.67 90.00 96.67 100.00 0.6698 N2/N3 27 0.00 43.33 83.33 100.00 100.00 Tumor Stage N Minimum Q1 Median Q3 Maximum p-value T1/T2 24 0.00 41.67 81.67 98.33 100.00 0.2399 T3/T4 12 5.00 65.00 92.50 100.00 100.00 Recurrence Status N Minimum Q1 Median Q3 Maximum p-value No Recurrence 20 0.00 55.00 86.67 100.00 100.00 0.7656 Recurrence 12 0.00 78.33 91.67 100.00 100.00 Recurrence Type N Minimum Q1 Median Q3 Maximum p-value Distant 7 0.00 5.00 80.00 100.00 100.00 0.4553 Locoregional 5 76.67 86.67 96.67 100.00 100.00 Smoking Status: Pack N Minimum Q1 Median Q3 Maximum p-value Years 10 pack years or less 8 3.33 43.33 78.33 89.17 100.00 0.3061 More than 10 pack years 28 0.00 55.00 88.33 100.00 100.00 Extranodal Extension N Minimum Q1 Median Q3 Maximum p-value No 16 0.00 73.33 95.83 100.00 100.00 0.1962 Yes 20 3.33 46.67 80.00 95.00 100.00 Perineural Invasion N Minimum Q1 Median Q3 Maximum p-value No 26 0.00 63.33 89.17 100.00 100.00 0.1189 Yes 10 0.00 43.33 78.33 90.00 100.00 HPV-/p16- Node Stage N Minimum Q1 Median Q3 Maximum p-value N0/N1 26 0.00 30.00 55.00 83.33 100.00 0.3630 N2/N3 20 0.00 19.17 40.83 71.67 100.00 Tumor Stage N Minimum Q1 Median Q3 Maximum p-value T1/T2 19 0.00 23.33 53.33 76.67 98.33 0.5842 T3/T4 27 0.00 20.00 53.33 83.33 100.00 Recurrence Status N Minimum Q1 Median Q3 Maximum p-value No Recurrence 22 0.00 23.33 51.67 88.33 100.00 0.7984 Recurrence 21 0.00 25.00 56.67 80.00 100.00 Recurrence Type N Minimum Q1 Median Q3 Maximum p-value

138

Distant 6 0.00 25.00 35.00 76.67 83.33 0.4832 Locoregional 15 0.00 18.33 60.00 80.00 100.00 Smoking Status: Pack N Minimum Q1 Median Q3 Maximum p-value Years 10 pack years or less 3 0.00 0.00 76.67 98.33 98.33 0.7846 More than 10 pack years 42 0.00 23.33 53.33 80.00 100.00 Extranodal Extension N Minimum Q1 Median Q3 Maximum p-value No 24 0.00 26.67 53.33 80.00 100.00 0.6109 Yes 18 0.00 25.00 48.33 80.00 100.00 Perineural Invasion N Minimum Q1 Median Q3 Maximum p-value No 25 0.00 50.00 66.67 86.67 100.00 0.0144 Yes 20 0.00 11.67 33.33 51.67 100.00

15Supplementary Table 5.4: Clinical characteristics associated with DEK staining,

based on HPV/p16 status (DEK stain proportion).

139

Predictor Hazard Ratio 95% CI p-value N p16 Positive DEK Stain Intensity 1.134 0.821 1.567 0.4451 146 DEK Stain Proportion 1.007 0.997 1.018 0.1443 146 DEK Quick Score 1.002 0.999 1.005 0.2534 146 p16 Negative DEK Stain Intensity 0.666 0.429 1.034 0.0704 47 DEK Stain Proportion 0.993 0.982 1.005 0.2471 47 DEK Quick Score 0.997 0.993 1.002 0.2579 47

16Supplementary Table 5.5: DEK is not correlated with survival in p16 mono-labeled

tumors (survival univariate models).

140

Chapter 6: Discussion and future directions

141

Due to its association with chemotherapy resistance, higher grade malignancies, and poor survival outcomes, the DEK oncogene has been implicated in the prognosis of many types of tumors and has been sought as a therapeutic target178,280. This dissertation work addresses both of these clinical interests in DEK. In chapter 4, the function of DEK in HR was examined. While much is understood about the biochemistry of DEK, the molecular details by which it drives chemotherapy resistance in tumors have been uncertain at best. By beginning to understand the mechanisms underpinning how DEK promotes chemotherapy resistance through HR and is synthetic lethal with

DNA-PK inhibition, we move closer to realizing its potential as a drug target. Likewise, in chapter 5, a moderate-sized retrospective cohort study was utilized to assess the potential of DEK as a prognostic biomarker in OPSCC. Similar scale-up studies have been reported for breast, melanoma, and colorectal cancers370,372,378. Akin to these reports, chapter 5 provides a rationale for pursuing a larger clinical trial to develop and validate a DEK biomarker test for clinical use in HPV+ OPSCC. The studies in this dissertation, while focusing on translating DEK to the clinic, also carry important implications for OPSCC and DEK biology while providing fertile ground for new research endeavors.

As far as HR factors go, DEK is highly unusual. Knockdown of the protein reduced HR DSB repair to a similar level as a BRCA1 deficiency, but DEK was expendable for repair of replication forks. Since most HR factors are essential for both

DSB and replication-associated repair, this sets DEK apart from canonical HR factors and indicates a specialized role for DSB repair. There are two potential mechanisms for how DEK functions in this arm of HR. First, we know that DEK can form a complex with

142

RAD51, and that recruitment of RAD51 to RPA foci is mildly attenuated, suggesting that

DEK may assist in loading RAD51 onto DNA. However, the mild impact on RAD51/RPA foci co-localization and the lack of interaction with BRCA1 indicate that RAD51 loading may not be the critical function of DEK. Rather, the affinity of DEK for cruciform DNA structures suggest that it may assist in stabilizing the D-loop structure during RAD51- mediated strand invasion. Examination of these hypotheses is a natural extension of this work and may uncover the specific, and potentially druggable, mechanism by which

DEK participates in and, is essential, for HR DSB repair.

As a well-appreciated histone chaperone and regulator of epigenetic changes, it is not surprising that DEK can regulate γH2AX phosphorylation during DNA repair.

While signaling through ATM appears to be intact in DEK-depleted cells after IR treatment, these cells have severely attenuated γH2AX formation, suggesting an important role for DEK in rendering H2AX as a substrate for (de)-phosphorylation.

Mechanistically, DEK may accomplish this through directly loading H2AX onto chromatin near sites of DSBs, in manner similar to how the protein constrains H3.3 to telomeres and PML bodies302, or by binding and reconfiguring H2AX into a conformation better accepted by ATM and other sensor kinases. Conversely, chemotherapeutic treatment in DEK-deficient cells results in increased γH2AX accumulation212,317, suggesting that DEK may have differing functions based on the type of DNA damage accrued.

The molecular findings discovered in this dissertation have revealed and highlighted paradoxes regarding DEK function in CRT resistance and repair. First, while it is well appreciated that DEK loss results in CRT sensitization212,314, it has yet to be

143 identified if DEK overexpression drives therapy resistance. In preliminary experiments not shown in this dissertation, DEK overexpression did not enhance radiation or individual agent chemotherapy resistance in MEFs or normal oral keratinocytes as measured by colony assay. This preliminary data indicates that the presence of DEK is necessary to allow resistance, but overexpression may not further enhance resistance.

However, this interpretation is in direct conflict with clinical data showing a correlation between DEK expression and chemotherapy resistance278, thus a paradox. A second potential paradox that was uncovered concerns the relationship between DEK and

H2AX phosphorylation. Following chemotherapy treatment, γH2AX levels are increased upon DEK loss212, but the opposite is seen following IR treatment (Figure 4.3).

Mechanistically, it is possible that DEK is required to chaperone H2AX to the DSB in response to IR damage, and that this may be accomplished by other factors in the context of stalled replication forks or complex DSBs induced by chemotherapeutics.

Alternatively, DEK may modulate the ability of phosphatases or sensor kinases to

(de)phosphorylate H2AX, promoting the generation of γH2AX in response to IR, but suppressing γH2AX activation in response to chemotherapy. This latter scenario also supports a model by which DEK promotes the restart of stalled replication forks317, potentially by avoiding fork collapse and HR repair. Solving these paradoxes is essential to understand how DEK promotes treatment resistance and should become a priority in the DEK field.

The synthetic lethal relationship between DEK and DNA-PK has important biological and clinical implications. First, DEK-deficient cells are clearly relying on DNA-

PK function, and by extension NHEJ, for survival. This suggests that the DSB repair

144 salvage pathways, single strand annealing and microhomology-mediated end-joining, are likely also compromised by DEK loss and/or are insufficient to repair endogenous

DSBs in the absence of NHEJ and DEK-mediated HR repair. It is still unknown how well a combination therapy of DEK and DNA-PK inhibition would be tolerated in vivo, especially considering that untransformed Dek knockout cells underwent apoptosis with

DNA-PK inhibition. However, developing a DEK inhibitor to use in such experiments has proven challenging for several reasons. DEK has no known enzymatic activity that can be readily targeted, and the lack of a full length crystal structure precludes in silico screening for inhibiting compounds. One way to overcome these obstacles is to understand the mechanism by which DEK drives a clinical characteristic, such as chemotherapy resistance. By isolating the specific step where DEK performs key function in HR, we are one step closer to developing a “DEK function assay” that can now be screened for inhibitors. Such an assay could measure the hypothetical DEK- mediated D-loop stabilization. Beyond small molecule development, novel methods of inhibiting DEK are being developed, such as the DNA-aptamer for extracellular DEK in the treatment of arthritis454. Designing similar systems to inactivate or block translation of intracellular DEK may supersede the need for small molecule inhibitors, but could still be validated with a future DEK function assay. If targetable to specific cell types, nanoparticle siRNA delivery therapeutics may also overcome toxicity that would likely be associated with dual DNA-PK and DEK inhibition.

A previous pilot study by the Susanne Wells’ laboratory found that DEK protein was highly expressed across 20 OPSCC patient samples318. The expanded retrospective case study (194 patients) described in chapter 5 sought to validate these

145 results and determine if DEK could be used as a potential prognostic biomarker in

OPSCC. For HPV16+ tumors, the oncoprotein was significantly elevated in larger tumors (T3/T4) and was a predictor of worse prognosis. The results of this study are promising and support the utility of the DEK biomarker screen in larger multi-institutional clinical trial testing.

For the overall HPV16- and double-negative HPV16-/p16- OPSCC cohorts, there was no correlation between DEK expression and survival or tumor stage. This is an important observation that suggests HPV16- OPSCC is one of the few solid tumors where DEK does not have prognostic utility. It is possible that this is true for other tumor subtypes, so care should be taken when assessing the prognostic value of DEK in other types of malignancies. In this study, similiar conclusions cannot be drawn in the HPV16-

/p16+ cohort due to low sample size and tumor heterogeneity. Many of these tumors are likely positive for HPV subtypes other than HPV16 and would be expected to behave similarly to HPV16+ tumors. However, a sizeable portion of the true HPV- population are predicted to have chromosomal 6p22.3 amplification, which has not been reported prior to this work453. The amplification event duplicates both the DEK and E2F3 genes, the latter of which is implicated in upregulating p16 expression451. Further studies that separate all true HPV-/p16+ would be required to assess a prognostic function of DEK in this population.

In summary, the work described in this dissertation supports the development of

DEK as a prognostic biomarker for HPV+ OPSCC and defines a mechanism for whereby the oncoprotein supports chemotherapy resistance. While further research is still required to realize the potential of DEK as a biomarker and to ascertain the

146 biochemical mechanism of the protein’s role in HR, these studies laid the groundwork and have defined the necessary hypotheses that need to be studied to further develop

DEK as a clinical tool and drug target.

147

REFERENCES

1 Haines, G. K. in Head & Neck Cancer: Current Perspectives, Advances, and Challenges (ed James A. Radosevich) 257-287 (Springer Netherlands, 2013). 2 Torre, L. A. et al. Global cancer statistics, 2012. CA: a cancer journal for clinicians 65, 87-108, doi:10.3322/caac.21262 (2015). 3 Stein, A. P., Saha, S., Yu, M., Kimple, R. J. & Lambert, P. F. Prevalence of human papillomavirus in oropharyngeal squamous cell carcinoma in the United States across time. Chem Res Toxicol 27, 462-469, doi:10.1021/tx500034c (2014). 4 Marur, S. & Forastiere, A. A. Head and Neck Squamous Cell Carcinoma: Update on Epidemiology, Diagnosis, and Treatment. Mayo Clin Proc 91, 386-396, doi:10.1016/j.mayocp.2015.12.017 (2016). 5 Chaturvedi, A. K. et al. Human papillomavirus and rising oropharyngeal cancer incidence in the United States. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 29, 4294-4301, doi:10.1200/JCO.2011.36.4596 (2011). 6 Gillison, M. L., Chaturvedi, A. K., Anderson, W. F. & Fakhry, C. Epidemiology of Human Papillomavirus-Positive Head and Neck Squamous Cell Carcinoma. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 33, 3235-3242, doi:10.1200/JCO.2015.61.6995 (2015). 7 Marur, S. & Forastiere, A. A. Head and neck cancer: changing epidemiology, diagnosis, and treatment. Mayo Clin Proc 83, 489-501, doi:10.4065/83.4.489 (2008). 8 Blot, W. J. et al. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 48, 3282-3287 (1988). 9 Wyss, A. et al. Cigarette, cigar, and pipe smoking and the risk of head and neck cancers: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. Am J Epidemiol 178, 679-690, doi:10.1093/aje/kwt029 (2013). 10 Kelly, J. R., Husain, Z. A. & Burtness, B. Treatment de-intensification strategies for head and neck cancer. Eur J Cancer 68, 125-133, doi:10.1016/j.ejca.2016.09.006 (2016). 11 Ahmed, M., Hansen, V. N., Harrington, K. J. & Nutting, C. M. Reducing the risk of xerostomia and mandibular osteoradionecrosis: the potential benefits of intensity modulated radiotherapy in advanced oral cavity carcinoma. Med Dosim 34, 217-224, doi:10.1016/j.meddos.2008.08.008 (2009). 12 Funk, G. F., Karnell, L. H. & Christensen, A. J. Long-term health-related quality of life in survivors of head and neck cancer. Archives of otolaryngology--head & neck surgery 138, 123-133, doi:10.1001/archoto.2011.234 (2012). 13 Bagri, P. K. et al. Comparative analysis of cisplatin-induced nephrotoxicity in head and neck cancer and carcinoma cervix during concurrent chemoradiotherapy. South Asian J Cancer 3, 217- 220, doi:10.4103/2278-330X.142979 (2014). 14 Zuur, C. L. et al. Ototoxicity in a randomized phase III trial of intra-arterial compared with intravenous cisplatin chemoradiation in patients with locally advanced head and neck cancer. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 25, 3759-3765, doi:10.1200/JCO.2006.08.9540 (2007). 15 Wang, X. & Eisbruch, A. IMRT for head and neck cancer: reducing xerostomia and dysphagia. Journal of radiation research 57 Suppl 1, i69-i75, doi:10.1093/jrr/rrw047 (2016). 16 Posner, M. R. et al. Survival and human papillomavirus in oropharynx cancer in TAX 324: a subset analysis from an international phase III trial. Ann Oncol 22, 1071-1077, doi:10.1093/annonc/mdr006 (2011).

