Molecular Mechanisms of Oncogenesis & Precision Medicine

University of Pennsylvania ScholarlyCommons

Publicly Accessible Penn Dissertations

2016

Molecular Mechanisms of Oncogenesis & Precision Medicine Approaches for Pediatric Low-Grade Gliomas

Payal Jain University of Pennsylvania, [email protected]

Follow this and additional works at: https://repository.upenn.edu/edissertations

Part of the Cell Biology Commons, Medicine and Health Sciences Commons, and the Molecular Biology Commons

Recommended Citation Jain, Payal, "Molecular Mechanisms of Oncogenesis & Precision Medicine Approaches for Pediatric Low- Grade Gliomas" (2016). Publicly Accessible Penn Dissertations. 1782. https://repository.upenn.edu/edissertations/1782

This paper is posted at ScholarlyCommons. https://repository.upenn.edu/edissertations/1782 For more information, please contact [email protected]. Molecular Mechanisms of Oncogenesis & Precision Medicine Approaches for Pediatric Low-Grade Gliomas

Abstract Pediatric low-grade gliomas (PLGGs) are a heterogeneous group of tumors that collectively represent the most common childhood brain cancer. Despite favorable outcomes with surgical and adjuvant therapies, majority of patients suffer from long-term treatment-related morbidities and recurrent/inoperable disease. This necessitates a deeper understanding of PLGG biology to aid development of molecular diagnostics and low-toxicity targeted therapeutics.

Hitherto, PLGGs have been defined by activating mutations that dysregulate the MAPK signaling pathway, leading to clinical testing of RAF/MAPK inhibitors for PLGGs. Interestingly, recent large-scale sequencing efforts discovered novel gene fusions in PLGGs and we identified the unique ecurrr ent association of tumor suppressor Quaking (QKI) with distinct proto-oncogenes, MYB and RAF1, in different PLGG subtypes. We hypothesized that MYB-QKI and QKI-RAF1 function via novel oncogenic mechanisms invoking a two-hit mechanism of gain-of-function in the MYB/RAF1 oncoproteins collaborating with QKI loss of putative tumor suppressor function, describing for the first time a unique gene fusion setting involving both fusion partners with implications for therapeutic targeting.

Utilizing heterologous cell model systems and in vivo mouse models, we found MYB-QKI and QKI-RAF1 are capable of driving oncogenesis. Furthermore, MYB-QKI is a specific driver mutation defining angiocentric gliomas and mediates tumorigenesis via a tri-partite mechanism: (1)MYB oncogenic activation via truncation, (2)rearrangement led enhancer translocation that drives MYB-QKI expression and (3)LOH of QKI tumor suppressor. In contrast, QKI-RAF1 drives some pilocytic astrocytomas via aberrant activation of the MAPK pathway in a QKI-dimerization dependent manner. We also found differential response to RAF targeted therapy in QKI-RAF1, compared to BRAF fusions in PLGGs, due to QKI-mediated dimerization. Hence, our study highlights distinct roles for the same gene, QKI in supporting the oncogenic functions of MYB and RAF1 in different PLGG-gene fusions.

Overall, our study has uncovered distinct molecular mechanisms associated with different QKI gene fusions in PLGGs. We show that MYB-QKI is specific ot angiocentric gliomas and mediates a unique oncogenic program, and with QKI-RAF1 we demonstrate how mutational context guides differential response to targeted therapy. Therefore, our study has important clinical implications on molecular diagnostics and targeted therapy for these rather understudied class of childhood brain tumors.

Degree Type Dissertation

Degree Name Doctor of Philosophy (PhD)

Graduate Group Cell & Molecular Biology

First Advisor Adam C. Resnick

Keywords Gene fusions, MYB-QKI, Pediatric brain tumor, Pediatric low-grade glioma, QKI-RAF1, Targeted therapy Subject Categories Cell Biology | Medicine and Health Sciences | Molecular Biology

This dissertation is available at ScholarlyCommons: https://repository.upenn.edu/edissertations/1782 MOLECULAR MECHANISMS OF ONCOGENESIS & PRECISION MEDICINE

APPROACHES FOR PEDIATRIC LOW-GRADE GLIOMAS

Payal Jain

A DISSERTATION

Cell and Molecular Biology

Presented to the Faculties of the University of Pennsylvania

Partial Fulfillment of the Requirements for the

Degree of Doctor of Philosophy

2016

Supervisor of Dissertation

______

Adam C. Resnick, Assistant Professor of Neurosurgery

Graduate Group Chairperson

______

Daniel S. Kessler, Associate Professor of Cell and Developmental Biology

Dissertation Committee

Garrett M. Brodeur, MD, Professor of Pediatrics, Audrey E. Evans Chair in Molecular Oncology and

Associate Director, Abramson Cancer Center, The Children’s Hospital of Philadelphia

Margaret M. Chou, Ph.D., Associate Professor of Pathology and Laboratory Medicine

Jean D. Boyer, Ph.D., Research Associate Professor of Pathology and Laboratory Medicine

David B. Weiner, PhD, Professor, WW Smith Chair in Cancer Research, Director, Wistar Vaccine Center

MOLECULAR MECHANISMS OF ONCOGENESIS & PRECISION MEDICINE

APPROACHES FOR PEDIATRIC LOW-GRADE GLIOMAS

2016

Payal Jain

This work is licensed under the Creative Commons Attribution- NonCommercial-ShareAlike 3.0 License

To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-sa/3.0/us/ Dedication page

To my mother, who always inspired me to pursue my dreams (Debika Jain, 69).

To the children who have lost their lives to brain tumors.

iii

ACKNOWLEDGMENT

I would like to start by thanking my dissertation mentor, Adam C. Resnick, for his unwavering commitment, never-ending enthusiasm and contagious optimism towards research and scientific collaborations for improving therapies for children with brain tumors. I will always be grateful to Adam for being a great teacher and helping me understand the most basic scientific concepts when I first started in lab. Even though I had no prior training in cancer research, Adam always believed in me and pushed me to perform to the best of my abilities. As a result of his mentorship and support, I have been able to complete two major projects within three years in his lab, could present my findings as posters and talks at major national and international conferences and most importantly, contribute in a small way to understand and possibly treat childhood brain tumors. Adam has been a truly inspirational role model, juggling many responsibilities but still always being available for his lab. No matter where I go, I truly hope to imbibe and take with me his immense dedication, hard working nature and positive and helpful attitude.

Next, I would like to thank current and past lab members of the Resnick lab. I thank Katie Boucher and

Namrata Choudhary for their technical help, insight and always being in lab for any help I needed in the past 3 years. I thank Harry Han for lending a helpful ear to all my lab related issues, reading all my proposal/paper drafts and always giving great scientific input. I thank Yunkaun Zhu for being a great friend and an amazing bio-informatician who always made sense of my random questions about handling big data. I thank Lucy Chong and Wanda Anselmo for being wonderful bay mates and friends, especially when working late nights in lab. I thank our past mouse technicians Amanda Silva and Aesha Vakil, and current mouse expert, Tiffany Smith for being so patient with me and helping me with my mouse experiments. I thank Angela J. Waanders for being a great mentor and supporter and the amazing female neurosurgeon,

Tamara M. Fierst for being my partner-in-crime in solving the QKI-RAF1 mystery. I am lucky to have been a part of the Resnick lab, where everyone has always been extremely helpful and supportive, and it has been so much fun to celebrate birthdays, baby showers, festivals and so much more.

I also thank our collaborators at Dana Faber Cancer Institute for their work on the MYB-QKI project. I thank Pratiti Bandopadhayay, Lori A. Ramkissoon and Guillaume Bergthold who are co-first authors on the publication and Keith L. Ligon and Rameen Beroukhim, who were co-PI’s for the project. I have learned the essence of scientific collaborations while working with the team where each contributor came forward with their expertise and made it possible for us to put together an impactful publication in Nature Genetics.

Next, I am grateful for all the helpful suggestions, advice, comments and criticisms provided by my thesis committee members Garrett M. Brodeur, Margaret M. Chou, Jean D. Boyer and David B. Weiner in the past years. I thank each of them for giving me their valuable time and helping me navigate the rocky roads of graduate school. I especially thank David Weiner for being our awesome GTV Chair and always being available to listen to my problems throughout graduate school. He has truly been available for all GTV-ers and I am so grateful to have him as my Program Chair. I also thank Matthew Weitzman for being a great PI on our floor with his invaluable suggestions and scientific input and giving me his valuable time by listening to me practice talks. I am grateful to senior postdocs in the neighboring lab, Grace Tan and

Mateusz Koptyra for always being available and providing me with instant solutions to any problem with my experiments/ lab-related problems/anything and everything.

I am also very thankful to another group of wonderful people, who have heard me cry, complain and have celebrated every step of my graduate school journey with me- my friends. I am grateful to Montreh

Tavakkoli for being my close friend and morale booster during difficult times in PhD, which was usually very often. I sincerely thank my ex-roomates and amazing friends Vanessa Putz, Sato Ortega, Delphine

Rason, Marta, Shailly Saini and Priya Sharma for being my family in Philadelphia and always bringing a smile to my face. I am also thankful to Meilin Fernandez for being my fellow-graduate student and confidante outside lab and all the fun trips we have made together. I am also very grateful to International

House of Philadlephia that has been my wonderful home away from home here in Philly.

Lastly but most importantly, I would like to thank my parents Jawahar Jain and Debika Jain, my brother

Abhinandan Jain, my sister-in-law Nisha Jain and baby niece, Jesal Jain. They have been my pillars of support and strength the past 4 years and I could not have gone through the ups and downs of graduate school without them. I am extremely grateful to my family who have always been just a phone call away, despite being 1000s of miles away in India. No words of gratitude would ever be enough to thank my amazing mother, who passed away on May 12, 2016, but had always been my greatest champion and had been really looking forward to my defense (and finally finishing school!). She, along with family members and relatives have always expressed their pride in what I do, and I am thankful to God for their infinite love and support, near or far.

ABSTRACT

MOLECULAR MECHANISMS OF ONCOGENESIS & PRECISION MEDICINE

APPROACHES FOR PEDIATRIC LOW-GRADE GLIOMAS

Payal Jain

Adam C. Resnick, PhD

Pediatric low-grade gliomas (PLGGs) are a heterogeneous group of tumors that collectively represent the most common childhood brain cancer. Despite favorable outcomes with surgical and adjuvant therapies, majority of patients suffer from long-term treatment-related morbidities and recurrent/inoperable disease. This necessitates a deeper understanding of PLGG biology to aid development of molecular diagnostics and low-toxicity targeted therapeutics.

MYB and RAF1, in different PLGG sub-types. We hypothesized that MYB-QKI and QKI-RAF1 function via novel oncogenic mechanisms invoking a two-hit mechanism of gain-of-function in the

MYB/RAF1 oncoproteins collaborating with QKI loss of putative tumor suppressor function, describing for the first time a unique gene fusion setting involving both fusion partners with implications for therapeutic targeting.

Utilizing heterologous cell model systems and in vivo mouse models, we found MYB-QKI and

QKI-RAF1 are capable of driving oncogenesis. Furthermore, MYB-QKI is a specific driver vii mutation defining angiocentric gliomas and mediates tumorigenesis via a tri-partite mechanism:

(1)MYB oncogenic activation via truncation, (2)rearrangement led enhancer translocation that drives MYB-QKI expression and (3)LOH of QKI tumor suppressor. In contrast, QKI-RAF1 drives some pilocytic astrocytomas via aberrant activation of the MAPK pathway in a QKI- dimerization dependent manner. We also found differential response to RAF targeted therapy in

QKI-RAF1, compared to BRAF fusions in PLGGs, due to QKI-mediated dimerization. Hence, our study highlights distinct roles for the same gene, QKI in supporting the oncogenic functions of

MYB and RAF1 in different PLGG-gene fusions.

Overall, our study has uncovered distinct molecular mechanisms associated with different QKI gene fusions in PLGGs. We show that MYB-QKI is specific to angiocentric gliomas and mediates a unique oncogenic program, and with QKI-RAF1 we demonstrate how mutational context guides differential response to targeted therapy. Therefore, our study has important clinical implications on molecular diagnostics and targeted therapy for these rather understudied class of childhood brain tumors.

viii

TABLE OF CONTENTS

ACKNOWLEDGMENT ...... IV

ABSTRACT ...... VII

LIST OF TABLES ...... XI

LIST OF FIGURES ...... XII

CHAPTER 1: INTRODUCTION ...... 1

Pediatric Brain Cancer ...... 1 Pediatric Low-Grade Gliomas ...... 2

Landscape of genomic alterations in PLGGs ...... 3 BRAF mutations and gene fusions ...... 4 Non-BRAF mutations and gene fusions ...... 6

Targeted therapy for PLGGs: Need for molecular-based diagnosis and patient stratification for personalized medicine ...... 10 MAPK/PI3K/mTOR targeted inhibitors ...... 11 Novel therapeutic avenues: Epigenetic modulators ...... 12

Tables and Figures ...... 14

CHAPTER 2: MYB-QKI REARRANGEMENTS IN ANGIOCENTRIC GLIOMAS DRIVE ONCOGENICITY THROUGH A TRIPARTITE MECHANISM ...... 22

Summary ...... 22

Introduction ...... 22

Results ...... 23

Discussion ...... 34

Materials and Methods ...... 36

Tables and Figures ...... 48

CHAPTER 3: RAF1 GENE FUSIONS IN PEDIATRIC LOW-GRADE GLIOMAS SHOW DIFFERENTIAL RESPONSE TO TARGETED THERAPEUTICS BASED ON DIMERIZATION PROFILES...... 68

Summary ...... 68

Introduction ...... 69

Results ...... 70

Discussion ...... 82

Materials and Methods ...... 86

Tables and Figures ...... 90

CHAPTER 4: DISCUSSION ...... 100

Takes two to tango: QKI gene fusions define novel oncogenic mechanisms driven by both fusion partners ...... 100 MYB-QKI: Synergism of a proto-oncogene and a tumor suppressor ...... 100 QKI-RAF1: Dimerization and activation of kinase mediated by non-kinase partner ...... 102

All about that grade: Molecular stratification of PLGG patients ...... 103

Rolling in the deep-sequencing: Personalized medicine for PLGG patients...... 104

Molecular Revolution: Utilizing concepts from PLGG gene fusions to inform adult cancers ...... 105

Conclusion...... 107

Tables and Figures ...... 108

REFERENCES ...... 110

LIST OF TABLES

TABLE 2.1. PLGGS INCLUDED IN ANALYSIS INCLUDING TUMOR DEMOGRAPHICS, METHOD OF PROFILING AND DRIVER ALTERATIONS IDENTIFIED...... 48

TABLE 2.2. GISTIC AMPLIFICATION PEAKS ...... 48

TABLE 2.3. GISTIC DELETION PEAKS ...... 48

TABLE 2.4. DIFFERENTIALLY EXPRESSED GENES IN NEURAL STEM CELLS EXPRESSING EGFP, MYBTR OR MYB-QKI ...... 48

TABLE 2.5. GENE-SET ENRICHMENT ANALYSIS IN NEURAL STEM CELLS EXPRESSING EGFP, MYBTR OR MYB-QKI ...... 48

TABLE 2.6. DIFFERENTIALLY EXPRESSED GENES IN NEURAL STEM CELLS EXPRESSING MYB-QKI WITH SHLACZ OR SHQK ...... 48

TABLE 2.7. MIRNAS THAT ARE DIFFERENTIALLY EXPRESSED IN MNSCS EXPRESSING MYB-QKI FOLLOWING SUPPRESSION OF WILD TYPE QK . 48

TABLE 2.8. MYB PROMOTER SEQUENCE AND SHQK CLONES USED IN EXPERIMENTS ...... 48

TABLE 2.9. PCR PRIMERS USED IN EXPERIMENTS...... 49

TABLE 2.10. CHIP-SEQ PEAKS (P1E-6) IDENTIFIED FOLLOWING MYB CHIP-SEQ IN MNSC OVER-EXPRESSING MYBQKI5. PEAKS THAT CONTAIN A MYB BINDING MOTIF (FRAGMENT SIZE USED IN ANALYSIS = 200) ARE SHOWN IN THE COLUMNS LABELED MYBL1 OR MYB...... 49

LIST OF FIGURES

FIGURE 1.1. COMPARISON OF INCIDENCE RATES OF VARIOUS CHILDHOOD CANCERS BY AGE...... 14

FIGURE 1.2. COMPARISON OF INCIDENCE RATES AND MORTALITY RATES BETWEEN ADULT AND CHILDHOOD CANCERS...... 15

FIGURE 1.3. CLASSIFICATION OF PEDIATRIC BRAIN TUMORS BY HISTOLOGY...... 16

FIGURE 1.4. GRADE I AND GRADE II PLGGS CLASSIFIED ON THE BASIS OF GENETIC/MOLECULAR ALTERATIONS...... 17

FIGURE 1.5. SIMPLIFIED LINEAR REPRESENTATION OF THE RAS/RAF/MEK/ERK (MAPK) PATHWAY ...... 18

FIGURE 1.6. BREAKPOINTS IN BRAF FORMING IN-FRAME GENE FUSIONS IN PLGGS...... 19

FIGURE 1.7. FUNCTIONAL PROTEIN DOMAINS ...... 20

FIGURE 2.1. SIGNIFICANT REGIONS OF RECURRENT COPY-NUMBER ALTERATIONS IN PLGG...... 50

FIGURE 2.2. GENOMIC ANALYSIS OF 172 WGS AND/OR RNA-SEQ OF PLGGS REVEALS A RECURRENT REARRANGEMENT INVOLVING MYB AND QKI IN ANGIOCENTRIC GLIOMAS...... 51

FIGURE 2.3. CO-MUT PLOT OF DRIVER ALTERATIONS ...... 52

FIGURE 2.4. ALTERATIONS OF MYB AND QKI OCCUR FREQUENTLY IN HUMAN CANCERS...... 53

FIGURE 2.5. MYB-QKI FUNCTIONS AS A TRANSCRIPTION FACTOR, AND ITS MOLECULAR EFFECTS ARE OBSERVED IN ANGIOCENTRIC GLIOMAS...... 54

xii

FIGURE 2.6. MYB FUNCTIONAL ASSAYS AND RELATED DATA ...... 55

FIGURE 2.7. ANGIOCENTRIC GLIOMAS EXHIBIT ABERRANT EXPRESSION OF MYB-QKI DUE TO H3K27AC ENHANCER TRANSLOCATION AND AN AUTO- REGULATORY FEEDBACK CIRCUIT IN WHICH MYB-QKI BINDS TO THE MYB PROMOTER...... 56

FIGURE 2.8. HUMAN ANGIOCENTRIC GLIOMAS EXHIBIT H3K27AC ENHANCER TRANSLOCATION WITH AN ABERRANT ENHANCER ASSOCIATED WITH THE MYB PROMOTER...... 57

FIGURE 2.9. H3K27AC ENHANCER PROFILING OF 3’QKI IN A HUMAN ANGIOCENTRIC GLIOMA. H3K27AC PEAKS ARE DEPICTED ...... 58

FIGURE 2.10. H3K27AC ENHANCER FORMATION AT THE MYB PROMOTER ..... 59

FIGURE 2.11. SUPPLEMENTARY FIGURES FOR MYB-QKI ASSAYS ...... 60

FIGURE 2.12. MYB-QKI AND MYBTR ARE ONCOGENIC...... 61

FIGURE 2.13 ONCOGENIC ASSAYS OF MYB-QKI EXPRESSING CELLS...... 62

FIGURE 2.14. MYB-QKI DISRUPTS EXPRESSION OF QKI, A TUMOR SUPPRESSOR GENE...... 63

FIGURE 2.15. QKI KNOCKDOWN IN MNSCS AND SUPPLEMENTARY ASSAYS ..... 64

FIGURE 2.16. THE MYB-QKI REARRANGEMENT CONTRIBUTES TO ONCOGENESIS THROUGH AT LEAST THREE MECHANISMS...... 65

FIGURE 2.17. VALIDATION OF MYB CHIP-SEQ...... 66

FIGURE 2.18. VALIDATION OF MYB DNA BINDING MOTIF IN MYB-QKI ...... 67

FIGURE 3.1. QKI-RAF1 AND SRGAP3-RAF1 ARE ONCOGENIC VIA ACTIVATION OF MAPK AND PI3K PATHWAYS...... 90

FIGURE 3.2. EFFECT OF RAF INHIBITORS ON KIAA1549-BRAF SIGNALING...... 91

xiii

FIGURE 3.3. EXISTING RAF INHIBITORS DO NOT SUPPRESS CRAF FUSION DRIVEN SIGNALING PATHWAYS AND ONCOGENIC PHENOTYPE...... 92

FIGURE 3.4. HOMO- AND HETERO-DIMERIZATION IS REQUIRED TO DRIVE CRAF FUSION DRIVEN GLIOMAGENESIS...... 93

FIGURE 3.5. SRGAP3-RAF1 MEDIATED DIMERIZATION AND ONCOGENICITY IS PARTLY MEDIATED BY RAF1...... 94

FIGURE 3.6. RAF1 FUSIONS ACTIVATE ONCOGENIC SIGNALING IN A DIMERIZATION-DEPENDENT MANNER...... 95

FIGURE 3.7. EXPRESSION LEVEL OF QKI AND QKI-RAF1 IN UV CROSS-LINKING ASSAY ...... 96

FIGURE 3.8. NOVEL RAF DIMER INHIBITOR LY3009120 STABILIZES AND INACTIVATES CRAF FUSION DIMERS ...... 97

FIGURE 3.9. MEK INHIBITORS, AZD6244 AND GSK1120212, PARTIALLY SUPPRESS QKI-RAF1 MEDIATED TUMOR GROWTH...... 98

FIGURE 3.10. COMBINATORIAL TARGETING OF MAPK AND PI3K PATHWAY .. 99

FIGURE 4.1. MYB-QKI DRIVES ONCOGENESIS IN ANGIOCENTRIC GLIOMAS VIA A TRIPARTITE MECHANISM...... 108

FIGURE 4.2. MAPK AND PI3K ONCOGENIC PATHWAYS DRIVEN BY QKI-RAF1 AND RESPONSE TO TARGETED THERAPIES...... 109

xiv

CHAPTER 1: Introduction

Pediatric Brain Cancer

Childhood brain cancer is defined as a disease where abnormal cells accumulate in the tissues of the brain and spinal cord, which together make up the central nervous system (CNS). CNS tumors represent the most common cancer in children 0-14 years with infants (<1 year old) showing the highest brain tumor incidence of all children, even surpassing leukemia (Ostrom, de Blank, et al.,

2015; Ostrom, Gittleman, et al., 2015) (Fig. 1.1). Furthermore, the overall incidence of brain and

CNS tumors in children has increased between the years of 2000 and 2010 (Gittleman et al.,

2015), and it is estimated that 4,630 cases will be diagnosed in 2016 (Ostrom, Gittleman, et al.,

2015).

Despite advances in sequencing technologies and the advent of molecular/cellular-targeted therapies in oncology, there is a significant lag in the field of pediatric brain tumors where novel targeted therapies are yet to be adopted in the clinic. As a consequence, pediatric brain tumors have surpassed leukemias as the leading cause of cancer-related death in children (Patel et al.,

2014; Smith, Altekruse, Adamson, Reaman, & Seibel, 2014). Even comparison with adult brain tumors, which are distinct histological and molecular entities from their pediatric counterparts, shows that pediatric brain and CNS tumors have a higher incidence and mortality rate than adult brain tumors (Fig. 1.2). This necessitates a deeper molecular understanding of pediatric brain tumors that can guide development of novel biomarkers and targeted therapies.

Pediatric brain tumors represent a diverse group of diseases that are classified by the World

Health Organization (WHO) into large groups based on predominant cell type, location of tumors and level of malignancy. Based on the WHO classification, pediatric brain tumors mainly

1 comprise of glial tumors/gliomas (astrocytomas, oligodendrogliomas, oligoastrocytomas, ependymomas), germ cell tumors (GCTs), embryonal tumors of neuronal origin

(medulloblastoma, atypical teratoid rhabdoid tumors [ATRTs], CNS primitive neuroectodermal tumors [PNETs] and epithelial tumors with multiple rosettes [EMTRs]), and other rare tumors such as craniopharyngiomas and tumors of mixed neuronal-glial lineages (Louis et al., 2007)

(Fig. 1.3). Each tumor type has been found to have preference for site of occurrence, with majority occurring below a structure called the tentorium cerebeli that divides the forebrain from posterior fossa and hindbrain structures. Pediatric tumors are mostly found in infratentorial regions such as cerebellum, the optic pathway and the brainstem (Pollack, 1999). Recent large- scale sequencing studies are helping to add annotation based on mutational context to these diverse tumor types, beginning an era of molecular sub-classification, mutation-based diagnosis and identification of novel targets.

Gliomas are the most common primary tumors found in children (52.9%) and are further sub- divided on the basis of aggressiveness, grade I (benign) to grade IV (most malignant and aggressive). Further sub-grouping has resulted in two broad clinical and commonly known categories: low-grade gliomas (LGG, grades I and II) and high-grade gliomas (HGG, grades III and IV). Amongst these diverse tumor sub-types, we studied the molecular mechanisms driving pediatric low-grade gliomas (PLGGs) that are often considered ‘low-risk’ tumors but can lead to devastating long-term consequences in children.

Pediatric Low-Grade Gliomas

PLGGs are the most commonly diagnosed brain tumor in children and represent a heterogeneous group of histological entities, ranging from grade I pilocytic astrocytomas (PAs) to grade II diffuse astrocytomas (DAs), angiocentric gliomas (AGs), pleomorphic xanthoastrocytomas

(PXAs), and gangliogliomas (Louis et al., 2007). These tumors account for 30-50% of the CNS tumors in children and are characterized by their slow growth and less aggressive behavior

(Qaddoumi, Sultan, & Gajjar, 2009). Most PLGG patients have favorable outcomes with surgical resection but therapeutic challenges remain in cases with incompletely resected tumors in inaccessible brain regions such as the hypothalamus/optic pathway, brain stem and basal ganglia.

Overall, up to 20% of patients suffer from recurrent, progressive or inoperable disease (Sievert &

Fisher, 2009).

Adjuvant radiation and chemotherapy improve progression-free survival, however, these regimens are associated with significant long-term morbidities in nearly 65% cases due to toxic side effects on the developing CNS (Merchant, Conklin, Wu, Lustig, & Xiong, 2009; Sievert &

Fisher, 2009; Stokland et al., 2010; Taphoorn et al., 1994). Current chemotherapy regimens used in the clinic include carboplatin. vincistrine, vinblastine, cisplatin with etoposide, and temozolomide (Ater et al., 2012; Bouffet et al., 2012; Massimino et al., 2010). Survivors often suffer from late effects of both disease and treatment such as severe cognitive and neurological defects, risk of secondary cancers, neuro-endocrine disorders, infertility and other issues in adulthood. Hence, despite a 20-year overall survival of 87% in PLGG patients with current treatment options (P. Bandopadhayay et al., 2014), there is a large unmet need for combined molecular-histopathological diagnosis of these diverse tumors, and improved treatment plans that can target patient-specific mutations leading to better quality of life for survivors.

Landscape of genomic alterations in PLGGs

More than a decade ago, the first insight into the molecular/genetic alterations of PLGGs came from tumors arising due to genetic predisposition syndromes such as neurofibromatosis 1 (NF1) associated pilocytic astrocytomas (Gutmann, Donahoe, Brown, James, & Perry, 2000; Kluwe et

3 al., 2001; Lau et al., 2000) and tuberous sclerosis complex (TSC) related LGGs (called

Subependymal Giant Cell Astrocytoma or SEGA) (Shepherd, Scheithauer, Gomez, Altermatt, &

Katzmann, 1991). These findings highlighted aberration of canonical signaling pathways driven by somatic mutations in NF1 or TSC-associated PLGGs. Subsequently, a plethora of studies utilizing cytogenetics, single nucleotide polymorphism (SNP) arrays, Sanger sequencing, next- generation sequencing (whole-genome, exome and RNA sequencing) and epigenetic mapping have helped define the molecular/genetic underpinnings of diverse PLGGs that are not associated with genetic predisposition syndromes. These large-scale studies have discovered various mutations, gene rearrangements forming gene fusions, and copy number alterations that are distinct to specific tumor sub-types.

As shown in figure 1.4, we have categorized PLGG-associated mutations and gene fusions that have been discovered in the past few years by our group and others (D. T. Jones et al., 2009; D.

T. W. Jones et al., 2013). Such categorization has shown promise in associating the unique molecular profiles with clinical progression and therapeutic response in some PLGG sub-types

(mostly grade I PAs) but further research is required to delineate molecular mechanisms and response to therapy for grade II tumors. Furthermore, very few studies have explored the therapeutic response of distinct mutations within the same PLGG sub-type (Fig. 1.4) and current clinical trials fail to address such detailed molecular classification that could critically determine success/failure of PLGG targeted therapies.

BRAF mutations and gene fusions

Large-scale profiling studies over the past few years have found grade I PAs to have relatively stable genomes with limited genetic alterations that often lead to aberrant activation of the mitogen associated protein kinase (MAPK) pathway. NF1 associated PLGGs have biallelic

4 inactivation of the NF1 gene resulting in loss of neurofibromin protein. NF1 is known to act as a tumor suppressor by inhibiting RAS in the pathway, thereby resulting in aberrant MAPK activation in NF1 affected PAs (Gutmann et al., 2000; Kluwe et al., 2001; Lau et al., 2000) (Fig

1.5). In contrast to NF1 associated tumors that account for 15-20% of PAs, sporadic PAs were found to harbor somatic gene rearrangements involving the kinase domain of BRAF gene with a gene called KIAA1549, also activating the MAPK pathway (D. T. Jones et al., 2008; Pfister et al.,

2008). BRAF is an upstream kinase involved in the MAPK signaling pathway (Fig 1.5) and the

KIAA1549-BRAF fusion arises due to tandem duplication at 7q34 followed by fusion of 5’

KIAA1549 and 3’ BRAF. The fusion transcript can also activate the PI3K/AKT/mTOR pathway via crosstalk between the MAPK and PI3K/mTOR pathways (Kaul, Chen, Emnett, Dahiya, &

Gutmann, 2012). Almost 65% of sporadic PAs harbor the KIAA1549-BRAF fusion but nearly all

PAs display constitutive activation of the MAPK pathway via other mutations in the MAPK signaling axis (Jacob et al., 2009; D. T. Jones et al., 2008; Sievert et al., 2009) as described below. Interestingly, better clinical outcomes are correlated with the occurrence of KIAA1549-

BRAF in cerebellar PAs (Hawkins et al., 2011).

Next generation sequencing studies have discovered alternative BRAF-activating mutations that occur at a lower frequency compared to KIAA1549-BRAF in cerebellar PAs, such as novel BRAF gene fusions (FAM131-BRAF, FXFR1-BRAF, MACF1-BRAF, RNF130-BRAF, MKRN130-BRAF,

GNAI1-BRAF), CRAF/RAF1 gene fusions (QKI-RAF1, SRGAP3-RAF1, FYCO-RAF1) and

BRAFV600E mutations (Bacon, Endris, & Rappold, 2009; D. T. Jones et al., 2009; D. T. W.