148

17 Melkane, A. E. et al. Human papillomavirus prevalence and prognostic implication in oropharyngeal squamous cell carcinomas. Head Neck 36, 257-265, doi:10.1002/hed.23302 (2014). 18 Liu, S. Z., Zandberg, D. P., Schumaker, L. M., Papadimitriou, J. C. & Cullen, K. J. Correlation of p16 expression and HPV type with survival in oropharyngeal squamous cell cancer. Oral oncology 51, 862-869, doi:10.1016/j.oraloncology.2015.06.014 (2015). 19 Moore, K. A., 2nd & Mehta, V. The Growing Epidemic of HPV-Positive Oropharyngeal Carcinoma: A Clinical Review for Primary Care Providers. J Am Board Fam Med 28, 498-503, doi:10.3122/jabfm.2015.04.140301 (2015). 20 Huang, S. H. et al. Refining American Joint Committee on Cancer/Union for International Cancer Control TNM stage and prognostic groups for human papillomavirus-related oropharyngeal carcinomas. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 33, 836-845, doi:10.1200/JCO.2014.58.6412 (2015). 21 Amin, B. M. S. B. G. F. L. in AJCC cancer staging manual (ed B.M.; S.B.; Greene F.L. Amin) 1024 (Springer, France, 2017). 22 Gillison, M. L. et al. Evidence for a causal association between human papillomavirus and a subset of head and neck cancers. Journal of the National Cancer Institute 92, 709-720 (2000). 23 Gillison, M. L. et al. Prevalence of oral HPV infection in the United States, 2009-2010. JAMA 307, 693-703, doi:10.1001/jama.2012.101 (2012). 24 Gillison, M. L. Human Papillomavirus and Oropharyngeal Cancer Stage. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 34, 1833-1835, doi:10.1200/JCO.2016.66.6081 (2016). 25 Dahlstrom, K. R., Garden, A. S., William, W. N., Jr., Lim, M. Y. & Sturgis, E. M. Proposed Staging System for Patients With HPV-Related Oropharyngeal Cancer Based on Nasopharyngeal Cancer N Categories. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 34, 1848-1854, doi:10.1200/JCO.2015.64.6448 (2016). 26 Platek, A. J. et al. Smoking cessation is associated with improved survival in oropharynx cancer treated by chemoradiation. Laryngoscope 126, 2733-2738, doi:10.1002/lary.26083 (2016). 27 Liskamp, C. P. et al. Adverse effect of smoking on prognosis in human papillomavirus-associated oropharyngeal carcinoma. Head Neck 38, 1780-1787, doi:10.1002/hed.24516 (2016). 28 Gensheimer, M. F., Liao, J. J., Garden, A. S., Laramore, G. E. & Parvathaneni, U. Submandibular gland-sparing radiation therapy for locally advanced oropharyngeal squamous cell carcinoma: patterns of failure and xerostomia outcomes. Radiation oncology 9, 255, doi:10.1186/s13014- 014-0255-x (2014). 29 van Rij, C. M. et al. Parotid gland sparing IMRT for head and neck cancer improves xerostomia related quality of life. Radiation oncology 3, 41, doi:10.1186/1748-717X-3-41 (2008). 30 Chajon, E. et al. Salivary gland-sparing other than parotid-sparing in definitive head-and-neck intensity-modulated radiotherapy does not seem to jeopardize local control. Radiation oncology 8, 132, doi:10.1186/1748-717X-8-132 (2013). 31 Duarte, V. M. et al. Comparison of dental health of patients with head and neck cancer receiving IMRT vs conventional radiation. Otolaryngol Head Neck Surg 150, 81-86, doi:10.1177/0194599813509586 (2014). 32 Nabil, S. & Samman, N. Incidence and prevention of osteoradionecrosis after dental extraction in irradiated patients: a systematic review. Int J Oral Maxillofac Surg 40, 229-243, doi:10.1016/j.ijom.2010.10.005 (2011). 33 Weinstein, G. S. et al. Transoral robotic surgery: a multicenter study to assess feasibility, safety, and surgical margins. Laryngoscope 122, 1701-1707, doi:10.1002/lary.23294 (2012).

149

34 Laccourreye, O. et al. Transoral lateral oropharyngectomy for squamous cell carcinoma of the tonsillar region: II. An analysis of the incidence, related variables, and consequences of local recurrence. Archives of otolaryngology--head & neck surgery 131, 592-599, doi:10.1001/archotol.131.7.592 (2005). 35 Hamilton, D., Khan, M. K., O'Hara, J. & Paleri, V. The changing landscape of oropharyngeal cancer management. J Laryngol Otol 131, 3-7, doi:10.1017/S0022215116009178 (2017). 36 Haughey, B. H. et al. Transoral laser microsurgery as primary treatment for advanced-stage oropharyngeal cancer: a United States multicenter study. Head Neck 33, 1683-1694, doi:10.1002/hed.21669 (2011). 37 Moore, E. J. et al. Long-term functional and oncologic results of transoral robotic surgery for oropharyngeal squamous cell carcinoma. Mayo Clin Proc 87, 219-225, doi:10.1016/j.mayocp.2011.10.007 (2012). 38 Blanchard, P. et al. Meta-analysis of chemotherapy in head and neck cancer (MACH-NC): a comprehensive analysis by tumour site. Radiother Oncol 100, 33-40, doi:10.1016/j.radonc.2011.05.036 (2011). 39 Magrini, S. M. et al. Cetuximab and Radiotherapy Versus Cisplatin and Radiotherapy for Locally Advanced Head and Neck Cancer: A Randomized Phase II Trial. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 34, 427-435, doi:10.1200/JCO.2015.63.1671 (2016). 40 Lorch, J. H. et al. Induction chemotherapy with cisplatin and fluorouracil alone or in combination with docetaxel in locally advanced squamous-cell cancer of the head and neck: long-term results of the TAX 324 randomised phase 3 trial. The lancet oncology 12, 153-159, doi:10.1016/S1470- 2045(10)70279-5 (2011). 41 Zhang, L. et al. Induction chemotherapy with concurrent chemoradiotherapy versus concurrent chemoradiotherapy for locally advanced squamous cell carcinoma of head and neck: a meta- analysis. Sci Rep 5, 10798, doi:10.1038/srep10798 (2015). 42 Guo, S. S. et al. Induction chemotherapy followed by concurrent chemoradiotherapy versus concurrent chemoradiotherapy alone in stage III-IVb nasopharyngeal carcinoma patients with Epstein-Barr virus DNA >/=4000 copies/ml: a matched study. Oncotarget 7, 29739-29748, doi:10.18632/oncotarget.8828 (2016). 43 Oksuz, D. C. et al. Recurrence patterns of locally advanced head and neck squamous cell carcinoma after 3D conformal (chemo)-radiotherapy. Radiation oncology 6, 54, doi:10.1186/1748-717X-6-54 (2011). 44 Brockstein, B. et al. Patterns of failure, prognostic factors and survival in locoregionally advanced head and neck cancer treated with concomitant chemoradiotherapy: a 9-year, 337- patient, multi-institutional experience. Ann Oncol 15, 1179-1186, doi:10.1093/annonc/mdh308 (2004). 45 Vermorken, J. B. & Specenier, P. Optimal treatment for recurrent/metastatic head and neck cancer. Ann Oncol 21 Suppl 7, vii252-261, doi:10.1093/annonc/mdq453 (2010). 46 Vermorken, J. B. et al. Open-label, uncontrolled, multicenter phase II study to evaluate the efficacy and toxicity of cetuximab as a single agent in patients with recurrent and/or metastatic squamous cell carcinoma of the head and neck who failed to respond to platinum-based therapy. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 25, 2171-2177, doi:10.1200/JCO.2006.06.7447 (2007). 47 Ritter, M. et al. Second-line treatment of recurrent HNSCC: tumor debulking in combination with high-dose-rate brachytherapy and a simultaneous cetuximab-paclitaxel protocol. Radiation oncology 11, 6, doi:10.1186/s13014-016-0583-0 (2016).

150

48 Gibson, M. K. et al. Randomized phase III evaluation of cisplatin plus fluorouracil versus cisplatin plus paclitaxel in advanced head and neck cancer (E1395): an intergroup trial of the Eastern Cooperative Oncology Group. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 23, 3562-3567, doi:10.1200/JCO.2005.01.057 (2005). 49 Forastiere, A. A. et al. Randomized comparison of cisplatin plus fluorouracil and carboplatin plus fluorouracil versus methotrexate in advanced squamous-cell carcinoma of the head and neck: a Southwest Oncology Group study. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 10, 1245-1251, doi:10.1200/JCO.1992.10.8.1245 (1992). 50 Pauloski, B. R. et al. Pretreatment swallowing function in patients with head and neck cancer. Head Neck 22, 474-482 (2000). 51 Bressmann, T., Sader, R., Whitehill, T. L. & Samman, N. Consonant intelligibility and tongue motility in patients with partial glossectomy. J Oral Maxillofac Surg 62, 298-303 (2004). 52 Stelzle, F. et al. Automatic quantification of speech intelligibility in patients after treatment for oral squamous cell carcinoma. J Oral Maxillofac Surg 69, 1493-1500, doi:10.1016/j.joms.2010.05.077 (2011). 53 Fingeret, M. C. et al. Associations among speech, eating, and body image concerns for surgical patients with head and neck cancer. Head Neck 35, 354-360, doi:10.1002/hed.22980 (2013). 54 Ridner, S. H. et al. A Prospective Study of the Lymphedema and Fibrosis Continuum in Patients with Head and Neck Cancer. Lymphat Res Biol 14, 198-205, doi:10.1089/lrb.2016.0001 (2016). 55 Jackson, L. K. et al. Internal Lymphedema Correlates with Subjective and Objective Measures of Dysphagia in Head and Neck Cancer Patients. J Palliat Med 19, 949-956, doi:10.1089/jpm.2016.0018 (2016). 56 El-Deiry, M. et al. Long-term quality of life for surgical and nonsurgical treatment of head and neck cancer. Archives of otolaryngology--head & neck surgery 131, 879-885, doi:10.1001/archotol.131.10.879 (2005). 57 Wijers, O. B. et al. Patients with head and neck cancer cured by radiation therapy: a survey of the dry mouth syndrome in long-term survivors. Head Neck 24, 737-747, doi:10.1002/hed.10129 (2002). 58 Bensadoun, R. J. et al. A systematic review of trismus induced by cancer therapies in head and neck cancer patients. Support Care Cancer 18, 1033-1038, doi:10.1007/s00520-010-0847-4 (2010). 59 Peterson, D. E. et al. Osteoradionecrosis in cancer patients: the evidence base for treatment- dependent frequency, current management strategies, and future studies. Support Care Cancer 18, 1089-1098, doi:10.1007/s00520-010-0898-6 (2010). 60 Guru, K., Manoor, U. K. & Supe, S. S. A comprehensive review of head and neck cancer rehabilitation: physical therapy perspectives. Indian J Palliat Care 18, 87-97, doi:10.4103/0973- 1075.100820 (2012). 61 Windebank, A. J. & Grisold, W. Chemotherapy-induced neuropathy. J Peripher Nerv Syst 13, 27- 46, doi:10.1111/j.1529-8027.2008.00156.x (2008). 62 Yao, M. et al. Health-related quality-of-life outcomes following IMRT versus conventional radiotherapy for oropharyngeal squamous cell carcinoma. International journal of radiation oncology, biology, physics 69, 1354-1360, doi:10.1016/j.ijrobp.2007.05.003 (2007). 63 Baujat, B. et al. Hyperfractionated or accelerated radiotherapy for head and neck cancer. Cochrane Database Syst Rev, CD002026, doi:10.1002/14651858.CD002026 (2010). 64 Scully, C. Oral squamous cell carcinoma; from an hypothesis about a virus, to concern about possible sexual transmission. Oral oncology 38, 227-234 (2002). 65 Ang, K. K. et al. Human papillomavirus and survival of patients with oropharyngeal cancer. The New England journal of medicine 363, 24-35, doi:10.1056/NEJMoa0912217 (2010).

151

66 Westra, W. H. The changing face of head and neck cancer in the 21st century: the impact of HPV on the epidemiology and pathology of oral cancer. Head and neck pathology 3, 78-81, doi:10.1007/s12105-009-0100-y (2009). 67 Vidal, L. & Gillison, M. L. Human papillomavirus in HNSCC: recognition of a distinct disease type. Hematology/oncology clinics of North America 22, 1125-1142, vii, doi:10.1016/j.hoc.2008.08.006 (2008). 68 Kreimer, A. R., Clifford, G. M., Boyle, P. & Franceschi, S. Human papillomavirus types in head and neck squamous cell carcinomas worldwide: a systematic review. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology 14, 467-475, doi:10.1158/1055- 9965.EPI-04-0551 (2005). 69 Scheffner, M., Huibregtse, J. M., Vierstra, R. D. & Howley, P. M. The HPV-16 E6 and E6-AP complex functions as a ubiquitin-protein ligase in the ubiquitination of p53. Cell 75, 495-505 (1993). 70 Vousden, K. H. & Prives, C. Blinded by the Light: The Growing Complexity of p53. Cell 137, 413- 431, doi:10.1016/j.cell.2009.04.037 (2009). 71 Patel, D., Huang, S. M., Baglia, L. A. & McCance, D. J. The E6 protein of human papillomavirus type 16 binds to and inhibits co-activation by CBP and p300. EMBO J 18, 5061-5072, doi:10.1093/emboj/18.18.5061 (1999). 72 Thomas, M. C. & Chiang, C. M. E6 oncoprotein represses p53-dependent gene activation via inhibition of protein acetylation independently of inducing p53 degradation. Molecular cell 17, 251-264, doi:10.1016/j.molcel.2004.12.016 (2005). 73 Xie, X. et al. Targeting HPV16 E6-p300 interaction reactivates p53 and inhibits the tumorigenicity of HPV-positive head and neck squamous cell carcinoma. Oncogene 33, 1037-1046, doi:10.1038/onc.2013.25 (2014). 74 Filippova, M., Song, H., Connolly, J. L., Dermody, T. S. & Duerksen-Hughes, P. J. The human papillomavirus 16 E6 protein binds to tumor necrosis factor (TNF) R1 and protects cells from TNF-induced apoptosis. The Journal of biological chemistry 277, 21730-21739, doi:10.1074/jbc.M200113200 (2002). 75 Filippova, M., Parkhurst, L. & Duerksen-Hughes, P. J. The human papillomavirus 16 E6 protein binds to Fas-associated death domain and protects cells from Fas-triggered apoptosis. The Journal of biological chemistry 279, 25729-25744, doi:10.1074/jbc.M401172200 (2004). 76 Garnett, T. O., Filippova, M. & Duerksen-Hughes, P. J. Accelerated degradation of FADD and procaspase 8 in cells expressing human papilloma virus 16 E6 impairs TRAIL-mediated apoptosis. Cell death and differentiation 13, 1915-1926, doi:10.1038/sj.cdd.4401886 (2006). 77 Gross-Mesilaty, S. et al. Basal and human papillomavirus E6 oncoprotein-induced degradation of Myc proteins by the ubiquitin pathway. Proc Natl Acad Sci U S A 95, 8058-8063 (1998). 78 Thomas, M. & Banks, L. Inhibition of Bak-induced apoptosis by HPV-18 E6. Oncogene 17, 2943- 2954, doi:10.1038/sj.onc.1202223 (1998). 79 Klingelhutz, A. J., Foster, S. A. & McDougall, J. K. Telomerase activation by the E6 gene product of human papillomavirus type 16. Nature 380, 79-82, doi:10.1038/380079a0 (1996). 80 McMurray, H. R. & McCance, D. J. Human papillomavirus type 16 E6 activates TERT gene transcription through induction of c-Myc and release of USF-mediated repression. J Virol 77, 9852-9861 (2003). 81 Gewin, L., Myers, H., Kiyono, T. & Galloway, D. A. Identification of a novel telomerase repressor that interacts with the human papillomavirus type-16 E6/E6-AP complex. Genes & development 18, 2269-2282, doi:10.1101/gad.1214704 (2004).

152

82 Katzenellenbogen, R. A., Vliet-Gregg, P., Xu, M. & Galloway, D. A. NFX1-123 increases hTERT expression and telomerase activity posttranscriptionally in human papillomavirus type 16 E6 keratinocytes. J Virol 83, 6446-6456, doi:10.1128/JVI.02556-08 (2009). 83 Liu, X. et al. HPV E6 protein interacts physically and functionally with the cellular telomerase complex. Proc Natl Acad Sci U S A 106, 18780-18785, doi:10.1073/pnas.0906357106 (2009). 84 Jung, Y. S., Kato, I. & Kim, H. R. A novel function of HPV16-E6/E7 in epithelial-mesenchymal transition. Biochemical and biophysical research communications 435, 339-344, doi:10.1016/j.bbrc.2013.04.060 (2013). 85 Duensing, S. et al. Centrosome abnormalities and genomic instability by episomal expression of human papillomavirus type 16 in raft cultures of human keratinocytes. J Virol 75, 7712-7716, doi:10.1128/JVI.75.16.7712-7716.2001 (2001). 86 Zhang, Y. et al. BRCA1 interaction with human papillomavirus oncoproteins. The Journal of biological chemistry 280, 33165-33177, doi:10.1074/jbc.M505124200 (2005). 87 Iftner, T. et al. Interference of papillomavirus E6 protein with single-strand break repair by interaction with XRCC1. EMBO J 21, 4741-4748 (2002). 88 Wallace, N. A., Robinson, K., Howie, H. L. & Galloway, D. A. HPV 5 and 8 E6 abrogate ATR activity resulting in increased persistence of UVB induced DNA damage. PLoS pathogens 8, e1002807, doi:10.1371/journal.ppat.1002807 (2012). 89 Shin, K. H. et al. HPV-16 E6 oncoprotein impairs the fidelity of DNA end-joining via p53- dependent and -independent pathways. International journal of oncology 28, 209-215 (2006). 90 Moody, C. A. & Laimins, L. A. Human papillomaviruses activate the ATM DNA damage pathway for viral genome amplification upon differentiation. PLoS pathogens 5, e1000605, doi:10.1371/journal.ppat.1000605 (2009). 91 Dyson, N., Howley, P. M., Munger, K. & Harlow, E. The human papilloma virus-16 E7 oncoprotein is able to bind to the retinoblastoma gene product. Science 243, 934-937 (1989). 92 Dyson, N. The regulation of E2F by pRB-family proteins. Genes & development 12, 2245-2262 (1998). 93 Stevaux, O. & Dyson, N. J. A revised picture of the E2F transcriptional network and RB function. Current opinion in cell biology 14, 684-691 (2002). 94 Boyer, S. N., Wazer, D. E. & Band, V. E7 protein of human papilloma virus-16 induces degradation of retinoblastoma protein through the ubiquitin-proteasome pathway. Cancer Res 56, 4620-4624 (1996). 95 McLaughlin-Drubin, M. E., Huh, K. W. & Munger, K. Human papillomavirus type 16 E7 oncoprotein associates with E2F6. J Virol 82, 8695-8705, doi:10.1128/JVI.00579-08 (2008). 96 Hwang, S. G., Lee, D., Kim, J., Seo, T. & Choe, J. Human papillomavirus type 16 E7 binds to E2F1 and activates E2F1-driven transcription in a retinoblastoma protein-independent manner. The Journal of biological chemistry 277, 2923-2930, doi:10.1074/jbc.M109113200 (2002). 97 Avvakumov, N., Torchia, J. & Mymryk, J. S. Interaction of the HPV E7 proteins with the pCAF acetyltransferase. Oncogene 22, 3833-3841, doi:10.1038/sj.onc.1206562 (2003). 98 Wittekindt, C., Wagner, S., Mayer, C. S. & Klussmann, J. P. [Basics of tumor development and importance of human papilloma virus (HPV) for head and neck cancer]. Laryngo- rhino- otologie 91 Suppl 1, S1-26, doi:10.1055/s-0031-1297241 (2012). 99 Brehm, A. et al. The E7 oncoprotein associates with Mi2 and histone deacetylase activity to promote cell growth. EMBO J 18, 2449-2458, doi:10.1093/emboj/18.9.2449 (1999). 100 Carro, M. S. et al. DEK Expression is controlled by E2F and deregulated in diverse tumor types. Cell Cycle 5, 1202-1207 (2006). 101 Deshpande, A., Sicinski, P. & Hinds, P. W. Cyclins and cdks in development and cancer: a perspective. Oncogene 24, 2909-2915, doi:10.1038/sj.onc.1208618 (2005).

153

102 Zerfass-Thome, K. et al. Inactivation of the cdk inhibitor p27KIP1 by the human papillomavirus type 16 E7 oncoprotein. Oncogene 13, 2323-2330 (1996). 103 Funk, J. O. et al. Inhibition of CDK activity and PCNA-dependent DNA replication by p21 is blocked by interaction with the HPV-16 E7 oncoprotein. Genes & development 11, 2090-2100 (1997). 104 Nguyen, C. L. & Munger, K. Direct association of the HPV16 E7 oncoprotein with cyclin A/CDK2 and cyclin E/CDK2 complexes. Virology 380, 21-25, doi:10.1016/j.virol.2008.07.017 (2008). 105 Katich, S. C., Zerfass-Thome, K. & Hoffmann, I. Regulation of the Cdc25A gene by the human papillomavirus Type 16 E7 oncogene. Oncogene 20, 543-550, doi:10.1038/sj.onc.1204130 (2001). 106 Nguyen, D. X., Westbrook, T. F. & McCance, D. J. Human papillomavirus type 16 E7 maintains elevated levels of the cdc25A tyrosine phosphatase during deregulation of cell cycle arrest. J Virol 76, 619-632 (2002). 107 Moody, C. A. & Laimins, L. A. Human papillomavirus oncoproteins: pathways to transformation. Nature reviews. Cancer 10, 550-560, doi:10.1038/nrc2886 (2010). 108 Zhong, Z. H. et al. Disruption of telomere maintenance by depletion of the MRE11/RAD50/NBS1 complex in cells that use alternative lengthening of telomeres. The Journal of biological chemistry 282, 29314-29322, doi:10.1074/jbc.M701413200 (2007). 109 Spardy, N., Duensing, A., Hoskins, E. E., Wells, S. I. & Duensing, S. HPV-16 E7 reveals a link between DNA replication stress, fanconi anemia D2 protein, and alternative lengthening of telomere-associated promyelocytic leukemia bodies. Cancer Res 68, 9954-9963, doi:10.1158/0008-5472.CAN-08-0224 (2008). 110 Freire, R., van Vugt, M. A., Mamely, I. & Medema, R. H. Claspin: timing the cell cycle arrest when the genome is damaged. Cell Cycle 5, 2831-2834 (2006). 111 Spardy, N. et al. Human papillomavirus 16 E7 oncoprotein attenuates DNA damage checkpoint control by increasing the proteolytic turnover of claspin. Cancer Res 69, 7022-7029, doi:10.1158/0008-5472.CAN-09-0925 (2009). 112 Duensing, S. & Munger, K. The human papillomavirus type 16 E6 and E7 oncoproteins independently induce numerical and structural chromosome instability. Cancer Res 62, 7075- 7082 (2002). 113 Gillespie, K. A., Mehta, K. P., Laimins, L. A. & Moody, C. A. Human papillomaviruses recruit cellular DNA repair and homologous recombination factors to viral replication centers. J Virol 86, 9520-9526, doi:10.1128/JVI.00247-12 (2012). 114 Hong, S. & Laimins, L. A. -The JAK-STAT Transcriptional Regulator, STAT-5, Activates the ATM DNA Damage Pathway to Induce HPV 31 Genome Amplification upon Epithelial Differentiation. PLoS pathogens 9, e1003295, doi:10.1371/journal.ppat.1003295 (2013). 115 Kaverina, I., Krylyshkina, O. & Small, J. V. Regulation of substrate adhesion dynamics during cell motility. The international journal of biochemistry & cell biology 34, 746-761 (2002). 116 Hellner, K., Mar, J., Fang, F., Quackenbush, J. & Munger, K. HPV16 E7 oncogene expression in normal human epithelial cells causes molecular changes indicative of an epithelial to mesenchymal transition. Virology 391, 57-63, doi:10.1016/j.virol.2009.05.036 (2009). 117 Todorovic, B. et al. The human papillomavirus E7 proteins associate with p190RhoGAP and alter its function. J Virol 88, 3653-3663, doi:10.1128/JVI.03263-13 (2014). 118 Mileo, A. M. et al. The human papillomavirus-16 E7 oncoprotein exerts antiapoptotic effects via its physical interaction with the actin-binding protein gelsolin. Carcinogenesis 34, 2424-2433, doi:10.1093/carcin/bgt192 (2013).

154

119 Liao, S. et al. Human papillomavirus 16/18 E5 promotes cervical cancer cell proliferation, migration and invasion in vitro and accelerates tumor growth in vivo. Oncology reports 29, 95- 102, doi:10.3892/or.2012.2106 (2013). 120 Leptak, C. et al. Tumorigenic transformation of murine keratinocytes by the E5 genes of bovine papillomavirus type 1 and human papillomavirus type 16. J Virol 65, 7078-7083 (1991). 121 Leechanachai, P., Banks, L., Moreau, F. & Matlashewski, G. The E5 gene from human papillomavirus type 16 is an oncogene which enhances growth factor-mediated signal transduction to the nucleus. Oncogene 7, 19-25 (1992). 122 Stoppler, M. C., Straight, S. W., Tsao, G., Schlegel, R. & McCance, D. J. The E5 gene of HPV-16 enhances keratinocyte immortalization by full-length DNA. Virology 223, 251-254, doi:10.1006/viro.1996.0475 (1996). 123 Barbaresi, S. et al. Effects of human papillomavirus type 16 E5 deletion mutants on epithelial morphology: functional characterization of each transmembrane domain. The Journal of general virology 91, 521-530, doi:10.1099/vir.0.016295-0 (2010). 124 Kivi, N., Greco, D., Auvinen, P. & Auvinen, E. Genes involved in cell adhesion, cell motility and mitogenic signaling are altered due to HPV 16 E5 protein expression. Oncogene 27, 2532-2541, doi:10.1038/sj.onc.1210916 (2008). 125 Pim, D., Collins, M. & Banks, L. Human papillomavirus type 16 E5 gene stimulates the transforming activity of the epidermal growth factor receptor. Oncogene 7, 27-32 (1992). 126 Conrad, M., Bubb, V. J. & Schlegel, R. The human papillomavirus type 6 and 16 E5 proteins are membrane-associated proteins which associate with the 16-kilodalton pore-forming protein. J Virol 67, 6170-6178 (1993). 127 Conrad, M., Goldstein, D., Andresson, T. & Schlegel, R. The E5 protein of HPV-6, but not HPV-16, associates efficiently with cellular growth factor receptors. Virology 200, 796-800, doi:10.1006/viro.1994.1244 (1994). 128 Hwang, E. S., Nottoli, T. & Dimaio, D. The HPV16 E5 protein: expression, detection, and stable complex formation with transmembrane proteins in COS cells. Virology 211, 227-233, doi:10.1006/viro.1995.1395 (1995). 129 Zhang, B., Srirangam, A., Potter, D. A. & Roman, A. HPV16 E5 protein disrupts the c-Cbl-EGFR interaction and EGFR ubiquitination in human foreskin keratinocytes. Oncogene 24, 2585-2588, doi:10.1038/sj.onc.1208453 (2005). 130 Straight, S. W., Herman, B. & McCance, D. J. The E5 oncoprotein of human papillomavirus type 16 inhibits the acidification of endosomes in human keratinocytes. J Virol 69, 3185-3192 (1995). 131 Thomsen, P., van Deurs, B., Norrild, B. & Kayser, L. The HPV16 E5 oncogene inhibits endocytic trafficking. Oncogene 19, 6023-6032, doi:10.1038/sj.onc.1204010 (2000). 132 Suprynowicz, F. A. et al. The human papillomavirus type 16 E5 oncoprotein inhibits epidermal growth factor trafficking independently of endosome acidification. J Virol 84, 10619-10629, doi:10.1128/JVI.00831-10 (2010). 133 Wetherill, L. F. et al. High-risk human papillomavirus E5 oncoprotein displays channel-forming activity sensitive to small-molecule inhibitors. J Virol 86, 5341-5351, doi:10.1128/JVI.06243-11 (2012). 134 Di Domenico, F. et al. Expression of human papilloma virus type 16 E5 protein in amelanotic melanoma cells regulates endo-cellular pH and restores tyrosinase activity. Journal of experimental & clinical cancer research : CR 28, 4, doi:10.1186/1756-9966-28-4 (2009). 135 Kim, S. H. et al. Human papillomavirus 16 E5 up-regulates the expression of vascular endothelial growth factor through the activation of epidermal growth factor receptor, MEK/ ERK1,2 and PI3K/Akt. Cellular and molecular life sciences : CMLS 63, 930-938, doi:10.1007/s00018-005-5561- x (2006).