Jones et al., 2013; J. Zhang et al., 2013) (Fig 1.6). These studies indicate that grade I PAs are primarily a single pathway disease where a single hit in the MAPK pathway is sufficient to induce neoplasia. Interestingly, studies have shown also that BRAF/other mutation driven MAPK

5 pathway activation can lead to oncogene induced senescence in LGGs (Jacob et al., 2011). This partly accounts for their indolent growth pattern and absence of malignant progression to HGGs.

Non-BRAF mutations and gene fusions

In contrast to the predominance of BRAF alterations in cerebellar PAs, non-cerebellar PAs are more commonly associated with FGFR1 and PTPN11 mutations, FGFR1-TACC1 fusions, and novel NTRK2 gene fusions (QKI-NTRK2 and NACC2-NTRK2) (Fig 1.6). These tumors occur in less surgically accessible brain regions and pose a major therapeutic challenge. Since FGFR and

NTRK are upstream receptors of the MAPK and PI3K/mTOR pathways, mutations and gene fusions can constitutively activate both signaling cascades (D. T. W. Jones et al., 2013).

Further histopathological annotations have classified another WHO grade I tumor called angiocentric glioma (AG) (Buccoliero et al., 2013). These epilepsy-associated, cerebral tumors are mostly found in children and young adults and alterations in the MYB gene (truncation of 3’ end and MYB-QKI gene fusion) have been found in a few cases of AGs (Ramkissoon et al., 2013;

J. Zhang et al., 2013). However there is no further information about the mechanism of action and specific correlation of MYB alterations with occurrence of AGs. We have addressed these questions revolving around MYB-QKI and our findings are described in chapter 3.

Grade II pediatric gliomas encompass diffuse astrocytomas (DAs), pleomorphic xanthoastrocytomas (PXAs) and ganglioglomas. While these tumors are rare in the pediatric population, they can lead to significant morbidity in patients due to occurrence at unresectable sites in the brain and having diffuse/infiltrative morphology, leading to incomplete surgical resection. Furthermore, DAs are found to harbor distinct genetic alterations as compared to

MAPK altered PAs. These include chromosomal duplication at 8q13.1 leading to truncation of negative regulatory domain of the transcription factor MYBL1 (Ramkissoon et al., 2013).

Approximately 25% of cerebral gliomas with a diffuse morphology harbor alterations in MYB or

MYBL1 whereas another 25% show FGFR1 duplication (J. Zhang et al., 2013) (Fig 1.6). FGFR1 mutated grade II tumors also depict MAPK and PI3K/mTOR alterations. Overall, MYB and

FGFR mutations are useful biomarkers and effective therapeutic targets for these clinically challenging tumors.

QKI gene fusions: MYB-QKI, QKI-RAF1 and QKI-NTRK2

In our analysis of whole genome sequencing (WGS), RNA sequencing or exome sequencing data, we also found the MAPK pathway to be ubiquitously affected in most PLGGs. Interestingly, we found non-BRAF gene fusions mediating changes in these well-known signaling pathways and activating other unique oncogenic mechanisms in PLGGs. In particular, we and others have identified the recurrence of the gene fusions MYB-QKI, QKI-RAF1 and QKI-NTRK2 that share the Quaking (QKI) fusion partner, an RNA binding protein and putative tumor suppressor, fused to known proto-oncogenes (D. T. W. Jones et al., 2013; J. Zhang et al., 2013) (Fig 1.7). We refer to this new class of gene fusions as Quaking (QKI) fusions and the recurrence of QKI as a fusion partner strongly indicates a role for QKI gene fusion mediated oncogenic processes in driving

PLGGs. This provides a unique platform for identifying the mechanisms by which QKI collaborates with distinct fusion partners to drive PLGG pathogenesis, thereby suggesting an emergent theme for their contribution to low-grade glioma biology. My thesis will discuss the oncogenic mechanism of action and therapeutic targeting of the MYB-QKI and QKI-RAF1 fusions.

QKI belongs to a family of RNA binding proteins known as the signal transduction and activation of RNA (STAR) proteins. These proteins share a single, highly conserved, maxi K- homology (KH)–RNA binding domain (RBD). Flanking the KH domain is the N-terminus

QUAKING1 (QUA1) domain, required for homodimerization (HD), and the C-terminus

QUAKING2 (QUA2) domain, necessary for RNA recognition and binding (Vernet & Artzt,

1997) (Fig 1.7). QKI isoforms (QKI 5/6/7) act at various levels of post-transcriptional gene regulation, including mRNA export, RNA stability, pre-mRNA splicing and protein translation, in several cell-specific contexts. The most well studied role of QKI has been its regulation of glial development and myelination in the CNS (Vernet & Artzt, 1997). Interestingly, alterations in

QKI expression are found in ~ 30% of HGGs (A. J. Chen et al., 2012) where QKI acts as a putative tumor suppressor gene. Furthermore, recent large-scale genomic analyses have revealed

QKI as a critical cancer gene lost in several cancer types (Lawrence et al., 2014). However, structural alterations of QKI seen in the MYB-QKI, QKI-RAF1 and QKI-NTRK2 gene fusions have not been previously described. My structural analysis of QKI-RAF1 and QKI-NTRK2 reveals that rearrangement of chromosome 6q results in a loss of QKI’s C-terminal Y-rich region in addition to loss of the QUA2 domain in QKI-RAF, whereas MYB-QKI loses QKI N-terminal regions, including RBD and HD domains (Fig 1.7). However, QKI retains the QUA1 region in

QKI-NTRK2 and QKI-RAF1 fusion proteins that has homo-dimerization properties.

MYB was first discovered as the cellular homolog (c-myb) of highly oncogenic v-myb found in the chicken leukemia viruses AMV and E26 (Gonda & Bishop, 1983; Klempnauer, Gonda, &

Bishop, 1982). This association between altered MYB and avian leukemias suggested a potential role for MYB as a proto-oncogene in human cancer. During normal development, MYB functions as a transcription factor that regulates proliferation and differentiation in specific cell lineages. In hematopoiesis, MYB plays a critical role in maintaining the undifferentiated state of hematopoietic stem cells (Sicurella et al., 2001). However, MYB expression declines postnatally in most differentiated tissues including the CNS, with some expression persisting in the neurogenic foci in adult brain (Malaterre et al., 2008). This implicates that MYB plays a role in

8 stimulating proliferation and suppressing differentiation. These functions are consistent with

MYB’s established role as a proto-oncogene in tumors. However, the mechanism of MYB dysregulation differs in different cancer types. In human leukemias and several epithelial cancers,

MYB is often deregulated via chromosomal translocations, activating truncations, and/or amplifications (Pelicci, Lanfrancone, Brathwaite, Wolman, & Dalla-Favera, 1984; Zhou & Ness,

2011). Provided that the N-terminus of MYB contains the DNA binding domain (DBD) and the

C-terminus contains a negative regulatory domain (NRD) (Fig 1.7), which interacts with the

DBD to auto-inhibit its function, alterations in the N or C terminus both enhance the transformative capacity of MYB. This has been evidenced by the up-regulation of MYB activity in cancers with C-terminally truncated MYB (Dash, Orrico, & Ness, 1996; Zhou & Ness, 2011).

In addition, loss of the 3’UTR of MYB, that harbors negative regulatory microRNA binding sites, has also been found in several cancers (Fehr et al., 2011; M. Persson et al., 2009). Interestingly, the MYB-QKI fusion leads to loss of MYB NRD (Fig 1.7) that could lead to constitutive activation of MYB in MYB-QKI. This observation, along with loss of QKI RBD, forms the basis of my hypothesis for Aim 1 of my thesis (chapter 2), as described below.

AIM 1: Investigate the oncogenic mechanism of action of MYB-QKI in PLGGs with a focus on independent and collaborative roles of MYB and QKI in mediating gliomagenesis. This aim will test the hypothesis that there is a gain-of-function of the MYB proto-oncogene via loss of its regulatory elements along with loss of tumor suppressor functions of QKI in a specific subset of PLGGs called angiocentric gliomas.

RAF-1 (or CRAF) is a serine-threonine kinase involved in the MAPK signaling cascade and is one of the three existing RAF isoforms (A-RAF, B-RAF, C-RAF). All three RAF proteins share similar protein domains, common mechanisms of activation and downstream effectors. Activation of up-stream RAS by tyrosine kinase auto-phosphorylation recruits RAF to the membrane where 9

RAF is activated and in turn activates the MAPK/ERK pathway by phosphorylating the

MAPK/ERK kinase (MEK) (Emerson et al., 1995; Moodie, Willumsen, Weber, & Wolfman,

1993; Nassar et al., 1995). The N-terminal domains of RAF1 physically inhibit the kinase activity of RAF1 (Cutler, Stephens, Saracino, & Morrison, 1998) and membrane recruitment by RAS relieves this negative regulation of RAF kinase activity (Fig. 1.5). The structural gene rearrangement in QKI-RAF1 leads to loss of N-terminal auto-inhibitory domains, thereby suggesting constitutive kinase activity. This observation, along with loss of portion of QKI RBD, forms the basis of my hypothesis for Aim 2 (chapter 3), of my thesis, as described below.

AIM 2: Investigate the oncogenic mechanism of action of QKI-RAF1 in PLGGs with a focus on independent and collaborative roles of QKI and RAF1 in mediating gliomagenesis. This aim will test the hypothesis that dimerization domains in truncated QKI contribute to constitutive homo/hetero-dimerization of QKI-RAF1, thereby causing activation of downstream MAPK and

PI3K pathway and oncogenic transformation.

These aims would explore an unaddressed avenue in the field of oncogenic gene fusions where usually one fusion partner is found to drive transformation. Our study will expose unique collaborative mechanisms between fusion partners that lead to pediatric brain tumors and could also provide mechanistic insight into the oncogenic mechanism mediated by gene fusions in adult cancers.

Targeted therapy for PLGGs: Need for molecular-based diagnosis and patient stratification for personalized medicine

Due to the tight association of specific mutations/gene fusions with PLGG tumor subtypes,

PLGGs are an excellent candidate for molecular diagnosis and targeted therapy. Some efforts towards this end and future possibilities are described below.

MAPK/PI3K/mTOR targeted inhibitors

Due to extensive involvement of the RAF proteins and subsequent MAPK/PI3K pathway activation in multiple cancers, several RAF and MAPK-targeted small molecule inhibitors have been developed in the past few years. Pharmacological intervention with RAF (vemurafenib and dabrafenib) and MEK (trametinib) inhibitors has significantly improved median survival of melanoma patients with BRAF alterations (Long et al., 2011). Despite its success against the constitutively active BRAF V600E mutation found in melanomas, when tested in PLGGs containing BRAF fusions (KIAA1549-BRAF) or MAPK pathway activation mutations other than

BRAFV600E, vemurafenib induced paradoxical activation (A. J. Sievert et al., 2013). This is due to stabilization of RAS-dependent RAF dimerization by RAF inhibitors (vemurafenib or its research analog PLX4720) in cells expressing wild-type RAF proteins (Cutler et al., 1998;

Hatzivassiliou et al., 2010; Heidorn et al., 2010; Poulikakos et al., 2011). Even a phase II study with sorafenib (multi-kinase inhibitor towards BRAF) for MAPK-altered PLGGs showed significant early tumor progression due to paradoxical activation of the MAPK pathway

(Karajannis et al., 2014). These findings highlight the therapeutic nuances of targeting PLGGs and how important it is to understand the distinct molecular mechanisms driving distinct tumor types, making universal therapy unattainable.

We have previously also shown that the KIAA1549-BRAF fusions act as dimers and lead to paradoxical activation of MAPK pathway with first generation RAF inhibitors, but are inhibited by paradox breaker RAF inhibitors such as PLX8394 (A. J. Sievert et al., 2013). Recently, pan-

RAF dimer inhibitors LY3009120 and BGB-283 have shown to be efficacious in RAF mutant tumors (Henry et al., 2015; Peng et al., 2015; Tang et al., 2015) but how these drugs will impact

RAF fusions in PLGGs, is yet to be known. In addition to targeting upstream in the MAPK and

PI3K pathway, downstream inhibitors are also being evaluated in clinical trials. Phase I/II studies

11 using the MEK1/2 inhibitor Selumetinib (AZD6244) for children with refractory/recurrent low- grade gliomas is currently underway (ClinicalTrials.gov Identifier: NCT01089101 and

NCT01386450). Another phase II study is evaluating efficacy of single-agent Everolimus (or

RAD001, an mTOR2 inhibitor) for progressive pediatric low-grade gliomas (ClinicalTrials.gov

Identifier: NCT01734512). These studies are the first to test targeted inhibitors on children with

PLGGs and assess benefit over standard therapies. However, given the diversity in molecular aberrations, such clinical studies need to be combined with comprehensive mutation screening to ensure that PLGG patients are receiving personalized and effective therapy.

Novel therapeutic avenues: Epigenetic modulators

Epigenetic modifications are genomic changes that do not directly alter the DNA sequence, including different mechanisms to affect the chemical properties of DNA or DNA packaging proteins (histones). Oncogenic alterations in epigenetic modulators have been described for several adult and pediatric tumors and are expanding the therapeutic and diagnostic toolkit for human cancers.

Pediatric HGGs (Grade IV glioblastoma) display a range of somatic mutations in the histone 3

(H3) variant H3.3 encoded by the H3F3A gene, such as the K27M and G34R/V mutations

(Schwartzentruber et al., 2012; G. Wu et al., 2012). K27M mutations in H3.3 and another histone variant H3.1 (HIST1H3B) are found in a large proportion of DIPGs, another devastating PHGG with a dismal 5-year survival of less than 1% (Khuong-Quang et al., 2012; G. Wu et al., 2012).

Recent studies have demonstrated, for the first time, some efficacy of epigenetic therapy in

DIPGs. The multi-histone deacetylase (HDAC) inhibitor panobinostat showed surprising efficacy in suppressing in vitro and in vivo DIPG mice xenograft growth (Grasso et al., 2015). Single- agent therapy with the histone demethylase inhibitor GSK-J4 can suppress DIPG cell lines

(Hashizume et al., 2014) but combination with panobinostat has shown synergistic inhibition of 12

DIPG patient-derived lines (Grasso et al., 2015). Such epigenetic modifying therapies are opening the preclinical realm for PHGGs where no targeted therapies exist.

PLGGs have been found to harbor H3F3A mutations in very rare cases of DAs and PAs (D. T. W.

Jones et al., 2013) with no other predominant epigenetic mutations. Furthermore, a group very recently found pilocytic astrocytomas (PAs) to have distinct methylation signatures with hypomethylation at multiple CpG sites, thereby suggesting uncontrolled gene transcription

(Jeyapalan et al., 2016). These findings suggest epigenetic modifications to precede gene mutation/chromosomal rearrangement events that could then drive aberrant expression of PLGG- specific oncogenes. As such, epigenetic targeting could hold some promise for PLGGs but this needs to be explored further. Our findings in aim 1 (chapter 2) with MYB-QKI will provide some mechanistic insight into how epigenetic modification can mediate PLGGs.

Tables and Figures

Figure 1.1

<1 year old 1-4 years old

5-9 years old 10-14 years old

Figure 1.1. Comparison of incidence rates of various childhood cancers by age. Plots showing average annual age-related incidence of all primary brain and CNS tumors in comparison to other cancers (leukemias, lymphomas and other solid tumors) in children. (a) Infants (<1 Year Old), (b) Children 1–4 Years, (c) Children 5–9 Years, and (d) Children 10–14 Years (source CBTRUS 2007–2011, USCS 2007–2011)(Ostrom, de Blank, et al., 2015). Figure adapted from Quinn T. Ostrom et al. Neuro Oncol 2015.

Figure 1.2

Figure 1.2. Comparison of incidence rates and mortality rates between adult and childhood cancers. Plots showing average annual age-related incidence of all primary brain and CNS tumors in comparison to other cancers in (a) adults (+20 years old) and (b) children and adolescents (0-19 years old) and mortality rates of all primary brain and CNS tumors as compared to other common cancers in c) adults (age 20+ years) and d) children and adolescents (age 0-19 years), CBTRUS statistical report: npcr and seer 2008-2012,USCS2008-2011b,NCVS 2008-2012 (Ostrom, de Blank, et al., 2015). Figure adapted from Quinn t. Qstrom et al. Neuro oncol 2015.

Figure 1.3

Figure 1.3. Classification of pediatric brain tumors by histology. Pie chart showing distribution of all primary brain and CNS tumors by histology groupings in patients 0–14 Years, N = 16,044 (CBTRUS 2007–2011) (Ostrom, de Blank, et al., 2015). Figure adapted from Quinn T. Ostrom et al. Neuro Oncol 2015.

Figure 1.4

Figure 1.4. Grade I and Grade II PLGGs classified on the basis of genetic/molecular alterations. Mutations grouped into BRAF, FGFR, CRAF/RAF1, NTRK and MYB families based on known oncogenes involved. Family of mutations can be further grouped within grade I and grade II astrocytomas as low-grade tumor sub-types have distinct molecular signatures. AG- angiocentric glioma, PXA- pleomorphic xanthoastrocytoma, GG- ganglioglioma.

Figure 1.5

Figure 1.5. Simplified linear representation of the RAS/RAF/MEK/ERK (MAPK) pathway. Upon receiving upstream signal from a growth factor activated receptor kinase, the RAS protein is activated and leads to further stimulates downstream RAF kinases. NF1 serves to negatively regulate RAS by preventing RAS activation leading to inhibition of the MAPK/ERK pathway. Activated RAS proteins recruit RAF kinases to the membrane that leads to their dimerization and consequent activation. RAF kinases then phosphorylate downstream MEK1/2 kinases, which then can phosphorylate ERK. The MAPK pathway eventually leads to transcriptional changes in the nucleus that leads to cell proliferation and growth.

Figure 1.6

Figure 1.6. Breakpoints in BRAF forming in-frame gene fusions in PLGGs. Linear structure of BRAF gene on chromosome 7q34 showing breakpoints at exon 9 (KIAA1549-BRAF, FAM131B-BRAF, MACF1-BRAF and RNF130-BRAF), exon 10 (KIAA1549-BRAF, FXR1-BRAF and GNAI-BRAF) and exon 11 (KIAA1549-BRAF, CLC6-BRAF and MKRN1-BRAF) to form the respective oncogenic gene fusions in PLGG patients. The conserved region 3 is retained in all BRAF fusions (shown in green). BRAF point mutations (R506VLRK507, R509H and V600E) are depicted in the ATP binding domain and activation segment of the kinase.

Figure 1.7

Figure 1.7. Functional protein domains of (A) wild type QKI, MYB, RAF1 and NTRK2 proteins, (B) QKI fusions found in PLGG patient samples, MYB-QKI5/6, QKI-RAF1 and QKI- NTRK2 and the retained protein domains in truncated QKI and truncated MYB/ RAF1/ NTRK2. DBD- DNA binding domain, TAD-trans-activation domain, NRD-negative-regulatory domain, TM-transmembrane domain.

The data in the following chapter are published in Nature Genetics 2016 Mar; 48(3):273-82. doi: 10.1038/ng.3500. The Nature Publishing Group allows for authors to own copyright to their primary research and data, so no permission was required from publisher, Kyle Vogan.

CHAPTER 2: MYB-QKI rearrangements in angiocentric gliomas drive oncogenicity through a tripartite mechanism

Summary

Angiocentric gliomas are pediatric low-grade gliomas (PLGGs) without known recurrent genetic drivers. We performed genomic analysis of new and published data from 249 PLGGs including

19 Angiocentric Gliomas. We identified MYB-QKI fusions as a specific and single candidate driver event in Angiocentric Gliomas. In vitro and in vivo functional studies show MYB-QKI rearrangements promote tumorigenesis through three mechanisms: MYB activation by truncation, enhancer translocation driving aberrant MYB-QKI expression, and hemizygous loss of the tumor suppressor QKI. This represents the first example of a single driver rearrangement simultaneously transforming cells via three genetic and epigenetic mechanisms in a tumor.

Introduction

Pediatric low-grade gliomas (PLGG) encompass a heterogeneous group of World Health

Organization grade I and II tumors that collectively represent the most common pediatric brain tumor. PLGGs undergo frequent alterations in the MAPK pathway and in MYB family genes, including MYBL1 and MYB (Ramkissoon et al., 2013; J. Zhang et al., 2013). Alterations in MYB are heterogeneous; several fusion partners have been reported as rare events in PLGGs

(Ramkissoon et al., 2013). The frequency of specific alterations and associations with histological subtypes are unknown.

Angiocentric Gliomas arise in the temporal lobe and share histologic features with astrocytomas and ependymomas (Lellouch-Tubiana et al., 2005; M. Wang et al., 2005). We previously identified one Angiocentric Glioma with deletion of the 3’ region of MYB (Ramkissoon et al.,

2013) and one other Angiocentric Glioma has been reported to harbor a MYB-QKI rearrangement

(J. Zhang et al., 2013). However, the nature and incidence of MYB alterations in Angiocentric

Glioma has not been determined. Furthermore, oncogenicity of MYB family transcription factors in the CNS and the mechanisms by which they contribute to gliomagenesis are yet to be defined.

To address these questions, we performed a combined analysis of newly generated and published

PLGG genomic datasets (D. T. W. Jones et al., 2013; Ramkissoon et al., 2013; J. Zhang et al.,

2013). We found MYB-QKI rearrangements to be the most common event involving a MYB family member and to be specific to Angiocentric Gliomas. We also found that this rearrangement contributes to oncogenicity through three mechanisms: generation of oncogenic

MYB-QKI, enhancer translocation that establishes an auto-regulatory feedback loop selectively driving MYB-QKI expression, and partial loss of expression of QKI, a tumor suppressor gene.

Results

Angiocentric Gliomas exhibit recurrent MYB-QKI rearrangements

Previously published genomic analyses of PLGGs did not individually contain sufficient numbers of rare histologic subtypes for statistical power to detect recurrent aberrations. To address this we performed a combined genomic analysis of whole-genome sequencing (WGS) and/or RNA- sequencing (RNA-seq) data from 172 PLGGs spanning ten histologic subtypes (Table 2.1), including 145 published samples (D. T. W. Jones et al., 2013; J. Zhang et al., 2013) and 27 rare

PLGGs that are new to this study. We performed analyses of significantly recurrent somatic genetic events across all samples with WGS or RNA-seq data (Fig. 2.1 and Tables 2.2 and 2.3).

We observed recurrent somatic alterations in 154 tumors (90%), including all 140 tumors subject to WGS. Rearrangements or structural alterations were observed in 129 tumors (83%; Fig. 2.2a,

Table 2.1). 23

Rearrangements involving MYB family members (MYB, MYBL1) were the second-most recurrent alteration, affecting 16 tumors (10%), predominantly Diffuse Astrocytomas and Angiocentric

Gliomas (Fig. 2.1 and 2.2a). Six of seven Angiocentric Gliomas, including all tumors subject to central pathology review, exhibited intra-chromosomal deletions resulting in MYB-QKI rearrangements. The other Angiocentric Glioma, which was not centrally reviewed, contained a

MYB-ESR1 rearrangement.

Although MYB rearrangements have been described in PLGGs (Ramkissoon et al., 2013; J. Zhang et al., 2013), we were struck by two novel findings: QKI was the most frequent fusion partner, and MYB-QKI fusions were near-universal in Angiocentric Gliomas. For validation we identified studied 12 additional Angiocentric Gliomas with only FFPE tissue using targeted assays. Nine

Angiocentric Gliomas were analyzed by FISH to detect MYB rearrangement or deletion (Fig.2.

2b), and three Angiocentric Gliomas were analyzed by WES and/or aCGH (Fig. 2.3). All 12 harbored MYB aberrations.

In total, all 19 Angiocentric Gliomas profiled by WGS, RNA-seq, WES, FISH, or aCGH displayed MYB alterations, and in six of the seven cases in which its fusion partner could be detected, MYB was fused to QKI. In tumors confirmed to harbor MYB-QKI, the genetic event appeared to be present in the majority of cells, although evidence of heterogeneity (aberration in

~50% of tumor cells) was observed by FISH in 2/5 tumors with sufficient cells for quantitative scoring.

MYB-QKI rearrangements appeared specific to Angiocentric Glioma. None of the 147 non-

Angiocentric Gliomas profiled with WGS or RNA-seq exhibited MYB-QKI fusions (p<0.0001,

Fig. 2.2c). We also evaluated MYB alterations in an additional 65 PLGGs from two separate cohorts: 10 non-Angiocentric Gliomas analyzed by FISH and 55 non-Angiocentric Gliomas

24 evaluated by whole-exome sequencing (WES) and/or array CGH. Only one of these tumors exhibited alterations of MYB (vs 19/19 Angiocentric Gliomas; p<0.0001) (Fig. 2.3 and Table

2.1). This tumor was designated not-otherwise-specified on research review but had been diagnosed as Angiocentric Glioma at the referring institution. Five tumors evaluated by WES or aCGH exhibited alterations of MYBL1; these were all Diffuse Astrocytomas. The FISH assays, aCGH, and WES, though able to detect MYB alterations, were unable to characterize its fusion partners.

All MYB-QKI rearrangements had breakpoints within intron 4 of QKI while the MYB breakpoint varied from intron 9 to 15; all were predicted to express an in-frame fusion protein MYB-QKI

(Fig. 2.2d). We identified fusion mRNA transcripts by RNA-seq (Fig. 2d) and observed copy- number breakpoints in these genes from WGS data (Fig. 2.2e).

In the WGS/RNA-seq cohort we also observed rearrangements involving QKI but not MYB in three supratentorial Pilocytic Astrocytomas (PAs), and rearrangements involving MYB or MYBL1 but not QKI in nine tumors, seven of which were Diffuse Astrocytomas. Across the entire cohort of 172 tumors profiled with WGS and/or RNA-seq, 10% harbored alterations of either MYB family members or QKI.

MYB and QKI have contrasting roles in cortical brain

MYB proteins are transcription factors characterized by highly conserved DNA-binding motifs.

First identified as v-myb (Klempnauer, Bonifer, & Sippel, 1986; Klempnauer et al., 1982; Ness,

Marknell, & Graf, 1989) the cellular proto-oncogene counterpart c-MYB is comprised of a N- terminus that contains helix-turn-helix (HTH) DNA binding motifs followed by a transcriptional activation domain and a C-terminal negative regulatory domain (Sakura et al., 1989). Full-length

MYB is non-transforming or only weakly transforming in vitro (Gonda, Buckmaster, & Ramsay, 25

1989), but C-terminal MYB truncations are oncogenic (Gonda et al., 1989; Grässer, Graf, &

Lipsick, 1991; Y. L. Hu, Ramsay, Kanei-Ishii, Ishii, & Gonda, 1991; Press, Reddy, & Ewert,

1994). MYB-QKI breakpoints in MYB intron 9 to 15 are predicted to result in C-terminal truncation of MYB.

MYB is not expressed in the postnatal brain cortex, where Angiocentric Gliomas occur. We examined RNA-seq data of normal tissues (G. Consortium, 2013) and found MYB expression to be negligible in human brain cortex and substantially lower than MYB expression in colon, breast, blood, esophagus, or skin (Fig. 2.4a). Likewise, immunohistochemistry of adult human frontal cortex and white matter were negative for MYB (Fig. 2.4b and 2.4c); however we detected high

MYB expression in human fetal neural progenitor cells generated from the ganglionic eminence at 22 weeks gestation (Fig. 2.4d and 2.4e).

In mice MYB is expressed in E14.5 neural progenitor cells of the ganglionic eminence subventricular region (Fig. 2.4f-i). In adult mice we detected expression in the ependyma/subventricular zone (Fig. 2.4j-k), consistent with previous reports of MYB expression in mouse progenitor cells but not in cortical brain (Malaterre et al., 2008).

QKI encodes the STAR (Signal transduction and activation of RNA) RNA-binding protein

Quaking, which plays an essential role in oligodendroglial differentiation (Friedrich, 1974) and is widely expressed in the nervous system. Deletions of QKI have been suggested to be oncogenic in a number of cancers including glioblastoma (Yin et al., 2009), prostate cancer (Y. Zhao et al.,

2014), and gastric cancer (Bian et al., 2012). In copy-number analyses of 10,570 cancers within the Cancer Genome Atlas (Zack et al., 2013), QKI was one of two genes in a deletion peak in adult glioblastomas (Fig. 2.4l), renal clear cell, and cervical squamous cell carcinomas. It was

26 also in larger peak regions of significant deletion in low-grade gliomas and bladder and adrenocortical carcinomas. Focal QKI deletions were observed in over 10% of glioblastomas.

The MYB-QKI fusion protein is expected to retain the MYB N-terminal HTH DNA binding motifs fused to the QKI C-terminus (Fig. 2.4m). The QKI N-terminal KH RNA-binding motif is lost, while C-terminal alternative splice sites are preserved. The splice variant MYB-QKI5 retains a nuclear localizing motif which is not present in the splice variant MYB-QKI6 (J. Wu, Zhou,

Tonissen, Tee, & Artzt, 1999). Fusions that contain only exons 1-9 of MYB also lose the MYB negative regulatory domain (designated short variant).

The findings that both MYB and QKI are disrupted suggest that MYB-QKI rearrangements may be oncogenic through the additive effects of alterations in both MYB and QKI. The lack of expression of MYB in normal post-natal human cortical brain regions also suggests that the rearrangement drives aberrant expression of the fusion allele. We therefore characterized mechanisms through which MYB-QKI rearrangements may contribute to aberrant MYB-QKI expression and evaluated the oncogenic potential of both genes.

MYB-QKI functions as a transcription factor

We performed genome-wide gene expression analyses of three independently-generated pools of mouse neural stem cells (mNSCs) engineered to stably over-express MYB-QKI5, MYB-QKI6, truncated MYB including exons 1-9 (MYBtrExon1-9), or eGFP. Relative to eGFP-expressing cells, those expressing MYB-QKI5 and MYB-QKI6 exhibited significantly different expression of

1621 and 1947 genes, respectively, with 1029 genes overlapping (p<0.0001; Table 2.4). Gene-set enrichment analysis revealed expression of either MYBtrExon1-9 or MYB-QKI was associated with enrichment of signatures of MYB pathway activation (p<0.0001, Table 2.5).

We defined a MYB-QKI gene expression signature comprising the 50 genes whose differential expression correlated most with its expression (Fig. 2.5a). These genes include KIT and CDK6, previously reported to be associated with MYB activation (Gao et al., 2014).

We performed chromatin immunoprecipitation with parallel sequencing (ChIP-seq) in mNSCs expressing MYB-QKI, using an antibody which recognizes the N’ terminus of MYB and another antibody against H3K27ac, which defines the location of enhancer regions. We found MYB-

QKI5 bound 3,672 sites across the genome (92% of these sites contain a MYB binding motif) and

H3K27ac bound 9,122 sites, with overlap at 1,907 sites (52% of MYB binding sites, p<0.0001)

(Fig. 2.5b). These findings are consistent with reports in T-cell ALL, where MYB binding was highly correlated with H3K27ac defined enhancers (Mansour et al., 2014). We also identified

MYB-QKI binding to the endogenous Myb promoter (Fig. 2.6a).

MYB-QKI5 binding sites were located within 100kb of 88% (22/25) of upregulated genes in the

MYB-QKI signature but only 40% (10/25) of downregulated genes (p<0.001; Fig. 2.5c). Each of the MYB-QKI binding sites associated with an upregulated gene was associated with an

H3K27ac enhancer peak, while only 70% of MYB-QKI binding sites at downregulated genes overlapped enhancers (p=0.003).

The MYB-QKI fusion protein can activate transcription through binding of MYB consensus binding motifs. We generated a luciferase-reporter construct using known MYB binding sites from the target promoter mim-1 (Ness et al., 1989), and co-transfected this reporter with

MYBtrExon1-9, MYB-QKI, or full-length MYB, in 293T cells. We observed a slight induction of mim-1 promoter activity with transfection with full length MYB compared to the control vector.

The greatest induction of mim-1 promoter activity was observed upon co-transfection with

MYBtrExon1-9 or MYB-QKI (Fig. 2.5d and 2.6b), with MYBtrExon1-9 having the highest level of activity.

Angiocentric Gliomas exhibited significantly higher expression of the MYB-QKI signature relative to normal pediatric brain (p=0.001) and PLGGs without MYB-QKI alterations

(p=0.0011) (Fig. 2.5e). PLGGs exhibited increased expression of genes associated with MYB pathway activation compared to normal brain (p=0.0003), but this was not specific to MYB-QKI rearranged tumors, and was of lower magnitude than the difference observed with the MYB-QKI signature (Fig. 2.6c).

MYB-QKI rearrangements drive aberrant expression of truncated MYB

Angiocentric Gliomas with MYB-QKI exhibit significantly higher MYB expression relative to normal pediatric cortical brain (p=0.0062) or to PLGGs with BRAF or FGFR alterations (p=0.03)

(Fig. 2.7a). The MYB that is expressed is truncated and corresponds to the exons retained in the rearranged MYB-QKI allele. Three Angiocentric Gliomas harbored MYB-QKI rearrangement breakpoints between exons 9 and 10 of MYB. These exhibit increased expression of MYB exons

1-9 relative to PLGGs that do not harbor MYB-QKI (p<0.05), but minimal expression of the remaining exons (Fig. 2.7b). These data support the selective, aberrant regulation of expression of truncated MYB via MYB-QKI.

The MYB-QKI rearrangement results in enhancer translocation

Aberrant oncogene expression can result from enhancer translocation (Northcott et al., 2014). In published H3K27ac enhancer profiles from normal human cortical brain samples (E. P.

Consortium, 2012), MYB is not associated with H3K27ac enhancer peaks, consistent with the finding that MYB is not expressed. In contrast, QKI, which is expressed, is associated with several

H3K27ac peaks, including sequences at the 3’ end of QKI (Fig. 2.7c, d, e). The MYB-QKI 29 rearrangement is predicted to bring these 3’ QKI-associated H3K27ac enhancer elements to within only 15kb of the MYB promoter (Fig. 2.7e).

H3K27ac enhancer profiling of two human Angiocentric Gliomas expressing MYB-QKI confirmed the presence of active enhancer elements that are translocated proximally towards the

MYB promoter (Figs. 2.8a and 2.9). ChIP-seq revealed multiple H3K27ac peaks associated with

3’ QKI, similar to the peaks observed in normal human brain, and in a BRAF-duplicated supratentorial pilocytic astrocytoma. We also observed enhancers within 10kb of 3’ QKI and a larger cluster of super-enhancers 100-500kb 3’ to QKI (Q3SE1 and Q3SE2). In Angiocentric

Gliomas with MYB-QKI, these enhancers are translocated proximally towards the MYB promoter.

We observed an aberrant enhancer associated with the MYB promoter in MYB-QKI defined

Angiocentric Glioma (Fig. 2.8a). Normal human cortical brain is not associated with H3K27ac

MYB-related enhancers, and indeed we did not observe formation of H3K27ac MYB enhancer peaks in the Pilocytic Astrocytoma (Fig. 2.10). However, in both Angiocentric Gliomas, we observed a large H3K27ac peak associated with the MYB promoter (M5E1). RNA-seq revealed expression of the first nine exons of MYB corresponding to those retained in the rearrangement, suggesting that the aberrant M5E1 enhancer is regulating expression of truncated MYB from the rearranged allele. The lack of full-length MYB expression indicates the aberrant enhancer does not regulate the remaining wild-type MYB allele.

We examined whether MYB-QKI was able to functionally activate the MYB promoter by creating a luciferase-reporter construct possessing the human MYB-promoter (MYB-luc). We observed significant induction of MYB promoter activity in U87 cells stably expressing MYB-QKI with

MYB-luc as compared to U87 cells containing MYB-luc or the promoter-less control luciferase construct alone (Fig. 2.8c). This suggests MYB-QKI contributes to an auto-regulatory feedback

30 loop, possibly by binding to the MYB promoter. MYB-QKI activated the MYB promoter in two additional cellular contexts (HEK 293T and NIH-3T3 cells, Fig. 2.11).

We predicted that enhancers in the QKI 3’ UTR could aberrantly activate the MYB promoter when translocated, thereby further driving MYB-QKI expression. We cloned the proximal QKI

3’UTR enhancer sequence (Q3E1) upstream of the human MYB promoter in the MYB-luc construct. Baseline activity of the Q3E1-MYB-luc promoter construct was higher than with MYB- luc alone in U87 glioma cells (Fig. 2.8c), increasing activation by approximately 1.5 fold, a level of activation shown to harbor biological relevance in other diseases (Menzel et al., 2007).

Expression of MYB-QKI with Q3E1-MYB-luc led to even higher activity, again consistent with an auto-regulatory feedback loop in the presence of the fusion protein (Fig. 2.8c).

The MYB-QKI fusion protein is oncogenic

Expression of truncated MYB has previously been reported to be oncogenic (Gonda et al., 1989;

Y. L. Hu et al., 1991; Press et al., 1994). In mNSCs, overexpression of MYB exons 1-9 (short variant) increased cell proliferation rates compared to eGFP controls (Figs. 2.12a and 2.11b), while in NIH-3T3 cells overexpression of MYBtr (exons 1-15), but not full-length MYB, induced tumors when injected into mouse flanks (Figs. 2.12b and 2.11b). Furthermore mNSCs expressing MYBtr induced diffuse gliomas on average 100 days post intracranial injection

(Figure 6e, f). These tumors expressed OLIG2 and GFAP in a subset of tumor cells, a pattern similar to that observed in human diffuse gliomas (Fig. 2.11c).

To test whether MYB-QKI fusions are oncogenic, we stably expressed MYB-QKI5 and MYB-

QKI6 in mNSCs and NIH3T3 cells. In mNSCs, overexpression of either isoform led to significantly increased proliferation compared to eGFP (Figs. 2.12c and 2.11b), p<0.0001.

Similarly both isoforms induced anchorage-independent growth in NIH-3T3 cells (Fig. 2.13a); in 31 vivo, overexpression of both MYB-QKI5 and MYB-QKI6, but not full-length MYB, led to tumor formation as flank xenografts (Fig. 2.12b). Intracranial injections of mNSCs overexpressing

MYB-QKI5 or MYB-QKI6 formed gliomas with infiltrating tumor cells with some evidence of enhanced growth around vessels and a clustered growth pattern, features similar to Angiocentric

Glioma and distinct from the histology seen adult glioblastoma models (e.g.

Ink4a/ARF:EGFRvIII) (Bachoo et al., 2002). However these tumors differed from human

Angiocentric Gliomas in that they had high-grade features with frequent mitoses and marked cytologic atypia (Fig. 2.12e). Immunohistochemical analysis showed diffuse GFAP expression and a subset of OLIG2 positive tumor cells, a pattern similar to that seen in human Angiocentric

Gliomas (Fig. 2.11c).

In total, we established flank injections in 15 mice with NIH-3T3 cells over-expressing either

MYBtr or MYB-QKI (and five vector controls), and 29 intracranial injections of mNSC expressing MYBtr or MYB-QKI (15 vector controls) (Fig. 2.13b). We observed flank tumors in all 15 mice injected with NIH-3T3 cells over-expressing either MYBtr or MYB-QKI and five intracranial tumors from mice injected with mNSCs expressing MYBtr or MYB-QKI. We did not observe tumors in any vector controls. These data represent a significant enrichment of tumor formation in cells expressing MYBtr or MYB-QKI (p<0.0001).

The MYB-QKI rearrangement disrupts QKI, a tumor suppressor

We were interested in understanding how disruption of QKI may contribute to oncogenicity in tumors that harbor MYB-QKI. Exon-specific RNA-seq analysis of Angiocentric Gliomas with

MYB-QKI (n=4) showed reduced expression of QKI compared to PLGGs that harbor BRAF alterations (n=5) (Fig. 2.14a). These data suggest that the MYB-QKI rearrangement may contribute to tumor formation through reduced expression of QKI, a tumor suppressor gene.

Indeed, suppressing wild-type Qk using shRNAs that target the first four exons of Qk led to increased proliferation of mNSCs, with the greatest increase observed in the context of pre- existing MYB-QKI expression. In mNSCs over-expressing MYBtr, MYB-QKI5 or MYB-QKI6, suppression of wild type Qk was sufficient to increase proliferation within only three days of suppression (Fig. 2.14b and 2.15a). The greatest effect was observed in cells overexpressing

MYB-QKI, despite a similar or lower degree of suppression of Qk in these cells compared to those over-expressing eGFP or MYBtr. We did not observe increased proliferation within three days in cells expressing eGFP, though we did observe a mild increase on day 5 (Fig. 2.15b).

These data suggest MYB-QKI overexpression and QKI suppression exert cooperative functional effects.

Suppression of Qk by shRNAs in mNSCs expressing MYB-QKI6 led to differential expression of

309 genes relative to shLacZ (q<0.25, Table 2.6). QKI has been previously reported to regulate expression of micro-RNAs (A. J. Chen et al., 2012; Y. Wang, Vogel, Yu, & Richard, 2013), and we also observed upregulation of 10 miRNAs with suppression of wild-type Qk, including

Mir717 (Table 2.7). The mouse Qk isoform 7 is predicted to contain a miRNA regulatory element (MRE) for Mir717 (Ji et al., 2013).

Angiocentric Gliomas exhibit molecular effects consistent with QKI suppression. In mNSCs, we defined a signature consisting of the 50 genes whose expression was most correlated with Qk suppression (Fig. 2.15c). This signature was significantly enriched in Angiocentric Gliomas relative to normal brain (p<0.0001, Fig. 2.14c).

Taken together, our data suggest three mechanisms through which the MYB-QKI rearrangement contributes to oncogenicity (Fig. 2.16). First, the alteration results in proximal translocation of

H3K27ac enhancers on 3’QKI towards the MYB promoter, resulting in MYB promoter activation.

Second, the MYB-QKI fusion protein that is expressed is oncogenic, functions as a transcription factor, and exhibits the ability to bind to and activate the MYB promoter, resulting in an auto- regulatory feedback loop. Third, hemizygous loss of QKI results in suppression of QKI, which functions as a tumor suppressor gene. Thus disruption of both MYB and QKI appear to contribute to tumor formation in a co-operative manner.

Discussion

We describe MYB-QKI as a novel recurrent diagnostic fusion in Angiocentric Glioma. It also represents the first example of a single driver translocation of two genes resulting in the aberrant expression of an activated oncogenic fusion protein which then participates in an auto-regulatory feedback loop, proximal translocation of enhancer elements regulating fusion-gene expression, and simultaneous functional loss of a tumor suppressor gene.

We found MYB-QKI to be a defining event in Angiocentric Glioma. This has important implications for treatment and diagnosis of this disease. The tight association of the translocation with this histology supports pathologic classification of Angiocentric Glioma as a separate biological entity. We propose that the presence of this fusion should be considered diagnostic of

Angiocentric Glioma. This could aid in distinction of Angiocentric Glioma from tumors with higher potential for recurrence or require further treatment, such as IDH-mutant diffuse gliomas or ependymomas.

MYB-QKI expression was sufficient to reproducibly generate intracranial tumors. Angiocentric

Gliomas are WHO grade 1 tumors and exhibit a very low mitotic index. Successful models of low-grade pediatric or adult gliomas are rare across all histologies. The low penetrance in mNSCs

(5 tumors in 29 attempts) relative to high-grade glioma models (e.g. EGFRvIII and Ink4a/Arf -/-

NSCs) (Bachoo et al., 2002) suggests genetic drivers of low-grade gliomas may be borderline sufficient for transformation.

Although the development of small molecule inhibitors to target MYB directly is likely to be challenging, MYB-QKI transcriptional targets such as KIT or CDK6 can be targeted. The association of MYB-QKI with H3K27ac enhancer elements also raises the possibility of therapeutically inhibiting its effects through indirect mechanisms, such as BET-bromodomain

(Pratiti Bandopadhayay et al., 2014; Delmore et al., 2011) or CDK7 inhibition (Chipumuro et al.,

2014).

Adenoid cystic carcinomas harbor MYB-NFIB alterations (D'Alfonso et al., 2014; Marta

Persson et al., 2009; Stephens et al., 2013). Like MYB-QKI, MYB-NFIB also results in high levels of MYB expression, although the mechanism underlying this, the functional role of NFIB, and oncogenicity of MYB-NFIB remain undefined.

We observed an additive effect of Qk suppression with MYB-QKI over-expression, confirming

QKI as a tumor suppressor and suggesting cooperation between its loss and expression of MYB-

QKI. Recent studies indicate a diversity of roles for QKI in cancer, including altered splicing events (Danan-Gotthold et al., 2015) and a role in epithelial-mesenchymal transition via regulation of circular RNAs (Conn et al., 2015). QKI regulates expression of genes implicated in cancer (Lu et al., 2014), and microRNA processing (Ji et al., 2013). Further investigation is required to evaluate mechanisms of cooperativity between QKI suppression and MYB-QKI expression.

Angiocentric Gliomas exhibit haploinsufficiency of QKI while GBMs demonstrate biallelic loss.

One explanation may be that complete loss of QKI exerts negative selection in the developing brain. This is supported by the essential role of QKI in oligodendrocytic differentiation. 35

Haploinsufficiency may also account for the lower-grade nature of Angiocentric Glioma compared to GBM.

Pediatric tumors are characterized by simple genomes with single driver alterations (Crompton et al., 2014; Kieran et al., 2012). Our findings that one rearrangement contributes to oncogenicity through multiple mechanisms may be applicable to a large number of pediatric tumors.

Materials and Methods

Ethics statement

Ethics approval was granted by relevant human IRB and/or animal research committees (IACUC) of Dana-Farber Cancer Institute (DFCI), Boston Children’s Hospital, The Broad Institute and

Children’s Hospital of Philadelphia (CHOP). IRB approval from all institutions was obtained, and all patients provided informed consent prior to collection of samples or were analyzed as de- identified samples with specific IRB waiver of informed consent.

Whole-genome sequencing and processing

PLGGs and normal controls from CBTTC/CHOP and DFHCC/PLGA Consortium were sequenced at BGI@CHOP, and The Broad Institute of MIT and Harvard. DNA was randomly fragmented, and libraries prepared for paired-end sequencing on an Illumina HiSeq 2000.

Sequencing files from recently published PLGG datasets were accessed (D. T. W. Jones et al.,

2013; J. Zhang et al., 2013). Read pairs were aligned to reference genome hg19 (Build 37) using the Burrows-Wheeler Aligner (bwa) with options −q 5 −l 32 −k 2 −o 1 (H. Li & Durbin, 2009).

Reads were sorted by coordinates, normalized, cleaned and duplicates were marked using

SAMtools and Picard. Base quality score assignments were recalibrated to control for biases due to flow cell, lane, dinucleotide context and machine cycle using the Genome Analysis Toolkit

(GATK) (McKenna et al., 2010). Copy-number alterations were evaluated using SegSeq (Chiang et al., 2009). GISTIC2 was used to identify recurrent copy-number alterations (Beroukhim et al.,

2007; Beroukhim et al., 2010; Zack et al., 2013). Somatic point mutations and short indels were called using Mutect (Cibulskis et al., 2013) and IndelLocator, and visual inspection in IGV

(Robinson et al., 2011). Mutsig (version 2.0) (Lawrence et al., 2014) was applied to detect significantly recurrent mutations. Rearrangements and breakpoints were identified using dRanger,

BreakPointer (Yotam Drier et al., 2013) and visual inspection. All analyses were performed within Firehose (M. A. Chapman et al., 2011). This work was conducted by our collaborating lab,

Rameen Beroukhim and Keith Ligon at Dana-Faber Cancer Institute.

RNA-sequencing and analysis pipeline

Following RNA extraction (RNeasy, Qiagen), library construction was performed using a non- strand specific Illumunia TruSeq protocol. Flowcell cluster amplification and sequencing were performed according to the manufacturer’s protocols using HiSeq 2000/2500, with a 76 bp paired-end run including an eight-base index barcode read. RNA-sequencing files were downloaded from published datasets (D. T. W. Jones et al., 2013; J. Zhang et al., 2013). RNA-seq bam files were transformed to fastq files using the Picard SamToFastq algorithm. Raw paired-end reads were aligned to the reference genome hg19 and preprocessed using PRADA (Pipeline for

RNA-sequencing Data Analysis) (Torres-García et al., 2014). We used PRADA within Firehose to determine gene-expression levels, exon expression levels, quality metrics, and for detection of fusion transcripts. BAM files were also assessed by visual inspection. This work was conducted in collaboration with Rameen Beroukhim and Keith Ligon’s labs at Dana-Faber Cancer Institute.

Array CGH

DNA was extracted from archival FFPE samples and aCGH performed as previously described

(Craig et al., 2012; Ramkissoon et al., 2013). GC-normalized copy-number data was cleaned of known germ-line copy-number variations and circular Binary Segmentation was used to segment the copy-number data (α = 0.001, undo.splits = sdundo, undo.SD = 1.5, minimum width = 5).

This work was conducted by Rameen Beroukhim and Keith Ligon labs at Dana-Faber Cancer

Institute.

Whole exome sequencing

WES was performed from FFPE samples (without matched control). These samples were used to confirm driver alterations identified by the WGS. DNA was extracted using the QIAGEN DNA

Blood and Tissue kit. Libraries with a 250 bp average insert size were prepared by Covaris sonication, followed by double-size selection (Agencourt AMPure XP beads) and ligation to specific barcoded adaptors (Illumina TruSeq) for multiplexed analysis. Exome hybrid capture was performed with the Agilent Human All Exon v2 (44 Mb) bait set.

Sequence data were aligned to the hg19 reference genome with the Burrows-Wheeler Aligner with parameters [-q 5 -l 32 -k 2 -t 4 -o 1]. Aligned data were sorted, duplicate-marked, and indexed with Picard tools. Base-quality score recalibration and local realignment around insertions and deletions was achieved with the Genome Analysis Toolkit.

Mutations were called with MuTect, filtered against a panel of normals, and annotated to genes with Oncotator. Likely germline SNPs were removed by filtering against the Exome Sequencing

Project (ESP) and Exome Aggregation Consortium (ExAC).

This work was conducted by Rameen Beroukhim and Keith Ligon labs at Dana-Faber Cancer

Institute.

Histological assignment

Histologic subtype assignments were according to previously published data. Samples not previously published were centrally reviewed and classified by a board-certified neuropathologist

(K.L.L., S.S, or S.R.) using W.H.O. 2007 criteria.

MYB FISH

FISH was performed as previously described (Firestein et al., 2008) using five micron FFPE tissue sections and Homebrew probes RP11-63K22 (5’ to MYB; directly labeled in

SpectrumOrange) and RP11-170P19 (3’ to MYB; directly labeled in SpectrumGreen) that map to

6q23.3. MYB status was assessed in 50 tumor nuclei per sample. A CEP6 aqua probe

(Invitrogen) mapping to the centromeric region of chromosome 6 was co-hybridized as a control.

This work was conducted by the Keith Ligon lab at Dana-Faber Cancer Institute.

Immunohistochemistry

Diaminobenzidine (DAB), bright-field staining was performed according to standard protocols on five-micron thick paraffin sections. Heat and 10 mM sodium citrate buffer (pH 6.0) were used for antigen retrieval for the MYB (Abcam for human tissue, Bethyl Laboratories for mouse tissue), OLIG2 (Chemicon) and GFAP (Millipore) antibodies. Counterstaining for nuclei was performed using Mayer’s hematoxylin stain and coverslips were mounted with Permount (Fisher

Scientific). Sections from the left occipital pole of a normal adult brain autopsy were used to assess MYB levels. This work was conducted by the Keith Ligon lab at Dana-Faber Cancer

Institute.

Analysis of QKI alterations in TCGA samples.

GISTIC 2.0 analyses were performed across 10,570 tumor samples from 31 lineages from The

Cancer Genome Atlas (TCGA), as previously described ack 20 (Zack et al., 2013). This work was conducted by the Beroukhim lab at Dana-Faber Cancer Institute.

Analysis of gene expression in normal tissues.

RNA-sequencing of normal pediatric brain samples was accessed from the BRAINSPAN Atlas of the human developing brain and processed as previously described (Bergthold et al., 2015). MYB expression levels from RNA-sequencing obtained from normal autopsy tissues were downloaded from the GTEx consortium (G. Consortium, 2013). Expression levels were compared using

ANOVA and t-tests. p-values <0.05 were considered significant. This work was conducted by the

Beroukhim lab at Dana-Faber Cancer Institute.

Vector construction and generation of NIH3T3 stable lines

MYB-QKI5 and MYB-QKI6 constructs were synthesized as Gateway compatible entry clones.

MYBtr constructs were generated via PCR mutagenesis using MYB-QKI fusions as templates.

Full-length MYB and QKI constructs were purchased as gateway entry clones from

PlasmID/DF/HCC DNA Resource Core. MYB-QKI5 and MYB-QKI6, MYBtr, full-length MYB and QKI constructs were sub-cloned into a Gateway-compatible N-MYC-tagged pMXs-Puro

Retroviral Vector (Cell Biolabs). Platinum-E retroviral packaging cells (Cell BioLabs) were used to generate retrovirus as per manufacturer protocols. NIH3T3 cells were infected with retrovirus containing media for 6 hours and puromycin selection commenced 48 hours post infection. Stable expression of MYC-tagged proteins was confirmed via western blot analysis (anti-MYC HRP

1:5000 (Invitrogen), anti-MYB antibody 1:5000 (Abcam) and anti-QKI 1:1000 (Bethyl Lab).

Soft Agar Colony Formation Assays and quantification

Anchorage-independent growth of NIH3T3 cells was assayed as previously described (A. J.

Sievert et al., 2013) with the following modifications: NIH3T3 cells expressing each of the MYB-

QKI5, MYB-QKI6, MYBtr, full-length MYB and full length QKI proteins and retroviral vector control were plated in 0.7% agar with DMEM and DBS in 96 well plates (in triplicates). Cell colonies were allowed to form for two weeks and images were taken. Images were analyzed using

ImageJ and colonies with area greater than 500 pixels quantified.

Generation of reporter construct containing MYB promoter and enhancer constructs

To assess the effect of candidate enhancer regions on MYB promoter activity, the human MYB promoter sequence (shown below) was cloned into the pLightSwitch_Prom Vector (Active Motif) that contains a multiple cloning site upstream of a Renilla luciferase reporter gene (RenSP) without a promoter. The MluI/BglII site on pLightSwitch_Prom Vector was used to clone the

MYB promoter sequence (Table 2.9) and the MluI site was further used to clone candidate enhancer regions upstream of the MYB promoter. The human QKI 3’UTR enhancer sequences

(hg18 chr6:163920360-163920809 and chr6:163921548-163921972) were synthesized by

Invitrogen and cloned into the reporter constructs as described above. The LightSwitch™

Random Promoter Control 1 (Active Motif) containing a 1 kb non-conserved, non-genic and non- repetitive fragment from the human genome cloned upstream of the RenSP luciferase reporter gene was used as a negative control. A housekeeping gene promoter vector, LightSwitch™

ACTB Promoter Control, was used as positive control for all assays. The luciferase reporter constructs containing either the MYB promoter or MYB promoter with enhancers were transfected into U87 glioma line (or MYB-QKI stably expressing U87 line) using Lipofectamine

3000 (n=5), or co-transfected with MYB-QKI or vector control into HEK 293s via Lipofectamine

2000 (Invitrogen), or in NIH3T3/ MYB-QKI stably expressing NIH-3T3 lines using PolyFect

(Qiagen). Luciferase activity was quantified 24 hours post transfection using the LightSwitch™

Luciferase Assay Reagent (Active Motif) according to the manufacturer’s protocol. mim-1 reporter construct generation and MYB transactivation assays Luciferase reporter constructs containing a consensus DNA-binding sequence for c-MYB were generated. The reporter construct was designed using the core MYB recognition element (MRE) consensus sequence PyAAC(G/T)G which is present in the mim-1 gene promoter, a previously described

MYB target (Ness et al., 1989). Double stranded oligos were generated by annealing primers mim-1 forward and mim-1 reverse (Table 2.9). The annealed oligo was ligated into pGL4.10[luc2] vector (Promega) digested with XhoI and HindIII. The pRL Renilla Luciferase

Reporter Vector (Promega E2261) served as an internal control in all assays. The mim-1 reporter construct and pRL renilla vector (ratio 30:1) were co-transfected into HEK-293 along with indicated fusions or controls via Lipofectamine 2000. Luciferase activity was quantified 24 hours post transfection using the Dual-Luciferase Reporter Assay System (Promega) according to manufacturer’s protocol.

Cell lines

NIH3T3, 293T and U87 MG cell lines were obtained directly from ACTT and not re- authenticated. All cell lines were routinely tested (at least every three months) for mycoplasma infection.

Generation of neural stem cells.

Embryonic murine neural stem cells (mNSC) were derived from C57BL6 wild-type E14.5 dpc mouse embryos (purchased from Taconic) as previously described (Reynolds & Weiss, 1992). mNSCs were maintained in culture media with 1:1 Dulbecco modified Eagle medium (Gibco) 42 and neural stem cell media (Gibco) supplemented with B27 (Gibco), EGF (02653, Stem Cell),

FGF (GF003, Millipore), and Heparin (07980, Stem Cell). This work was conducted by the

Beroukhim lab at Dana-Faber Cancer Institute.

Overexpression of transcripts in mouse neural stem cells

293T cells were transfected with 10 μg lentiviral pLEX307 expression vectors (gift from David

Root, Addgene plasmid #41392) with packaging plasmids encoding PSPAX2 and VSVG using

Lipofectamine. Lentivirus-containing supernatant was collected 48 hour after transfection, pooled and concentrated (ultrafiltration). Target neural stem cells underwent infection using a spin protocol (2000rpm for 120 minutes at 30C with no polybrene). Puromycin selection (0.5mcg/ml) commenced 48 hours after infection. This work was conducted by the Beroukhim lab at Dana-

Faber Cancer Institute.

ShQk experiments and proliferation assays

Lentiviral vectors (pLKO) encoding shRNAs specific for mouse Qk, targeting sequences in the first four exons of Qk, and the control shLacZ were obtained from The RNAi Consortium (Table

9). Lentivirus was produced by transfection of 293T cells with vectors encoding each shRNA

(10 μg) with packaging plasmids encoding PSPAX2 and VSVG using Lipofectamine (Invitrogen,

56532). Lentivirus-containing supernatant was collected 48 hours after transfection, and concentrated. Target mNSC underwent infection using a spin protocol (2000rpm for 120 minutes at 30C with no polybrene). Cells were placed into proliferation assays 48 hours after infection.

This work was conducted by the Beroukhim lab at Dana-Faber Cancer Institute.

Cell Proliferation Assays

1000 cells/well were plated in 96-well plates, with five replicates. Cell viability was measured by

43 assessing ATP content using Cell Titre-Glo (Promega). Mean SEM was calculated. This work was conducted by the Beroukhim lab at Dana-Faber Cancer Institute.

Western immunoblotting

Cells were lysed and subjected to SDS-PAGE gradient gels as previously described (Pratiti

Bandopadhayay et al., 2014). Blots were probed with antibodies against MYB (ab45150,

Abcam), QKI (A300-183A, Bethyl Laboratories) and actin (sc-1615, Santa Cruz).

RNA extraction and Real-Time RT-PCR

RNA was extracted with the RNeasy kit (Qiagen). cDNA was synthesized from 1µg RNA using

High Capacity RNA to cDNA kits (Applied Biosystems). Real-time RT-PCR was performed as previously described (Pratiti Bandopadhayay et al., 2014). Primers for MYB, QKI and -actin are listed in Table 9. Samples were amplified in triplicate and data analyzed using the CT method.

This work was conducted by the Beroukhim lab at Dana-Faber Cancer Institute.

Gene-expression analysis of neural stem cells expressing MYB-QKI

RNA was extracted from three independently generated pools of mNSC expressing one of eGFP,

MYBtr, MYB-QKI5, MYB-QKI6 or QKItr. Gene expression profiles were assayed using

Affymetrix Mouse Gene 2.0 ST microarrays (Affymetrix, Santa Clara, CA). CEL files were

RMA normalized (Bolstad, Irizarry, Astrand, & Speed, 2003). Comparative marker selection analysis (Gould, Getz, Monti, Reich, & Mesirov, 2006) was performed in GenePattern using default settings. Genes with p-value <0.05 and q-value <0.35 were considered significant. GSEA was performed using the C2 (CP) gene sets (MSigDB). Genesets with nominal p-values <0.05 were considered significant. The MYB-QKI signature was defined using the ClassNeighbours module of GenePattern (default settings). This work was conducted by the Beroukhim lab at

Dana-Faber Cancer Institute.

Antibody optimization and ChIP-seq

We systematically determined the antibody and concentrations that produce the highest signal to noise ratio for MYB ChIP-seq using our automated ChIP-seq methodology (Etchegaray et al.,

2015; Garber et al., 2012). We tested two MYB antibodies: abcam ab45150 and Sigma

SAB4501936. Abcam45150 has been previously used to ChIP MYB (Mansour et al., 2014). We split the sheared chromatin between 3 ratios of antibody/chromatin (0.5µl, 1 µl and 5 µl of each antibody/1,000,000 cells), and performed ChIP-seq as previously described (Garber et al., 2012).

As a positive control, we included an antibody targeting H3K27ac (Cell Signaling Technologies

D5E4, optimized at 1 µl/1,000,000 cells). We found 1µl ab45150/1,000,000 cells to be optimal.

Results from the MYB ChIP-seq were validated in three ways. First, we performed MYB ChIP- seq in K562 cells and confirmed enrichment at genes reported to be target genes in a prior study of these cells (Fig. 2.17) (Bengtsen et al., 2015). Second, we used Homer (Heinz et al., 2010) to perform an unbiased motif analysis across peaks identified (peak detection threshold of ) p1-e6)) in mNSC over-expressing MYBQKI, and identified MYB motifs to be the most enriched motifs across all peaks (p= 1e-681) (Fig. 2.18a). We observed 92% of all peaks (p threshold 1e-6,

3392/3672 peaks) to contain a MYB motif (Fig. 2.18b, Table 2.10). Enrichment of MYB motifs was significantly higher in data generated with the MYB antibody (p= 1e-681) compared to those generated from enrichment with H3K27ac (p=1e-25). Third, we compared our results from mNSCs to other published MYB ChIP-seq results. We determined whether MYB bound genes identified in our study (MYB peaks containing a MYB motif) were overlapping target genes reported in these studies, using a Chi-Square test with Yates Correction. We observed significant enrichment (p<0.0001; Fig. 2.18c) (Quintana, Liu, O'Rourke, & Ness, 2011; L. Zhao et al.,

2011). Of 2879 genes identified in our set 20% (Zhao) and 42% (Quintana) were also identified in these studies.