155

136 Kim, S. H. et al. Involvement of NF-kappaB and AP-1 in COX-2 upregulation by human papillomavirus 16 E5 oncoprotein. Carcinogenesis 30, 753-757, doi:10.1093/carcin/bgp066 (2009). 137 Ricciotti, E. & FitzGerald, G. A. Prostaglandins and inflammation. Arteriosclerosis, thrombosis, and vascular biology 31, 986-1000, doi:10.1161/ATVBAHA.110.207449 (2011). 138 Oh, J. M. et al. Human papillomavirus E5 protein induces expression of the EP4 subtype of prostaglandin E2 receptor in cyclic AMP response element-dependent pathways in cervical cancer cells. Carcinogenesis 30, 141-149, doi:10.1093/carcin/bgn236 (2009). 139 Goel, H. L. & Mercurio, A. M. VEGF targets the tumour cell. Nature reviews. Cancer 13, 871-882, doi:10.1038/nrc3627 (2013). 140 Tsao, Y. P., Li, L. Y., Tsai, T. C. & Chen, S. L. Human papillomavirus type 11 and 16 E5 represses p21(WafI/SdiI/CipI) gene expression in fibroblasts and keratinocytes. J Virol 70, 7535-7539 (1996). 141 Pedroza-Saavedra, A., Lam, E. W., Esquivel-Guadarrama, F. & Gutierrez-Xicotencatl, L. The human papillomavirus type 16 E5 oncoprotein synergizes with EGF-receptor signaling to enhance cell cycle progression and the down-regulation of p27(Kip1). Virology 400, 44-52, doi:10.1016/j.virol.2010.01.009 (2010). 142 Liao, S. et al. HPV16/18 E5, a promising candidate for cervical cancer vaccines, affects SCPs, cell proliferation and cell cycle, and forms a potential network with E6 and E7. International journal of molecular medicine 31, 120-128, doi:10.3892/ijmm.2012.1168 (2013). 143 Oh, J. M. et al. Human papillomavirus type 16 E5 protein inhibits hydrogen-peroxide-induced apoptosis by stimulating ubiquitin-proteasome-mediated degradation of Bax in human cervical cancer cells. Carcinogenesis 31, 402-410, doi:10.1093/carcin/bgp318 (2010). 144 Crusius, K., Kaszkin, M., Kinzel, V. & Alonso, A. The human papillomavirus type 16 E5 protein modulates phospholipase C-gamma-1 activity and phosphatidyl inositol turnover in mouse fibroblasts. Oncogene 18, 6714-6718, doi:10.1038/sj.onc.1203075 (1999). 145 Crusius, K., Rodriguez, I. & Alonso, A. The human papillomavirus type 16 E5 protein modulates ERK1/2 and p38 MAP kinase activation by an EGFR-independent process in stressed human keratinocytes. Virus genes 20, 65-69 (2000). 146 Rossjohn, J., Pellicci, D. G., Patel, O., Gapin, L. & Godfrey, D. I. Recognition of CD1d-restricted antigens by natural killer T cells. Nature reviews. Immunology 12, 845-857, doi:10.1038/nri3328 (2012). 147 Miura, S. et al. CD1d, a sentinel molecule bridging innate and adaptive immunity, is downregulated by the human papillomavirus (HPV) E5 protein: a possible mechanism for immune evasion by HPV. J Virol 84, 11614-11623, doi:10.1128/JVI.01053-10 (2010). 148 Gao, G. F. & Jakobsen, B. K. Molecular interactions of coreceptor CD8 and MHC class I: the molecular basis for functional coordination with the T-cell receptor. Immunology today 21, 630- 636 (2000). 149 Paulsson, K. & Wang, P. Chaperones and folding of MHC class I molecules in the endoplasmic reticulum. Biochimica et biophysica acta 1641, 1-12 (2003). 150 Regan, J. A. & Laimins, L. A. Bap31 is a novel target of the human papillomavirus E5 protein. J Virol 82, 10042-10051, doi:10.1128/JVI.01240-08 (2008). 151 Gruener, M., Bravo, I. G., Momburg, F., Alonso, A. & Tomakidi, P. The E5 protein of the human papillomavirus type 16 down-regulates HLA-I surface expression in calnexin-expressing but not in calnexin-deficient cells. Virology journal 4, 116, doi:10.1186/1743-422X-4-116 (2007). 152 Campo, M. S. et al. HPV-16 E5 down-regulates expression of surface HLA class I and reduces recognition by CD8 T cells. Virology 407, 137-142, doi:10.1016/j.virol.2010.07.044 (2010).

156

153 DiMaio, D. & Petti, L. M. The E5 proteins. Virology 445, 99-114, doi:10.1016/j.virol.2013.05.006 (2013). 154 Bernard, B. A. et al. The human papillomavirus type 18 (HPV18) E2 gene product is a repressor of the HPV18 regulatory region in human keratinocytes. J Virol 63, 4317-4324 (1989). 155 Goodwin, E. C. & DiMaio, D. Repression of human papillomavirus oncogenes in HeLa cervical carcinoma cells causes the orderly reactivation of dormant tumor suppressor pathways. Proc Natl Acad Sci U S A 97, 12513-12518, doi:10.1073/pnas.97.23.12513 (2000). 156 Xue, Y. et al. Loss of HPV16 E2 Protein Expression Without Disruption of the E2 ORF Correlates with Carcinogenic Progression. The open virology journal 6, 163-172, doi:10.2174/1874357901206010163 (2012). 157 Baker, C. C. et al. Structural and transcriptional analysis of human papillomavirus type 16 sequences in cervical carcinoma cell lines. J Virol 61, 962-971 (1987). 158 Jeon, S., Allen-Hoffmann, B. L. & Lambert, P. F. Integration of human papillomavirus type 16 into the human genome correlates with a selective growth advantage of cells. J Virol 69, 2989-2997 (1995). 159 Dowhanick, J. J., Mcbride, A. A. & Howley, P. M. Suppression of Cellular Proliferation by the Papillomavirus E2 Protein. Journal of Virology 69, 7791-7799 (1995). 160 Desaintes, C., Demeret, C., Goyat, S., Yaniv, M. & Thierry, F. Expression of the papillomavirus E2 protein in HeLa cells leads to apoptosis. EMBO J 16, 504-514, doi:10.1093/emboj/16.3.504 (1997). 161 Desaintes, C., Goyat, S., Garbay, S., Yaniv, M. & Thierry, F. Papillomavirus E2 induces p53- independent apoptosis in HeLa cells. Oncogene 18, 4538-4545, doi:10.1038/sj.onc.1202818 (1999). 162 Wells, S. I. et al. Papillomavirus E2 induces senescence in HPV-positive cells via pRB- and p21(CIP)-dependent pathways. EMBO J 19, 5762-5771, doi:10.1093/emboj/19.21.5762 (2000). 163 Thierry, F. & Yaniv, M. The BPV1-E2 trans-acting protein can be either an activator or a repressor of the HPV18 regulatory region. EMBO J 6, 3391-3397 (1987). 164 Vinokurova, S. et al. Type-dependent integration frequency of human papillomavirus genomes in cervical lesions. Cancer Res 68, 307-313, doi:10.1158/0008-5472.CAN-07-2754 (2008). 165 Akagi, K. et al. Genome-wide analysis of HPV integration in human cancers reveals recurrent, focal genomic instability. Genome research 24, 185-199, doi:10.1101/gr.164806.113 (2014). 166 Nishimura, A. et al. Mechanisms of human papillomavirus E2-mediated repression of viral oncogene expression and cervical cancer cell growth inhibition. J Virol 74, 3752-3760 (2000). 167 Kanno, T. et al. Selective recognition of acetylated histones by bromodomain proteins visualized in living cells. Molecular cell 13, 33-43 (2004). 168 You, J., Croyle, J. L., Nishimura, A., Ozato, K. & Howley, P. M. Interaction of the bovine papillomavirus E2 protein with Brd4 tethers the viral DNA to host mitotic chromosomes. Cell 117, 349-360 (2004). 169 Wu, S. Y. et al. Brd4 links chromatin targeting to HPV transcriptional silencing. Genes & development 20, 2383-2396, doi:10.1101/gad.1448206 (2006). 170 Schweiger, M. R., Ottinger, M., You, J. & Howley, P. M. Brd4-independent transcriptional repression function of the papillomavirus e2 proteins. J Virol 81, 9612-9622, doi:10.1128/JVI.00447-07 (2007). 171 Lee, A. Y. & Chiang, C. M. Chromatin adaptor Brd4 modulates E2 transcription activity and protein stability. The Journal of biological chemistry 284, 2778-2786, doi:10.1074/jbc.M805835200 (2009).

157

172 Smith, J. A. et al. Genome-wide siRNA screen identifies SMCX, EP400, and Brd4 as E2-dependent regulators of human papillomavirus oncogene expression. Proc Natl Acad Sci U S A 107, 3752- 3757, doi:10.1073/pnas.0914818107 (2010). 173 Asumalahti, K. et al. Coding haplotype analysis supports HCR as the putative susceptibility gene for psoriasis at the MHC PSORS1 locus. Human molecular genetics 11, 589-597 (2002). 174 Tiala, I. et al. The PSORS1 locus gene CCHCR1 affects keratinocyte proliferation in transgenic mice. Human molecular genetics 17, 1043-1051, doi:10.1093/hmg/ddm377 (2008). 175 Muller, M. & Demeret, C. CCHCR1 Interacts Specifically with the E2 Protein of Human Papillomavirus Type 16 on a Surface Overlapping BRD4 Binding. PLoS One 9, e92581, doi:10.1371/journal.pone.0092581 (2014). 176 Abbate, E. A., Voitenleitner, C. & Botchan, M. R. Structure of the papillomavirus DNA-tethering complex E2:Brd4 and a peptide that ablates HPV chromosomal association. Molecular cell 24, 877-889, doi:10.1016/j.molcel.2006.11.002 (2006). 177 Lai, D. et al. Localization of HPV-18 E2 at mitochondrial membranes induces ROS release and modulates host cell metabolism. PLoS One 8, e75625, doi:10.1371/journal.pone.0075625 (2013). 178 Wise-Draper, T. M. et al. Future directions and treatment strategies for head and neck squamous cell carcinomas. Transl Res 160, 167-177, doi:10.1016/j.trsl.2012.02.002 (2012). 179 Kaur, P. & McDougall, J. K. Characterization of primary human keratinocytes transformed by human papillomavirus type 18. J Virol 62, 1917-1924 (1988). 180 Hanahan, D. & Weinberg, Robert A. Hallmarks of Cancer: The Next Generation. Cell 144, 646- 674, doi:10.1016/j.cell.2011.02.013 (2011). 181 Halbert, C. L., Demers, G. W. & Galloway, D. A. The E7 gene of human papillomavirus type 16 is sufficient for immortalization of human epithelial cells. J Virol 65, 473-478 (1991). 182 Hawley-Nelson, P., Vousden, K. H., Hubbert, N. L., Lowy, D. R. & Schiller, J. T. HPV16 E6 and E7 proteins cooperate to immortalize human foreskin keratinocytes. EMBO J 8, 3905-3910 (1989). 183 Munger, K., Phelps, W. C., Bubb, V., Howley, P. M. & Schlegel, R. The E6 and E7 genes of the human papillomavirus type 16 together are necessary and sufficient for transformation of primary human keratinocytes. J Virol 63, 4417-4421 (1989). 184 Boulenouar, S. et al. Effects of HPV-16 E5, E6 and E7 proteins on survival, adhesion, migration and invasion of trophoblastic cells. Carcinogenesis 31, 473-480, doi:10.1093/carcin/bgp281 (2010). 185 Riley, R. R. et al. Dissection of human papillomavirus E6 and E7 function in transgenic mouse models of cervical carcinogenesis. Cancer Res 63, 4862-4871 (2003). 186 Arbeit, J. M., Howley, P. M. & Hanahan, D. Chronic estrogen-induced cervical and vaginal squamous carcinogenesis in human papillomavirus type 16 transgenic mice. Proc Natl Acad Sci U S A 93, 2930-2935 (1996). 187 Brake, T. & Lambert, P. F. Estrogen contributes to the onset, persistence, and malignant progression of cervical cancer in a human papillomavirus-transgenic mouse model. Proc Natl Acad Sci U S A 102, 2490-2495, doi:10.1073/pnas.0409883102 (2005). 188 Genther Williams, S. M. et al. Requirement of epidermal growth factor receptor for hyperplasia induced by E5, a high-risk human papillomavirus oncogene. Cancer Res 65, 6534-6542, doi:10.1158/0008-5472.CAN-05-0083 (2005). 189 Lee, S. S., Glaunsinger, B., Mantovani, F., Banks, L. & Javier, R. T. Multi-PDZ domain protein MUPP1 is a cellular target for both adenovirus E4-ORF1 and high-risk papillomavirus type 18 E6 oncoproteins. J Virol 74, 9680-9693 (2000). 190 Pim, D., Thomas, M., Javier, R., Gardiol, D. & Banks, L. HPV E6 targeted degradation of the discs large protein: evidence for the involvement of a novel ubiquitin ligase. Oncogene 19, 719-725, doi:10.1038/sj.onc.1203374 (2000).

158

191 Massimi, P., Shai, A., Lambert, P. & Banks, L. HPV E6 degradation of p53 and PDZ containing substrates in an E6AP null background. Oncogene 27, 1800-1804, doi:10.1038/sj.onc.1210810 (2008). 192 Spanos, W. C. et al. The PDZ binding motif of human papillomavirus type 16 E6 induces PTPN13 loss, which allows anchorage-independent growth and synergizes with ras for invasive growth. J Virol 82, 2493-2500, doi:10.1128/JVI.02188-07 (2008). 193 Watson, R. A., Thomas, M., Banks, L. & Roberts, S. Activity of the human papillomavirus E6 PDZ- binding motif correlates with an enhanced morphological transformation of immortalized human keratinocytes. Journal of cell science 116, 4925-4934, doi:10.1242/jcs.00809 (2003). 194 Rey, O., Lee, S. & Park, N. H. Human papillomavirus type 16 E7 oncoprotein represses transcription of human fibronectin. J Virol 74, 4912-4918 (2000). 195 Chen, S. L., Huang, C. H., Tsai, T. C., Lu, K. Y. & Tsao, Y. P. The regulation mechanism of c-jun and junB by human papillomavirus type 16 E5 oncoprotein. Archives of virology 141, 791-800 (1996). 196 Matsukura, T., Koi, S. & Sugase, M. Both episomal and integrated forms of human papillomavirus type 16 are involved in invasive cervical cancers. Virology 172, 63-72 (1989). 197 Schwarz, E. et al. Structure and transcription of human papillomavirus sequences in cervical carcinoma cells. Nature 314, 111-114 (1985). 198 Shirasawa, H., Tomita, Y., Sekiya, S., Takamizawa, H. & Simizu, B. Integration and transcription of human papillomavirus type 16 and 18 sequences in cell lines derived from cervical carcinomas. The Journal of general virology 68 ( Pt 2), 583-591 (1987). 199 Ashrafi, G. H., Haghshenas, M., Marchetti, B. & Campo, M. S. E5 protein of human papillomavirus 16 downregulates HLA class I and interacts with the heavy chain via its first hydrophobic domain. International journal of cancer. Journal international du cancer 119, 2105- 2112, doi:10.1002/ijc.22089 (2006). 200 Huang, S. M. & McCance, D. J. Down regulation of the interleukin-8 promoter by human papillomavirus type 16 E6 and E7 through effects on CREB binding protein/p300 and P/CAF. J Virol 76, 8710-8721 (2002). 201 Cho, Y. S. et al. Down modulation of IL-18 expression by human papillomavirus type 16 E6 oncogene via binding to IL-18. FEBS letters 501, 139-145 (2001). 202 Richards, K. H. et al. Human papillomavirus e7 oncoprotein increases production of the anti- inflammatory interleukin-18 binding protein in keratinocytes. J Virol 88, 4173-4179, doi:10.1128/JVI.02546-13 (2014). 203 Ronco, L. V., Karpova, A. Y., Vidal, M. & Howley, P. M. Human papillomavirus 16 E6 oncoprotein binds to interferon regulatory factor-3 and inhibits its transcriptional activity. Genes & development 12, 2061-2072 (1998). 204 Hebner, C. M., Wilson, R., Rader, J., Bidder, M. & Laimins, L. A. Human papillomaviruses target the double-stranded RNA protein kinase pathway. The Journal of general virology 87, 3183- 3193, doi:10.1099/vir.0.82098-0 (2006). 205 Kleine-Lowinski, K. et al. Selective suppression of monocyte chemoattractant protein-1 expression by human papillomavirus E6 and E7 oncoproteins in human cervical epithelial and epidermal cells. International journal of cancer. Journal international du cancer 107, 407-415, doi:10.1002/ijc.11411 (2003). 206 Niebler, M. et al. Post-translational control of IL-1beta via the human papillomavirus type 16 E6 oncoprotein: a novel mechanism of innate immune escape mediated by the E3-ubiquitin ligase E6-AP and p53. PLoS pathogens 9, e1003536, doi:10.1371/journal.ppat.1003536 (2013). 207 Antonsson, A., Payne, E., Hengst, K. & McMillan, N. A. The human papillomavirus type 16 E7 protein binds human interferon regulatory factor-9 via a novel PEST domain required for transformation. Journal of interferon & cytokine research : the official journal of the