ChIP libraries were indexed, pooled and sequenced on Illumina Hi-seq-2000 sequencers. Raw data was aligned to the mouse reference genome MM9 using Picard tools. Raw sequencing data was mapped to the reference genome using bowtie2 version 2.2.1 with parameters -p 4 -k 1.

Peaks were called using MACS version 1.4.2 over an input control. A p-value threshold of enrichment of 1E-6 was used. Density of genomic regions was calculated using bamliquidator_batch, version 1.1.0. Reads were extended 200-bp and normalized to read-density in units of reads per million mapped reads per bp (rpm/bp). To calculate genome-wide overlap, all enriched H3K27ac peaks were extended 5kb in each direction, divided into 50 bins, and read density was calculated in each bin. Density was normalized to the largest value observed in each experiment genome-wide and plotted as a heat map. Peaks and alignments were converted to

TDFs by IGV tools and visualized by IGV. Bed files of published ChIP-seq data of H3K27ac chromatin maps from normal brain (Zhu et al., 2013) were downloaded and visualized in IGV.

ChIP-seq enriching for H3K27ac was performed on human pediatric low-grade gliomas by

Active Motif as recently described (Northcott et al., 2014). Analysis was performed as above using a p-value threshold of enrichment of 1E-5. Super-enhancer analysis was performed as previously described (Chapuy et al., 2013).

This work was conducted by the Ligon and Beroukhim labs at Dana-Faber Cancer Institute.

In vivo experiments

Mouse flank tumor studies with NIH3T3 Stable Cell Lines: NIH3T3 cell lines were injected subcutaneously into the flanks of NSG mice (5 mice for each cell line). Mice were 6-10 weeks of

46 age, with equal representation of male and female mice. Tumor growth was measured biweekly.

Ellipsoid tumor volume was calculated using the formula: volume = 1/2(length × width2).

Intracranial mouse injections: Neurospheres were dissociated and resuspended at 100,000 viable

immunocompromised ICR-SCID mice. Animals were monitored and sacrificed at the onset of neurological symptoms. Brains were subjected to routine histological analysis. Tumors were scored as present based on identification of atypical cells by a neuropathologist. 4-6 week old, male IcrTac:ICRPrkdc-Scid mice from Taconic were used. A total of 44 mice were used. This MNSC work was conducted by the Ligon and Beroukhim labs at Dana-Faber Cancer Institute.

Mouse injections were not randomized or blinded. Sample size was not predetermined.

Qualitative assessment of tumorogenicity was the primary outcome measure. Neuropathologists were blinded to group allocation.

Statistical analysis

For statistical analysis (unless otherwise described), p values were calculated using Fisher’s, T- tests or Pearson’s as appropriate. ANOVA with correction was used for comparison of multiple groups. Log-rank (Mantel-Cox) survival analysis was performed for animal studies and Kaplan

Meier curves generated. Error bars shown depict standard error of the mean.

Tables and Figures

Table 2.1. PLGGs included in analysis including tumor demographics, method of profiling and driver alterations identified. Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S2.xlsx

Table 2.2. GISTIC amplification peaks Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S3.txt

Table 2.3. GISTIC deletion peaks Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S4.txt

Table 2.4. Differentially expressed genes in neural stem cells expressing eGFP, MYBtr or MYB-QKI Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S5.xlsx

Table 2.5. Gene-Set Enrichment Analysis in neural stem cells expressing eGFP, MYBtr or MYB-QKI Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S6.xls

Table 2.6. Differentially expressed genes in neural stem cells expressing MYB-QKI with shLacZ or shQk Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S7.xlsx

Table 2.7. miRNAs that are differentially expressed in mNSCs expressing MYB- QKI following suppression of wild type Qk Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S8.xlsx

Table 2.8. MYB promoter sequence and shQk clones used in experiments Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S9.xlsx

Table 2.9. PCR primers used in experiments. Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S10.xlsx

Table 2.10. ChIP-seq peaks (p1e-6) identified following MYB ChIP-seq in mNSC over-expressing MYBQKI5. Peaks that contain a MYB binding motif (fragment size used in analysis = 200) are shown in the columns labeled MYBL1 or MYB. Online link to table: http://www.nature.com/ng/journal/v48/n3/extref/ng.3500-S11.xls

Figure 2.1

Amplifications Deletions

1 1

2 2

3 3

4 4 TERT 5 5

6 6

7 7 BRAF 8 8 8q24.3 8q24.3 CDKN2A, CDKN2B 9 9 10 10 11p15.5 11 11

12 12 13 13 14 14 15 16p13.3 15 16 16 17 17 18 18 19 19 20 20q13.33 20 2122 21 -5 -10 -30 -80 -300 22 0.25 10 10 10 10 10 0.25 10-1 10-2 10-3 10-4 q value q value

Figure 2.1. Significant regions of recurrent copy-number alterations in PLGG. GISTIC q values for amplifications (left) and deletions (right) are plotted across the genome.

Figure 2.2

Figure 2.2. Genomic analysis of 172 WGS and/or RNA-seq of PLGGs reveals a recurrent rearrangement involving MYB and QKI in Angiocentric Gliomas. (a) Driver alterations were identified in 154 of 172 PLGGs profiled with WGS and/or RNA-seq. Histological subtypes include Pilocytic Astrocytoma (PA). Pilomyxoid Astrocytoma (PMA), Angiocentric Glioma (AG), Oligodendroglioma (OD), Diffuse Astrocytoma (DA), Dysembryoplastic Neuroepithelial Tumor (DNT), Ganglioglioma (GG), Pleomorphic Xanthoastrocytoma (PXA), and PLGG not otherwise specified (NOS). Tumors for which histology is unavailable are designated NA. (b) FISH using probes flanking MYB reveal three patterns in PLGG: disomy, MYB rearrangement, or 3’ MYB deletion. Scale bars = 5 microns. (c) Frequency of MYB alterations or MYB-QKI rearrangements in Diffuse Astrocytoma and Angiocentric Glioma. p value represents enrichment of MYB-QKI rearrangements in Angiocentric Glioma. MYB-QKI alterations were identified with WGS alone (n=1), WGS and RNA-seq (n=2) or RNA-seq alone (n=3). (d) Breakpoints observed in MYB and QKI in Angiocentric Gliomas. Sequence across the breakpoints as determined by RNA-seq is shown for each rearrangement. (e) Copy-number profiles from WGS data of MYB and QKI in three Angiocentric Gliomas. 51

Figure 2.3

Location

Sequencing

Histology

BRAF

MYB/MYBL1 or 6q deletion

H3F3A

TP53

IDH1

Location Alteration Histology Sequencing Supratentorial Rearrangement PA DA WES Infratentorial Mutation PMA DNT aCGH NA AG GG aCGH and WES OD PXA NOS NA

Figure 2.3. Co-mut plot of driver alterations identified in 54 PLGGs profiled with WES and/or array CGH.

Figure 2.4

Figure 2.4. Alterations of MYB and QKI occur frequently in human cancers. (a) Expression of MYB (mean ± SEM) in normal human colon (n=12), breast (n=27), whole blood (n=51), esophagus (n=38), skin (n=25), and brain cortex (n=47). (b) MYB immunohistochemistry on human adult frontal cortex. Scale bar = 100 microns (c) MYB immunohistochemistry on human adult white matter. Scale bar = 100 microns (d) Hematoxylin and eosin (H&E) on human fetal neural stem cells generated from the ganglionic eminence at 22 weeks gestation. Scale bar = 100 microns (e) MYB immunohistochemistry demonstrates positive staining in a subset of cells. Scale bar = 100 microns (f) Sagittal section from embryonic 14.5 days post coitus (E14.5) mouse brain. Scale bar = 500 microns. (g) H&E of E14.5 ganglionic eminence (ge) including ventricular (ge-vz) and subventricular (ge-svz) zones. Scale bar = 50 microns. (h) Immunohistochemical analysis for MYB on the E14.5 ganglionic eminence. Scale bar = 50 microns. (i) MYB immunohistochemistry demonstrates positive staining in a subset of cells subventricular zone (ge- svz) but not the ventricular zone (ge-vz). Scale bars = 50 microns. (j) H&E from periventricular region of adult mouse brain. Scale bar = 100 microns. (k) Immunohistochemistry for MYB demonstrates positive cells (arrows) in the ependymal/SVZ layer. Scale bars = 100 microns. (l) (Left panel) Significance of deletions (x-axis) and (middle and right panels) heatmaps indicating copy-number profiles of individual tumors, for chromosome 6q of adult Glioblastomas profiled by TCGA. (m) Predicted structure of the MYB-QKI fusion protein, with fusion of the C-terminus of QKI to truncated MYB. TAD denotes transactivating domain. The C-terminus of QKI includes QUA2 domains and undergoes alternate splicing. MYB-QKI5 retains a nuclear localizing sequence (NLS) that is lost in MYB-QKI6. Two variants of MYB-QKI are depicted (Short and Long) corresponding to the breakpoint of MYB within the rearrangement.

Figure 2.5

a b c eGFP MYB-QKI

ANKS1A MYB H3K27ac k g g a n FRAS1 n i i

e 100 ** d d KIT p

n n I i PRELP i K b b

ZMYND10 Q y y I I - b b K

SH3RF3 K

d d Y Q CDK6 Q 60 - - e e M k

ANKS1A k B

B n h n Y Y FLNB t

i 40 a a r r M M w

LRRC8B

f f s s ITM2A s o o k k

e 20 a KIF26B a n e e MYLIP e p p g

0 STOM f A A d o d

N e UACA N t t a a % D TMOD1 D DGKG PIWIL4 0 72 0 65 SRGAP1 Up-regul PRUNE2 reads/bin reads/bin Down-regul SNCAIP FRMD4B ANKRD28 SLC9A3R1 LRRC8C d e TTYH1 PTGDS SOX8 LIFR *** KCNMA1 *** **

CSPG4 s t i 1500 ** XYLT1 n *** 0.20 )

REV3L U

m e EDNRB k s p r

NXPH1 a

r 0.15

ACAP3 e m 1000 f i u

HEPACAM c s u

B3GAT1 ( L

0.10

ANKH n d o e ATP1A2 *** i z s i 500 PLXNA2 l s a

IGSF21 e 0.05 r m p MTSS1L r x SMOC1 o E N LRIG1 0 0 GJC3 l r l I o B t I5 I6 a I K BCHE tr Y B K K K Q n M Y Q - TNFRSF21 o M -Q -Q - B C B B Norm B Y PTPRE Y Y Y M M M M LRRC61 n No

Figure 2.5. MYB-QKI functions as a transcription factor, and its molecular effects are observed in Angiocentric Gliomas. (a) MYB-QKI expression signature in mouse neural stem cells relative to cells expressing eGFP controls. (b) Heatmap of H3K27ac and MYB-QKI levels at MYB-QKI regions. Each row shows +/-5 kb centered on MYB-QKI peaks. These regions are rank-ordered by MYB-QKI signal. Scaled intensities are in units of rpm/bp (c) ChIP-seq binding of MYB-QKI to genes included in the MYB-QKI gene expression signature. (d) mim-1 reporter induction following transfection of MYBtr, MYB-QKI5, MYBQKI6 or full length MYB in 293T cells. Values shown represent mean of three independent measurements ± SEM. (e) Expression of MYB-QKI signature in normal pediatric brain samples, PLGGs without MYB-QKI, or Angiocentric Gliomas with MYB-QKI. Values represent mean expression of signature in tumors ± SEM. Expression of signature within each tumor is the sum of rpkm of each gene in the signature.

a Figure 2.6 Peak a H3K27ac PromoterPeak

MYB H3K27ac Promoter

MYB

b c l

o Anti- Myb Anti- Qki

6 5

I I

K K

r r r

r f

t t

Q Q

5 6

B B B B

b I I c

Y Y Y Y

K K e

M Q Q

V M M M 98 kDa

62 kDa l

o Anti- Myb Anti- Qki

6 5

I 38 kDa I

K K

r r r

r f

t t

Q Q

5 6

B B B B

I I

Y Y Y Y K K 17 kDa e

M Q Q

V M M M 98 kDa Anti- GAPDH 62 kDa 38 kDa d 17 kDa ** 300

) ** Anti- GAPDH m k p r ( 200 n o i s s e d r p 100 x E ** 300 ) 0 ** m Normal non Angiocentric k

p Brain MYB-QKI Glioma r ( 200 PLGG n o

i s s e r

Figurep 100 2.6. MYB functional assays and related data (a) Identified MYB-QKI binding peaks x at MYBE in mNSC over-expressing MYB-QKI. (b) Expression of MYB, MYBtr, MYB-QKI and QKItr0 in mim-promoter assays (c) Linear increases in mim-1 promoter activity with increases in MYB-QKINormal expression.non (d) AngiocentricEnrichment of MYB pathway activation in normal brain, PLGGs Brain MYB-QKI Glioma without MYB-QKI PLGGor Angiocentric Gliomas with MYB-QKI. Values represent mean expression ± SEM

Figure 2.7

a b ** 800 n )

40 o Breakpoint m n

k o

p ** t r

(

e 600

30 I v n i K t o i a Q l s - s e B

20 r e

r Y 400 n p o M i x

E 10

s n e r a

p 200 e 0 x M Normal Non MYB-QKI E MYB-QKI 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 c 255 kb H3K27ac Peak MYB-QKI Peak Qk Chr 17q

d e

400 ** b K

/ 300 c

a MYB 15 kb QKI 7 200 2 H3K27ac Temporal K H3K27ac Frontal 3 100 H

0 Q3E1 MYB QKI

Figure 2.7. Angiocentric Gliomas exhibit aberrant expression of MYB-QKI due to H3K27ac enhancer translocation and an auto-regulatory feedback circuit in which MYB-QKI binds to the MYB promoter. (a) MYB RPKM expression levels of MYB-QKI tumors (n=5) relative to normal brain (n=10) or BRAF or FGFR driven PLGGs (n=10). Values shown represent mean ± SEM. (b) Exon-specific expression of MYB in Angiocentric Gliomas that harbor MYB-QKI relative to PLGGs that harbor BRAF alterations (n=3). (c) H3K27ac binding in proximity to MYB (upper) and QKI (lower) in mouse neural stem cells. MACs peaks are shown. (d) H3K27ac binding to enhancer elements within MYB and QKI in human frontal and temporal lobes (Encode). Values shown depict the mean number of nucleotides that are H3K27ac bound (per kilobase) in MYB and QKI across both locations. (e) Predicted H3K27ac enhancer elements in MYB-QKI, with translocation of genomic enhancers on 3’QKI within 15kb of 5’MYB. Enhancer maps shown are derived from Encode data from normal brain (frontal lobe and temporal lobe). Q3E1 shows an H3K27ac enhancer present in the Encode data from normal brain data.

Figure 2.8 a Chr 6 135.3 Mb 135.5 Mb 135.8 Mb 163.8 Mb 164 Mb 164.4 Mb

HBS1L MYB AHI1 LINC00271 QKI BRAF duplicated Tumor H3K27ac

Peaks

MYB-QKI Tumor H3K27ac

Peaks M5E1 Q3E1 Q3SE1 Q3SE2 MYB Exon 1-9 QKI Exon 5 -8 9 Kb

H3K27ac

Peaks M5E1 Q3E1 RNA-seq reads b c *** *** 120 U87

164 Mb 164.3 Mb 164.55 s U87 with MYB-QKI t i QKI n U

AG1 H3K27ac y r 80 a ***

AG1 Enhancer peaks t i

b *** *** r

AG2 H3K27ac A ***

AG2 Enhancer peaks e s

a r 40 e

SE Peaks AG1 f i c

Q3SE1 Q3SE2 u

SE Peaks AG2 L 0 Control Control promoter MYB MYB promoter promoter and Q3E1 promoter and Q3E1 Figure 2.8. Human Angiocentric Gliomas exhibit H3K27ac enhancer translocation with an aberrant enhancer associated with the MYB promoter. (a) H3K27ac enhancer peaks in proximity to MYB and QKI in a BRAF-duplicated Pilocytic Astrocytoma (top) and MYB-QKI Angiocentric Glioma (lower). Q3E1 is an enhancer associated with the 3’UTR of QKI. Two super-enhancer clusters (Q3SE1 and Q3SE2) are located within 500kb of QKI. Angiocentric Gliomas are associated with aberrant enhancer formation at the MYB promoter (M5E1), which is not detected in the BRAF driven pilocytic astrocytoma. The breakpoints for the MYB-QKI rearrangement are between exons 1-9 MYB and 5-8 QKI. Expression as determined by RNA- sequencing is depicted for the MYB-QKI Angiocentric Glioma. (b) 3’ QKI associated super- enhancers (Q3SE1/2) presented in two Angiocentric Gliomas. (c) MYB promoter activation following transfection of the MYB-luc construct in U87 cells and U87 cells over-expressing MYB-QKI with and without Q3E1 enhancer cloned into MYB-luc construct. Changes in luciferase activity of the MYB-luc reporter is shown as mean of three individual replicate experiments with n=5.

Figure 2.9

Figure 2.9. H3K27ac enhancer profiling of 3’QKI in a human Angiocentric Glioma. H3K27ac peaks are depicted.

Figure 2.10

MYB

Figure 2.10. H3K27ac enhancer formation at the MYB promoter in two human MYB-QKI Angiocentric Gliomas (top tracks) and a BRAF altered Pilocytic Astrocytoma (lower track).

Figure 2.11

a 1000 20 HEK 293t NIH 3T3 HEK 293t expressing MYB-QKI NIH 3T3 expressing MYB-QKI s s t t i 800 i n n U

U 15 ***

e e s s a a r

r ***

e 600 e f f i i c c u u 10 L L

r r a

400 a e e G G h h c c t t i

i 5 w

200 w S S

0 0 Control MYB promoter Control MYB promoter Luciferase reporter constructs Luciferase reporter constructs

b c

Expression of MYB-QKIl in NIH3T3 cells Expression of MYB-QKI in mouse NSC +eGFP +MYBtr +MYB-QKI5 +MYB-QKI6 o r t n 6 5 5 6 o 2 I I I I G C I K K K K

L r P Q Q Q Q O o F t B B B B B c G Y Y Y Y Y e e M M V M M M MYB

MYB P A

Actin F GAPDH G

Expression of MYBtr in NIH3T3 cells Expression of MYBtr in mouse NSC

I I

K K o r t r

Q Q t t P

B B

B B B n F

Y Y

Y Y Y o G

e M

M M C M M MYB 98kDa 62kDa MYB 38kDa GAPDH Actin

Figure 2.11. Supplementary figures for MYB-QKI assays (a) MYB promoter activation in HEK293T and NIH3T3 cells with and without expression of MYB-QKI. (b) Exogenous expression of MYB-QKI and MYBtr in NIH3T3 cells and mouse neural stem cells via retroviral transduction. (c) Representative OLIG2 and GFAP immunohistochemical analysis of tumors from mNSCs overexpressing MYBtr, MYBQKI5, MYB-QKI compared to eGFP controls. Positive staining for both OLIG2 and GFAP is observed in a subset of tumors cells in tumors generated from MYBtr or MYBQKI5 overexpression, while tumors formed by MYB-QKI6 overexpression are negative for OLIG2 and have low GFAP expression. Scale bars=50 microns.

Figure 2.12

Figure 2.12. MYB-QKI and MYBtr are oncogenic. (a) In vitro cell proliferation rates of mNSCs over-expressing eGFP, MYB-QKI5 (Short) or MYB-QKI6 (Short). Mean of five independent pools are depicted. Error bars represent SEM. Representative images intracranial mNSC-MYB-QKI6 tumors. (b) Tumor growth following flank injections of NIH3T3 cells overexpressing MYB, MYB-QKI5 (Long), MYB-QKI6 (Long) or a vector control. Mean of five measurements are depicted. Error bars represent SEM. (c) In vitro cell proliferation of mouse neural stem cells which over-express MYBtr or eGFP controls. Mean of five independent pools is depicted. Error bars represent SEM. Representative images of intracranial mNSC-MYBtr tumors. (d) Tumor growth following flank injections of NIH3T3 cells overexpressing MYB, MYBtr or vector control. Mean of five measurements is depicted. Error bars represent SEM. (e) Hematoxylin and eosin (H&E) analysis of SCID mouse brain after striatal injections with mNSCs expressing eGFP, MYBtr, MYBQKI5 or MYBQKI6. Scale bars=2.0mm (top) or 50 microns (bottom). (f) Kaplan-Meier survival analysis of orthotopic SCID mice injected with mNSCs overexpressing MYBtr, MYBQKI5, or MYBQKI6 develop tumors with short latency compared to mNSCs expressing eGFP which never developed tumors, p<0.01. 61

Figure 2.13

a b

mNSCs 3T3

) 30 p=0.02 Vector Control BRAF V600E 2 l +GFP 0/15 (0%) 0/5 (0%) e x i +MYBtr 2/10 (20%) 5/5 (100%) p

0 p=0.005 +MYBQKI5 2/9 (22%) 5/5 (100%) 0 20 MYB MYB-QKI5 5 +MYBQKI6 1/10 (10%) 5/5 (100%) > (

e Overall p<0.001 b 10

m MYB-QKI6 u n

y n o l 0 o Vector MYB MYB-QKI5MYB-QKI6BRAF C Control V600E

Figure 2.13 Oncogenic assays of MYB-QKI expressing cells. (a) Number of colonies of NIH3T3 cells expressing MYB, MYB-QKI or a vector control in soft agar (Left) and representative images (right). NIH3T3 cells over-expressing BRAFV600E are shown as a positive control. Mean of three replicate measurements are shown. Error bars represent SEM. (b) Table of tumor penetrance from both 3T3 and mNSC models with overall p-value<0.001.

Figure 2.14

a 1.5 o t

e G v i G t L a l P e

I 1

I K K Q - Q

B n Y o i M s

n 0.5 e o r n p

x E

0 1 2 3 4 5 6 7 Exon b 10 *** e n i

l *** e s

a 8 b

*** o ) t

e e g v i n 6 t a a l h e c r

d l e

c o 4 n F ( e c s e n

i 2 m u L 0 r P P t tr I5 I5 I6 I6 F F B B K K K K Y Y Q G G -Q -Q - -Q e e M M B Z I Z I B B B c K c K Y Y Y Y a Q a Q M M M M L h L h Z I Z I h s h s c K c K s s a a L Q L Q h h h h s s s s c h i t 400 w

p<0.0001 d e t a i c e

o 300 r s u s t a a

n s g e i n s

e 200 I g K

f Q o

h n s o i

s 100 s e r p x E 0 Normal Angiocentric Brain Glioma

Figure 2.14. MYB-QKI disrupts expression of QKI, a tumor suppressor gene. (a) Exon specific expression of in Angiocentric Gliomas (n=5) relative to BRAF-driven PLGGs (n=5). Values represent mean ± SEM. RNA-sequencing data of Exon 8 of QKI revealed a high number of duplicate reads and thus is not shown. (b) Cell proliferation of mouse neural stem cells expressing MYBtr, MYB-QKI5, MYB-QKI6 or eGFP control with suppression of wild-type Qk. Values represent mean of four independent experiments ± SEM. (c) Expression of signature within each tumor is the sum of rpkm of each gene in the signature.

Figure 2.15

c shLacZ shQKI a HEATR5B ZNF347 CCRL1 Z

c shLacZ PSD3

a MITD1

L shQk OR9M1P

o IL17RB

t 1.0

SGCD

e OR10W1 v

i TNFSF18

t 0.75

a ATP10B l PDSS2 e

r RPE65

0.5 C19ORF80 A WDSUB1 N 0.25 UNC50 R TM2D1

m OR2K2

k 0 CYP3A7 eGFP MYBtr MYBQKI5 MYBQKI6 DCUN1D1 Q SLC22A5 ALS2CR8 OR10A6 C17ORF80 FAHD2A b OR10D3 DNAJB4 e n

i 6 RBP7 l FOPNL e ** eGFP shLacZ

s MBLAC2

a PRPF4B eGFP shQKI MED21 B C19ORF46 o t 4 ** MYBQKI5 shLacZ OR9G4

e UROS

v SDR16C6P i MYBQKI5 shQKI t CCDC88A a l DSC3

e MYBQKI6_shLacz CD59

R ** MYBQKI6_shQKI DZANK1 e 2 ULBP1

g RP2 n PHLDA2 a OR2Y1 h MEOX2 C IL7 d l 0 STK11IP o GALNT5 F CDH26 RUFY3 MMRN2 Z

H2AFY2

c ALDH3B2

a SLC25A47

L 1.0 GKN1 o

t APBB1IP MTUS2 e 0.8

v DCAF7 i

t CHAT a l 0.6 NPPC e BHLHE23

R GCNT4 FAM70B e 0.4

g ROR2

n TCF3

a PDE6B h 0.2 SORBS3 C

AOPEP

d RTP1 l 0.0 o eGFP MYB-QKI5 MYB-QKI6 ARRB1 F KRT80 IL27 GRIK1 LRRC15 STAT5A ADM2 SIRPB2 FGD2 C19ORF47 RUNX1T1 AP1G1 FLJ42875 CYP21A1P COO5 GNAI2 MYOZ3 CDK18 CSTF2 RBM22 CPSF4 RAB25 FOXO4 KLF14 ZNF143 HNF4A TM4SF19 KLHL30 SPIRE2 MUC3A IL18RAP Figure 2.15. QKI knockdown in MNSCs and supplementary assays (a) Suppression of wild type Qk in mouse NSC expressing MYB-QKI following infection with shQKI vectors (or shLacZ control). (b) Proliferation of mouse NSC over-expressing eGFP, MYBtr or MYB-QKI with suppression of wild type Qk over five days in Pool 4 (Values represent mean of 5 replicate measurements with SEM). Expression of wild type Qk relative to shLacZ control is shown. (c) shQK signature defined following suppression of wild type Qk in mouse neural stem cells that over-express MYB-QKI6.

Figure 2.16

Figure 2.16. The MYB-QKI rearrangement contributes to oncogenesis through at least three mechanisms. The MYB-QKI rearrangement disrupts both MYB and QKI, resulting in hemizygous deletion of 3’MYB and 5’ QKI. This results in proximal translocation of H3K27ac enhancers on 3’QKI towards the MYB promoter, resulting in MYB promoter activation (i). The MYB-QKI fusion protein that is expressed is oncogenic, functions as a transcription factor, and exhibits the ability to bind to and activate the MYB promoter, resulting in an auto-regulatory feedback loop (ii). Hemizygous loss of QKI results in suppression of QKI expression, which functions as a tumor suppressor gene (iii).

Figure 2.17

MYB DUS3L

GRSF1 IKZF1

KCNH2 LMO2

MYADM RABEPK

RUNX1 RUNX2

SENP1

Figure 2.17. Validation of MYB ChIP-seq. MYB binding to reported MYB target genes in K562 cells. ChIP-seq enrichment was performed using Abcam 45150.

Figure 2.18

Figure 2.18. Validation of MYB DNA binding motif in MYB-QKI (a) Top four enriched DNA binding motifs in MYB ChIP-seq peaks obtained from mNSC over-expressing MYB-QKI. (b) MYB DNA-binding motifs. (c) Venn diagrams showing overlap of identified MYBQKI peaks identified (p threshold 1e-6) compared to those reported in Zhao et al and Quintana et al. p values shown were calcuated by Chi Square with Yates correction. Of 2879 genes identified in our set 20% (Zhao) and 42% (Quintana) were also identified in these studies.

CHAPTER 3: RAF1 gene fusions in pediatric low-grade gliomas show differential response to targeted therapeutics based on dimerization profiles.

Summary

Pediatric low-grade gliomas (PLGGs) are associated with dysregulated MAPK signaling, most commonly driven by gene fusions involving BRAF. This has led to the initiation of several clinical trials utilizing RAF and MAPK targeted therapeutics for PLGGs. While the majority of infratentorial PLGGs harbor BRAF gene fusions, a small number have been recently found to harbor gene fusions involving another RAF isoform, CRAF/RAF1. In this study, we unexpectedly found distinct responsiveness of RAF1 fusions to clinically relevant RAF-targeted therapies. We show that CRAF fusions, QKI-RAF1 and SRGAP3-RAF1 are oncogenic in multiple cellular contexts through activation of the MAPK and PI3K signaling pathways. While both BRAF and CRAF fusions resist suppression with first-generation RAF inhibitor

Vemurafenib, only BRAF fusions exhibited growth inhibition and reduced MAPK signaling in response to second-generation paradox breaking RAF inhibitor PLX8394. We further demonstrate that unlike BRAF fusions, RAF1 fusions retained robust homo- and heterodimerization in the presence of PLX8394 and disrupting dimerization through point mutations or with the novel LY3009120 RAF dimer inhibitor significantly reduced the oncogenicity of RAF1 fusions. Intriguingly, using dimerization point mutants of QKI-RAF1, we found that altering the dimerization domains in N-terminal QKI had a significantly larger effect on oncogenic phenotype compared to disruption of C-terminal RAF1 dimerization. This suggests a previously unrecognized role for N-terminal partner protein in mediating dimerization of RAF1 gene fusions and potential resistance mechanism to paradox breaker RAF inhibitors. Finally, we tested therapeutic targeting of downstream MAPK pathway in RAF1 fusions using the MEK inhibitor,

Trametinib and observed partial suppression of tumor growth. Using combinatorial targeting of both MAPK and PI3K pathway was more successful in suppressing RAF1 fusion induced tumor growth.

Introduction

As the most common set of brain tumors in children (Kleihues et al., 2002), pediatric low-grade gliomas (PLGGs) represent a heterogeneous group of tumors with histologies ranging from pilocytic astrocytomas (WHO grade I) to diffuse fibrillary astrocytomas (WHO grade II). Despite having good overall prognosis, children with PLGGs often suffer from long-term treatment related and tumor related complications (Sievert & Fisher, 2009). Mutations involving BRAF, mostly the KIAA1549-BRAF gene fusion that activates the MAPK pathway, have predominantly been associated with pilocytic astrocytomas (PAs) (Jacob et al., 2009; Sievert et al., 2009; A. J.

Sievert et al., 2013). However, recent sequencing studies have discovered the occurrence of rare

CRAF (or RAF1) gene mutations, QKI-RAF1 and SRGAP3-RAF1, in PLGGs (D. T. Jones et al.,

2009; J. Zhang et al., 2013). But the prevalence, oncogenic mechanism of action and sensitivity of the novel RAF1 gene fusions to targeted therapeutics remains unknown.

Whole-genome sequencing studies first detected QKI-RAF1 as an infrequent gene fusion in pilocytic astrocytomas (J. Zhang et al., 2013) whereas SRGAP3-RAF1 was first reported as a tandem duplication event that leads to a copy number gain at chromosome 3p25 in a pilocytic astrocytoma (D. T. Jones et al., 2006; D. T. Jones et al., 2009). While SRGAP3-RAF1 has been reported to activate the MAPK pathway, no further studies have been conducted on these RAF1 gene fusions to assess response to therapy. Another RAF1 gene fusion, called FYCO1-RAF1, has also been detected in PLGGs (J. Zhang et al., 2013). These studies add further complexity to the molecular classification schemes that are currently being developed for PLGGs, as there seems to be genetic diversity even within the same histologic subtype. 69

To study the PLGG-associated RAF1 fusions, QKI-RAF1 and SRGAP3-RAF1, we performed cellular, molecular, biochemical and in vivo assays to test oncogenic mechanisms activated by these fusions. Like the BRAF fusions, RAF1 fusions also activate the MAPK signaling pathway and also a parallel PI3K/mTOR pathway. While first-generation RAF inhibitors such as vemurafenib do not target PLGGs with BRAF fusion, second-generation PLX8394 inhibits

BRAF fusion’s oncogenicity (A. J. Sievert et al., 2013). Interestingly, we found the RAF1 fusion to be unresponsive to both first- and second-generation RAF inhibitors. Upon assessing their dimerization profile, we found dependence of RAF1 fusions on both homodimerization with itself and heterodimerization with wild type RAF proteins. These protein-protein interactions remained resistant to second-generation RAF inhibitors but were sensitive to and modified by novel RAF dimer inhibitor, LY3009120. We also tested the sensitivity of RAF1 fusions to MAPK and

PI3K/mTOR targeted therapeutics as single-agent and combinatorial therapies.

Results

QKI-RAF1 and SRGAP3-RAF1 fusions are oncogenic proteins that activate the MAPK and

PI3K signaling pathways

QKI-RAF1 contains exons 1-3 of the N-terminus of QKI, which belongs to a family of RNA binding proteins known as the signal transduction and activation of RNA (STAR) proteins (Fig.

3.1a). In this fusion, QKI loses part of its RNA binding domain from the C terminus but retains an intact homodimerization domain at the N-terminus (Vernet & Artzt, 1997). While QKI is known to regulate glial development and myelination in the CNS (Friedrich, 1974), loss of QKI has been detected in glioblastomas (Lawrence et al., 2014; Z. Z. Li et al., 2002; Yin et al., 2009), angiocentric gliomas (Bandopadhayay et al., 2016), prostate cancer (Y. Zhao et al., 2014), lung cancer (Zong et al., 2014) and gastric cancer (Bian et al., 2012).

CRAF/RAF1 is a serine/threonine kinase that was first discovered as the v-raf oncogene in the transforming mouse sarcoma virus (Rapp et al., 1983; Rapp & Todaro, 1978). Both QKI-RAF1 and SRGAP3-RAF1 fusions have lost the N-terminal non-catalytic domain of RAF1 and the functional kinase domain is retained (Fig. 3.1a) suggesting auto-activation of the kinase activity.

Several studies have shown that loss of N-terminal domains of RAF1 can lead to oncogenic activation and tumorigenesis (Cutler et al., 1998; Heidecker et al., 1990; Stanton, Nichols,

Laudano, & Cooper, 1989) but few studies have implicated oncogenic mechanism of RAF1 gene fusions by focusing on both gene fusion partners.

SRGAP3 is a member of the SLIT-ROBO Rho GTPase-activating protein (srGAP) family and is involved in several neurodevelopmental processes via regulation of actin cytoskeleton dynamics

(Bacon et al., 2009; Carlson et al., 2011). SRGAP3 possess an N-terminal Fes-CIP4 homology

(FCH) domain and a coiled-coil domain (together called F-BAR domain), a central Rho-GAP domain and C-terminal SH3 domain (Wong et al., 2001). The SRGAP3-RAF1 gene fusion retains the FCH domains of SRGAP3 that has been suggested to have dimerization properties (Henne et al., 2007) whereas the other C-terminal domains are lost. This also suggests a potential role for the N-terminal fusion protein SRGAP3 to contribute to fusion dimerization, which is a critical component for activation of RAF1 kinase activity.

While RAF1 gene fusions are rare genetic events in PLGGs, the overall prevalence is not known due to lack of genomic studies that can enrich for such fusions. However, the occurrence of different RAF1 gene fusions in adult cancers along with PLGGs (Fig. 3.1b) highlights the growing importance of studying RAF1 gene fusions and their therapeutic targeting. Majority of the RAF1 fusions listed in the table have lost the N-terminal auto-inhibitory domains of RAF1 thereby supporting the hypothesis that RAF1 kinase could be constitutively active in RAF1

71 fusions, thereby driving tumorigenesis. Therefore, we sought to test the oncogenic potential and mechanism of action of the PLGG associated RAF1 fusions- SRGAP3-RAF1 and QKI-RAF1.

To study these fusions, we stably expressed SRGAP3-RAF1 and QKI-RAF1 in two different heterologous cell model systems: NIH3T3 cells and p53-null primary mouse astrocyte cells

(PMAs) (J. Zhang et al., 2013). To assess the oncogenic potential of RAF1 gene fusions, we performed soft agar assays and measured colony formation. Over-expression of QKI-RAF1 and

SRGAP3-RAF1 in NIH3T3 cells and PMAs was sufficient to drive significant anchorage independent growth in soft agar compared to vector controls (Fig. 3.1c and 1e, p-value<0.05 to

0.001).

Upon assessing downstream signaling pathways in the stable NIH3T3 cell lines, we found the mitogen associated protein kinase (MAPK) and phosphoinositide-3 kinase (PI3K) signaling pathway to be aberrantly activated by QKI-RAF1 and SRGAP3-RAF1 as monitored by elevated phosphorylated MEK land AKT/S6 levels, respectively (Fig. 3.1d). The MAPK signaling pathway was aberrantly activated by QKI-RAF1 and SRGAP3-RAF1 in PMA cell lines (Fig.

3.1f). The differences in aberrant activation of MAPK and PI3K pathway in the NIH3T3 cell lines versus the PMA cells is likely due to variation in signaling programs in different cellular contexts.

To test tumor-forming capacity in vivo, we injected stable NIH3T3 cell lines expressing QKI-

RAF1 and SRGAP3-RAF1 into flanks of NSG mice. Both RAF1 fusions formed robust flank xenograft tumors compared to vector controls (Fig. 3.1g, p-value<0.001). Using the stable PMA cell lines, we performed intra-cranial injections into NSG mice brains. Interestingly, QKI-RAF1 and SRGAP3-RAF1 expressing PMAs but not vector control, caused tumors leading to poor survival in all mice injected with RAF1 fusion expressing PMAs (Fig. 3.1h, p-value<0.01). These

72 findings demonstrate that QKI-RAF1 and SRGAP3-RAF1 are driver oncogenes in PLGGs as they drive robust oncogenesis in different cell model systems.

First- and second-generation RAF inhibitors do not suppress QKI-RAF1 and SRGAP3-

RAF1

First generation RAF inhibitors such as vemurafenib (PLX4032) (Bollag et al., 2010) have shown clinical efficacy and survival benefit in melanomas that harbor the most commonly found BRAF mutation, BRAF V600E, that changes valine to glutamate at residue 600 in exon 15 (P. B.

Chapman et al., 2011; Flaherty et al., 2010; Sosman et al., 2012). However, we have previously shown that PLGGs with BRAF fusions such as KIAA1549-BRAF also demonstrate resistance to

PLX4720 (research analog of vemurafenib) via paradoxical activation of the MAPK pathway (A.

J. Sievert et al., 2013) (Fig. 3.2). We also demonstrated that KIAA1549-BRAF was responsive to a second-generation RAF inhibitor, PLX8394, which is termed as a ‘paradox breaker’ due to evasion of paradoxical MAPK pathway activation (A. J. Sievert et al., 2013; C. Zhang et al.,

2015). Although several RAF inhibitors are currently being tested in clinical trials for PLGGs, there are no existing studies examining how the RAF1 gene fusions would respond to these first- and second-generation RAF inhibitors, compared to BRAF fusions.

We tested the response of NIH3T3 and PMA cell models expressing QKI-RAF1 and SRGAP3-

RAF1 to increasing concentrations of first- and second-generation RAF inhibitors, PLX4720 and

PLX8394, respectively. In NIH3T3 lines expressing QKI-RAF1, both the first- and second- generation inhibitors caused paradoxical activation of MAPK pathway as evident by elevated phospho-MEK and phospho-ERK levels with increasing drug concentrations (Fig. 3.3a).

Interestingly, we observed a partial suppression of phospho-S6 signal at high concentrations of the drugs despite paradoxical phosphorylation of AKTT308 with PLX8394 (Fig. 3.3a), suggesting

73 some downregulation of PI3K pathway. We also observed similar paradoxical activation of

MAPK pathway in stable PMA cell lines with QKI-RAF1 (Fig. 3.3e). However, both RAF inhibitors failed to suppress QKI-RAF1 driven soft agar colony growth and instead showed paradoxical increase in oncogenic growth (Fig. 3.3b, f).

SRGAP3-RAF1 expressing NIH3T3 lines showed a similar paradoxical increase in MAPK pathway with increasing PLX4720 and PLX8394 concentrations (Fig. 3.3c) with some PI3K pathway suppression. Again, there was no decrease in SRGAP3-RAF1 driven soft agar colony growth in the presence of the RAF inhibitors in both NIH3T3 cells and PMAs (Fig. 3.3 d, g, respectively). These findings show a direct correlation of drug-induced paradoxical activation of the MAPK pathway in RAF1 fusion cells and unresponsiveness to first- and second-generation

RAF inhibitors. Even though the PI3K pathway was affected by high RAF inhibitor concentrations by some unknown mechanism, lack of suppression of soft agar growth suggests that it could be secondary to MAPK pathway in driving the RAF1 fusion cells.

RAF1 fusions function as homodimers and heterodimers with B/CRAF that are resistant to second-generation paradox breakers

The chemical structure of PLX8394 has been shown to prevent RAF dimerization by interacting with residues on the regulatory spine of RAF kinases (C. Zhang et al., 2015) but our findings suggest that RAF1 fusions respond differentially compared to wild type proteins. Since RAF homodimerization and heterodimerization is important for activation of RAF kinase activity

(Rajakulendran, Sahmi, Lefrancois, Sicheri, & Therrien, 2009) and the oncogenic kinase activity of KIAA1549-BRAF fusion has been attributed to homodimerization (A. J. Sievert et al., 2013), we hypothesized that the RAF1 gene fusions could be functioning via homodimerization and/or heterodimerization dependent activation. The presence of dimerization domains in both QKI (Ali

& Broadhurst, 2013; T. Chen & S. Richard, 1998)/ SRGAP3 (Henne et al., 2007) and RAF1

(Hatzivassiliou et al., 2010; Rajakulendran et al., 2009) suggests that both fusion partners in

RAF1 fusions could be contributing to homodimerization and/or heterodimerization. Therefore, we characterized how such protein-protein interactions could be responsible for oncogenic activity and RAF inhibitor resistance in RAF1 gene fusions.

To test the dimerization potential of RAF1 fusions, we assessed the ability of QKI-RAF1 to interact as homodimers and also heterodimerize with wild type BRAF, CRAF, QKI and truncated

QKI (exons 1-3) by using myc- and flag-tagged constructs. Upon immunoprecipitation (IP) of

QKI-RAF1 with a Myc antibody, we detected QKI-RAF1 protein-protein interactions with QKI-

RAF1 itself, wild type BRAF, CRAF, QKI and truncated QKI but not the vector control (Fig.

3.4). Similarly, SRGAP3-RAF1 also shows home-dimerization and heterodimerization with wild type BRAF, CRAF (Fig. 3.5). These results suggest that dimerization may underlie the activation and functionality of RAF1 fusion kinases.

Previous studies have shown that first-generation RAF inhibitor vemurafenib induces paradoxical activation of MAPK pathway by binding to one protomer of wild type RAF dimer and transactivating the other uninhibited RAF protomer (Hatzivassiliou et al., 2010). Modifications to structure of vemurafenib resulted in PLX8394 that has been shown to evade paradoxical activation in wild type BRAF and BRAFV600E mutant cells (C. Zhang et al., 2015). Since

PLX8394 does not suppress RAF1 fusion mediated oncogenic growth, we next evaluated whether this unresponsiveness is due to inability of the drug to bind and disrupt RAF1 fusion’s protein- protein interactions. We found that PLX8394 was unable to disrupt QKI-RAF1 mediated homodimerization and heterodimerization with wild type B/CRAF or QKI (Fig. 3B). This is in contrast to PLX8394 disrupting the homodimerization of KIAA1549-BRAF fusion (Fig. 3.4c), where we have previously shown that the paradox breaker (PLX8394) is functional in suppressing 75 oncogenicity (A. J. Sievert et al., 2013). This suggests a correlation between dimer inhibition in the presence of PLX8394 and responsiveness to drug, thereby rationalizing why clinically relevant RAF inhibitors do not suppress CRAF fusions but BRAF fusions can be suppressed with paradox breaker drugs.

RAF1 fusions activate oncogenic signaling in a dimerization-dependent manner

QKI-RAF1 and SRGAP3-RAF1 both retain the active RAF1 kinase domain, including specific residues critical for forming the dimer interface of RAF1 (Bollag et al., 2010; Rajakulendran et al., 2009). Point mutations in key dimer interface residues of RAF1 (R401H) have been shown to disrupt RAF1 dimerization with little to no effect on basal kinase activity (Freeman, Ritt, &

Morrison, 2013a, 2013b). We assessed whether mutating the critical dimerization residues of

RAF1 would disrupt the fusion’s dimerization capacity using QKI-RAF1 (denoted as QKI-

RAF1R401H). In co-IP assays, we observed that QKI-RAF1 homodimerization was unaffected with

QKI-RAF1R401H (Fig. 3.6a) indicating that RAF1 fusion partner was not necessary for dimerization. QKI-RAF1R401H showed decreased dimerization with wild type BRAF and CRAF proteins (Fig. 3.6b, c) and also with wild type QKI (Fig. 3.6d) by unknown mechanisms. These findings with the R401H mutants suggest that while RAF1 dimer interface is critical for interaction with wild type B/CRAF, it is only partly responsible for QKI-RAF1 mediated dimerization.

We next assessed the dimerization motif of the N-terminal partner, QKI. The N-terminal QUA1 region of QKI contains dimerization elements and a point mutation E48G in the key dimerization residue results in disruption of QKI dimers and RNA binding function (Ali & Broadhurst, 2013;

Beuck, Qu, Fagg, Ares, & Williamson, 2012). This mutation (QKIE48G-RAF1) robustly decreased homodimerization with QKI-RAF1 (Fig. 3.6a) and heterodimerization with CRAF and QKI (Fig.

3.6 c, d). These data suggest a critical role for the QKI fusion partner in mediating dimerization and downstream functionality of QKI-RAF1. Even CRAF interaction was affected with QKIE48G-

RAF1 whereas CRAF interaction is expected to be mediated by RAF1 and not QKI. This hints at the possibility that QKI-RAF1 forms homodimers followed by CRAF binding to form complex oligomers but this needs to be tested further. It could be possible that due to loss of QKIE48G-

RAF1 homo-dimerization, CRAF does not interact with QKIE48G-RAF1 (Fig. 3.6 c).

Next, we tested double point mutants with mutations in dimerization residues of both QKI and

RAF1 (denoted as QKIE48G-RAF1R401H). In co-IP assays, QKIE48G-RAF1R401H double mutant showed severe loss of interaction with QKI-RAF1, wild type BRAF, CRAF and QKI (Fig. 3.6).

Overall, these results suggest that the N-terminal fusion partner (QKI) primarily mediates RAF1 fusion’s dimerization with some contribution of RAF1 as mutating dimerization residues in both

QKI and RAF1 results in robust loss of all protein-protein dimerizations.

We further tested the effect of QKI, RAF1 and double dimer mutants on oncogenic signaling and tumorigenic phenotype of QKI-RAF1. We transduced mouse astrocyte cells with various dimer mutant constructs and tested downstream MAPK signaling. As compared to QKI-RAF1, QKI-

RAF1R401H showed similar phospho-MEK, phospho-ERK and phospho-S6 levels whereas

QKIE48G-RAF1 and QKIE48G-RAF1R401H had robust reduction in MAPK signaling (Fig. 3.6e).

QKIE48G-RAF1R401H also demonstrated reduction in phospho-S6 beyond the basal level. These results strongly suggest the importance of QKI fusion partner driving QKI-RAF1 dimerization and subsequent activation of oncogenic MAPK signaling. When tested in soft agar assays for tumorigenic potential, QKI-RAF1R401H showed a decrease in colony formation but was not significantly different from QKI-RAF1 (Fig. 3.6f). In contrast, QKIE48G-RAF1 and QKIE48G-

RAF1R401H displayed little anchorage-independent growth compared to QKI-RAF1 (Fig. 3.6f, p- value <0.01), demonstrating that QKI-RAF1 driven soft agar growth is primarily dependent on 77

QKI dimerization residues. Hence, disruption of QKI-driven dimerization would be required to suppress oncogenic properties of QKI-RAF1 expressing PLGGs.

We also evaluated effects of R401H mutation on SRGAP3-RAF1. MAPK signaling was decreased in mouse astrocyte cells expressing SRGAP3-RAF1R401H as compared to SRGAP3-

RAF1 but was more than baseline vector control (Fig. 3.5b). In soft agar assays SRGAP3-

RAF1R401H showed a significant decrease in oncogenic colony formation indicating the role RAF1 dimerization motif in SRGAP3-RAF1 (Fig. 3.5c).

N-terminal fusion partner looses functionality in RAF1 fusions

The loss of RNA binding function of QKI has been associated with increased oncogenicity in adult glioblastomas (A. J. Chen et al., 2012) and lung cancers (Zong et al., 2014) and decreased

QKI expression has been correlated with occurrence of prostate cancer (Y. Zhao et al., 2014) and gastric cancer (Bian et al., 2012). Since part of the QKI RNA binding motif is retained in QKI-

RAF1, we tested whether QKI has lost its RNA binding function in QKI-RAF1 and how that could correlate with oncogenicity in PLGGs. We performed UV cross-linking experiments where well-known RNA targets of QKI, myelin basic protein (MBP) and early growth response gene-2

(EGR-2) (Galarneau & Richard, 2005), were radiolabeled and incubated with vector control, full length QKI and QKI-RAF1. Compared to negative vector control, QKI bound to both RNA targets but QKI-RAF1 failed to bind either RNA target (Fig. 3.6g, 3.7) thereby indicating that

QKI-RAF1 has lost RNA binding function. We have previously shown that partial loss of QKI expression in MYB-QKI expressing PLGGs leads to increased oncogenicity (Bandopadhayay et al., 2016) and such a mechanism could also be possible in QKI-RAF1 expressing tumors. These results suggest the possible contribution of QKI loss of function in mediating QKI-RAF1 oncogenicity.

SRGAP3 has also been reported as a tumor suppressor-like gene with low expression in breast cancer cells (Lahoz & Hall, 2013). Loss of the functional Rho-GTPase domains in SRGAP3-

RAF1 suggests a loss of function for SRGAP3, similar to QKI in QKI-RAF1. This needs to be further explored as a common theme for N-terminal fusion partners contributing to RAF1 fusion’s oncogenicity.

RAF1 fusions are sensitive to LY3009120, a RAF dimer inhibitor

Since dimerization is a key component of RAF1 fusion’s oncogenicity, we investigated the effect of novel dimer inhibitor LY3009120 that has recently shown efficacy in suppressing mutant

BRAF dimers in adult cancers (S. H. Chen et al., 2016; Henry et al., 2015; Peng et al., 2015).

LY3009120 is a pan-RAF inhibitor with potent anti-tumor effects and does not result in paradoxical activation unlike vemurafenib. In stable NIH3T3 cells, LY3009120 showed potent, dose-dependent inhibition of both MAPK and PI3K pathways driven by QKI-RAF1 and

SRGAP3-RAF1 (Fig. 3.8a). Even low doses of LY3009120 were successful at suppressing anchorage independent growth driven by RAF1 fusions in NIH3T3 cells (Fig. 3.8b, p- value<0.001). Similar suppression of MAPK pathway was observed at high concentrations of

LY3009120 in RAF1 fusion expressing mouse astrocyte cell lines (Fig. 3.8c) whereas PI3K pathway was not abolished as seen in NIH3T3 cell lines. Nevertheless, LY3009120 potently suppressed soft agar colony growth of mouse astrocyte cell lines at low doses (Fig. 3.8d) verifying the specific anti-tumor effects of LY3009120 on RAF1 fusion expressing cell lines.

LY3009120 induces RAF1 fusion homodimerization and heterodimerization with B/CRAF but prevents QKI heterodimerization

To further evaluate the inhibitory mechanism of LY3009120 in RAF1 fusions, we performed drug co-IP assays as before. Treatment with 10μM of LY3009120 induced QKI-RAF1 79 homodimers and QKI-RAF1/BRAF heterodimers (Fig. 3.8e) with minimal effect on QKI-

RAF1/CRAF heterodimers. Interestingly, we observed complete disruption of QKI-RAF1/full- length QKI heterodimerization in the presence of LY3009120 but interaction with N-terminal truncated QKI is retained (Fig. 3.8e). These results suggest that while LY3009120 induces and stabilizes QKI-RAF1 homo-dimers and hetero-dimers with other RAF proteins, it suppresses the

RAF1 kinase domains and prevents downstream MAPK/PI3K signaling (Fig. 3.8a, c). However,

LY3009120 disrupts interaction of full length QKI via some unknown mechanism suggesting that

LY3009120 can inhibit QKI dimerization residues in QKI-RAF1, subsequently decreasing QKI-

RAF1 dimerization and downstream MAPK/PI3K signaling. The unique ability of dimer inhibitor

LY3009120 to suppress RAF1 fusions, where existing RAF inhibitors (PLX4720 and PLX8394) fail, could be used as a therapeutic option that warrants more testing.

RAF1 fusions are partially suppressed by MAPK pathway inhibitors in vivo

Because of the lack of preclinical efficacy of existing RAF inhibitors in RAF1 fusions, we next tested effects of clinically available MAPK pathway inhibitors, AZD6244 and trametinib

(GSK1120212), as single agent therapies against RAF1 fusions. NIH3T3 cell lines over- expressing QKI-RAF1 and SRGAP3-RAF1 showed a dose dependent decrease in phospho-ERK levels with AZD6244 indicating on-target inhibition of MEK1/2 (Fig. 3.9 a, d respectively, left panels). In soft agar assays, AZD6244 showed robust growth inhibition against NIH3T3 cells expressing QKI-RAF1 and to a lower extent for SRGAP3-RAF1 (Fig. 3.9 b, e respectively, top panel, p-value<0.01, 0.1). We also evaluated a more potent MEK inhibitor and FDA approved

MEK inhibitor, Trametinib. We observed robust decrease in phospho-ERK levels with trametinib treatment in NIH3T3 cells expressing QKI-RAF1 and SRGAP3-RAF1 (Fig. 3.9 a, d respectively, middle panels) and also in mouse astrocyte cell lines stably transduced with QKI-RAF1 and

SRGAP3-RAF1 (Fig. 3.9 a, d respectively, right panels). Even low doses of trametinib were 80 successful in inhibiting anchorage independent growth driven by QKI-RAF1 and SRGAP3-RAF1 in NIH3T3 cells (Fig. 3.9 b, e respectively, top panels, p-value<0.01) and in mouse astrocyte cell lines (Fig. 3.9 b, e respectively, bottom panels, p-value<0.01). These in vitro studies suggest that trametinib can suppress MAPK signaling and in vitro colony formation driven by RAF1 fusions; hence we tested this mono-therapy in vivo.

We used stable NIH3T3 cell lines expressing RAF1 fusions to form flank xenograft tumors in immuno-compromised mice and treated mice daily with trametinib at 1 mg/kg. Despite significant activity in vitro, Trametinib mono-therapy only partially suppressed QKI-RAF1 and

SRGAP3-RAF1 tumor growth (Fig. 3.9 c, d, respectively). The recurrence of tumors in the setting of continued trametinib treatment suggests that alternate resistance pathways were activated by RAF1 fusions in vivo.

Combinatorial targeting of MAPK and PI3K/mTOR pathways is more potent in suppressing RAF1 fusions than single agent therapies

Because we observed PI3K/mTOR pathway activation in RAF1 fusion expressing cell lines, we sought to test combination therapy targeting both MAPK and PI3K pathways for RAF1 fusions using MEK inhibitor trametinib and PI3K/mTOR pathway inhibitor RAD001 (Everolimus).

NIH3T3 and mouse astrocyte cells expressing QKI-RAF1 (Fig. 3.10a, left and right panel respectively) revealed decrease in phospho-ERK and phospho-S6 at very low doses of trametinib and 10μM RAD001 indicating targeting of oncogenic signaling. Similar decrease in MAPK and

PI3K/mTOR signaling was observed in SRGAP3-RAF1 expressing NIH3T3 and mouse astrocyte cells (Fig. 3.10c, left and right panel respectively).

To test the effect of combinatorial treatment on oncogenic growth inhibition in vitro, we performed soft agar assays with RAF1 fusion expressing NIH3T3 and mouse astrocyte cells. In 81 both cell model systems, combinatorial targeting was successful at inhibiting anchorage independent growth mediated by RAF1 fusions (Fig. 3.10 b, d).

We then evaluated sensitivity to combinatorial therapy in in vivo models with NIH3T3 flank xenografts. As seen in Figure 3.10 e, mice with QKI-RAF1 xenografts treated with single-agent

RAD001 or trametinib show initial tumor suppression but tumor eventually resurges. QKI-RAF1 xenograft mice treated with trametinib and RAD001 at 1mg/kg and 10mg/kg, respectively show prolonged suppression of tumor growth (Fig. 3.10e). We observed similar responsiveness in

SRGAP3-RAF1 expressing NIH3T3 flank xenografts, with single-agent therapy against MAPK or PI3K/mTOR leading to only partial tumor suppression and combinatorial targeting causing more prominent, prolonged tumor regression (Fig. 3.10f). These in vivo results strongly suggest the significance of using combinatorial therapy with trametinib and RAD001 for RAF1 fusion expressing PLGGs as targeting both signaling pathways would be essential for effective targeted therapy.

Discussion

In this study, we describe the oncogenic mechanism of action and therapeutic sensitivity of

PLGG associated RAF1 gene fusions. To our knowledge, this is the first report to distinguish between BRAF and CRAF/RAF1 gene fusions, both found in the grade I PLGG, pilocytic astrocytomas (PAs) (J. Zhang et al., 2013). While BRAF fusions result in paradoxical activation with vemurafenib, a first generation RAF inhibitor, second generation PLX8394 suppresses

BRAF fusion driven oncogenesis. Interestingly, RAF1 fusions are not suppressed by either vemurafenib or PLX8394 but are responsive to novel dimer inhibitor LY3009120 or combinatorial targeting with MAPK and PI3K/mTOR pathway inhibitors, highlighting important clinical implications. We also report a critical role of the N-terminal fusion partner, QKI in

82 mediating QKI-RAF1 fusion’s oncogenic nature and resistance to existing RAF inhibitors, a finding that might be relevant to pan-cancer RAF1 gene fusions.

The BRAFV600E mutation is found in 6-10% of PAs with other subtypes displaying higher incidences (Olow et al., 2016) whereas the KIAA1549-BRAF fusion has been reported in majority of PAs, and novel BRAF mutations have also being discovered by deep sequencing (J. Zhang et al., 2013). The overall frequency of BRAF mutations is much higher in PAs with a small fraction of patients containing RAF1 gene fusions, however their overall prevalence remains unknown.

Since both RAF1 fusions and BRAF fusions occur in the same PLGG sub-type (J. Zhang et al.,

2013), patients with a clinical diagnosis of PAs might not have the same molecular diagnosis.

This necessitates development of molecular stratification schemes for PLGGs as patients with similar histological diagnosis might have different molecular drivers.

We found QKI-RAF1 and SRGAP3-RAF1 to be oncogenic proteins in different cellular contexts that drive robust in vivo flank xenograft and intra-cranial tumor growth. Relevant patient-derived models are extremely rare for PLGGs, with the few reported cell lines not expressing RAF1 fusions (Bax et al., 2009; Olow et al., 2016). In our model systems, RAF1 fusions were found to aberrantly activate both the MAPK and PI3K/mTOR pathways. Despite similar downstream signaling to BRAF fusions, RAF1 fusions were found to function differentially and exist as RAF inhibitor resistant homodimers and heterodimers with B/CRAF. This is in contrast to KIAA1549-

BRAF fusion that exists solely as homodimers (A. J. Sievert et al., 2013) that we found to be disrupted in response to PLX8394. Alternatively, BRAFV600E exists as monomers that are targetable by both vemurafenib and PLX8394 (Freeman et al., 2013a; Poulikakos et al., 2011; C.

Zhang et al., 2015). Differences in response to RAF inhibitor therapy and BRAF mutational status has been reported previously but our findings suggest unique differences between BRAF and CRAF fusions based on dimerization status, primarily mediated by the non-kinase fusion 83 partner in RAF1 fusions. These mechanistic findings have important implications when assessing therapeutic options for patients with RAF1 fusions, as clinically available RAF inhibitors would be unsuccessful in disrupting oncogenic dimers of RAF1 fusions, rendering patients resistant to therapy.

LY3009120 is a novel RAF dimer inhibitor drug that has shown promise in inhibiting models with various NRAS, KRAS and BRAF mutations (S. H. Chen et al., 2016; Peng et al., 2015). We demonstrate ability of LY3009120 to disrupt QKI mediated interaction between QKI-RAF1 fusions as well as stabilize inactive homodimers and heterodimers with B/CRAF. This strongly supports our hypothesis that RAF inhibitor can be effective against RAF1 fusions if would it targets dimerization mediated by both fusion partners. Due to distinct binding properties of

LY3009120, its clinical use could extend from BRAF mutant cancers to RAF1 fusion expressing cancers but this would require further in vivo testing.

The RAS/RAF/MEK signaling cascade is altered in a variety of cancers leading to the successful development of potent targeted inhibitors for the downstream kinases, MEK1/2, such as

Selumetinib (AZD6244) and Trametinib (GSK1120212) (Gilmartin et al., 2011; Henne et al.,

2007; Yamaguchi, Kakefuda, Tajima, Sowa, & Sakai, 2011). While AZD6244 monotherapy has been tested against adult cancers (Catalanotti et al., 2013) and is also currently in clinical trial for

PLGGs (ClinicalTrials.gov Identifier: NCT01089101), we have shown limited response to

AZD6244 monotherapy in RAF1 fusion expressing cell models. This led us to focus efforts on the more potent Trametinib that has also been approved for advanced, refractory melanomas

(Flaherty et al., 2012). In BRAF mutant cancers, monotherapy with MEK inhibitor is more potent than RAF inhibitors but multi-agent therapy combining MEKin with either Vemurafenib or mTOR inhibitor (Everolimus) has shown the most promising pre-clinical results (Olow et al.,

2016). In our study, Trametinib monotherapy showed partial suppression of RAF1 fusion driven 84 tumor growth, highlighting the importance of the parallel PI3K/mTOR pathway in RAF1 fusion driven PLGGs. Therefore, we tested combinatorial therapy using MEKin Trametinib and mTORin Everolimus (RAD001) and found combinatorial treatment to be most successful at prolonged suppression of QKI-RAF1 and SRGAP3-RAF1 driven tumors. This provides a strong rationale for clinical use of dual therapy for PLGGs with RAF1 fusion. Safety and tolerability of combining MEK- and mTOR-inhibitors is currently being in adult, advanced cancers

(ClinicalTrials.gov Identifier: NCT01347866) with pediatric testing in the future (Olow et al.,

2016).