159

International Society for Interferon and Cytokine Research 26, 455-461, doi:10.1089/jir.2006.26.455 (2006). 208 Wallace, N. A. & Galloway, D. A. Manipulation of cellular DNA damage repair machinery facilitates propagation of human papillomaviruses. Seminars in cancer biology, doi:10.1016/j.semcancer.2013.12.003 (2014). 209 Smith, J., Tho, L. M., Xu, N. & Gillespie, D. A. The ATM-Chk2 and ATR-Chk1 pathways in DNA damage signaling and cancer. Advances in cancer research 108, 73-112, doi:10.1016/B978-0-12- 380888-2.00003-0 (2010). 210 Cheng, S., Schmidt-Grimminger, D. C., Murant, T., Broker, T. R. & Chow, L. T. Differentiation- dependent up-regulation of the human papillomavirus E7 gene reactivates cellular DNA replication in suprabasal differentiated keratinocytes. Genes & development 9, 2335-2349 (1995). 211 Korzeniewski, N., Spardy, N., Duensing, A. & Duensing, S. Genomic instability and cancer: lessons learned from human papillomaviruses. Cancer Lett 305, 113-122, doi:10.1016/j.canlet.2010.10.013 (2011). 212 Kavanaugh, G. M. et al. The human DEK oncogene regulates DNA damage response signaling and repair. Nucleic Acids Res 39, 7465-7476, doi:10.1093/nar/gkr454 (2011). 213 Yim, E. K. et al. Novel interaction between HPV E6 and BARD1 (BRCA1-associated ring domain 1) and its biologic roles. DNA and cell biology 26, 753-761, doi:10.1089/dna.2007.0588 (2007). 214 Romick-Rosendale, L. E., Lui, V. W., Grandis, J. R. & Wells, S. I. The Fanconi anemia pathway: repairing the link between DNA damage and squamous cell carcinoma. Mutation research 743- 744, 78-88, doi:10.1016/j.mrfmmm.2013.01.001 (2013). 215 Kottemann, M. C. & Smogorzewska, A. Fanconi anaemia and the repair of Watson and Crick DNA crosslinks. Nature 493, 356-363, doi:10.1038/nature11863 (2013). 216 Kutler, D. I. et al. Human papillomavirus DNA and p53 polymorphisms in squamous cell carcinomas from Fanconi anemia patients. Journal of the National Cancer Institute 95, 1718- 1721 (2003). 217 van Zeeburg, H. J. et al. Generation and molecular characterization of head and neck squamous cell lines of fanconi anemia patients. Cancer Res 65, 1271-1276, doi:10.1158/0008-5472.CAN-04- 3665 (2005). 218 van Zeeburg, H. J. et al. Clinical and molecular characteristics of squamous cell carcinomas from Fanconi anemia patients. Journal of the National Cancer Institute 100, 1649-1653, doi:10.1093/jnci/djn366 (2008). 219 Alter, B. P. et al. Squamous cell carcinomas in patients with Fanconi anemia and dyskeratosis congenita: a search for human papillomavirus. International journal of cancer. Journal international du cancer 133, 1513-1515, doi:10.1002/ijc.28157 (2013). 220 Park, J. W., Shin, M. K. & Lambert, P. F. High incidence of female reproductive tract cancers in FA-deficient HPV16-transgenic mice correlates with E7's induction of DNA damage response, an activity mediated by E7's inactivation of pocket proteins. Oncogene, doi:10.1038/onc.2013.327 (2013). 221 Spardy, N. et al. The human papillomavirus type 16 E7 oncoprotein activates the Fanconi anemia (FA) pathway and causes accelerated chromosomal instability in FA cells. J Virol 81, 13265- 13270, doi:10.1128/JVI.01121-07 (2007). 222 Hoskins, E. E. et al. Fanconi anemia deficiency stimulates HPV-associated hyperplastic growth in organotypic epithelial raft culture. Oncogene 28, 674-685, doi:10.1038/onc.2008.416 (2009). 223 Andreassen, P. R., D'Andrea, A. D. & Taniguchi, T. ATR couples FANCD2 monoubiquitination to the DNA-damage response. Genes & development 18, 1958-1963, doi:10.1101/gad.1196104 (2004).

160

224 D'Andrea, A. D. & Grompe, M. The Fanconi anaemia/BRCA pathway. Nature reviews. Cancer 3, 23-34, doi:10.1038/nrc970 (2003). 225 Hovest, M. G., Krieg, T. & Herrmann, G. Differential roles for Chk1 and FANCD2 in ATR-mediated signalling for psoralen photoactivation-induced senescence. Experimental dermatology 20, 883- 889, doi:10.1111/j.1600-0625.2011.01365.x (2011). 226 Hoskins, E. E. et al. The fanconi anemia pathway limits human papillomavirus replication. J Virol 86, 8131-8138, doi:10.1128/JVI.00408-12 (2012). 227 Rieckmann, T. et al. HNSCC cell lines positive for HPV and p16 possess higher cellular radiosensitivity due to an impaired DSB repair capacity. Radiother Oncol, doi:10.1016/j.radonc.2013.03.013 (2013). 228 Kimple, R. J. et al. Enhanced radiation sensitivity in HPV-positive head and neck cancer. Cancer Res 73, 4791-4800, doi:10.1158/0008-5472.CAN-13-0587 (2013). 229 Chaurushiya, M. S. & Weitzman, M. D. Viral manipulation of DNA repair and cell cycle checkpoints. DNA repair 8, 1166-1176, doi:10.1016/j.dnarep.2009.04.016 (2009). 230 Adey, A. et al. The haplotype-resolved genome and epigenome of the aneuploid HeLa cancer cell line. Nature 500, 207-211, doi:10.1038/nature12064 (2013). 231 zur Hausen, H. Immortalization of human cells and their malignant conversion by high risk human papillomavirus genotypes. Seminars in cancer biology 9, 405-411, doi:10.1006/scbi.1999.0144 (1999). 232 Schiffman, M., Castle, P. E., Jeronimo, J., Rodriguez, A. C. & Wacholder, S. Human papillomavirus and cervical cancer. Lancet 370, 890-907, doi:10.1016/S0140-6736(07)61416-0 (2007). 233 Winkler, B. et al. Koilocytotic lesions of the cervix. The relationship of mitotic abnormalities to the presence of papillomavirus antigens and nuclear DNA content. Cancer 53, 1081-1087 (1984). 234 Duensing, S. et al. The human papillomavirus type 16 E6 and E7 oncoproteins cooperate to induce mitotic defects and genomic instability by uncoupling centrosome duplication from the cell division cycle. Proc Natl Acad Sci U S A 97, 10002-10007, doi:10.1073/pnas.170093297 (2000). 235 Guarguaglini, G. et al. The forkhead-associated domain protein Cep170 interacts with Polo-like kinase 1 and serves as a marker for mature centrioles. Molecular biology of the cell 16, 1095- 1107, doi:10.1091/mbc.E04-10-0939 (2005). 236 Duensing, S., Duensing, A., Crum, C. P. & Munger, K. Human papillomavirus type 16 E7 oncoprotein-induced abnormal centrosome synthesis is an early event in the evolving malignant phenotype. Cancer Res 61, 2356-2360 (2001). 237 Duensing, A. et al. Centriole overduplication through the concurrent formation of multiple daughter centrioles at single maternal templates. Oncogene 26, 6280-6288, doi:10.1038/sj.onc.1210456 (2007). 238 Duensing, S. & Munger, K. Human papillomavirus type 16 E7 oncoprotein can induce abnormal centrosome duplication through a mechanism independent of inactivation of retinoblastoma protein family members. Journal of virology 77, 12331-12335 (2003). 239 Korzeniewski, N., Treat, B. & Duensing, S. The HPV-16 E7 oncoprotein induces centriole multiplication through deregulation of Polo-like kinase 4 expression. Molecular cancer 10, 61, doi:10.1186/1476-4598-10-61 (2011). 240 Sadasivam, S. & DeCaprio, J. A. The DREAM complex: master coordinator of cell cycle-dependent gene expression. Nature reviews. Cancer 13, 585-595, doi:10.1038/nrc3556 (2013). 241 Fischer, M., Quaas, M., Wintsche, A., Muller, G. A. & Engeland, K. Polo-like kinase 4 transcription is activated via CRE and NRF1 elements, repressed by DREAM through CDE/CHR sites and deregulated by HPV E7 protein. Nucleic acids research 42, 163-180, doi:10.1093/nar/gkt849 (2014).

161

242 Nguyen, C. L., Eichwald, C., Nibert, M. L. & Munger, K. Human papillomavirus type 16 E7 oncoprotein associates with the centrosomal component gamma-tubulin. Journal of virology 81, 13533-13543, doi:10.1128/JVI.01669-07 (2007). 243 Loncarek, J., Hergert, P., Magidson, V. & Khodjakov, A. Control of daughter centriole formation by the pericentriolar material. Nature cell biology 10, 322-328, doi:10.1038/ncb1694 (2008). 244 Thompson, D. A. et al. The human papillomavirus-16 E6 oncoprotein decreases the vigilance of mitotic checkpoints. Oncogene 15, 3025-3035, doi:10.1038/sj.onc.1201495 (1997). 245 Thomas, J. T. & Laimins, L. A. Human papillomavirus oncoproteins E6 and E7 independently abrogate the mitotic spindle checkpoint. Journal of virology 72, 1131-1137 (1998). 246 Schvartzman, J. M., Sotillo, R. & Benezra, R. Mitotic chromosomal instability and cancer: mouse modelling of the human disease. Nature reviews. Cancer 10, 102-115, doi:10.1038/nrc2781 (2010). 247 Sluder, G., Thompson, E. A., Miller, F. J., Hayes, J. & Rieder, C. L. The checkpoint control for anaphase onset does not monitor excess numbers of spindle poles or bipolar spindle symmetry. Journal of cell science 110 ( Pt 4), 421-429 (1997). 248 Vaziri, C. et al. A p53-dependent checkpoint pathway prevents rereplication. Molecular cell 11, 997-1008 (2003). 249 Machida, Y. J. & Dutta, A. The APC/C inhibitor, Emi1, is essential for prevention of rereplication. Genes & development 21, 184-194, doi:10.1101/gad.1495007 (2007). 250 Mihaylov, I. S. et al. Control of DNA replication and chromosome ploidy by geminin and cyclin A. Molecular and cellular biology 22, 1868-1880 (2002). 251 Fan, X., Liu, Y., Heilman, S. A. & Chen, J. J. Human papillomavirus E7 induces rereplication in response to DNA damage. J Virol 87, 1200-1210, doi:10.1128/JVI.02038-12 (2013). 252 Patel, D. & McCance, D. J. Compromised spindle assembly checkpoint due to altered expression of Ubch10 and Cdc20 in human papillomavirus type 16 E6- and E7-expressing keratinocytes. J Virol 84, 10956-10964, doi:10.1128/JVI.00259-10 (2010). 253 Schmitz, M., Driesch, C., Jansen, L., Runnebaum, I. B. & Durst, M. Non-random integration of the HPV genome in cervical cancer. PLoS One 7, e39632, doi:10.1371/journal.pone.0039632 (2012). 254 Cheng, L. et al. Aurora B regulates formin mDia3 in achieving metaphase chromosome alignment. Developmental cell 20, 342-352, doi:10.1016/j.devcel.2011.01.008 (2011). 255 Espinosa, A. M. et al. Mitosis is a source of potential markers for screening and survival and therapeutic targets in cervical cancer. PloS one 8, e55975, doi:10.1371/journal.pone.0055975 (2013). 256 Stransky, N. et al. The mutational landscape of head and neck squamous cell carcinoma. Science 333, 1157-1160, doi:10.1126/science.1208130 (2011). 257 Agrawal, N. et al. Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1. Science 333, 1154-1157, doi:10.1126/science.1206923 (2011). 258 Helfer, C. M., Wang, R. & You, J. Analysis of the papillomavirus E2 and bromodomain protein Brd4 interaction using bimolecular fluorescence complementation. PLoS One 8, e77994, doi:10.1371/journal.pone.0077994 (2013). 259 Suprynowicz, F. A. et al. HPV-16 E5 oncoprotein upregulates lipid raft components caveolin-1 and ganglioside GM1 at the plasma membrane of cervical cells. Oncogene 27, 1071-1078, doi:10.1038/sj.onc.1210725 (2008). 260 Riaz, N., Sherman, E. J., Fury, M. & Lee, N. Should cetuximab replace Cisplatin for definitive chemoradiotherapy in locally advanced head and neck cancer? Journal of clinical oncology : official journal of the American Society of Clinical Oncology 31, 287-288, doi:10.1200/JCO.2012.46.9049 (2013).