Most studies on kinase gene fusions have focused on the downstream effects of truncated, activated kinases, whereas our study depicts a novel oncogenic role for the non-kinase fusion partner. We found that stable dimerization of QKI-RAF1 is mediated by QKI and is essential for resistance to RAF inhibitor therapy. This adds to the growing list oncogenic roles being attributed to QKI by recent studies including loss of expression/function and altered regulation of target

RNAs (Abedalthagafi et al., 2015; Bandopadhayay et al., 2016; Conn et al., 2015; Danan-

Gotthold et al., 2015) and its growing importance in regulating brain tumors. However, the role of

N-terminal fusion partner might be gene-specific, as the ESRP1-RAF1 fusion exists as a constitutive dimer where dimerization is mediated by RAF1 and R401H mutation causes loss of

MAPK signaling (Palanisamy et al., 2010; Yao et al., 2015).

The distinct sub-types of pediatric low-grade gliomas are beginning to be associated with specific mutations such as the association of MYB-QKI with angiocentric gliomas and KIAA1549-BRAF fusions with pilocytic astrocytomas but this study adds another layer of complexity with same subtype displaying different mutations and distinct therapeutic responses. We highlight the importance of molecular classification of PLGGs in assisting targeted therapeutic approaches.

Materials and Methods

Cell Culture: NIH3T3 and HEK 293 cells were obtained from ATCC and maintained in 1X

DMEM media containing 10% DBS and 10% FBS respectively. We obtained Tp53-null early- passage primary mouse astrocytes (PMAs) from Suzi Baker’s group at St. Jude Children’s

Hospital, and these cells were maintained in DMEM- F12 media supplemented with 10% FBS and EGF. Cell lines were not re-authenticated and were routinely tested (at least every three months) for mycoplasma infection.

Vector Construction and Generation of Stable Cell Lines: SRGAP3-RAF1 and QKI-RAF1 constructs were synthesized as Gateway compatible entry clones. Full length RAF1, QKI and

SRGAP3 plasmids were purchased as gateway entry clones from PlasmID/DF/HCC DNA

Resource Core. Sub-cloning was done to integrate SRGAP3-RAF1, QKI-RAF1, full-length QKI,

RAF and SRGAP3 constructs into a Gateway-compatible N-MYC-tagged pMXs-Puro Retroviral

Vector (Cell Biolabs). Empty MYC-tagged retroviral construct was used as control. Retrovirus was generated using Platinum-E retroviral packaging cells (Cell BioLabs) as per manufacturer protocols. NIH3T3 cells and early-passage PMAs were transduced with retroviral media for 6 hours and infected cells were selected with puromycin 48 hours post infection. Stable expression of MYC-tagged proteins was confirmed via western blot analysis (anti-MYC HRP 1:5000;

Invitrogen). For protein-protein interaction assays, we also cloned the above-mentioned constructs into Gateway destination vectors with either an N-terminal MYC tag or N-terminal

FLAG tag (Invitrogen).

Dimerization mutants of QKI-RAF1 were generated by PCR-based site-directed mutagenesis of

MYC- and FLAG-tagged constructs described above. RAFR401H dimerization mutants (Freeman et al., 2013a, 2013b) in QKI-RAF1 were generated using primers: Forward

CGCAAAACACACCATGTGAACA and Reverse CAGAACAGCCACCTCATTCCT. QKIE48G 86 dimerization mutants in QKI-RAF1 were generated using primers: Forward

CTGGACGAAGGAATTAGCAGAG and Reverse CAGCCGCTCGAGGTGGTT. QKIE48G dimerization mutants (T Chen & S Richard, 1998) in QKI-RAF1 were generated using primers:

Forward CTGGACGAAGGAATTAGCAGAG and Reverse CAGCCGCTCGAGGTGGTT.

Cellular kinase drug assays: For pathway suppression experiments using RAF inhibitors,

PLX4720, a first-generation RAF inhibitor, or PLX8394/PB-3, a second-generation paradox- breaking RAF inhibitor (Plexxikon) or LY3009120, a pan-RAF dimer inhibitor (Eli Lilly; drug purchased from Selleck Chemicals) were used at indicated concentrations. The MEK inhibitor,

GSK1120212 was provided by GlaxoSmithKline, and AZD6244 (Astra Zeneca) was purchased from Selleck Chemicals. These drugs were used alone or in combination with RAD001 (mTOR inhibitor). All drug aliquots were dissolved in DMSO and stored at −20 °C. Cells were plated at

1x106 cells/ml in six-well dishes with serum-containing media. Before drug treatment, cells were serum starved. Thereafter, cells were exposed to indicated drug concentrations for indicated time periods.

Western Blot Analysis: Stable NIH 3T3 cells and stable PMA cell lines used in drug assays were lysed in 1X RIPA (50 mM Tris, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1%

Nonidet P-40) containing protease and phosphatase inhibitor cocktail (Thermo Fisher).

Supernatant from centrifuged lysates were normalized for protein concentration using a Pierce

660nm protein assay and run on NuPAGE precast gels (4–12% Bis-Tris; Life Technologies). For

MAPK pathway analysis, pMEK, MEK, pERK, ERK antibodies from Cell Signaling were used at

1:1000 dilutions in BSA. To analyze PI3K/mTOR pathway, pAKT Ser473, pAKT Thr308, AKT, pS6, S6 antibodies from Cell Signaling were used.

Soft-Agar Cellular Transformation Assays and quantification: Anchorage-independent growth of SRGAP3-RAF1 and QKI-RAF1 expressing stable NIH 3T3 cells and PMAs was assessed as previously described (Bandopadhayay et al., 2016; Angela J Sievert et al., 2013).

Images were analyzed using ImageJ and colonies with area greater than 500 pixels quantified.

Co-immunoprecipitation Assays: Interactions of SRGAP3-RAF1 and QKI-RAF1 with itself, wild-type BRAF, CRAF, QKI or SRGAP3 were assessed via co-transfections of MYC- and

FLAG-tagged constructs into HEK 293 cells using Lipofectamine 2000. Anti-MYC tag antibody coated magnetic beads (MBL) were used to immunoprecipitate the transfected proteins from cell lysates in 1X RIPA at 4 °C over-night. Beads were washed 3 times for 15 minutes each at 4 °C to remove non-specific interactions, followed by a final 1xPBS wash. Proteins were eluted off beads using 2xLDS and heating at 70°C for 10 minutes and used for in western blot analysis. Anti-

MYC antibody (Invitrogen, 1:5000) and anti-FLAG antibody (Thermo Fisher, 1:10,000) were used to assess co-immunoprecipitation.

Animal studies: The Children’s Hospital of Philadelphia Institutional Animal Care and Use

Committee approved all animal protocols. Homozygous NSG strain immuno-deficient mice were bred in our animal facility and housed under aseptic conditions. Mice were 6-10 weeks of age, with equal representation of male and female mice.

Mouse flank xenograft studies with NIH3T3 Stable Cell Lines: To assess in vivo tumor growth,

NIH3T3 cell lines were injected subcutaneously into the flanks of NSG mice (10 mice for each cell line, per drug condition) and tumor growth was measured. The GSK1120212 and RAD001 combinatorial drug study was performed in a xenograft mouse model by pre-treating with daily oral gavage (1 mg/kg/dose of GSK1120212 and 10mg/kg of RAD001) for one week prior to injecting cell lines subcutaneously into the flanks of NSG mice. Tumor growth was measured

88 daily while on drug treatment. Ellipsoid tumor volume was calculated using the formula: volume

= 1/2(length × width2).

Intracranial mouse injections: 1x106 PMAs expressing SRGAP3-RAF1, QKI-RAF1 and vector control were resuspended in Matrigel basement matrix (BD Sciences) and 2 microliter was injected stereo-tactically into the right striatum of immuno-compromised NSG mice (5 mice each). Animals were monitored and sacrificed at the onset of neurological symptoms, QKI-RAF1 at day 39 post-injection and SRGAP3-RAF1 at day 42 post-injection. Brains were isolated post intra-cardiac perfusion with 4% paraformaldehyde solution and stored in 4% paraformaldehyde solution in 1X PBS at 4°C over-night, and subjected to routine histological analysis.

UV cross-linking and co-immunoprecipitation assays: To assess the RNA binding capacity of

QKI in QKI-RAF1, known RNA targets of QKI were used- myelin basic protein (MBP) and early growth Response 2 (EGR-2) (Galarneau & Richard, 2005). In vitro transcription of P32 radiolabeled RNA probes (MBP and EGR-2) and UV cross-linking with QKI or QKI-RAF1 expressing HEK293 lysates was done as described previously (Lynch & Maniatis, 1996;

Rothrock, House, & Lynch, 2005). Cross-linked protein-RNA conjugates were immuo- precipitated using anti-Myc tagged beads and separated on an SDS-PAGE gel.

Tables and Figures Figure 3.1

Figure 3.1. QKI-RAF1 and SRGAP3-RAF1 are oncogenic via activation of MAPK and PI3K pathways. (a) Structure of RAF1 fusions in PLGGs. In QKI-RAF1, exons 1-3 of QKI contain QUA1 dimerization domain and a truncated K- homology domain (KH-Tr), and exons 8-17 of RAF1 retain the protein kinase domain. In SRGAP3-RAF1, exons 1-10 of SRGAP3 contain the Fes/CIP4-Homology (FCH) domain or also called the F-BAR homodimerization domain and, RAF1 exons 9-17 encode the RAF1 kinase domain. (b) Table showing different RAF1 fusions present in various adult cancers and pediatric cancer. (c) Soft agar assay using NIH3T3 cell lines stabling expressing RAF1 fusions. Mean of three measurements is depicted. Error bars represent SEM. (d) Western blot analysis of MAPK and PI3K pathway in stable NIH3T3 cell lines under serum starvation. (e) Soft agar assay using p53-null mouse astrocyte cells (PMAs) stabling expressing RAF1 fusions. Mean of five measurements is depicted. Error bars represent SEM. (f) Western blot analysis of MAPK and PI3K pathway in stable PMA cell lines under serum starvation. (g) Flank xenograft tumor measurements with stable NIH3T3 cell lines expressing RAF1 fusions. Mean of ten measurements is depicted. Error bars represent SEM. (h) Kaplan- Meier survival plot of NSG mice orthotopically injected with PMAs overexpressing RAF1 fusions develop tumors with short latency compared to PMAs expressing vector control which never developed tumors, n=5, p<0.01. p-value * <0.05, ** <0.01, *** <0.001.

Figure 3.2

pMEK pERK pS6 tMEK tERK tAKT tS6 PLX4720 PLX8394 0,0.1,1,10 µM 0,1,5,10 µM KIAA1549-BRAF- NIH 3T3

Figure 3.2. Effect of RAF inhibitors on KIAA1549-BRAF signaling. Western blots showing paradoxical activation of PLX4720 on growth pathways regulated by KIAA1549-BRAF expressing NIH 3T3 cells, PLX8394 can suppress KIAA1549-BRAF signaling.

Figure 3.3

A. B. C. D. ) ) pMEK pMEK 2 l 150 2

l PLX4720/Vemurafenib e e 150 x pERK i x pERK

i PLX4720/Vemurafenib

p PLX8394/PB3

S473 PLX8394/PB3 S473 0 0 pAKT pAKT 0 0 5 100 5 >

T308 100 T308 ( > pAKT ( pAKT r

r e e b

pS6 b pS6 m m u 50

u 50 tMEK n

tMEK n

y y n

tERK n

tERK o l o l o o tAKT 0 tAKT C 0 C 0 0.1 1 10 0 1 5 10 0 0.1 1 10 0 1 5 10 tS6 tS6 Drug concentrations (µM) Drug concentrations (µM) PLX4720 PLX8394 PLX4720 PLX8394 0,0.1,1,10 µM 0,1,5,10 µM 0,0.1,1,10 µM 0,1,5,10 µM

QKI-RAF1- NIH 3T3 QKI-RAF1- NIH 3T3 SRGAP3-RAF1- NIH 3T3 SRGAP3-RAF1- NIH 3T3

E. F. G.

pMEK )

PLX4720 )

2 PLX4720

80 2 80 l l e pERK PLX8394/PB3 e x x PLX8394/PB3 i i p p

S473 0 pAKT 60 0 60 0 0 5 T308 5 > > ( pAKT (

r r e 40 e 40 b

pS6 b m m u tMEK u n n

20 20 y tERK y n n o o l l o tAKT o C 0 C 0 tS6 0 0.1 1 10 0 1 5 10 0 0.1 1 10 0 1 5 10 QKI-RAF1 SRGAP3-RAF1 Drug concentrations (µM) Drug concentrations (µM) PLX8394 0,1,5,10 µM Mouse astrocyte lines QKI-RAF1- mouse astrocyte lines SRGAP3-RAF1- mouse astrocyte lines

Figure 3.3. Existing RAF inhibitors do not suppress CRAF fusion driven signaling pathways and oncogenic phenotype. Current RAF inhibitors, PLX8394 and PLX4720, have minimal effect of on (a) growth pathways and (b) soft agar growth regulated by QKI-RAF1 expressed in NIH 3T3 cells. Similar unresponsiveness seen on (c) growth pathway and (d) soft agars driven by SRGAP3-RAF in NIH 3T3 cells. (e) p53-/- mouse astrocyte cell lines (PMAs) with QKI-RAF1 and SRGAP3-RAF1 are unresponsive to current RAF inhibitors. Soft agar growth of PMAs expressing (f) QKI-RAF1 and (g) SRGAP3-RAF1

Figure 3.4

A. QKI-RAF1 F/M F F F F F BRAF - M - - - - RAF1 - - M - - - Vector Control - - - M - - QKI - - - - M - QKI.Truncated - - - - - M

WB:Flag IP:Myc WB:Myc

Lysates WB:Flag

B. PLX8394 (PB3) 10µM PLX8394 (PB3) QKI-RAF1 F/M F F F F/M F F F 10µM BRAF - M - - - M - - QKI-RAF1 F F F F RAF1 - - M - - - M - QKI M - M - Vector Control - - - M - - - M QKI.Truncated - M - M

WB:Flag WB:Flag IP:Myc IP:Myc WB:Myc WB:Myc

Lysates WB:Flag Lysates WB:Flag

C. PLX8394 (PB3) 10µM KIAA1549-BRAF F/M F F/M F Vector Control - M - M

WB:Flag IP:Myc WB:Myc

Lysates WB:Flag

Figure 3.4. Homo- and hetero-dimerization is required to drive CRAF fusion driven gliomagenesis. (a) Co-immunoprecipitation assays show QKI-RAF1 homo-dimerization with itself, and hetero-dimerization with wild type BRAF, CRAF and QKI in HEK 293t cells. (b) Presence of ‘paradox breaker’ RAF inhibitor, PLX8394, does not affect QKI-RAF1 dimerization in co-immunoprecipitation assays in HEK 293t cells. (c) KIAA1549-BRAF homo-dimerization is suppressed by PLX8394.

Figure 3.5

H d

A. B. 1

1 1

n F

SRGAP3-RAF1 F/M F F F F

n R R T

BRAF - M - - d

- - .

e o

3 3 3 3 a c

- - M C

P P P - P

RAF1 n r u

A A A o T

G G G - - - G

Vector Control M R c .

R R R e

V S S S WB:Flag Q S pMEK IP:Myc WB:Myc pERK

pS6

Lysates WB:Flag tMEK

tERK

tS6

C. )

2 15 l

e ** x i p

0 0

5 10 > (

r N.S. e b m

u 5 n

y n o l o

C 0 Vector RAF1 RAF1R401H SRGAP3SRGAP3- R401H Control -RAF1 RAF1

Figure 3.5. SRGAP3-RAF1 mediated dimerization and oncogenicity is partly mediated by RAF1. (a) SRGAP3-RAF1 homo-dimerizes with itself, and hetero-dimerizes with wild type BRAF and CRAF. (b) Single dimerization mutants of SRGAP3-RAF1 (RAF1R401H) expressed in PMAs show reduction in MAPK pathway. (c) Soft agar assays with SRGAP3- RAF1R401H shows decrease in oncogenicity.

Figure 3.6

4 A. B. C. D. BRAF F F F F RAF-1 F F F F QKI F F F F QKI-RAF1 F/M F F F QKI-RAF1 M - - - QKI-RAF1 M - - - QKI-RAF1 M - - - R401H - M - - - M - - R401H QKI-RAF1 QKI-RAF1R401H - M - - QKI-RAF1R401H QKI-RAF1 - M - - E48G QKI -RAF1 - - M - QKIE48G-RAF1 - - M - QKIE48G-RAF1 - - M - QKIE48G-RAF1 - - M - E48G R401H E48G R401H QKI -RAF1 - - - M QKI -RAF1 - - - M QKIE48G-RAF1R401H - - - M QKIE48G-RAF1R401H - - - M WB:Flag IP:Myc WB:Myc

Lysates WB:Flag

Over-expression in HEK293 cells

E. F. G.

H 1

0 RNA probe: MBP EGR-2

1 1

1 TE IP TE IP

F F 4

t 100kDa -

A A

1 1

o 75kDa -

R R

- -

F F

G G G

A **

8 8 8 R

r Vector

4 4 4

R R

1 1

o 50kDa -

- - E

E E

t **

I I I I I I

F F

c Control

)

K K K K

K K

A A

2 e

l 60 N.S.

V Q Q Q Q Q Q R R

e 37kDa - x i p

pMEK 0

0 100kDa - pERK 5 40 > 75kDa - (

e QKI pS6 b 50kDa - m

u 20

tMEK n

37kDa - y n o tERK l 100kDa - o 75kDa - C 0 tS6 Vector QKI- E48G QKIE48G- QKI-RAF1 QKI QKI-RAF1 50kDa - Mouse astrocyte lines Control RAF1R401H -RAF1 RAF1R401H 37kDa - Mouse astrocyte lines Figure 3.6. RAF1 fusions activate oncogenic signaling in a dimerization-dependent manner. Co-IP assays using single dimerization mutants of QKI (QKI E48G-RAF1) and RAF1 (QKI-RAF1R401H) and double dimerization mutants (QKI E48G -RAF1R401H) of QKI-RAF1 against (a) QKI-RAF1, (b) BRAF, (c) CRAF/RAF-1 and (d) QKI. PMAs expressing QKI-RAF1 mutants were assessed for (e) MAPK and PI3K pathway signaling and (f) soft agar growth. (g) UV crosslinking assay with wild type QKI and QKI-RAF1 assessing RNA binding to probes, MBP and EGR-2.

Figure 3.7

t R

- -

y e

QKI - NT RI K2

V M 75kDa - 67kDa- 40kDa -

WB:Myc

Figure 3.7. Expression level of QKI and QKI-RAF1 in UV cross-linking assay. Expression level in UV cross-linking experiment

Figure 3.8

C. A. B. C. LY3009120 (0, 1, 5,10 µM) LY3009120 (0, 1, 5,10 µM) 30 QKI-RAF1 )

20 2 l pMEK

pMEK e x i 10 p pERK pERK 0 *** 0

S473 5 0 *** *** pAKTS473

pAKT > (

0 1 5 10 r T308 pAKT T308 e LY3009120(µM) pAKT b m

pS6 u pS6 n SRGAP3-RAF1 y 30 tMEK n tMEK o l o tERK tERK C 20 tAKT tAKT 10 *** tS6 tS6 *** *** 0 QKI-RAF1 SRGAP3-RAF1 QKI-RAF1 SRGAP3-RAF1 0 1 5 10 NIH3T3 cell lines LY3009120 (µM) Mouse astrocyte lines D. 80 E. QKI-RAF1 60

) LY3009120

l 10µM e 40 LY3009120 x i

p 10µM QKI-RAF1 F/M F F F F/M F F F

0 20 0

5 *** BRAF - M - - - M - - QKI-RAF1 F F F F > 0 *** *** (

r 0 1 5 10 CRAF - - M - - - M - QKI M - M - e

b LY3009120(µM) Vector Control - - - M - - - M QKI.Truncated - M - M

m 80 u n WB:Flag WB:Flag y

n 60 SRGAP3-RAF1

o IP:Myc

l IP:Myc o

C 40 WB:Myc WB:Myc

20 WB:Flag *** Lysates Lysates WB:Flag 0 *** *** 0 1 5 10 LY3009120(µM) Over-expression in HEK293 cells Mouse astrocyte lines Figure 3.8. Novel RAF dimer inhibitor LY3009120 stabilizes and inactivates CRAF fusion dimers. Western blot showing suppression of QKI-RAF1 and SRGAP3-RAF1 mediated growth pathways in (a) NIH 3T3 cells and (c) PMAs. Soft agar assays showing robust response to LY3009120 in (b) NIH 3T3 cells and (d) PMAs. (e) Co-IP assays showing effects of LY3009120 on QKI-RAF1 dimerization in HEK 293t cells.

Figure 3.9

A. B. C. AZD6244 Trametinib 80 AZD6244 0, 0.005,0.01,0.1 µM 0, 0.005,0.01,0.1 µM Trametinib 60 2 Vehicle Control

pMEK ) ) 2 40 3 l Trametinib 1mg/kg e m x i pERK c 1.5

** ( p 20

e S473 0 ** 0 pAKT ** ** m

5 ** 0 u l > 0 5 0 5 1 1 1 1 1 . . ( T308 o 0 0 0 0

. 0 . 0

pAKT r V 0 0

. . 0 0 e r * b pS6 Drug concentrations (µM) o m m 0.5 u 80 u * n T tMEK y

n 60 * o * tERK l 0 o 1 14 16 18 21 24 28 tAKT C 40 Days post flank injection of 20 tS6 ** ** ** QKI-RAF1 NIH 3T3 cell QKI-RAF1- NIH 3T3 QKI-RAF1- 0 0 0.005 0.01 0.1 Mouse Astrocyte Trametinib(µM) D. E. AZD6244 Trametinib 150 AZD6244 F. 0, 0.005,0.01,0.1 µM 0, 0.005,0.01,0.1 µM Trametinib 100

) 3 2

l * Vehicle Control

pMEK )

e * 3 x

i * Trametinib * 2.5

p 50 m pERK c 1mg/kg 0 (

0 e

5 2 S473 ** > pAKT m (

0 u l r 0 0 1 1 5 5 1 1 . . o T308 e 1.5 0 0 0 0 0 0 . . b

pAKT V 0 0

. . 0 0 r m o pS6 u Drug concentrations (µM) 1 n m

u y T tMEK n 60 0.5

o * l * o 40 * tERK C 0 * * 20 1 14 16 18 21 24 28 30 tAKT ** ** ** 0 Days post flank injection of tS6 0 0.005 0.01 0.1 SRGAP3-RAF1 NIH 3T3 cells SRGAP3-RAF1- NIH 3T3 SRGAP3-RAF1- Trametinib(µM) Mouse Astrocyte

Figure 3.9. MEK inhibitors, AZD6244 and GSK1120212, partially suppress QKI-RAF1 mediated tumor growth. (a) Westerns showing effect of MEK inhibitors on growth pathways by QKI-RAF1 in NIH 3T3 cells. (b) MEK inhibitors partially suppress QKI-RAF1 driven soft agar growth and mouse flank xenografts. (c)Westerns showing effect of MEK inhibitors on growth pathways by SRGAP3-RAF1 in NIH 3T3 cells. (d) MEK inhibitors partially suppress SRGAP3-RAF1 driven soft agar growth and mouse flank xenografts.

Figure 3.10

Figure 3.10. Combinatorial targeting of MAPK and PI3K pathway (MEK inhibitors+ mTOR inhibitors) suppresses CRAF fusion driven tumor growth. (a) Westerns showing effect of MEK+ mTOR inhibitors on growth pathways by QKI-RAF1 and SRGAP3-RAF1 in NIH 3T3 cells. (b) Combination of MEK + mTOR inhibitors robustly suppress CRAF fusion driven soft agar growth in NIH 3T3 cells. (c) CRAF fusion driven mouse flank xenografts are suppressed by combinatorial targeting of MAPK and PI3K pathways.

CHAPTER 4: Discussion

Takes two to tango: QKI gene fusions define novel oncogenic mechanisms driven by both fusion partners

Rapid advances in deep sequencing technologies and fusion-detection platforms over the past few years have raised the total number of pan-cancer gene fusions to be around 10,000 (Latysheva &

Babu, 2016; Mertens, Johansson, Fioretos, & Mitelman, 2015). The prevalence of gene fusions varies widely across adult cancers with the overall frequency of gene fusion being lower than occurrence of somatic mutations (Lawrence et al., 2013). But for pediatric brain tumors, especially low-grade gliomas, recent studies suggest that genes fusions are the predominant pathognomonic events that mediate disease occurrence and progression. Historically, gene fusions have mostly been characterized to involve kinases and transcriptional control genes

(Mitelman, Johansson, & Mertens, 2004) that exert oncogenesis via constitutive activation of kinases/signaling pathways or association of a proto-oncogene with a strong promoter, respectively. In contrast to these mutually exclusive mechanisms, our studies shed light on novel, collaborative processes that gene fusions utilize to drive pediatric brain tumors. Our findings show the importance of RNA binding protein QKI as a functional gene fusion partner, highlighting an important, understudied role for the non-kinase/non-TF related fusion proteins.

Furthermore, we highlight how the combination of specific fusion partners in a gene fusion can affect not just the pathogenesis and tumor sub-type but also response to targeted therapy.

MYB-QKI: Synergism of a proto-oncogene and a tumor suppressor

MYB-QKI is an archetypal example of a gene fusion that is characteristic of a tumor sub-type and can be used for diagnostic and molecular subtyping purposes. Our in vitro and in vivo studies

100 show that MYB-QKI is a driver mutation in angiocentric gliomas and mediates gliomagenesis via multiple mechanisms (Fig. 4.1). Truncations in MYB result in constitutively active transcriptional activity of MYB-QKI, as shown in figure 2.5. Chromosomal rearrangement at 6q results in juxta- positioning of MYB and QKI, causing association of QKI-related enhancers with MYB promoters leading to MYB-QKI expression. Interestingly, MYB-QKI is capable of driving its own expression leading to a unique positive feedback loop. This finding represents a novel mechanism utilized by an oncogenic gene fusion and also provides a possible opportunity for epigenetic- targeted therapies that could possibly block expression of MYB-QKI. Due to lack of patient derived lines/models, we could not directly assess if there are any other mechanisms driving the

MYB promoter, especially since it is inactive in most brain regions.

In addition to activating the MYB transcriptional axis, MYB-QKI also involves the putative tumor suppressor gene QKI. Our findings suggest a strong additive effect due to QKI loss of heterozygosity and oncogenic MYB activation in neural stem cells. To our knowledge, this is the first demonstration that a proto-oncogene and a tumor suppressor collaborate in a gene fusion setting to enhance oncogenesis. From a therapeutic standpoint, these findings suggest that it might be insufficient to target only of the fusion partners to get complete efficacy. While no approved therapies exist for LOH mutations, epigenetic therapies are emerging as described below, and might be useful to target aberrant enhancer driven MYB-QKI expression.

Overall these findings have broad implications for distinguishing angiocentric gliomas accurately from other PLGG sub-types using deep sequencing, RT-PCR or FISH techniques. Other than diagnostic benefits, novel targeted therapies can also be designed against MYB-QKI expressing

AGs, either targeting epigenetic mechanisms mediating fusion expression or target downstream pathways driven by MYB-QKI. Furthermore, MYB-QKI can also be used to assess residual tumor post-surgery, given tumors express the fusion protein homogenously. Future aims of this 101 project would be to assess therapeutic agents against MYB-QKI, such as JQ1 that is an inhibitor of bromodomain-containing protein BRD4 (Filippakopoulos et al., 2010). This is because BRD4 remains bound to transcriptional start sites of genes (at acetyl-lysine residues) of many genes and is also a critical mediator of transcriptional elongation (Yang, 2005) and it has shown therapeutic efficacy in castration-resistant prostrate cancer (Asangani et al., 2014). BRD4 inhibition has also shown efficacy in suppressing another MYB rearrangement (MYB-NF1b) expressing tumor called adenoid cystic carcinoma (ACC) (Y. Drier et al., 2016). JQ1 could prevent MYB-QKI driven oncogenic transcriptional program by disrupting the MYB promoter-enhancer loop but this hypothesis needs to be tested.

QKI-RAF1: Dimerization and activation of kinase mediated by non-kinase partner

The CRAF fusion, QKI-RAF1, demonstrates another aspect of how QKI can contribute to gene fusion biology. Since QKI-RAF1 retains the N-terminal portion of QKI (as compared to C terminus in MYB-QKI), QKI homo-dimerization domain is retained along with portion of the

RBD. This suggested a different mechanism of action for QKI in QKI-RAF1 as opposed to QKI

LOH and associated loss of tumor suppressor function that we see in MYB-QKI. While loosing it’s RNA binding properties in QKI-RAF1 (Fig. 3.6g), QKI supports the oncogenic kinase activity of RAF1 via QKI-RAF1 dimerization. We have shown that QKI-mediated dimerization is key to maintaining stable QKI-RAF1 homo-dimers and subsequent transformation (Fig. 3.6).

Future work in this project will assess the cooperativity of QKI loss of functions and QKI-RAF1 oncogenicity, as we have shown in MYB-QKI tumors.

RAF1/CRAF has lower basal kinase activity compared to BRAF and requires additional phosphorylation/dephosphorylation steps for activation (Dhillon, Meikle, Yazici, Eulitz, & Kolch,

2002; J. Hu et al., 2013). RAF activation requires phosphorylation at 2 key positions: the C- terminal activation loop (AL) and the N-terminal acidic NtA) motif (Diaz et al., 1997; Mason et 102 al., 1999). BRAF has acidic residues in the NtA along with constitutively phosphorylated serines

(SSDD), leaving BRAF in a partially activated state (Mason et al., 1999). RAF1 lacks such phosphorylation and needs membrane recruitment and dimerization with BRAF to be activated (J.

Hu et al., 2013). RAF1 truncations and QKI mediated dimerization might be helping to bypass some of these activation steps. This hypothesis needs to be tested further.

Lastly, our findings highlight the evolving role of CRAF in cancer, in contrast to BRAF being considered the RAF protein of choice for oncogenic mutations. Even though BRAF is one of the most frequently mutated genes in cancer and is a major focus of targeted therapeutics, this study strongly suggests the need to delineate the rare CRAF mutations from BRAF mutations as there could be driving distinct molecular mechanisms to activate the MAPK pathway. While vemurafenib (PLX4720) doesn’t suppress either BRAF or CRAF fusions in PLGGs, PLX8394 can target only BRAF fusions but not CRAF fusions (Fig. 3.3). Since both BRAF and CRAF fusions occur in pilocytic astrocytomas, our findings point out a key therapeutic distinction with differential response to therapy within the same tumor sub-type. While combinatorial targeting with a MEK and mTOR inhibitor might be necessary/sufficient for targeting CRAF fusions, it is also essential to test upstream RAF targeting options due to evolving resistance mechanisms with

MAPK targeting (Waanders et al., manuscript under revision with Clinical Cancer Research).

Overall, QKI-RAF1 is the first RAF fusion in PLGGs where the N-terminal non-kinase partner has been shown to be involved in pathogenesis and therapeutic resistance making QKI-RAF1 a distinct molecular and therapeutic target (Figure 4.2).

All about that grade: Molecular stratification of PLGG patients

Current clinical practice utilizes the WHO classification as the standard of care for PLGGs diagnosis, treatment and management (Bergthold et al., 2014) but there is growing consensus that histopathological classification is not sufficient to discern between heterogeneous tumor subtypes 103

(Louis et al., 2014). Lack of accurate diagnosis might occur due to significant overlap in histologic and clinical features, scare tumor tissue availability, inter-physician variability and intra-tumoral heterogeneity (Bergthold et al., 2014). Furthermore, our pre-clinical study suggests that histological grade I tumor, PA, would respond differentially to RAF targeted therapy based on its mutational profile. Hence, a paradigm shift is impending to integrate traditional histopathological PLGG classification with emerging molecular and genomic data, as shown in figure 1.4.

Multiple large-scale sequencing studies have uncovered immense molecular complexity within

PLGGs. To address this, further work needs to be done to develop comprehensive mutational profiles that correlate with histological tumor grades. Furthermore, our study shows that even within the same tumor type, distinct molecular entities could exist, requiring even further tumor sub-typing. Lastly, application in the clinic would require augmenting current diagnostics with deep sequencing technologies so each patient’s tumor can be sequenced and driving mutations identified, thereby leading to better clinical outcomes and patient management.

Rolling in the deep-sequencing: Personalized medicine for PLGG patients

Despite good overall survival, long-term PLGG survivors suffer from significant morbidities, especially when complete surgical resection is difficult due to tumor location, such as in the optic pathway, hypothalamus, diencephalon, and brainstem. Therefore, development of novel, mutation-specific therapies is essential to deliver less toxic treatments to the developing nervous system of young PLGG patients.

Based on the findings from our study and others, it will become imperative to use deep sequencing techniques to first assess the mutational status of PLGG patients before assigning therapy. Future work on evaluating the molecular mechanisms of other PLLG-associated gene

104 fusions/mutations will add to the growing body of work for devising mutation-specific targeted therapies. Current clinical trials are assessing RAF inhibitors (Dabrafenib, NCT01677741), MEK inhibitors (MEK162, NCT02285439; trametinib, NCT02124772 and selumetinib, NCT01089101) and mTOR inhibitor (Everolimus, NCT01734512) against children with low-grade gliomas.

However, future clinical trials should incorporate genomic sequencing to identify patient-specific mutations to decide trial eligibility and have maximal efficacy.

PLGG treatment has evolved from radiation therapy once being the standard-of care to surgery and/or chemotherapy being the first line of therapy currently (Bergthold et al., 2014). Radiation therapy might still be used in non-surgical, chemotherapy refractory tumors but it can lead to severe long-term morbidities as discussed earlier. As such, there is increased awareness that non- specific therapeutic options should be avoided in treating these children. Our ongoing efforts to define the molecular mechanisms driving distinct PLGGs provide hope that there would soon be better, less toxic targeted therapies for these clinically challenging childhood tumors.

Molecular Revolution: Utilizing concepts from PLGG gene fusions to inform adult cancers

Adult LGGs have distinct molecular drivers (for example, IDH1/2 mutations, ATRX mutations and 1p/19q co-deletions) and share negligible overlap with the mutation profile of pediatric

LGGs. In addition to vast differences in the molecular pathogenesis, adult and pediatric brain tumors also differ in terms of predilection or tumor site in the brain, cellular microenvironment and cell of origin (Baker, Ellison, & Gutmann, 2016). However, the genes implicated in PLGGs are also mutated in several adult cancers, such as the MYB, BRAF, CRAF and QKI genes that were under investigation in this study undergo various alterations in adult cancers. Our

105 mechanistic findings have the potential to inform tumor mechanisms in adult cancers that involve similar genetic mutations and have shared biology.

MYB is involved in frequent gene rearrangements, point mutations and deletion events in adult cancers. Adenoid cystic carcinoma (ACC), a malignant cancer arising in salivary glands, is associated with a hallmark MYB-NF1B fusion event (Ho et al., 2013). Interestingly, a recent study has shown that similar to MYB-QKI, the gene rearrangement of MYB-NF1B also results in translocation of enhancers close to the MYB promoter, resulting in activation of MYB promoter

(Y. Drier et al., 2016). Furthermore, MYB-NF1B was also found to bind and activate the MYB promoter. These findings suggest that some of our mechanistic findings in MYB-QKI could be universally applicable to other adult/pediatric MYB gene rearrangements and other gene fusions that involve a proto-oncogene and tumor suppressor gene.

Our previous studies on therapeutic targeting of BRAF gene fusions have shown similar resistance mechanisms to therapy in BRAF mutant melanomas (A. J. Sievert et al., 2013).

Paradoxical activation of the KIAA1549-BRAF fusion occurred due to vemurafenib therapy, and it was later found that BRAFV600E mutant melanomas also exhibit paradoxical activation of pathway as a resistance mechanism post initial response. These findings highlight that the basic biological mechanisms retain some homogeneity across cancers of different age, cell of origin, location, malignancy and prognosis. Our study on CRAF fusions also holds promise to inform therapy for CRAF fusions identified in adult cancers such as prostrate cancer (ESRP1-RAF1,

AGGF1-RAF1), breast cancer (RAF1-DAZL, CMTM8-RAF1), thyroid cancer (AGGF1-RAF1) and pancreatic cancer (HACL1-RAF1) (Fig. 3.1b). While deep sequencing studies have highlighted the prevalence of CRAF fusions in several adult cancers, therapeutic targeting for majority of these fusions remains unknown. Furthermore, a recent study found a rare CRAF point mutation

(R391W) as a driver mutation in melanoma where dimerization-dependent oncogenic 106 mechanisms are also at play (Atefi et al., 2016). Therefore, our mechanistic findings on how

PLGG- associated CRAF fusions display distinct dimerization properties in response to inhibitors could be extended to/tested for adult CRAF fusions and mutations, thereby guiding universal personalized medicine approaches for all oncogenic CRAF fusions.

Conclusion

Overall, our study describes the unique mechanism of action of a new class of gene fusions in pediatric low grade gliomas. QKI gene fusions, MYB-QKI and QKI-RAF1 are found to function via involvement of both fusion partners thereby opening the field of fusion biology to new possibilities. First, we found MYB-QKI to be a single, specific driver event in angiocentric gliomas, which have not been associated with a molecular marker thus far. Second, we uncovered the oncogenic mechanism of action of MYB-QKI to be a collaboration of gain-of-function of MYB oncogene and partial loss of function of QKI tumor suppressor, along with involvement of epigenetic events to drive aberrant MYB-QKI expression. Third, our study suggests that QKI-

RAF1 can activate the MAPK and PI3K signaling pathways but they are dependent on QKI- mediated dimerization, a finding that distinguishes them from other RAF fusions found in

PLGGs. Lastly, we report differential response to RAF targeted inhibitors in QKI-RAF1 expressing model systems due to stable dimerization induced by QKI fusion partner. Therefore, we have shown how QKI gene fusions function to drive gliomagenesis in PLGGs, with broader implications on biological mechanisms adopted by pan-cancer gene fusions and personalized medicine approaches for PLGGs.

107

Tables and Figures

Figure 4.1

Chr 6q23-27 MYB-QKI gene rearrangement MYB QKI Exon 1 9 15 17 ~30Mb 1 4 8 H3K27Ac

QKI 3’UTR enhancer translocation

3’UTR MYB promoter activation enhancer MYB-QKI

Auto-regulatory Expression of oncogenic feed back loop MYB-QKI fusion protein

Oncogenic MYB QKI tumor MYB-QKI pathway active suppressor loss

Angiocentric Gliomas

Figure 4.1. MYB-QKI drives oncogenesis in angiocentric gliomas via a tripartite mechanism. MYB and QKI are located on chromosome 6q and intra-chromosomal deletion leads to juxta-positioning of 5’ end of MYB with 3’ end of QKI, causing translocation of 3’UTR enhancers. QKI-related enhancers can drive MYB promoter activation, leading to MYB-QKI expression. MYB-QKI is oncogenic via additive effects of MYB gain of function and LOH of QKI.

108

Figure 4.2

Extracellular Fluid

RAF inhibitors QKI RAF1 QKI RAF1 or QKI RAF1 (PLX8394) BRAF RAF1 or QKI RAF1 LY3009120 (RAF RAF1 QKI PI3K dimer inhibitor) QKI AK MEK T GSK1120212/Trametinib RAD001 (mTOR (MEK inhibitor) mTOR inhibitor)

Cytoplasm ERK S6

Nucleus

Regulate cell proliferation/growth

Figure 4.2. MAPK and PI3K oncogenic pathways driven by QKI-RAF1 and response to targeted therapies. QKI-RAF1 exists as homodimers and heterodimers (with BRAF, CRAF and QKI) that can activate the MAPK/ERK and PI3K/mTOR signaling pathways. While RAF inhibitors (PLX4720 and PLX8394) cannot inhibit CRAF fusions, the novel RAF dimer inhibitor LY3009120 could be a possible therapeutic option. Combinatorial targeting of downstream MAPK and PI3K/mTOR pathways with MEK inhibitor trametinib and mTOR inhibitor everolimus, respectively, has shown significant in vitro and in vivo efficacy.

109

REFERENCES

Abedalthagafi, M. S., Bi, W. L., Merrill, P. H., Gibson, W. J., Rose, M. F., Du, Z., . . . Santagata, S. (2015). ARID1A and TERT promoter mutations in dedifferentiated meningioma. Cancer genetics, 208(6), 345-350. doi:10.1016/j.cancergen.2015.03.005 Ali, M., & Broadhurst, R. W. (2013). Solution structure of the QUA1 dimerization domain of pXqua, the Xenopus ortholog of Quaking. PLoS One, 8(3), e57345. doi:10.1371/journal.pone.0057345 Asangani, I. A., Dommeti, V. L., Wang, X., Malik, R., Cieslik, M., Yang, R., . . . Chinnaiyan, A. M. (2014). Therapeutic targeting of BET bromodomain proteins in castration-resistant prostate cancer. Nature, 510(7504), 278-282. doi:10.1038/nature13229 Atefi, M., Titz, B., Tsoi, J., Avramis, E., Le, A., Ng, C., . . . Graeber, T. G. (2016). CRAF R391W is a melanoma driver oncogene. Sci Rep, 6, 27454. doi:10.1038/srep27454 Ater, J. L., Zhou, T., Holmes, E., Mazewski, C. M., Booth, T. N., Freyer, D. R., . . . Pollack, I. F. (2012). Randomized study of two chemotherapy regimens for treatment of low-grade glioma in young children: a report from the Children's Oncology Group. J Clin Oncol, 30(21), 2641-2647. doi:10.1200/jco.2011.36.6054 Bachoo, R. M., Maher, E. A., Ligon, K. L., Sharpless, N. E., Chan, S. S., You, M. J., . . . Depinho, R. A. (2002). Epidermal growth factor receptor and Ink4a/Arf: convergent mechanisms governing terminal differentiation and transformation along the neural stem cell to astrocyte axis. Cancer Cell, 1(3), 269-277. Bacon, C., Endris, V., & Rappold, G. (2009). Dynamic expression of the Slit-Robo GTPase activating protein genes during development of the murine nervous system. J Comp Neurol, 513(2), 224-236. doi:10.1002/cne.21955 Baker, S. J., Ellison, D. W., & Gutmann, D. H. (2016). Pediatric gliomas as neurodevelopmental disorders. Glia, 64(6), 879-895. doi:10.1002/glia.22945 Bandopadhayay, P., Bergthold, G., London, W. B., Goumnerova, L. C., Morales La Madrid, A., Marcus, K. J., . . . Manley, P. E. (2014). Long-term outcome of 4,040 children diagnosed with pediatric low-grade gliomas: an analysis of the Surveillance Epidemiology and End Results (SEER) database. Pediatr Blood Cancer, 61(7), 1173-1179. doi:10.1002/pbc.24958 Bandopadhayay, P., Bergthold, G., Nguyen, B., Schubert, S., Gholamin, S., Tang, Y., . . . Cho, Y.-J. (2014). BET bromodomain inhibition of MYC-amplified medulloblastoma. Clin Cancer Res, 20(4), 912-925. doi:10.1158/1078-0432.CCR-13-2281 Bandopadhayay, P., Ramkissoon, L. A., Jain, P., Bergthold, G., Wala, J., Zeid, R., . . . Resnick, A. C. (2016). MYB-QKI rearrangements in angiocentric glioma drive tumorigenicity through a tripartite mechanism. Nat Genet, 48(3), 273-282. doi:10.1038/ng.3500 Bax, D. A., Little, S. E., Gaspar, N., Perryman, L., Marshall, L., Viana-Pereira, M., . . . Jones, C. (2009). Molecular and phenotypic characterisation of paediatric glioma cell lines as models for preclinical drug development. PLoS One, 4(4), e5209. doi:10.1371/journal.pone.0005209 Bengtsen, M., Klepper, K., Gundersen, S., Cuervo, I., Drabløs, F., Hovig, E., . . . Eskeland, R. (2015). c-Myb Binding Sites in Haematopoietic Chromatin Landscapes. PLoS One, 10(7), e0133280. doi:10.1371/journal.pone.0133280

110

Bergthold, G., Bandopadhayay, P., Bi, W. L., Ramkissoon, L., Stiles, C., Segal, R. A., . . . Kieran, M. W. (2014). Pediatric low-grade gliomas: how modern biology reshapes the clinical field. Biochimica et biophysica acta, 1845(2), 294-307. doi:10.1016/j.bbcan.2014.02.004 Bergthold, G., Bandopadhayay, P., Hoshida, Y., Ramkissoon, S., Ramkissoon, L., Rich, B., . . . Beroukhim, R. (2015). Expression profiles of 151 pediatric low-grade gliomas reveal molecular differences associated with location and histological subtype. Neuro- oncology. doi:10.1093/neuonc/nov045 Beroukhim, R., Getz, G., Nghiemphu, L., Barretina, J., Hsueh, T., Linhart, D., . . . Sellers, W. R. (2007). Assessing the significance of chromosomal aberrations in cancer: methodology and application to glioma. PNAS, 104(50), 20007-20012. doi:10.1073/pnas.0710052104 Beroukhim, R., Mermel, C. H., Porter, D., Wei, G., Raychaudhuri, S., Donovan, J., . . . Meyerson, M. (2010). The landscape of somatic copy-number alteration across human cancers. Nature, 463(7283), 899-905. doi:10.1038/nature08822 Beuck, C., Qu, S., Fagg, W. S., Ares, M., Jr., & Williamson, J. R. (2012). Structural analysis of the quaking homodimerization interface. J Mol Biol, 423(5), 766-781. doi:10.1016/j.jmb.2012.08.027 Bian, Y., Wang, L., Lu, H., Yang, G., Zhang, Z., Fu, H., . . . Lu, Z. (2012). Downregulation of tumor suppressor QKI in gastric cancer and its implication in cancer prognosis. Biochem Biophys Res Commun, 422(1), 187-193. doi:10.1016/j.bbrc.2012.04.138 Bollag, G., Hirth, P., Tsai, J., Zhang, J., Ibrahim, P. N., Cho, H., . . . Nolop, K. (2010). Clinical efficacy of a RAF inhibitor needs broad target blockade in BRAF-mutant melanoma. Nature, 467(7315), 596-599. doi:10.1038/nature09454 Bolstad, B. M., Irizarry, R. A., Astrand, M., & Speed, T. P. (2003). A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics (Oxford, England), 19(2), 185-193. Bouffet, E., Jakacki, R., Goldman, S., Hargrave, D., Hawkins, C., Shroff, M., . . . Baruchel, S. (2012). Phase II study of weekly vinblastine in recurrent or refractory pediatric low-grade glioma. J Clin Oncol, 30(12), 1358-1363. doi:10.1200/jco.2011.34.5843 Buccoliero, A. M., Castiglione, F., Degl'innocenti, D. R., Moncini, D., Spacca, B., Giordano, F., . . . Taddei, G. L. (2013). Angiocentric glioma: clinical, morphological, immunohistochemical and molecular features in three pediatric cases. Clin Neuropathol, 32(2), 107-113. doi:10.5414/np300500 Carlson, B. R., Lloyd, K. E., Kruszewski, A., Kim, I. H., Rodriguiz, R. M., Heindel, C., . . . Soderling, S. H. (2011). WRP/srGAP3 facilitates the initiation of spine development by an inverse F- BAR domain, and its loss impairs long-term memory. J Neurosci, 31(7), 2447-2460. doi:10.1523/jneurosci.4433-10.2011 Catalanotti, F., Solit, D. B., Pulitzer, M. P., Berger, M. F., Scott, S. N., Iyriboz, T., . . . Chapman, P. B. (2013). Phase II trial of MEK inhibitor selumetinib (AZD6244, ARRY-142886) in patients with BRAFV600E/K-mutated melanoma. Clin Cancer Res, 19(8), 2257-2264. doi:10.1158/1078-0432.ccr-12-3476 Chapman, M. A., Lawrence, M. S., Keats, J. J., Cibulskis, K., Sougnez, C., Schinzel, A. C., . . . Golub, T. R. (2011). Initial genome sequencing and analysis of multiple myeloma. Nature, 471(7339), 467-472. doi:10.1038/nature09837 Chapman, P. B., Hauschild, A., Robert, C., Haanen, J. B., Ascierto, P., Larkin, J., . . . McArthur, G. A. (2011). Improved survival with vemurafenib in melanoma with BRAF V600E mutation. N Engl J Med, 364(26), 2507-2516. doi:10.1056/NEJMoa1103782 111

Chapuy, B., McKeown, M. R., Lin, C. Y., Monti, S., Roemer, M. G. M., Qi, J., . . . Bradner, J. E. (2013). Discovery and characterization of super-enhancer-associated dependencies in diffuse large B cell lymphoma. Cancer Cell, 24(6), 777-790. doi:10.1016/j.ccr.2013.11.003 Chen, A. J., Paik, J. H., Zhang, H., Shukla, S. A., Mortensen, R., Hu, J., . . . Depinho, R. A. (2012). STAR RNA-binding protein Quaking suppresses cancer via stabilization of specific miRNA. Genes Dev, 26(13), 1459-1472. doi:10.1101/gad.189001.112 Chen, S. H., Zhang, Y., Van Horn, R. D., Yin, T., Buchanan, S., Yadav, V., . . . Peng, S. B. (2016). Oncogenic BRAF Deletions That Function as Homodimers and Are Sensitive to Inhibition by RAF Dimer Inhibitor LY3009120. Cancer Discov, 6(3), 300-315. doi:10.1158/2159- 8290.cd-15-0896 Chen, T., & Richard, S. (1998). Structure-function analysis of Qk1: a lethal point mutation in mouse quaking prevents homodimerization. Mol Cell Biol, 18(8), 4863-4871. Chen, T., & Richard, S. (1998). Structure-function analysis of Qk1: a lethal point mutation in mouse quaking prevents homodimerization. Mol Cell Biol, 18(8), 4863-4871. Chiang, D. Y., Getz, G., Jaffe, D. B., O'Kelly, M. J. T., Zhao, X., Carter, S. L., . . . Lander, E. S. (2009). High-resolution mapping of copy-number alterations with massively parallel sequencing. Nat Methods, 6(1), 99-103. doi:10.1038/nmeth.1276 Chipumuro, E., Marco, E., Christensen, C. L., Kwiatkowski, N., Zhang, T., Hatheway, C. M., . . . George, R. E. (2014). CDK7 inhibition suppresses super-enhancer-linked oncogenic transcription in MYCN-driven cancer. Cell, 159(5), 1126-1139. doi:10.1016/j.cell.2014.10.024 Cibulskis, K., Lawrence, M. S., Carter, S. L., Sivachenko, A., Jaffe, D., Sougnez, C., . . . Getz, G. (2013). Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nature Biotechnology, 31(3), 213-219. doi:10.1038/nbt.2514 Conn, S. J., Pillman, K. A., Toubia, J., Conn, V. M., Salmanidis, M., Phillips, C. A., . . . Goodall, G. J. (2015). The RNA binding protein quaking regulates formation of circRNAs. Cell, 160(6), 1125-1134. doi:10.1016/j.cell.2015.02.014 Consortium, E. P. (2012). An integrated encyclopedia of DNA elements in the human genome. Nature, 489(7414), 57-74. doi:10.1038/nature11247 Consortium, G. (2013). The Genotype-Tissue Expression (GTEx) project. Nat Genet, 45(6), 580- 585. doi:10.1038/ng.2653 Craig, J. M., Vena, N., Ramkissoon, S., Idbaih, A., Fouse, S. D., Ozek, M., . . . Ligon, A. H. (2012). DNA fragmentation simulation method (FSM) and fragment size matching improve aCGH performance of FFPE tissues. PLoS One, 7(6), e38881. doi:10.1371/journal.pone.0038881 Crompton, B. D., Stewart, C., Taylor-Weiner, A., Alexe, G., Kurek, K. C., Calicchio, M. L., . . . Stegmaier, K. (2014). The genomic landscape of pediatric Ewing sarcoma. Cancer Discov, 4(11), 1326-1341. doi:10.1158/2159-8290.CD-13-1037 Cutler, R. E., Jr., Stephens, R. M., Saracino, M. R., & Morrison, D. K. (1998). Autoregulation of the Raf-1 serine/threonine kinase. Proc Natl Acad Sci U S A, 95(16), 9214-9219. D'Alfonso, T. M., Mosquera, J. M., MacDonald, T. Y., Padilla, J., Liu, Y.-F., Rubin, M. A., & Shin, S. J. (2014). MYB-NFIB gene fusion in adenoid cystic carcinoma of the breast with special focus paid to the solid variant with basaloid features. Human pathology, 45(11), 2270-2280. doi:10.1016/j.humpath.2014.07.013

112

Danan-Gotthold, M., Golan-Gerstl, R., Eisenberg, E., Meir, K., Karni, R., & Levanon, E. Y. (2015). Identification of recurrent regulated alternative splicing events across human solid tumors. Nucleic Acids Research. doi:10.1093/nar/gkv210 Dash, A. B., Orrico, F. C., & Ness, S. A. (1996). The EVES motif mediates both intermolecular and intramolecular regulation of c-Myb. Genes Dev, 10(15), 1858-1869. Delmore, J. E., Issa, G. C., Lemieux, M. E., Rahl, P. B., Shi, J., Jacobs, H. M., . . . Mitsiades, C. S. (2011). BET bromodomain inhibition as a therapeutic strategy to target c-Myc. Cell, 146(6), 904-917. doi:10.1016/j.cell.2011.08.017 Dhillon, A. S., Meikle, S., Yazici, Z., Eulitz, M., & Kolch, W. (2002). Regulation of Raf-1 activation and signalling by dephosphorylation. EMBO J, 21(1-2), 64-71. Diaz, B., Barnard, D., Filson, A., MacDonald, S., King, A., & Marshall, M. (1997). Phosphorylation of Raf-1 serine 338-serine 339 is an essential regulatory event for Ras-dependent activation and biological signaling. Mol Cell Biol, 17(8), 4509-4516. Drier, Y., Cotton, M. J., Williamson, K. E., Gillespie, S. M., Ryan, R. J., Kluk, M. J., . . . Bernstein, B. E. (2016). An oncogenic MYB feedback loop drives alternate cell fates in adenoid cystic carcinoma. Nat Genet, 48(3), 265-272. doi:10.1038/ng.3502 Drier, Y., Lawrence, M. S., Carter, S. L., Stewart, C., Gabriel, S. B., Lander, E. S., . . . Getz, G. (2013). Somatic rearrangements across cancer reveal classes of samples with distinct patterns of DNA breakage and rearrangement-induced hypermutability. Genome Research, 23(2), 228-235. doi:10.1101/gr.141382.112 Emerson, S. D., Madison, V. S., Palermo, R. E., Waugh, D. S., Scheffler, J. E., Tsao, K. L., . . . Fry, D. C. (1995). Solution structure of the Ras-binding domain of c-Raf-1 and identification of its Ras interaction surface. Biochemistry, 34(21), 6911-6918. Etchegaray, J.-P., Chavez, L., Huang, Y., Ross, K. N., Choi, J., Martinez-Pastor, B., . . . Mostoslavsky, R. (2015). The histone deacetylase SIRT6 controls embryonic stem cell fate via TET-mediated production of 5-hydroxymethylcytosine. Nature cell biology, 17(5), 545-557. doi:10.1038/ncb3147 Fehr, A., Kovacs, A., Loning, T., Frierson, H., Jr., van den Oord, J., & Stenman, G. (2011). The MYB-NFIB gene fusion-a novel genetic link between adenoid cystic carcinoma and dermal cylindroma. J Pathol, 224(3), 322-327. doi:10.1002/path.2909 Filippakopoulos, P., Qi, J., Picaud, S., Shen, Y., Smith, W. B., Fedorov, O., . . . Bradner, J. E. (2010). Selective inhibition of BET bromodomains. Nature, 468(7327), 1067-1073. doi:10.1038/nature09504 Firestein, R., Bass, A. J., Kim, S. Y., Dunn, I. F., Silver, S. J., Guney, I., . . . Hahn, W. C. (2008). CDK8 is a colorectal cancer oncogene that regulates beta-catenin activity. Nature, 455(7212), 547-551. doi:10.1038/nature07179 Flaherty, K. T., Puzanov, I., Kim, K. B., Ribas, A., McArthur, G. A., Sosman, J. A., . . . Chapman, P. B. (2010). Inhibition of mutated, activated BRAF in metastatic melanoma. N Engl J Med, 363(9), 809-819. doi:10.1056/NEJMoa1002011 Flaherty, K. T., Robert, C., Hersey, P., Nathan, P., Garbe, C., Milhem, M., . . . Schadendorf, D. (2012). Improved survival with MEK inhibition in BRAF-mutated melanoma. N Engl J Med, 367(2), 107-114. doi:10.1056/NEJMoa1203421 Freeman, A. K., Ritt, D. A., & Morrison, D. K. (2013a). Effects of Raf dimerization and its inhibition on normal and disease-associated Raf signaling. Mol Cell, 49(4), 751-758. doi:10.1016/j.molcel.2012.12.018

113

Freeman, A. K., Ritt, D. A., & Morrison, D. K. (2013b). The importance of Raf dimerization in cell signaling. Small GTPases, 4(3), 180-185. doi:10.4161/sgtp.26117 Friedrich, V. L., Jr. (1974). The myelin deficit in quacking mice. Brain Res, 82(1), 168-172. Galarneau, A., & Richard, S. (2005). Target RNA motif and target mRNAs of the Quaking STAR protein. Nat Struct Mol Biol, 12(8), 691-698. doi:10.1038/nsmb963 Gao, R., Cao, C., Zhang, M., Lopez, M.-C., Yan, Y., Chen, Z., . . . Kaye, F. J. (2014). A unifying gene signature for adenoid cystic cancer identifies parallel MYB-dependent and MYB- independent therapeutic targets. Oncotarget, 5(24), 12528-12542. Garber, M., Yosef, N., Goren, A., Raychowdhury, R., Thielke, A., Guttman, M., . . . Amit, I. (2012). A high-throughput chromatin immunoprecipitation approach reveals principles of dynamic gene regulation in mammals. Mol Cell, 47(5), 810-822. doi:10.1016/j.molcel.2012.07.030 Gilmartin, A. G., Bleam, M. R., Groy, A., Moss, K. G., Minthorn, E. A., Kulkarni, S. G., . . . Laquerre, S. G. (2011). GSK1120212 (JTP-74057) is an inhibitor of MEK activity and activation with favorable pharmacokinetic properties for sustained in vivo pathway inhibition. Clin Cancer Res, 17(5), 989-1000. doi:10.1158/1078-0432.ccr-10-2200 Gittleman, H. R., Ostrom, Q. T., Rouse, C. D., Dowling, J. A., de Blank, P. M., Kruchko, C. A., . . . Barnholtz-Sloan, J. S. (2015). Trends in central nervous system tumor incidence relative to other common cancers in adults, adolescents, and children in the United States, 2000 to 2010. Cancer, 121(1), 102-112. doi:10.1002/cncr.29015 Gonda, T. J., & Bishop, J. M. (1983). Structure and transcription of the cellular homolog (c-myb) of the avian myeloblastosis virus transforming gene (v-myb). Journal of virology, 46(1), 212-220. Gonda, T. J., Buckmaster, C., & Ramsay, R. G. (1989). Activation of c-myb by carboxy-terminal truncation: relationship to transformation of murine haemopoietic cells in vitro. EMBO J, 8(6), 1777-1783. Gould, J., Getz, G., Monti, S., Reich, M., & Mesirov, J. P. (2006). Comparative gene marker selection suite. Bioinformatics (Oxford, England), 22(15), 1924-1925. doi:10.1093/bioinformatics/btl196 Grässer, F. A., Graf, T., & Lipsick, J. S. (1991). Protein truncation is required for the activation of the c-myb proto-oncogene. Mol Cell Biol, 11(8), 3987-3996. Grasso, C. S., Tang, Y., Truffaux, N., Berlow, N. E., Liu, L., Debily, M.-A., . . . Monje, M. (2015). Functionally defined therapeutic targets in diffuse intrinsic pontine glioma. Nat Med, 21(6), 555-559. doi:10.1038/nm.3855 Gutmann, D. H., Donahoe, J., Brown, T., James, C. D., & Perry, A. (2000). Loss of neurofibromatosis 1 (NF1) gene expression in NF1-associated pilocytic astrocytomas. Neuropathol Appl Neurobiol, 26(4), 361-367. Hashizume, R., Andor, N., Ihara, Y., Lerner, R., Gan, H., Chen, X., . . . James, C. D. (2014). Pharmacologic inhibition of histone demethylation as a therapy for pediatric brainstem glioma. Nat Med, 20(12), 1394-1396. doi:10.1038/nm.3716 Hatzivassiliou, G., Song, K., Yen, I., Brandhuber, B. J., Anderson, D. J., Alvarado, R., . . . Malek, S. (2010). RAF inhibitors prime wild-type RAF to activate the MAPK pathway and enhance growth. Nature, 464(7287), 431-435. doi:10.1038/nature08833 Hawkins, C., Walker, E., Mohamed, N., Zhang, C., Jacob, K., Shirinian, M., . . . Tabori, U. (2011). BRAF-KIAA1549 fusion predicts better clinical outcome in pediatric low-grade astrocytoma. Clin Cancer Res, 17(14), 4790-4798. doi:10.1158/1078-0432.ccr-11-0034 114