162

261 Garnett, T. O. & Duerksen-Hughes, P. J. Modulation of apoptosis by human papillomavirus (HPV) oncoproteins. Archives of virology 151, 2321-2335, doi:10.1007/s00705-006-0821-0 (2006). 262 Du, M., Fan, X., Hong, E. & Chen, J. J. Interaction of oncogenic papillomavirus E6 proteins with fibulin-1. Biochemical and biophysical research communications 296, 962-969 (2002). 263 Liu, Y., Heilman, S. A., Illanes, D., Sluder, G. & Chen, J. J. p53-independent abrogation of a postmitotic checkpoint contributes to human papillomavirus E6-induced polyploidy. Cancer Res 67, 2603-2610, doi:10.1158/0008-5472.CAN-06-3436 (2007). 264 Boon, S. S. & Banks, L. High-risk human papillomavirus E6 oncoproteins interact with 14-3-3zeta in a PDZ binding motif-dependent manner. J Virol 87, 1586-1595, doi:10.1128/JVI.02074-12 (2013). 265 Jones, D. L., Alani, R. M. & Munger, K. The human papillomavirus E7 oncoprotein can uncouple cellular differentiation and proliferation in human keratinocytes by abrogating p21Cip1- mediated inhibition of cdk2. Genes & development 11, 2101-2111 (1997). 266 Huh, K. W. et al. Association of the human papillomavirus type 16 E7 oncoprotein with the 600- kDa retinoblastoma protein-associated factor, p600. Proc Natl Acad Sci U S A 102, 11492-11497, doi:10.1073/pnas.0505337102 (2005). 267 McKenna, D. J., Patel, D. & McCance, D. J. miR-24 and miR-205 expression is dependent on HPV onco-protein expression in keratinocytes. Virology 448, 210-216, doi:10.1016/j.virol.2013.10.014 (2014). 268 Bernat, A., Avvakumov, N., Mymryk, J. S. & Banks, L. Interaction between the HPV E7 oncoprotein and the transcriptional coactivator p300. Oncogene 22, 7871-7881, doi:10.1038/sj.onc.1206896 (2003). 269 Byg, L. M. et al. NF-kappaB signalling is attenuated by the E7 protein from cutaneous human papillomaviruses. Virus research 169, 48-53, doi:10.1016/j.virusres.2012.06.028 (2012). 270 Yaginuma, Y., Eguchi, A., Yoshimoto, M. & Ogawa, K. The PxDLLCxE sequence in conserved region 2 of human papilloma virus 18 protein E7 is required for E7 binding to centromere protein C. Oncology 83, 210-217, doi:10.1159/000341360 (2012). 271 Heilman, S. A., Nordberg, J. J., Liu, Y., Sluder, G. & Chen, J. J. Abrogation of the postmitotic checkpoint contributes to polyploidization in human papillomavirus E7-expressing cells. J Virol 83, 2756-2764, doi:10.1128/JVI.02149-08 (2009). 272 Szalmas, A. et al. Activation of Src, Fyn and Yes non-receptor tyrosine kinases in keratinocytes expressing human papillomavirus (HPV) type 16 E7 oncoprotein. Virology journal 10, 79, doi:10.1186/1743-422X-10-79 (2013). 273 Chen, S. L., Lin, S. T., Tsai, T. C., Hsiao, W. C. & Tsao, Y. P. ErbB4 (JM-b/CYT-1)-induced expression and phosphorylation of c-Jun is abrogated by human papillomavirus type 16 E5 protein. Oncogene 26, 42-53, doi:10.1038/sj.onc.1209768 (2007). 274 French, D. et al. Expression of HPV16 E5 down-modulates the TGFbeta signaling pathway. Molecular cancer 12, 38, doi:10.1186/1476-4598-12-38 (2013). 275 Tomakidi, P., Cheng, H., Kohl, A., Komposch, G. & Alonso, A. Connexin 43 expression is downregulated in raft cultures of human keratinocytes expressing the human papillomavirus type 16 E5 protein. Cell and tissue research 301, 323-327 (2000). 276 Sunthamala, N. et al. E2 proteins of high risk human papillomaviruses down-modulate STING and IFN-kappa transcription in keratinocytes. PLoS One 9, e91473, doi:10.1371/journal.pone.0091473 (2014). 277 von Lindern, M. et al. The translocation (6;9), associated with a specific subtype of acute myeloid leukemia, results in the fusion of two genes, dek and can, and the expression of a chimeric, leukemia-specific dek-can mRNA. Mol Cell Biol 12, 1687-1697 (1992).

163

278 Ying, G. & Wu, Y. DEK: A novel early screening and prognostic marker for breast cancer. Mol Med Rep 12, 7491-7495, doi:10.3892/mmr.2015.4380 (2015). 279 Martinez-Useros, J. et al. DEK is a potential marker for aggressive phenotype and irinotecan- based therapy response in metastatic colorectal cancer. BMC cancer 14, 965, doi:10.1186/1471- 2407-14-965 (2014). 280 Privette Vinnedge, L. M., Kappes, F., Nassar, N. & Wells, S. I. Stacking the DEK: from chromatin topology to cancer stem cells. Cell Cycle 12, 51-66, doi:10.4161/cc.23121 (2013). 281 Kappes, F., Burger, K., Baack, M., Fackelmayer, F. O. & Gruss, C. Subcellular localization of the human proto-oncogene protein DEK. The Journal of biological chemistry 276, 26317-26323, doi:10.1074/jbc.M100162200 (2001). 282 Alexiadis, V. et al. The protein encoded by the proto-oncogene DEK changes the topology of chromatin and reduces the efficiency of DNA replication in a chromatin-specific manner. Genes & development 14, 1308-1312 (2000). 283 Waldmann, T., Eckerich, C., Baack, M. & Gruss, C. The ubiquitous chromatin protein DEK alters the structure of DNA by introducing positive supercoils. The Journal of biological chemistry 277, 24988-24994, doi:10.1074/jbc.M204045200 (2002). 284 Kappes, F., Scholten, I., Richter, N., Gruss, C. & Waldmann, T. Functional domains of the ubiquitous chromatin protein DEK. Mol Cell Biol 24, 6000-6010, doi:10.1128/MCB.24.13.6000- 6010.2004 (2004). 285 Devany, M., Kotharu, N. P. & Matsuo, H. Solution NMR structure of the C-terminal domain of the human protein DEK. Protein science : a publication of the Protein Society 13, 2252-2259, doi:10.1110/ps.04797104 (2004). 286 Devany, M., Kappes, F., Chen, K. M., Markovitz, D. M. & Matsuo, H. Solution NMR structure of the N-terminal domain of the human DEK protein. Protein science : a publication of the Protein Society 17, 205-215, doi:10.1110/ps.073244108 (2008). 287 Aravind, L. & Koonin, E. V. SAP - a putative DNA-binding motif involved in chromosomal organization. Trends in biochemical sciences 25, 112-114 (2000). 288 Fu, G. K., Grosveld, G. & Markovitz, D. M. DEK, an autoantigen involved in a chromosomal translocation in acute myelogenous leukemia, binds to the HIV-2 enhancer. Proc Natl Acad Sci U S A 94, 1811-1815 (1997). 289 Sawatsubashi, S. et al. A histone chaperone, DEK, transcriptionally coactivates a nuclear receptor. Genes & development 24, 159-170, doi:10.1101/gad.1857410 (2010). 290 Ko, S. I. et al. Regulation of histone acetyltransferase activity of p300 and PCAF by proto- oncogene protein DEK. FEBS letters 580, 3217-3222, doi:10.1016/j.febslet.2006.04.081 (2006). 291 Waldmann, T., Baack, M., Richter, N. & Gruss, C. Structure-specific binding of the proto- oncogene protein DEK to DNA. Nucleic Acids Res 31, 7003-7010 (2003). 292 Brazda, V., Laister, R. C., Jagelska, E. B. & Arrowsmith, C. Cruciform structures are a common DNA feature important for regulating biological processes. BMC Mol Biol 12, 33, doi:10.1186/1471-2199-12-33 (2011). 293 Kappes, F. et al. Phosphorylation by protein kinase CK2 changes the DNA binding properties of the human chromatin protein DEK. Mol Cell Biol 24, 6011-6020, doi:10.1128/MCB.24.13.6011- 6020.2004 (2004). 294 Faulkner, N. E., Hilfinger, J. M. & Markovitz, D. M. Protein phosphatase 2A activates the HIV-2 promoter through enhancer elements that include the pets site. The Journal of biological chemistry 276, 25804-25812, doi:10.1074/jbc.M006454200 (2001). 295 Cleary, J. et al. p300/CBP-associated factor drives DEK into interchromatin granule clusters. The Journal of biological chemistry 280, 31760-31767, doi:10.1074/jbc.M500884200 (2005).

164

296 Gamble, M. J. & Fisher, R. P. SET and PARP1 remove DEK from chromatin to permit access by the transcription machinery. Nature structural & molecular biology 14, 548-555, doi:10.1038/nsmb1248 (2007). 297 Kappes, F. et al. DEK is a poly(ADP-ribose) acceptor in apoptosis and mediates resistance to genotoxic stress. Mol Cell Biol 28, 3245-3257, doi:10.1128/MCB.01921-07 (2008). 298 Mor-Vaknin, N. et al. The DEK nuclear autoantigen is a secreted chemotactic factor. Mol Cell Biol 26, 9484-9496, doi:10.1128/MCB.01030-06 (2006). 299 Saha, A. K. et al. Intercellular trafficking of the nuclear oncoprotein DEK. Proc Natl Acad Sci U S A 110, 6847-6852, doi:10.1073/pnas.1220751110 (2013). 300 Kappes, F. et al. The DEK oncoprotein is a Su(var) that is essential to heterochromatin integrity. Genes & development 25, 673-678, doi:10.1101/gad.2036411 (2011). 301 Hu, H. G., Scholten, I., Gruss, C. & Knippers, R. The distribution of the DEK protein in mammalian chromatin. Biochemical and biophysical research communications 358, 1008-1014, doi:10.1016/j.bbrc.2007.05.019 (2007). 302 Ivanauskiene, K. et al. The PML-associated protein DEK regulates the balance of H3.3 loading on chromatin and is important for telomere integrity. Genome research 24, 1584-1594, doi:10.1101/gr.173831.114 (2014). 303 Hollenbach, A. D., McPherson, C. J., Mientjes, E. J., Iyengar, R. & Grosveld, G. Daxx and histone deacetylase II associate with chromatin through an interaction with core histones and the chromatin-associated protein Dek. Journal of cell science 115, 3319-3330 (2002). 304 Campillos, M., Garcia, M. A., Valdivieso, F. & Vazquez, J. Transcriptional activation by AP-2alpha is modulated by the oncogene DEK. Nucleic Acids Res 31, 1571-1575 (2003). 305 Adams, A. K. et al. IRAK1 is a novel DEK transcriptional target and is essential for head and neck cancer cell survival. Oncotarget 6, 43395-43407, doi:10.18632/oncotarget.6028 (2015). 306 Koleva, R. I. et al. C/EBPalpha and DEK coordinately regulate myeloid differentiation. Blood 119, 4878-4888, doi:10.1182/blood-2011-10-383083 (2012). 307 Zhang, Y. et al. The DEK oncogene activates VEGF expression and promotes tumor angiogenesis and growth in HIF-1alpha-dependent and -independent manners. Oncotarget 7, 23740-23756, doi:10.18632/oncotarget.8060 (2016). 308 Karam, M. et al. Chromatin redistribution of the DEK oncoprotein represses hTERT transcription in leukemias. Neoplasia 16, 21-30 (2014). 309 Wise-Draper, T. M. et al. Apoptosis inhibition by the human DEK oncoprotein involves interference with p53 functions. Mol Cell Biol 26, 7506-7519, doi:10.1128/MCB.00430-06 (2006). 310 Iannetti, A. et al. Regulation of p53 and Rb links the alternative NF-kappaB pathway to EZH2 expression and cell senescence. PLoS genetics 10, e1004642, doi:10.1371/journal.pgen.1004642 (2014). 311 Soares, L. M., Zanier, K., Mackereth, C., Sattler, M. & Valcarcel, J. Intron removal requires proofreading of U2AF/3' splice site recognition by DEK. Science 312, 1961-1965, doi:10.1126/science.1128659 (2006). 312 Delaunay, S. et al. Elp3 links tRNA modification to IRES-dependent translation of LEF1 to sustain metastasis in breast cancer. J Exp Med 213, 2503-2523, doi:10.1084/jem.20160397 (2016). 313 Lee, K. S. et al. Caspase-dependent apoptosis induction by targeted expression of DEK in Drosophila involves histone acetylation inhibition. Journal of cellular biochemistry 103, 1283- 1293, doi:10.1002/jcb.21511 (2008). 314 Khodadoust, M. S. et al. Melanoma proliferation and chemoresistance controlled by the DEK oncogene. Cancer Res 69, 6405-6413, doi:10.1158/0008-5472.CAN-09-1063 (2009).

165

315 Privette Vinnedge, L. M. et al. The DEK oncogene promotes cellular proliferation through paracrine Wnt signaling in Ron receptor-positive breast cancers. Oncogene 34, 2325-2336, doi:10.1038/onc.2014.173 (2015). 316 Privette Vinnedge, L. M. et al. The human DEK oncogene stimulates beta-catenin signaling, invasion and mammosphere formation in breast cancer. Oncogene 30, 2741-2752, doi:10.1038/onc.2011.2 (2011). 317 Deutzmann, A. et al. The human oncoprotein and chromatin architectural factor DEK counteracts DNA replication stress. Oncogene 34, 4270-4277, doi:10.1038/onc.2014.346 (2015). 318 Adams, A. K. et al. DEK promotes HPV-positive and -negative head and neck cancer cell proliferation. Oncogene 34, 868-877, doi:10.1038/onc.2014.15 (2015). 319 Matrka, M. C. et al. DEK over-expression promotes mitotic defects and micronucleus formation. Cell Cycle, 0, doi:10.1080/15384101.2015.1044177 (2015). 320 Waldmann, T., Scholten, I., Kappes, F., Hu, H. G. & Knippers, R. The DEK protein--an abundant and ubiquitous constituent of mammalian chromatin. Gene 343, 1-9, doi:10.1016/j.gene.2004.08.029 (2004). 321 Fenech, M. et al. Molecular mechanisms of micronucleus, nucleoplasmic bridge and nuclear bud formation in mammalian and human cells. Mutagenesis 26, 125-132, doi:10.1093/mutage/geq052 (2011). 322 Yamazaki, H. et al. SiRNA knockdown of the DEK nuclear protein mRNA enhances apoptosis and chemosensitivity of canine transitional cell carcinoma cells. Vet J 204, 60-65, doi:10.1016/j.tvjl.2015.02.009 (2015). 323 Riveiro-Falkenbach, E. & Soengas, M. S. Control of tumorigenesis and chemoresistance by the DEK oncogene. Clinical cancer research : an official journal of the American Association for Cancer Research 16, 2932-2938, doi:10.1158/1078-0432.CCR-09-2330 (2010). 324 Meyn, M. S., Lu-Kuo, J. M. & Herzing, L. B. Expression cloning of multiple human cDNAs that complement the phenotypic defects of ataxia-telangiectasia group D fibroblasts. American journal of human genetics 53, 1206-1216 (1993). 325 Rothkamm, K., Kruger, I., Thompson, L. H. & Lobrich, M. Pathways of DNA double-strand break repair during the mammalian cell cycle. Mol Cell Biol 23, 5706-5715 (2003). 326 Walker, J. R., Corpina, R. A. & Goldberg, J. Structure of the Ku heterodimer bound to DNA and its implications for double-strand break repair. Nature 412, 607-614, doi:10.1038/35088000 (2001). 327 Kass, E. M. & Jasin, M. Collaboration and competition between DNA double-strand break repair pathways. FEBS letters 584, 3703-3708, doi:10.1016/j.febslet.2010.07.057 (2010). 328 Menon, V. & Povirk, L. F. End-processing nucleases and phosphodiesterases: An elite supporting cast for the non-homologous end joining pathway of DNA double-strand break repair. DNA repair 43, 57-68, doi:10.1016/j.dnarep.2016.05.011 (2016). 329 Boettcher, M. & McManus, M. T. Choosing the Right Tool for the Job: RNAi, TALEN, or CRISPR. Molecular cell 58, 575-585, doi:10.1016/j.molcel.2015.04.028 (2015). 330 Ghezraoui, H. et al. Chromosomal translocations in human cells are generated by canonical nonhomologous end-joining. Molecular cell 55, 829-842, doi:10.1016/j.molcel.2014.08.002 (2014). 331 Shibata, A. et al. Factors determining DNA double-strand break repair pathway choice in G2 phase. EMBO J 30, 1079-1092, doi:10.1038/emboj.2011.27 (2011). 332 Stavnezer, J. & Schrader, C. E. IgH chain class switch recombination: mechanism and regulation. J Immunol 193, 5370-5378, doi:10.4049/jimmunol.1401849 (2014). 333 Krejci, L., Altmannova, V., Spirek, M. & Zhao, X. Homologous recombination and its regulation. Nucleic Acids Res 40, 5795-5818, doi:10.1093/nar/gks270 (2012).

166

334 Dupre, A., Boyer-Chatenet, L. & Gautier, J. Two-step activation of ATM by DNA and the Mre11- Rad50-Nbs1 complex. Nature structural & molecular biology 13, 451-457, doi:10.1038/nsmb1090 (2006). 335 Lee, J. H. & Paull, T. T. ATM activation by DNA double-strand breaks through the Mre11-Rad50- Nbs1 complex. Science 308, 551-554, doi:10.1126/science.1108297 (2005). 336 So, S., Davis, A. J. & Chen, D. J. Autophosphorylation at serine 1981 stabilizes ATM at DNA damage sites. The Journal of cell biology 187, 977-990, doi:10.1083/jcb.200906064 (2009). 337 Yuan, J., Adamski, R. & Chen, J. Focus on histone variant H2AX: to be or not to be. FEBS letters 584, 3717-3724, doi:10.1016/j.febslet.2010.05.021 (2010). 338 Ceccaldi, R., Rondinelli, B. & D'Andrea, A. D. Repair Pathway Choices and Consequences at the Double-Strand Break. Trends Cell Biol 26, 52-64, doi:10.1016/j.tcb.2015.07.009 (2016). 339 Morrical, S. W. DNA-pairing and annealing processes in homologous recombination and homology-directed repair. Cold Spring Harb Perspect Biol 7, a016444, doi:10.1101/cshperspect.a016444 (2015). 340 Ward, A., Khanna, K. K. & Wiegmans, A. P. Targeting homologous recombination, new pre- clinical and clinical therapeutic combinations inhibiting RAD51. Cancer Treat Rev 41, 35-45, doi:10.1016/j.ctrv.2014.10.006 (2015). 341 Krajewska, M., Fehrmann, R. S., de Vries, E. G. & van Vugt, M. A. Regulators of homologous recombination repair as novel targets for cancer treatment. Front Genet 6, 96, doi:10.3389/fgene.2015.00096 (2015). 342 Norquist, B. et al. Secondary somatic mutations restoring BRCA1/2 predict chemotherapy resistance in hereditary ovarian carcinomas. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 29, 3008-3015, doi:10.1200/JCO.2010.34.2980 (2011). 343 Dhillon, K. K., Swisher, E. M. & Taniguchi, T. Secondary mutations of BRCA1/2 and drug resistance. Cancer Sci 102, 663-669, doi:10.1111/j.1349-7006.2010.01840.x (2011). 344 Taniguchi, T. et al. Disruption of the Fanconi anemia-BRCA pathway in cisplatin-sensitive ovarian tumors. Nat Med 9, 568-574, doi:10.1038/nm852 (2003). 345 Xie, C. et al. Panobinostat enhances cytarabine and daunorubicin sensitivities in AML cells through suppressing the expression of BRCA1, CHK1, and Rad51. PLoS One 8, e79106, doi:10.1371/journal.pone.0079106 (2013). 346 Abbott, D. W., Freeman, M. L. & Holt, J. T. Double-strand break repair deficiency and radiation sensitivity in BRCA2 mutant cancer cells. Journal of the National Cancer Institute 90, 978-985 (1998). 347 Graeser, M. et al. A marker of homologous recombination predicts pathologic complete response to neoadjuvant chemotherapy in primary breast cancer. Clinical cancer research : an official journal of the American Association for Cancer Research 16, 6159-6168, doi:10.1158/1078-0432.CCR-10-1027 (2010). 348 Kiyohara, E., Tamai, K., Katayama, I. & Kaneda, Y. The combination of chemotherapy with HVJ-E containing Rad51 siRNA elicited diverse anti-tumor effects and synergistically suppressed melanoma. Gene Ther 19, 734-741, doi:10.1038/gt.2011.123 (2012). 349 Short, S. C. et al. Rad51 inhibition is an effective means of targeting DNA repair in glioma models and CD133+ tumor-derived cells. Neuro Oncol 13, 487-499, doi:10.1093/neuonc/nor010 (2011). 350 Takenaka, T. et al. Combined evaluation of Rad51 and ERCC1 expressions for sensitivity to platinum agents in non-small cell lung cancer. International journal of cancer. Journal international du cancer 121, 895-900, doi:10.1002/ijc.22738 (2007). 351 Yang, K., Guo, R. & Xu, D. Non-homologous end joining: advances and frontiers. Acta Biochim Biophys Sin (Shanghai) 48, 632-640, doi:10.1093/abbs/gmw046 (2016).

167

352 Kakarougkas, A. & Jeggo, P. A. DNA DSB repair pathway choice: an orchestrated handover mechanism. Br J Radiol 87, 20130685, doi:10.1259/bjr.20130685 (2014). 353 Bothmer, A. et al. Characterization of the interplay between DNA repair and CRISPR/Cas9- induced DNA lesions at an endogenous locus. Nat Commun 8, 13905, doi:10.1038/ncomms13905 (2017). 354 Cox, D. B., Platt, R. J. & Zhang, F. Therapeutic genome editing: prospects and challenges. Nat Med 21, 121-131, doi:10.1038/nm.3793 (2015). 355 Chapman, J. R. et al. RIF1 Is Essential for 53BP1-Dependent Nonhomologous End Joining and Suppression of DNA Double-Strand Break Resection. Molecular cell, doi:10.1016/j.molcel.2013.01.002 (2013). 356 Bothmer, A. et al. 53BP1 regulates DNA resection and the choice between classical and alternative end joining during class switch recombination. J Exp Med 207, 855-865, doi:10.1084/jem.20100244 (2010). 357 Zimmermann, M., Lottersberger, F., Buonomo, S. B., Sfeir, A. & de Lange, T. 53BP1 regulates DSB repair using Rif1 to control 5' end resection. Science 339, 700-704, doi:10.1126/science.1231573 (2013). 358 Chen, L., Nievera, C. J., Lee, A. Y. & Wu, X. Cell cycle-dependent complex formation of BRCA1.CtIP.MRN is important for DNA double-strand break repair. The Journal of biological chemistry 283, 7713-7720, doi:10.1074/jbc.M710245200 (2008). 359 Feng, L. et al. Cell cycle-dependent inhibition of 53BP1 signaling by BRCA1. Cell Discov 1, 15019, doi:10.1038/celldisc.2015.19 (2015). 360 Neal, J. A. et al. Inhibition of homologous recombination by DNA-dependent protein kinase requires kinase activity, is titratable, and is modulated by autophosphorylation. Mol Cell Biol 31, 1719-1733, doi:10.1128/MCB.01298-10 (2011). 361 Allen, C., Halbrook, J. & Nickoloff, J. A. Interactive competition between homologous recombination and non-homologous end joining. Molecular cancer research : MCR 1, 913-920 (2003). 362 Pierce, A. J., Hu, P., Han, M., Ellis, N. & Jasin, M. Ku DNA end-binding protein modulates homologous repair of double-strand breaks in mammalian cells. Genes & development 15, 3237- 3242, doi:10.1101/gad.946401 (2001). 363 Bunting, S. F. et al. 53BP1 inhibits homologous recombination in Brca1-deficient cells by blocking resection of DNA breaks. Cell 141, 243-254, doi:10.1016/j.cell.2010.03.012 (2010). 364 Kee, Y. & D'Andrea, A. D. Expanded roles of the Fanconi anemia pathway in preserving genomic stability. Genes & development 24, 1680-1694, doi:10.1101/gad.1955310 (2010). 365 Pace, P. et al. Ku70 corrupts DNA repair in the absence of the Fanconi anemia pathway. Science 329, 219-223, doi:10.1126/science.1192277 (2010). 366 Adamo, A. et al. Preventing nonhomologous end joining suppresses DNA repair defects of Fanconi anemia. Molecular cell 39, 25-35, doi:10.1016/j.molcel.2010.06.026 (2010). 367 Escribano-Diaz, C. et al. A Cell Cycle-Dependent Regulatory Circuit Composed of 53BP1-RIF1 and BRCA1-CtIP Controls DNA Repair Pathway Choice. Molecular cell, doi:10.1016/j.molcel.2013.01.001 (2013). 368 Sandahl, J. D. et al. t(6;9)(p22;q34)/DEK-NUP214-rearranged pediatric myeloid leukemia: an international study of 62 patients. Haematologica 99, 865-872, doi:10.3324/haematol.2013.098517 (2014). 369 Datta, A. et al. Oncoprotein DEK as a tissue and urinary biomarker for bladder cancer. BMC cancer 11, 234, doi:10.1186/1471-2407-11-234 (2011). 370 Liu, S. et al. DEK overexpression is correlated with the clinical features of breast cancer. Pathology international 62, 176-181, doi:10.1111/j.1440-1827.2011.02775.x (2012).

168

371 Wu, Q. et al. DEK overexpression in uterine cervical cancers. Pathology international 58, 378- 382, doi:10.1111/j.1440-1827.2008.02239.x (2008). 372 Lin, L. et al. DEK over expression as an independent biomarker for poor prognosis in colorectal cancer. BMC cancer 13, 366, doi:10.1186/1471-2407-13-366 (2013). 373 Piao, J. et al. High expression of DEK predicts poor prognosis of gastric adenocarcinoma. Diagn Pathol 9, 67, doi:10.1186/1746-1596-9-67 (2014). 374 Ou, Y. et al. Overexpression of DEK is an indicator of poor prognosis in patients with gastric adenocarcinoma. Oncol Lett 11, 1823-1828, doi:10.3892/ol.2016.4147 (2016). 375 Lin, L. J. & Chen, L. T. The role of DEK protein in hepatocellular carcinoma for progression and prognosis. Pak J Med Sci 29, 778-782 (2013). 376 Lin, D. et al. Identification of DEK as a potential therapeutic target for neuroendocrine prostate cancer. Oncotarget 6, 1806-1820 (2015). 377 Kappes, F. et al. DEK expression in melanocytic lesions. Human pathology 42, 932-938, doi:10.1016/j.humpath.2010.10.022 (2011). 378 Riveiro-Falkenbach, E. et al. DEK oncogene is overexpressed during melanoma progression. Pigment Cell Melanoma Res, doi:10.1111/pcmr.12563 (2016). 379 Liu, X. et al. Significance of DEK overexpression for the prognostic evaluation of non-small cell lung carcinoma. Oncology reports 35, 155-162, doi:10.3892/or.2015.4365 (2016). 380 Sun, J. et al. DEK protein overexpression predicts poor prognosis in pancreatic ductal adenocarcinoma. Oncology reports 37, 857-864, doi:10.3892/or.2016.5302 (2017). 381 Wang, X. et al. High expression of oncoprotein DEK predicts poor prognosis of small cell lung cancer. Int J Clin Exp Pathol 7, 5016-5023 (2014). 382 Han, S. et al. Clinicopathological significance of DEK overexpression in serous ovarian tumors. Pathology international 59, 443-447, doi:10.1111/j.1440-1827.2009.02392.x (2009). 383 Tabbert, A. et al. Hypophosphorylation of the architectural chromatin protein DEK in death- receptor-induced apoptosis revealed by the isotope coded protein label proteomic platform. Proteomics 6, 5758-5772, doi:10.1002/pmic.200600197 (2006). 384 Wise-Draper, T. M. et al. Overexpression of the cellular DEK protein promotes epithelial transformation in vitro and in vivo. Cancer Res 69, 1792-1799, doi:10.1158/0008-5472.CAN-08- 2304 (2009). 385 Krithivas, A., Fujimuro, M., Weidner, M., Young, D. B. & Hayward, S. D. Protein interactions targeting the latency-associated nuclear antigen of Kaposi's sarcoma-associated herpesvirus to cell chromosomes. J Virol 76, 11596-11604 (2002). 386 Wise-Draper, T. M. et al. DEK proto-oncogene expression interferes with the normal epithelial differentiation program. Am J Pathol 174, 71-81, doi:10.2353/ajpath.2009.080330 (2009). 387 Broxmeyer, H. E. et al. DEK regulates hematopoietic stem engraftment and progenitor cell proliferation. Stem cells and development 21, 1449-1454, doi:10.1089/scd.2011.0451 (2012). 388 Logan, G. E. et al. DEK oncogene expression during normal hematopoiesis and in Acute Myeloid Leukemia (AML). Blood Cells Mol Dis 54, 123-131, doi:10.1016/j.bcmd.2014.07.009 (2015). 389 Qiao, M. X. et al. Regulation of DEK expression by AP-2alpha and methylation level of DEK promoter in hepatocellular carcinoma. Oncology reports 36, 2382-2390, doi:10.3892/or.2016.4984 (2016). 390 Privette Vinnedge, L. M., Ho, S. M., Wikenheiser-Brokamp, K. A. & Wells, S. I. The DEK oncogene is a target of steroid hormone receptor signaling in breast cancer. PLoS One 7, e46985, doi:10.1371/journal.pone.0046985 (2012). 391 Mor-Vaknin, N. et al. DEK in the synovium of patients with juvenile idiopathic arthritis: characterization of DEK antibodies and posttranslational modification of the DEK autoantigen. Arthritis Rheum 63, 556-567, doi:10.1002/art.30138 (2011).