Heidecker, G., Huleihel, M., Cleveland, J. L., Kolch, W., Beck, T. W., Lloyd, P., . . . Rapp, U. R. (1990). Mutational activation of c-raf-1 and definition of the minimal transforming sequence. Mol Cell Biol, 10(6), 2503-2512. Heidorn, S. J., Milagre, C., Whittaker, S., Nourry, A., Niculescu-Duvas, I., Dhomen, N., . . . Marais, R. (2010). Kinase-dead BRAF and oncogenic RAS cooperate to drive tumor progression through CRAF. Cell, 140(2), 209-221. doi:10.1016/j.cell.2009.12.040 Heinz, S., Benner, C., Spann, N., Bertolino, E., Lin, Y. C., Laslo, P., . . . Glass, C. K. (2010). Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell, 38(4), 576-589. doi:10.1016/j.molcel.2010.05.004 Henne, W. M., Kent, H. M., Ford, M. G., Hegde, B. G., Daumke, O., Butler, P. J., . . . McMahon, H. T. (2007). Structure and analysis of FCHo2 F-BAR domain: a dimerizing and membrane recruitment module that effects membrane curvature. Structure, 15(7), 839-852. doi:10.1016/j.str.2007.05.002 Henry, J. R., Kaufman, M. D., Peng, S. B., Ahn, Y. M., Caldwell, T. M., Vogeti, L., . . . Flynn, D. L. (2015). Discovery of 1-(3,3-dimethylbutyl)-3-(2-fluoro-4-methyl-5-(7-methyl-2- (methylamino)pyrido[2,3- d]pyrimidin-6-yl)phenyl)urea (LY3009120) as a pan-RAF inhibitor with minimal paradoxical activation and activity against BRAF or RAS mutant tumor cells. J Med Chem, 58(10), 4165-4179. doi:10.1021/acs.jmedchem.5b00067 Ho, A. S., Kannan, K., Roy, D. M., Morris, L. G., Ganly, I., Katabi, N., . . . Chan, T. A. (2013). The mutational landscape of adenoid cystic carcinoma. Nat Genet, 45(7), 791-798. doi:10.1038/ng.2643 Hu, J., Stites, E. C., Yu, H., Germino, E. A., Meharena, H. S., Stork, P. J., . . . Shaw, A. S. (2013). Allosteric activation of functionally asymmetric RAF kinase dimers. Cell, 154(5), 1036- 1046. doi:10.1016/j.cell.2013.07.046 Hu, Y. L., Ramsay, R. G., Kanei-Ishii, C., Ishii, S., & Gonda, T. J. (1991). Transformation by carboxyl-deleted Myb reflects increased transactivating capacity and disruption of a negative regulatory domain. Oncogene, 6(9), 1549-1553. Jacob, K., Albrecht, S., Sollier, C., Faury, D., Sader, E., Montpetit, A., . . . Jabado, N. (2009). Duplication of 7q34 is specific to juvenile pilocytic astrocytomas and a hallmark of cerebellar and optic pathway tumours. Br J Cancer, 101(4), 722-733. doi:10.1038/sj.bjc.6605179 Jacob, K., Quang-Khuong, D. A., Jones, D. T., Witt, H., Lambert, S., Albrecht, S., . . . Jabado, N. (2011). Genetic aberrations leading to MAPK pathway activation mediate oncogene- induced senescence in sporadic pilocytic astrocytomas. Clin Cancer Res, 17(14), 4650- 4660. doi:10.1158/1078-0432.ccr-11-0127 Jeyapalan, J. N., Doctor, G. T., Jones, T. A., Alberman, S. N., Tep, A., Haria, C. M., . . . Sheer, D. (2016). DNA methylation analysis of paediatric low-grade astrocytomas identifies a tumour-specific hypomethylation signature in pilocytic astrocytomas. Acta Neuropathol Commun, 4(1), 54. doi:10.1186/s40478-016-0323-6 Ji, S., Ye, G., Zhang, J., Wang, L., Wang, T., Wang, Z., . . . Yang, J. Y. (2013). miR-574-5p negatively regulates Qki6/7 to impact β-catenin/Wnt signalling and the development of colorectal cancer. Gut, 62(5), 716-726. doi:10.1136/gutjnl-2011-301083 Jones, D. T., Ichimura, K., Liu, L., Pearson, D. M., Plant, K., & Collins, V. P. (2006). Genomic analysis of pilocytic astrocytomas at 0.97 Mb resolution shows an increasing tendency

115

toward chromosomal copy number change with age. J Neuropathol Exp Neurol, 65(11), 1049-1058. doi:10.1097/01.jnen.0000240465.33628.87 Jones, D. T., Kocialkowski, S., Liu, L., Pearson, D. M., Backlund, L. M., Ichimura, K., & Collins, V. P. (2008). Tandem duplication producing a novel oncogenic BRAF fusion gene defines the majority of pilocytic astrocytomas. Cancer Res, 68(21), 8673-8677. doi:10.1158/0008- 5472.can-08-2097 Jones, D. T., Kocialkowski, S., Liu, L., Pearson, D. M., Ichimura, K., & Collins, V. P. (2009). Oncogenic RAF1 rearrangement and a novel BRAF mutation as alternatives to KIAA1549:BRAF fusion in activating the MAPK pathway in pilocytic astrocytoma. Oncogene, 28(20), 2119-2123. doi:10.1038/onc.2009.73 Jones, D. T. W., Hutter, B., Jäger, N., Korshunov, A., Kool, M., Warnatz, H.-J., . . . Project, I. C. G. C. P. T. (2013). Recurrent somatic alterations of FGFR1 and NTRK2 in pilocytic astrocytoma. Nat Genet, 45(8), 927-932. doi:10.1038/ng.2682 Karajannis, M. A., Legault, G., Fisher, M. J., Milla, S. S., Cohen, K. J., Wisoff, J. H., . . . Allen, J. C. (2014). Phase II study of sorafenib in children with recurrent or progressive low-grade astrocytomas. Neuro Oncol, 16(10), 1408-1416. doi:10.1093/neuonc/nou059 Kaul, A., Chen, Y. H., Emnett, R. J., Dahiya, S., & Gutmann, D. H. (2012). Pediatric glioma- associated KIAA1549:BRAF expression regulates neuroglial cell growth in a cell type- specific and mTOR-dependent manner. Genes Dev, 26(23), 2561-2566. doi:10.1101/gad.200907.112 Khuong-Quang, D. A., Buczkowicz, P., Rakopoulos, P., Liu, X. Y., Fontebasso, A. M., Bouffet, E., . . . Hawkins, C. (2012). K27M mutation in histone H3.3 defines clinically and biologically distinct subgroups of pediatric diffuse intrinsic pontine gliomas. Acta Neuropathol, 124(3), 439-447. doi:10.1007/s00401-012-0998-0 Kieran, M. W., Roberts, C. W. M., Chi, S. N., Ligon, K. L., Rich, B. E., MacConaill, L. E., . . . Biegel, J. A. (2012). Absence of oncogenic canonical pathway mutations in aggressive pediatric rhabdoid tumors. Pediatric blood & cancer, 59(7), 1155-1157. doi:10.1002/pbc.24315 Kleihues, P., Louis, D. N., Scheithauer, B. W., Rorke, L. B., Reifenberger, G., Burger, P. C., & Cavenee, W. K. (2002). The WHO classification of tumors of the nervous system. J Neuropathol Exp Neurol, 61(3), 215-225; discussion 226-219. Klempnauer, K. H., Bonifer, C., & Sippel, A. E. (1986). Identification and characterization of the protein encoded by the human c-myb proto-oncogene. EMBO J, 5(8), 1903-1911. Klempnauer, K. H., Gonda, T. J., & Bishop, J. M. (1982). Nucleotide sequence of the retroviral leukemia gene v-myb and its cellular progenitor c-myb: the architecture of a transduced oncogene. Cell, 31(2 Pt 1), 453-463. Kluwe, L., Hagel, C., Tatagiba, M., Thomas, S., Stavrou, D., Ostertag, H., . . . Mautner, V. F. (2001). Loss of NF1 alleles distinguish sporadic from NF1-associated pilocytic astrocytomas. J Neuropathol Exp Neurol, 60(9), 917-920. Lahoz, A., & Hall, A. (2013). A tumor suppressor role for srGAP3 in mammary epithelial cells. Oncogene, 32(40), 4854-4860. doi:10.1038/onc.2012.489 Latysheva, N. S., & Babu, M. M. (2016). Discovering and understanding oncogenic gene fusions through data intensive computational approaches. Nucleic Acids Res, 44(10), 4487-4503. doi:10.1093/nar/gkw282 Lau, N., Feldkamp, M. M., Roncari, L., Loehr, A. H., Shannon, P., Gutmann, D. H., & Guha, A. (2000). Loss of neurofibromin is associated with activation of RAS/MAPK and PI3-K/AKT 116

signaling in a neurofibromatosis 1 astrocytoma. J Neuropathol Exp Neurol, 59(9), 759- 767. Lawrence, M. S., Stojanov, P., Mermel, C. H., Robinson, J. T., Garraway, L. A., Golub, T. R., . . . Getz, G. (2014). Discovery and saturation analysis of cancer genes across 21 tumour types. Nature, 505(7484), 495-501. doi:10.1038/nature12912 Lawrence, M. S., Stojanov, P., Polak, P., Kryukov, G. V., Cibulskis, K., Sivachenko, A., . . . Getz, G. (2013). Mutational heterogeneity in cancer and the search for new cancer-associated genes. Nature, 499(7457), 214-218. doi:10.1038/nature12213 Lellouch-Tubiana, A., Boddaert, N., Bourgeois, M., Fohlen, M., Jouvet, A., Delalande, O., . . . Sainte-Rose, C. (2005). Angiocentric neuroepithelial tumor (ANET): a new epilepsy- related clinicopathological entity with distinctive MRI. Brain Pathol, 15(4), 281-286. Li, H., & Durbin, R. (2009). Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics (Oxford, England), 25(14), 1754-1760. doi:10.1093/bioinformatics/btp324 Li, Z. Z., Kondo, T., Murata, T., Ebersole, T. A., Nishi, T., Tada, K., . . . Abe, K. (2002). Expression of Hqk encoding a KH RNA binding protein is altered in human glioma. Jpn J Cancer Res, 93(2), 167-177. Long, G. V., Menzies, A. M., Nagrial, A. M., Haydu, L. E., Hamilton, A. L., Mann, G. J., . . . Kefford, R. F. (2011). Prognostic and clinicopathologic associations of oncogenic BRAF in metastatic melanoma. J Clin Oncol, 29(10), 1239-1246. doi:10.1200/jco.2010.32.4327 Louis, D. N., Ohgaki, H., Wiestler, O. D., Cavenee, W. K., Burger, P. C., Jouvet, A., . . . Kleihues, P. (2007). The 2007 WHO classification of tumours of the central nervous system. Acta Neuropathol, 114(2), 97-109. doi:10.1007/s00401-007-0243-4 Louis, D. N., Perry, A., Burger, P., Ellison, D. W., Reifenberger, G., von Deimling, A., . . . Wesseling, P. (2014). International Society Of Neuropathology--Haarlem consensus guidelines for nervous system tumor classification and grading. Brain Pathol, 24(5), 429- 435. doi:10.1111/bpa.12171 Lu, W., Feng, F., Xu, J., Lu, X., Wang, S., Wang, L., . . . Lei, X. (2014). QKI impairs self-renewal and tumorigenicity of oral cancer cells via repression of SOX2. Cancer biology & therapy, 15(9), 1174-1184. doi:10.4161/cbt.29502 Lynch, K. W., & Maniatis, T. (1996). Assembly of specific SR protein complexes on distinct regulatory elements of the Drosophila doublesex splicing enhancer. Genes Dev, 10(16), 2089-2101. Malaterre, J., Mantamadiotis, T., Dworkin, S., Lightowler, S., Yang, Q., Ransome, M. I., . . . Ramsay, R. G. (2008). c-Myb is required for neural progenitor cell proliferation and maintenance of the neural stem cell niche in adult brain. Stem cells (Dayton, Ohio), 26(1), 173-181. doi:10.1634/stemcells.2007-0293 Mansour, M. R., Abraham, B. J., Anders, L., Berezovskaya, A., Gutierrez, A., Durbin, A. D., . . . Look, A. T. (2014). An oncogenic super-enhancer formed through somatic mutation of a noncoding intergenic element. Science. doi:10.1126/science.1259037 Mason, C. S., Springer, C. J., Cooper, R. G., Superti-Furga, G., Marshall, C. J., & Marais, R. (1999). Serine and tyrosine phosphorylations cooperate in Raf-1, but not B-Raf activation. EMBO J, 18(8), 2137-2148. doi:10.1093/emboj/18.8.2137 Massimino, M., Spreafico, F., Riva, D., Biassoni, V., Poggi, G., Solero, C., . . . Giangaspero, F. (2010). A lower-dose, lower-toxicity cisplatin-etoposide regimen for childhood

117

progressive low-grade glioma. J Neurooncol, 100(1), 65-71. doi:10.1007/s11060-010- 0136-6 McKenna, A., Hanna, M., Banks, E., Sivachenko, A., Cibulskis, K., Kernytsky, A., . . . DePristo, M. A. (2010). The Genome Analysis Toolkit: a MapReduce framework for analyzing next- generation DNA sequencing data. Genome Research, 20(9), 1297-1303. doi:10.1101/gr.107524.110 Menzel, S., Jiang, J., Silver, N., Gallagher, J., Cunningham, J., Surdulescu, G., . . . Thein, S. L. (2007). The HBS1L-MYB intergenic region on chromosome 6q23.3 influences erythrocyte, platelet, and monocyte counts in humans. Blood, 110(10), 3624-3626. doi:10.1182/blood-2007-05-093419 Merchant, T. E., Conklin, H. M., Wu, S., Lustig, R. H., & Xiong, X. (2009). Late effects of conformal radiation therapy for pediatric patients with low-grade glioma: prospective evaluation of cognitive, endocrine, and hearing deficits. J Clin Oncol, 27(22), 3691-3697. doi:10.1200/jco.2008.21.2738 Mertens, F., Johansson, B., Fioretos, T., & Mitelman, F. (2015). The emerging complexity of gene fusions in cancer. Nat Rev Cancer, 15(6), 371-381. doi:10.1038/nrc3947 Mitelman, F., Johansson, B., & Mertens, F. (2004). Fusion genes and rearranged genes as a linear function of chromosome aberrations in cancer. Nat Genet, 36(4), 331-334. doi:10.1038/ng1335 Moodie, S. A., Willumsen, B. M., Weber, M. J., & Wolfman, A. (1993). Complexes of Ras.GTP with Raf-1 and mitogen-activated protein kinase kinase. Science, 260(5114), 1658-1661. Nassar, N., Horn, G., Herrmann, C., Scherer, A., McCormick, F., & Wittinghofer, A. (1995). The 2.2 A crystal structure of the Ras-binding domain of the serine/threonine kinase c-Raf1 in complex with Rap1A and a GTP analogue. Nature, 375(6532), 554-560. doi:10.1038/375554a0 Ness, S. A., Marknell, A., & Graf, T. (1989). The v-myb oncogene product binds to and activates the promyelocyte-specific mim-1 gene. Cell, 59(6), 1115-1125. Northcott, P. A., Lee, C., Zichner, T., Stütz, A. M., Erkek, S., Kawauchi, D., . . . Pfister, S. M. (2014). Enhancer hijacking activates GFI1 family oncogenes in medulloblastoma. Nature, 511(7510), 428-434. doi:10.1038/nature13379 Olow, A., Mueller, S., Yang, X., Hashizume, R., Meyerowitz, J., Weiss, W. A., . . . Haas-Kogan, D. A. (2016). BRAF status in personalizing treatment approaches for pediatric gliomas. Clin Cancer Res. doi:10.1158/1078-0432.ccr-15-1101 Ostrom, Q. T., de Blank, P. M., Kruchko, C., Petersen, C. M., Liao, P., Finlay, J. L., . . . Barnholtz- Sloan, J. S. (2015). Alex's Lemonade Stand Foundation Infant and Childhood Primary Brain and Central Nervous System Tumors Diagnosed in the United States in 2007-2011. Neuro Oncol, 16 Suppl 10, x1-x36. doi:10.1093/neuonc/nou327 Ostrom, Q. T., Gittleman, H., Fulop, J., Liu, M., Blanda, R., Kromer, C., . . . Barnholtz-Sloan, J. S. (2015). CBTRUS Statistical Report: Primary Brain and Central Nervous System Tumors Diagnosed in the United States in 2008-2012. Neuro Oncol, 17 Suppl 4, iv1-iv62. doi:10.1093/neuonc/nov189 Palanisamy, N., Ateeq, B., Kalyana-Sundaram, S., Pflueger, D., Ramnarayanan, K., Shankar, S., . . . Chinnaiyan, A. M. (2010). Rearrangements of the RAF kinase pathway in prostate cancer, gastric cancer and melanoma. Nat Med, 16(7), 793-798. doi:10.1038/nm.2166 Patel, S., Bhatnagar, A., Wear, C., Osiro, S., Gabriel, A., Kimball, D., . . . Loukas, M. (2014). Are pediatric brain tumors on the rise in the USA? Significant incidence and survival findings 118

from the SEER database analysis. Childs Nerv Syst, 30(1), 147-154. doi:10.1007/s00381- 013-2307-1 Pelicci, P. G., Lanfrancone, L., Brathwaite, M. D., Wolman, S. R., & Dalla-Favera, R. (1984). Amplification of the c-myb oncogene in a case of human acute myelogenous leukemia. Science, 224(4653), 1117-1121. Peng, S. B., Henry, J. R., Kaufman, M. D., Lu, W. P., Smith, B. D., Vogeti, S., . . . Flynn, D. L. (2015). Inhibition of RAF Isoforms and Active Dimers by LY3009120 Leads to Anti-tumor Activities in RAS or BRAF Mutant Cancers. Cancer Cell, 28(3), 384-398. doi:10.1016/j.ccell.2015.08.002 Persson, M., Andren, Y., Mark, J., Horlings, H. M., Persson, F., & Stenman, G. (2009). Recurrent fusion of MYB and NFIB transcription factor genes in carcinomas of the breast and head and neck. Proc Natl Acad Sci U S A, 106(44), 18740-18744. doi:10.1073/pnas.0909114106 Persson, M., Andrén, Y., Mark, J., Horlings, H. M., Persson, F., & Stenman, G. (2009). Recurrent fusion of MYB and NFIB transcription factor genes in carcinomas of the breast and head and neck. PNAS, 106(44), 18740-18744. doi:10.1073/pnas.0909114106 Pfister, S., Janzarik, W. G., Remke, M., Ernst, A., Werft, W., Becker, N., . . . Lichter, P. (2008). BRAF gene duplication constitutes a mechanism of MAPK pathway activation in low- grade astrocytomas. The Journal of clinical investigation, 118(5), 1739-1749. doi:10.1172/JCI33656 Pollack, I. F. (1999). Pediatric brain tumors. Semin Surg Oncol, 16(2), 73-90. Poulikakos, P. I., Persaud, Y., Janakiraman, M., Kong, X., Ng, C., Moriceau, G., . . . Solit, D. B. (2011). RAF inhibitor resistance is mediated by dimerization of aberrantly spliced BRAF(V600E). Nature, 480(7377), 387-390. doi:10.1038/nature10662 Press, R. D., Reddy, E. P., & Ewert, D. L. (1994). Overexpression of C-terminally but not N- terminally truncated Myb induces fibrosarcomas: a novel nonhematopoietic target cell for the myb oncogene. Mol Cell Biol, 14(4), 2278-2290. Qaddoumi, I., Sultan, I., & Gajjar, A. (2009). Outcome and prognostic features in pediatric gliomas: a review of 6212 cases from the Surveillance, Epidemiology, and End Results database. Cancer, 115(24), 5761-5770. doi:10.1002/cncr.24663 Quintana, A. M., Liu, F., O'Rourke, J. P., & Ness, S. A. (2011). Identification and regulation of c-Myb target genes in MCF-7 cells. BMC cancer, 11, 30. doi:10.1186/1471-2407-11-30 Rajakulendran, T., Sahmi, M., Lefrancois, M., Sicheri, F., & Therrien, M. (2009). A dimerization- dependent mechanism drives RAF catalytic activation. Nature, 461(7263), 542-545. doi:10.1038/nature08314 Ramkissoon, L. A., Horowitz, P. M., Craig, J. M., Ramkissoon, S. H., Rich, B. E., Schumacher, S. E., . . . Ligon, K. L. (2013). Genomic analysis of diffuse pediatric low-grade gliomas identifies recurrent oncogenic truncating rearrangements in the transcription factor MYBL1. Proc Natl Acad Sci U S A, 110(20), 8188-8193. doi:10.1073/pnas.1300252110 Rapp, U. R., Goldsborough, M. D., Mark, G. E., Bonner, T. I., Groffen, J., Reynolds, F. H., Jr., & Stephenson, J. R. (1983). Structure and biological activity of v-raf, a unique oncogene transduced by a retrovirus. Proc Natl Acad Sci U S A, 80(14), 4218-4222. Rapp, U. R., & Todaro, C. (1978). Generation of new mouse sarcoma viruses in cell culture. Science, 201(4358), 821-824. Reynolds, B. A., & Weiss, S. (1992). Generation of neurons and astrocytes from isolated cells of the adult mammalian central nervous system. Science, 255(5052), 1707-1710. 119

Robinson, J. T., Thorvaldsdóttir, H., Winckler, W., Guttman, M., Lander, E. S., Getz, G., & Mesirov, J. P. (2011). Integrative genomics viewer. Nature Biotechnology, 29(1), 24-26. doi:10.1038/nbt.1754 Rothrock, C. R., House, A. E., & Lynch, K. W. (2005). HnRNP L represses exon splicing via a regulated exonic splicing silencer. EMBO J, 24(15), 2792-2802. doi:10.1038/sj.emboj.7600745 Sakura, H., Kanei-Ishii, C., Nagase, T., Nakagoshi, H., Gonda, T. J., & Ishii, S. (1989). Delineation of three functional domains of the transcriptional activator encoded by the c-myb protooncogene. PNAS, 86(15), 5758-5762. Schwartzentruber, J., Korshunov, A., Liu, X. Y., Jones, D. T., Pfaff, E., Jacob, K., . . . Jabado, N. (2012). Driver mutations in histone H3.3 and chromatin remodelling genes in paediatric glioblastoma. Nature, 482(7384), 226-231. doi:10.1038/nature10833 Shepherd, C. W., Scheithauer, B. W., Gomez, M. R., Altermatt, H. J., & Katzmann, J. A. (1991). Subependymal giant cell astrocytoma: a clinical, pathological, and flow cytometric study. Neurosurgery, 28(6), 864-868. Sicurella, C., Freeman, R., Micallef, S., Mucenski, M. L., Bertoncello, I., & Ramsay, R. G. (2001). Defective stem cell factor expression in c-myb null fetal liver stroma. Blood Cells Mol Dis, 27(2), 470-478. doi:10.1006/bcmd.2001.0407 Sievert, A. J., & Fisher, M. J. (2009). Pediatric low-grade gliomas. J Child Neurol, 24(11), 1397- 1408. doi:10.1177/0883073809342005 Sievert, A. J., Jackson, E. M., Gai, X., Hakonarson, H., Judkins, A. R., Resnick, A. C., . . . Biegel, J. A. (2009). Duplication of 7q34 in pediatric low-grade astrocytomas detected by high- density single-nucleotide polymorphism-based genotype arrays results in a novel BRAF fusion gene. Brain Pathol, 19(3), 449-458. doi:10.1111/j.1750-3639.2008.00225.x Sievert, A. J., Lang, S.-S., Boucher, K. L., Madsen, P. J., Slaunwhite, E., Choudhari, N., . . . Resnick, A. C. (2013). Paradoxical activation and RAF inhibitor resistance of BRAF protein kinase fusions characterizing pediatric astrocytomas. PNAS, 110(15), 5957-5962. doi:10.1073/pnas.1219232110 Sievert, A. J., Lang, S. S., Boucher, K. L., Madsen, P. J., Slaunwhite, E., Choudhari, N., . . . Resnick, A. C. (2013). Paradoxical activation and RAF inhibitor resistance of BRAF protein kinase fusions characterizing pediatric astrocytomas. Proc Natl Acad Sci U S A, 110(15), 5957- 5962. doi:10.1073/pnas.1219232110 Smith, M. A., Altekruse, S. F., Adamson, P. C., Reaman, G. H., & Seibel, N. L. (2014). Declining childhood and adolescent cancer mortality. Cancer, 120(16), 2497-2506. doi:10.1002/cncr.28748 Sosman, J. A., Kim, K. B., Schuchter, L., Gonzalez, R., Pavlick, A. C., Weber, J. S., . . . Ribas, A. (2012). Survival in BRAF V600-mutant advanced melanoma treated with vemurafenib. N Engl J Med, 366(8), 707-714. doi:10.1056/NEJMoa1112302 Stanton, V. P., Jr., Nichols, D. W., Laudano, A. P., & Cooper, G. M. (1989). Definition of the human raf amino-terminal regulatory region by deletion mutagenesis. Mol Cell Biol, 9(2), 639-647. Stephens, P. J., Davies, H. R., Mitani, Y., Van Loo, P., Shlien, A., Tarpey, P. S., . . . Futreal, P. A. (2013). Whole exome sequencing of adenoid cystic carcinoma. The Journal of clinical investigation, 123(7), 2965-2968. doi:10.1172/JCI67201 Stokland, T., Liu, J. F., Ironside, J. W., Ellison, D. W., Taylor, R., Robinson, K. J., . . . Walker, D. A. (2010). A multivariate analysis of factors determining tumor progression in childhood 120

low-grade glioma: a population-based cohort study (CCLG CNS9702). Neuro Oncol, 12(12), 1257-1268. doi:10.1093/neuonc/noq092 Tang, Z., Yuan, X., Du, R., Cheung, S. H., Zhang, G., Wei, J., . . . Luo, L. (2015). BGB-283, a Novel RAF Kinase and EGFR Inhibitor, Displays Potent Antitumor Activity in BRAF-Mutated Colorectal Cancers. Mol Cancer Ther, 14(10), 2187-2197. doi:10.1158/1535-7163.mct- 15-0262 Taphoorn, M. J., Schiphorst, A. K., Snoek, F. J., Lindeboom, J., Wolbers, J. G., Karim, A. B., . . . Heimans, J. J. (1994). Cognitive functions and quality of life in patients with low-grade gliomas: the impact of radiotherapy. Ann Neurol, 36(1), 48-54. doi:10.1002/ana.410360111 Torres-García, W., Zheng, S., Sivachenko, A., Vegesna, R., Wang, Q., Yao, R., . . . Verhaak, R. G. W. (2014). PRADA: pipeline for RNA sequencing data analysis. Bioinformatics (Oxford, England), 30(15), 2224-2226. doi:10.1093/bioinformatics/btu169 Vernet, C., & Artzt, K. (1997). STAR, a gene family involved in signal transduction and activation of RNA. Trends Genet, 13(12), 479-484. Wang, M., Tihan, T., Rojiani, A. M., Bodhireddy, S. R., Prayson, R. A., Iacuone, J. J., . . . Burger, P. C. (2005). Monomorphous angiocentric glioma: a distinctive epileptogenic neoplasm with features of infiltrating astrocytoma and ependymoma. J Neuropathol Exp Neurol, 64(10), 875-881. Wang, Y., Vogel, G., Yu, Z., & Richard, S. (2013). The QKI-5 and QKI-6 RNA binding proteins regulate the expression of microRNA 7 in glial cells. Mol Cell Biol, 33(6), 1233-1243. doi:10.1128/MCB.01604-12 Wong, K., Ren, X. R., Huang, Y. Z., Xie, Y., Liu, G., Saito, H., . . . Rao, Y. (2001). Signal transduction in neuronal migration: roles of GTPase activating proteins and the small GTPase Cdc42 in the Slit-Robo pathway. Cell, 107(2), 209-221. Wu, G., Broniscer, A., McEachron, T. A., Lu, C., Paugh, B. S., Becksfort, J., . . . Baker, S. J. (2012). Somatic histone H3 alterations in pediatric diffuse intrinsic pontine gliomas and non- brainstem glioblastomas. Nat Genet, 44(3), 251-253. doi:10.1038/ng.1102 Wu, J., Zhou, L., Tonissen, K., Tee, R., & Artzt, K. (1999). The quaking I-5 protein (QKI-5) has a novel nuclear localization signal and shuttles between the nucleus and the cytoplasm. The Journal of biological chemistry, 274(41), 29202-29210. Yamaguchi, T., Kakefuda, R., Tajima, N., Sowa, Y., & Sakai, T. (2011). Antitumor activities of JTP- 74057 (GSK1120212), a novel MEK1/2 inhibitor, on colorectal cancer cell lines in vitro and in vivo. Int J Oncol, 39(1), 23-31. doi:10.3892/ijo.2011.1015 Yang, X. J. (2005). Multisite protein modification and intramolecular signaling. Oncogene, 24(10), 1653-1662. doi:10.1038/sj.onc.1208173 Yao, Z., Torres, N. M., Tao, A., Gao, Y., Luo, L., Li, Q., . . . Rosen, N. (2015). BRAF Mutants Evade ERK-Dependent Feedback by Different Mechanisms that Determine Their Sensitivity to Pharmacologic Inhibition. Cancer Cell, 28(3), 370-383. doi:10.1016/j.ccell.2015.08.001 Yin, D., Ogawa, S., Kawamata, N., Tunici, P., Finocchiaro, G., Eoli, M., . . . Koeffler, H. P. (2009). High-resolution genomic copy number profiling of glioblastoma multiforme by single nucleotide polymorphism DNA microarray. Mol Cancer Res, 7(5), 665-677. doi:10.1158/1541-7786.mcr-08-0270 Zack, T. I., Schumacher, S. E., Carter, S. L., Cherniack, A. D., Saksena, G., Tabak, B., . . . Beroukhim, R. (2013). Pan-cancer patterns of somatic copy number alteration. Nat Genet, 45(10), 1134-1140. doi:10.1038/ng.2760 121

Zhang, C., Spevak, W., Zhang, Y., Burton, E. A., Ma, Y., Habets, G., . . . Bollag, G. (2015). RAF inhibitors that evade paradoxical MAPK pathway activation. Nature, 526(7574), 583- 586. doi:10.1038/nature14982 Zhang, J., Wu, G., Miller, C. P., Tatevossian, R. G., Dalton, J. D., Tang, B., . . . St. Jude Children's Research Hospital-Washington University Pediatric Cancer Genome, P. (2013). Whole- genome sequencing identifies genetic alterations in pediatric low-grade gliomas. Nat Genet, 45(6), 602-612. doi:10.1038/ng.2611 Zhao, L., Glazov, E. A., Pattabiraman, D. R., Al-Owaidi, F., Zhang, P., Brown, M. A., . . . Gonda, T. J. (2011). Integrated genome-wide chromatin occupancy and expression analyses identify key myeloid pro-differentiation transcription factors repressed by Myb. Nucleic Acids Research, 39(11), 4664-4679. doi:10.1093/nar/gkr024 Zhao, Y., Zhang, G., Wei, M., Lu, X., Fu, H., Feng, F., . . . Yuan, J. (2014). The tumor suppressing effects of QKI-5 in prostate cancer: a novel diagnostic and prognostic protein. Cancer Biol Ther, 15(1), 108-118. doi:10.4161/cbt.26722 Zhou, Y., & Ness, S. A. (2011). Myb proteins: angels and demons in normal and transformed cells. Front Biosci (Landmark Ed), 16, 1109-1131. Zhu, J., Adli, M., Zou, J. Y., Verstappen, G., Coyne, M., Zhang, X., . . . Bernstein, B. E. (2013). Genome-wide chromatin state transitions associated with developmental and environmental cues. Cell, 152(3), 642-654. doi:10.1016/j.cell.2012.12.033 Zong, F. Y., Fu, X., Wei, W. J., Luo, Y. G., Heiner, M., Cao, L. J., . . . Hui, J. (2014). The RNA-binding protein QKI suppresses cancer-associated aberrant splicing. PLoS Genet, 10(4), e1004289. doi:10.1371/journal.pgen.1004289

122