169

392 Yu, L. et al. Critical role of DEK and its regulation in tumorigenesis and metastasis of hepatocellular carcinoma. Oncotarget 7, 26844-26855, doi:10.18632/oncotarget.8565 (2016). 393 Wang, J. et al. DEK Depletion Negatively Regulates Rho/ROCK/MLC Pathway in Non-Small Cell Lung Cancer. The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society 61, 510-521, doi:10.1369/0022155413488120 (2013). 394 Wise-Draper, T. M. et al. The human DEK proto-oncogene is a senescence inhibitor and an upregulated target of high-risk human papillomavirus E7. J Virol 79, 14309-14317, doi:10.1128/JVI.79.22.14309-14317.2005 (2005). 395 Qin, H., Malek, S., Cowell, J. K. & Ren, M. Transformation of human CD34+ hematopoietic progenitor cells with DEK-NUP214 induces AML in an immunocompromised mouse model. Oncogene 35, 5686-5691, doi:10.1038/onc.2016.118 (2016). 396 Bohm, F. et al. The SAF-box domain of chromatin protein DEK. Nucleic Acids Res 33, 1101-1110, doi:10.1093/nar/gki258 (2005). 397 Kotnis, A., Du, L., Liu, C., Popov, S. W. & Pan-Hammarstrom, Q. Non-homologous end joining in class switch recombination: the beginning of the end. Philos Trans R Soc Lond B Biol Sci 364, 653- 665, doi:10.1098/rstb.2008.0196 (2009). 398 Cheung-Ong, K., Giaever, G. & Nislow, C. DNA-damaging agents in cancer chemotherapy: serendipity and chemical biology. Chem Biol 20, 648-659, doi:10.1016/j.chembiol.2013.04.007 (2013). 399 Hashimoto, Y., Ray Chaudhuri, A., Lopes, M. & Costanzo, V. Rad51 protects nascent DNA from Mre11-dependent degradation and promotes continuous DNA synthesis. Nature structural & molecular biology 17, 1305-1311, doi:10.1038/nsmb.1927 (2010). 400 Schlacher, K., Wu, H. & Jasin, M. A distinct replication fork protection pathway connects Fanconi anemia tumor suppressors to RAD51-BRCA1/2. Cancer cell 22, 106-116, doi:10.1016/j.ccr.2012.05.015 (2012). 401 Lossaint, G. et al. FANCD2 binds MCM proteins and controls replisome function upon activation of s phase checkpoint signaling. Molecular cell 51, 678-690, doi:10.1016/j.molcel.2013.07.023 (2013). 402 Gonzalez-Prieto, R., Munoz-Cabello, A. M., Cabello-Lobato, M. J. & Prado, F. Rad51 replication fork recruitment is required for DNA damage tolerance. EMBO J 32, 1307-1321, doi:10.1038/emboj.2013.73 (2013). 403 Bakkenist, C. J. & Kastan, M. B. DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421, 499-506, doi:10.1038/nature01368 (2003). 404 Zannini, L., Delia, D. & Buscemi, G. CHK2 kinase in the DNA damage response and beyond. J Mol Cell Biol 6, 442-457, doi:10.1093/jmcb/mju045 (2014). 405 Prakash, R., Zhang, Y., Feng, W. & Jasin, M. Homologous recombination and human health: the roles of BRCA1, BRCA2, and associated proteins. Cold Spring Harb Perspect Biol 7, a016600, doi:10.1101/cshperspect.a016600 (2015). 406 Scully, R. & Xie, A. Double strand break repair functions of histone H2AX. Mutation research 750, 5-14, doi:10.1016/j.mrfmmm.2013.07.007 (2013). 407 Jasin, M. & Rothstein, R. Repair of strand breaks by homologous recombination. Cold Spring Harb Perspect Biol 5, a012740, doi:10.1101/cshperspect.a012740 (2013). 408 Yun, M. H. & Hiom, K. CtIP-BRCA1 modulates the choice of DNA double-strand-break repair pathway throughout the cell cycle. Nature 459, 460-463, doi:10.1038/nature07955 (2009). 409 Polato, F. et al. CtIP-mediated resection is essential for viability and can operate independently of BRCA1. J Exp Med 211, 1027-1036, doi:10.1084/jem.20131939 (2014).

170

410 Xu, G. et al. REV7 counteracts DNA double-strand break resection and affects PARP inhibition. Nature 521, 541-544, doi:10.1038/nature14328 (2015). 411 Onyango, D. O., Howard, S. M., Neherin, K., Yanez, D. A. & Stark, J. M. Tetratricopeptide repeat factor XAB2 mediates the end resection step of homologous recombination. Nucleic Acids Res 44, 5702-5716, doi:10.1093/nar/gkw275 (2016). 412 Lombardi, A. J. et al. Acquisition of Relative Interstrand Crosslinker Resistance and PARP Inhibitor Sensitivity in Fanconi Anemia Head and Neck Cancers. Clinical cancer research : an official journal of the American Association for Cancer Research 21, 1962-1972, doi:10.1158/1078-0432.CCR-14-2616 (2015). 413 Leahy, J. J. et al. Identification of a highly potent and selective DNA-dependent protein kinase (DNA-PK) inhibitor (NU7441) by screening of chromenone libraries. Bioorg Med Chem Lett 14, 6083-6087, doi:10.1016/j.bmcl.2004.09.060 (2004). 414 Veuger, S. J., Curtin, N. J., Richardson, C. J., Smith, G. C. & Durkacz, B. W. Radiosensitization and DNA repair inhibition by the combined use of novel inhibitors of DNA-dependent protein kinase and poly(ADP-ribose) polymerase-1. Cancer Res 63, 6008-6015 (2003). 415 Callen, E. et al. Essential role for DNA-PKcs in DNA double-strand break repair and apoptosis in ATM-deficient lymphocytes. Molecular cell 34, 285-297, doi:10.1016/j.molcel.2009.04.025 (2009). 416 Bohringer, M. & Wiesmuller, L. Fluorescence-based quantification of pathway-specific DNA double-strand break repair activities: a powerful method for the analysis of genome destabilizing mechanisms. Subcell Biochem 50, 297-306, doi:10.1007/978-90-481-3471-7_15 (2010). 417 Willis, N. A. et al. BRCA1 controls homologous recombination at Tus/Ter-stalled mammalian replication forks. Nature 510, 556-559, doi:10.1038/nature13295 (2014). 418 Chandramouly, G. et al. BRCA1 and CtIP suppress long-tract gene conversion between sister chromatids. Nat Commun 4, 2404, doi:10.1038/ncomms3404 (2013). 419 Petermann, E., Orta, M. L., Issaeva, N., Schultz, N. & Helleday, T. Hydroxyurea-stalled replication forks become progressively inactivated and require two different RAD51-mediated pathways for restart and repair. Molecular cell 37, 492-502, doi:10.1016/j.molcel.2010.01.021 (2010). 420 Raynard, S. & Sung, P. Assay for human Rad51-mediated DNA displacement loop formation. Cold Spring Harb Protoc 2009, pdb prot5120, doi:10.1101/pdb.prot5120 (2009). 421 Ciszewski, W. M., Tavecchio, M., Dastych, J. & Curtin, N. J. DNA-PK inhibition by NU7441 sensitizes breast cancer cells to ionizing radiation and doxorubicin. Breast Cancer Res Treat 143, 47-55, doi:10.1007/s10549-013-2785-6 (2014). 422 Matrka, M. C. et al. DEK over-expression promotes mitotic defects and micronucleus formation. Cell Cycle 14, 3939-3953, doi:10.1080/15384101.2015.1044177 (2015). 423 Wang, X., Andreassen, P. R. & D'Andrea, A. D. Functional interaction of monoubiquitinated FANCD2 and BRCA2/FANCD1 in chromatin. Mol Cell Biol 24, 5850-5862, doi:10.1128/MCB.24.13.5850-5862.2004 (2004). 424 Siegel, R. L., Miller, K. D. & Jemal, A. Cancer statistics, 2015. CA: a cancer journal for clinicians 65, 5-29, doi:10.3322/caac.21254 (2015). 425 Sturgis, E. M. & Cinciripini, P. M. Trends in head and neck cancer incidence in relation to smoking prevalence: an emerging epidemic of human papillomavirus-associated cancers? Cancer 110, 1429-1435, doi:10.1002/cncr.22963 (2007). 426 Fakhry, C. et al. Improved survival of patients with human papillomavirus-positive head and neck squamous cell carcinoma in a prospective clinical trial. Journal of the National Cancer Institute 100, 261-269, doi:10.1093/jnci/djn011 (2008).

171

427 Masterson, L. et al. De-escalation treatment protocols for human papillomavirus-associated oropharyngeal squamous cell carcinoma: a systematic review and meta-analysis of current clinical trials. Eur J Cancer 50, 2636-2648, doi:10.1016/j.ejca.2014.07.001 (2014). 428 Sharma, A. et al. Human papillomavirus-positive oral cavity and oropharyngeal cancer patients do not have better quality-of-life trajectories. Otolaryngol Head Neck Surg 146, 739-745, doi:10.1177/0194599811434707 (2012). 429 Baxi, S. S. et al. Employment and return to work following chemoradiation in patient with HPV- related oropharyngeal cancer. Cancers of the Head & Neck 1, 1-8, doi:10.1186/s41199-016- 0002-0 (2016). 430 Bosch, F. X., Lorincz, A., Munoz, N., Meijer, C. J. & Shah, K. V. The causal relation between human papillomavirus and cervical cancer. Journal of clinical pathology 55, 244-265 (2002). 431 Wen, Z. et al. Orphan nuclear receptor PNR/NR2E3 stimulates p53 functions by enhancing p53 acetylation. Mol Cell Biol 32, 26-35, doi:10.1128/MCB.05513-11 (2012). 432 Scheffner, M., Werness, B. A., Huibregtse, J. M., Levine, A. J. & Howley, P. M. The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 63, 1129-1136 (1990). 433 Cancer Genome Atlas, N. Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature 517, 576-582, doi:10.1038/nature14129 (2015). 434 Kalavrezos, N. & Bhandari, R. Current trends and future perspectives in the surgical management of oral cancer. Oral oncology 46, 429-432, doi:10.1016/j.oraloncology.2010.03.007 (2010). 435 Hauswald, H., Simon, C., Hecht, S., Debus, J. & Lindel, K. Long-term outcome and patterns of failure in patients with advanced head and neck cancer. Radiation oncology 6, 70, doi:10.1186/1748-717X-6-70 (2011). 436 Seiwert, T. Y. Ties that bind: p16 as a prognostic biomarker and the need for high-accuracy human papillomavirus testing. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 32, 3914-3916, doi:10.1200/JCO.2014.57.9268 (2014). 437 Smeets, S. J. et al. A novel algorithm for reliable detection of human papillomavirus in paraffin embedded head and neck cancer specimen. International journal of cancer. Journal international du cancer 121, 2465-2472, doi:10.1002/ijc.22980 (2007). 438 Lewis, J. S., Jr. et al. p16 positive oropharyngeal squamous cell carcinoma:an entity with a favorable prognosis regardless of tumor HPV status. Am J Surg Pathol 34, 1088-1096, doi:10.1097/PAS.0b013e3181e84652 (2010). 439 Rietbergen, M. M. et al. Molecular characterization of p16-immunopositive but HPV DNA- negative oropharyngeal carcinomas. International journal of cancer. Journal international du cancer 134, 2366-2372, doi:10.1002/ijc.28580 (2014). 440 Enders, G. H., Koh, J., Missero, C., Rustgi, A. K. & Harlow, E. p16 inhibition of transformed and primary squamous epithelial cells. Oncogene 12, 1239-1245 (1996). 441 Cody, D. T., 2nd, Huang, Y., Darby, C. J., Johnson, G. K. & Domann, F. E. Differential DNA methylation of the p16 INK4A/CDKN2A promoter in human oral cancer cells and normal human oral keratinocytes. Oral oncology 35, 516-522 (1999). 442 Darbro, B. W., Schneider, G. B. & Klingelhutz, A. J. Co-regulation of p16INK4A and migratory genes in culture conditions that lead to premature senescence in human keratinocytes. J Invest Dermatol 125, 499-509, doi:10.1111/j.0022-202X.2005.23844.x (2005). 443 Simpson, D. S., Mason-Richie, N. A., Gettler, C. A. & Wikenheiser-Brokamp, K. A. Retinoblastoma family proteins have distinct functions in pulmonary epithelial cells in vivo critical for suppressing cell growth and tumorigenesis. Cancer Res 69, 8733-8741, doi:10.1158/0008- 5472.CAN-09-1359 (2009).

172

444 Kotake, Y. et al. pRB family proteins are required for H3K27 trimethylation and Polycomb repression complexes binding to and silencing p16INK4alpha tumor suppressor gene. Genes & development 21, 49-54, doi:10.1101/gad.1499407 (2007). 445 Munger, K., Gwin, T. K. & McLaughlin-Drubin, M. E. p16 in HPV-associated cancers. Oncotarget 4, 1864-1865, doi:10.18632/oncotarget.1523 (2013). 446 Loughran, O. et al. Association of CDKN2A/p16INK4A with human head and neck keratinocyte replicative senescence: relationship of dysfunction to immortality and neoplasia. Oncogene 13, 561-568 (1996). 447 von Lindern, M. et al. Translocation t(6;9) in acute non-lymphocytic leukaemia results in the formation of a DEK-CAN fusion gene. Baillieres Clin Haematol 5, 857-879 (1992). 448 Pease, N. A., Wise-Draper, T. & Privette Vinnedge, L. Dissecting the Potential Interplay of DEK Functions in Inflammation and Cancer. J Oncol 2015, 106517, doi:10.1155/2015/106517 (2015). 449 Kumar, B. et al. High expression of myoferlin is associated with poor outcome in oropharyngeal squamous cell carcinoma patients and is inversely associated with HPV-status. Oncotarget, doi:10.18632/oncotarget.7625 (2016). 450 Yadav, A., Kumar, B., Datta, J., Teknos, T. N. & Kumar, P. IL-6 promotes head and neck tumor metastasis by inducing epithelial-mesenchymal transition via the JAK-STAT3-SNAIL signaling pathway. Molecular cancer research : MCR 9, 1658-1667, doi:10.1158/1541-7786.MCR-11-0271 (2011). 451 Li, J., Poi, M. J. & Tsai, M. D. Regulatory mechanisms of tumor suppressor P16(INK4A) and their relevance to cancer. Biochemistry 50, 5566-5582, doi:10.1021/bi200642e (2011). 452 Jeon, S. & Lambert, P. F. Integration of human papillomavirus type 16 DNA into the human genome leads to increased stability of E6 and E7 mRNAs: implications for cervical carcinogenesis. Proc Natl Acad Sci U S A 92, 1654-1658 (1995). 453 Santos, G. C., Zielenska, M., Prasad, M. & Squire, J. A. Chromosome 6p amplification and cancer progression. Journal of clinical pathology 60, 1-7, doi:10.1136/jcp.2005.034389 (2007). 454 Mor-Vaknin, N. et al. DEK-targeting DNA aptamers as therapeutics for inflammatory arthritis. Nat Commun 8, 14252, doi:10.1038/ncomms14252 (2017).

173